MPC8540

MPC8540 PowerQUICC III™ Integrated Host Processor Reference Manual MPC8540RM Rev. 1 07/2004 How to Reach Us: USA/Europe/Locations Not Listed: Freescale Semiconductor Literature Distribution Center P.O. Box 5405, Denver, Colorado 80217 1-480-768-2130 (800) 521-6274 Japan: Freescale Semiconductor Japan Ltd. Technical Information Center 3-20-1, Minami-Azabu, Minato-ku Tokyo 106-8573, Japan 81-3-3440-3569 Asia/Pacific: Freescale Semiconductor Hong Kong Ltd. 2 Dai King Street Tai Po Industrial Estate Tai Po, N.T. Hong Kong 852-26668334 Home Page: www.freescale.com Information in this document is provided solely to enable system and software implementers to use Freescale Semiconductor products. There are no express or implied copyright licenses granted hereunder to design or fabricate any integrated circuits or integrated circuits based on the information in this document. Freescale Semiconductor reserves the right to make changes without further notice to any products herein. Freescale Semiconductor makes no warranty, representation or guarantee regarding the suitability of its products for any particular purpose, nor does Freescale Semiconductor assume any liability arising out of the application or use of any product or circuit, and specifically disclaims any and all liability, including without limitation consequential or incidental damages. “Typical” parameters which may be provided in Freescale Semiconductor data sheets and/or specifications can and do vary in different applications and actual performance may vary over time. All operating parameters, including “Typicals” must be validated for each customer application by customer’s technical experts. Freescale Semiconductor does not convey any license under its patent rights nor the rights of others. Freescale Semiconductor products are not designed, intended, or authorized for use as components in systems intended for surgical implant into the body, or other applications intended to support or sustain life, or for any other application in which the failure of the Freescale Semiconductor product could create a situation where personal injury or death may occur. Should Buyer purchase or use Freescale Semiconductor products for any such unintended or unauthorized application, Buyer shall indemnify and hold Freescale Semiconductor and its officers, employees, subsidiaries, affiliates, and distributors harmless against all claims, costs, damages, and expenses, and reasonable attorney fees arising out of, directly or indirectly, any claim of personal injury or death associated with such unintended or unauthorized use, even if such claim alleges that Freescale Semiconductor was negligent regarding the design or manufacture of the part. Learn More: For more information about Freescale Semiconductor products, please visit www.freescale.com Freescale™ and the Freescale logo are trademarks of Freescale Semiconductor, Inc. The described product is a PowerPC microprocessor. The PowerPC name is a trademark of IBM Corp. and used under license. All other product or service names are the property of their respective owners. © Freescale Semiconductor, Inc. 2004. MPC8540RM Rev. 1 07/2004 Part I—Overview Overview Memory Map Signal Descriptions Reset, Clocking, and Initialization Part II—e500 Core Complex and L2 Cache e500 Core Complex Overview e500 Register Summary L2 Look-Aside Cache/SRAM Part III—Memory and I/O Interfaces e500 Coherency Module DDR Memory Controller Programmable Interrupt Controller I2C Interface DUART Local Bus Controller Three-Speed Ethernet Controllers DMA Controller PCI/PCI-X Bus Interface RapidIO Interface Part IV—Global Functions and Debug Global Utilities Performance Monitor Debug Features and Watchpoint Facility 10/100 Fast Ethernet Controller Appendix A—Revision History Glossary of Terms and Abbreviations Register Index (Memory-Mapped Registers) General Index I 1 2 3 4 II 5 6 7 III 8 9 10 11 12 13 14 15 16 17 IV 18 19 20 21 A GLO REG IND I 1 2 3 4 II 5 6 7 III 8 9 10 11 12 13 14 15 16 17 IV 18 19 20 21 A GLO REG IND Part I—Overview Overview Memory Map Signal Descriptions Reset, Clocking, and Initialization Part II—e500 Core Complex and L2 Cache e500 Core Complex Overview e500 Register Summary L2 Look-Aside Cache/SRAM Part III—Memory and I/O Interfaces e500 Coherency Module DDR Memory Controller Programmable Interrupt Controller I2C Interface DUART Local Bus Controller Three-Speed Ethernet Controllers DMA Controller PCI/PCI-X Bus Interface RapidIO Interface Part IV—Global Functions and Debug Global Utilities Performance Monitor Debug Features and Watchpoint Facility 10/100 Fast Ethernet Controller Appendix A—Revision History Glossary of Terms and Abbreviations Register Index (Memory-Mapped Registers) General Index Contents Paragraph Number Title Contents Page Number About This Book Audience ..................................................................................................................... lxxxiii Organization................................................................................................................ lxxxiii Suggested Reading...................................................................................................... lxxxvi General Information............................................................................................ lxxxvi Related Documentation ...................................................................................... lxxxvi Conventions ............................................................................................................... lxxxvii Signal Conventions ................................................................................................... lxxxviii Acronyms and Abbreviations ................................................................................... lxxxviii Part I Overview Chapter 1 Overview 1.1 1.2 1.2.1 1.3 1.3.1 1.3.2 1.3.2.1 1.3.2.2 1.3.3 1.3.4 1.3.5 1.3.6 1.3.7 1.3.8 1.3.9 1.3.10 1.3.11 1.3.12 1.3.13 1.3.14 Introduction...................................................................................................................... 1-1 MPC8540 Overview ........................................................................................................ 1-1 Key Features ................................................................................................................ 1-2 MPC8540 Architecture Overview ................................................................................... 1-8 e500 Core Overview .................................................................................................... 1-8 On-Chip Memory Unit............................................................................................... 1-12 On-Chip Memory as Memory-Mapped SRAM..................................................... 1-13 On-Chip Memory as L2 Cache.............................................................................. 1-13 e500 Coherency Module (ECM)................................................................................ 1-14 DDR SDRAM Controller .......................................................................................... 1-14 Programmable Interrupt Controller (PIC).................................................................. 1-15 I2C Controller ............................................................................................................ 1-15 Boot Sequencer .......................................................................................................... 1-16 Dual Universal Asynchronous Receiver/Transmitter (DUART) ............................... 1-16 10/100 Fast Ethernet Controller................................................................................. 1-16 Local Bus Controller (LBC) ...................................................................................... 1-17 Three-Speed Ethernet Controllers (10/100/1Gb)....................................................... 1-17 Integrated DMA......................................................................................................... 1-18 PCI Controller............................................................................................................ 1-18 RapidIO Controller .................................................................................................... 1-18 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor v Contents Paragraph Number 1.3.14.1 1.3.15 1.3.16 1.3.17 1.3.18 1.3.19 1.3.20 1.4 1.4.1 Title Page Number RapidIO Message Unit .......................................................................................... 1-19 Power Management ................................................................................................... 1-19 Clocking..................................................................................................................... 1-20 Address Map .............................................................................................................. 1-20 OCeaN Switch Fabric ................................................................................................ 1-20 Processing Across the On-Chip Fabric...................................................................... 1-20 Data Processing with the e500 Coherency Module ................................................... 1-21 MPC8540 Application Examples .................................................................................. 1-21 MPC8540 Applications.............................................................................................. 1-21 Chapter 2 Memory Map 2.1 2.2 2.2.1 2.2.2 2.2.3 2.2.3.1 2.2.3.2 2.2.3.3 2.2.3.4 2.2.3.5 2.2.3.6 2.2.3.7 2.2.4 2.2.5 2.2.5.1 2.2.5.2 2.2.5.3 2.3 2.3.1 2.3.2 2.3.3 2.3.4 2.3.5 Local Memory Map Overview and Example .................................................................. 2-1 Address Translation and Mapping ................................................................................... 2-3 SRAM Windows.......................................................................................................... 2-4 Window into Configuration Space............................................................................... 2-4 Local Access Windows................................................................................................ 2-4 Local Access Register Memory Map ...................................................................... 2-5 Local Access Window n Base Address Registers (LAWBAR0–LAWBAR7)................................................................................... 2-6 Local Access Window n Attributes Registers (LAWAR0–LAWAR7)......................................................................................... 2-6 Precedence of Local Access Windows .................................................................... 2-7 Configuring Local Access Windows ....................................................................... 2-7 Distinguishing Local Access Windows from Other Mapping Functions ................ 2-8 Illegal Interaction Between Local Access Windows and DDR SDRAM Chip Selects.......................................................................................... 2-8 Outbound Address Translation and Mapping Windows.............................................. 2-8 Inbound Address Translation and Mapping Windows ................................................ 2-9 RapidIO Inbound ATMU......................................................................................... 2-9 PCI/PCI-X Inbound ATMU..................................................................................... 2-9 Illegal Interaction Between Inbound ATMUs and Local Access Windows .............................................................................................................. 2-9 Configuration, Control, and Status Register Map.......................................................... 2-10 Accessing CCSR Memory from External Masters .................................................... 2-11 Organization of CCSR Memory ................................................................................ 2-11 General Utilities Registers ......................................................................................... 2-13 Interrupt Controller and CCSR .................................................................................. 2-14 RapidIO and CCSR.................................................................................................... 2-15 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 vi Freescale Semiconductor Contents Paragraph Number 2.3.6 2.4 Title Page Number Device-Specific Utilities............................................................................................ 2-15 Complete CCSR Map .................................................................................................... 2-16 Chapter 3 Signal Descriptions 3.1 3.2 3.3 Signals Overview ............................................................................................................. 3-1 Configuration Signals Sampled at Reset ....................................................................... 3-15 Output Signal States During Reset ................................................................................ 3-16 Chapter 4 Reset, Clocking, and Initialization 4.1 4.2 4.2.1 4.2.2 4.3 4.3.1 4.3.1.1 4.3.1.1.1 4.3.1.1.2 4.3.1.2 4.3.1.2.1 4.3.1.2.2 4.3.1.3 4.3.1.3.1 4.3.2 4.4 4.4.1 4.4.1.1 4.4.1.2 4.4.2 4.4.3 4.4.3.1 4.4.3.2 4.4.3.3 4.4.3.4 4.4.3.5 Overview.......................................................................................................................... 4-1 External Signal Descriptions ........................................................................................... 4-1 System Control Signals................................................................................................ 4-2 Clock Signals ............................................................................................................... 4-3 Memory Map/Register Definition ................................................................................... 4-3 Local Configuration Control........................................................................................ 4-3 Accessing Configuration, Control, and Status Registers......................................... 4-4 Updating CCSRBAR ........................................................................................... 4-4 Configuration, Control, and Status Base Address Register (CCSRBAR) .................................................................................................... 4-5 Accessing Alternate Configuration Space ............................................................... 4-5 Alternate Configuration Base Address Register (ALTCBAR)............................ 4-6 Alternate Configuration Attribute Register (ALTCAR)...................................... 4-6 Boot Page Translation.............................................................................................. 4-7 Boot Page Translation Register (BPTR).............................................................. 4-8 Boot Sequencer ............................................................................................................ 4-8 Functional Description..................................................................................................... 4-8 Reset Operations .......................................................................................................... 4-8 Soft Reset................................................................................................................. 4-9 Hard Reset ............................................................................................................... 4-9 Power-On Reset Sequence........................................................................................... 4-9 Power-On Reset Configuration.................................................................................. 4-11 System PLL Ratio.................................................................................................. 4-12 e500 Core PLL Ratio ............................................................................................. 4-13 Boot ROM Location .............................................................................................. 4-14 Host/Agent Configuration ..................................................................................... 4-14 CPU Boot Configuration ....................................................................................... 4-15 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor vii Contents Paragraph Number 4.4.3.6 4.4.3.7 4.4.3.8 4.4.3.9 4.4.3.10 4.4.3.11 4.4.3.12 4.4.3.13 4.4.3.14 4.4.3.15 4.4.3.16 4.4.3.17 4.4.3.18 4.4.3.19 4.4.3.20 4.4.3.21 4.4.4 4.4.4.1 4.4.4.2 4.4.4.3 4.4.4.4 Title Page Number Boot Sequencer Configuration .............................................................................. 4-16 TSEC Width........................................................................................................... 4-16 TSEC1 Protocol ..................................................................................................... 4-17 TSEC2 Protocol ..................................................................................................... 4-17 RapidIO Transmit Clock Source............................................................................ 4-18 RapidIO Device ID ................................................................................................ 4-18 PCI Width Configuration....................................................................................... 4-19 PCI I/O Impedance ................................................................................................ 4-19 PCI Arbiter Configuration ..................................................................................... 4-19 PCI Debug Configuration ...................................................................................... 4-20 PCI-X Configuration ............................................................................................. 4-20 Memory Debug Configuration .............................................................................. 4-20 DDR Debug Configuration.................................................................................... 4-21 PCI/PCI-X Output Hold Configuration ................................................................. 4-21 Local Bus Output Hold Configuration .................................................................. 4-22 General-Purpose POR Configuration .................................................................... 4-23 Clocking..................................................................................................................... 4-23 System Clock/PCI Clock ....................................................................................... 4-23 RapidIO Clocks ..................................................................................................... 4-24 Ethernet Clocks...................................................................................................... 4-25 Real Time Clock .................................................................................................... 4-25 Part II e500 Core Complex and L2 Cache Chapter 5 Core Complex Overview 5.1 5.1.1 5.1.2 5.2 5.3 5.4 5.5 5.5.1 5.5.2 5.5.3 5.6 5.7 Overview.......................................................................................................................... 5-1 Upward Compatibility ................................................................................................. 5-3 Core Complex Summary ............................................................................................. 5-3 e500 Processor and System Version Numbers................................................................. 5-4 Features ............................................................................................................................ 5-5 Instruction Set ................................................................................................................ 5-12 Instruction Flow ............................................................................................................. 5-14 Initial Instruction Fetch.............................................................................................. 5-14 Branch Detection and Prediction ............................................................................... 5-14 e500 Execution Pipeline ............................................................................................ 5-15 Programming Model ...................................................................................................... 5-18 On-Chip Cache Implementation .................................................................................... 5-20 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 viii Freescale Semiconductor Contents Paragraph Number 5.8 5.8.1 5.8.2 5.8.3 5.8.4 5.8.5 5.9 5.9.1 5.9.2 5.9.3 5.9.4 5.10 5.10.1 5.10.2 5.10.3 5.10.4 5.11 5.12 5.12.1 5.12.2 5.12.3 5.13 5.13.1 5.13.1.1 5.13.1.2 5.13.2 5.13.3 5.13.4 5.13.5 5.13.6 5.14 Title Page Number Interrupts and Exception Handling ................................................................................ 5-20 Exception Handling ................................................................................................... 5-20 Interrupt Classes ........................................................................................................ 5-21 Interrupt Types ........................................................................................................... 5-21 Upper Bound on Interrupt Latencies ......................................................................... 5-22 Interrupt Registers...................................................................................................... 5-22 Memory Management.................................................................................................... 5-24 Address Translation ................................................................................................... 5-26 MMU Assist Registers (MAS1–MAS4 and MAS6) ................................................. 5-27 Process ID Registers (PID0–PID2)............................................................................ 5-27 TLB Coherency.......................................................................................................... 5-27 Memory Coherency ....................................................................................................... 5-28 Atomic Update Memory References ......................................................................... 5-28 Memory Access Ordering.......................................................................................... 5-28 Cache Control Instructions ........................................................................................ 5-28 Programmable Page Characteristics .......................................................................... 5-29 Core Complex Bus (CCB) ............................................................................................. 5-29 Performance Monitoring................................................................................................ 5-29 Global Control Register ............................................................................................. 5-30 Performance Monitor Counter Registers ................................................................... 5-30 Local Control Registers ............................................................................................. 5-30 Legacy Support of PowerPC Architecture..................................................................... 5-31 Instruction Set Compatibility..................................................................................... 5-31 User Instruction Set ............................................................................................... 5-31 Supervisor Instruction Set...................................................................................... 5-31 Memory Subsystem ................................................................................................... 5-32 Exception Handling ................................................................................................... 5-32 Memory Management................................................................................................ 5-32 Reset........................................................................................................................... 5-32 Little-Endian Mode.................................................................................................... 5-33 PowerQUICC III Implementation Details ..................................................................... 5-33 Chapter 6 Core Register Summary 6.1 6.1.1 6.2 6.2.1 6.3 Overview.......................................................................................................................... 6-1 Register Set .................................................................................................................. 6-1 Register Model for 32-Bit Implementations .................................................................... 6-3 Special-Purpose Registers (SPRs) ............................................................................... 6-4 Registers for Computational Operations.......................................................................... 6-8 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor ix Contents Paragraph Number 6.3.1 6.3.2 6.4 6.4.1 6.4.2 6.4.3 6.5 6.5.1 6.5.2 6.5.3 6.5.4 6.6 6.6.1 6.6.2 6.6.3 6.6.4 6.6.5 6.7 6.7.1 6.7.1.1 6.7.1.2 6.7.1.3 6.7.1.4 6.7.1.5 6.7.1.6 6.7.1.7 6.7.1.8 6.7.2 6.7.2.1 6.7.2.2 6.7.2.3 6.7.2.4 6.8 6.9 6.9.1 6.9.2 6.9.3 6.10 6.10.1 6.10.2 Title Page Number General-Purpose Registers (GPRs).............................................................................. 6-8 Integer Exception Register (XER)............................................................................... 6-8 Registers for Branch Operations...................................................................................... 6-9 Condition Register (CR) .............................................................................................. 6-9 Link Register (LR)..................................................................................................... 6-11 Count Register (CTR)................................................................................................ 6-11 Processor Control Registers........................................................................................... 6-11 Machine State Register (MSR) .................................................................................. 6-12 Processor ID Register (PIR) ...................................................................................... 6-14 Processor Version Register (PVR)............................................................................. 6-14 System Version Register (SVR)................................................................................. 6-14 Timer Registers .............................................................................................................. 6-15 Timer Control Register (TCR)................................................................................... 6-15 Timer Status Register (TSR)...................................................................................... 6-16 Time Base Registers .................................................................................................. 6-17 Decrementer Register ................................................................................................ 6-17 Decrementer Auto-Reload Register (DECAR).......................................................... 6-17 Interrupt Registers.......................................................................................................... 6-18 Interrupt Registers Defined by Book E...................................................................... 6-18 Save/Restore Register 0 (SRR0)............................................................................ 6-18 Save/Restore Register 1 (SRR1)............................................................................ 6-18 Critical Save/Restore Register 0 (CSRR0) ............................................................ 6-18 Critical Save/Restore Register 1 (CSRR1) ............................................................ 6-19 Data Exception Address Register (DEAR)............................................................ 6-19 Interrupt Vector Prefix Register (IVPR) ................................................................ 6-19 Interrupt Vector Offset Registers (IVORn)............................................................ 6-19 Exception Syndrome Register (ESR) .................................................................... 6-20 EIS-Defined Interrupt Registers ................................................................................ 6-21 Machine Check Save/Restore Register 0 (MCSRR0) ........................................... 6-21 Machine Check Save/Restore Register 1 (MCSRR1) ........................................... 6-22 Machine Check Address Register (MCAR) .......................................................... 6-22 Machine Check Syndrome Register (MCSR)........................................................ 6-23 Software-Use SPRs (SPRG0–SPRG7 and USPRG0) ................................................... 6-24 Branch Target Buffer (BTB) Registers .......................................................................... 6-25 Branch Buffer Entry Address Register (BBEAR) ..................................................... 6-25 Branch Buffer Target Address Register (BBTAR) .................................................... 6-25 Branch Unit Control and Status Register (BUCSR).................................................. 6-26 Hardware Implementation-Dependent Registers........................................................... 6-27 Hardware Implementation-Dependent Register 0 (HID0)......................................... 6-27 Hardware Implementation-Dependent Register 1 (HID1)......................................... 6-28 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 x Freescale Semiconductor Contents Paragraph Number 6.11 6.11.1 6.11.2 6.11.3 6.11.4 6.12 6.12.1 6.12.2 6.12.3 6.12.4 6.12.4.1 6.12.4.2 6.12.5 6.12.5.1 6.12.5.2 6.12.5.3 6.12.5.4 6.12.5.5 6.12.5.6 6.13 6.13.1 6.13.1.1 6.13.1.2 6.13.1.3 6.13.2 6.13.3 6.13.4 6.14 6.14.1 6.15 6.15.1 6.15.2 6.15.3 6.15.4 Title Page Number L1 Cache Configuration Registers................................................................................. 6-30 L1 Cache Control and Status Register 0 (L1CSR0) .................................................. 6-30 L1 Cache Control and Status Register 1 (L1CSR1) .................................................. 6-31 L1 Cache Configuration Register 0 (L1CFG0) ......................................................... 6-32 L1 Cache Configuration Register 1 (L1CFG1) ......................................................... 6-33 MMU Registers.............................................................................................................. 6-34 Process ID Registers (PID0–PID2)............................................................................ 6-34 MMU Control and Status Register 0 (MMUCSR0) .................................................. 6-34 MMU Configuration Register (MMUCFG) .............................................................. 6-35 TLB Configuration Registers (TLBnCFG)................................................................ 6-35 TLB0 Configuration Register 0 (TLB0CFG) ........................................................ 6-35 TLB1 Configuration Register 1 (TLB1CFG) ........................................................ 6-36 MMU Assist Registers............................................................................................... 6-37 MAS Register 0 (MAS0) ....................................................................................... 6-37 MAS Register 1 (MAS1) ....................................................................................... 6-38 MAS Register 2 (MAS2) ....................................................................................... 6-39 MAS Register 3 (MAS3) ....................................................................................... 6-40 MAS Register 4 (MAS4) ....................................................................................... 6-40 MAS Register 6 (MAS6) ....................................................................................... 6-41 Debug Registers ............................................................................................................. 6-42 Debug Control Registers (DBCR0–DBCR2) ............................................................ 6-42 Debug Control Register 0 (DBCR0)...................................................................... 6-42 Debug Control Register 1 (DBCR1)...................................................................... 6-43 Debug Control Register 2 (DBCR2)...................................................................... 6-45 Debug Status Register (DBSR).................................................................................. 6-46 Instruction Address Compare Registers (IAC1–IAC2) ............................................. 6-47 Data Address Compare Registers (DAC1–DAC2).................................................... 6-48 Signal Processing and Embedded Floating-Point Status and Control Register (SPEFSCR) .................................................................................... 6-48 Accumulator (ACC)................................................................................................... 6-50 Performance Monitor Registers (PMRs) ....................................................................... 6-51 Global Control Register 0 (PMGC0, UPMGC0)....................................................... 6-52 Local Control A Registers (PMLCa0–PMLCa3, UPMLCa0–UPMLCa3) ...................................................... 6-52 Local Control B Registers (PMLCb0–PMLCb3, UPMLCb0–UPMLCb3) ..................................................... 6-53 Performance Monitor Counter Registers (PMC0–PMC3, UPMC0–UPMC3) ....................................................................... 6-54 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor xi Contents Paragraph Number Title Chapter 7 L2 Look-Aside Cache/SRAM 7.1 7.1.1 7.2 7.3 7.3.1 7.3.1.1 7.3.1.2 7.3.1.3 7.3.1.4 7.3.1.5 7.3.1.5.1 7.3.1.5.2 7.4 7.5 7.6 7.6.1 7.6.2 7.7 7.7.1 7.7.2 7.7.3 7.7.4 7.7.5 7.7.6 7.8 7.8.1 7.8.2 7.9 7.9.1 7.9.2 7.9.3 7.10 7.10.1 7.10.1.1 7.10.1.2 7.10.2 L2 Cache Overview ......................................................................................................... 7-1 L2 Cache and SRAM Features .................................................................................... 7-2 Cache Organization.......................................................................................................... 7-3 Memory Map/Register Definition ................................................................................... 7-6 L2/SRAM Register Descriptions ................................................................................. 7-7 L2 Control Register (L2CTL).................................................................................. 7-7 L2 Cache External Write Address Registers 0–3 (L2CEWARn) .......................... 7-10 L2 Cache External Write Control Registers 0–3 (L2CEWCRn) ........................... 7-10 L2 Memory-Mapped SRAM Base Address Registers 0–1 (L2SRBARn) ..................................................................................................... 7-11 L2 Error Registers.................................................................................................. 7-12 Error Injection Registers.................................................................................... 7-13 Error Control and Capture Registers ................................................................. 7-15 External Writes to the L2 Cache (Cache Stashing)........................................................ 7-21 L2 Cache Timing ........................................................................................................... 7-21 L2 Cache and SRAM Coherency................................................................................... 7-22 L2 Cache Coherency Rules........................................................................................ 7-22 Memory-Mapped SRAM Coherency Rules .............................................................. 7-24 L2 Cache Locking.......................................................................................................... 7-24 Locking the Entire L2 Cache ..................................................................................... 7-25 Locking Programmed Memory Ranges..................................................................... 7-25 Locking Selected Lines.............................................................................................. 7-25 Clearing Locks on Selected Lines ............................................................................. 7-26 Flash Clearing of Instruction and Data Locks ........................................................... 7-27 Locks with Stale Data ................................................................................................ 7-27 PLRU L2 Replacement Policy....................................................................................... 7-27 PLRU Bit Update Considerations.............................................................................. 7-28 Allocation of Lines .................................................................................................... 7-29 L2 Cache Operation ....................................................................................................... 7-29 L2 Cache States ......................................................................................................... 7-29 Flash Invalidation of the L2 Cache............................................................................ 7-30 L2 State Transitions ................................................................................................... 7-30 Initialization/Application Information ........................................................................... 7-34 Initialization ............................................................................................................... 7-35 L2 Cache Initialization .......................................................................................... 7-35 Memory-Mapped SRAM Initialization ................................................................. 7-35 Managing Errors ........................................................................................................ 7-35 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 xii Freescale Semiconductor Page Number Contents Paragraph Number 7.10.2.1 7.10.2.2 Title Page Number ECC Errors............................................................................................................. 7-35 Tag Parity Errors.................................................................................................... 7-36 Part III Memory and I/O Interfaces Chapter 8 e500 Coherency Module 8.1 8.1.1 8.1.2 8.2 8.2.1 8.2.1.1 8.2.1.2 8.2.1.3 8.2.1.4 8.2.1.5 8.2.1.6 8.3 8.3.1 8.3.2 8.3.3 8.3.4 8.3.5 8.4 Introduction...................................................................................................................... 8-1 Overview...................................................................................................................... 8-1 Features........................................................................................................................ 8-2 Memory Map/Register Definition ................................................................................... 8-2 Register Descriptions................................................................................................... 8-3 ECM CCB Address Configuration Register (EEBACR) ........................................ 8-3 ECM CCB Port Configuration Register (EEBPCR) ............................................... 8-4 ECM Error Detect Register (EEDR) ....................................................................... 8-5 ECM Error Enable Register (EEER) ....................................................................... 8-6 ECM Error Attributes Capture Register (EEATR) .................................................. 8-7 ECM Error Address Capture Register (EEADR) .................................................... 8-8 Functional Description..................................................................................................... 8-9 I/O Arbiter.................................................................................................................... 8-9 CCB Arbiter................................................................................................................. 8-9 Transaction Queue ....................................................................................................... 8-9 Global Data Multiplexor............................................................................................ 8-10 CCB Interface ............................................................................................................ 8-10 Initialization/Application Information ........................................................................... 8-10 Chapter 9 DDR Memory Controller 9.1 9.2 9.2.1 9.3 9.3.1 9.3.2 9.3.2.1 9.3.2.2 9.3.2.3 Introduction...................................................................................................................... 9-1 Features ............................................................................................................................ 9-2 Modes of Operation ..................................................................................................... 9-3 External Signal Descriptions ........................................................................................... 9-3 Signals Overview......................................................................................................... 9-3 Detailed Signal Descriptions ....................................................................................... 9-5 Memory Interface Signals........................................................................................ 9-5 Clock Interface Signals............................................................................................ 9-8 Debug Signals.......................................................................................................... 9-8 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor xiii Contents Paragraph Number 9.4 9.4.1 9.4.1.1 9.4.1.2 9.4.1.3 9.4.1.4 9.4.1.5 9.4.1.6 9.4.1.7 9.4.1.8 9.4.1.9 9.4.1.10 9.4.1.11 9.4.1.12 9.4.1.13 9.4.1.14 9.4.1.15 9.4.1.16 9.4.1.17 9.4.1.18 9.4.1.19 9.5 9.5.1 9.5.1.1 9.5.2 9.5.3 9.5.4 9.5.4.1 9.5.5 9.5.6 9.5.7 9.5.8 9.5.8.1 9.5.8.2 9.5.8.2.1 9.5.9 9.5.10 Title Page Number Memory Map/Register Definition ................................................................................... 9-9 Register Descriptions................................................................................................... 9-9 Chip Select Memory Bounds (CSn_BNDS).......................................................... 9-10 Chip Select Configuration (CSn_CONFIG).......................................................... 9-10 DDR SDRAM Timing Configuration 1 (TIMING_CFG_1)................................. 9-11 DDR SDRAM Timing Configuration 2 (TIMING_CFG_2)................................. 9-13 DDR SDRAM Control Configuration (DDR_SDRAM_CFG)............................. 9-14 DDR SDRAM Mode Configuration (DDR_SDRAM_MODE)............................ 9-16 DDR SDRAM Interval Configuration (DDR_SDRAM_INTERVAL) ................. 9-16 Memory Data Path Error Injection Mask High (DATA_ERR_INJECT_HI) ............................................................................... 9-17 Memory Data Path Error Injection Mask Low (DATA_ERR_INJECT_LO).............................................................................. 9-18 Memory Data Path Error Injection Mask ECC (ECC_ERR_INJECT)........................................................................................ 9-18 Memory Data Path Read Capture High (CAPTURE_DATA_HI)......................... 9-19 Memory Data Path Read Capture Low (CAPTURE_DATA_LO) ........................ 9-20 Memory Data Path Read Capture ECC (CAPTURE_ECC).................................. 9-20 Memory Error Detect (ERR_DETECT)................................................................ 9-21 Memory Error Disable (ERR_DISABLE)............................................................. 9-21 Memory Error Interrupt Enable (ERR_INT_EN).................................................. 9-22 Memory Error Attributes Capture (CAPTURE_ATTRIBUTES).......................... 9-23 Memory Error Address Capture (CAPTURE_ADDRESS) .................................. 9-24 Single-Bit ECC Memory Error Management (ERR_SBE) ................................... 9-25 Functional Description................................................................................................... 9-25 DDR SDRAM Interface Operation............................................................................ 9-30 Supported DDR SDRAM Organizations............................................................... 9-31 DDR SDRAM Address Multiplexing........................................................................ 9-31 JEDEC Standard DDR SDRAM Interface Commands ............................................. 9-32 SDRAM Interface Timing ......................................................................................... 9-34 Clock Distribution ................................................................................................. 9-38 DDR SDRAM Mode-Set Command Timing............................................................. 9-38 DDR SDRAM Registered DIMM Mode ................................................................... 9-39 DDR SDRAM Write Timing Adjustments ................................................................ 9-40 DDR SDRAM Refresh .............................................................................................. 9-41 DDR SDRAM Refresh Timing.............................................................................. 9-42 DDR SDRAM Refresh and Power-Saving Modes ................................................ 9-43 Self-Refresh in Sleep Mode............................................................................... 9-44 DDR Data Beat Ordering........................................................................................... 9-45 Page Mode and Logical Bank Retention ................................................................... 9-45 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 xiv Freescale Semiconductor Contents Paragraph Number 9.5.11 9.5.12 9.6 9.6.1 Title Page Number Error Checking and Correcting (ECC) ...................................................................... 9-46 Error Management ..................................................................................................... 9-48 Initialization/Application Information ........................................................................... 9-49 DDR SDRAM Initialization Sequence ...................................................................... 9-50 Chapter 10 Programmable Interrupt Controller 10.1 10.1.1 10.1.2 10.1.3 10.1.4 10.1.4.1 10.1.4.2 10.1.5 10.1.5.1 10.1.5.2 10.2 10.2.1 10.2.2 10.3 10.3.1 10.3.1.1 10.3.1.2 10.3.1.3 10.3.1.4 10.3.1.5 10.3.1.6 10.3.2 10.3.2.1 10.3.2.2 10.3.2.3 10.3.2.4 10.3.2.5 10.3.2.6 10.3.3 10.3.3.1 10.3.3.2 10.3.3.3 Introduction.................................................................................................................... 10-1 Overview.................................................................................................................... 10-2 Features...................................................................................................................... 10-3 Interrupts to the Processor Core................................................................................. 10-4 Modes of Operation ................................................................................................... 10-5 Mixed Mode (GCR[M] = 1) .................................................................................. 10-5 Pass-Through Mode (GCR[M] = 0) ...................................................................... 10-6 Interrupt Sources........................................................................................................ 10-6 Interrupt Routing—Mixed Mode........................................................................... 10-7 Internal Interrupt Sources ...................................................................................... 10-7 External Signal Descriptions ......................................................................................... 10-8 Signal Overview ........................................................................................................ 10-8 Detailed Signal Descriptions ..................................................................................... 10-8 Memory Map/Register Definition ................................................................................. 10-9 Global Registers....................................................................................................... 10-16 Feature Reporting Register (FRR)....................................................................... 10-16 Global Configuration Register (GCR)................................................................. 10-17 Vendor Identification Register (VIR) .................................................................. 10-17 Processor Initialization Register (PIR) ................................................................ 10-18 IPI Vector/Priority Registers (IPIVPRn) ............................................................. 10-19 Spurious Vector Register (SVR).......................................................................... 10-19 Global Timer Registers ............................................................................................ 10-20 Timer Frequency Reporting Register (TFRR)..................................................... 10-20 Global Timer Current Count Registers (GTCCRn) ............................................. 10-21 Global Timer Base Count Registers (GTBCRn) ................................................. 10-21 Global Timer Vector/Priority Registers (GTVPRn) ............................................ 10-22 Global Timer Destination Registers (GTDRn) .................................................... 10-23 Timer Control Register (TCR)............................................................................. 10-24 IRQ_OUT and Critical Interrupt Summary Registers ............................................. 10-26 IRQ_OUT Summary Register 0 (IRQSR0) ......................................................... 10-26 IRQ_OUT Summary Register 1 (IRQSR1) ......................................................... 10-27 Critical Interrupt Summary Register 0 (CISR0).................................................. 10-27 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor xv Contents Paragraph Number 10.3.3.4 10.3.4 10.3.4.1 10.3.4.2 10.3.5 10.3.5.1 10.3.5.2 10.3.5.3 10.3.6 10.3.6.1 10.3.6.2 10.3.6.3 10.3.6.4 10.3.6.5 10.3.6.6 10.3.7 10.3.7.1 10.3.7.2 10.3.7.3 10.3.7.4 10.3.7.5 10.4 10.4.1 10.4.1.1 10.4.1.2 10.4.1.3 10.4.2 10.4.3 10.4.4 10.4.5 10.4.6 10.4.7 10.5 10.5.1 10.5.1.1 10.5.1.2 Title Page Number Critical Interrupt Summary Register 1 (CISR1).................................................. 10-28 Performance Monitor Mask Registers (PMMRs).................................................... 10-28 Performance Monitor Mask Register (Lower) (PMnMR0)................................. 10-29 Performance Monitor Mask Registers (Upper) (PMnMR1)................................ 10-30 Message Registers.................................................................................................... 10-30 Message Registers (MSGR0–MSGR3) ............................................................... 10-30 Message Enable Register (MER)......................................................................... 10-31 Message Status Register (MSR) .......................................................................... 10-31 Interrupt Source Configuration Registers ................................................................ 10-32 External Interrupt Vector/Priority Registers (EIVPR0–EIVPR11) ..................... 10-32 External Interrupt Destination Registers (EIDR0–EIDR11) ............................... 10-33 Internal Interrupt Vector/Priority Registers (IIVPR0–IIVPR31)......................... 10-34 Internal Interrupt Destination Registers (IIDR0–IIDR31) .................................. 10-35 Messaging Interrupt Vector/Priority Registers (MIVPR0–MIVPR3) ................. 10-36 Messaging Interrupt Destination Registers (MIDR0–MIDR3) ........................... 10-37 Per-CPU Registers ................................................................................................... 10-38 Interprocessor Interrupt Dispatch Register (IPIDR0–IPIDR3) ........................... 10-39 Processor Current Task Priority Register (CTPR)............................................... 10-40 Who Am I Register (WHOAMI) ......................................................................... 10-41 Processor Interrupt Acknowledge Register (IACK)............................................ 10-41 Processor End of Interrupt Register (EOI) .......................................................... 10-42 Functional Description................................................................................................. 10-43 Flow of Interrupt Control......................................................................................... 10-43 Interrupt Source Priority ...................................................................................... 10-45 Processor Current Task Priority........................................................................... 10-45 Interrupt Acknowledge ........................................................................................ 10-46 Nesting of Interrupts ................................................................................................ 10-46 Processor Initialization ............................................................................................ 10-46 Spurious Vector Generation ..................................................................................... 10-46 Messaging Interrupts................................................................................................ 10-47 Global Timers .......................................................................................................... 10-47 Reset of the PIC ....................................................................................................... 10-47 Initialization/Application Information ......................................................................... 10-48 Programming Guidelines ......................................................................................... 10-48 PIC Registers ....................................................................................................... 10-48 Changing Interrupt Source Configuration ........................................................... 10-50 Chapter 11 I2C Interface MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 xvi Freescale Semiconductor Contents Paragraph Number 11.1 11.1.1 11.1.2 11.1.3 11.2 11.2.1 11.2.2 11.3 11.3.1 11.3.1.1 11.3.1.2 11.3.1.3 11.3.1.4 11.3.1.5 11.3.1.6 11.4 11.4.1 11.4.1.1 11.4.1.2 11.4.1.3 11.4.1.4 11.4.1.5 11.4.1.5.1 11.4.1.5.2 11.4.1.6 11.4.2 11.4.2.1 11.4.3 11.4.4 11.4.4.1 11.4.4.2 11.4.4.2.1 11.4.4.2.2 11.4.4.3 11.4.5 11.4.5.1 11.4.5.2 11.5 11.5.1 11.5.2 Title Page Number Introduction.................................................................................................................... 11-1 Overview.................................................................................................................... 11-2 Features...................................................................................................................... 11-2 Modes of Operation ................................................................................................... 11-2 External Signal Descriptions ......................................................................................... 11-3 Signal Overview ........................................................................................................ 11-3 Detailed Signal Descriptions ..................................................................................... 11-3 Memory Map/Register Definition ................................................................................. 11-4 Register Descriptions................................................................................................. 11-4 I2C Address Register (I2CADR) ........................................................................... 11-5 I2C Frequency Divider Register (I2CFDR)........................................................... 11-5 I2C Control Register (I2CCR) ............................................................................... 11-6 I2C Status Register (I2CSR) .................................................................................. 11-7 I2C Data Register (I2CDR).................................................................................... 11-9 Digital Filter Sampling Rate Register (I2CDFSRR) ........................................... 11-10 Functional Description................................................................................................. 11-10 Transaction Protocol ................................................................................................ 11-10 START Condition .................................................................................................11-11 Slave Address Transmission.................................................................................11-11 Repeated START Condition ................................................................................ 11-12 STOP Condition................................................................................................... 11-12 Protocol Implementation Details ......................................................................... 11-13 Transaction Monitoring—Implementation Details.......................................... 11-13 Control Transfer—Implementation Details ..................................................... 11-13 Address Compare—Implementation Details ....................................................... 11-14 Arbitration Procedure .............................................................................................. 11-14 Arbitration Control .............................................................................................. 11-15 Handshaking ............................................................................................................ 11-15 Clock Control........................................................................................................... 11-15 Clock Synchronization......................................................................................... 11-16 Input Synchronization and Digital Filter ............................................................. 11-16 Input Signal Synchronization .......................................................................... 11-16 Filtering of SCL and SDA Lines ..................................................................... 11-16 Clock Stretching .................................................................................................. 11-17 Boot Sequencer Mode.............................................................................................. 11-17 EEPROM Calling Address .................................................................................. 11-17 EEPROM Data Format ........................................................................................ 11-18 Initialization/Application Information ......................................................................... 11-20 Initialization Sequence............................................................................................. 11-20 Generation of START .............................................................................................. 11-21 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor xvii Contents Paragraph Number 11.5.3 11.5.4 11.5.5 11.5.6 11.5.7 11.5.7.1 11.5.7.2 11.5.8 Title Page Number Post-Transfer Software Response ............................................................................ 11-21 Generation of STOP................................................................................................. 11-22 Generation of Repeated START .............................................................................. 11-22 Generation of SCL When SDA Low ....................................................................... 11-22 Slave Mode Interrupt Service Routine..................................................................... 11-23 Slave Transmitter and Received Acknowledge ................................................... 11-23 Loss of Arbitration and Forcing of Slave Mode.................................................. 11-23 Interrupt Service Routine Flowchart........................................................................ 11-23 Chapter 12 DUART 12.1 12.1.1 12.1.2 12.2 12.2.1 12.2.2 12.3 12.3.1 12.3.1.1 12.3.1.2 12.3.1.3 12.3.1.4 12.3.1.5 12.3.1.6 12.3.1.7 12.3.1.8 12.3.1.9 12.3.1.10 12.3.1.11 12.3.1.12 12.3.1.13 12.4 12.4.1 12.4.1.1 12.4.1.2 12.4.1.3 12.4.1.4 Overview........................................................................................................................ 12-1 Features...................................................................................................................... 12-2 Modes of Operation ................................................................................................... 12-3 External Signal Descriptions ......................................................................................... 12-3 Signal Overview ........................................................................................................ 12-3 Detailed Signal Descriptions ..................................................................................... 12-3 Memory Map/Register Definition ................................................................................. 12-4 Register Descriptions................................................................................................. 12-6 Receiver Buffer Registers (URBR0, URBR1) ...................................................... 12-6 Transmitter Holding Registers (UTHR0, UTHR1) ............................................... 12-6 Divisor Most and Least Significant Byte Registers (UDMB and UDLB).......................................................................................... 12-7 Interrupt Enable Register (UIER).......................................................................... 12-9 Interrupt ID Registers (UIIR0, UIIR1) ................................................................ 12-10 FIFO Control Registers (UFCR0, UFCR1) ......................................................... 12-11 Line Control Registers (ULCR0, ULCR1).......................................................... 12-12 MODEM Control Registers (UMCR0, UMCR1)................................................ 12-14 Line Status Registers (ULSR0, ULSR1) ............................................................. 12-15 MODEM Status Registers (UMSR0, UMSR1) ................................................... 12-16 Scratch Registers (USCR0, USCR1) ................................................................... 12-17 Alternate Function Registers (UAFR0, UAFR1) ................................................ 12-17 DMA Status Registers (UDSR0, UDSR1) .......................................................... 12-18 Functional Description................................................................................................. 12-20 Serial Interface......................................................................................................... 12-20 START Bit ........................................................................................................... 12-21 Data Transfer ....................................................................................................... 12-21 Parity Bit .............................................................................................................. 12-22 STOP Bit.............................................................................................................. 12-22 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 xviii Freescale Semiconductor Contents Paragraph Number 12.4.2 12.4.3 12.4.4 12.4.4.1 12.4.4.2 12.4.4.3 12.4.5 12.4.5.1 12.4.5.2 12.4.5.3 12.5 Title Page Number Baud-Rate Generator Logic ..................................................................................... 12-22 Local Loop-Back Mode ........................................................................................... 12-22 Errors ....................................................................................................................... 12-23 Framing Error ...................................................................................................... 12-23 Parity Error .......................................................................................................... 12-23 Overrun Error....................................................................................................... 12-23 FIFO Mode .............................................................................................................. 12-24 FIFO Interrupts .................................................................................................... 12-24 DMA Mode Select ............................................................................................... 12-24 Interrupt Control Logic........................................................................................ 12-25 DUART Initialization/Application Information .......................................................... 12-25 Chapter 13 Local Bus Controller 13.1 13.1.1 13.1.2 13.1.3 13.1.3.1 13.1.3.2 13.1.4 13.1.5 13.2 13.3 13.3.1 13.3.1.1 13.3.1.2 13.3.1.2.1 13.3.1.2.2 13.3.1.2.3 13.3.1.2.4 13.3.1.3 13.3.1.4 13.3.1.5 13.3.1.6 13.3.1.7 13.3.1.8 13.3.1.9 13.3.1.10 Introduction.................................................................................................................... 13-1 Overview.................................................................................................................... 13-2 Features...................................................................................................................... 13-2 Modes of Operation ................................................................................................... 13-3 LBC Bus Clock and Clock Ratios ......................................................................... 13-4 Source ID Debug Mode ......................................................................................... 13-4 Power-Down Mode.................................................................................................... 13-4 References.................................................................................................................. 13-4 External Signal Descriptions ......................................................................................... 13-5 Memory Map/Register Definition ................................................................................. 13-9 Register Descriptions............................................................................................... 13-10 Base Registers (BR0–BR7) ................................................................................. 13-11 Option Registers (OR0–OR7).............................................................................. 13-12 Address Mask .................................................................................................. 13-13 Option Registers (ORn)—GPCM Mode ......................................................... 13-14 Option Registers (ORn)—UPM Mode ............................................................ 13-16 Option Registers (ORn)—SDRAM Mode ...................................................... 13-17 UPM Memory Address Register (MAR)............................................................. 13-18 UPM Mode Registers (MxMR) ........................................................................... 13-19 Memory Refresh Timer Prescaler Register (MRTPR) ........................................ 13-21 UPM Data Register (MDR) ................................................................................. 13-22 SDRAM Machine Mode Register (LSDMR) ...................................................... 13-22 UPM Refresh Timer (LURT)............................................................................... 13-24 SDRAM Refresh Timer (LSRT).......................................................................... 13-25 Transfer Error Status Register (LTESR).............................................................. 13-26 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor xix Contents Paragraph Number 13.3.1.11 13.3.1.12 13.3.1.13 13.3.1.14 13.3.1.15 13.3.1.16 13.4 13.4.1 13.4.1.1 13.4.1.2 13.4.1.3 13.4.1.4 13.4.1.5 13.4.1.6 13.4.1.7 13.4.2 13.4.2.1 13.4.2.2 13.4.2.2.1 13.4.2.2.2 13.4.2.2.3 13.4.2.2.4 13.4.2.2.5 13.4.2.3 13.4.2.4 13.4.3 13.4.3.1 13.4.3.2 13.4.3.3 13.4.3.4 13.4.3.5 13.4.3.6 13.4.3.7 13.4.3.7.1 13.4.3.7.2 13.4.3.7.3 13.4.3.7.4 13.4.3.7.5 13.4.3.7.6 13.4.3.8 Title Page Number Transfer Error Check Disable Register (LTEDR)................................................ 13-27 Transfer Error Interrupt Enable Register (LTEIR) .............................................. 13-28 Transfer Error Attributes Register (LTEATR) ..................................................... 13-29 Transfer Error Address Register (LTEAR).......................................................... 13-30 Local Bus Configuration Register (LBCR) ......................................................... 13-31 Clock Ratio Register (LCRR).............................................................................. 13-32 Functional Description................................................................................................. 13-33 Basic Architecture.................................................................................................... 13-34 Address and Address Space Checking ................................................................ 13-34 External Address Latch Enable Signal (LALE) .................................................. 13-35 Data Transfer Acknowledge (TA) ....................................................................... 13-36 Data Buffer Control (LBCTL)............................................................................. 13-37 Atomic Operation ................................................................................................ 13-37 Parity Generation and Checking (LDP)............................................................... 13-38 Bus Monitor ......................................................................................................... 13-38 General-Purpose Chip-Select Machine (GPCM)..................................................... 13-38 Timing Configuration .......................................................................................... 13-39 Chip-Select Assertion Timing ............................................................................. 13-44 Programmable Wait State Configuration......................................................... 13-44 Chip-Select and Write Enable Negation Timing ............................................. 13-44 Relaxed Timing ............................................................................................... 13-45 Output Enable (LOE) Timing.......................................................................... 13-48 Extended Hold Time on Read Accesses .......................................................... 13-48 External Access Termination (LGTA) ................................................................. 13-49 Boot Chip-Select Operation................................................................................. 13-50 SDRAM Machine .................................................................................................... 13-51 Supported SDRAM Configurations..................................................................... 13-51 SDRAM Power-On Initialization ........................................................................ 13-52 Intel PC133 and JEDEC-Standard SDRAM Interface Commands ..................... 13-52 Page Hit Checking ............................................................................................... 13-53 Page Management................................................................................................ 13-54 SDRAM Address Multiplexing ........................................................................... 13-54 SDRAM Device-Specific Parameters.................................................................. 13-55 Precharge-to-Activate Interval......................................................................... 13-55 Activate-to-Read/Write Interval ...................................................................... 13-56 Column Address to First Data Out—CAS Latency......................................... 13-56 Last Data In to Precharge—Write Recovery ................................................... 13-56 Refresh Recovery Interval (RFRC) ................................................................. 13-57 External Address and Command Buffers (BUFCMD).................................... 13-57 SDRAM Interface Timing ................................................................................... 13-58 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 xx Freescale Semiconductor Contents Paragraph Number 13.4.3.9 13.4.3.10 13.4.3.11 13.4.3.11.1 13.4.4 13.4.4.1 13.4.4.1.1 13.4.4.1.2 13.4.4.1.3 13.4.4.1.4 13.4.4.2 13.4.4.3 13.4.4.4 13.4.4.4.1 13.4.4.4.2 13.4.4.4.3 13.4.4.4.4 13.4.4.4.5 13.4.4.4.6 13.4.4.4.7 13.4.4.4.8 13.4.4.4.9 13.4.4.4.10 13.4.4.5 13.4.4.6 13.4.4.7 13.5 13.5.1 13.5.1.1 13.5.1.2 13.5.1.3 13.5.1.4 13.5.2 13.5.2.1 13.5.2.2 13.5.2.3 13.5.2.4 13.5.3 13.5.4 13.5.4.1 Title Page Number SDRAM Read/Write Transactions....................................................................... 13-60 SDRAM MODE-SET Command Timing............................................................ 13-60 SDRAM Refresh.................................................................................................. 13-61 SDRAM Refresh Timing ................................................................................. 13-61 User-Programmable Machines (UPMs)................................................................... 13-62 UPM Requests ..................................................................................................... 13-63 Memory Access Requests................................................................................ 13-64 UPM Refresh Timer Requests ......................................................................... 13-65 Software Requests—RUN Command ............................................................. 13-65 Exception Requests.......................................................................................... 13-66 Programming the UPMs ...................................................................................... 13-66 UPM Signal Timing............................................................................................. 13-66 RAM Array .......................................................................................................... 13-67 RAM Words..................................................................................................... 13-68 Chip-Select Signal Timing (CSTn) ................................................................. 13-72 Byte Select Signal Timing (BSTn) .................................................................. 13-72 General-Purpose Signals (GnTn, GOn)........................................................... 13-73 Loop Control (LOOP) ..................................................................................... 13-73 Repeat Execution of Current RAM Word (REDO) ......................................... 13-74 Address Multiplexing (AMX) ......................................................................... 13-74 Data Valid and Data Sample Control (UTA) ................................................... 13-75 LGPL[0:5] Signal Negation (LAST) ............................................................... 13-76 Wait Mechanism (WAEN) ............................................................................... 13-76 Synchronous Sampling of LUPWAIT for Early Transfer Acknowledge ............ 13-77 Extended Hold Time on Read Accesses .............................................................. 13-77 Memory System Interface Example Using UPM ................................................ 13-78 Initialization/Application Information ......................................................................... 13-84 Interfacing to Peripherals......................................................................................... 13-84 Multiplexed Address/Data Bus and Unmultiplexed Address Signals ................. 13-84 Peripheral Hierarchy on the Local Bus................................................................ 13-84 Peripheral Hierarchy on the Local Bus for Very High Bus Speeds..................... 13-85 GPCM Timings.................................................................................................... 13-86 Bus Turnaround ....................................................................................................... 13-87 Address Phase After Previous Read .................................................................... 13-87 Read Data Phase After Address Phase ................................................................ 13-88 Read-Modify-Write Cycle for Parity Protected Memory Banks ......................... 13-88 UPM Cycles with Additional Address Phases..................................................... 13-88 Interface to Different Port-Size Devices.................................................................. 13-88 Interfacing to SDRAM............................................................................................. 13-90 Basic SDRAM Capabilities of the Local Bus...................................................... 13-90 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor xxi Contents Paragraph Number 13.5.4.2 13.5.4.3 13.5.4.3.1 13.5.4.3.2 13.5.4.3.3 13.5.4.3.4 13.5.4.3.5 13.5.4.3.6 13.5.4.3.7 13.5.4.3.8 13.5.4.4 13.5.5 13.5.6 13.5.6.1 13.5.6.1.1 13.5.6.1.2 13.5.6.1.3 13.5.6.1.4 13.5.6.2 13.5.6.2.1 13.5.6.2.2 13.5.6.2.3 13.5.6.3 13.5.6.3.1 Title Page Number Maximum Amount of SDRAM Supported.......................................................... 13-91 SDRAM Machine Limitations............................................................................. 13-92 Analysis of Maximum Row Number Due to Bank Select Multiplexing ................................................................................................ 13-92 Bank Select Signals ......................................................................................... 13-92 128-Mbyte SDRAM ........................................................................................ 13-92 256-Mbyte SDRAM ........................................................................................ 13-95 512-Mbyte SDRAM ........................................................................................ 13-95 Power-Down Mode.......................................................................................... 13-97 Self-Refresh ..................................................................................................... 13-97 SDRAM Timing .............................................................................................. 13-99 Parity Support for SDRAM ............................................................................... 13-101 Interfacing to ZBT SRAM..................................................................................... 13-101 Interfacing to DSP Host Ports................................................................................ 13-103 Interfacing to MSC8101 HDI16 ........................................................................ 13-103 HDI16 Peripherals ......................................................................................... 13-103 Physical Interconnections .............................................................................. 13-105 Supporting Burst Transfers............................................................................ 13-106 Host 60x Bus: HDI16 Peripheral Interface Hardware Timings..................... 13-107 Interfacing to MSC8102 DSI............................................................................. 13-107 DSI in Asynchronous SRAM-Like Mode ..................................................... 13-108 DSI in Synchronous Mode ............................................................................ 13-110 Broadcast Accesses........................................................................................ 13-118 Interfacing to EHPI from Texas Instruments TMS320Cxxxx DSPs ................. 13-118 Expansion to Multiple DSPs.......................................................................... 13-121 Chapter 14 Three-Speed Ethernet Controllers 14.1 14.1.1 14.2 14.3 14.4 14.4.1 14.5 14.5.1 14.5.2 14.5.3 14.5.3.1 Introduction.................................................................................................................... 14-1 Three-Speed Ethernet Controller Overview .............................................................. 14-7 Features .......................................................................................................................... 14-7 Modes of Operation ....................................................................................................... 14-8 External Signal Descriptions ......................................................................................... 14-9 Detailed Signal Descriptions ................................................................................... 14-10 Memory Map/Register Definition ............................................................................... 14-13 Top-Level Module Memory Map ............................................................................ 14-13 Detailed Memory Map—Control/Status Registers.................................................. 14-14 Memory-Mapped Register Descriptions.................................................................. 14-20 TSEC General Control and Status Registers ....................................................... 14-20 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 xxii Freescale Semiconductor Contents Paragraph Number 14.5.3.1.1 14.5.3.1.2 14.5.3.1.3 14.5.3.1.4 14.5.3.1.5 14.5.3.1.6 14.5.3.1.7 14.5.3.1.8 14.5.3.2 14.5.3.2.1 14.5.3.2.2 14.5.3.2.3 14.5.3.2.4 14.5.3.3 14.5.3.3.1 14.5.3.3.2 14.5.3.3.3 14.5.3.3.4 14.5.3.3.5 14.5.3.3.6 14.5.3.3.7 14.5.3.3.8 14.5.3.3.9 14.5.3.4 14.5.3.4.1 14.5.3.4.2 14.5.3.4.3 14.5.3.4.4 14.5.3.4.5 14.5.3.4.6 14.5.3.4.7 14.5.3.4.8 14.5.3.5 14.5.3.5.1 14.5.3.5.2 14.5.3.5.3 14.5.3.5.4 14.5.3.5.5 14.5.3.6 Title Page Number Interrupt Event Register (IEVENT) ................................................................ 14-20 Interrupt Mask Register (IMASK) .................................................................. 14-23 Error Disabled Register (EDIS)....................................................................... 14-24 Ethernet Control Register (ECNTRL) ............................................................. 14-25 Minimum Frame Length Register (MINFLR)................................................. 14-26 Pause Time Value Register (PTV) ................................................................... 14-27 DMA Control Register (DMACTRL) ............................................................. 14-28 TBI Physical Address Register (TBIPA) ......................................................... 14-29 TSEC FIFO Control and Status Registers ........................................................... 14-30 FIFO Pause Control Register (FIFO_PAUSE_CTRL) ................................... 14-30 FIFO Transmit Threshold Register (FIFO_TX_THR) .................................... 14-31 FIFO Transmit Starve Register (FIFO_TX_STARVE) ................................... 14-32 FIFO Transmit Starve Shutoff Register (FIFO_TX_STARVE_SHUTOFF).............................................................. 14-32 TSEC Transmit Control and Status Registers...................................................... 14-33 Transmit Control Register (TCTRL) ............................................................... 14-33 Transmit Status Register (TSTAT) .................................................................. 14-34 TxBD Data Length Register (TBDLEN)......................................................... 14-35 Transmit Interrupt Coalescing Configuration Register (TXIC) ...................... 14-35 Current Transmit Buffer Descriptor Pointer Register (CTBPTR) ................... 14-36 Transmit Buffer Descriptor Pointer Register (TBPTR) ................................... 14-37 Transmit Descriptor Base Address Register (TBASE) ................................... 14-37 Out-of-Sequence TxBD Register (OSTBD).................................................... 14-38 Out-of-Sequence Tx Data Buffer Pointer Register (OSTBDP)....................... 14-40 TSEC Receive Control and Status Registers ....................................................... 14-40 Receive Control Register (RCTRL) ................................................................ 14-41 Receive Status Register (RSTAT).................................................................... 14-41 RxBD Data Length Register (RBDLEN) ........................................................ 14-42 Receive Interrupt Coalescing Configuration Register (RXIC)........................ 14-43 Current Receive Buffer Descriptor Pointer Register (CRBPTR) .................... 14-44 Maximum Receive Buffer Length Register (MRBLR) ................................... 14-44 Receive Buffer Descriptor Pointer Register (RBPTR) .................................... 14-45 Receive Descriptor Base Address Register (RBASE)..................................... 14-45 MAC Functionality.............................................................................................. 14-46 Configuring the MAC...................................................................................... 14-46 Controlling CSMA/CD.................................................................................... 14-46 Handling Packet Collisions ............................................................................. 14-47 Controlling Packet Flow.................................................................................. 14-47 Controlling PHY Links.................................................................................... 14-48 MAC Registers .................................................................................................... 14-49 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor xxiii Contents Paragraph Number 14.5.3.6.1 14.5.3.6.2 14.5.3.6.3 14.5.3.6.4 14.5.3.6.5 14.5.3.6.6 14.5.3.6.7 14.5.3.6.8 14.5.3.6.9 14.5.3.6.10 14.5.3.6.11 14.5.3.6.12 14.5.3.6.13 14.5.3.6.14 14.5.3.7 14.5.3.7.1 14.5.3.7.2 14.5.3.7.3 14.5.3.7.4 14.5.3.7.5 14.5.3.7.6 14.5.3.7.7 14.5.3.7.8 14.5.3.7.9 14.5.3.7.10 14.5.3.7.11 14.5.3.7.12 14.5.3.7.13 14.5.3.7.14 14.5.3.7.15 14.5.3.7.16 14.5.3.7.17 14.5.3.7.18 14.5.3.7.19 Title Page Number MAC Configuration Register 1 (MACCFG1)................................................. 14-49 MAC Configuration Register 2 (MACCFG2)................................................. 14-51 Inter-Packet Gap/Inter-Frame Gap Register (IPGIFG) ................................... 14-52 Half-Duplex Register (HAFDUP) ................................................................... 14-53 Maximum Frame Length Register (MAXFRM) ............................................. 14-54 MII Management Configuration Register (MIIMCFG) .................................. 14-54 MII Management Command Register (MIIMCOM)....................................... 14-55 MII Management Address Register (MIIMADD)........................................... 14-56 MII Management Control Register (MIIMCON)............................................ 14-57 MII Management Status Register (MIIMSTAT) ............................................. 14-57 MII Management Indicator Register (MIIMIND)........................................... 14-58 Interface Status Register (IFSTAT) ................................................................. 14-58 Station Address Register Part 1 (MACSTNADDR1) ..................................... 14-59 Station Address Register Part 2 (MACSTNADDR2) ..................................... 14-60 MIB Registers...................................................................................................... 14-60 Transmit and Receive 64-Byte Frame Counter Register (TR64) .................... 14-61 Transmit and Receive 65- to 127-Byte Frame Counter Register (TR127) ....................................................................................................... 14-61 Transmit and Receive 128- to 255-Byte Frame Counter Register (TR255) ....................................................................................................... 14-62 Transmit and Receive 256- to 511-Byte Frame Counter Register (TR511)........................................................................................................ 14-62 Transmit and Receive 512- to 1023-Byte Frame Counter Register (TR1K) ........................................................................................................ 14-63 Transmit and Receive 1024- to 1518-Byte Frame Counter Register (TRMAX) .................................................................................................... 14-63 Transmit and Receive 1519- to 1522-Byte VLAN Frame Counter Register (TRMGV)...................................................................................... 14-64 Receive Byte Counter Register (RBYT) ......................................................... 14-64 Receive Packet Counter Register (RPKT)....................................................... 14-65 Receive FCS Error Counter Register (RFCS) ................................................. 14-65 Receive Multicast Packet Counter Register (RMCA) ..................................... 14-66 Receive Broadcast Packet Counter Register (RBCA) ..................................... 14-66 Receive Control Frame Packet Counter Register (RXCF) .............................. 14-67 Receive Pause Frame Packet Counter Register (RXPF) ................................. 14-67 Receive Unknown Opcode Packet Counter Register (RXUO) ....................... 14-68 Receive Alignment Error Counter Register (RALN) ...................................... 14-68 Receive Frame Length Error Counter Register (RFLR) ................................. 14-69 Receive Code Error Counter Register (RCDE) ............................................... 14-69 Receive Carrier Sense Error Counter Register (RCSE) .................................. 14-70 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 xxiv Freescale Semiconductor Contents Paragraph Number 14.5.3.7.20 14.5.3.7.21 14.5.3.7.22 14.5.3.7.23 14.5.3.7.24 14.5.3.7.25 14.5.3.7.26 14.5.3.7.27 14.5.3.7.28 14.5.3.7.29 14.5.3.7.30 14.5.3.7.31 14.5.3.7.32 14.5.3.7.33 14.5.3.7.34 14.5.3.7.35 14.5.3.7.36 14.5.3.7.37 14.5.3.7.38 14.5.3.7.39 14.5.3.7.40 14.5.3.7.41 14.5.3.7.42 14.5.3.7.43 14.5.3.7.44 14.5.3.7.45 14.5.3.7.46 14.5.3.7.47 14.5.3.8 14.5.3.8.1 14.5.3.8.2 14.5.3.9 14.5.3.9.1 14.5.3.9.2 14.5.4 14.5.4.1 14.5.4.2 14.5.4.3 14.5.4.4 14.5.4.5 Title Page Number Receive Undersize Packet Counter Register (RUND) .................................... 14-70 Receive Oversize Packet Counter Register (ROVR) ...................................... 14-71 Receive Fragments Counter Register (RFRG) ................................................ 14-71 Receive Jabber Counter Register (RJBR) ....................................................... 14-72 Receive Dropped Packet Counter Register (RDRP) ....................................... 14-72 Transmit Byte Counter Register (TBYT) ........................................................ 14-73 Transmit Packet Counter Register (TPKT) ..................................................... 14-73 Transmit Multicast Packet Counter Register (TMCA).................................... 14-74 Transmit Broadcast Packet Counter Register (TBCA).................................... 14-74 Transmit Pause Control Frame Counter Register (TXPF) .............................. 14-75 Transmit Deferral Packet Counter Register (TDFR)....................................... 14-75 Transmit Excessive Deferral Packet Counter Register (TEDF) ...................... 14-76 Transmit Single Collision Packet Counter Register (TSCL)........................... 14-76 Transmit Multiple Collision Packet Counter Register (TMCL)...................... 14-77 Transmit Late Collision Packet Counter Register (TLCL).............................. 14-77 Transmit Excessive Collision Packet Counter Register (TXCL) .................... 14-78 Transmit Total Collision Counter Register (TNCL)........................................ 14-78 Transmit Drop Frame Counter Register (TDRP) ............................................ 14-79 Transmit Jabber Frame Counter Register (TJBR) ........................................... 14-79 Transmit FCS Error Counter Register (TFCS) ................................................ 14-80 Transmit Control Frame Counter Register (TXCF) ........................................ 14-80 Transmit Oversize Frame Counter Register (TOVR)...................................... 14-81 Transmit Undersize Frame Counter Register (TUND) ................................... 14-81 Transmit Fragment Counter Register (TFRG) ................................................ 14-82 Carry Register 1 (CAR1) ................................................................................. 14-82 Carry Register 2 (CAR2) ................................................................................. 14-84 Carry Mask Register 1 (CAM1) ...................................................................... 14-85 Carry Mask Register 2 (CAM2) ...................................................................... 14-86 Hash Function Registers ...................................................................................... 14-87 Individual Address Registers 0–7 (IADDRn) ................................................. 14-88 Group Address Registers 0–7 (GADDRn) ...................................................... 14-88 Attribute Registers ............................................................................................... 14-89 Attribute Register (ATTR)............................................................................... 14-89 Attribute Extract Length and Extract Index Register (ATTRELI) .................. 14-90 Ten-Bit Interface (TBI) ............................................................................................ 14-91 TBI MII Set Register Descriptions ...................................................................... 14-91 Control Register (CR).......................................................................................... 14-92 Status Register (SR) ........................................................................................... 14-93 AN Advertisement Register (ANA) .................................................................... 14-94 AN Link Partner Base Page Ability Register (ANLPBPA)................................. 14-96 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor xxv Contents Paragraph Number 14.5.4.6 14.5.4.7 14.5.4.8 14.5.4.9 14.5.4.10 14.5.4.11 14.6 14.6.1 14.6.1.1 14.6.1.2 14.6.1.3 14.6.1.4 14.6.1.5 14.6.2 14.6.2.1 14.6.2.1.1 14.6.2.1.2 14.6.2.2 14.6.2.3 14.6.2.4 14.6.2.5 14.6.2.6 14.6.2.6.1 14.6.2.6.2 14.6.2.6.3 14.6.2.7 14.6.2.8 14.6.2.8.1 14.6.2.8.2 14.6.2.8.3 14.6.2.9 14.6.2.10 14.6.2.11 14.6.3 14.6.3.1 14.6.3.2 14.6.4 14.7 14.7.1 14.7.1.1 Title Page Number AN Expansion Register (ANEX)......................................................................... 14-98 AN Next Page Transmit Register (ANNPT)........................................................ 14-98 AN Link Partner Ability Next Page Register (ANLPANP) ................................ 14-99 Extended Status Register (EXST) ..................................................................... 14-100 Jitter Diagnostics Register (JD) ......................................................................... 14-101 TBI Control Register (TBICON)....................................................................... 14-102 Functional Description............................................................................................... 14-103 Connecting to Physical Interfaces.......................................................................... 14-103 Media-Independent Interface (MII) ................................................................... 14-104 Gigabit Media-Independent Interface (GMII) ................................................... 14-104 Reduced Gigabit Media-Independent Interface (RGMII) ................................. 14-105 Ten-Bit Interface (TBI)...................................................................................... 14-106 Reduced Ten-Bit Interface (RTBI) .................................................................... 14-107 Gigabit Ethernet Channel Operation.......................................................................14-111 Initialization Sequence........................................................................................14-111 Hardware-Controlled Initialization.................................................................14-111 User Initialization ...........................................................................................14-111 Soft Reset and Reconfiguring Procedure........................................................... 14-112 Gigabit Ethernet Frame Transmission ............................................................... 14-113 Gigabit Ethernet Frame Reception .................................................................... 14-115 RMON Support.................................................................................................. 14-116 Frame Recognition............................................................................................. 14-116 Destination Address Recognition .................................................................. 14-116 Hash Table Algorithm.................................................................................... 14-118 CRC Computation Examples ........................................................................ 14-118 Flow Control...................................................................................................... 14-119 Interrupt Handling ............................................................................................. 14-120 Interrupt Coalescing ...................................................................................... 14-121 Interrupt Coalescing By Frame Count Threshold.......................................... 14-122 Interrupt Coalescing By Timer Threshold ..................................................... 14-122 Inter-Packet Gap Time....................................................................................... 14-123 Internal and External Loop Back....................................................................... 14-123 Error-Handling Procedure.................................................................................. 14-123 Buffer Descriptors.................................................................................................. 14-125 Transmit Data Buffer Descriptor (TxBD).......................................................... 14-126 Receive Buffer Descriptor (RxBD) ................................................................... 14-128 Data Extraction to the L2 Cache............................................................................ 14-130 Initialization/Application Information ....................................................................... 14-131 Interface Mode Configuration ............................................................................... 14-131 MII Interface Mode............................................................................................ 14-131 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 xxvi Freescale Semiconductor Contents Paragraph Number 14.7.1.2 14.7.1.3 14.7.1.4 14.7.1.5 Title Page Number GMII Interface Mode......................................................................................... 14-135 TBI Interface Mode ........................................................................................... 14-138 RGMII Interface Mode ...................................................................................... 14-143 RTBI Interface Mode......................................................................................... 14-147 Chapter 15 DMA Controller 15.1 15.1.1 15.1.2 15.1.3 15.1.4 15.2 15.2.1 15.2.2 15.3 15.3.1 15.3.2 15.3.2.1 15.3.2.2 15.3.2.3 15.3.2.4 15.3.2.5 15.3.2.5.1 15.3.2.6 15.3.2.7 15.3.2.7.1 15.3.2.8 15.3.2.9 15.3.2.10 15.3.2.11 15.3.2.12 15.3.2.13 15.3.2.14 15.4 15.4.1 15.4.1.1 Introduction.................................................................................................................... 15-1 Block Diagram........................................................................................................... 15-1 Overview.................................................................................................................... 15-1 Features...................................................................................................................... 15-2 Modes of Operation ................................................................................................... 15-2 External Signal Descriptions ......................................................................................... 15-5 Signal Overview ........................................................................................................ 15-5 Detailed Signal Descriptions ..................................................................................... 15-6 Memory Map/Register Definition ................................................................................. 15-6 Module Memory Map................................................................................................ 15-7 DMA Register Descriptions..................................................................................... 15-10 Mode Registers (MRn) ........................................................................................ 15-10 Status Registers (SRn) ......................................................................................... 15-13 Current Link Descriptor Address Registers (CLNDARn)................................... 15-14 Source Attributes Registers (SATRn).................................................................. 15-16 Source Address Registers (SARn)....................................................................... 15-18 Source Address Registers for RapidIO Maintenance Reads (SARn)......................................................................................................... 15-19 Destination Attributes Registers (DATRn).......................................................... 15-19 Destination Address Registers (DARn)............................................................... 15-21 Destination Address Registers for RapidIO Maintenance Writes (DARn) ........................................................................................................ 15-22 Byte Count Registers (BCRn) ............................................................................. 15-23 Next Link Descriptor Address Registers (NLNDARn)....................................... 15-23 Current List Descriptor Address Registers (CLSDARn)..................................... 15-24 Next List Descriptor Address Registers (NLSDARn)......................................... 15-25 Source Stride Registers (SSRn) ........................................................................... 15-25 Destination Stride Registers (DSRn) ................................................................... 15-26 DMA General Status Register (DGSR) ............................................................... 15-26 Functional Description................................................................................................. 15-28 DMA Channel Operation......................................................................................... 15-28 Basic DMA Mode Transfer ................................................................................. 15-29 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor xxvii Contents Paragraph Number 15.4.1.1.1 15.4.1.1.2 15.4.1.1.3 15.4.1.1.4 15.4.1.2 15.4.1.2.1 15.4.1.2.2 15.4.1.2.3 15.4.1.2.4 15.4.1.3 15.4.1.4 15.4.1.4.1 15.4.1.4.2 15.4.1.5 15.4.1.6 15.4.1.7 15.4.1.8 15.4.2 15.4.3 15.4.4 15.4.5 15.5 15.5.1 15.5.1.1 15.5.1.2 15.5.1.3 15.5.1.4 15.5.1.5 Title Page Number Basic Direct Mode ........................................................................................... 15-29 Basic Direct Single-Write Start Mode ............................................................. 15-30 Basic Chaining Mode ...................................................................................... 15-31 Basic Chaining Single-Write Start Mode ........................................................ 15-31 Extended DMA Mode Transfer ........................................................................... 15-32 Extended Direct Mode..................................................................................... 15-32 Extended Direct Single-Write Start Mode....................................................... 15-32 Extended Chaining Mode ................................................................................ 15-32 Extended Chaining Single-Write Start Mode .................................................. 15-33 External Control Mode Transfer.......................................................................... 15-33 Channel Continue Mode for Cascading Transfer Chains .................................... 15-34 Basic Mode ...................................................................................................... 15-35 Extended Mode................................................................................................ 15-35 Channel Abort...................................................................................................... 15-35 Bandwidth Control............................................................................................... 15-36 Channel State ....................................................................................................... 15-36 Illustration of Stride Size and Stride Distance..................................................... 15-36 DMA Transfer Interfaces ......................................................................................... 15-37 DMA Errors ............................................................................................................. 15-37 DMA Descriptors..................................................................................................... 15-37 Limitations and Restrictions .................................................................................... 15-40 DMA System Considerations ...................................................................................... 15-41 Unusual DMA Scenarios ......................................................................................... 15-43 DMA to e500 Core .............................................................................................. 15-43 DMA to Ethernet ................................................................................................. 15-44 DMA to Configuration and Control Registers..................................................... 15-44 DMA to I2C ......................................................................................................... 15-44 DMA to DUART ................................................................................................. 15-44 Chapter 16 PCI/PCI-X Bus Interface 16.1 16.1.1 16.1.1.1 16.1.1.2 16.1.2 16.1.3 16.1.3.1 16.1.3.1.1 Introduction.................................................................................................................... 16-1 Overview.................................................................................................................... 16-2 MPC8540 as a PCI/X Initiator............................................................................... 16-3 MPC8540 as a PCI/X Target ................................................................................. 16-4 Features...................................................................................................................... 16-4 Modes of Operation ................................................................................................... 16-5 Host/Agent Modes ................................................................................................. 16-5 Host Mode ......................................................................................................... 16-6 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 xxviii Freescale Semiconductor Contents Paragraph Number 16.1.3.1.2 16.1.3.1.3 16.1.3.2 16.1.3.3 16.1.3.4 16.1.3.5 16.1.3.6 16.2 16.3 16.3.1 16.3.1.1 16.3.1.1.1 16.3.1.1.2 16.3.1.1.3 16.3.1.2 16.3.1.2.1 16.3.1.2.2 16.3.1.2.3 16.3.1.2.4 16.3.1.3 16.3.1.3.1 16.3.1.3.2 16.3.1.3.3 16.3.1.3.4 16.3.1.4 16.3.1.4.1 16.3.1.4.2 16.3.1.4.3 16.3.1.4.4 16.3.1.4.5 16.3.1.4.6 16.3.1.4.7 16.3.1.4.8 16.3.1.4.9 16.3.1.4.10 16.3.2 16.3.2.1 Title Page Number Agent Mode ....................................................................................................... 16-6 Agent Configuration Lock Mode ...................................................................... 16-6 PCI/X Bus Width (64-/32-Bit Bus) ....................................................................... 16-6 PCI/X Arbiter (Internal/External Arbiter) ............................................................. 16-6 PCI/X Bus Mode.................................................................................................... 16-6 PCI/X Signal Output Hold Timing ........................................................................ 16-6 PCI/X Impedance................................................................................................... 16-7 External Signal Descriptions ......................................................................................... 16-7 Memory Map/Register Definitions .............................................................................. 16-15 PCI/X Memory Mapped Registers .......................................................................... 16-15 PCI/X Configuration Access Registers ............................................................... 16-18 PCI/X Configuration Address Register (CFG_ADDR) .................................. 16-18 PCI/X Configuration Data Register (CFG_DATA) ......................................... 16-19 PCI/X Interrupt Acknowledge Register (INT_ACK)...................................... 16-20 PCI/X ATMU Outbound Registers...................................................................... 16-20 PCI/X Outbound Translation Address Registers (POTARn) .......................... 16-21 PCI/X Outbound Translation Extended Address Registers (POTEARn) ................................................................................................. 16-21 PCI/X Outbound Window Base Address Registers (POWBARn).................. 16-22 PCI/X Outbound Window Attributes Registers (POWARn)........................... 16-22 PCI/X ATMU Inbound Registers......................................................................... 16-24 PCI/X Inbound Translation Address Registers (PITARn)............................... 16-25 PCI/X Inbound Window Base Address Registers (PIWBARn) ...................... 16-25 PCI/X Inbound Window Base Extended Address Registers (PIWBEARn) .............................................................................................. 16-26 PCI/X Inbound Window Attributes Registers (PIWARn) ............................... 16-26 PCI/X Error Management Registers.................................................................... 16-28 PCI/X Error Detect Register (ERR_DR)......................................................... 16-29 PCI/X Error Capture Disable Register (ERR_CAP_DR)................................ 16-30 PCI/X Error Enable Register (ERR_EN) ........................................................ 16-31 PCI/X Error Attributes Capture Register (ERR_ATTRIB) ............................. 16-32 PCI/X Error Address Capture Register (ERR_ADDR)................................... 16-33 PCI/X Error Extended Address Capture Register (ERR_EXT_ADDR).................................................................................... 16-34 PCI/X Error Data Low Capture Register (ERR_DL) ...................................... 16-34 PCI/X Error Data High Capture Register (ERR_DH)..................................... 16-34 PCI/X Gasket Timer Register (GAS_TIMR) .................................................. 16-35 PCI-X Split Completion Timer Register (PCIX_TIMR)................................. 16-36 PCI/X Configuration Header ................................................................................... 16-36 PCI Vendor ID Register—Offset 0x00 ................................................................ 16-37 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor xxix Contents Paragraph Number 16.3.2.2 16.3.2.3 16.3.2.4 16.3.2.5 16.3.2.6 16.3.2.7 16.3.2.8 16.3.2.9 16.3.2.10 16.3.2.11 16.3.2.12 16.3.2.13 16.3.2.14 16.3.2.15 16.3.2.16 16.3.2.17 16.3.2.18 16.3.2.19 16.3.2.20 16.3.2.21 16.3.2.22 16.3.2.23 16.3.2.24 16.4 16.4.1 16.4.1.1 16.4.1.2 16.4.1.3 16.4.1.4 16.4.2 16.4.2.1 16.4.2.2 16.4.2.3 16.4.2.3.1 16.4.2.3.2 16.4.2.3.3 16.4.2.4 16.4.2.5 16.4.2.6 16.4.2.7 Title Page Number PCI Device ID Register—Offset 0x02 ................................................................ 16-38 PCI Bus Command Register—Offset 0x04 ......................................................... 16-38 PCI Bus Status Register—Offset 0x06................................................................ 16-40 PCI Revision ID Register—Offset 0x08 ............................................................. 16-41 PCI Bus Programming Interface Register—Offset 0x09 .................................... 16-41 PCI Subclass Code Register—Offset 0x0A......................................................... 16-42 PCI Bus Base Class Code Register—Offset 0x0B .............................................. 16-42 PCI Bus Cache Line Size Register—Offset 0x0C............................................... 16-43 PCI Bus Latency Timer Register—0x0D ............................................................ 16-43 PCI Base Address Registers ................................................................................ 16-43 PCI Subsystem Vendor ID Register..................................................................... 16-46 PCI Subsystem ID Register ................................................................................. 16-46 PCI Bus Capabilities Pointer Register ................................................................. 16-47 PCI Bus Interrupt Line Register .......................................................................... 16-47 PCI Bus Interrupt Pin Register ............................................................................ 16-48 PCI Bus Minimum Grant (MIN_GNT) Register................................................. 16-48 PCI Bus Maximum Latency (MAX_LAT) Register............................................ 16-49 PCI Bus Function Register (PBFR) ..................................................................... 16-49 PCI Bus Arbiter Configuration Register (PBACR)............................................. 16-50 PCI-X Next Capabilities ID Register—0x60 ...................................................... 16-51 PCI-X Capability Pointer Register—0x61 .......................................................... 16-51 PCI-X Command Register—0x62 ....................................................................... 16-52 PCI-X Status Register—0x64.............................................................................. 16-52 Functional Description................................................................................................. 16-53 PCI/X Bus Arbitration ............................................................................................. 16-53 PCI/X Bus Arbiter Operation .............................................................................. 16-54 PCI/X Bus Parking .............................................................................................. 16-56 Broken Master Lock-Out (PCI only) ................................................................... 16-56 Power-Saving Modes and the PCI Arbiter .......................................................... 16-56 PCI Bus Protocol ..................................................................................................... 16-56 Basic Transfer Control......................................................................................... 16-57 PCI Bus Commands............................................................................................. 16-57 Addressing ........................................................................................................... 16-59 Memory Space Addressing.............................................................................. 16-59 I/O Space Addressing ...................................................................................... 16-60 Configuration Space Addressing ..................................................................... 16-60 Device Selection .................................................................................................. 16-60 Byte Alignment.................................................................................................... 16-61 Bus Driving and Turnaround ............................................................................... 16-61 PCI Bus Transactions........................................................................................... 16-61 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 xxx Freescale Semiconductor Contents Paragraph Number 16.4.2.7.1 16.4.2.7.2 16.4.2.8 16.4.2.8.1 16.4.2.8.2 16.4.2.9 16.4.2.10 16.4.2.11 16.4.2.11.1 16.4.2.11.2 16.4.2.11.3 16.4.2.11.4 16.4.2.11.5 16.4.2.12 16.4.2.12.1 16.4.2.12.2 16.4.2.13 16.4.2.13.1 16.4.2.13.2 16.4.3 16.4.3.1 16.4.3.2 16.4.3.3 16.4.3.4 16.4.3.5 16.4.3.6 16.4.3.6.1 16.4.3.6.2 16.4.3.6.3 16.4.3.6.4 16.4.3.7 16.4.3.8 16.4.3.8.1 16.5 16.5.1 16.5.1.1 16.5.1.2 16.5.1.3 16.5.2 16.5.3 Title Page Number PCI Read Transactions .................................................................................... 16-62 PCI Write Transactions.................................................................................... 16-63 Transaction Termination ...................................................................................... 16-64 Master-Initiated Termination ........................................................................... 16-65 Target-Initiated Termination ............................................................................ 16-65 Fast Back-to-Back Transactions .......................................................................... 16-68 Dual Address Cycles............................................................................................ 16-68 Configuration Cycles ........................................................................................... 16-70 PCI Configuration Space Header .................................................................... 16-71 Accessing the PCI Configuration Space in Host Mode................................... 16-72 PCI Configuration in Agent and Agent Lock Modes ...................................... 16-74 PCI Type 0 Configuration Translation............................................................. 16-74 Type 1 Configuration Translation.................................................................... 16-76 Other Bus Transactions........................................................................................ 16-76 Interrupt-Acknowledge Transactions .............................................................. 16-76 Special-Cycle Transactions ............................................................................. 16-77 PCI Error Functions............................................................................................. 16-78 PCI Parity ........................................................................................................ 16-78 Error Reporting................................................................................................ 16-79 PCI-X Bus Protocol ................................................................................................. 16-81 PCI-X Terminology ............................................................................................. 16-81 PCI-X Command Encodings ............................................................................... 16-82 PCI-X Attribute Phase ......................................................................................... 16-83 PCI-X Transactions.............................................................................................. 16-84 PCI-X Wait State and Termination Rules ............................................................ 16-88 PCI-X Split Transactions ..................................................................................... 16-89 Split Response ................................................................................................. 16-89 Completion Address ........................................................................................ 16-90 Completer Attributes ....................................................................................... 16-91 Split Completion Messages ............................................................................. 16-92 PCI-X Configuration Transactions ...................................................................... 16-93 PCI-X Error Functions......................................................................................... 16-96 Error Reporting................................................................................................ 16-96 Initialization/Application Information ......................................................................... 16-99 Power-On Reset Configuration Modes.................................................................... 16-99 Host Mode ......................................................................................................... 16-100 Agent Mode ....................................................................................................... 16-100 Agent Configuration Lock Mode....................................................................... 16-100 Nonposted Writes in PCI-X ................................................................................... 16-100 PCI-X Outbound Read Transaction Alignment ..................................................... 16-101 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor xxxi Contents Paragraph Number 16.5.4 16.5.5 Title Page Number PCI-X Inbound Reads that Cross Window Boundaries ......................................... 16-101 Bridging Between RapidIO and PCI/X ................................................................. 16-101 Chapter 17 RapidIO Interface 17.1 17.1.1 17.1.2 17.1.3 17.1.3.1 17.1.3.2 17.1.4 17.1.4.1 17.1.4.2 17.1.4.3 17.1.4.4 17.1.4.5 17.1.4.6 17.1.4.7 17.1.4.8 17.2 17.2.1 17.2.2 17.3 17.3.1 17.3.1.1 17.3.1.2 17.3.1.3 17.3.1.4 17.3.1.5 17.3.1.6 17.3.1.7 17.3.1.8 17.3.1.9 17.3.1.10 17.3.1.11 17.3.1.12 Introduction.................................................................................................................... 17-1 Overview.................................................................................................................... 17-1 RapidIO Terminology ................................................................................................ 17-4 RapidIO Interface Features........................................................................................ 17-4 Supported Features ................................................................................................ 17-4 Features Not Implemented..................................................................................... 17-5 RapidIO Modes of Operation .................................................................................... 17-5 Transmit Clock-Select Mode ................................................................................. 17-5 CRC Checking Modes ........................................................................................... 17-5 Error Checking Disable Mode ............................................................................... 17-6 Output Port Enable Mode ...................................................................................... 17-6 Output Port Driver Disable Mode.......................................................................... 17-6 Accept All Mode.................................................................................................... 17-6 Packet Response Time-Out Disable Mode ............................................................ 17-6 Link Response Time-Out Disable Mode ............................................................... 17-6 External Signal Descriptions ......................................................................................... 17-7 Signals Overview....................................................................................................... 17-7 Detailed Signal Descriptions ..................................................................................... 17-8 Memory Map/Register Definition ................................................................................. 17-9 Architectural Registers ............................................................................................ 17-13 Device Identity Capability Register (DIDCAR).................................................. 17-13 Device Information Capability Register (DICAR).............................................. 17-14 Assembly Identity Capability Register (AIDCAR) ............................................. 17-14 Assembly Information Capability Register (AICAR) ......................................... 17-15 Processing Element Features Capability Register (PEFCAR) ............................ 17-15 Switch Port Information Capability Register (SPICAR)..................................... 17-17 Source Operations Capability Register (SOCAR)............................................... 17-17 Destination Operations Capability Register (DOCAR)....................................... 17-19 Mailbox Command and Status Register (MSR) .................................................. 17-22 Port-Write and Doorbell Command and Status Register (PWDCSR)................. 17-22 Processing Element Logic Layer Control Command and Status Register (PELLCCSR) .................................................................................... 17-24 Local Configuration Space Base Address 1 Command and Status Register (LCSBA1CSR).................................................................................. 17-24 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 xxxii Freescale Semiconductor Contents Paragraph Number 17.3.1.13 17.3.1.14 17.3.1.15 17.3.1.16 17.3.1.17 17.3.1.18 17.3.1.19 17.3.1.20 17.3.1.21 17.3.1.22 17.3.1.23 17.3.1.24 17.3.2 17.3.2.1 17.3.2.1.1 17.3.2.1.2 17.3.2.1.3 17.3.2.2 17.3.2.2.1 17.3.2.2.2 17.3.2.2.3 17.3.2.2.4 17.3.2.2.5 17.3.2.2.6 17.3.2.3 17.3.2.3.1 17.3.2.3.2 17.3.2.3.3 Title Page Number Base Device ID Command and Status Register (BDIDCSR) ............................. 17-25 Host Base Device ID Lock Command and Status Register (HBDIDLCSR)................................................................................................ 17-26 Component Tag Command and Status Register (CTCSR).................................. 17-26 8/16 LP-LVDS Port Maintenance Block Header 0 Command and Status Register (PMBH0CSR)......................................................................... 17-27 Port Link Time-Out Control Command and Status Register (PLTOCCSR)................................................................................................... 17-27 Port Response Time-Out Control Command and Status Register (PRTOCCSR) .................................................................................................. 17-28 Port General Control Command and Status Register (PGCCSR) ....................... 17-29 Port Link Maintenance Request Command and Status Register (PLMREQCSR)............................................................................................... 17-29 Port Link Maintenance Response Command and Status Register (PLMRESPCSR) ............................................................................................. 17-30 Port Local AckID Status Command and Status Register (PLASCSR) ............... 17-31 Port Error and Status Command and Status Register (PESCSR) ........................ 17-31 Port Control Command and Status Register (PCCSR)........................................ 17-33 Implementation Registers ........................................................................................ 17-34 General Registers................................................................................................. 17-34 Configuration Register (CR) ........................................................................... 17-34 Port Configuration Register (PCR).................................................................. 17-35 Port Error Injection Register (PEIR) ............................................................... 17-35 ATMU Registers .................................................................................................. 17-36 RapidIO Outbound Window Translation Address Registers 0–8 (ROWTARn) ............................................................................................... 17-37 RapidIO Outbound Window Base Address Registers 1–8 (ROWBARn) ............................................................................................... 17-38 RapidIO Outbound Window Attributes Registers 0–8 (ROWARn)................ 17-39 RapidIO Inbound Window Translation Address Registers 0–4 (RIWTARn) ................................................................................................. 17-40 RapidIO Inbound Window Base Address Registers 1–4 (RIWBARn) ................................................................................................ 17-41 RapidIO Inbound Window Attributes Registers 0–4 (RIWARn) .................... 17-42 Error Management Registers ............................................................................... 17-43 Port Notification/Fatal Error Detect Register (PNFEDR) ............................... 17-44 Port Notification/Fatal Error Detect Disable Register (PNFEDiR) .................................................................................................. 17-47 Port Notification/Fatal Error Interrupt Enable Register (PNFEIER) .................................................................................................. 17-49 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor xxxiii Contents Paragraph Number 17.3.2.3.4 17.3.2.3.5 17.3.2.3.6 17.3.2.3.7 17.3.2.3.8 17.3.2.3.9 17.3.2.3.10 17.3.2.3.11 17.3.2.3.12 17.3.2.3.13 17.3.2.3.14 17.3.2.3.15 17.3.2.3.16 17.3.2.3.17 17.3.2.3.18 17.3.2.3.19 17.3.2.3.20 17.3.2.3.21 17.3.2.3.22 17.3.2.3.23 17.3.2.3.24 17.3.2.3.25 17.3.3 17.3.3.1 17.3.3.1.1 17.3.3.1.2 17.3.3.1.3 17.3.3.1.4 17.3.3.1.5 17.3.3.1.6 17.3.3.1.7 17.3.3.1.8 17.3.3.2 17.3.3.2.1 17.3.3.2.2 17.3.3.2.3 Title Page Number Port Error Capture Status Register (PECSR) .................................................. 17-52 Error Packet Capture Register 0 (EPCR0) ...................................................... 17-52 Error Packet Capture Register 1 (EPCR1) ...................................................... 17-53 EPCR1—Type 1 Packet Format ...................................................................... 17-53 EPCR1—Type 2 Packet Format ...................................................................... 17-54 EPCR1—Type 5 Packet Format ...................................................................... 17-54 EPCR1—Type 6 Packet Format ...................................................................... 17-55 EPCR1—Type 8 Request Packet Format ........................................................ 17-55 EPCR1—Type 8 Response Packet Format...................................................... 17-56 EPCR1—Type 10 Packet Format .................................................................... 17-56 EPCR1—Type 11 Packet Format .................................................................... 17-57 EPCR1—Type 13 Packet Format .................................................................... 17-58 Error Packet Capture Register 2 (EPCR2) ...................................................... 17-58 EPCR2—Type 1, 2, or 5 Packet Format.......................................................... 17-58 EPCR2—Type 6 Packet Format ...................................................................... 17-59 EPCR2—Type 8 Request Packet Format ........................................................ 17-59 EPCR2—Type 8 Response Packet Format...................................................... 17-60 EPCR2—Type 10 Packet Format .................................................................... 17-60 EPCR2—Type 11 or 13 Packet Format ........................................................... 17-61 Port Recoverable Error Detect Register (PREDR).......................................... 17-61 Port Error Recovery Threshold Register (PERTR) ......................................... 17-64 Port Retry Threshold Register (PRTR)............................................................ 17-64 RapidIO Message Unit Registers............................................................................. 17-65 RapidIO Outbound Message Registers................................................................ 17-65 Outbound Mode Register (OMR).................................................................... 17-65 Outbound Status Register (OSR)..................................................................... 17-67 Outbound Descriptor Queue Dequeue Pointer Address Register (ODQDPAR)................................................................................................ 17-68 Outbound Unit Source Address Register (OSAR) .......................................... 17-69 Outbound Destination Port Register (ODPR) ................................................. 17-70 Outbound Destination Attributes Register (ODATR) ..................................... 17-70 Outbound Double-Word Count Register (ODCR) .......................................... 17-71 Outbound Descriptor Queue Enqueue Pointer Address Register (ODQEPAR) ................................................................................................ 17-72 RapidIO Inbound Message Registers .................................................................. 17-72 Inbound Mode Register (IMR) ........................................................................ 17-72 Inbound Status Register (ISR) ......................................................................... 17-74 Inbound Frame Queue Dequeue Pointer Address Register (IFQDPAR).................................................................................................. 17-75 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 xxxiv Freescale Semiconductor Contents Paragraph Number 17.3.3.2.4 17.3.3.3 17.3.3.3.1 17.3.3.3.2 17.3.3.3.3 17.3.3.3.4 17.3.3.4 17.3.3.4.1 17.3.3.4.2 17.3.3.4.3 17.4 17.4.1 17.5 17.5.1 17.5.1.1 17.5.2 17.5.3 17.5.3.1 17.5.3.2 17.6 17.7 17.7.1 17.7.1.1 17.7.1.2 17.7.2 17.7.2.1 17.7.2.2 17.7.2.3 17.7.2.4 17.8 17.8.1 17.8.2 17.8.3 17.8.4 17.8.4.1 17.8.4.2 17.8.4.2.1 17.8.4.2.2 17.8.4.2.3 Title Page Number Inbound Frame Queue Enqueue Pointer Address Register (IFQEPAR) .................................................................................................. 17-76 RapidIO Doorbell Registers ................................................................................ 17-77 Doorbell Mode Register (DMR)...................................................................... 17-77 Doorbell Status Register (DSR) ...................................................................... 17-78 Doorbell Queue/Dequeue Pointer Address Register (DQDPAR).................... 17-79 Doorbell Queue Enqueue Pointer Address Register (DQEPAR) .................... 17-80 RapidIO Port-Write Registers.............................................................................. 17-81 Port-Write Mode Register (PWMR)................................................................ 17-81 Port-Write Status Register (PWSR)................................................................. 17-82 Port-Write Queue Base Address Register (PWQBAR) ................................... 17-83 Functional Description................................................................................................. 17-83 RapidIO Transaction, Packet, and Control Symbol Summary ................................ 17-83 RapidIO Functionality ................................................................................................. 17-89 General Functionality Lists...................................................................................... 17-89 8/16 LP-LVDS Layer Functionality Lists............................................................ 17-91 Common Transport Layer Functionality Lists......................................................... 17-98 Logical Layer Functionality Lists.......................................................................... 17-100 Logical Layer Source Transaction Support List ................................................ 17-101 Logical Layer Extended Functionality .............................................................. 17-101 RapidIO Errors........................................................................................................... 17-102 ATMU (Address Translation and Mapping Unit) ...................................................... 17-106 Outbound ATMU Translation ................................................................................ 17-106 Outbound ATMU Bypass Mode ........................................................................ 17-107 Outbound Special Transactions and Requirements ........................................... 17-107 Inbound ATMU Translation................................................................................... 17-107 Inbound ATMU LCSBA1CSR Window............................................................ 17-108 Inbound ATMU Bypass Mode........................................................................... 17-108 Inbound Special Transactions and Requirements .............................................. 17-108 ATMU Boundary Crossing Errors ..................................................................... 17-109 RapidIO Message Unit............................................................................................... 17-109 Overview................................................................................................................ 17-109 Message Unit Features........................................................................................... 17-110 Message Unit Modes of Operation .........................................................................17-111 RapidIO Messaging Description.............................................................................17-111 Data Message Controller ....................................................................................17-111 Outbox Controller Operation..............................................................................17-111 Direct Mode Operation.................................................................................. 17-112 Chaining Mode Operation ............................................................................. 17-112 Switching Between Direct and Chaining Modes........................................... 17-115 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor xxxv Contents Paragraph Number 17.8.4.2.4 17.8.4.2.5 17.8.4.2.6 17.8.4.3 17.8.4.3.1 17.8.4.3.2 17.8.4.3.3 17.8.4.3.4 17.8.4.4 17.8.4.4.1 17.8.4.4.2 17.8.4.4.3 17.8.4.4.4 17.8.4.4.5 17.8.4.5 17.9 Title Page Number Descriptor Format.......................................................................................... 17-115 Outbox Controller Interrupts ......................................................................... 17-116 Special Error Case Condition ........................................................................ 17-117 Inbox Controller Operation................................................................................ 17-117 Retry Response Conditions ........................................................................... 17-118 Error Response Conditions............................................................................ 17-118 Inbox Controller Interrupts............................................................................ 17-118 Data Message Controller Limitations and Restrictions ................................. 17-119 Doorbell Message Controller............................................................................. 17-119 Inbound Doorbell Reception ......................................................................... 17-120 Doorbell Queue Entry Format ....................................................................... 17-120 Retry Response Conditions ........................................................................... 17-121 Error Response Conditions............................................................................ 17-121 Doorbell Controller Interrupts ....................................................................... 17-121 Port-Write Controller Structure ......................................................................... 17-121 Initialization and Application Information ................................................................ 17-122 Part IV Global Functions and Debug Chapter 18 Global Utilities 18.1 18.2 18.2.1 18.2.2 18.2.3 18.2.4 18.2.5 18.3 18.3.1 18.3.2 18.4 18.4.1 18.4.1.1 18.4.1.2 18.4.1.3 18.4.1.4 18.4.1.5 Overview........................................................................................................................ 18-1 Global Utilities Features ................................................................................................ 18-1 Power Management and Block Disables ................................................................... 18-1 Accessing Current POR Configuration Settings........................................................ 18-1 General-Purpose I/O .................................................................................................. 18-1 Interrupt and Local Bus Signal Multiplexing ............................................................ 18-2 Clock Control............................................................................................................. 18-2 External Signal Descriptions ......................................................................................... 18-2 Signals Overview....................................................................................................... 18-2 Detailed Signal Descriptions ..................................................................................... 18-2 Memory Map/Register Definition ................................................................................. 18-3 Register Descriptions................................................................................................. 18-4 POR PLL Status Register (PORPLLSR) ............................................................... 18-4 POR Boot Mode Status Register (PORBMSR)..................................................... 18-5 POR I/O Impedance Status and Control Register (PORIMPSCR) ....................... 18-6 POR Device Status Register (PORDEVSR).......................................................... 18-7 POR Debug Mode Status Register (PORDBGMSR) ............................................ 18-9 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 xxxvi Freescale Semiconductor Contents Paragraph Number 18.4.1.6 18.4.1.7 18.4.1.8 18.4.1.9 18.4.1.10 18.4.1.11 18.4.1.12 18.4.1.13 18.4.1.14 18.4.1.15 18.4.1.16 18.4.1.17 18.4.1.18 18.5 18.5.1 18.5.1.1 18.5.1.2 18.5.1.3 18.5.1.4 18.5.1.5 18.5.1.5.1 18.5.1.5.2 18.5.1.5.3 18.5.1.6 18.5.1.7 18.5.1.8 18.5.1.8.1 18.5.1.8.2 18.5.1.9 18.5.1.10 18.5.1.11 18.5.2 18.5.3 Title Page Number General-Purpose POR Configuration Register (GPPORCR) ................................ 18-9 General-Purpose I/O Control Register (GPIOCR) .............................................. 18-10 General-Purpose Output Data Register (GPOUTDR) ......................................... 18-11 General-Purpose Input Data Register (GPINDR)................................................ 18-12 Alternate Function Signal Multiplex Control Register (PMUXCR) ................... 18-12 Device Disable Register (DEVDISR) ................................................................. 18-13 Power Management Control and Status Register (POWMGTCSR) ................... 18-15 Machine Check Summary Register (MCPSUMR).............................................. 18-16 Processor Version Register (PVR)....................................................................... 18-17 System Version Register (SVR)........................................................................... 18-18 Clock Out Control Register (CLKOCR) ............................................................. 18-18 DDR DLL Control Register (DDRDLLCR) ....................................................... 18-19 Local Bus DLL Control Register (LBDLLCR)................................................... 18-20 Functional Description................................................................................................. 18-21 Power Management ................................................................................................. 18-21 Relationship Between Core and Device Power Management States................... 18-21 CKSTP_IN Is Not Power Management............................................................... 18-22 Dynamic Power Management.............................................................................. 18-22 Shutting Down Unused Blocks............................................................................ 18-22 Software-Controlled Power-Down States............................................................ 18-23 Doze Mode ...................................................................................................... 18-23 Nap Mode ........................................................................................................ 18-23 Sleep Mode ...................................................................................................... 18-24 Power Management Control Fields ..................................................................... 18-24 Power-Down Sequence Coordination.................................................................. 18-24 Interrupts and Power Management ...................................................................... 18-27 Interrupts and Power Management Controlled by MSR[WE] ........................ 18-27 Interrupts and Power Management Controlled by POWMGTCSR ................ 18-27 Snooping in Power-Down Modes........................................................................ 18-28 Software Considerations for Power Management ............................................... 18-28 Requirements for Reaching and Recovering from Sleep State ........................... 18-28 General-Purpose I/O Signals ................................................................................... 18-29 Interrupt and Local Bus Signal Multiplexing .......................................................... 18-29 Chapter 19 Performance Monitor 19.1 19.1.1 19.1.2 Introduction.................................................................................................................... 19-1 Overview.................................................................................................................... 19-1 Features...................................................................................................................... 19-3 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor xxxvii Contents Paragraph Number 19.2 19.3 19.3.1 19.3.2 19.3.2.1 19.3.2.2 19.3.3 19.3.3.1 19.4 19.4.1 19.4.2 19.4.3 19.4.4 19.4.5 19.4.6 19.4.7 19.4.8 Title Page Number External Signal Descriptions ......................................................................................... 19-3 Memory Map and Register Definition........................................................................... 19-3 Register Summary...................................................................................................... 19-3 Control Registers ....................................................................................................... 19-5 Performance Monitor Global Control Register (PMGC0) .................................... 19-5 Performance Monitor Local Control Registers (PMLCAn and PMLCBn) ................................................................................. 19-5 Counter Registers....................................................................................................... 19-9 Performance Monitor Counters (PMC0–PMC8)................................................... 19-9 Functional Description................................................................................................. 19-10 Performance Monitor Interrupt................................................................................ 19-10 Event Counting ........................................................................................................ 19-11 Threshold Events ..................................................................................................... 19-11 Chaining................................................................................................................... 19-12 Triggering ................................................................................................................ 19-12 Burstiness Counting................................................................................................. 19-13 Performance Monitor Events ................................................................................... 19-15 Performance Monitor Examples .............................................................................. 19-27 Chapter 20 Debug Features and Watchpoint Facility 20.1 20.1.1 20.1.2 20.1.3 20.1.3.1 20.1.3.2 20.1.3.3 20.1.3.4 20.1.3.5 20.2 20.2.1 20.2.2 20.2.2.1 20.2.2.2 20.2.2.3 20.3 20.3.1 20.3.1.1 Introduction.................................................................................................................... 20-1 Overview.................................................................................................................... 20-2 Features...................................................................................................................... 20-3 Modes of Operation ................................................................................................... 20-3 Local Bus (LBC) Debug Mode.............................................................................. 20-4 DDR SDRAM Interface Debug Modes ................................................................. 20-4 PCI/PCI-X Interface Debug Modes....................................................................... 20-5 Watchpoint Monitor Modes ................................................................................... 20-5 Trace Buffer Modes ............................................................................................... 20-5 External Signal Descriptions ......................................................................................... 20-6 Overview.................................................................................................................... 20-6 Detailed Signal Descriptions ..................................................................................... 20-7 Debug Signals—Details......................................................................................... 20-7 Watchpoint Monitor Trigger Signals—Details...................................................... 20-8 Test Signals—Details............................................................................................. 20-9 Memory Map/Register Definition ............................................................................... 20-10 Watchpoint Monitor Register Descriptions ............................................................. 20-11 Watchpoint Monitor Control Registers 0–1 (WMCR0, WMCR1)...................... 20-11 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 xxxviii Freescale Semiconductor Contents Paragraph Number 20.3.1.2 20.3.1.3 20.3.1.4 20.3.1.5 20.3.2 20.3.2.1 20.3.2.2 20.3.2.3 20.3.2.4 20.3.2.5 20.3.2.6 20.3.2.7 20.3.2.8 20.3.3 20.3.3.1 20.3.3.2 20.3.4 20.3.4.1 20.4 20.4.1 20.4.2 20.4.3 20.4.3.1 20.4.3.2 20.4.4 20.4.5 20.4.5.1 20.4.6 20.4.6.1 20.5 Title Page Number Watchpoint Monitor Address Register (WMAR)................................................ 20-13 Watchpoint Monitor Address Mask Register (WMAMR) ................................. 20-14 Watchpoint Monitor Transaction Mask Register (WMTMR) ............................. 20-14 Watchpoint Monitor Status Register (WMSR).................................................... 20-16 Trace Buffer Register Descriptions.......................................................................... 20-16 Trace Buffer Control Registers (TBCR0, TBCR1) ............................................. 20-16 Trace Buffer Address Register (TBAR) .............................................................. 20-19 Trace Buffer Address Mask Register (TBAMR)................................................. 20-19 Trace Buffer Transaction Mask Register (TBTMR)............................................ 20-20 Trace Buffer Status Register (TBSR) .................................................................. 20-21 Trace Buffer Access Control Register (TBACR) ................................................ 20-22 Trace Buffer Access Data High Register (TBADHR)......................................... 20-22 Trace Buffer Access Data Register (TBADR)..................................................... 20-23 Context ID Registers................................................................................................ 20-24 Programmed Context ID Register (PCIDR) ........................................................ 20-24 Current Context ID Register (CCIDR) ................................................................ 20-24 Trigger Out Function ............................................................................................... 20-25 Trigger Out Source Register (TOSR) .................................................................. 20-25 Functional Description................................................................................................. 20-26 Source and Target ID ............................................................................................... 20-26 PCI/PCI-X Interface Debug..................................................................................... 20-27 DDR SDRAM Interface Debug ............................................................................... 20-28 Debug Information on Debug Pins ...................................................................... 20-28 Debug Information on ECC Pins......................................................................... 20-28 Local Bus Interface Debug ...................................................................................... 20-28 Watchpoint Monitor ................................................................................................. 20-28 Watchpoint Monitor Performance Monitor Events ............................................. 20-29 Trace Buffer ............................................................................................................. 20-29 Traced Data Formats (as a Function of TBCR1[IFSEL]).................................... 20-30 Initialization ................................................................................................................. 20-33 Chapter 21 10/100 Fast Ethernet Controller 21.1 21.1.1 21.2 21.3 21.4 21.4.1 Introduction.................................................................................................................... 21-1 10/100 Fast Ethernet Controller Overview................................................................ 21-5 Features .......................................................................................................................... 21-6 Modes of Operation ....................................................................................................... 21-7 External Signal Descriptions ......................................................................................... 21-7 Detailed Signal Descriptions ..................................................................................... 21-8 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor xxxix Contents Paragraph Number 21.5 21.5.1 21.5.2 21.5.3 21.5.3.1 21.5.3.1.1 21.5.3.1.2 21.5.3.1.3 21.5.3.1.4 21.5.3.1.5 21.5.3.1.6 21.5.3.2 21.5.3.2.1 21.5.3.2.2 21.5.3.2.3 21.5.3.2.4 21.5.3.3 21.5.3.3.1 21.5.3.3.2 21.5.3.3.3 21.5.3.3.4 21.5.3.3.5 21.5.3.3.6 21.5.3.3.7 21.5.3.3.8 21.5.3.4 21.5.3.4.1 21.5.3.4.2 21.5.3.4.3 21.5.3.4.4 21.5.3.4.5 21.5.3.4.6 21.5.3.4.7 21.5.3.5 21.5.3.5.1 21.5.3.5.2 21.5.3.5.3 21.5.3.5.4 21.5.3.6 Title Page Number Memory Map/Register Definition ................................................................................. 21-9 Top-Level Module Memory Map .............................................................................. 21-9 Detailed Memory Map............................................................................................. 21-10 Memory-Mapped Register Descriptions.................................................................. 21-13 FEC General Control and Status Registers.......................................................... 21-13 Interrupt Event Register (IEVENT) ................................................................ 21-13 Interrupt Mask Register (IMASK) .................................................................. 21-16 Error Disabled Register (EDIS)....................................................................... 21-18 Minimum Frame Length Register (MINFLR)................................................. 21-18 Pause Time Value Register (PTV) ................................................................... 21-19 DMA Control Register (DMACTRL) ............................................................. 21-20 FEC FIFO Control and Status Registers.............................................................. 21-21 FIFO Pause Control Register (FIFO_PAUSE_CTRL).................................... 21-22 FIFO Transmit Threshold Register (FIFO_TX_THR) .................................... 21-23 FIFO Transmit Starve Register (FIFO_TX_STARVE) ................................... 21-24 FIFO Transmit Starve Shutoff Register (FIFO_TX_STARVE_SHUTOFF).............................................................. 21-24 FEC Transmit Control and Status Registers ........................................................ 21-25 Transmit Control Register (TCTRL) ............................................................... 21-25 Transmit Status Register (TSTAT) .................................................................. 21-26 TxBD Data Length Register (TBDLEN)......................................................... 21-27 Current Transmit Buffer Descriptor Pointer Register (CTBPTR) ................... 21-27 Transmit Buffer Descriptor Pointer Register (TBPTR) ................................... 21-28 Transmit Descriptor Base Address Register (TBASE) ................................... 21-28 Out-of-Sequence TxBD Register (OSTBD).................................................... 21-29 Out-of-Sequence Tx Data Buffer Pointer Register (OSTBDP)....................... 21-31 FEC Receive Control and Status Registers ......................................................... 21-31 Receive Control Register (RCTRL) ................................................................ 21-31 Receive Status Register (RSTAT).................................................................... 21-32 RxBD Data Length Register (RBDLEN) ........................................................ 21-33 Current Receive Buffer Descriptor Pointer Register (CRBPTR) .................... 21-33 Maximum Receive Buffer Length Register (MRBLR) ................................... 21-34 Receive Buffer Descriptor Pointer Register (RBPTR) .................................... 21-35 Receive Descriptor Base Address Register (RBASE)..................................... 21-35 MAC Functionality.............................................................................................. 21-36 Configuring the MAC...................................................................................... 21-36 Controlling CSMA/CD.................................................................................... 21-36 Handling Packet Collisions ............................................................................. 21-36 Controlling Packet Flow.................................................................................. 21-37 MAC Registers .................................................................................................... 21-38 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 xl Freescale Semiconductor Contents Paragraph Number 21.5.3.6.1 21.5.3.6.2 21.5.3.6.3 21.5.3.6.4 21.5.3.6.5 21.5.3.6.6 21.5.3.6.7 21.5.3.6.8 21.5.3.7 21.5.3.7.1 21.5.3.7.2 21.5.3.8 21.5.3.8.1 21.5.3.8.2 21.6 21.6.1 21.6.1.1 21.6.2 21.6.2.1 21.6.2.1.1 21.6.2.1.2 21.6.2.2 21.6.2.3 21.6.2.4 21.6.2.5 21.6.2.5.1 21.6.2.5.2 21.6.2.5.3 21.6.2.6 21.6.2.7 21.6.2.8 21.6.2.9 21.6.2.10 21.6.3 21.6.3.1 21.6.3.2 21.6.4 21.7 21.7.1 21.7.1.1 Title Page Number MAC Configuration Register 1 (MACCFG1)................................................. 21-38 MAC Configuration Register 2 (MACCFG2)................................................. 21-40 Inter-Packet Gap/Inter-Frame Gap Register (IPGIFG) ................................... 21-41 Half-Duplex Register (HAFDUP) ................................................................... 21-41 Maximum Frame Length Register (MAXFRM) ............................................. 21-42 Interface Status Register (IFSTAT) ................................................................. 21-43 Station Address Register Part 1 (MACSTNADDR1) ..................................... 21-43 Station Address Register Part 2 (MACSTNADDR2) ..................................... 21-44 Hash Function Registers ...................................................................................... 21-45 Individual Address Registers 0–7 (IADDRn) ................................................. 21-45 Group Address Registers 0–7 (GADDRn) ...................................................... 21-46 Attribute Registers ............................................................................................... 21-46 Attribute Register (ATTR)............................................................................... 21-46 Attribute Extract Length and Extract Index Register (ATTRELI) .................. 21-47 Functional Description................................................................................................. 21-48 Connecting to the Physical Interface ....................................................................... 21-48 Media-Independent Interface (MII) ..................................................................... 21-48 FEC Channel Operation........................................................................................... 21-49 Initialization Sequence......................................................................................... 21-50 Hardware Controlled Initialization.................................................................. 21-50 User Initialization ............................................................................................ 21-50 Soft Reset and Reconfiguring Procedure............................................................. 21-51 FEC Frame Transmission .................................................................................... 21-52 FEC Frame Reception.......................................................................................... 21-53 Frame Recognition............................................................................................... 21-54 Destination Address Recognition .................................................................... 21-55 Hash Table Algorithm...................................................................................... 21-57 CRC Computation Examples .......................................................................... 21-57 Flow Control........................................................................................................ 21-59 Interrupt Handling ............................................................................................... 21-59 Inter-Packet Gap Time......................................................................................... 21-61 Internal and External Loop Back......................................................................... 21-61 Error-Handling Procedure.................................................................................... 21-61 Buffer Descriptors.................................................................................................... 21-63 Transmit Data Buffer Descriptor (TxBD)............................................................ 21-65 Receive Buffer Descriptor (RxBD) ..................................................................... 21-67 Data Extraction to the L2 Cache.............................................................................. 21-69 Application Information .............................................................................................. 21-69 Interface Mode Configuration ................................................................................. 21-69 MII Interface Mode.............................................................................................. 21-70 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor xli Contents Paragraph Number Title Appendix A Revision History A.1 Changes From Revision 0 to Revision 1 ........................................................................ A-1 Glossary of Terms and Abbreviations Index 1 Register Index (Memory-Mapped Registers) Index 2 General Index Page Number MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 xlii Freescale Semiconductor Figures Figure Number 1-1 1-2 1-3 1-4 1-5 1-6 1-7 1-8 2-1 2-2 2-3 2-4 2-5 2-6 2-7 2-8 3-1 4-1 4-2 4-3 4-4 4-5 4-6 4-7 4-8 5-1 5-2 5-3 5-4 5-5 5-6 5-7 5-8 6-1 6-2 Title Figures Page Number MPC8540 Block Diagram....................................................................................................... 1-2 Processing Transactions Across the On-Chip Fabric............................................................ 1-21 High-Performance Communication System Using SMP...................................................... 1-22 High-Performance Communication System Using MPC8540 ............................................. 1-22 RAID Controller Application Using MPC8540.................................................................... 1-23 VPN Access Point Enabled by RapidIO and Ethernet.......................................................... 1-23 MPC8540 with SERDES ...................................................................................................... 1-24 MPC8540 with a DSP Farm Enabled by General-Purpose Chip Select Machine or User Programmable Machine on the Local Bus Controller Port .................. 1-24 Local Memory Map Example ................................................................................................. 2-2 Local Access Window n Base Address Registers (LAWBAR0–LAWBAR7) ....................... 2-6 Local Access Window n Attributes Registers (LAWAR0–LAWAR7) ................................... 2-6 Top-Level Register Map Example ........................................................................................ 2-10 General Utilities Registers Mapping to Configuration, Control, and Status Memory Block.................................................................................................................. 2-13 PIC Mapping to Configuration, Control, and Status Memory Block ................................... 2-14 RapidIO Mapping to Configuration, Control, and Status Memory Block............................ 2-15 Device-Specific Register Mapping to Configuration, Control, and Status Memory Block.................................................................................................................. 2-16 MPC8540 Signal Groupings ................................................................................................... 3-3 Configuration, Control, and Status Register Base Address Register (CCSRBAR)................ 4-5 Alternate Configuration Base Address Register (ALTCBAR) ............................................... 4-6 Alternate Configuration Attribute Register (ALTCAR) ......................................................... 4-6 Boot Page Translation Register (BPTR) ................................................................................. 4-8 Power-On Reset Sequence .................................................................................................... 4-11 Clock Subsystem Block Diagram ......................................................................................... 4-24 RapidIO Transmit Clock Options ......................................................................................... 4-24 RTC and Core Timer Facilities Clocking Options ................................................................ 4-26 e500 Core Complex Block Diagram ....................................................................................... 5-2 Four-Stage MU Pipeline, Showing Divide Bypass................................................................. 5-8 Three-Stage Load/Store Unit .................................................................................................. 5-9 Instruction Pipeline Flow ...................................................................................................... 5-15 GPR Issue Queue (GIQ) ....................................................................................................... 5-17 e500 Core Programming Model............................................................................................ 5-19 MMU Structure ..................................................................................................................... 5-25 Effective-to-Real Address Translation Flow......................................................................... 5-26 Core Register Model ............................................................................................................... 6-2 Integer Exception Register (XER) .......................................................................................... 6-8 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor xliii Figures Figure Number 6-3 6-4 6-5 6-6 6-7 6-8 6-9 6-10 6-11 6-12 6-13 6-14 6-15 6-16 6-17 6-18 6-19 6-20 6-21 6-22 6-23 6-24 6-25 6-26 6-27 6-28 6-29 6-30 6-31 6-32 6-33 6-34 6-35 6-36 6-37 6-38 6-39 6-40 6-41 6-42 Title Page Number Condition Register (CR) ......................................................................................................... 6-9 Link Register (LR) ................................................................................................................ 6-11 Count Register (CTR) ........................................................................................................... 6-11 Machine State Register (MSR) ............................................................................................. 6-12 Processor ID Register (PIR).................................................................................................. 6-14 Processor Version Register (PVR) ........................................................................................ 6-14 System Version Register (SVR)............................................................................................ 6-14 Timer Control Register (TCR) .............................................................................................. 6-15 Timer Status Register (TSR) ................................................................................................. 6-16 Time Base Upper/Lower Registers (TBU/TBL)................................................................... 6-17 Decrementer Register (DEC) ................................................................................................ 6-17 Decrementer Auto-Reload Register (DECAR)..................................................................... 6-17 Save/Restore Register 0 (SRR0) ........................................................................................... 6-18 Save/Restore Register 1 (SRR1) ........................................................................................... 6-18 Critical Save/Restore Register 0 (CSRR0) ........................................................................... 6-18 Critical Save/Restore Register 1 (CSRR1) ........................................................................... 6-19 Data Exception Address Register (DEAR) ........................................................................... 6-19 Interrupt Vector Prefix Register (IVPR) ............................................................................... 6-19 Interrupt Vector Offset Registers (IVORn) ........................................................................... 6-19 Exception Syndrome Register (ESR).................................................................................... 6-20 Machine Check Save/Restore Register 0 (MCSRR0)........................................................... 6-21 Machine Check Save/Restore Register 1 (MCSRR1)........................................................... 6-22 Machine Check Address Register (MCAR).......................................................................... 6-22 Machine Check Syndrome Register (MCSR) ....................................................................... 6-23 Software-Use SPRs (SPRG0–SPRG7 and USPRG0)........................................................... 6-24 Branch Buffer Entry Address Register (BBEAR) ................................................................ 6-25 Branch Buffer Target Address Register (BBTAR)................................................................ 6-25 Branch Unit Control and Status Register (BUCSR) ............................................................. 6-26 Hardware Implementation-Dependent Register 0 (HID0).................................................... 6-27 Hardware Implementation-Dependent Register 1 (HID1).................................................... 6-28 L1 Cache Control and Status Register 0 (L1CSR0) ............................................................. 6-30 L1 Cache Control and Status Register 1 (L1CSR1) ............................................................. 6-31 L1 Cache Configuration Register 0 (L1CFG0)..................................................................... 6-32 L1 Cache Configuration Register 1 (L1CFG1)..................................................................... 6-33 Process ID Registers (PID0–PID2)....................................................................................... 6-34 MMU Control and Status Register 0 (MMUCSR0) ............................................................. 6-34 MMU Configuration Register (MMUCFG) ......................................................................... 6-35 TLB Configuration Register 0 (TLB0CFG) ......................................................................... 6-35 TLB Configuration Register 1 (TLB1CFG) ......................................................................... 6-36 MAS Register 0 (MAS0) ...................................................................................................... 6-37 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 xliv Freescale Semiconductor Figures Figure Number 6-43 6-44 6-45 6-46 6-47 6-48 6-49 6-50 6-51 6-52 6-53 6-54 6-55 6-56 6-57 6-58 6-59 7-1 7-2 7-3 7-4 7-5 7-6 7-7 7-8 7-9 7-10 7-11 7-12 7-13 7-14 7-15 7-16 7-17 7-18 Title Page Number MAS Register 1 (MAS1) ...................................................................................................... 6-38 MAS Register 2 (MAS2) ...................................................................................................... 6-39 MAS Register 3 (MAS3) ...................................................................................................... 6-40 MAS Register 4 (MAS4) ...................................................................................................... 6-40 MAS Register 6 (MAS6) ...................................................................................................... 6-41 Debug Control Register 0 (DBCR0) ..................................................................................... 6-42 Debug Control Register 1 (DBCR1) ..................................................................................... 6-43 Debug Control Register 2 (DBCR2) ..................................................................................... 6-45 Debug Status Register (DBSR)............................................................................................. 6-46 Instruction Address Compare Registers (IAC1–IAC2) ........................................................ 6-47 Data Address Compare Registers (DAC1–DAC2) ............................................................... 6-48 Signal Processing and Embedded Floating-Point Status and Control Register (SPEFSCR) ........................................................................................................ 6-48 Accumulator (ACC) .............................................................................................................. 6-50 Performance Monitor Global Control Register 0 (PMGC0), User Performance Monitor Global Control Register 0 (UPMGC0) ................................ 6-52 Local Control A Registers (PMLCa0–PMLCa3), User Local Control A Registers (UPMLCa0–UPMLCa3) .............................................. 6-52 Local Control B Registers (PMLCb0–PMLCb3), User Local Control B Registers (UPMLCb0–UPMLCb3) .............................................. 6-53 Performance Monitor Counter Registers (PMC0–PMC3), User Performance Monitor Counter Registers (UPMC0–UPMC3)................................. 6-54 L2 Cache/SRAM Configuration ............................................................................................. 7-1 Cache Organization ................................................................................................................. 7-3 256-Kbyte L2 Cache Address Configuration—Full Cache Mode.......................................... 7-4 128-Kbyte L2 Cache Address Configuration—Half SRAM, Half Cache Mode.................... 7-5 Data Bus Connection of CCB ................................................................................................. 7-5 Address Bus Connection of CCB............................................................................................ 7-6 L2 Control Register (L2CTL) ................................................................................................. 7-7 L2 Cache External Write Address Registers (L2CEWARn)................................................. 7-10 L2 Cache External Write Control Registers (L2CEWCR0–L2CEWCR3)........................... 7-10 L2 Memory-Mapped SRAM Base Address Registers (L2SRBARn)................................... 7-11 L2 Error Injection Mask High Register (L2ERRINJHI) ...................................................... 7-13 L2 Error Injection Mask Low Register (L2ERRINJLO) ...................................................... 7-14 L2 Error Injection Mask Control Register (L2ERRINJCTL) ............................................... 7-14 L2 Error Capture Data High Register (L2CAPTDATAHI)................................................... 7-15 L2 Error Capture Data Low Register (L2CAPTDATALO) .................................................. 7-15 L2 Error Syndrome Register (L2CAPTECC) ....................................................................... 7-16 L2 Error Detect Register (L2ERRDET) ............................................................................... 7-16 L2 Error Disable Register (L2ERRDIS) ............................................................................... 7-17 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor xlv Figures Figure Number 7-19 7-20 7-21 7-22 7-23 8-1 8-2 8-3 8-4 8-5 8-6 8-7 9-1 9-2 9-3 9-4 9-5 9-6 9-7 9-8 9-9 9-10 9-11 9-12 9-13 9-14 9-15 9-16 9-17 9-18 9-19 9-20 9-21 9-22 9-23 9-24 9-25 9-26 9-27 9-28 Title Page Number L2 Error Interrupt Enable Register (L2ERRINTEN) ........................................................... 7-18 L2 Error Attributes Capture Register (L2ERRATTR) .......................................................... 7-19 L2 Error Address Capture Register (L2ERRADDR) ........................................................... 7-20 L2 Error Control Register (L2ERRCTL).............................................................................. 7-20 L2 Cache Line Replacement Algorithm ............................................................................... 7-28 e500 Coherency Module Block Diagram................................................................................ 8-1 ECM CCB Address Configuration Register (EEBACR)........................................................ 8-3 ECM CCB Port Configuration Register (EEBPCR)............................................................... 8-4 ECM Error Detect Register (EEDR)....................................................................................... 8-5 ECM Error Enable Register (EEER) ...................................................................................... 8-6 ECM Error Attributes Capture Register (EEATR) ................................................................. 8-7 ECM Error Address Capture Register (EEADR) ................................................................... 8-8 DDR Memory Controller Simplified Block Diagram............................................................. 9-2 Chip Select Bounds Registers (CSn_BNDS)........................................................................ 9-10 Chip Select Configuration Register (CSn_CONFIG) ........................................................... 9-10 DDR SDRAM Timing Configuration Register 1 (TIMING_CFG_1).................................. 9-11 DDR SDRAM Timing Configuration Register 2 (TIMING_CFG_2).................................. 9-13 DDR SDRAM Control Configuration Register (DDR_SDRAM_CFG) .............................. 9-14 DDR SDRAM Mode Configuration Register (DDR_SDRAM_MODE)............................. 9-16 DDR SDRAM Interval Configuration Register (DDR_SDRAM_INTERVAL) .................. 9-16 Memory Data Path Error Injection Mask High Register (DATA_ERR_INJECT_HI) ......... 9-17 Memory Data Path Error Injection Mask Low Register (DATA_ERR_INJECT_LO)......... 9-18 Memory Data Path Error Injection Mask ECC Register (ECC_ERR_INJECT) .................. 9-18 Memory Data Path Read Capture High Register (CAPTURE_DATA_HI).......................... 9-19 Memory Data Path Read Capture Low Register (CAPTURE_DATA_LO) ......................... 9-20 Memory Data Path Read Capture ECC Register (CAPTURE_ECC)................................... 9-20 Memory Error Detect Register (ERR_DETECT) ................................................................. 9-21 Memory Error Disable Register (ERR_DISABLE).............................................................. 9-22 Memory Error Interrupt Enable Register (ERR_INT_EN)................................................... 9-22 Memory Error Attributes Capture Register (CAPTURE_ATTRIBUTES)........................... 9-23 Memory Error Address Capture Register (CAPTURE_ADDRESS) ................................... 9-24 Single-Bit ECC Memory Error Management Register (ERR_SBE) .................................... 9-25 DDR Memory Controller Block Diagram ............................................................................ 9-26 Controller DLL Timing Loop ............................................................................................... 9-27 Typical Dual Data Rate SDRAM Internal Organization....................................................... 9-28 Typical DDR SDRAM Interface Signals .............................................................................. 9-28 Example 256-Mbyte DDR SDRAM Configuration with ECC............................................. 9-29 DDR SDRAM Burst Read Timing—ACTTORW = 3, MCAS Latency = 2 ........................ 9-36 DDR SDRAM Single-Beat (Double-Word) Write Timing—ACTTORW = 3 ..................... 9-37 DDR SDRAM Burst Write Timing—ACTTORW = 4 ......................................................... 9-37 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 xlvi Freescale Semiconductor Figures Figure Number 9-29 9-30 9-31 9-32 9-33 9-34 9-35 9-36 10-1 10-2 10-3 10-4 10-5 10-6 10-7 10-8 10-9 10-10 10-11 10-12 10-13 10-14 10-15 10-16 10-17 10-18 10-19 10-20 10-21 10-22 10-23 10-24 10-25 10-26 10-27 10-28 10-29 10-30 10-31 10-32 Title Page Number DDR SDRAM Clock Distribution Example ......................................................................... 9-38 DDR SDRAM Mode-Set Command Timing........................................................................ 9-39 Registered DDR SDRAM DIMM Burst Write Timing ........................................................ 9-40 Write Timing Adjustments Example..................................................................................... 9-41 DDR SDRAM Bank-Staggered Auto-Refresh Timing......................................................... 9-42 DDR SDRAM Power-Down Mode ...................................................................................... 9-43 DDR SDRAM Self-Refresh Entry Timing ........................................................................... 9-44 DDR SDRAM Self-Refresh Exit Timing ............................................................................. 9-44 MPC8540 Interrupt Sources Block Diagram ........................................................................ 10-3 Pass-Through Mode Example ............................................................................................... 10-6 Feature Reporting Register (FRR) ...................................................................................... 10-16 Global Configuration Register (GCR) ................................................................................ 10-17 Vendor Identification Register (VIR).................................................................................. 10-17 Processor Initialization Register (PIR) ............................................................................... 10-18 IPI Vector/Priority Registers (IPIVPRn)............................................................................. 10-19 Spurious Vector Register (SVR) ......................................................................................... 10-20 Timer Frequency Reporting Register (TFRR) .................................................................... 10-20 Global Timer Current Count Registers (GTCCRn) ............................................................ 10-21 Global Timer Base Count Registers (GTBCRn)................................................................. 10-22 Global Timer Vector/Priority Registers (GTVPRn)............................................................ 10-22 Global Timer Destination Registers (GTDRn) ................................................................... 10-23 Example Calculation for Cascaded Timers......................................................................... 10-24 Timer Control Register (TCR) ............................................................................................ 10-24 IRQ_OUT Summary Register 0 (IRQSR0) ........................................................................ 10-26 IRQ_OUT Summary Register 1 (IRQSR1) ........................................................................ 10-27 Critical Interrupt Summary Register 0 (CISR0) ................................................................. 10-27 Critical Interrupt Summary Register 1 (CISR1) ................................................................. 10-28 Performance Monitor Mask Registers (PMnMR0)............................................................. 10-29 Performance Monitor Mask Registers (PMnMR1)............................................................. 10-30 Message Registers (MSGRs) .............................................................................................. 10-30 Message Enable Register (MER) ........................................................................................ 10-31 Message Status Register (MSR) ......................................................................................... 10-31 External Interrupt Vector/Priority Registers (EIVPR0–EIVPR11) ..................................... 10-32 External Interrupt Destination Registers (EIDRn).............................................................. 10-33 Internal Interrupt Vector/Priority Registers (IIVPRn) ........................................................ 10-34 Internal Interrupt Destination Registers (IIDRn) ................................................................ 10-35 Messaging Interrupt Vector/Priority Registers (MIVPRn) ................................................. 10-36 Messaging Interrupt Destination Registers (MIDRn) ......................................................... 10-37 Per-CPU Register Address Decoding in a Four-Core Device............................................. 10-39 Interprocessor Interrupt Dispatch Registers (IPIDR0–IPIDR3) ......................................... 10-39 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor xlvii Figures Figure Number 10-33 10-34 10-35 10-36 10-37 11-1 11-2 11-3 11-4 11-5 11-6 11-7 11-8 11-9 11-10 11-11 12-1 12-2 12-3 12-4 12-5 12-6 12-7 12-8 12-9 12-10 12-11 12-12 12-13 12-14 12-15 12-16 13-1 13-2 13-3 13-4 13-5 13-6 13-7 13-8 Title Page Number Processor Current Task Priority Register (CTPR) .............................................................. 10-40 Processor Who Am I Register (WHOAMI)........................................................................ 10-41 Processor Interrupt Acknowledge Register (IACK) ........................................................... 10-42 End of Interrupt Register (EOI) .......................................................................................... 10-42 PIC Interrupt Processing Flow Diagram ............................................................................. 10-44 I2C Block Diagram................................................................................................................ 11-1 I2C Address Register (I2CADR)........................................................................................... 11-5 I2C Frequency Divider Register (I2CFDR) .......................................................................... 11-5 I2C Control Register (I2CCR)............................................................................................... 11-6 I2C Status Register (I2CSR) ................................................................................................. 11-8 I2C Data Register (I2CDR) ................................................................................................... 11-9 I2C Digital Filter Sampling Rate Register (I2CDFSRR).................................................... 11-10 I2C Interface Transaction Protocol.......................................................................................11-11 EEPROM Data Format for One Register Preload Command............................................. 11-19 EEPROM Contents ............................................................................................................. 11-19 Example I2C Interrupt Service Routine Flowchart ............................................................. 11-24 UART Block Diagram .......................................................................................................... 12-2 Receiver Buffer Registers (URBR0, URBR1)...................................................................... 12-6 Transmitter Holding Registers (UTHR0, UTHR1)............................................................... 12-7 Divisor Most Significant Byte Registers (UDMB0, UDMB1)............................................. 12-7 Divisor Least Significant Byte Registers (UDLB0, UDLB1)............................................... 12-8 Interrupt Enable Register (UIER) ......................................................................................... 12-9 Interrupt ID Registers (UIIR).............................................................................................. 12-10 FIFO Control Registers (UFCR0, UFCR1) ........................................................................ 12-11 Line Control Register (ULCR) ........................................................................................... 12-13 Modem Control Register (UMCR) ..................................................................................... 12-14 Line Status Register (ULSR) .............................................................................................. 12-15 Modem Status Register (UMSR) ........................................................................................ 12-16 Scratch Register (USCR) .................................................................................................... 12-17 Alternate Function Register (UAFR) .................................................................................. 12-18 DMA Status Register (UDSR)............................................................................................ 12-18 UART Bus Interface Transaction Protocol Example .......................................................... 12-21 Local Bus Controller Block Diagram ................................................................................... 13-1 Base Registers (BRn) .......................................................................................................... 13-11 Option Registers (ORn) in GPCM Mode............................................................................ 13-14 Option Registers (ORn) in UPM Mode .............................................................................. 13-16 Option Registers (ORn) in SDRAM Mode......................................................................... 13-17 UPM Memory Address Register (MAR) ............................................................................ 13-18 UPM Mode Registers (MxMR)........................................................................................... 13-19 Memory Refresh Timer Prescaler Register (MRTPR)........................................................ 13-21 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 xlviii Freescale Semiconductor Figures Figure Number 13-9 13-10 13-11 13-12 13-13 13-14 13-15 13-16 13-17 13-18 13-19 13-20 13-21 13-22 13-23 13-24 13-25 13-26 13-27 13-28 13-29 13-30 13-31 13-32 13-33 13-34 13-35 13-36 13-37 13-38 13-39 13-40 13-41 Title Page Number UPM Data Register (MDR) ................................................................................................ 13-22 SDRAM Machine Mode Register (LSDMR) ..................................................................... 13-22 UPM Refresh Timer (LURT) .............................................................................................. 13-24 LSRT SDRAM Refresh Timer (LSRT)............................................................................... 13-25 Transfer Error Status Register (LTESR) ............................................................................. 13-26 Transfer Error Check Disable Register (LTEDR) ............................................................... 13-27 Transfer Error Interrupt Enable Register (LTEIR).............................................................. 13-28 Transfer Error Attributes Register (LTEATR) .................................................................... 13-29 Transfer Error Address Register (LTEAR) ......................................................................... 13-30 Local Bus Configuration Register....................................................................................... 13-31 Clock Ratio Register (LCRR) ............................................................................................. 13-32 Basic Operation of Memory Controllers in the LBC.......................................................... 13-34 Example of 8-Bit GPCM Writing 32 Bytes to Address 0x5420 ......................................... 13-36 Basic LBC Bus Cycle with LALE, TA, and LCSn ............................................................. 13-37 Local Bus to GPCM Device Interface ................................................................................ 13-39 GPCM Basic Read Timing (XACS = 0, ACS = 1x, TRLX = 0, CLKDIV = 4,8) .............................................................................................................. 13-39 GPCM Basic Write Timing (XACS = 0, ACS = 00, CSNT = 1, SCY = 1, TRLX = 0, CLKDIV = 4 or 8) ....................................................................................... 13-45 GPCM Relaxed Timing Read (XACS = 0, ACS = 1x, SCY = 1, CSNT = 0, TRLX = 1, CLKDIV = 4 or 8) ....................................................................................... 13-46 GPCM Relaxed Timing Back-to-Back Writes (XACS = 0, ACS = 1x, SCY = 0, CSNT = 0, TRLX = 1, CLKDIV = 4 or 8)..................................................... 13-46 GPCM Relaxed Timing Write (XACS = 0, ACS = 10, SCY = 0, CSNT = 1, TRLX = 1, CLKDIV = 4 or 8) ....................................................................................... 13-47 GPCM Relaxed Timing Write (XACS = 0, ACS = 00, SCY = 1, CSNT = 1, TRLX = 1, CLKDIV = 4 or 8) ....................................................................................... 13-48 GPCM Read Followed by Read (TRLX = 0, EHTR = 0, Fastest Timing) ......................... 13-49 GPCM Read Followed by Write (TRLX = 0, EHTR = 1, 1-Cycle Extended Hold Time on Reads) ..................................................................................................... 13-49 External Termination of GPCM Access.............................................................................. 13-50 Connection to a 32-Bit SDRAM with 12 Address Lines.................................................... 13-52 SDRAM Address Multiplexing .......................................................................................... 13-54 PRETOACT = 2 (2 Clock Cycles)...................................................................................... 13-55 ACTTORW = 2 (2 Clock Cycles)....................................................................................... 13-56 CL = 2 (2 Clock Cycles) ..................................................................................................... 13-56 WRC = 2 (2 Clock Cycles) ................................................................................................. 13-57 RFRC = 4 (6 Clock Cycles) ................................................................................................ 13-57 BUFCMD = 1, LCRR[BUFCMDC] = 2............................................................................. 13-58 SDRAM Single-Beat Read, Page Closed, CL = 3 .............................................................. 13-58 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor xlix Figures Figure Number 13-42 13-43 13-44 13-45 13-46 13-47 13-48 13-49 13-50 13-51 13-52 13-53 13-54 13-55 13-56 13-57 13-58 13-59 13-60 13-61 13-62 13-63 13-64 13-65 13-66 13-67 13-68 13-69 13-70 13-71 13-72 13-73 13-74 13-75 13-76 13-77 13-78 13-79 13-80 13-81 Title Page Number SDRAM Single-Beat Read, Page Hit, CL = 3 .................................................................... 13-58 SDRAM Two-Beat Burst Read, Page Closed, CL = 3........................................................ 13-58 SDRAM Four-Beat Burst Read, Page Miss, CL = 3........................................................... 13-59 SDRAM Single-Beat Write, Page Hit................................................................................. 13-59 SDRAM Three-Beat Write, Page Closed............................................................................ 13-59 SDRAM Read-after-Read Pipelined, Page Hit, CL = 3...................................................... 13-59 SDRAM Write-after-Write Pipelined, Page Hit.................................................................. 13-60 SDRAM Read-after-Write Pipelined, Page Hit .................................................................. 13-60 SDRAM MODE-SET Command........................................................................................ 13-61 SDRAM Bank-Staggered Auto-Refresh Timing ................................................................ 13-62 User-Programmable Machine Functional Block Diagram.................................................. 13-62 RAM Array Indexing .......................................................................................................... 13-64 Memory Refresh Timer Request Block Diagram ............................................................... 13-65 UPM Clock Scheme for LCRR[CLKDIV] = 2................................................................... 13-67 UPM Clock Scheme for LCRR[CLKDIV] = 4 or 8 ........................................................... 13-67 RAM Array and Signal Generation .................................................................................... 13-68 RAM Word Field Descriptions ........................................................................................... 13-68 LCSn Signal Selection ........................................................................................................ 13-72 LBS Signal Selection .......................................................................................................... 13-73 UPM Read Access Data Sampling...................................................................................... 13-76 Effect of LUPWAIT Signal ................................................................................................. 13-77 Single-Beat Read Access to FPM DRAM .......................................................................... 13-79 Single-Beat Write Access to FPM DRAM ......................................................................... 13-80 Burst Read Access to FPM DRAM Using LOOP (Two Beats Shown).............................. 13-81 Refresh Cycle (CBR) to FPM DRAM ................................................................................ 13-82 Exception Cycle .................................................................................................................. 13-83 Multiplexed Address/Data Bus ........................................................................................... 13-84 Local Bus Peripheral Hierarchy.......................................................................................... 13-85 Local Bus Peripheral Hierarchy for Very High Bus Speeds ............................................... 13-86 GPCM Address Timings ..................................................................................................... 13-86 GPCM Data timings............................................................................................................ 13-87 Interface to Different Port-Size Devices ............................................................................. 13-89 128-Mbyte SDRAM Diagram............................................................................................. 13-93 SDRAM Power-Down Timing............................................................................................ 13-97 SDRAM Self-Refresh Mode Timing .................................................................................. 13-98 Local Bus DLL Operation................................................................................................. 13-100 Parity Support for SDRAM............................................................................................... 13-101 Interface to ZBT SRAM ................................................................................................... 13-102 MSC8101 HDI16 Peripheral Registers............................................................................. 13-104 Interface to MSC8101 HDI16........................................................................................... 13-105 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 l Freescale Semiconductor Figures Figure Number 13-82 13-83 13-84 13-85 13-86 13-87 13-88 13-89 13-90 13-91 13-92 13-93 14-1 14-2 14-3 14-4 14-5 14-6 14-7 14-8 14-9 14-10 14-11 14-12 14-13 14-14 14-15 14-16 14-17 14-18 14-19 14-20 14-21 14-22 14-23 14-24 14-25 14-26 14-27 14-28 Title Page Number Interface to MSC8102 DSI in Asynchronous Mode ......................................................... 13-108 Asynchronous Write to MSC8102 DSI............................................................................. 13-109 Asynchronous Read from MSC8102 DSI......................................................................... 13-110 Interface to MSC8102 DSI in Synchronous Mode ............................................................13-111 UPM Synchronization Cycle ............................................................................................ 13-112 Synchronous Single Write to MSC8102 DSI.................................................................... 13-114 Synchronous Single Read from MSC8102 DSI................................................................ 13-115 Synchronous Burst Write to MSC8102 DSI ..................................................................... 13-116 Synchronous Burst Read from MSC8102 DSI ................................................................. 13-117 Interface to Texas Instruments EHPI in Non-Multiplexed Mode ..................................... 13-120 EHPI Non-Multiplexed Read Timings.............................................................................. 13-121 EHPI Non-Multiplexed Write Timings............................................................................. 13-121 Ethernet Protocol in Relation to the OSI Protocol Stack ...................................................... 14-2 IEEE 802.3z and 802.3ab Physical Standards ...................................................................... 14-3 Ethernet/IEEE 802.3 Frame Structure .................................................................................. 14-3 Ethernet/IEEE 802.3 Frame Structure With More Details.................................................... 14-5 TSEC Block Diagram ........................................................................................................... 14-6 IEVENT Register Definition .............................................................................................. 14-21 IMASK Register Definition ................................................................................................ 14-23 Error Disabled Register (EDIS) .......................................................................................... 14-24 ECNTRL Register Definition ............................................................................................. 14-25 MINFLR Register Definition.............................................................................................. 14-26 PTV Register Definition...................................................................................................... 14-27 DMACTRL Register Definition ......................................................................................... 14-28 TBIPA Register Definition.................................................................................................. 14-29 FIFO_PAUSE_CTRL Register Definition......................................................................... 14-31 FIFO_TX_THR Register Definition ................................................................................... 14-31 FIFO_TX_STARVE Register Definition ............................................................................ 14-32 FIFO_TX_STARVE_SHUTOFF Register Definition ........................................................ 14-33 TCTRL Register Definition ................................................................................................ 14-33 TSTAT Register Definition ................................................................................................. 14-34 TBDLEN Register Definition ............................................................................................. 14-35 TXIC Register Definition.................................................................................................... 14-35 CTBPTR Register Definition .............................................................................................. 14-36 TBPTR Register Definition................................................................................................. 14-37 TBASE Register Definition ................................................................................................ 14-38 OSTBD Register Definition................................................................................................ 14-38 OSTBDP Register Definition.............................................................................................. 14-40 RCTRL Register Definition ................................................................................................ 14-41 RSTAT Register Definition ................................................................................................. 14-42 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor li Figures Figure Number 14-29 14-30 14-31 14-32 14-33 14-34 14-35 14-36 14-37 14-38 14-39 14-40 14-41 14-42 14-43 14-44 14-45 14-46 14-47 14-48 14-49 14-50 14-51 14-52 14-53 14-54 14-55 14-56 14-57 14-58 14-59 14-60 14-61 14-62 14-63 14-64 14-65 14-66 14-67 14-68 Title Page Number RBDLEN Register Definition............................................................................................. 14-42 RXIC Register Definition ................................................................................................... 14-43 CRBPTR Register Definition.............................................................................................. 14-44 MRBL Register Definition.................................................................................................. 14-44 RBPTR Register Definition ................................................................................................ 14-45 RBASE Register Definition ................................................................................................ 14-46 MACCFG1 Register Definition .......................................................................................... 14-49 MACCFG2 Register Definition .......................................................................................... 14-51 IPGIFG Register Definition ................................................................................................ 14-52 Half-Duplex Register Definition......................................................................................... 14-53 Maximum Frame Length Register Definition..................................................................... 14-54 MII Management Configuration Register Definition ......................................................... 14-54 MIIMCOM Register Definition .......................................................................................... 14-55 MIIMADD Register Definition .......................................................................................... 14-56 MII Management Control Register Definition.................................................................... 14-57 MIIMSTAT Register Definition.......................................................................................... 14-57 MII Management Indicator Register Definition ................................................................. 14-58 Interface Status Register Definition.................................................................................... 14-58 Station Address Part 1 Register Definition ......................................................................... 14-59 Station Address Part 2 Register Definition ......................................................................... 14-60 Transmit and Receive 64-Byte Frame Register Definition ................................................. 14-61 Transmit and Receive 65- to 127-Byte Frame Register Definition .................................... 14-61 Transmit and Receive 128- to 255-Byte Frame Register Definition .................................. 14-62 Transmit and Receive 256- to 511-Byte Frame Register Definition................................... 14-62 Transmit and Receive 512- to 1023-Byte Frame Register Definition ................................ 14-63 Transmit and Receive 1024- to 1518-Byte Frame Register Definition .............................. 14-63 Transmit and Receive 1519- to 1522-Byte VLAN Frame Register Definition .................. 14-64 Receive Byte Counter Register Definition.......................................................................... 14-64 Receive Packet Counter Register Definition ...................................................................... 14-65 Receive FCS Error Counter Register Definition................................................................. 14-65 Receive Multicast Packet Counter Register Definition ...................................................... 14-66 Receive Broadcast Packet Counter Register Definition ..................................................... 14-66 Receive Control Frame Packet Counter Register Definition .............................................. 14-67 Receive Pause Frame Packet Counter Register Definition ................................................. 14-67 Receive Unknown Opcode Packet Counter Register Definition ........................................ 14-68 Receive Alignment Error Counter Register Definition....................................................... 14-68 Receive Frame Length Error Counter Register Definition ................................................. 14-69 Receive Code Error Counter Register Definition ............................................................... 14-69 Receive Carrier Sense Error Counter Register Definition .................................................. 14-70 Receive Undersize Packet Counter Register Definition ..................................................... 14-70 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 lii Freescale Semiconductor Figures Figure Number 14-69 14-70 14-71 14-72 14-73 14-74 14-75 14-76 14-77 14-78 14-79 14-80 14-81 14-82 14-83 14-84 14-85 14-86 14-87 14-88 14-89 14-90 14-91 14-92 14-93 14-94 14-95 14-96 14-97 14-98 14-99 14-100 14-101 14-102 14-103 14-104 14-105 14-106 14-107 14-108 Title Page Number Receive Oversize Packet Counter Register Definition ....................................................... 14-71 Receive Fragments Counter Register Definition ................................................................ 14-71 Receive Jabber Counter Register Definition....................................................................... 14-72 Receive Dropped Packet Counter Register Definition ....................................................... 14-72 Transmit Byte Counter Register Definition ........................................................................ 14-73 Transmit Packet Counter Register Definition ..................................................................... 14-73 Transmit Multicast Packet Counter Register Definition ..................................................... 14-74 Transmit Broadcast Packet Counter Register Definition .................................................... 14-74 Transmit Pause Control Frame Counter Register Definition .............................................. 14-75 Transmit Deferral Packet Counter Register Definition....................................................... 14-75 Transmit Excessive Deferral Packet Counter Register Definition...................................... 14-76 Transmit Single Collision Packet Counter Register Definition .......................................... 14-76 Transmit Multiple Collision Packet Counter Register Definition....................................... 14-77 Transmit Late Collision Packet Counter Register Definition ............................................. 14-77 Transmit Excessive Collision Packet Counter Register Definition .................................... 14-78 Transmit Total Collision Counter Register Definition........................................................ 14-78 Transmit Drop Frame Counter Register Definition ............................................................ 14-79 Transmit Jabber Frame Counter Register Definition .......................................................... 14-79 Transmit FCS Error Counter Register Definition ............................................................... 14-80 Transmit Control Frame Counter Register Definition ........................................................ 14-80 Transmit Oversized Frame Counter Register Definition .................................................... 14-81 Transmit Undersize Frame Counter Register Definition .................................................... 14-81 Transmit Fragment Counter Register Definition ................................................................ 14-82 Carry Register 1 (CAR1) Register Definition..................................................................... 14-82 Carry Register 2 (CAR2) Register Definition..................................................................... 14-84 Carry Mask Register 1 (CAM1) Register Definition.......................................................... 14-85 Carry Mask Register 2 (CAM2) Register Definition.......................................................... 14-86 IADDRn Register Definition .............................................................................................. 14-88 GADDRn Register Definition............................................................................................. 14-88 ATTR Register Definition ................................................................................................... 14-89 ATTRELI Register Definition............................................................................................. 14-90 Control Register Definition................................................................................................. 14-92 Status Register Definition ................................................................................................... 14-93 AN Advertisement Register Definition............................................................................... 14-94 AN Link Partner Base Page Ability Register Definition .................................................... 14-96 AN Expansion Register Definition ..................................................................................... 14-98 AN Next Page Transmit Register Definition ...................................................................... 14-98 AN Link Partner Ability Next Page Register Definition .................................................... 14-99 Extended Status Register Definition ................................................................................. 14-100 Jitter Diagnostics Register Definition ............................................................................... 14-101 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor liii Figures Figure Number 14-109 14-110 14-111 14-112 14-113 14-114 14-115 14-116 14-117 14-118 14-119 15-1 15-2 15-3 15-4 15-5 15-6 15-7 15-8 15-9 15-10 15-11 15-12 15-13 15-14 15-15 15-16 15-17 15-18 15-19 15-20 15-21 15-22 15-23 15-24 15-25 15-26 15-27 16-1 16-2 Title Page Number TBI Control Register Definition ....................................................................................... 14-102 TSEC-MII Connection...................................................................................................... 14-104 TSEC-GMII Connection ................................................................................................... 14-105 TSEC-RGMII Connection ................................................................................................ 14-106 TSEC-TBI Connection...................................................................................................... 14-107 TSEC-RTBI Connection ................................................................................................... 14-108 Ethernet Address Recognition Flowchart ......................................................................... 14-117 Example of TSEC Memory Structure for BD................................................................... 14-126 Buffer Descriptor Ring...................................................................................................... 14-126 Transmit Buffer Descriptor ............................................................................................... 14-127 Receive Buffer Descriptor................................................................................................. 14-129 DMA Block Diagram............................................................................................................ 15-1 DMA Operational Flow Chart .............................................................................................. 15-4 DMA Signal Summary.......................................................................................................... 15-5 DMA Register Space (Memory-Mapped)............................................................................. 15-7 DMA Mode Registers (MRn) ............................................................................................. 15-10 Status Registers (SRn) ........................................................................................................ 15-13 Basic Chaining Mode Flow Chart....................................................................................... 15-15 Current Link Descriptor Address Registers (CLNDARn) .................................................. 15-16 Source Attributes Registers (SATRn) ................................................................................. 15-17 Source Address Registers (SARn) ...................................................................................... 15-18 Source Address Registers for RapidIO Maintenance Reads (SARn) ................................. 15-19 Destination Attributes Registers (DATRn) ......................................................................... 15-20 Destination Address Registers (DARn) .............................................................................. 15-21 Destination Address Registers for RapidIO Maintenance Writes (DARn)......................... 15-22 Byte Count Registers (BCRn)............................................................................................. 15-23 Next Link Descriptor Address Registers (NLNDARn) ...................................................... 15-23 Current List Descriptor Address Registers (CLSDARn) .................................................... 15-24 Next List Descriptor Address Registers (NLSDARn) ........................................................ 15-25 Source Stride Registers (SSRn) .......................................................................................... 15-25 Destination Stride Registers (DSRn) .................................................................................. 15-26 DMA General Status Register (DGSR) .............................................................................. 15-27 External Control Interface Timing ...................................................................................... 15-34 Stride Size and Stride Distance ........................................................................................... 15-37 DMA Transaction Flow with DMA Descriptors ................................................................ 15-39 List Descriptor Format ........................................................................................................ 15-40 Link Descriptor Format....................................................................................................... 15-40 DMA Data Paths ................................................................................................................. 15-42 PCI/X Controller Block Diagram ......................................................................................... 16-2 PCI/X Interface External Signals.......................................................................................... 16-7 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 liv Freescale Semiconductor Figures Figure Number 16-3 16-4 16-5 16-6 16-7 16-8 16-10 16-9 16-11 16-12 16-13 16-14 16-15 16-16 16-17 16-18 16-19 16-20 16-21 16-22 16-23 16-24 16-25 16-26 16-27 16-28 16-29 16-30 16-31 16-32 16-33 16-34 16-35 16-36 16-37 16-38 16-39 16-40 16-41 16-42 Title Page Number PCI/X CFG_ADDR Register .............................................................................................. 16-18 PCI/X CFG_DATA Register ............................................................................................... 16-19 PCI/X INT_ACK Register .................................................................................................. 16-20 PCI/X Outbound Translation Address Registers ................................................................ 16-21 PCI/X Outbound Translation Extended Address Registers ................................................ 16-21 PCI/X Outbound Window Base Address Registers ............................................................ 16-22 PCI/X Outbound Window Attributes Registers 1–4 ........................................................... 16-23 PCI/X Outbound Window Attributes Register 0 (Default) ................................................. 16-23 PCI/X Inbound Translation Address Registers ................................................................... 16-25 PCI/X Inbound Window Base Address Registers............................................................... 16-25 PCI/X Inbound Window Base Extended Address Registers............................................... 16-26 PCI/X Inbound Window Attributes Registers..................................................................... 16-27 PCI/X Error Detect Register (ERR_DR) ............................................................................ 16-29 PCI/X Error Capture Disable Register (ERR_CAP_DR) ................................................... 16-30 PCI/X Error Enable Register (ERR_EN)............................................................................ 16-31 PCI/X Error Attributes Capture Register (ERR_ATTRIB)................................................. 16-32 PCI/X Error Address Capture Register (ERR_ADDR) ...................................................... 16-33 PCI/X Error Extended Address Capture Register (ERR_EXT_ADDR) ............................ 16-34 PCI/X Error Data Low Capture Register (ERR_DL) ......................................................... 16-34 PCI/X Error Data High Capture Register (ERR_DH) ........................................................ 16-35 PCI/X Gasket Timer Register (GAS_TIMR)...................................................................... 16-35 PCI-X Split Completion Timer Register (PCIX_TIMR) .................................................... 16-36 MPC8540 PCI/X Configuration Header ............................................................................. 16-37 PCI-X Additional Configuration Registers......................................................................... 16-37 PCI Vendor ID Register ...................................................................................................... 16-38 PCI Device ID Register....................................................................................................... 16-38 PCI Bus Command Register ............................................................................................... 16-39 PCI Bus Status Register ...................................................................................................... 16-40 PCI Revision ID Register.................................................................................................... 16-41 PCI Bus Programming Interface Register........................................................................... 16-41 PCI Subclass Code Register................................................................................................ 16-42 PCI Bus Base Class Code Register ..................................................................................... 16-42 PCI Bus Cache Line Size Register (PCLSR)...................................................................... 16-43 PCI Bus Latency Timer Register ........................................................................................ 16-43 PCI Configuration and Status Register Base Address Register (PCSRBAR) .................... 16-44 32–Bit Memory Base Address Register.............................................................................. 16-44 64-Bit Low Memory Base Address Register ...................................................................... 16-45 64-Bit High Memory Base Address Register ..................................................................... 16-46 PCI Subsystem Vendor ID Register .................................................................................... 16-46 PCI Subsystem ID Register................................................................................................. 16-47 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor lv Figures Figure Number 16-43 16-44 16-45 16-46 16-47 16-48 16-49 16-50 16-51 16-52 16-53 16-54 16-55 16-56 16-57 16-58 16-59 16-60 16-61 16-62 16-63 16-64 16-65 16-66 16-67 16-68 16-69 16-70 16-71 16-72 16-73 16-74 16-75 16-76 16-77 16-78 16-79 16-80 17-1 17-2 Title Page Number PCI Bus Capabilities Pointer Register ................................................................................ 16-47 PCI Bus Interrupt Line Register.......................................................................................... 16-47 PCI Bus Interrupt Pin Register............................................................................................ 16-48 PCI Bus Minimum Grant Register...................................................................................... 16-48 PCI Bus Maximum Latency Register ................................................................................. 16-49 PCI Bus Function Register.................................................................................................. 16-49 PCI Bus Arbiter Configuration Register ............................................................................. 16-50 PCI-X Next Capabilities ID Register.................................................................................. 16-51 PCI-X Capability Pointer Register...................................................................................... 16-51 PCI-X Command Register .................................................................................................. 16-52 PCI-X Status Register ......................................................................................................... 16-53 PCI/X Arbitration Example ................................................................................................ 16-55 PCI Single-Beat Read Transaction...................................................................................... 16-63 PCI Burst Read Transaction................................................................................................ 16-63 PCI Single-Beat Write Transaction..................................................................................... 16-64 PCI Burst Write Transaction ............................................................................................... 16-64 PCI Target-Initiated Terminations....................................................................................... 16-67 DAC Single-Beat Read Example........................................................................................ 16-69 DAC Burst Read Example .................................................................................................. 16-69 DAC Single-Beat Write Example ....................................................................................... 16-69 DAC Burst Write Example ................................................................................................. 16-70 64-Bit Dual Address Read Cycle ........................................................................................ 16-70 Standard PCI Configuration Header ................................................................................... 16-71 PCI Type 0 Configuration Translation ................................................................................ 16-75 PCI Parity Operation........................................................................................................... 16-79 AD[31:0] During PCI-X Burst/DWORD Attribute Phase.................................................. 16-83 Typical PCI-X Write Transaction........................................................................................ 16-85 Burst Memory Write Transaction........................................................................................ 16-86 Memory Read Block Transaction ....................................................................................... 16-87 DWORD (4-Byte) Write Transaction ................................................................................. 16-88 DWORD (4-Byte) Read Transaction .................................................................................. 16-88 Split Response to a Read Transaction ................................................................................. 16-90 Split Response to a DWORD (4-Byte) Write Transaction.................................................. 16-90 Split Completion Transaction Address AD[31:0]............................................................... 16-91 AD[31:0] During PCI-X Split Completion Attribute Phase ............................................... 16-92 PCI-X Split Completion Message Format .......................................................................... 16-92 PCI-X Type 0 Configuration Translation............................................................................ 16-94 AD[31:0] During PCI-X Configuration Attribute Phase .................................................... 16-95 RapidIO High Level Block Diagram .................................................................................... 17-2 RapidIO Interface Signals Model ......................................................................................... 17-7 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 lvi Freescale Semiconductor Figures Figure Number 17-3 17-4 17-5 17-6 17-7 17-8 17-9 17-10 17-11 17-12 17-13 17-14 17-15 17-16 17-17 17-18 17-19 17-20 17-21 17-22 17-23 17-24 17-25 17-26 17-27 17-28 17-29 17-30 17-31 17-32 17-33 17-34 17-35 17-36 17-37 Title Page Number Device Identity Capability Register (DIDCAR) ................................................................. 17-13 Device Information Capability Register (DICAR) ............................................................. 17-14 Assembly Identity Capability Register (AIDCAR) ............................................................ 17-14 Assembly Information Capability Register (AICAR) ........................................................ 17-15 Processing Element Features Capability Register (PEFCAR)............................................ 17-15 Switch Port Information Capability Register (SPICAR) .................................................... 17-17 Source Operations Capability Register (SOCAR) .............................................................. 17-17 Destination Operations Capability Register (DOCAR) ...................................................... 17-20 Mailbox Status Register (MSR).......................................................................................... 17-22 Port-Write and Doorbell Command and Status Register (PWDCSR) ................................ 17-23 Processing Element Logic Layer Control Command and Status Register (PELLCCSR) ................................................................................................................. 17-24 Local Configuration Space Base Address 1 Command and Status Register (LCSBA1CSR) ................................................................................................ 17-25 Base Device ID Command and Status Register (BDIDCSR)............................................. 17-25 Host Base Device ID Lock Command and Status Register (HBDIDLCSR)...................... 17-26 Component Tag Command and Status Register (CTCSR) ................................................. 17-26 8/16 LP-LVDS Port Maintenance Block Header 0 Command and Status Register (PMBH0CSR) .................................................................................................. 17-27 Port Link Time-Out Control Command and Status Register (PLTOCCSR)....................... 17-28 Port Response Time-Out Control Command and Status Register (PRTOCCSR) .............. 17-28 Port General Control Command and Status Register (PGCCSR)....................................... 17-29 Port Link Maintenance Request Command and Status Register (PLMREQCSR)............. 17-30 Port Link Maintenance Response Command and Status Register (PLMRESPCSR)......... 17-30 Port Local AckID Status Command and Status Register (PLASCSR)............................... 17-31 Port Error and Status Command and Status Register (PESCSR) ....................................... 17-32 Port Control Command and Status Register (PCCSR) ....................................................... 17-33 Configuration Register (CR) ............................................................................................... 17-34 Port Configuration Register (PCR) ..................................................................................... 17-35 Port Error Injection Register (PEIR)................................................................................... 17-36 RapidIO Outbound Window Translation Address Registers 0–8 for Standard Transactions (ROWTARn) ............................................................................................. 17-37 RapidIO Outbound Window Translation Address Registers 0–8 for Maintenance Transactions (ROWTARn) ....................................................................... 17-38 RapidIO Outbound Window Base Address Registers 1–8 (ROWBARn) .......................... 17-38 RapidIO Outbound Window Attributes Register 0 (ROWAR0) ......................................... 17-39 RapidIO Outbound Window Attributes Registers 1–8 (ROWARn) ................................... 17-39 RapidIO Inbound Window Translation Address Registers 0–4 (RIWTARn)..................... 17-41 RapidIO Inbound Window Base Address Registers 1–4 (RIWBARn)............................... 17-41 RapidIO Inbound Window Attributes Register 0 (RIWAR0) ............................................. 17-42 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor lvii Figures Figure Number 17-38 17-39 17-40 17-41 17-42 17-43 17-44 17-45 17-46 17-47 17-48 17-49 17-50 17-51 17-52 17-53 17-54 17-55 17-56 17-57 17-58 17-59 17-60 17-61 17-62 17-63 17-64 17-65 17-66 17-67 17-68 17-69 17-70 17-71 17-72 17-73 17-74 17-75 17-76 17-77 Title Page Number RapidIO Inbound Window Attributes Registers 1–4 (RIWARn)........................................ 17-42 Port Notification/Fatal Error Detect Register (PNFEDR) .................................................. 17-44 Port Notification/Fatal Error Detect Disable Register (PNFEDiR) .................................... 17-47 Port Notification/Fatal Error Interrupt Enable Register (PNFEIER).................................. 17-49 Port Error Capture Status Register (PECSR) ...................................................................... 17-52 Error Packet Capture Register 0 (EPCR0) .......................................................................... 17-52 Error Packet Capture Register 1 (EPCR1)—Type 1 Packet Format................................... 17-53 Error Packet Capture Register 1 (EPCR1)—Type 2 Packet Format................................... 17-54 Error Packet Capture Register 1 (EPCR1)—Type 5 Packet Format................................... 17-54 Error Packet Capture Register 1 (EPCR1)—Type 6 Packet Format................................... 17-55 Error Packet Capture Register 1 (EPCR1)—Type 8 Request Packet Format..................... 17-55 Error Packet Capture Register 1 (EPCR1)—Type 8 Response Packet Format .................. 17-56 Error Packet Capture Register 1 (EPCR1)—Type 10 Packet Format................................. 17-57 Error Packet Capture Register 1 (EPCR1)—Type 11 Packet Format ................................. 17-57 Error Packet Capture Register 1 (EPCR1)—Type 13 Packet Format................................. 17-58 Error Packet Capture Register 2 (EPCR2)—Type 1, 2, or 5 Packet Format ...................... 17-58 Error Packet Capture Register 2 (EPCR2)—Type 6 Packet Format................................... 17-59 Error Packet Capture Register 2 (EPCR2)—Type 8 Request Packet Format..................... 17-59 Error Packet Capture Register 2 (EPCR2)—Type 8 Response Packet Format .................. 17-60 Error Packet Capture Register 2 (EPCR2)—Type 10 Packet Format................................. 17-60 Error Packet Capture Register 2 (EPCR2)—Type 11 or 13 Packet Format........................ 17-61 Port Recoverable Error Detect Register (PREDR) ............................................................. 17-61 Port Error Recovery Threshold Register (PERTR)............................................................. 17-64 Port Retry Threshold Register (PRTR) ............................................................................... 17-64 Outbound Mode Register (OMR) ....................................................................................... 17-65 Outbound Status Register (OSR) ........................................................................................ 17-67 Outbound Descriptor Queue Dequeue Pointer Address Register (ODQDPAR)................. 17-69 Outbound Unit Source Address Registers (OSAR) ............................................................ 17-69 Outbound Destination Port Registers (ODPR) ................................................................... 17-70 Outbound Destination Attributes Register (ODATR) ......................................................... 17-70 Outbound Double-Word Count Register (ODCR) .............................................................. 17-71 Outbound Descriptor Queue Enqueue Pointer Register (ODQEPAR) ............................... 17-72 Inbound Mode Register (IMR) ........................................................................................... 17-73 Inbound Status Register (ISR) ............................................................................................ 17-74 Inbound Frame Queue Dequeue Pointer Address Registers (IFQDPAR)........................... 17-75 Inbound Frame Queue Enqueue Pointer Address Registers (IFQEPAR) ........................... 17-76 Doorbell Mode Register (DMR) ......................................................................................... 17-77 Doorbell Status Register (DSR).......................................................................................... 17-78 Doorbell Queue Dequeue Pointer Address Registers (DQDPAR)...................................... 17-80 Doorbell Queue Enqueue Pointer Address Register (DQEPAR)........................................ 17-80 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 lviii Freescale Semiconductor Figures Figure Number 17-78 17-79 17-80 17-81 17-82 17-83 17-84 17-85 17-86 17-87 18-1 18-2 18-3 18-4 18-5 18-6 18-7 18-8 18-9 18-10 18-11 18-12 18-13 18-14 18-15 18-16 18-17 18-18 18-19 18-20 19-1 19-2 19-3 19-4 19-5 19-6 19-7 19-8 19-9 19-10 Title Page Number Port-Write Mode Register (PWMR) ................................................................................... 17-81 Port-Write Status Register (PWSR) .................................................................................... 17-82 Port-Write Queue Base Address Register (PWQBAR) ...................................................... 17-83 Physical Maintenance Field (PMF) Definition ................................................................... 17-86 Outbound Message Queue Structure................................................................................. 17-113 Descriptor Dequeue Pointer and Descriptor ..................................................................... 17-116 Inbound Mailbox Structure ............................................................................................... 17-117 Inbound Doorbell Queue and Pointer Structure................................................................ 17-119 Doorbell Entry Format ...................................................................................................... 17-120 Inbound Port-Write Structure............................................................................................ 17-122 POR PLL Status Register (PORPLLSR) .............................................................................. 18-4 POR Boot Mode Status Register (PORBMSR) .................................................................... 18-5 POR I/O Impedance Status and Control Register (PORIMPSCR)....................................... 18-6 POR Device Status Register (PORDEVSR) ......................................................................... 18-8 POR Debug Mode Status Register (PORDBGMSR) ........................................................... 18-9 POR Configuration Register (GPPORCR) ........................................................................... 18-9 General-Purpose I/O Control Register (GPIOCR).............................................................. 18-10 General-Purpose Output Data Register (GPOUTDR) ........................................................ 18-11 General-Purpose Output Data Register (GPINDR) ............................................................ 18-12 Alternate Function Pin Multiplex Control Register (PMUXCR) ....................................... 18-13 Device Disable Register (DEVDISR)................................................................................. 18-13 Power Management Control and Status Register (POWMGTCSR) .................................. 18-15 Machine Check Summary Register (MCPSUMR) ............................................................. 18-17 Processor Version Register (PVR) ...................................................................................... 18-17 System Version Register (SVR).......................................................................................... 18-18 Clock Out Control Register (CLKOCR)............................................................................. 18-18 DDR DLL Control Register (DDRDLLCR)....................................................................... 18-19 Local Bus DLL Control Register (LBDLLCR) .................................................................. 18-20 e500 Core Power Management State Diagram ................................................................... 18-21 MPC8540 Power Management Handshaking Signals ........................................................ 18-26 Performance Monitor Block Diagram................................................................................... 19-2 Performance Monitor Global Control Register (PMGC0).................................................... 19-5 Performance Monitor Local Control Register A0 (PMLCA0) ............................................. 19-6 Performance Monitor Local Control Registers A1–A8 (PMLCA1–PMLCA8) ................... 19-6 Performance Monitor Local Control Register B0 (PMLCB0).............................................. 19-7 Performance Monitor Local Control Registers B1–B8 (PMLCB1–PMLCB8).................... 19-8 Performance Monitor Counter Register 0 (PMC0)............................................................. 19-10 Performance Monitor Counter Registers 1–8 (PMC1–PMC8)........................................... 19-10 Duration Threshold Event Sequence Timing Diagram ....................................................... 19-12 Burst Size, Distance, Granularity, and Burstiness Counting............................................... 19-13 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor lix Figures Figure Number 19-11 20-1 20-2 20-3 20-4 20-5 20-6 20-7 20-8 20-9 20-10 20-11 20-12 20-13 20-14 20-15 20-16 20-17 20-18 20-19 20-20 20-21 20-22 20-23 21-1 21-2 21-3 21-4 21-5 21-6 21-7 21-8 21-9 21-10 21-11 21-12 21-13 21-14 21-15 21-16 Title Page Number Burstiness Counting Timing Diagram ................................................................................ 19-15 Debug and Watchpoint Monitor Block Diagram .................................................................. 20-2 Watchpoint Monitor Control Register 0 (WMCR0) ........................................................... 20-11 Watchpoint Monitor Control Register 1 (WMCR1) ........................................................... 20-13 Watchpoint Monitor Address Register (WMAR) ............................................................... 20-13 Watchpoint Monitor Address Mask Register (WMAMR).................................................. 20-14 Watchpoint Monitor Transaction Mask Register (WMTMR)............................................. 20-15 Watchpoint Monitor Status Register (WMSR) ................................................................... 20-16 Trace Buffer Control Register 0 (TBCR0).......................................................................... 20-17 Trace Buffer Control Register 1 (TBCR1).......................................................................... 20-18 Trace Buffer Address Register (TBAR).............................................................................. 20-19 Trace Buffer Address Mask Register (TBAMR) ................................................................ 20-20 Trace Buffer Transaction Mask Register (TBTMR) ........................................................... 20-20 Trace Buffer Status Register (TBSR).................................................................................. 20-21 Trace Buffer Access Control Register (TBACR) ............................................................... 20-22 Trace Buffer Read High Register (TBADHR).................................................................... 20-23 Trace Buffer Access Data Register (TBADR) .................................................................... 20-23 Programmed Context ID Register (PCIDR) ....................................................................... 20-24 Current Context ID Register (CCIDR) ............................................................................... 20-24 Trigger Out Source Register (TOSR).................................................................................. 20-25 e500 Coherency Module Dispatch (CMD) Trace Buffer Entry.......................................... 20-30 DDR Trace Buffer Entry ..................................................................................................... 20-31 PCI Trace Buffer Entry ....................................................................................................... 20-32 RapidIO Trace Buffer Entry................................................................................................ 20-32 10/100 Fast Ethernet Controller in Relation to the OSI Protocol Stack ............................... 21-1 Ethernet/IEEE 802.3 Frame Structure .................................................................................. 21-2 Ethernet/IEEE 802.3 Frame Structure With More Details.................................................... 21-4 FEC Block Diagram.............................................................................................................. 21-5 IEVENT Register Definition .............................................................................................. 21-14 IMASK Register Definition ................................................................................................ 21-17 EDIS Register Definition .................................................................................................... 21-18 MINFLR Register Definition.............................................................................................. 21-18 PTV Register Definition...................................................................................................... 21-19 DMACTRL Register Definition ......................................................................................... 21-20 FIFO_PAUSE_CTRL Register Definition.......................................................................... 21-22 FIFO_TX_THR Register Definition ................................................................................... 21-23 FIFO_TX_STARVE Register Definition ............................................................................ 21-24 FIFO_TX_STARVE_SHUTOFF Register Definition ........................................................ 21-25 TCTRL Register Definition ................................................................................................ 21-25 TSTAT Register Definition ................................................................................................. 21-26 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 lx Freescale Semiconductor Figures Figure Number 21-17 21-18 21-19 21-20 21-21 21-22 21-23 21-24 21-25 21-26 21-27 21-28 21-29 21-30 21-31 21-32 21-33 21-34 21-35 21-36 21-37 21-38 21-39 21-40 21-41 21-42 21-43 21-44 21-45 21-46 21-47 Title Page Number TBDLEN Register Definition ............................................................................................. 21-27 CTBPTR Register Definition .............................................................................................. 21-27 TBPTR Register Definition................................................................................................. 21-28 TBASE Register Definition ................................................................................................ 21-28 OSTBD Register Definition................................................................................................ 21-29 OSTBDP Register Definition.............................................................................................. 21-31 RCTRL Register Definition ................................................................................................ 21-31 RSTAT Register Definition ................................................................................................. 21-32 RBDLEN Register Definition............................................................................................. 21-33 CRBPTR Register Definition.............................................................................................. 21-34 MRBLR Register Definition ............................................................................................... 21-34 Receive Buffer Descriptor Pointer Definition..................................................................... 21-35 RBASE Register Definition ................................................................................................ 21-35 MACCFG1 Register Definition .......................................................................................... 21-38 MACCFG2 Register Definition .......................................................................................... 21-40 IPGIFG Register Definition ................................................................................................ 21-41 Half-Duplex Register Definition......................................................................................... 21-41 Maximum Frame Length Register Definition..................................................................... 21-42 Interface Status Register Definition.................................................................................... 21-43 Station Address Part 1 Register Definition ......................................................................... 21-44 Station Address Part 2 Register Definition ......................................................................... 21-44 IADDRn Register Definition .............................................................................................. 21-45 GADDRn Register Definition............................................................................................. 21-46 ATTR Register Definition ................................................................................................... 21-47 ATTRELI Register Definition............................................................................................. 21-48 FEC–MII Connection.......................................................................................................... 21-49 Ethernet Address Recognition Flowchart ........................................................................... 21-56 Example of FEC Memory Structure for BD ....................................................................... 21-64 Buffer Descriptor Ring........................................................................................................ 21-64 Transmit Buffer Descriptor ................................................................................................. 21-65 Receive Buffer Descriptor................................................................................................... 21-67 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor lxi Figures Figure Number Title Page Number MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 lxii Freescale Semiconductor Tables Table Number 1 2-1 2-2 2-3 2-4 2-5 2-6 2-7 2-8 2-9 3-1 3-2 3-3 3-4 4-1 4-2 4-3 4-4 4-5 4-6 4-7 4-8 4-9 4-10 4-11 4-12 4-13 4-14 4-15 4-16 4-17 4-18 4-19 4-20 4-21 4-22 4-23 4-24 Title Tables Page Number Acronyms and Abbreviated Terms................................................................................... lxxxviii Target Interface Codes ............................................................................................................ 2-1 Local Access Windows Example............................................................................................ 2-2 Format of ATMU Window Definitions................................................................................... 2-3 Local Access Register Memory Map...................................................................................... 2-5 LAWBARn Field Descriptions ............................................................................................... 2-6 LAWARn Field Descriptions .................................................................................................. 2-6 Overlapping Local Access Windows ...................................................................................... 2-7 Local Memory Configuration, Control, and Status Register Summary................................ 2-12 Memory Map......................................................................................................................... 2-16 MPC8540 Signal Reference by Functional Block .................................................................. 3-4 MPC8540 Alphabetical Signal Reference .............................................................................. 3-9 MPC8540 Reset Configuration Signals ................................................................................ 3-16 Output Signal States During System Reset........................................................................... 3-17 Signal Summary ...................................................................................................................... 4-1 System Control Signals—Detailed Signal Descriptions ......................................................... 4-2 Clock Signals—Detailed Signal Descriptions ........................................................................ 4-3 Local Configuration Control Register Map ............................................................................ 4-4 CCSRBAR Field Descriptions................................................................................................ 4-5 ALTCBAR Field Descriptions ................................................................................................ 4-6 ALTCAR Field Descriptions................................................................................................... 4-7 BPTR Field Descriptions ........................................................................................................ 4-8 CCB Clock PLL Ratio .......................................................................................................... 4-13 e500 Core Clock PLL Ratios ................................................................................................ 4-13 Boot ROM Location.............................................................................................................. 4-14 Host/Agent Configuration..................................................................................................... 4-15 CPU Boot Configuration....................................................................................................... 4-15 Boot Sequencer Configuration.............................................................................................. 4-16 TSEC Width Configuration................................................................................................... 4-17 TSEC1 Protocol Configuration............................................................................................. 4-17 TSEC2 Protocol Configuration............................................................................................. 4-18 RapidIO Transmit Clock Source ........................................................................................... 4-18 RapidIO Device ID ............................................................................................................... 4-19 PCI-32 Configuration............................................................................................................ 4-19 PCI I/O Impedance................................................................................................................ 4-19 PCI Arbiter Configuration .................................................................................................... 4-20 PCI Debug Configuration ..................................................................................................... 4-20 PCI/PCI-X Configuration ..................................................................................................... 4-20 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor lxiii Tables Table Number 4-25 4-26 4-27 4-28 4-29 4-30 5-1 5-2 5-3 5-4 5-5 5-6 5-7 5-8 6-1 6-2 6-3 6-4 6-5 6-6 6-7 6-8 6-9 6-10 6-11 6-12 6-13 6-14 6-15 6-16 6-17 6-18 6-19 6-20 6-21 6-22 6-23 6-24 6-25 6-26 Title Page Number Memory Debug Configuration.............................................................................................. 4-21 DDR Debug Configuration ................................................................................................... 4-21 PCI Output Hold Configuration (cfg_pci_mode = 1) ........................................................... 4-22 PCI-X Output Hold Configuration (cfg_pci_mode = 0) ....................................................... 4-22 Local Bus Output Hold Configuration.................................................................................. 4-22 General-Purpose POR Configuration.................................................................................... 4-23 Revision Level-to-Device Marking Cross-Reference............................................................. 5-5 Performance Monitor APU Instructions ............................................................................... 5-12 Cache Block Lock and Unlock APU Instructions ................................................................ 5-12 Scalar and Vector Embedded Floating-Point APU Instructions ........................................... 5-13 BTB Locking APU Instructions............................................................................................ 5-13 Interrupt Registers................................................................................................................. 5-22 Interrupt Vector Registers and Exception Conditions........................................................... 5-24 Differences Between the e500 Core and the PowerQUICC III Core Implementation ......... 5-33 Book E Special-Purpose Registers (by SPR Abbreviation).................................................... 6-4 Implementation-Specific SPRs (by SPR Abbreviation) ......................................................... 6-6 XER Field Description............................................................................................................ 6-9 BI Operand Settings for CR Fields ......................................................................................... 6-9 CR0 Bit Descriptions ............................................................................................................ 6-11 MSR Field Descriptions........................................................................................................ 6-12 PVR Field Descriptions ........................................................................................................ 6-14 TCR Field Descriptions ........................................................................................................ 6-15 TSR Field Descriptions ......................................................................................................... 6-16 IVOR Assignments ............................................................................................................... 6-20 ESR Field Descriptions ......................................................................................................... 6-21 MCSR Field Descriptions ..................................................................................................... 6-23 BBEAR Field Descriptions ................................................................................................... 6-25 BBTAR Field Descriptions ................................................................................................... 6-25 BUCSR Field Descriptions ................................................................................................... 6-26 HID0 Field Descriptions ....................................................................................................... 6-27 HID1 Field Descriptions ....................................................................................................... 6-28 L1CSR0 Field Descriptions .................................................................................................. 6-30 L1CSR1 Field Descriptions .................................................................................................. 6-31 L1CFG0 Field Descriptions .................................................................................................. 6-32 L1CFG1 Field Descriptions .................................................................................................. 6-33 MMUCSR0 Field Descriptions............................................................................................. 6-34 MMUCFG Field Descriptions .............................................................................................. 6-35 TLB0CFG Field Descriptions ............................................................................................... 6-36 TLB1CFG Field Descriptions ............................................................................................... 6-36 MAS0 Field Descriptions—MMU Read/Write and Replacement Control .......................... 6-37 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 lxiv Freescale Semiconductor Tables Table Number 6-27 6-28 6-29 6-30 6-31 6-32 6-33 6-34 6-35 6-36 6-37 6-38 6-39 6-40 6-41 6-42 6-43 7-1 7-2 7-3 7-4 7-5 7-6 7-7 7-8 7-9 7-10 7-11 7-12 7-13 7-14 7-15 7-16 7-17 7-18 7-19 7-20 7-21 7-22 7-23 Title Page Number MAS1 Field Descriptions—Descriptor Context and Configuration Control........................ 6-38 MAS2 Field Descriptions—EPN and Page Attributes ......................................................... 6-39 MAS3 Field Descriptions–RPN and Access Control ........................................................... 6-40 MAS4 Field Descriptions—Hardware Replacement Assist Configuration.......................... 6-40 MAS6—TLB Search Context Register 0.............................................................................. 6-41 DBCR0 Field Descriptions ................................................................................................... 6-42 DBCR1 Field Descriptions ................................................................................................... 6-44 DBCR2 Field Descriptions ................................................................................................... 6-45 DBSR Field Descriptions...................................................................................................... 6-46 SPEFSCR Field Descriptions................................................................................................ 6-48 ACC FIeld Descriptions........................................................................................................ 6-50 Supervisor-Level PMRs (PMR[5] = 1) ................................................................................. 6-51 User-Level PMRs (PMR[5] = 0) (Read Only)...................................................................... 6-51 PMGC0 Field Descriptions ................................................................................................... 6-52 PMLCa0–PMLCa3 Field Descriptions ................................................................................. 6-53 PMLCb0–PMLCb3 Field Descriptions ................................................................................ 6-54 PMC0–PMC3 Field Descriptions ......................................................................................... 6-54 L2/SRAM Memory-Mapped Registers................................................................................... 7-6 L2CTL Field Descriptions ...................................................................................................... 7-8 L2CEWARn Field Descriptions............................................................................................ 7-10 L2CEWCRn Field Descriptions............................................................................................ 7-11 L2SRBARn Field Descriptions............................................................................................. 7-12 L2ERRINJHI Field Description............................................................................................ 7-13 L2ERRINJLO Field Description .......................................................................................... 7-14 L2ERRINJCTL Field Descriptions....................................................................................... 7-14 L2CAPTDATAHI Field Description..................................................................................... 7-15 L2CAPTDATALO Field Description.................................................................................... 7-16 L2CAPTECC Field Descriptions .......................................................................................... 7-16 L2ERRDET Field Descriptions ............................................................................................ 7-16 L2ERRDIS Field Descriptions.............................................................................................. 7-17 L2ERRINTEN Field Descriptions ........................................................................................ 7-18 L2ERRATTR Field Descriptions .......................................................................................... 7-19 L2ERRADDR Field Description .......................................................................................... 7-20 L2ERRCTL Field Descriptions ............................................................................................ 7-20 Fastest Read Timing—Hit in L2 ........................................................................................... 7-22 PLRU Bit Update Algorithm ................................................................................................ 7-28 PLRU-Based Victim Selection Mechanism .......................................................................... 7-29 L2 Cache States..................................................................................................................... 7-30 State Transitions Due to Core-Initiated Transactions ........................................................... 7-31 State Transitions Due to System-Initiated Transactions ....................................................... 7-33 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor lxv Tables Table Number 8-1 8-2 8-3 8-4 8-5 8-6 8-7 9-1 9-2 9-3 9-4 9-5 9-6 9-7 9-8 9-9 9-10 9-11 9-12 9-13 9-14 9-15 9-16 9-17 9-18 9-19 9-20 9-21 9-22 9-23 9-24 9-25 9-26 9-27 9-28 9-29 9-30 9-31 9-32 9-33 Title Page Number ECM Memory Map................................................................................................................. 8-3 EEBACR Field Descriptions .................................................................................................. 8-3 EEBPCR Field Descriptions ................................................................................................... 8-4 EEDR Field Descriptions........................................................................................................ 8-6 EEER Field Descriptions ........................................................................................................ 8-7 EEATR Field Descriptions...................................................................................................... 8-7 EEADR Field Descriptions ..................................................................................................... 8-8 DDR Memory Interface Signal Summary .............................................................................. 9-4 Memory Address Signal Mappings......................................................................................... 9-4 Memory Interface Signals—Detailed Signal Descriptions ..................................................... 9-5 Clock Signals—Detailed Signal Descriptions ........................................................................ 9-8 DDR Memory Controller Memory Map................................................................................. 9-9 CSn_BNDS Field Descriptions............................................................................................. 9-10 CSn_CONFIG Field Descriptions ........................................................................................ 9-11 TIMING_CFG_1 Field Descriptions .................................................................................... 9-12 TIMING_CFG_2 Register Field Descriptions...................................................................... 9-13 DDR_SDRAM_CFG Field Descriptions.............................................................................. 9-15 DDR_SDRAM_MODE Field Descriptions.......................................................................... 9-16 DDR_SDRAM_INTERVAL Field Descriptions .................................................................. 9-17 DATA_ERR_INJECT_HI Field Descriptions....................................................................... 9-17 DATA_ERR_INJECT_LO Field Descriptions ..................................................................... 9-18 ECC_ERR_INJECT Field Descriptions ............................................................................... 9-19 CAPTURE_DATA_HI Field Descriptions............................................................................ 9-19 CAPTURE_DATA_LO Field Descriptions........................................................................... 9-20 CAPTURE_ECC Field Descriptions .................................................................................... 9-20 ERR_DETECT Field Descriptions ....................................................................................... 9-21 ERR_DISABLE Field Descriptions...................................................................................... 9-22 ERR_INT_EN Field Descriptions ........................................................................................ 9-23 CAPTURE_ATTRIBUTES Field Descriptions .................................................................... 9-23 CAPTURE_ADDRESS Field Descriptions .......................................................................... 9-24 ERR_SBE Field Descriptions ............................................................................................... 9-25 Byte Lane to Data Relationship ............................................................................................ 9-30 Supported DDR SDRAM Device Configurations ................................................................ 9-31 DDR SDRAM Address Multiplexing ................................................................................... 9-32 SDRAM Command Table ..................................................................................................... 9-34 DDR SDRAM Interface Timing Intervals ............................................................................ 9-35 DDR SDRAM Power-Saving Modes Refresh Configuration............................................... 9-43 Memory Controller–Data Beat Ordering .............................................................................. 9-45 DDR SDRAM ECC Syndrome Encoding ............................................................................ 9-46 DDR SDRAM ECC Syndrome Encoding (Check Bits) ....................................................... 9-48 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 lxvi Freescale Semiconductor Tables Table Number 9-34 9-35 10-1 10-2 10-3 10-4 10-5 10-6 10-7 10-8 10-9 10-10 10-11 10-12 10-13 10-14 10-15 10-16 10-17 10-18 10-19 10-20 10-21 10-22 10-23 10-24 10-25 10-26 10-27 10-28 10-29 10-30 10-31 10-32 10-33 10-34 10-35 10-36 10-37 10-38 Title Page Number Memory Controller Errors .................................................................................................... 9-49 Memory Interface Configuration Register Initialization Parameters.................................... 9-49 Interrupt Source Signalling Options ..................................................................................... 10-2 Processor Interrupts Generated Outside the Core—Types and Sources ............................... 10-4 e500 Core-Generated Interrupts that Cause a Wake-up ........................................................ 10-5 Internal Interrupt Assignments.............................................................................................. 10-7 PIC Interface Signals ............................................................................................................ 10-8 Interrupt Signals—Detailed Signal Descriptions .................................................................. 10-8 PIC Register Address Map.................................................................................................. 10-10 FRR Field Descriptions....................................................................................................... 10-16 GCR Field Descriptions ...................................................................................................... 10-17 VIR Field Descriptions ....................................................................................................... 10-18 PIR Field Descriptions ........................................................................................................ 10-18 IPIVPRn Field Descriptions................................................................................................ 10-19 SVR Field Descriptions ...................................................................................................... 10-20 TFRR Field Descriptions .................................................................................................... 10-21 GTCCRn Field Descriptions............................................................................................... 10-21 GTBCRn Field Descriptions............................................................................................... 10-22 GTVPRn Field Descriptions ............................................................................................... 10-23 GTDRn Field Descriptions ................................................................................................. 10-23 Parameters for Hourly Interrupt Timer Cascade Example.................................................. 10-24 TCR Field Descriptions ...................................................................................................... 10-25 IRQSR0 Field Descriptions ................................................................................................ 10-26 IRQSR1 Field Descriptions ................................................................................................ 10-27 CISR0 Field Descriptions ................................................................................................... 10-28 CISR1 Field Descriptions ................................................................................................... 10-28 PMnMR0 Field Descriptions .............................................................................................. 10-29 PMnMR1 Field Descriptions .............................................................................................. 10-30 MSGRn Field Descriptions................................................................................................. 10-30 MER Field Descriptions...................................................................................................... 10-31 MSR Field Descriptions...................................................................................................... 10-32 EIVPRn Field Descriptions................................................................................................. 10-33 EIDRn Field Descriptions................................................................................................... 10-34 IIVPRn Field Descriptions.................................................................................................. 10-35 IIDRn Field Descriptions .................................................................................................... 10-36 MIVPRn Field Descriptions................................................................................................ 10-36 MIDRn Field Descriptions.................................................................................................. 10-37 Per-CPU Registers—Private Access Address Offsets ........................................................ 10-38 IPIDRn Field Descriptions.................................................................................................. 10-40 CTPR Field Descriptions .................................................................................................... 10-40 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor lxvii Tables Table Number 10-39 10-40 10-41 11-1 11-2 11-3 11-4 11-5 11-6 11-7 11-8 11-9 12-1 12-2 12-3 12-4 12-5 12-6 12-7 12-8 12-9 12-10 12-11 12-12 12-13 12-14 12-15 12-16 12-17 12-18 12-19 12-20 12-21 12-22 12-23 12-24 13-1 13-2 13-3 13-4 Title Page Number WHOAMI Field Descriptions ............................................................................................. 10-41 IACK Field Descriptions .................................................................................................... 10-42 EOI Field Descriptions........................................................................................................ 10-42 I2C Interface Signal Description........................................................................................... 11-3 I2C Interface Signal—Detailed Signal Descriptions............................................................. 11-4 I2C Memory Map .................................................................................................................. 11-4 I2CADR Field Descriptions.................................................................................................. 11-5 I2CFDR Field Descriptions .................................................................................................. 11-6 I2CCR Field Descriptions ..................................................................................................... 11-7 I2CSR Field Descriptions ..................................................................................................... 11-8 I2CDR Field Description ...................................................................................................... 11-9 I2CDFSRR Field Descriptions............................................................................................ 11-10 DUART Signal Overview ..................................................................................................... 12-3 DUART Signals—Detailed Signal Descriptions .................................................................. 12-4 DUART Register Summary .................................................................................................. 12-5 URBR Field Definition ......................................................................................................... 12-6 UTHR Field Definition ......................................................................................................... 12-7 UDMB Field Definition ........................................................................................................ 12-7 UDLB Field Definition ......................................................................................................... 12-8 Baud Rate Examples ............................................................................................................. 12-8 UIER Field Definitions ......................................................................................................... 12-9 UIIR Field Definitions ........................................................................................................ 12-10 UIIR IID Bits Summary...................................................................................................... 12-11 UFCR Field Definitions ...................................................................................................... 12-12 ULCR Field Definitions...................................................................................................... 12-13 Parity Selection Using ULCR[PEN], ULCR[SP], and ULCR[EPS] .................................. 12-14 UMCR Field Definitions..................................................................................................... 12-14 ULSR Field Definitions ...................................................................................................... 12-15 UMSR Field Definitions ..................................................................................................... 12-17 USCR Field Definitions ...................................................................................................... 12-17 UAFR Field Definitions...................................................................................................... 12-18 UDSR Field Definitions...................................................................................................... 12-19 UDSR[TXRDY] Set Conditions ......................................................................................... 12-19 UDSR[TXRDY] Cleared Conditions.................................................................................. 12-19 UDSR[RXRDY] Set Conditions......................................................................................... 12-19 UDSR[RXRDY] Cleared .................................................................................................... 12-20 Signal Properties—Summary................................................................................................ 13-5 Local Bus Controller Detailed Signal Descriptions .............................................................. 13-6 Local Bus Controller Memory Map...................................................................................... 13-9 BRn Field Descriptions....................................................................................................... 13-11 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 lxviii Freescale Semiconductor Tables Table Number 13-5 13-6 13-7 13-8 13-9 13-10 13-11 13-12 13-13 13-14 13-15 13-16 13-17 13-18 13-19 13-20 13-21 13-22 13-23 13-24 13-25 13-26 13-27 13-28 13-29 13-30 13-31 13-32 13-33 13-34 13-35 13-36 13-37 13-38 13-39 13-40 13-41 13-42 13-43 13-44 Title Page Number Memory Bank Sizes in Relation to Address Mask ............................................................. 13-13 ORn—GPCM Field Descriptions ....................................................................................... 13-14 ORn—UPM Field Descriptions .......................................................................................... 13-16 ORn—SDRAM Field Descriptions .................................................................................... 13-18 MAR Field Description....................................................................................................... 13-19 MxMR Field Descriptions................................................................................................... 13-19 MRTPR Field Descriptions ................................................................................................. 13-22 MDR Field Description....................................................................................................... 13-22 LSDMR Field Descriptions ................................................................................................ 13-23 LURT Field Descriptions .................................................................................................... 13-25 LSRT Field Descriptions..................................................................................................... 13-25 LTESR Field Descriptions .................................................................................................. 13-26 LTEDR Field Descriptions.................................................................................................. 13-27 LTEIR Field Descriptions ................................................................................................... 13-29 LTEATR Field Descriptions................................................................................................ 13-30 LTEAR Field Descriptions.................................................................................................. 13-30 LBCR Field Descriptions.................................................................................................... 13-31 LCRR Field Descriptions.................................................................................................... 13-32 GPCM Write Control Signal Timing for LCRR[CLKDIV] = 4 or 8.................................. 13-40 GPCM Read Control Signal Timing for LCRR[CLKDIV] = 4 or 8................................... 13-41 GPCM Write Control Signal Timing for LCRR[CLKDIV] = 2 ......................................... 13-42 GPCM Read Control Signal Timing for LCRR[CLKDIV] = 2.......................................... 13-43 Boot Bank Field Values after Reset .................................................................................... 13-51 SDRAM Interface Commands ............................................................................................ 13-53 UPM Routines Start Addresses........................................................................................... 13-64 RAM Word Field Descriptions ........................................................................................... 13-69 MxMR Loop Field Use ....................................................................................................... 13-74 UPM Address Multiplexing ................................................................................................ 13-75 Data Bus Requirements For Read Cycle............................................................................. 13-89 Micron SDRAM Devices.................................................................................................... 13-91 LADn Signal Connections to 128-Mbyte SDRAM ............................................................ 13-93 Logical Address Bus Partitioning ....................................................................................... 13-94 SDRAM Device Address Port during Address Phase ........................................................ 13-94 SDRAM Device Address Port during READ/WRITE Command...................................... 13-94 Register Settings for 128-Mbytes SDRAMs....................................................................... 13-95 Logical Address Partitioning .............................................................................................. 13-96 SDRAM Device Address Port During Address Phase........................................................ 13-96 SDRAM Device Address Port During READ/WRITE Command..................................... 13-96 Register Settings for 512-Mbyte SDRAMs ........................................................................ 13-96 SDRAM Capacitance .......................................................................................................... 13-99 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor lxix Tables Table Number 13-45 13-46 13-47 13-48 14-1 14-2 14-3 14-4 14-5 14-6 14-7 14-8 14-9 14-10 14-11 14-12 14-13 14-14 14-15 14-16 14-17 14-18 14-19 14-20 14-21 14-22 14-23 14-24 14-25 14-26 14-27 14-28 14-29 14-30 14-31 14-32 14-33 14-34 14-35 14-36 Title Page Number SDRAM AC Characteristics ............................................................................................... 13-99 Local Bus to MSC8101 HDI16 Connections.................................................................... 13-105 UPM Synchronization Cycles ........................................................................................... 13-112 EHPI Signals .................................................................................................................... 13-119 TSEC Signals—Detailed Signal Descriptions .................................................................... 14-10 Module Memory Map Summary......................................................................................... 14-13 Module Memory Map ......................................................................................................... 14-14 IEVENT Field Descriptions................................................................................................ 14-21 IMASK Field Descriptions ................................................................................................. 14-24 EDIS Field Descriptions ..................................................................................................... 14-25 ECNTRL Field Descriptions............................................................................................... 14-25 MINFLR Field Descriptions ............................................................................................... 14-27 PTV Field Descriptions ....................................................................................................... 14-27 DMACTRL Field Descriptions........................................................................................... 14-28 TBIPA Field Descriptions ................................................................................................... 14-30 FIFO_PAUSE_CTRL Field Descriptions .......................................................................... 14-31 FIFO_TX_THR Field Descriptions .................................................................................... 14-32 FIFO_TX_STARVE Field Descriptions ............................................................................. 14-32 FIFO_TX_STARVE_SHUTOFF Field Descriptions.......................................................... 14-33 TCTRL Field Descriptions.................................................................................................. 14-33 TSTAT Field Descriptions................................................................................................... 14-34 TBDLEN Field Descriptions .............................................................................................. 14-35 TXIC Field Descriptions ..................................................................................................... 14-36 CTBPTR Field Descriptions ............................................................................................... 14-37 TBPTR Field Descriptions .................................................................................................. 14-37 TBASE Field Descriptions.................................................................................................. 14-38 OSTBD Field Descriptions ................................................................................................. 14-39 OSTBDP Field Descriptions ............................................................................................... 14-40 RCTRL Field Descriptions ................................................................................................. 14-41 RSTAT Field Descriptions .................................................................................................. 14-42 RBDLEN Field Descriptions .............................................................................................. 14-42 RXIC Field Descriptions..................................................................................................... 14-43 CRBPTR Field Descriptions ............................................................................................... 14-44 MRBLR Field Descriptions ................................................................................................ 14-45 RBPTR Field Descriptions.................................................................................................. 14-45 RBASE Field Descriptions ................................................................................................. 14-46 MACCFG1 Field Descriptions ........................................................................................... 14-49 MACCFG2 Field Descriptions ........................................................................................... 14-51 IPGIFG Field Descriptions ................................................................................................. 14-52 HAFDUP Field Descriptions .............................................................................................. 14-53 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 lxx Freescale Semiconductor Tables Table Number 14-37 14-38 14-39 14-40 14-41 14-42 14-43 14-44 14-45 14-46 14-47 14-48 14-49 14-50 14-51 14-52 14-53 14-54 14-55 14-56 14-57 14-58 14-59 14-60 14-61 14-62 14-63 14-64 14-65 14-66 14-67 14-68 14-69 14-70 14-71 14-72 14-73 14-74 14-75 14-76 Title Page Number MAXFRM Field Descriptions ............................................................................................ 14-54 MIIMCFG Field Descriptions............................................................................................. 14-55 MIIMCOM Field Descriptions ........................................................................................... 14-56 MIIMADD Field Descriptions............................................................................................ 14-56 MIIMCON Field Descriptions ............................................................................................ 14-57 MIIMSTAT Field Descriptions ........................................................................................... 14-57 MIIMIND Field Descriptions ............................................................................................. 14-58 IFSTAT Field Descriptions ................................................................................................. 14-59 MACSTNADDR1 Field Descriptions ................................................................................ 14-60 MACSTNADDR2 Field Descriptions ................................................................................ 14-60 TR64 Field Descriptions ..................................................................................................... 14-61 TR127 Field Descriptions ................................................................................................... 14-62 TR255 Field Descriptions ................................................................................................... 14-62 TR511 Field Descriptions ................................................................................................... 14-63 TR1K Field Descriptions .................................................................................................... 14-63 TRMAX Field Descriptions................................................................................................ 14-64 TRMGV Field Descriptions................................................................................................ 14-64 RBYT Field Descriptions.................................................................................................... 14-65 RPKT Field Descriptions .................................................................................................... 14-65 RFCS Field Descriptions .................................................................................................... 14-66 RMCA Field Descriptions .................................................................................................. 14-66 RBCA Field Descriptions ................................................................................................... 14-67 RXCF Field Descriptions.................................................................................................... 14-67 RXPF Field Descriptions .................................................................................................... 14-68 RXUO Field Descriptions ................................................................................................... 14-68 RALN Field Descriptions ................................................................................................... 14-69 RFLR Field Descriptions .................................................................................................... 14-69 RCDE Field Descriptions.................................................................................................... 14-70 RCSE Field Descriptions .................................................................................................... 14-70 RUND Field Descriptions ................................................................................................... 14-71 ROVR Field Descriptions ................................................................................................... 14-71 RFRG Field Descriptions.................................................................................................... 14-72 RJBR Field Descriptions..................................................................................................... 14-72 RDRP Field Descriptions.................................................................................................... 14-73 TBYT Field Descriptions.................................................................................................... 14-73 TPKT Field Descriptions .................................................................................................... 14-74 TMCA Field Descriptions................................................................................................... 14-74 TBCA Field Descriptions.................................................................................................... 14-75 TXPF Field Descriptions .................................................................................................... 14-75 TDFR Field Descriptions .................................................................................................... 14-76 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor lxxi Tables Table Number 14-77 14-78 14-79 14-80 14-81 14-82 14-83 14-84 14-85 14-86 14-87 14-88 14-89 14-90 14-91 14-92 14-93 14-94 14-95 14-96 14-97 14-98 14-99 14-100 14-101 14-102 14-103 14-104 14-105 14-106 14-107 14-108 14-109 14-110 14-111 14-112 14-113 14-114 14-115 14-116 Title Page Number TEDF Field Descriptions .................................................................................................... 14-76 TSCL Field Descriptions .................................................................................................... 14-77 TMCL Field Descriptions ................................................................................................... 14-77 TLCL Field Descriptions .................................................................................................... 14-78 TXCL Field Descriptions.................................................................................................... 14-78 TNCL Field Descriptions.................................................................................................... 14-79 TDRP Field Descriptions .................................................................................................... 14-79 TJBR Field Descriptions ..................................................................................................... 14-80 TFCS Field Descriptions..................................................................................................... 14-80 TXCF Field Descriptions .................................................................................................... 14-81 TOVR Field Descriptions ................................................................................................... 14-81 TUND Field Descriptions ................................................................................................... 14-82 TFRG Field Descriptions .................................................................................................... 14-82 CAR1 Field Descriptions .................................................................................................... 14-83 CAR2 Field Descriptions .................................................................................................... 14-84 CAM1 Field Descriptions ................................................................................................... 14-85 CAM2 Field Descriptions ................................................................................................... 14-87 IADDRn Field Description ................................................................................................. 14-88 GADDRn Field Description................................................................................................ 14-89 ATTR Field Descriptions .................................................................................................... 14-89 ATTRELI Field Descriptions .............................................................................................. 14-90 TBI MII Register Set........................................................................................................... 14-91 CR Field Descriptions ......................................................................................................... 14-92 SR Descriptions................................................................................................................... 14-93 ANA Field Descriptions...................................................................................................... 14-94 PAUSE Priority Resolution................................................................................................. 14-96 ANLPBPA Field Descriptions ............................................................................................ 14-97 ANEX Field Descriptions ................................................................................................... 14-98 ANNPT Field Descriptions ................................................................................................. 14-99 ANLPANP Field Descriptions ............................................................................................ 14-99 EXST Field Descriptions .................................................................................................. 14-100 JD Field Description ......................................................................................................... 14-102 TBICON Field Descriptions ............................................................................................. 14-103 GMII, MII, and TBI Signal Multiplexing ......................................................................... 14-108 RGMII and RTBI Signal Multiplexing ............................................................................. 14-109 Shared Signals....................................................................................................................14-111 Steps of Minimum Register Initialization ..........................................................................14-111 Flow Control Frame Structure .......................................................................................... 14-120 Non-Error Transmit Interrupts .......................................................................................... 14-121 Non-Error Receive Interrupts............................................................................................ 14-121 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 lxxii Freescale Semiconductor Tables Table Number 14-117 14-118 14-119 14-120 14-121 14-122 14-123 14-124 14-125 14-126 14-127 14-128 14-129 14-130 14-131 14-132 14-133 14-134 14-135 14-136 15-1 15-2 15-3 15-4 15-5 15-6 15-7 15-8 15-9 15-10 15-11 15-12 15-13 15-14 15-15 15-16 15-17 15-18 15-19 15-20 Title Page Number Interrupt Coalescing Timing Threshold Ranges ............................................................... 14-123 Transmission Errors .......................................................................................................... 14-124 Reception Errors ............................................................................................................... 14-124 Transmit Data Buffer Descriptor (TxBD) Field Descriptions .......................................... 14-127 Receive Buffer Descriptor Field Descriptions .................................................................. 14-129 MII Interface Mode Signal Configuration ........................................................................ 14-131 Shared MII Signals............................................................................................................ 14-132 MII Mode Register Initialization Steps............................................................................. 14-132 GMII Interface Mode Signal Configuration ..................................................................... 14-135 Shared GMII Signals......................................................................................................... 14-136 GMII Mode Register Initialization Steps.......................................................................... 14-136 TBI Interface Mode Signal Configuration ........................................................................ 14-138 Shared TBI Signals ........................................................................................................... 14-139 TBI Mode Register Initialization Steps ............................................................................ 14-140 RGMII Interface Mode Signal Configuration................................................................... 14-143 Shared RGMII Signals ...................................................................................................... 14-144 RGMII Mode Register Initialization Steps ....................................................................... 14-144 RTBI Interlace Mode Signal Configuration...................................................................... 14-147 Shared RTBI Signals ......................................................................................................... 14-148 RTBI Mode Register Initialization Steps .......................................................................... 14-148 Relationship of Modes and Features ..................................................................................... 15-3 DMA Mode Field Descriptions............................................................................................. 15-3 DMA Signals—Detailed Signal Descriptions....................................................................... 15-6 DMA Register Summary ...................................................................................................... 15-7 MRn Field Descriptions ...................................................................................................... 15-11 SRn Field Descriptions ....................................................................................................... 15-13 CLNDARn Field Descriptions............................................................................................ 15-16 SATRn Field Descriptions .................................................................................................. 15-17 SARn Field Descriptions .................................................................................................... 15-19 SARn Field Descriptions .................................................................................................... 15-19 DATRn Field Descriptions.................................................................................................. 15-20 DARn Field Descriptions.................................................................................................... 15-22 DARn Field Descriptions.................................................................................................... 15-22 BCRn Field Descriptions .................................................................................................... 15-23 NLNDARn Field Descriptions............................................................................................ 15-23 CLSDARn Field Descriptions ............................................................................................ 15-24 NLSDARn Field Descriptions ............................................................................................ 15-25 SSRn Field Descriptions ..................................................................................................... 15-26 DSRn Field Descriptions .................................................................................................... 15-26 DGSR Field Descriptions.................................................................................................... 15-27 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor lxxiii Tables Table Number 15-21 15-22 15-23 15-24 16-1 16-2 16-3 16-4 16-5 16-6 16-7 16-8 16-9 16-10 16-11 16-12 16-13 16-14 16-15 16-16 16-17 16-18 16-19 16-20 16-21 16-22 16-23 16-24 16-25 16-26 16-27 16-28 16-29 16-30 16-31 16-32 16-33 16-34 16-35 16-36 Title Page Number Channel State Table............................................................................................................. 15-36 DMA List Descriptor Summary.......................................................................................... 15-38 DMA List Descriptor Summary.......................................................................................... 15-38 DMA Paths.......................................................................................................................... 15-43 POR Parameters for PCI/X Controller.................................................................................. 16-5 PCI Interface Signals—Detailed Signal Descriptions .......................................................... 16-8 PCI/X Memory-Mapped Register Map............................................................................... 16-15 PCI/X CFG_ADDR Field Descriptions.............................................................................. 16-19 PCI/X CFG_DATA Field Descriptions ............................................................................... 16-19 PCI/X INT_ACK Field Descriptions.................................................................................. 16-20 POTARn Field Descriptions ............................................................................................... 16-21 POTEARn Field Descriptions............................................................................................. 16-22 POWBARn Field Descriptions ........................................................................................... 16-22 POWARn Field Descriptions .............................................................................................. 16-23 PITARn Field Descriptions ................................................................................................. 16-25 PIWBARn Field Descriptions............................................................................................. 16-26 PIWBEARn Field Descriptions .......................................................................................... 16-26 PIWARn Field Descriptions................................................................................................ 16-27 ERR_DR Field Descriptions ............................................................................................... 16-30 ERR_CAP_DR Field Descriptions ..................................................................................... 16-31 ERR_EN Field Descriptions ............................................................................................... 16-32 ERR_ATTRIB Field Descriptions ...................................................................................... 16-32 ERR_ADDR Field Descriptions ......................................................................................... 16-33 ERR_EXT_ADDR Field Descriptions ............................................................................... 16-34 ERR_DL Field Description................................................................................................. 16-34 ERR_DH Field Description ................................................................................................ 16-35 GAS_TIMR Field Descriptions .......................................................................................... 16-35 PCIX_TIMR Field Descriptions ......................................................................................... 16-36 PCI Vendor ID Register Field Description ......................................................................... 16-38 PCI Device ID Register Field Description.......................................................................... 16-38 PCI Bus Command Register Field Descriptions................................................................. 16-39 PCI Bus Status Register Field Descriptions........................................................................ 16-40 PCI Revision ID Register Field Descriptions ..................................................................... 16-41 PCI Bus Programming Interface Register Field Description.............................................. 16-42 PCI Subclass Code Register Field Description................................................................... 16-42 PCI Bus Base Class Code Register Field Description ........................................................ 16-42 PCI Bus Cache Line Size Register (PCLSR) Field Descriptions ....................................... 16-43 PCI Bus Latency Timer Register Field Descriptions .......................................................... 16-43 PCSRBAR Field Descriptions ............................................................................................ 16-44 32-Bit Memory Base Address Register Field Descriptions ................................................ 16-45 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 lxxiv Freescale Semiconductor Tables Table Number 16-37 16-38 16-39 16-40 16-41 16-42 16-43 16-44 16-45 16-46 16-47 16-48 16-49 16-50 16-51 16-52 16-53 16-54 16-55 16-56 16-57 16-58 16-59 16-60 16-61 16-62 16-63 16-64 16-65 16-66 16-67 17-1 17-2 17-3 17-4 17-5 17-6 17-7 17-8 17-9 Title Page Number 64-Bit Low Memory Base Address Register Field Descriptions........................................ 16-45 64-Bit High Memory Base Address Register Field Description......................................... 16-46 PCI Subsystem Vendor ID Register Field Description ....................................................... 16-46 PCI Subsystem ID Register Field Description.................................................................... 16-47 PCI Bus Capabilities Pointer Register Field Description ................................................... 16-47 PCI Bus Interrupt Line Register Field Description............................................................. 16-48 PCI Bus Interrupt Pin Register Field Description............................................................... 16-48 PCI Bus Minimum Grant Register Field Description ......................................................... 16-48 PCI Bus Maximum Latency Register Field Description .................................................... 16-49 PCI Bus Function Register Field Descriptions ................................................................... 16-49 PCI Bus Arbiter Configuration Register Field Descriptions .............................................. 16-50 PCI-X Next Capabilities ID Register Field Descriptions ................................................... 16-51 PCI-X Capability Pointer Register Field Description......................................................... 16-52 PCI-X Command Register Field Descriptions.................................................................... 16-52 PCI-X Status Register Field Descriptions........................................................................... 16-53 PCI Bus Commands ............................................................................................................ 16-58 Supported Combinations of PCI_AD[1:0].......................................................................... 16-59 PCI Configuration Space Header Summary ....................................................................... 16-72 PCI Type 0 Configuration—Device Number to ADn Translation...................................... 16-75 Special-Cycle Message Encodings ..................................................................................... 16-77 PCI Mode Error Actions ..................................................................................................... 16-80 PCI-X Command Encodings............................................................................................... 16-82 Burst/DWORD Transaction Attribute Summary ................................................................ 16-84 Split Completion Transaction Address ............................................................................... 16-91 PCI-X Split Completion Transaction Attribute Summary .................................................. 16-92 PCI-X Split Completion Message Summary ...................................................................... 16-93 PCI-X Type 0 Configuration—Device Number to ADn Translation ................................. 16-95 PCI-X Configuration Transaction Attribute Summary ....................................................... 16-96 PCI-X Mode Error Actions ................................................................................................. 16-97 Affected Configuration Register Bits for POR ................................................................... 16-99 Power-On Reset Values for Affected Configuration Bits ................................................. 16-100 RapidIO Terminology ........................................................................................................... 17-4 CRC Checking Mode ............................................................................................................ 17-6 Accept All Mode Behavior ................................................................................................... 17-6 RapidIO Interface Signals Summary .................................................................................... 17-7 RapidIO Interface Signals—Detailed Signal Descriptions ................................................... 17-8 RapidIO Module Memory Map ............................................................................................ 17-9 DIDCAR Field Descriptions............................................................................................... 17-13 DICAR Field Description ................................................................................................... 17-14 AIDCAR Field Descriptions ............................................................................................... 17-14 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor lxxv Tables Table Number 17-10 17-11 17-12 17-13 17-14 17-15 17-16 17-17 17-18 17-19 17-20 17-21 17-22 17-23 17-24 17-25 17-26 17-27 17-28 17-29 17-30 17-31 17-32 17-33 17-34 17-35 17-36 17-37 17-38 17-39 17-40 17-41 17-42 17-43 17-44 17-45 17-46 17-47 17-48 17-49 Title Page Number AICAR Field Descriptions.................................................................................................. 17-15 PEFCAR Field Descriptions ............................................................................................... 17-15 SPICAR Field Descriptions ................................................................................................ 17-17 SOCAR Field Descriptions ................................................................................................. 17-18 DOCAR Field Descriptions ................................................................................................ 17-20 MSR Field Definitions ........................................................................................................ 17-22 PWDCSR Field Descriptions.............................................................................................. 17-23 PELLCCSR Field Descriptions .......................................................................................... 17-24 LCSBA1CSR Field Descriptions........................................................................................ 17-25 BDIDCSR Field Descriptions ............................................................................................. 17-25 HBDIDLCSR Field Descriptions........................................................................................ 17-26 CTCSR Field Descriptions.................................................................................................. 17-27 PMBH0CSR Field Descriptions ......................................................................................... 17-27 PLTOCCSR Field Descriptions .......................................................................................... 17-28 PRTOCCSR Field Descriptions .......................................................................................... 17-28 PGCCSR Field Descriptions ............................................................................................... 17-29 PLMREQCSR Field Descriptions ...................................................................................... 17-30 PLMRESPCSR Field Descriptions ..................................................................................... 17-30 PLASCSR Field Descriptions ............................................................................................. 17-31 PESCSR Field Descriptions................................................................................................ 17-32 PCCSR Field Descriptions.................................................................................................. 17-33 CR Field Descriptions ......................................................................................................... 17-35 PCR Field Descriptions....................................................................................................... 17-35 PEIR Field Descriptions...................................................................................................... 17-36 ROWTARn Standard Field Descriptions ............................................................................ 17-37 ROWTARn Maintenance Field Descriptions...................................................................... 17-38 ROWBARn Field Descriptions........................................................................................... 17-39 ROWARn Field Descriptions.............................................................................................. 17-39 RIWTARn Field Descriptions............................................................................................. 17-41 RIWBARn Field Descriptions ............................................................................................ 17-42 RIWARn Field Descriptions ............................................................................................... 17-42 PNFEDR Field Descriptions ............................................................................................... 17-44 PNFEDiR Field Descriptions.............................................................................................. 17-47 PNFEIER Field Descriptions .............................................................................................. 17-50 PECSR Field Descriptions .................................................................................................. 17-52 EPCR0 Field Descriptions .................................................................................................. 17-53 EPCR1 Type 1 Field Descriptions ...................................................................................... 17-54 EPCR1 Type 2 Field Descriptions ...................................................................................... 17-54 EPCR1 Type 5 Field Descriptions ...................................................................................... 17-55 EPCR1 Type 6 Field Descriptions ...................................................................................... 17-55 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 lxxvi Freescale Semiconductor Tables Table Number 17-50 17-51 17-52 17-53 17-54 17-55 17-56 17-57 17-58 17-59 17-60 17-61 17-62 17-63 17-64 17-65 17-66 17-67 17-68 17-69 17-70 17-71 17-72 17-73 17-74 17-75 17-76 17-77 17-78 17-79 17-80 17-81 17-82 17-83 17-84 17-85 17-86 17-87 17-88 17-89 Title Page Number EPCR1 Type 8 Request Field Descriptions ........................................................................ 17-56 EPCR1 Type 8 Response Field Descriptions ...................................................................... 17-56 EPCR1 Type 10 Field Descriptions .................................................................................... 17-57 EPCR1 Type 11 Field Descriptions .................................................................................... 17-57 EPCR1 Type 13 Field Descriptions .................................................................................... 17-58 EPCR2 Type 1, 2, or 5 Field Descriptions .......................................................................... 17-59 EPCR2 Type 6 Field Descriptions ...................................................................................... 17-59 EPCR2 Type 8 Request Field Descriptions ........................................................................ 17-60 EPCR2 Type 8 Response Field Descriptions ...................................................................... 17-60 EPCR2 Type 10 Field Descriptions .................................................................................... 17-61 EPCR2 Type 11 or 13 Field Description............................................................................. 17-61 PREDR Field Descriptions ................................................................................................. 17-62 PERTR Field Descriptions .................................................................................................. 17-64 PRTR Field Descriptions .................................................................................................... 17-65 OMR Field Descriptions ..................................................................................................... 17-66 OSR Field Descriptions ...................................................................................................... 17-67 ODQDPAR Field Descriptions ........................................................................................... 17-69 OSAR Field Descriptions.................................................................................................... 17-69 ODPR Field Descriptions.................................................................................................... 17-70 ODATR Field Descriptions ................................................................................................. 17-70 ODCR Field Descriptions ................................................................................................... 17-71 ODQEPAR Field Descriptions............................................................................................ 17-72 IMR Field Descriptions....................................................................................................... 17-73 ISR Field Descriptions ........................................................................................................ 17-74 IFQDPAR Field Descriptions ............................................................................................. 17-76 IFQEPAR Field Descriptions .............................................................................................. 17-76 DMR Field Descriptions ..................................................................................................... 17-77 DSR Field Descriptions ...................................................................................................... 17-78 DQDPAR Field Descriptions .............................................................................................. 17-80 DQEPAR Field Descriptions............................................................................................... 17-81 PWMR Field Descriptions .................................................................................................. 17-81 PWSR Field Descriptions ................................................................................................... 17-82 PWQBAR Field Descriptions ............................................................................................. 17-83 RapidIO Transaction Summary........................................................................................... 17-84 RapidIO Packet Format Summary (TT = 0b00) ................................................................. 17-86 Control Symbol Formats ..................................................................................................... 17-88 General Device Functionality List ...................................................................................... 17-90 General Device 8/16 LP-LVDS Physical Layer Basic Functionality List .......................... 17-91 General Device 8/16 LP-LVDS Physical Layer Link Maintenance List ............................ 17-93 General Device 8/16 LP-LVDS Physical Layer Packet Transmission List ........................ 17-93 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor lxxvii Tables Table Number 17-90 17-91 17-92 17-93 17-94 17-95 17-96 17-97 17-98 17-99 17-100 17-101 17-102 17-103 17-104 17-105 18-1 18-2 18-3 18-4 18-5 18-6 18-7 18-8 18-9 18-10 18-11 18-12 18-13 18-14 18-15 18-16 18-17 18-18 18-19 18-20 18-21 18-22 18-23 19-1 Title Page Number General Device 8/16 LP-LVDS Physical Layer Packet Reception List.............................. 17-94 General Device 8/16 LP-LVDS Physical Layer Recoverable Errors List .......................... 17-96 General Device 8/16 LP-LVDS Physical Layer Nonrecoverable Errors List..................... 17-98 General Device Common Transport Layer Basic Functionality List.................................. 17-99 General Device Common Transport Layer Packet Transmission List................................ 17-99 General Device Common Transport Layer Packet Reception List ..................................... 17-99 General Device Common Transport Layer Detectable Errors List..................................... 17-99 General Device Logical Layer Basic Functionality List................................................... 17-100 General Device Logical Layer Target Transaction Support List ...................................... 17-101 General Device Logical Layer Detectable Errors List...................................................... 17-101 Recoverable Errors Detected by RapidIO Controller ....................................................... 17-102 Notification Errors Detected by RapidIO Controller ........................................................ 17-104 Fatal Errors Detected by RapidIO Controller ................................................................... 17-105 Outbound Message Unit Descriptor Summary ................................................................. 17-116 Target Information Definition ........................................................................................... 17-120 Source Information Definition .......................................................................................... 17-120 External Signal Summary ..................................................................................................... 18-2 Detailed Signal Descriptions................................................................................................. 18-2 Global Utilities Block Register Summary ............................................................................ 18-3 PORPLLSR Field Descriptions ............................................................................................ 18-5 PORBMSR Field Descriptions ............................................................................................. 18-6 PORIMPSCR Field Descriptions.......................................................................................... 18-7 PORDEVSR Field Descriptions ........................................................................................... 18-8 PORDBGMSR Field Descriptions........................................................................................ 18-9 GPPORCR Field Descriptions ............................................................................................ 18-10 GPIOCR Field Descriptions................................................................................................ 18-10 GPOUTDR Field Descriptions ........................................................................................... 18-11 GPINDR Field Descriptions ............................................................................................... 18-12 PMUXCR Field Descriptions ............................................................................................. 18-13 DEVDISR Field Descriptions ............................................................................................. 18-14 POWMGTCSR Field Descriptions ..................................................................................... 18-15 MCPSUMR Field Descriptions .......................................................................................... 18-17 PVR Field Descriptions ...................................................................................................... 18-18 SVR Field Descriptions ...................................................................................................... 18-18 CLKOCR Field Descriptions .............................................................................................. 18-19 DDRDLLCR Field Descriptions......................................................................................... 18-19 LBDLLCR Field Descriptions ............................................................................................ 18-20 MPC8540 Power Management Modes—Basic Description............................................... 18-22 Power Management Entry Protocol and Initiating Functional Units .................................. 18-25 Control Register Memory Map ............................................................................................. 19-4 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 lxxviii Freescale Semiconductor Tables Table Number 19-2 19-3 19-4 19-5 19-6 19-7 19-8 19-9 19-10 19-11 19-12 20-1 20-2 20-3 20-4 20-5 20-6 20-7 20-8 20-9 20-10 20-11 20-12 20-13 20-14 20-15 20-16 20-17 20-18 20-19 20-20 20-21 20-22 20-23 20-24 20-25 20-26 20-27 20-28 20-29 Title Page Number PMGC0 Field Descriptions ................................................................................................... 19-5 PMLCA0 Field Descriptions ................................................................................................ 19-6 PMLCA1–PMLC8 Field Descriptions.................................................................................. 19-6 PMLCB0 Field Descriptions................................................................................................. 19-7 PMLCBn Field Descriptions................................................................................................. 19-8 PMC0 Field Descriptions.................................................................................................... 19-10 PMCn Field Descriptions.................................................................................................... 19-10 Burst Definition................................................................................................................... 19-13 Performance Monitor Events .............................................................................................. 19-15 PMGC0 and PMLCAn Settings.......................................................................................... 19-27 Register Settings for Counting Examples MPC8540.......................................................... 19-28 POR Configuration Settings and Debug Modes ................................................................... 20-4 Debug, Watchpoint and Test Signal Summary...................................................................... 20-6 Debug Signals—Detailed Signal Descriptions ..................................................................... 20-7 Watchpoint and Trigger Signals—Detailed Signal Descriptions .......................................... 20-8 JTAG Test and Other Signals—Detailed Signal Descriptions .............................................. 20-9 Debug and Watchpoint Monitor Memory Map................................................................... 20-10 WMCR0 Field Descriptions................................................................................................ 20-12 WMCR1 Field Descriptions................................................................................................ 20-13 WMAR Field Descriptions ................................................................................................. 20-14 WMAMR Field Descriptions.............................................................................................. 20-14 WMTMR Field Descriptions .............................................................................................. 20-15 Transaction Types by Interface ........................................................................................... 20-15 WMSR Field Descriptions .................................................................................................. 20-16 TBCR0 Field Descriptions.................................................................................................. 20-17 TBCR1 Field Descriptions.................................................................................................. 20-19 TBAR Field Descriptions.................................................................................................... 20-19 TBAMR Field Descriptions ................................................................................................ 20-20 TBTMR Field Descriptions ................................................................................................ 20-21 TBSR Field Descriptions .................................................................................................... 20-21 TBACR Field Descriptions ................................................................................................. 20-22 TBADHR Field Descriptions.............................................................................................. 20-23 TBADR Field Descriptions................................................................................................. 20-23 PCIDR Field Descriptions .................................................................................................. 20-24 CCIDR Field Descriptions .................................................................................................. 20-25 TOSR Field Descriptions .................................................................................................... 20-26 Source and Target ID Values............................................................................................... 20-26 CMD Trace Buffer Entry Field Descriptions (TBCR1[IFSEL] = 000) .............................. 20-30 DDR Trace Buffer Entry Field Descriptions (TBCR1[IFSEL] = 001) ............................... 20-31 PCI Trace Buffer Entry Field Descriptions (TBCR1[IFSEL] = 010) ................................. 20-32 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor lxxix Tables Table Number 20-30 21-1 21-2 21-3 21-4 21-5 21-6 21-7 21-8 21-9 21-10 21-11 21-12 21-13 21-14 21-15 21-16 21-17 21-18 21-19 21-20 21-21 21-22 21-23 21-24 21-25 21-26 21-27 21-28 21-29 21-30 21-31 21-32 21-33 21-34 21-35 21-36 21-37 21-38 21-39 Title Page Number RapidIO Trace Buffer Entry Field Descriptions (TBCR1[IFSEL] = 011).......................... 20-32 FEC Signals—Detailed Signal Descriptions......................................................................... 21-8 Module Memory Map Summary......................................................................................... 21-10 Module Memory Map ......................................................................................................... 21-10 IEVENT Field Descriptions................................................................................................ 21-14 IMASK Field Descriptions ................................................................................................. 21-17 EDIS Field Descriptions ..................................................................................................... 21-18 MINFLR Field Descriptions ............................................................................................... 21-19 PTV Field Descriptions ....................................................................................................... 21-20 DMACTRL Field Descriptions........................................................................................... 21-20 FIFO_PAUSE_CTRL Field Descriptions ........................................................................... 21-23 FIFO_TX_THR Field Descriptions .................................................................................... 21-24 FIFO_TX_STARVE Field Descriptions ............................................................................. 21-24 FIFO_TX_STARVE_SHUTOFF Field Descriptions.......................................................... 21-25 TCTRL Field Descriptions.................................................................................................. 21-26 TSTAT Field Descriptions................................................................................................... 21-27 TBDLEN Field Descriptions .............................................................................................. 21-27 CTBPTR Field Descriptions ............................................................................................... 21-28 TBPTR Field Descriptions .................................................................................................. 21-28 TBASE Field Descriptions.................................................................................................. 21-29 OSTBD Field Descriptions ................................................................................................. 21-29 OSTBDP Field Descriptions ............................................................................................... 21-31 RCTRL Field Descriptions ................................................................................................. 21-32 RSTAT Field Descriptions .................................................................................................. 21-33 RBDLEN Field Descriptions .............................................................................................. 21-33 CRBPTR Field Descriptions ............................................................................................... 21-34 MRBLR Field Descriptions ................................................................................................ 21-34 RBPTR Field Descriptions.................................................................................................. 21-35 RBASE Field Descriptions ................................................................................................. 21-35 MACCFG1 Field Descriptions ........................................................................................... 21-38 MACCFG2 Field Descriptions ........................................................................................... 21-40 IPGIFG Field Descriptions ................................................................................................. 21-41 HAFDUP Field Descriptions .............................................................................................. 21-42 MAXFRM Descriptions...................................................................................................... 21-43 IFSTAT Field Descriptions ................................................................................................. 21-43 MACSTNADDR1 Field Descriptions ................................................................................ 21-44 MACSTNADDR2 Field Descriptions ................................................................................ 21-45 IADDRn Field Description ................................................................................................. 21-45 GADDRn Field Description................................................................................................ 21-46 ATTR Field Descriptions .................................................................................................... 21-47 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 lxxx Freescale Semiconductor Tables Table Number 21-40 21-41 21-42 21-43 21-44 21-45 21-46 21-47 21-48 21-49 21-50 21-51 Title Page Number ATTRELI Field Descriptions .............................................................................................. 21-48 MII Interface Signals .......................................................................................................... 21-49 Steps of Minimum Register Initialization ........................................................................... 21-50 Flow Control Frame Structure ............................................................................................ 21-59 Non-Error Transmit Interrupts ............................................................................................ 21-60 Non-Error Receive Interrupts.............................................................................................. 21-60 Transmission Errors ............................................................................................................ 21-61 Reception Errors ................................................................................................................. 21-63 Transmit Data Buffer Descriptor (TxBD) Field Descriptions ............................................ 21-65 Receive Buffer Descriptor Field Descriptions .................................................................... 21-67 MII Interface Mode Signal Configuration .......................................................................... 21-70 MII Mode Register Initialization Steps............................................................................... 21-71 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor lxxxi Tables Table Number Title Page Number MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 lxxxii Freescale Semiconductor About This Book The primary objective of this reference manual is to define the functionality of the MPC8540. The MPC8540 integrates a PowerPC™ processor with system logic required for networking, storage, and general purpose embedded applications. The e500 processor core is a low-power implementation of the family of reduced instruction set computing (RISC) embedded processors that implement the Book E definition of the PowerPC architecture. This book is intended as a companion to the PowerPC e500 Core Complex Reference Manual. Audience It is assumed that the reader understands operating systems, microprocessor system design, and the basic principles of RISC processing. Organization Following is a summary and a brief description of the major parts of this reference manual: Part I, “Overview,” describes the many features of the MPC8540 integrated host processor at an overview level. The following chapters are included: • Chapter 1, “Overview,” provides a high-level description of features and functionality of the MPC8540 integrated host processor. It describes the MPC8540, its interfaces, and its programming model. The functional operation of the MPC8540 with emphasis on peripheral functions is also described. Chapter 2, “Memory Map,” describes the memory map of the MPC8540. An overview of the local address map is followed by a description of how local access windows are used to define the local address map. The inbound and outbound address translation mechanisms used to map to and from external memory spaces are described next. Finally, the configuration, control, and status registers are described, including a complete listing of all memory-mapped registers with cross references to the sections detailing descriptions of each. Chapter 3, “Signal Descriptions,” provides a listing of all the external signals, cross-references for signals that serve multiple functions, output signal states at reset, and reset configuration signals (and the modes they define). Chapter 4, “Reset, Clocking, and Initialization,” describes the hard and soft resets, the power-on reset sequence, power-on reset (POR) configuration, clocking, and initialization of the MPC8540. • • • MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor lxxxiii Part II, “e500 Core Complex and L2 Cache,” describes the many features of the MPC8540 core processor at an overview level and the interaction between the core complex and the L2 cache. The following chapters are included: • • • Chapter 5, “Core Complex Overview,” provides an overview of the e500 core processor and the L1 caches and MMU that, together with the core, comprise the core complex. Chapter 6, “Core Register Summary,” provides a listing of the e500 registers in reference form. Chapter 7, “L2 Look-Aside Cache/SRAM,” describes the L2 cache of the MPC8540. Note that the L2 cache can also be addressed directly as memory-mapped SRAM. Part III, “Memory and I/O Interfaces,” defines the memory and I/O interfaces of the MPC8540 and how these blocks interact with one another and with other blocks on the device. The following chapters are included: • Chapter 8, “e500 Coherency Module,” defines the e500 coherency module and how it facilitates communication between the e500 core complex, the L2 cache, and the other blocks that comprise the coherent memory domain of the MPC8540. The ECM provides a mechanism for I/O-initiated transactions to snoop the core complex bus (CCB) of the e500 core in order to maintain coherency across cacheable local memory. It also provides a flexible, easily expandable switch-type structure for e500- and I/O-initiated transactions to be routed (dispatched) to target modules on the MPC8540. Chapter 9, “DDR Memory Controller,” describes the DDR SDRAM memory controller of the MPC8540. This fully programmable controller supports most DDR memories available today, including both buffered and unbuffered devices. The built-in error checking and correction (ECC) ensures very low bit-error rates for reliable high-frequency operation. Dynamic power management and auto-precharge modes simplify memory system design. A large set of special features like DLL software override, crawl mode, and ECC error injection support rapid system debug. Chapter 10, “Programmable Interrupt Controller,” describes the embedded programmable interrupt controller (PIC) of the MPC8540. This controller is an OpenPIC-compliant interrupt controller that provides interrupt management, and is responsible for receiving hardware-generated interrupts from different sources (both internal and external), prioritizing them and delivering them to the CPU for servicing. Chapter 11, “I2C Interface,” describes the inter-IC (IIC or I2C) bus controller of the MPC8540. This synchronous, serial, bidirectional, multi-master bus allows two-wire connection of devices, such as microcontrollers, EEPROMs, real-time clock devices, A/D converters and LCDs. The MPC8540 powers up in boot sequencer mode which allows the I2C controller to initialize configuration registers. Chapter 12, “DUART,” describes the (dual) universal asynchronous receiver/transmitters (UARTs) which feature a PC16552D-compatible programming model. These independent UARTs are provided specifically to support system debugging. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 lxxxiv Freescale Semiconductor • • • • • • • • • Chapter 13, “Local Bus Controller,” describes the local bus controller of the MPC8540. The main component of the local bus controller (LBC) is its memory controller which provides a seamless interface to many types of memory devices and peripherals. The memory controller is responsible for controlling eight memory banks shared by a high performance SDRAM machine, a general-purpose chip-select machine (GPCM), and up to three user-programmable machines (UPMs). As such, it supports a minimal glue logic interface to synchronous DRAM (SDRAM), SRAM, EPROM, flash EPROM, burstable RAM, regular DRAM devices, extended data output DRAM devices, and other peripherals. Chapter 14, “Three-Speed Ethernet Controllers,” describes the 2 three-speed Ethernet controllers on the MPC8540. These controllers provide 10/100/1000 Mbps Ethernet support with a complete set of media-independent interface options including GMII, RGMII, TBI, and RTBI. Each controller provides very high throughput using a captive DMA channel and direct connection to the MPC8540 memory coherency module. Chapter 15, “DMA Controller,” describes the four-channel general-purpose DMA controller of the MPC8540. The DMA controller transfers blocks of data, independent of the e500 core or external hosts. Data movement occurs among RapidIO and the local address space. The DMA controller has four high-speed channels. Both the e500 core and external masters can initiate a DMA transfer. All channels are capable of complex data movement and advanced transaction chaining. Chapter 16, “PCI/PCI-X Bus Interface,” describes the PCI/PCI-X controller of the MPC8540. Chapter 17, “RapidIO Interface,” describes the RapidIO controller of the MPC8540. Part IV, “Global Functions and Debug,” defines other global blocks of the MPC8540. The following chapters are included: • Chapter 18, “Global Utilities,” defines the global utilities of the MPC8540. These include power management, I/O device enabling, power-on-reset (POR) configuration monitoring, general-purpose I/O signal use, and multiplexing for the interrupt and local bus chip select signals Chapter 19, “Performance Monitor,” describes the performance monitor of the MPC8540. Note that the MPC8540 performance monitor is similar to but separate from the performance monitor implemented on the e500 core. Chapter 20, “Debug Features and Watchpoint Facility,” describes the debug features and watchpoint monitor of the MPC8540. Chapter 21, “10/100 Fast Ethernet Controller,” includes information about using the dual-speed Ethernet controller of the MPC8540 for debug purposes. Appendix A, “Revision History,” lists the major differences between revisions of the MPC8540 PowerQUICC III Integrated Host Processor Reference Manual. • • • • MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor lxxxv • • This document also contains the following indexes: — Register index. Contains listings of all memory-mapped registers implemented by the integrated logic. It does not include the following: – e500 core registers. These registers are listed in the general index under e500 core, registers. – Configuration header registers defined by the PCI specification. These registers are listed under PCI/PCI-X controller, register descriptions. — General index. Contains all entries except those for memory-mapped registers. This reference manual also includes a glossary. Suggested Reading This section lists additional reading that provides background for the information in this manual as well as general information about the architecture. General Information The following documentation, published by Morgan-Kaufmann Publishers, 340 Pine Street, Sixth Floor, San Francisco, CA, provides useful information about the PowerPC architecture and computer architecture in general: • • • The PowerPC Architecture: A Specification for a New Family of RISC Processors, Second Edition, by International Business Machines, Inc. Computer Architecture: A Quantitative Approach, Third Edition, by John L. Hennessy and David A. Patterson Computer Organization and Design: The Hardware/Software Interface, Second Edition, by David A. Patterson and John L. Hennessy Related Documentation Freescale Semiconductor documentation is available from the sources listed on the back cover of this manual; the document order numbers are included in parentheses for ease in ordering: • EREF: A Reference for Freescale Semiconductor Book E and the e500 Core—This book provides a higher-level view of the programming model as it is defined by Book E, the Freescale Semiconductor Book E implementation standards, and the e500 microprocessor. Reference manuals (formerly called user’s manuals)—These books provide details about individual implementations. Addenda/errata to reference or user’s manuals—Because some processors have follow-on parts, an addendum is provided that describes the additional features and functionality MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 lxxxvi Freescale Semiconductor • • • • • changes. These addenda are intended for use with the corresponding reference or user’s manuals. Hardware specifications—Hardware specifications provide specific data regarding bus timing, signal behavior, and AC, DC, and thermal characteristics, as well as other design considerations. Technical summaries—Each device has a technical summary that provides an overview of its features. This document is roughly equivalent to the overview (Chapter 1) of an implementation’s reference or user’s manual. Application notes—These short documents address specific design issues useful to programmers and engineers working with Freescale Semiconductor processors. Additional literature is published as new processors become available. For a current list of documentation, refer to http://www.freescale.com. Conventions This document uses the following notational conventions: cleared/set mnemonics italics When a bit takes the value zero, it is said to be cleared; when it takes a value of one, it is said to be set. Instruction mnemonics are shown in lowercase bold. Italics indicate variable command parameters, for example, bcctrx. Book titles in text are set in italics Internal signals are set in lowercase italics, for example, core int 0x0 0b0 rA, rB rD REG[FIELD] Prefix to denote hexadecimal number Prefix to denote binary number Instruction syntax used to identify a source GPR Instruction syntax used to identify a destination GPR Abbreviations for registers are shown in uppercase text. Specific bits, fields, or ranges appear in brackets. For example, MSR[LE] refers to the little-endian mode enable bit in the machine state register. In some contexts, such as signal encodings, an unitalicized x indicates a don’t care. An italicized x indicates an alphanumeric variable. An italicized n indicates a numeric variable. NOT logical operator AND logical operator MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor lxxxvii x x n ¬ & | || — R W R FIELDNAME W R W FIELDNAME 0 OR logical operator Concatenation, for example TCR[WPEXT]||TCR[WP] Indicates a reserved bit field in an e500 register. Although these bits can be written to as ones or zeros, they are always read as zeros. Indicates a reserved bit field in a memory-mapped register. Although these bits can be written to as ones or zeros, they are always read as zeros. Indicates a read-only bit field in a memory-mapped register. Indicates a write-only bit field in a memory-mapped register. Although these bits can be written to as ones or zeros, they are always read as zeros. Signal Conventions OVERBAR lowercase_italics An overbar indicates that a signal is active-low. Lowercase italics is used to indicate internal signals. lowercase_plaintext Lowercase plain text is used to indicate signals that are used for configuration. For more information, see Section 3.2, “Configuration Signals Sampled at Reset.” Acronyms and Abbreviations Table i contains acronyms and abbreviations used in this document. Table 1. Acronyms and Abbreviated Terms Term ADB ATM ATMU BD BIST BRI BTB BUID CAM CCB Allowable disconnect boundary Asynchronous transfer mode Address translation and mapping unit Buffer descriptor Built-in self test Basic rate interface Branch target buffer Bus unit ID Content-addressable memory Core complex bus Meaning MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 lxxxviii Freescale Semiconductor Table 1. Acronyms and Abbreviated Terms (continued) Term CCSR CEPT COL CRC CRS DDR DMA DPLL DRAM DUART EA ECC ECM EEST EHPI EPROM FCS GCI GMII GPCM GPIO GPR GUI HDLC I2 C Configuration control and status register Conference des administrations Europeanes des Postes et Telecommunications (European Conference of Postal and Telecommunications Administrations). Collision Cyclic redundancy check Carrier sense Double data rate Direct memory access Digital phase-locked loop Dynamic random access memory Dual universal asynchronous receiver/transmitter Effective address Error checking and correction e500 coherency module Enhanced Ethernet serial transceiver Enhanced host port interface Erasable programmable read-only memory Frame-check sequence General circuit interface Gigabit media independent interface General-purpose chip-select machine General-purpose I/O General-purpose register Graphical user interface High-level data link control Inter-integrated circuit Inter-chip digital link Institute of Electrical and Electronics Engineers Interpacket gap Infrared Data Association Integrated services digital network Instruction translation lookaside buffer Integer unit Meaning IDL IEEE IPG IrDA ISDN ITLB IU MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor lxxxix Table 1. Acronyms and Abbreviated Terms (continued) Term JTAG LAE LAW LBC LIFO LRU LSB lsb LSU MAC MDI MESI MII MMU MSB msb NMSI No-op OCeaN OSI PCI PCI/X PCI-X PCMCIA PCS PIC PMA PMD POR PRI RGMII RISC Joint Test Action Group Local access error Local access window Local bus controller Last-in-first-out Least recently used Least-significant byte Least-significant bit Load/store unit Multiply accumulate, media access control Medium-dependent interface Modified/exclusive/shared/invalid—cache coherency protocol Media independent interface Memory management unit Most-significant byte Most-significant bit Nonmultiplexed serial interface No operation On-chip network Open systems interconnection Peripheral component interconnect Abbreviation used to describe operation for both the PCI and PCI-X bus functionality PCI extended Personal Computer Memory Card International Association Physical coding sublayer Programmable interrupt controller Physical medium attachment Physical medium dependent Power-on reset Primary rate interface Reduced gigabit media independent interface Reduced instruction set computing Meaning MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 xc Freescale Semiconductor Table 1. Acronyms and Abbreviated Terms (continued) Term RIO RTOS RWITM RWM Rx RxBD SCP SDLC SDMA SFD SI SIU SMC SNA SPI SPR SRAM TAP TBI TDM TLB TSA TSEC Tx TxBD UART UPM USB UTP VA ZBT Meaning Abbreviation occasionally used to refer to the RapidIO interface Real-time operating system Read with intent to modify Read modify write Receive Receive buffer descriptor Serial control port Synchronous data link control Serial DMA Start frame delimiter Serial interface System interface unit Serial management controller Systems network architecture Serial peripheral interface Special-purpose register Static random access memory Test access port Ten-bit interface Time-division multiplexed Translation lookaside buffer Time-slot assigner Three-speed Ethernet controller Transmit Transmit buffer descriptor Universal asynchronous receiver/transmitter User-programmable machine Universal serial bus Unshielded twisted pair Virtual address Zero bus turnaround MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor xci MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 xcii Freescale Semiconductor Part I Overview Part I describes the many features of the MPC8540 integrated host processor at an overview level. The following chapters are included: • • • • Chapter 1, “Overview” Chapter 2, “Memory Map” Chapter 3, “Signal Descriptions” Chapter 4, “Reset, Clocking, and Initialization” Chapter 1, “Overview,” provides a high-level description of features and functionality of the MPC8540 integrated host processor. It describes the MPC8540, its interfaces, and its programming model. The functional operation of the MPC8540 with emphasis on peripheral functions is also described. Chapter 2, “Memory Map,” describes the MPC8540 memory map. An overview of the local address map is followed by a description of how local access windows are used to define the local address map. The inbound and outbound address translation mechanisms used to map to and from external memory spaces are described next. Finally, the configuration, control, and status registers are described, including a complete listing of all memory mapped registers with cross references to the sections detailing descriptions of each. Chapter 3, “Signal Descriptions,” provides a listing of all the external signals, cross-references for signals that serve multiple functions, output signal states at reset, and reset configuration signals (and the modes they define). Chapter 4, “Reset, Clocking, and Initialization,” describes the hard and soft resets, the power-on reset sequence, power-on reset (POR) configuration, clocking, and initialization of the MPC8540. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor I-1 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 I-2 Freescale Semiconductor Chapter 1 Overview The MPC8540 integrates a PowerPC™ processor core with system logic required for networking, storage, and general-purpose embedded applications. The MPC8540 is a member of a growing family of products that combine system-level support for industry standard interfaces to processors that implement the PowerPC architecture. This chapter provides a high-level description of the features and functionality of the MPC8540 integrated microprocessor. 1.1 Introduction The MPC8540 uses the e500 core and RapidIO interconnect technology to balance processor performance with I/O system throughput. The e500 core implements the enhanced Book E instruction set architecture and provides unprecedented levels of hardware and software debugging support. In addition, the MPC8540 offers 256 Kbytes of L2 cache, 2 integrated 10/100/1Gb three-speed Ethernet controllers (TSECs), a 10/100 maintenance port, a DDR SDRAM memory controller, a 64-bit PCI/PCI-X controller, an 8-bit RapidIO port, a programmable interrupt controller (PIC), an I2C controller, a 4-channel DMA controller, a general-purpose I/O port, and a dual universal asynchronous receiver/transmitter (DUART). The high level of integration in the MPC8540 simplifies board design and offers significant bandwidth and performance. 1.2 MPC8540 Overview The following section provides a high-level overview of the features of the MPC8540. Figure 1-1 shows the major functional units in the MPC8540. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 1-1 Overview DDR SDRAM ROM, SDRAM, GPIO IRQs DDR Memory Controller 256-Kbyte L2 Cache/ SRAM e500 Coherency Module e500 Core 32-Kbyte L1 Instruction Cache 32-Kbyte L1 Data Cache Local Bus Controller Programmable Interrupt Controller Core Complex Bus MII 10/100 Fast Ethernet Controller OCeaN Switch Fabric RapidIO Interface PCI/PCI-X Bus Interface 4-Channel DMA Controller TSEC RapidIO-8 16 Gb/s PCI-X 64b 133 MHz Serial DUART 10/100/1Gb MII, GMII,TBI, RTBI, RGMII I2C I2C Interface TSEC 10/100/1Gb MII, GMII,TBI, RTBI, RGMII Figure 1-1. MPC8540 Block Diagram 1.2.1 • Key Features High-performance, 32-bit Book E–enhanced core that implements the PowerPC architecture — 32-Kbyte L1 instruction cache and 32-Kbyte L1 data cache with parity protection. Caches can be locked entirely or on a per-line basis, with separate locking for instructions and data. — Signal-processing engine (SPE) auxiliary processing unit (APU) provides an extensive instruction set for vector (64-bit) integer, single-precision floating-point, and fractional operations. These instructions use both the upper and lower words of the 64-bit GPRs as they are defined by the SPE APU. — The single-precision floating-point (SPFP) APU provides an instruction set for single-precision (32-bit) floating-point instructions. — Memory management unit (MMU) especially designed for embedded applications — Enhanced hardware and software debug support — Performance monitor facility (similar to but different from the MPC8540 performance monitor described in Chapter 19, “Performance Monitor.”) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 The following is an overview of the MPC8540 feature set: 1-2 Freescale Semiconductor M PC8540 Overview • • The e500 defines features that are not implemented on the MPC8540. It also generally defines some features that the MPC8540 implements more specifically. An understanding of these differences can be critical to ensure proper operations. These differences are summarized in Section 5.14, “MPC8540 Implementation Details,” in the MPC8540 Integrated Processor Reference Manual. Section 1.3.1, “e500 Core Overview,” includes a comprehensive list of e500 core features. 256-Kbyte L2 cache/SRAM — Can be configured as follows: – Full cache mode (256-Kbyte cache) – Full memory-mapped SRAM mode (256-Kbyte SRAM mapped as a single 256-Kbyte block or two 128-Kbyte blocks) – Half SRAM and half cache mode (128-Kbyte cache and 128-Kbyte memory-mapped SRAM) — Full error checking and correction (ECC) support on 64-bit boundary in both cache and SRAM modes — Cache mode supports instruction caching, data caching, or both — External masters can force data to be allocated into the cache through programmed memory ranges or special transaction types (stashing). — Eight-way set-associative cache organization (1024 sets of 32-byte cache lines) — Supports locking the entire cache or selected lines. Individual line locks are set and cleared through Book E instructions or by externally mastered transactions. — Global locking and flash clearing done through writes to L2 configuration registers — Instruction and data locks can be flash cleared separately — Read and write buffering for internal bus accesses — SRAM features include the following: – I/O devices access SRAM regions by marking transactions as snoopable (global). – Regions can reside at any aligned location in the memory map. – Byte-accessible ECC is protected using read-modify-write transactions accesses for smaller than cache-line accesses Address translation and mapping unit (ATMU) — Eight local access windows define mapping within local 32-bit address space — Inbound and outbound ATMUs map to larger external address spaces – Three inbound windows plus a configuration window on PCI/PCI-X – Four inbound windows plus a default and configuration window on RapidIO – Four outbound windows plus default translation for PCI – Eight outbound windows plus default translation for RapidIO MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 1-3 Overview • • • DDR memory controller — Programmable timing supporting DDR-1 SDRAM — 64-bit data interface, up to 333-MHz data rate — Four banks of memory supported, each up to 1 Gbyte — DRAM chip configurations from 64 Mbits to 1 Gbit with x8/x16 data ports — Full ECC support — Page mode support (up to 16 simultaneous open pages) — Contiguous or discontiguous memory mapping — Read-modify-write support for RapidIO atomic increment, decrement, set, and clear transactions — Sleep mode support for self-refresh SDRAM — Supports auto refreshing — On-the-fly power management using CKE signal — Registered DIMM support — Fast memory access through JTAG port — 2.5-V SSTL2 compatible I/O RapidIO interface unit — 8-bit RapidIO I/O and messaging protocols — Source-synchronous double data rate (DDR) interfaces — Supports small type systems (small domain, 8-bit device ID) — Supports four priority levels (ordering within a level) — Reordering across priority levels — Maximum data payload of 256 bytes per packet — Packet pacing support at the physical layer — CRC protection for packets — Supports atomic operations increment, decrement, set, and clear — LVDS signaling RapidIO compliant message unit — — — — — — One inbound data message structure (inbox) One outbound data message structure (outbox) Supports chaining and direct modes in the outbox Support of up to 16 packets per message Support of up to 256 bytes per packet and up to 4 Kbytes of data per message Supports one inbound doorbell message structure MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 1-4 Freescale Semiconductor M PC8540 Overview • • • • • • Programmable interrupt controller (PIC) — Programming model is compliant with the OpenPIC architecture. — Supports 16 programmable interrupt and processor task priority levels — Supports 12 discrete external interrupts — Supports 4 message interrupts with 32-bit messages — Supports connection of an external interrupt controller such as the 8259 programmable interrupt controller — Four global high resolution timers/counters that can generate interrupts — Supports 22 other internal interrupt sources — Supports fully nested interrupt delivery — Interrupts can be routed to external pin for external processing. — Interrupts can be routed to the e500 core’s standard or critical interrupt inputs. — Interrupt summary registers allow fast identification of interrupt source. I2C controller — Two-wire interface — Multiple-master support — Master or slave I2C mode support — On-chip digital filtering rejects spikes on the bus Boot sequencer — Optionally loads configuration data from serial ROM at reset through the I2C interface — Can be used to initialize configuration registers and/or memory — Supports extended I2C addressing mode — Data integrity checked with preamble signature and CRC DUART — Two 4-wire interfaces (SIN, SOUT, RTS, CTS) — Programming model compatible with the original 16450 UART and the PC16550D 10/100 fast Ethernet controller (FEC) — Operates at 10 to 100 megabits per second (Mbps) as a device debug and maintenance port Local bus controller (LBC) — Multiplexed 32-bit address and data operating at up to 166 MHz — Eight chip selects support eight external slaves — Four- and eight-beat burst transfers MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 1-5 Overview • • • — The 32-, 16-, and 8-bit port sizes are controlled by an on-chip memory controller. — Three protocol engines available on a per chip select basis: – General-purpose chip select machine (GPCM) – Three user programmable machines (UPMs) – Dedicated single data rate SDRAM controller — Parity support — Default boot ROM chip select with configurable bus width (8-,16-, or 32-bit) Two three-speed (10/100/1Gb) Ethernet controllers (TSECs) — Dual IEEE 802.3, 802.3u, 802.3x, 802.3z, 802.3ac, 802.3ab compliant controllers — Support for different Ethernet physical interfaces: – 10/100/1Gb IEEE 802.3 GMII – 10/100 Mbps IEEE 802.3 MII – 10-Mbps IEEE 802.3 MII – 1-Gbps IEEE 802.3z TBI – 10/100/1Gb RGMII/RTBI — Full- and half-duplex support — Buffer descriptors are backward compatible with MPC8260 and MPC860T 10/100 programming models — 9.6-Kbyte jumbo frame support — RMON statistics support — 2-Kbyte internal transmit and receive FIFOs — MII management interface for control and status — Programmable CRC generation and checking — Ability to force allocation of header information and buffer descriptors into L2 cache OCeaN switch fabric — Four-port crossbar packet switch — Reorders packets from a source based on priorities — Reorders packets to bypass blocked packets — Implements starvation avoidance algorithms — Supports packets with payloads of up to 256 bytes Integrated DMA controller — Four-channel controller — All channels accessible by both the local and remote masters MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 1-6 Freescale Semiconductor M PC8540 Overview — Extended DMA functions (advanced chaining and striding capability) — Support for scatter and gather transfers — Misaligned transfer capability — Interrupt on completed segment, link, list, and error — Supports transfers to or from any local memory or I/O port — Selectable hardware-enforced coherency (snoop/no-snoop) — Ability to start and flow control each DMA channel from external 3-pin interface — Ability to launch DMA from single write transaction PCI/PCI-X controller — PCI 2.2 and PCI-X 1.0 compatible — 64- or 32-bit PCI port supports at 16 to 66 MHz — 64-bit PCI-X support up to 133 MHz — Host and agent mode support — 64-bit dual address cycle (DAC) support — PCI-X supports multiple split transactions — Supports PCI-to-memory and memory-to-PCI streaming — Memory prefetching of PCI read accesses — Supports posting of processor-to-PCI and PCI-to-memory writes — PCI 3.3-V compatible — Selectable hardware-enforced coherency Power management — Fully static 1.2-V CMOS design with 3.3- and 2.5-V I/O — Supports power saving modes: doze, nap, and sleep — Employs dynamic power management, which automatically minimizes power consumption of blocks when they are idle System performance monitor — Supports eight 32-bit counters that count the occurrence of selected events — Ability to count up to 512 counter-specific events — Supports 64 reference events that can be counted on any of the 8 counters — Supports duration and quantity threshold counting — Burstiness feature that permits counting of burst events with a programmable time between bursts — Triggering and chaining capability — Ability to generate an interrupt on overflow MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 1-7 • • • Overview • • • System access port — Uses JTAG interface and a TAP controller to access entire system memory map — Supports 32-bit accesses to configuration registers — Supports cache-line burst accesses to main memory — Supports large block (4-Kbyte) uploads and downloads — Supports continuous bit streaming of entire block for fast upload and download IEEE 1149.1-compliant, JTAG boundary scan 783 FC-PBGA package 1.3 1.3.1 MPC8540 Architecture Overview e500 Core Overview The following sections describe the major functional units of the MPC8540. The MPC8540 uses the e500 microprocessor core complex. The e500 core has an internal PLL that allows independent optimization of the operating frequencies. The core frequencies are derived from either the primary PCI clock input or an external oscillator. For information regarding the e500 core refer to the following documents: • • • EREF: A Reference for Freescale Semiconductor Book E and the e500 Core PowerPC e500 Core Complex Reference Manual PowerPC e500 Application Binary Interface User's Guide NOTE The e500 defines features that are not implemented on the MPC8540. It also generally defines some features that the MPC8540 implements more specifically. An understanding of these differences can be critical to ensure proper operation. These differences are summarized in Section 5.14, “MPC8540 Implementation Details,” in the MPC8540 Integrated Processor Reference Manual. The following is a brief list of some of the key features of the e500 core complex: • • Implements full Book E 32-bit architecture Implements additional instructions, registers, and interrupts defined by APUs. The SPE provides an extensive instruction set for 64-bit vector integer, single-precision floating-point, and fractional operations. The SPFP APU provides scalar (32-bit) single-precision, floating-point instructions. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 1-8 Freescale Semiconductor M PC8540 Architecture Overview NOTE The SPE APU and SPFP APU functionality will be implemented in the MPC8540, the MPC8560 and in their derivatives (that is, in all PowerQUICC III devices). However, these instructions will not be supported in devices subsequent to PowerQUICC III. Freescale Semiconductor strongly recommends that use of these instructions be confined to libraries and device drivers. Customer software that uses SPE or SPFP APU instructions at the assembly level or that uses SPE intrinsics will require rewriting for upward compatibility with next-generation PowerQUICC devices. Freescale Semiconductor offers a libmoto_e500 library that uses SPE and SPFP APU instructions. Freescale will also provide future libraries to support next generation PowerQUICC devices. • L1 cache structure — 32-Kbyte, 32-byte line, eight-way set-associative instruction cache — 32-Kbyte, 32-byte line, eight-way set-associative data cache — 1.5-cycle cache array access, 3-cycle load-to-use latency — Pseudo-LRU replacement algorithm — Copy-back data cache Dual-dispatch superscalar Precise exception handling Seven-stage pipeline control Instruction unit — Twelve-entry instruction queue — Full hardware detection of interlocks — Dispatch up to two instructions per cycle — Dispatch serialization control — Register dependency resolution and renaming Branch unit (BU) — Dynamic branch prediction — Two-entry branch instruction queue (BIQ) — Executes all branch and CR logical instruction Completion unit — As many as 14 instructions allowed in 14-entry completion queue — In-order retirement of up to two instructions per cycle MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 1-9 • • • • • • Overview • • • • — Completion and refetch serialization control — Synchronization for all instruction flow changes—interrupts and mispredicted branches Two simple execution units that perform the following: — Single-cycle add and subtract — Single-cycle shift and rotate — Single-cycle logical operations — Supports integer signal processing operations Multiple-cycle execution unit (MU) — Four-cycle latency for integer and floating-point multiplication (including integer, fractional, and both vector and scalar floating-point multiply instructions). — Variable-latency divide: 4, 11, 19, and 35 cycles for all Book E, SPE, and SPFP divide instructions. Note that the MU allows divide instructions to bypass the second two MU pipeline stages, freeing those stages for other MU instructions to execute in parallel. — Four-cycle floating-point multiply — Four-cycle floating-point add and subtract Signal processing engine APU (SPE APU). The SIMD capability provided by the 64-bit execution units (MIU, LSU, SIU1) is not a separate execution unit. The hardware that executes 32-bit Book E instructions also executes the lower half of 64-bit SPU instructions. — Single-cycle integer add and subtract — Single-cycle logical operations — Single-cycle shift and rotate — Four-cycle integer pipelined multiplies — 4-, 11-, 19-, and 35-cycle integer divides — Four-cycle single instruction multiple data (SIMD) pipelined multiply-accumulate (MAC) — 64-bit accumulator for MAC operations — Single-precision floating-point operations Load/store unit (LSU) — Three-cycle load latency — Fully pipelined — Four-entry load queue allows up to four load misses before stalling — Can continue servicing load hits when load queue is full — Six-entry store queue allows full pipelining of stores MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 1-10 Freescale Semiconductor M PC8540 Architecture Overview • • • • Cache coherency — Bus support for hardware-enforced coherency (bus snooping) Core complex bus (CCB) — High-speed, on-chip local bus with data tagging — 32-bit address bus — 60x-like address protocol with address pipelining and retry/copyback — Two general-purpose read data, one write data bus — 128-bit data plus parity/tags (each data bus) — Supports out-of-order reads, in-order writes — Little to no data bus arbitration logic required for native systems — Easily adaptable to 60x-like environments — Supports one-level pipelining of addresses with address-retry responses Extended exception handling — Supports Book E interrupt model – Interrupt vector prefix register (IVPR) – Vector offset registers (IVORs) 0–15 as defined in Book E, plus e500-defined IVORs 32–35 – Exception syndrome register (ESR) – Book E–defined preempting critical interrupt, including critical interrupt status registers (CSRR0 and CSRR1) and an rfci instruction — e500-specific interrupts not defined in Book E architecture – SPE APU unavailable exception – Floating-point data exception – Floating-point round exception – Performance monitor Memory management unit (MMU) — Data L1 MMU – Four-entry, fully-associative TLB array for variable-sized pages – 64-entry, four-way set-associative TLB for 4-Kbyte pages — Instruction L1 MMU – Four-entry, fully-associative TLB array for variable-sized pages – 64-entry, four-way set-associative TLB for 4-Kbyte pages MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 1-11 Overview • • • — Unified L2 MMU – 16-entry, fully-associative TLB array for variable-sized pages – 256-entry, two-way set-associative TLB for 4-Kbyte pages — Software reload for TLBs — Virtual memory support for as much as 4 Gbytes (232) of virtual memory — Real memory support for as much as 4 Gbytes (232) of physical memory — Support for big-endian and true little-endian memory on a per-page basis Power management — Low power, 1.2-V design — Dynamic power management on the core minimizes power consumption of functional units, such as execution units, caches, and MMUs, when they are idle. — Core power-saving modes: core-halted and core-stopped — NAP, DOZE, and SLEEP bits in HID0 that can be used to assert nap, doze, and sleep core output signals to initiate power-saving modes at the integrated-device level — Internal clock multipliers of 2x, 2.5x, and 3x from bus clock Testability — LSSD scan design — JTAG interface — ESP support — ABIST for arrays — LBIST Reliability and serviceability — Internal code parity — Parity checking on e500 local bus 1.3.2 On-Chip Memory Unit The MPC8540 contains an internal 256-Kbyte memory array that can be configured as memory-mapped SRAM or as a look-aside L2 cache. The array can also be divided into two 128-Kbyte arrays, one of which may be used as cache and the other as SRAM. The memory controller for this array connects to the core complex bus (CCB) and communicates through 128-bit read and write buses to the e500 core and the MPC8540 system logic. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 1-12 Freescale Semiconductor M PC8540 Architecture Overview The on-chip memory unit contains: • 256 Kbytes of on-chip memory — L2 cache partitioning is configurable – Can act as a 256-Kbyte L2 cache –256-Kbyte array organized as 1024 8-way sets of 32-byte cache lines – Array can be partitioned into 128-Kbyte L2 cache and 128-Kbyte memory mapped SRAM – Can act as two 128-Kbyte memory-mapped SRAM arrays or a 256-Kbyte SRAM region — SRAM operation is byte-accessible — Data ECC on 64-bit boundaries (single-error correction, double-error detection) — Tag parity (1 bit covering all tag bits) — Cache mode supports instruction caching, data caching, or both — External masters can force data to be allocated into the cache through programmed memory ranges or special transaction types — Separate locking for instructions and data so that locks can be set and cleared separately — Supports locking the entire cache or selected lines – Individual line locks are set and cleared through core-initiated instructions, by external reads or writes, or by accesses to programmed memory ranges — Flash clearing done through writes to L2 configuration registers — Locks for the entire cache may be set and cleared by accesses to memory-mapped control registers 1.3.2.1 On-Chip Memory as Memory-Mapped SRAM When the on-chip memory is configured as an SRAM, the 256 Kbytes of memory can be configured to reside at any aligned location in the memory map. It is byte-accessible and fully ECC-protected, using read-modify-write transactions for sub-cacheline transactions. I/O devices can access the SRAM by marking transactions global so that they are directed to the CCB. 1.3.2.2 On-Chip Memory as L2 Cache The MPC8540 on-chip memory arrays include a 256-Kbyte data array, an address tag array, and a status array. The data array is organized as 1024 sets of 8 cache lines. Each cache line size is 32 bytes. The replacement policy in each eight-way set is governed by a pseudo-LRU algorithm. The data is protected with ECC, and the tag array is protected by parity. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 1-13 Overview The L2 cache tags are non-blocking for efficient load/store and snooping operations. The L2 cache can be accessed internally while a load miss is pending (allowing hits under misses). Subsequent to a load miss updating the memory, loads or stores can occur to that line on the very next cycle. The L2 status array maintains status bits for each line to determine the status of the line. Different combinations of these bits result in different L2 states. Note that because the cache is always write-through, there is no modified state. The status bits include the following: • • • V—Valid IL—Instruction locked DL—Data locked All accesses to the L2 memory are fully pipelined so back-to-back loads and stores can have single-cycle throughput. The cache can be configured to allocate instructions-only, data-only, or both. It can also be configured to allocate global I/O writes that correspond to a programmable address window or that use a special transaction type (stashing). In this way, DMA engines or I/O devices can force data into the cache. Line locks can be set in a variety of ways. The Book E architecture defines instructions that explicitly set and clear locks in the L2. These instructions are supported by the core complex and the L2 controller. In addition, the L2 controller can be configured to lock all lines that fall into either of two specified address ranges when the line is allocated. Finally, the entire cache can be locked by writing to a configuration register in the L2 cache controller. The status array tracks line locks as either instruction locks or data locks for each line, and the status array supports flash clearing of all instruction locks or data locks separately by writes to configuration registers in the L2 controller. 1.3.3 e500 Coherency Module (ECM) The e500 coherency module (ECM) provides a mechanism for I/O-initiated transactions to snoop the bus between the e500 core and the integrated L2 cache in order to maintain coherency across local cacheable memory. It also provides a flexible switch-type structure for core and I/O-initiated transactions to be routed or dispatched to target modules on the device. 1.3.4 DDR SDRAM Controller The MPC8540 supports DDR-I SDRAM that operates at up to 166 MHz (333-MHz data rate). The memory interface controls main memory accesses and provides for a maximum of 3.5 Gbytes of main memory. The memory controller can be configured to support the various memory sizes through software initialization of on-chip configuration registers. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 1-14 Freescale Semiconductor M PC8540 Architecture Overview The MPC8540 supports a variety of SDRAM configurations. SDRAM banks can be built using DIMMs or directly-attached memory devices. Fifteen multiplexed address signals provide for device densities of 64 Mbits, 128 Mbits, 256 Mbits, and 512 Mbits, and 1 Gbit. Four chip select signals support up to four banks of memory. The MPC8540 supports bank sizes from 64 Mbytes to 1 Gbyte. Nine column address strobes (MDM[0:8]) are used to provide byte selection for memory bank writes. The MPC8540 can be configured to retain the currently active SDRAM page for pipelined burst accesses. Page mode support of up to 16 simultaneously open pages can dramatically reduce access latencies for page hits. Depending on the memory system design and timing parameters, using page mode can save 3 to 4 clock cycles from subsequent burst accesses that hit in an active page. The MPC8540 supports ECC for system memory. Using ECC, the MPC8540 detects and corrects all single-bit errors and detects all double-bit errors and all errors within a nibble. The MPC8540 can invoke a level of system power management by asserting the MCKE SDRAM signal on-the-fly to put the memory into a low-power sleep mode. 1.3.5 Programmable Interrupt Controller (PIC) The programmable interrupt controller (PIC) implements the necessary functions to provide a flexible solution for a general-purpose interrupt control. The interrupt controller unit implements the logic and programming structures of the OpenPIC architecture. The MPC8540 interrupt controller unit supports its processor core and provides for 12 external interrupts (with fully nested interrupt delivery), 4 message interrupts, internal-logic driven interrupts, and 4 global high resolution timers. Up to 16 programmable interrupt priority levels are supported. The interrupt controller unit can be bypassed to allow use of an external interrupt controller. Inter-processor interrupt (IPI) communication is supported through the external interrupt and core reset signals of different processor cores on the same device. The four IPIs are only used for self-interrupt in a single-core device such as the MPC8540. 1.3.6 I2C Controller The inter-IC (IIC or I2C) bus is a two-wire, bidirectional serial bus that provides a simple and efficient method of data exchange between devices. The synchronous, multiple-master bus of the I2C allows the MPC8540 to exchange data with other I2C devices, such as microcontrollers, EEPROMs, real-time clock devices, A/D converters, and LCDs. The two-wire bus (serial data SDA and serial clock SCL) minimizes the interconnections between devices. The synchronous, multiple-master bus of the I2C allows the connection of additional devices to the bus for expansion and system development. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 1-15 Overview The I2C controller is a true multiple-master bus; it includes collision detection and arbitration that prevents data corruption if two or more masters attempt to control the bus simultaneously. This feature allows for complex applications with multiprocessor control. The I2C controller consists of a transmitter/receiver unit, a clocking unit, and a control unit. The I2C unit supports general broadcast mode, and on-chip filtering rejects spikes on the bus. 1.3.7 Boot Sequencer The MPC8540 provides a boot sequencer that uses the I2C interface to access an external serial ROM and loads the data into the MPC8540’s configuration registers. The boot sequencer is enabled by a configuration pin sampled at the negation of the MPC8540 hardware reset signal. If enabled, the boot sequencer holds the MPC8540 processor core in reset until the boot sequence is complete. If the boot sequencer is not enabled, the processor core exits reset and fetches boot code in default configurations. 1.3.8 Dual Universal Asynchronous Receiver/Transmitter (DUART) The MPC8540 includes a DUART intended for use in maintenance, bringing-up, and debugging of systems. The MPC8540 provides a standard four-wire handshake (SIN, SOUT, RTS, CTS) for each port. The DUART is a slave interface. An interrupt is provided to the interrupt controller or optionally steered externally to allow device handshakes. Interrupts are generated for transmit, receive, line status, and MODEM status. The MPC8540 DUART supports full-duplex operation. It is compatible with the PC16450 and PC16550 programming models. Also, 16-byte FIFOs are supported for both the transmitter and the receiver. Software programmable baud generators divide the system clock to generate a 16x clock. Serial interface data formats (data length, parity, 1/1.5/2 STOP bit, baud rate) are also software selectable. 1.3.9 10/100 Fast Ethernet Controller The MPC8540 provides a 10/100 fast Ethernet controller operating at 10 to 100 Mbps as a device debug and maintenance port. The gigabit three-speed Ethernet controllers (TSECs) are separate from this 10/100 fast Ethernet controller and are defined and described in Chapter 14, “Three-Speed Ethernet Controllers.” The 10/100 Ethernet standard supports the IEEE 802.3 Ethernet frame format, backward compatibility for installed media, and the use of full- or half-duplex CSMA/CD. The Ethernet protocol implements the bottom two layers of the open systems interconnection (OSI) 7-layer model, that is, the data-link and physical sublayers. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 1-16 Freescale Semiconductor M PC8540 Architecture Overview 1.3.10 Local Bus Controller (LBC) The MPC8540 local bus controller (LBC) port allows connections with a wide variety of external memories, DSPs, and ASICs. Three separate state machines share the same external pins and can be programmed separately to access different types of devices. The general-purpose chip select machine (GPCM) controls accesses to asynchronous devices using a simple handshake protocol. The user programmable machine (UPM) can be programmed to interface to synchronous devices or custom ASIC interfaces. The SDRAM controller provides access to standard SDRAM. Each chip select can be configured so that the associated chip interface can be controlled by the GPCM, UPM, or SDRAM controller. All may exist in the same system. The GPCM provides a flexible asynchronous interface to SRAM, EPROM, FEPROM, ROM, and other devices such as asynchronous DSP host interfaces and CAMs. Minimal glue logic is required. Handshake signals can be configured to transition on fractions of the system clock. The GPCM does not support bursting. The UPM allows an extremely flexible interface in which the programmer configures each of a set of general-purpose protocol signals by writing the transition pattern into a memory array. The UPM supports synchronous and bursting interfaces. It also supports multiplexed addressing so that a simple DRAM interface can be implemented. The UPM is entirely flexible in order to provide a very high degree of customization with respect to both asynchronous and burst-synchronous interfaces, which permits glueless or almost glueless connection to burst SRAM, custom ASIC, and synchronous DSP interfaces. The LBC provides a synchronous DRAM (SDRAM) machine that supplies the control functions and signals for glueless connection to JEDEC-compliant SDRAM devices. An internal DLL (delay-locked loop) for bus clock generation ensures improved data setup margins for board designs. The SDRAM machine can optimize burst transfers and exploits interleaving to maximize data transfer bandwidth and minimize access latency. Programmable row and column address multiplexing allows a variety of SDRAM configurations and sizes to be supported without hardware changes. 1.3.11 Three-Speed Ethernet Controllers (10/100/1Gb) The MPC8540 has two on-chip three-speed Ethernet controllers (TSECs). The TSECs incorporate a media access control sublayer (MAC) that supports 10- and 100-Mbps, and 1-Gbps Ethernet/802.3 networks with MII, GMII, RGMII, RTBI, and TBI physical interfaces. The TSECs include 2-Kbyte receive and transmit FIFOs, and DMA functions. The buffer descriptors are based on the MPC8260 and MPC860T 10/100 programming models. The MPC8540 TSECs support programmable CRC generation and checking, RMON statistics, and jumbo frames of up to 9.6 Kbytes. Frame headers and buffer descriptors can be forced into the L2 cache to speed classification or other frame processing. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 1-17 Overview 1.3.12 Integrated DMA The MPC8540 DMA engine is capable of transferring blocks of data from any legal address range to any other legal address range. Therefore, it can perform a DMA transfer between any of its I/O or memory ports or even between two devices or locations on the same port. The four-channel DMA controller allows chaining (both extended and direct) through local memory-mapped chain descriptors. Scattering, gathering, and misaligned transfers are supported. In addition, advanced capabilities such as stride transfers and complex transaction chaining are supported. DMA transfers can be initiated by a single write to a configuration register. There is also support for external control of transfers using DMA_DREQ, DMA_DACK, and DMA_DDONE handshake signals. DMA descriptors encompass a rich set of attributes that allow DMA transfers to bypass outbound address translation and supply external addresses and attributes directly to the RapidIO port. Local attributes such as snoop and L2-write stashing can be specified by descriptors. Interrupts are provided on a completed segment, link, list, chain, or on an error condition. Coherency is selectable and hardware enforced (snoop/no snoop). 1.3.13 PCI Controller The MPC8540 64-bit PCI controller is compatible with the PCI Local Bus Specification, Revision 2.2 and the PCI-X Addendum, Revision 1.0. The interface can function as a host or agent bridge interface in either PCI or PCI-X mode. Both PCI and PCI-X modes support 64-bit addressing and 32-bit or 64-bit data buses. As a master, the MPC8540 supports read and write operations to the PCI memory space, the PCI I/O space, and the PCI configuration space. Also, the MPC8540 can generate PCI special-cycle and interrupt-acknowledge commands. As a target, the MPC8540 supports read and write operations to system memory as well as configuration accesses. PCI-X functionality includes split transaction support for four outstanding split transactions. Split response data is returned in order without interleaving. As a target, the MPC8540 supports all PCI-X sizes. As a master it internally combines transactions up to 256 bytes. An internal arbiter can be used to support up to five external masters. A round robin arbitration algorithm with two priority levels is used. 1.3.14 RapidIO Controller The RapidIO interconnect unit on the MPC8540 is based on the RapidIO Interconnect Specification, Revision 1.1. RapidIO is a high-performance, point-to-point, low-pin-count, MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 1-18 Freescale Semiconductor M PC8540 Architecture Overview packet-switched system-level interconnect that can be used in a variety of applications as an open standard. The RapidIO architecture provides a rich variety of features including high data bandwidth, low-latency capability, and support for high-performance I/O devices, as well as providing message-passing and software-managed programming models. The RapidIO unit on the MPC8540 supports the I/O and message-passing logical specifications, the common transport specification, and the 8/16 LP-LVDS physical layer specification of the RapidIO Interconnect Specification. It does not support the globally shared memory logical specification. Highlights of the implementation include: support for four priority levels and ordering within a priority level, CRC error management, 32- to 256-byte transactions and 8-bit data width ports. The physical layer of the RapidIO unit can operate at up to 500 MHz. Because the interface is defined as a source-synchronous, double-data-rate, LVDS-signaling interconnect, the theoretical unidirectional peak bandwidth is 8 Gbps. Receive and transmit ports operate independently, resulting in an aggregate theoretical bandwidth of 16 Gbps. 1.3.14.1 RapidIO Message Unit The MPC8540’s RapidIO messaging supports one inbox/outbox structure for data and one doorbell structure for messages. Both chaining and direct modes are provided for the outbox, and messages can hold up to 16 packets of 256 bytes, or a total of 4 Kbytes. 1.3.15 Power Management In addition to low-voltage operation and dynamic power management in its execution units, the MPC8540 supports four power consumption modes: full-on, doze, nap, and sleep. The three low-power modes: doze, nap, and sleep, can be entered under software control in the e500 core or by external masters accessing a configuration register. Doze mode suspends execution of instructions in the e500 core. The core is left in a standby mode in which cache snooping and time base interrupts are still enabled. Device logic external to the processor core is fully functional in this mode. Nap mode shuts down clocks to all the e500 functional units except the time base, which can be disabled separately. No snooping is performed in nap mode, but the device logic external to the processor core is fully functional. Sleep mode shuts down not only the e500 core, but all of the MPC8540 I/O interfaces as well. Only the interrupt controller and power management logic remain enabled so that the device can be awakened. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 1-19 Overview 1.3.16 Clocking The MPC8540 takes in the PCI_CLK/SYSCLK signal as an input to the device PLL and multiplies it by an integer from 1 to 16 to generate the core complex bus clock (the platform clock), which operates at the same frequency as the DDR DRAM data rate (for example, 266 or 333 MHz). The L2 cache also operates at this frequency. The e500 core uses the CCB clock as an input to its PLL, which multiplies it again by 2, 2.5, 3, or 3.5 to generate the core clock. DLLs are used in the DDR SDRAM controller and the local bus memory controller (LBC) to generate memory clocks. Six differential clock pairs are generated for DDR SDRAMs. Two clock outputs are generated for the LBC. The RapidIO transmit clock may be sourced from one of three locations: the platform clock, the RapidIO receive clock, or a special differential clock input. This input is designed to receive inputs from an external clock synthesis device driving a clock with a frequency of up to 500 MHz. 1.3.17 Address Map The MPC8540 supports a flexible physical address map. Conceptually, the address map consists of local space and external address space. The local address map is 4 Gbytes. The MPC8540 can be made part of a larger system address space through the mapping of translation windows. This functionality is included in the address translation and mapping units (ATMUs). Both inbound and outbound translation windows are provided. The ATMUs allows the MPC8540 to be part of larger address maps such as the PCI 64-bit address environment and the RapidIO environment. 1.3.18 OCeaN Switch Fabric In order to reduce the strain on the core interconnects with the addition of new functional blocks in this generation of the PowerQUICC family, an on-chip non-blocking crossbar switch fabric called OCeaN (on-chip network) has been integrated to decrease contention, decrease latency, and increase bandwidth. This revolutionary non-blocking crossbar fabric allows for full-duplex port connections at 128 Gbps concurrent throughput and independent per-port transaction queuing and flow control. 1.3.19 Processing Across the On-Chip Fabric When processing across the on-chip fabric, the ATMUs at each fabric port are used to determine the flow of data across the MPC8540. The ATMUs at each fabric port are responsible for generating a fabric port destination ID as well as a new local device address. The port ID and local MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 1-20 Freescale Semiconductor MPC8540 Application Examples address are based on the programmed destination of the transaction. The following is a general overview of how the ATMUs process transactions over the on-chip fabric. (Refer to Figure 1-2.) 1. When a transaction on one of the fabric ports begins, the ATMU on the origination port translates the programmed destination address into both a destination fabric port ID and a local device address. 2. The data is then processed across the on-chip fabric from the origination port to the destination port. 3. If the destination port connects off-chip (for example, to a PCI or RapidIO device), the local device address is translated by the destination port ATMU to an outbound address with respect to the destination port’s memory map, and the data is processed accordingly. 2 1 3 ATMU ATMU Figure 1-2. Processing Transactions Across the On-Chip Fabric 1.3.20 Data Processing with the e500 Coherency Module Processing through the ECM is similar to processing across the on-chip fabric (in the sense of how data is received and transmitted) with the exception that the transaction passes through the ECM. The purpose of the ECM is to provide a means for any I/O transaction to maintain coherency with the cacheable DDR SDRAM and the local bus memory. However, simply using the ECM does not make transactions across it coherent. The e500 and L2 cache are snooped to maintain coherency only if the transaction across the ECM is designated as global (GBL bit set). Otherwise, the transaction passes through the ECM using the ECM as a simple conduit to get to its destination. In essence, only global transactions across the ECM are coherent transactions; all other transactions (across the on-chip fabric) are non-coherent. 1.4 MPC8540 Application Examples The following section provides block diagrams of different MPC8540 applications. The MPC8540 is a very flexible device and can be configured to meet many system application needs. In order to build a system, many factors should be considered. 1.4.1 MPC8540 Applications The MPC8540 can be used for control processing in applications such as routers, switches, network memory, internet access devices, storage systems, and image display systems (IADs). Figure 1-3 shows an MPC8540 as a control plane processor in a high-performance communication system using symmetric multiprocessing (SMP). MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 1-21 Overview DDR Memory Local Bus Controller ASIC Accelerator Hardware RapidIO Switch Task Processor Processor 256K L2 Core Memory Complex Block RapidIO Interface To Network Forwarding Plane ASIC MPC8540 Processor 256K L2 Memory Core Block Complex RapidIO 2x GEnet Memory Control Dual Gigabit Ethernet DDR Memory 66-MHz PCI Tundra Tsi 400 PCI To Control Plane 133-MHz PCI-X Figure 1-3. High-Performance Communication System Using SMP Figure 1-4 shows the MPC8540 as part of a high-end network card used in a system area network that is enabled by PCI/PCI-X. Server Data Control Dual 10/100/1Gb Ethernet MPC8540 Processor Core Complex RapidIO PCI 256K L2 Memory Block 2x GEnet Memory Control FPGA ROM DDR Memory GPCM/UPM/ SDRAM RTC/NVRAM PCI-X 133-MHz Backplane RAID Controller Figure 1-4. High-Performance Communication System Using MPC8540 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 1-22 Freescale Semiconductor MPC8540 Application Examples Figure 1-5 shows the MPC8540 in a redundant array of independent disks (RAID) controller application. MPC8540 Processor 256K L2 Core Memory Complex Block RapidIO PCI 2x GEnet Memory Control Disk Cache SCSI Control RapidIO Switch ASIC XOR DMA DDR Memory MPC8540 Processor 256K L2 Core Memory Complex Block RapidIO PCI 2x GEnet Memory Control Disk Cache SCSI Control ASIC XOR DMA PCI-X to System DDR Memory DDR Memory Descriptor Memory Flash DDR Memory Descriptor Memory PCI-X Flash PCI-X to System Figure 1-5. RAID Controller Application Using MPC8540 Figure 1-6 illustrates the MPC8540 in a virtual private network (VPN) access that is enabled through RapidIO and Ethernet. MPC8540 Processor Core Complex RapidIO High Bandwidth Aggregators Data In Data Out RapidIO PCI 190 Security 64-Bit PCI Interface 256K L2 Memory Block 2x GEnet Memory Control DDR Memory 64-Bit 66 MHz PCI Data In Data Out 10/100/1G Ethernet Figure 1-6. VPN Access Point Enabled by RapidIO and Ethernet MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 1-23 Overview Figure 1-7 shows the MPC8540 in a serializer/deserializer (SERDES) data transfer mode application enabled by Ethernet. MPC8540 Processor Core Complex RapidIO PCI 256K L2 Memory Block 2x GEnet Memory Control SERDES Interface 10/100/1Gb Ethernet Data In Using 10-Bit Interface Data Out Using 10-Bit Interface SERDES DDR Memory Figure 1-7. MPC8540 with SERDES Figure 1-8 shows a DSP farm enabled by the MPC8540 using its local bus controller port. Core Complex Unit DSP DSP RapidIO DSP EPROM DSP Local Bus Controller DSP Farm 190 Security 64-Bit PCI Interface DDR Memory 64-Bit 66-MHz PCI PCI 2x GEnet Memory Control GPCM/UPM Local Bus Controller Port 256K L2 Memory Block Data In Data Out 10/100/1Gb Ethernet Figure 1-8. MPC8540 with a DSP Farm Enabled by General-Purpose Chip Select Machine or User Programmable Machine on the Local Bus Controller Port MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 1-24 Freescale Semiconductor Chapter 2 Memory Map This chapter describes the MPC8540 memory map. An overview of the local address map is followed by a description of how local access windows are used to define the local address map. The inbound and outbound address translation mechanisms used to map to and from external memory spaces are described next. Finally, the configuration, control, and status registers are described, including a complete listing of all memory-mapped registers with cross references to the sections detailing descriptions of each. 2.1 Local Memory Map Overview and Example The MPC8540 provides an extremely flexible local memory map. The local memory map refers to the 32-bit address space seen by the processor as it accesses memory and I/O space. DMA engines also see this same local memory map. All memory accessed by the MPC8540 DDR SDRAM and local bus memory controllers exists in this memory map, as do all memory-mapped configuration, control, and status registers. The local memory map is defined by a set of eight local access windows. Each of these windows map a region of memory to a particular target interface, such as the DDR SDRAM controller or the PCI controller. Note that the local access windows do not perform any address translation. The size of each window can be configured from 4 Kbytes to 2 Gbytes. The target interface is specified using the codes shown in Table 2-1. . Table 2-1. Target Interface Codes Target Interface PCI/PCI-X Local bus RapidIO DDR SDRAM Target Code 0000 0100 1100 1111 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 2-1 Memory Map Figure 2-1 shows an example memory map. 0 Example Local Memory Map 0xA000_0000 DDR SDRAM Memory PCI-X 0xB000_0000 0x8000_000 Local Bus SRAM 0xA000_000 I/O Configuration Registers 0xC000_0000 Local Bus DSP RapidIO Local Bus Flash 0xFFFF_FFFF CCSR Boot ROM Figure 2-1. Local Memory Map Example Table 2-2 shows one corresponding set of local access window settings. Table 2-2. Local Access Windows Example Window 0 1 2 3 4 5–7 Base Address 0x0000_0000 0x8000_0000 0xA000_0000 0xB000_0000 0xC000_0000 Size 2 Gbytes 1 Mbyte 256 Mbytes 256 Mbytes 256 Mbytes Unused Target Interface 0b1111 (DDR SDRAM) 0b0100 (local bus) 0b0000 (PCI/PCI-X) 0b1100 (RapidIO) 0b0100 (local bus) In this example, it is not necessary to use a local access window to specify the location of the boot ROM because it is in the default location at the highest 8 Mbytes of memory (see Section 4.4.3.3, “Boot ROM Location”). Neither is it required to define a local access window to describe the range of memory used for memory-mapped registers because this is a fixed 1-Mbyte space pointed to by CCSRBAR. See Section 4.3.1.1.2, “Configuration, Control, and Status Base Address Register MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 2-2 Freescale Semiconductor Address Translation and Mapping (CCSRBAR).” However, note that the e500 core only provides one default TLB entry to access boot code and it allows for accesses within the highest 4 Kbytes of memory. In order to access the full 8 Mbytes of default boot space (and the 1 Mbyte of CCSR space), additional TLB entries must be set up within the e500 core for mapping these regions. 2.2 Address Translation and Mapping Four distinct types of translation and mapping operations are performed on transactions in the MPC8540. These are as follows: • • • • Mapping a local address to a target interface Assigning attributes to transactions Translating the local 32-bit address to an external address space Translating external addresses to the local 32-bit address space The local access windows perform target mapping for transactions within the local address space. No address translation is performed by the local access windows. Outbound ATMU windows perform the mapping from the local 32-bit address space to the address spaces of RapidIO or PCI/PCI-X, which may be much larger than the local space. Outbound ATMU windows also map attributes such as transaction type or priority level. Inbound ATMU windows perform the address translation from the external address space to the local address space, attach attributes and transaction types to the transaction, and also map the transaction to its target interface. Note that in mapping the transaction to the target interface, an inbound ATMU window performs a similar function as the local access windows. The target mappings created by an inbound ATMU must be consistent with those of the local access windows. That is, if an inbound ATMU maps a transaction to a given local address and a given target, a local access window must also map that same local address to the same target. All of the configuration registers that define translation and mapping functions use the concept of translation or mapping windows, and all follow the same register format. Table 2-3 summarizes the general format of these window definitions. Table 2-3. Format of ATMU Window Definitions Register Translation address Base address Window size/attributes Function High-order address bits defining location of the window in the target address space High-order address bits defining location of the window in the initial address space Window enable, window size, target interface, and transaction attributes Windows must be a power-of-two size. To perform a translation or mapping function, the address of the transaction is compared with the base address register of each window. The number of bits used in the comparison is dictated by each window’s size attribute. When an address hits a MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 2-3 Memory Map window, if address translation is being performed, the new translated address is created by concatenating the window offset to the translation address. Again, the windows size attribute dictates how many bits are translated. 2.2.1 SRAM Windows The on-chip memory array of the MPC8540 can be configured as a memory-mapped SRAM of 128 or 256 Kbytes. Configuration registers in the L2 cache controller set the base addresses and sizes for these windows. When enabled, these windows supersede all other mappings of these addresses for processor and global (snoopable) I/O transactions. Therefore, SRAM windows must never overlap configuration space as defined by CCSRBAR. It is possible to have SRAM windows overlap local access windows, but this is discouraged because processor and snoopable I/O transactions would map to the SRAM while non-snooped I/O transactions would be mapped by the local access windows. Only if all accesses to the SRAM address range are snoopable can results be consistent if the SRAM window overlaps a local access window. See Section 7.3.1.4, “L2 Memory-Mapped SRAM Base Address Registers 0–1 (L2SRBARn),” for information about configuring SRAM windows. 2.2.2 Window into Configuration Space CCSRBAR defines a window used to access all memory-mapped configuration, control, and status registers. No address translation is done, so there are no associated translation address registers. The window is always enabled with a fixed size of 1 Mbyte; no other attributes are attached, so there is no associated size/attribute register. This window always takes precedence over all local access windows. See Section 4.3.1.1.2, “Configuration, Control, and Status Base Address Register (CCSRBAR),” and Section 2.3, “Configuration, Control, and Status Register Map.” 2.2.3 Local Access Windows As demonstrated in the address map overview in Section 2.1, “Local Memory Map Overview and Example,” local access windows associate a range of the local 32-bit address space with a particular target interface. This allows the internal interconnections of the MPC8540 to route a transaction from its source to the proper target. No address translation is performed. The base address defines the high order address bits that give the location of the window in the local address space. The window attributes enable the window, define its size, and specify the target interface. With the exception of configuration space (mapped by CCSRBAR), on-chip SRAM regions (mapped by L2SRBAR registers), and default boot ROM, all addresses used by the system must be mapped by a local access window. This includes addresses that are mapped by inbound ATMU windows; target mappings of inbound ATMU windows and local access windows must be consistent. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 2-4 Freescale Semiconductor Address Translation and Mapping The local access window registers exist as part of the local access block in the general utilities registers. See Section 2.3.3, “General Utilities Registers.” A detailed description of the local access window registers is given in the following sections. Note that the minimum size of a window is 4 Kbytes, so the low order 12 bits of the base address cannot be specified. 2.2.3.1 Local Access Register Memory Map Table 2-4. Local Access Register Memory Map Table 2-4 shows the memory map for the local access registers. Local Memory Offset (Hex) 0x0_0C08 0x0_0C10 0x0_0C28 0x0_0C30 0x0_0C48 0x0_0C50 0x0_0C68 0x0_0C70 0x0_0C88 0x0_0C90 0x0_0CA8 0x0_0CB0 0x0_0CC8 0x0_0CD0 0x0_0CE8 0x0_0CF0 Register LAWBAR0—Local access window 0 base address register LAWAR0—Local access window 0 attribute register LAWBAR1—Local access window 1 base address register LAWAR1—Local access window 1 attribute register LAWBAR2—Local access window 2 base address register LAWAR2—Local access window 2 attribute register LAWBAR3—Local access window 3 base address register LAWAR3—Local access window 3 attribute register LAWBAR4—Local access window 4 base address register LAWAR4—Local access window 4 attribute register LAWBAR5—Local access window 5 base address register LAWAR5—Local access window 5 attribute register LAWBAR6—Local access window 6 base address register LAWAR6—Local access window 6 attribute register LAWBAR7—Local access window 7 base address register LAWAR7—Local access window 7 attribute register Access R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W Reset 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 Section/Page 2.2.3.2/2-6 2.2.3.3/2-6 2.2.3.2/2-6 2.2.3.3/2-6 2.2.3.2/2-6 2.2.3.3/2-6 2.2.3.2/2-6 2.2.3.3/2-6 2.2.3.2/2-6 2.2.3.3/2-6 2.2.3.2/2-6 2.2.3.3/2-6 2.2.3.2/2-6 2.2.3.3/2-6 2.2.3.2/2-6 2.2.3.3/2-6 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 2-5 Memory Map 2.2.3.2 Local Access Window n Base Address Registers (LAWBAR0–LAWBAR7) Figure 2-2 shows the bit fields of the LAWBARn registers. 0 11 12 31 R0 W Reset Offset 0 0000000 0 00 BASE_ADDR 0000_0000_0000_0000_0000_0000_0000_0000 0xC08, 0xC28, 0xC48, 0xC68, 0xC88, 0xCA8, 0xCC8, 0xCE8 Figure 2-2. Local Access Window n Base Address Registers (LAWBAR0–LAWBAR7) Table 2-5 describes LAWBARn field descriptions. Table 2-5. LAWBARn Field Descriptions Bits 0–11 Name — Write reserved, read = 0 Description 12–31 BASE_ADDR Identifies the 20 most-significant address bits of the base of local access window n. The specified base address should be aligned to the window size, as defined by LAWAR n[SIZE]. 2.2.3.3 Local Access Window n Attributes Registers (LAWAR0–LAWAR7) Figure 2-3 shows the bit fields of the LAWARn registers. 0 1 7 8 11 12 25 26 31 RE 0 0 0 0 0 0 0 WN Reset Offset 00000000000000 TRGT_IF 0000_0000_0000_0000_0000_0000_0000_0000 SIZE 0xC10, 0xC30, 0xC50, 0xC70, 0xC90, 0xCB0, 0xCD0, 0xCF0 Figure 2-3. Local Access Window n Attributes Registers (LAWAR0–LAWAR7) Table 2-6 describes LAWARn field descriptions. Table 2-6. LAWARn Field Descriptions Bits 0 Name EN Description 0 The local access window n (and all other LAWARn and LAWBARn fields) are disabled. 1 The local access window n is enabled and other LAWARn and LAWBARn fields combine to identify an address range for this window. Write reserved, read = 0 1–7 — MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 2-6 Freescale Semiconductor Address Translation and Mapping Table 2-6. LAWARn Field Descriptions (continued) Bits 8–11 Name Description TRGT_IF Identifies the target interface ID when a transaction hits in the address range defined by this window. Note that configuration registers and SRAM regions are mapped by the windows defined by CCSRBAR and L2SRBAR. These mappings supersede local access window mappings, so configuration registers and SRAM do not appear as a target for local access windows. 0000 PCI/PCI-X 0001–0011Reserved 0100 Local bus memory controller 0101–1011Reserved 1100 RapidIO 1101–1110Reserved 1111 DDR SDRAM — SIZE Write reserved, read = 0 Identifies the size of the window from the starting address. Window size is 2(SIZE+1) bytes. 000000–001010Reserved 001011 4 Kbytes 001100 8 Kbytes 00110116 Kbytes . . . . . . . 2(SIZE+1) bytes 011110 2 Gbytes 011111–111111 Reserved 12–25 26–31 2.2.3.4 Precedence of Local Access Windows If two local access windows overlap, the lower numbered window takes precedence. For instance, if two windows are setup as shown in Table 2-7, local access window 1 governs the mapping of the 1-Mbyte region from 0x7FF0_0000 to 0x7FFF_FFF, even though the window described in local access window 2 also encompasses that memory region. Table 2-7. Overlapping Local Access Windows Window 1 2 Base Address 0x7FF0_0000 0x0000_0000 Size 1 Mbyte 2 Gbytes Target Interface 0b0100 (Local bus controller —LBC) 0b1111 (DDR SDRAM) 2.2.3.5 Configuring Local Access Windows Once a local access window is enabled, it should not be modified while any device in the system may be using the window. Neither should a new window be used until the effect of the write to the window is visible to all blocks that use the window. This can be guaranteed by completing a read of the last local access window configuration register before enabling any other devices to use the window. For instance, if local access windows 0–3 are being configured in order during the initialization process, the last write (to LAWAR3) should be followed by a read of LAWAR3 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 2-7 Memory Map before any devices try to use any of these windows. If the configuration is being done by the local e500 processor, the read of LAWAR3 should be followed by an isync instruction. 2.2.3.6 Distinguishing Local Access Windows from Other Mapping Functions It is important to distinguish between the mapping function performed by the local access windows and the additional mapping functions that happen at the target interface. The local access windows define how a transaction is routed through the MPC8540 internal interconnects from the transactions source to its target. After the transaction has arrived at its target interface, that interface controller may perform additional mapping. For instance, the DDR SDRAM controller has chip select registers that map a memory request to a particular external device. Similarly, the local bus controller has base registers that perform a similar function. The RapidIO and PCI interfaces have outbound address translation and mapping units that map the local address into an external address space. These other mapping functions are configured by programming the configuration, control, and status registers of the individual interfaces. Note that there is no need to have a one-to-one correspondence between local access windows and chip select regions or outbound ATMU windows. A single local access window can be further decoded to any number of chip selects or to any number or outbound ATMU windows at the target interface. 2.2.3.7 Illegal Interaction Between Local Access Windows and DDR SDRAM Chip Selects If a local access window maps an address to an interface other than the DDR SDRAM controller, then there should not be a valid chip select configured for the same address in the DDR SDRAM controller. Because DDR SDRAM chip selects boundaries are defined by a beginning and ending address, it is easy to define them so that they do not overlap with local access windows that map to other interfaces. 2.2.4 Outbound Address Translation and Mapping Windows Outbound address translation and mapping refers to the translation of addresses from the local 32-bit address space to the external address space and attributes of a particular I/O interface. On the MPC8540, both the RapidIO and the PCI/PCI-X blocks have outbound address translation and mapping units (ATMUs). The RapidIO controller has eight outbound ATMU windows plus a default window. If a transaction’s address does not hit any of the eight outbound ATMU windows, the translation actions defined by the default window are used. The default window is always enabled. See Section 17.3.2.2, “ATMU Registers,” for a detailed description of the RapidIO outbound ATMU windows. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 2-8 Freescale Semiconductor Address Translation and Mapping The PCI/PCI-X controller has four outbound ATMU windows plus a default window. The PCI/PCI-X outbound ATMU registers include an extended translation address register so that up to 64 bits of external address space can be supported. See Section 16.3.1.2, “PCI/X ATMU Outbound Registers,” for a detailed description of the PCI/PCI-X outbound ATMU windows. 2.2.5 Inbound Address Translation and Mapping Windows Inbound address translation and mapping refers to the translation of an address from the external address space of an I/O interface (such as PCI or RapidIO address space) to the local address space understood by the internal interfaces of the MPC8540. It also refers to the mapping of transactions to a particular target interface and the assignment of transaction attributes. Both the RapidIO controller and the PCI/PCI-X controller have inbound address translation and mapping units (ATMUs). 2.2.5.1 RapidIO Inbound ATMU The RapidIO controller has four inbound ATMU windows plus a default. If the inbound transaction’s address does not hit any of the four inbound ATMU windows, the translation actions defined by the default window are used. The default window is always enabled. See Section 17.3.2.2, “ATMU Registers,” for a detailed description of the RapidIO inbound ATMU windows. 2.2.5.2 PCI/PCI-X Inbound ATMU The PCI/PCI-X controller has three general inbound ATMU windows plus a dedicated window for memory mapped configuration accesses (PCSRBAR). These windows have a one-to-one correspondence with the base address registers in the PCI/PCI-X programming model. Updating one automatically updates the other. There is no default inbound window; if a PCI/PCI-X address does not match one of the inbound ATMU windows, the MPC8540 does not respond with an assertion of PCI_DEVSEL. See Section 16.3.1.3, “PCI/X ATMU Inbound Registers,” for a detailed description of the PCI/PCI-X inbound ATMU windows. 2.2.5.3 Illegal Interaction Between Inbound ATMUs and Local Access Windows Since both local access windows and inbound ATMUs map transactions to a target interface, it is essential that they not contradict one another. For instance, it is a programming error to have an inbound ATMU map a transaction to the DDR SDRAM memory controller (target interface 0b1111) if the resulting translated local address is mapped to PCI (target interface 0b0000) by a local access window. Such a programming error may result in unpredictable system deadlocks. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 2-9 Memory Map 2.3 Configuration, Control, and Status Register Map All of the memory mapped configuration, control, and status registers in the MPC8540 are contained within a 1-Mbyte address region. To allow for flexibility, the configuration, control, and status block is relocatable in the local address space. The local address map location of this register block is controlled by the configuration, control, and status registers base address register (CCSRBAR), see Section 4.3.1.1.2, “Configuration, Control, and Status Base Address Register (CCSRBAR).” The default value for CCSRBAR is 4 Gbytes–9 Mbytes, or 0xFF70_0000. NOTE The configuration, control, and status window must not overlap a local access window that maps to the DDR controller. Otherwise, undefined behavior occurs. An example of a top-level memory map with the default location of the configuration, control, and status registers is shown in Figure 2-4. Example Local Memory Map 0 0x0_0000 General Utilities CCSR Register Memory Block Memory 0x4_0000 PIC I/O 0x8_0000 Reserved 0xC_0000 CCSRBAR (4G-9M) CCSR 4G-8M Only reserved if Local Boot ROM 0xE_0000 0xF_FFFF Shaded area indicates locations not allowed for CCSRBAR. RapidIO Device-Specific Utilities 4G Figure 2-4. Top-Level Register Map Example MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 2-10 Freescale Semiconductor Configuration, Control, and Status Register Map 2.3.1 Accessing CCSR Memory from External Masters In addition to being accessible by the e500 processor, the configuration, control, and status registers are accessible from external interfaces. This allows external masters on the I/O ports to configure the MPC8540. External masters do not need to know the location of the CCSR memory in the local address map. Rather, they access this region of the local memory map through a window defined by a register in the interface’s programming model that is accessible to the external master from its external memory map. The PCI base address for accessing the local CCSR memory is selectable through the PCI configuration and status register base address register (PCSRBAR), at offset 0x10, described in Section 16.3.2.11, “PCI Base Address Registers.” An external PCI master sets this register by running a PCI configuration cycle to the MPC8540. Subsequent memory accesses by a PCI master to the PCI address range indicated by PCSRBAR are translated to the local address indicated by the current setting of CCSRBAR. The RapidIO base address for accessing the local CCSR memory is selectable through the RapidIO LCSBA1CSR, defined in the RapidIO programming model, see Section 17.3.1.12, “Local Configuration Space Base Address 1 Command and Status Register (LCSBA1CSR).” An external RapidIO master can set the value of LCSBA1CSR with a maintenance packet. Then subsequent read and write packets whose RapidIO addresses match the window defined by LCSBA1CSR are translated to the local address range indicated by CCSRBAR. 2.3.2 Organization of CCSR Memory The configuration, control, and status registers of the MPC8540 are grouped according to functional units. Most functional blocks are allocated for a 4-Kbyte address space for registers. Registers that fall into this category are referred to as general utilities registers. These registers occupy the first 256 Kbytes of CCSR memory. Registers that control functions that are not particular to a functional unit but to the device as a whole occupy the highest 256 Kbytes of CCSR memory. These are referred to as device-specific registers. Some functional units, such as RapidIO and the OpenPIC-based interrupt controller have larger address spaces as defined by their programming models. The registers for these blocks are given their own large regions of CCSR memory. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 2-11 Memory Map Table 2-8. Local Memory Configuration, Control, and Status Register Summary Offset from CCSRBAR 0x0_0000–0x3_FFFF 0x4_0000–0x7_FFFF 0x8_0000–0xB_FFFF 0xC_0000–0xD_FFFF 0xE_0000–0xF_FFFF Register Grouping General utilities Programmable interrupt controller (PIC) Reserved RapidIO Device-specific utilities MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 2-12 Freescale Semiconductor Configuration, Control, and Status Register Map 2.3.3 General Utilities Registers Figure 2-5 provides an overview of the general utilities registers. General Utilities Registers Memory Block 0x0 0000 0x0 1000 0x0 2000 0x0 3000 CCSR Memory Block 0x0 0000 General Utilities 0x0 4000 0x0 5000 0x0 6000 0x0 7000 0x0 8000 PCI General Utility Block 0x4 0000 0xn n000 Local Access ECM Memory I2C DUART Local Bus OCeaN PIC General Registers 0x2 0000 0x2 1000 L2 Cache DMA 0xn nC00 0x2 4000 TSEC 1 TSEC 2 10_100 FEC 0xn nE00 0xn nF00 0x8 0000 Reserved ATMU Error Mgmt Debug 0xC 0000 RIO 0xE 0000 0xF FFFF Device-Specific Utilities 0x2 5000 0x2 6000 0x3 FFFF Figure 2-5. General Utilities Registers Mapping to Configuration, Control, and Status Memory Block Figure 2-5 also shows the organization of registers inside the 4-Kbyte register space allocated to an individual functional block. The first 3 Kbytes are available for general registers. The next 512 bytes are dedicated to address translation and mapping registers, if applicable to that particular MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 2-13 Memory Map functional unit (for example, PCI). If a unit has error management registers, they are typically placed starting at offset 0xE00 from the beginning of the block’s 4-Kbyte space, and any debug registers are typically placed in the final 256 bytes of the unit’s register space starting at offset 0xF00. General utilities registers are accessed as 32-bit quantities except for the DUART and I2C registers, which are accessed as bytes. NOTE Refer to detailed register descriptions for each functional unit for exact locations, sizes, and access requirements. Some blocks may have exceptions to the above guidelines. 2.3.4 Interrupt Controller and CCSR The programmable interrupt controller (PIC) registers are at offset 0x4_0000 from CCSRBAR, see Figure 2-6. Its programming model follows the OpenPIC architecture. The interrupt controller registers should only be accessed with 32-bit accesses. CCSR Register Memory Block 0x0 0000 General Utilities 0x4 0000 PIC Registers 0x4 0000 0x4 1000 0x4 1100 0x4 1200 0x5 0000 0x5 1020 Global Cfg Timers External IRQs Internal IRQs PIC 0x6 0000 Processor 0x8 0000 0x7 FFFF Reserved 0xC 0000 RIO 0xE 0000 0xF FFFF Device-Specific Utilities Figure 2-6. PIC Mapping to Configuration, Control, and Status Memory Block MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 2-14 Freescale Semiconductor Configuration, Control, and Status Register Map 2.3.5 RapidIO and CCSR The RapidIO module uses 128 Kbytes of CCSR memory; refer to Figure 2-7. All registers are 32-bits wide and should only be accessed with 32-bit accesses. The 4-Kbyte RapidIO implementation block has the same internal organization as those defined for the general utilities described above. CCSR Register Memory Block 0x0 0000 General Utilities 0xC 0000 0xC 4100 RapidIO Extended Features RapidIO Configuration RapidIO Arch. 0x4 0000 PIC 0x8 0000 0xD 0000 0xD 1000 Reserved RapidIO Impl. RapidIO Messaging 0xC 0000 RIO 0xE 0000 0xF FFFF Device-Specific Utilities Figure 2-7. RapidIO Mapping to Configuration, Control, and Status Memory Block 2.3.6 Device-Specific Utilities The device-specific registers consist of power management, performance monitors, and device-wide debug utilities (refer to Figure 2-8). These registers are accessible with 32-bit accesses only. Transactions of other than 32-bit are considered a programming error and operation is undefined. Reserved bits in the following register descriptions are not guaranteed to have predictable values. Software must preserve the values of reserved bits when writing to a register. Also, when reading from a register, software should not rely on the value of any reserved bit remaining consistent. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 2-15 Memory Map Device-Specific Registers 0x0 0000 General Utilities 0xE 0000 0xE 1000 0xE 2000 Global Utilities Perf. Monitor Watchpoint/Debug 0x4 0000 PIC 0x8 0000 Reserved 0xC 0000 RIO 0xE 0000 0xF FFFF Device-Specific Utilities 0xF FFFC Figure 2-8. Device-Specific Register Mapping to Configuration, Control, and Status Memory Block 2.4 Complete CCSR Map Table 2-9. Memory Map Offset Register Access Reset Section/Page Table 2-9 lists the MPC8540 memory-mapped registers. Local-Access Registers—Configuration, Control, and Status Registers 0x0_0000 0x0_0008 0x0_0010 0x0_0020 CCSRBAR—Configuration, control, and status registers base address register ALTCBAR—Alternate configuration base address register ALTCAR—Alternate configuration attribute register BPTR—Boot page translation register R/W R/W R/W R/W 0x000F_F700 0x0000_0000 0x0000_0000 0x0000_0000 4.3.1.1.2/4-5 4.3.1.2.1/4-6 4.3.1.2.2/4-6 4.3.1.3.1/4-8 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 2-16 Freescale Semiconductor Complete CCSR Map Table 2-9. Memory Map (continued) Offset Register Access Reset Section/Page Local-Access Registers—Local-Access Window Base and Size Registers 0x0_0C08 0x0_0C10 0x0_0C28 0x0_0C30 0x0_0C48 0x0_0C50 0x0_0C68 0x0_0C70 0x0_0C88 0x0_0C90 0x0_0CA8 0x0_0CB0 0x0_0CC8 0x0_0CD0 0x0_0CE8 0x0_0CF0 LAWBAR0—Local access window 0 base address register LAWAR0—Local access window 0 attribute register LAWBAR1—Local access window 1 base address register LAWAR1—Local access window 1 attribute register LAWBAR2—Local access window 2 base address register LAWAR2—Local access window 2 attribute register LAWBAR3—Local access window 3 base address register LAWAR3—Local access window 3 attribute register LAWBAR4—Local access window 4 base address register LAWAR4—Local access window 4 attribute register LAWBAR5—Local access window 5 base address register LAWAR5—Local access window 5 attribute register LAWBAR6—Local access window 6 base address register LAWAR6—Local access window 6 attribute register LAWBAR7—Local access window 7 base address register LAWAR7—Local access window 7 attribute register Registers 0x0_1000 0x0_1010 0x0_1E00 0x0_1E08 0x0_1E0C 0x0_1E10 EEBACR—ECM CCB address configuration register EEBPCR—ECM CCB port configuration register EEDR—ECM error detect register EEER—ECM error enable register EEATR—ECM error attributes capture register EEADR—ECM error address capture register R/W R/W R/W R/W R R 0x0000_0003 0x0n00_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 8.2.1.1/8-3 8.2.1.2/8-4 8.2.1.3/8-5 8.2.1.4/8-6 8.2.1.5/8-7 8.2.1.6/8-8 R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 2.2.3.2/2-6 2.2.3.3/2-6 2.2.3.2/2-6 2.2.3.3/2-6 2.2.3.2/2-6 2.2.3.3/2-6 2.2.3.2/2-6 2.2.3.3/2-6 2.2.3.2/2-6 2.2.3.3/2-6 2.2.3.2/2-6 2.2.3.3/2-6 2.2.3.2/2-6 2.2.3.3/2-6 2.2.3.2/2-6 2.2.3.3/2-6 DDR Memory Controller Memory Map 0x0_2000 0x0_2008 0x0_2010 0x0_2018 CS0_BNDS—Chip select 0 memory bounds CS1_BNDS—Chip select 1 memory bounds CS2_BNDS—Chip select 2 memory bounds CS3_BNDS—Chip select 3 memory bounds R/W 0x0000_0000 9.4.1.1/9-10 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 2-17 Memory Map Table 2-9. Memory Map (continued) Offset 0x0_2080 0x0_2084 0x0_2088 0x0_208C 0x0_2108 0x0_210C 0x0_2110 0x0_2118 0x0_2124 0x0_2E00 0x0_2E04 0x0_2E08 0x0_2E20 0x0_2E24 0x0_2E28 0x0_2E40 0x0_2E44 0x0_2E48 0x0_2E4C 0x0_2E50 0x0_2E58 Register CS0_CONFIG—Chip select 0 configuration CS1_CONFIG—Chip select 1 configuration CS2_CONFIG—Chip select 2 configuration CS3_CONFIG—Chip select 3 configuration TIMING_CFG_1—DDR SDRAM timing configuration 1 TIMING_CFG_2—DDR SDRAM timing configuration 2 DDR_SDRAM_CFG—DDR SDRAM control configuration DDR_SDRAM_MODE—DDR SDRAM mode configuration DDR_SDRAM_INTERVAL—DDR SDRAM interval configuration DATA_ERR_INJECT_HI—Memory data path error injection mask high DATA_ERR_INJECT_LO—Memory data path error injection mask low ECC_ERR_INJECT—Memory data path error injection mask ECC CAPTURE_DATA_HI—Memory data path read capture high CAPTURE_DATA_LO—Memory data path read capture low CAPTURE_ECC—Memory data path read capture ECC ERR_DETECT—Memory error detect ERR_DISABLE—Memory error disable ERR_INT_EN—Memory error interrupt enable CAPTURE_ATTRIBUTES—Memory error attributes capture CAPTURE_ADDRESS—Memory error address capture ERR_SBE—Single-Bit ECC memory error management I2C 0x0_3000 0x0_3004 0x0_3008 0x0_300C I2CADR—I2C address register I2CFDR—I C frequency divider register I2CCR—I C control register I2CSR—I C status register 2 2 2 Access R/W Reset 0x0000_0000 Section/Page 9.4.1.2/9-10 R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W 0x0000_0000 0x0000_0000 0x0200_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 9.4.1.3/9-11 9.4.1.4/9-13 9.4.1.5/9-14 9.4.1.6/9-16 9.4.1.7/9-16 9.4.1.8/9-17 9.4.1.9/9-18 9.4.1.10/9-18 9.4.1.11/9-19 9.4.1.12/9-20 9.4.1.13/9-20 9.4.1.14/9-21 9.4.1.15/9-21 9.4.1.16/9-22 9.4.1.17/9-23 9.4.1.18/9-24 9.4.1.19/9-25 R/W R/W R/W R/W 0x00 0x00 0x00 0x81 11.3.1.1/11-5 11.3.1.2/11-5 11.3.1.3/11-6 11.3.1.4/11-7 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 2-18 Freescale Semiconductor Complete CCSR Map Table 2-9. Memory Map (continued) Offset 0x0_3010 0x0_3014 Register I2CDR—I2C data register I2CDFSRR—I C digital filter sampling rate register DUART Registers 0x0_4500 0x0_4500 0x0_4500 0x0_4501 0x0_4501 0x0_4502 0x0_4502 0x0_4502 0x0_4503 0x0_4504 0x0_4505 0x0_4506 0x0_4507 0x0_4510 0x0_4600 0x0_4600 0x0_4600 0x0_4601 0x0_4601 0x0_4602 0x0_4602 URBR—ULCR[DLAB] = 0 UART0 receiver buffer register UTHR—ULCR[DLAB] = 0 UART0 transmitter holding register UDLB—ULCR[DLAB] = 1 UART0 divisor least significant byte register UIER—ULCR[DLAB] = 0 UART0 interrupt enable register UDMB—ULCR[DLAB] = 1 UART0 divisor most significant byte register UIIR—ULCR[DLAB] = 0 UART0 interrupt ID register UFCR—ULCR[DLAB] = 0 UART0 FIFO control register UAFR—ULCR[DLAB] = 1 UART0 alternate function register ULCR—ULCR[DLAB] = x UART0 line control register UMCR—ULCR[DLAB] = x UART0 MODEM control register ULSR—ULCR[DLAB] = x UART0 line status register UMSR—ULCR[DLAB] = x UART0 MODEM status register USCR—ULCR[DLAB] = x UART0 scratch register UDSR—ULCR[DLAB] = x UART0 DMA status register URBR—ULCR[DLAB] = 0 UART1 receiver buffer register UTHR—ULCR[DLAB] = 0 UART1 transmitter holding register UDLB—ULCR[DLAB] = 1 UART1 divisor least significant byte register UIER—ULCR[DLAB] = 0 UART1 interrupt enable register UDMB_ULCR[DLAB] = 1 UART1 divisor most significant byte register UIIR—ULCR[DLAB] = 0 UART1 interrupt ID register UFCR—ULCR[DLAB] = 0 UART1 FIFO control register R W R/W R/W R/W R W R/W R/W R/W R R R/W R R W R/W R/W R/W R W 0x00 0x00 0x00 0x00 0x00 0x01 0x00 0x00 0x00 0x00 0x60 0x00 0x00 0x01 0x00 0x00 0x00 0x00 0x00 0x01 0x00 12.3.1.1/12-6 12.3.1.2/12-6 12.3.1.3/12-7 12.3.1.4/12-9 12.3.1.3/12-7 12.3.1.5/12-10 12.3.1.6/12-11 12.3.1.12/12-17 12.3.1.7/12-12 12.3.1.8/12-14 12.3.1.9/12-15 12.3.1.10/12-16 12.3.1.11/12-17 12.3.1.13/12-18 12.3.1.1/12-6 12.3.1.2/12-6 12.3.1.3/12-7 12.3.1.4/12-9 12.3.1.3/12-7 12.3.1.5/12-10 12.3.1.6/12-11 2 Access R/W R/W Reset 0x00 0x10 Section/Page 11.3.1.5/11-9 11.3.1.6/11-10 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 2-19 Memory Map Table 2-9. Memory Map (continued) Offset 0x0_4602 0x0_4603 0x0_4604 0x0_4605 0x0_4606 0x0_4607 0x0_4610 Register UAFR—ULCR[DLAB] = 1 UART1 alternate function register ULCR—ULCR[DLAB] = x UART1 line control register UMCR—ULCR[DLAB] = x UART1 MODEM control register ULSR—ULCR[DLAB] = x UART1 line status register UMSR—ULCR[DLAB] = x UART1 MODEM status register USCR—ULCR[DLAB] = x UART1 scratch register UDSR—ULCR[DLAB] = x UART1 DMA status register Access R/W R/W R/W R R R/W R Reset 0x00 0x00 0x00 0x60 0x00 0x00 0x01 Section/Page 12.3.1.12/12-17 12.3.1.7/12-12 12.3.1.8/12-14 12.3.1.9/12-15 12.3.1.10/12-16 12.3.1.11/12-17 12.3.1.13/12-18 Local Bus Controller Registers 0x0_5000 0x0_5008 0x0_5010 0x0_5018 0x0_5020 0x0_5028 0x0_5030 0x0_5038 0x0_5004 0x0_500C 0x0_5014 0x0_501C 0x0_5024 0x0_502C 0x0_5034 0x0_503C 0x0_5068 0x0_5070 0x0_5074 0x0_5078 0x0_5084 0x0_5088 0x0_5094 0x0_50A0 BR0—Base register 0 BR1—Base register 1 BR2—Base register 2 BR3—Base register 3 BR4—Base register 4 BR5—Base register 5 BR6—Base register 6 BR7—Base register 7 OR0—Options register 0 OR1—Options register 1 OR2—Options register 2 OR3—Options register 3 OR4—Options register 4 OR5—Options register 5 OR6—Options register 6 OR7—Options register 7 MAR—UPM address register MAMR—UPMA mode register MBMR—UPMB mode register MCMR—UPMC mode register MRTPR—Memory refresh timer prescaler register MDR—UPM data register LSDMR—SDRAM mode register LURT—UPM refresh timer R/W R/W R/W R/W R/W R/W R/W R/W 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 13.3.1.3/13-18 13.3.1.4/13-19 13.3.1.4/13-19 13.3.1.4/13-19 13.3.1.5/13-21 13.3.1.6/13-22 13.3.1.7/13-22 13.3.1.8/13-24 R/W 0x0000_0FF7 0x0000_0000 13.3.1.2/13-12 R/W 0x0000_nn011 0x0000_0000 13.3.1.1/13-11 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 2-20 Freescale Semiconductor Complete CCSR Map Table 2-9. Memory Map (continued) Offset 0x0_50A4 0x0_50B0 0x0_50B4 0x0_50B8 0x0_50BC 0x0_50C0 0x0_50D0 0x0_50D4 Register LSRT—SDRAM refresh timer LTESR—Transfer error status register LTEDR—Transfer error disable register LTEIR—Transfer error interrupt register LTEATR—Transfer error attributes register LTEAR—Transfer error address register LBCR—Configuration register LCRR—Clock ratio register PCI/X Registers PCI/X Configuration Access Registers 0x0_8000 0x0_8004 0x0_8008 0x0_800C– 0x0_8BFC CFG_ADDR—PCI/X configuration address CFG_DATA—PCI/X configuration data INT_ACK—PCI/X interrupt acknowledge Reserved R/W R/W R — PCI/X ATMU Registers—Outbound and Inbound 0x0_8C00–0x0_8C3C–Outbound Window 0 (default) 0x0_8C00 0x0_8C04 0x0_8C08 0x0_8C0C 0x0_8C10 0x0_8C14– 0x0_8C1C POTAR0—PCI/X outbound window 0 (default) translation address register POTEAR0—PCI/X outbound window 0 (default) translation extended address register Reserved Reserved POWAR0—PCI/X outbound window 0 (default) attributes register Reserved R/W R/W — — R/W — 0x0_8C20–0x0_8C3C—Outbound Window 1 0x0_8C20 0x0_8C24 0x0_8C28 0x0_8C2C 0x0_8C30 POTAR1—PCI/X outbound window 1 translation address register POTEAR1—PCI/X outbound window 1 translation extended address register POWBAR1—PCI/X outbound window 1 base address register Reserved POWAR1—PCI/X outbound window 1 attributes register R/W R/W R/W — R/W 0x0000_0000 0x0000_0000 0x0000_0000 — 0x0000_0000 16.3.1.2.4/16-22 16.3.1.2.1/16-21 16.3.1.2.2/16-21 16.3.1.2.3/16-22 0x0000_0000 0x0000_0000 — — 0x8004_401F — 16.3.1.2.4/16-22 16.3.1.2.1/16-21 16.3.1.2.2/16-21 0x0000_0000 0x0000_0000 0x0000_0000 — 16.3.1.1.1/16-18 16.3.1.1.1/16-18 16.3.1.1.3/16-20 — Access R/W Read/ Bit-reset Reset 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x8000_0008 Section/Page 13.3.1.9/13-25 13.3.1.10/13-26 13.3.1.11/13-27 13.3.1.12/13-28 13.3.1.13/13-29 13.3.1.14/13-30 13.3.1.15/13-31 13.3.1.16/13-32 R/W R/W R/W R/W R/W R/W MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 2-21 Memory Map Table 2-9. Memory Map (continued) Offset 0x0_8C34– 0x0_8C3C Reserved Register Access — 0x0_8C40–0x0_8C5C—Outbound Window 2 0x0_8C40 0x0_8C44 0x0_8C48 0x0_8C4C 0x0_8C50 0x0_8C54– 0x0_8C5C POTAR2—PCI/X outbound window 2 translation address register POTEAR2—PCI/X outbound window 2 translation extended address register POWBAR2—PCI/X outbound window 2 base address register Reserved POWAR2—PCI/X outbound window 2 attributes register Reserved R/W R/W R/W — R/W — 0x0_8C60–0x0_8C7C—Outbound Window 3 0x0_8C60 0x0_8C64 0x0_8C68 0x0_8C6C 0x0_8C70 0x0_8C74– 0x0_8C7C POTAR3—PCI/X outbound window 3 translation address register POTEAR3—PCI/X outbound window 3 translation extended address register POWBAR3—PCI/X outbound window 3 base address register Reserved POWAR3—PCI/X outbound window 3 attributes register Reserved R/W R/W R/W — R/W — 0x0_8C80–0x0_8C9C—Outbound Window 4 0x0_8C80 0x0_8C84 0x0_8C88 0x0_8C8C POTAR4—PCI/X outbound window 4 translation address register POTEAR4—PCI/X outbound window 4 translation extended address register POWBAR4—PCI/X outbound window 4 base address register Reserved R/W R/W R/W — 0x0000_0000 0x0000_0000 0x0000_0000 — 16.3.1.2.1/16-21 16.3.1.2.2/16-21 16.3.1.2.3/16-22 0x0000_0000 0x0000_0000 0x0000_0000 — 0x0000_0000 — 16.3.1.2.4/16-22 16.3.1.2.1/16-21 16.3.1.2.2/16-21 16.3.1.2.3/16-22 0x0000_0000 0x0000_0000 0x0000_0000 — 0x0000_0000 — 16.3.1.2.4/16-22 16.3.1.2.1/16-21 16.3.1.2.2/16-21 16.3.1.2.3/16-22 Reset — Section/Page MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 2-22 Freescale Semiconductor Complete CCSR Map Table 2-9. Memory Map (continued) Offset 0x0_8C90 0x0_8C94– 0x0_8D9C Register POWAR4—PCI/X outbound window 4 attributes register Reserved Access R/W — 0x0_8DA0–0x0_8DBC–Inbound Window 3 0x0_8DA0 0x0_8DA4 0x0_8DA8 0x0_8DAC 0x0_8DB0 0x0_8DB4– 0x0_8DBC PITAR3—PCI/X inbound window 3 translation address register Reserved PIWBAR3—PCI/X inbound window 3 base address register PIWBEAR3—PCI/X inbound window 3 base extended address register PIWAR3—PCI/X inbound window 3 attributes register Reserved R/W — R/W R/W R/W — 0x0_8DC0–0x0_8DDC–Inbound Window 2 0x0_8DC0 0x0_8DC4 0x0_8DC8 0x0_8DCC 0x0_8DD0 0x0_8DD4– 0x0_8DDC PITAR2—PCI/X inbound window 2 translation address register Reserved PIWBAR2—PCI/X inbound window 2 base address register PIWBEAR2—PCI/X inbound window 2 base extended address register PIWAR2—PCI/X inbound window 2 attributes register Reserved R/W — R/W R/W R/W — 0x0_8DE0–0x0_8DFC–Inbound Window 1 0x0_8DE0 0x0_8DE4 0x0_8DE8 0x0_8DEC 0x0_8DF0 0x0_8DF4– 0x0_8DFC PITAR1—PCI/X inbound window 1 translation address register Reserved PIWBAR1—PCI/X inbound window 1 base address register Reserved PIWAR1—PCI/X inbound window 1 attributes register Reserved R/W — R/W — R/W — 0x0000_0000 — 0x0000_0000 — 0x0000_0000 — 16.3.1.3.4/16-26 16.3.1.3.2/16-25 16.3.1.3.1/16-25 0x0000_0000 — 0x0000_0000 0x0000_0000 0x0000_0000 — 16.3.1.3.2/16-25 16.3.1.3.3/16-26 16.3.1.3.4/16-26 16.3.1.3.1/16-25 0x0000_0000 — 0x0000_0000 0x0000_0000 0x0000_0000 — 16.3.1.3.2/16-25 16.3.1.3.3/16-26 16.3.1.3.4/16-26 16.3.1.3.1/16-25 Reset 0x0000_0000 — Section/Page 16.3.1.2.4/16-22 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 2-23 Memory Map Table 2-9. Memory Map (continued) Offset Register Access Reset Section/Page PCI/X Error Management Registers 0x0_8E00 0x0_8E04 0x0_8E08 0x0_8E0C 0x0_8E10 0x0_8E14 0x0_8E18 0x0_8E1C 0x0_8E20 0x0_8E24 0x0_8E28– 0x0_8EFC 0x0_8F00– 0x0_8FFC ERR_DR—PCI/X error detect register ERR_CAP_DR—PCI/X error capture disabled register ERR_EN—PCI/X error enable register ERR_ATTRIB—PCI/X error attributes capture register ERR_ADDR—PCI/X error address capture register ERR_EXT_ADDR—PCI/X error extended address capture register ERR_DL—PCI/X error data low capture register ERR_DH—PCI/X error data high capture register GAS_TIMR—PCI/X gasket timer register PCIX_TIMR—PCIX split completion timer register Reserved Reserved for debug Special R/W R/W R/W R/W R/W R/W R/W R/W R/W — — 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 — — 16.3.1.4.1/16-29 16.3.1.4.2/16-30 16.3.1.4.3/16-31 16.3.1.4.4/16-32 16.3.1.4.5/16-33 16.3.1.4.6/16-34 16.3.1.4.7/16-34 16.3.1.4.8/16-34 16.3.1.4.9/16-35 16.3.1.4.10/16-36 L2/SRAM Memory-Mapped Configuration Registers 0x2_0000 0x2_0010 0x2_0018 0x2_0020 0x2_0028 0x2_0030 0x2_0038 0x2_0040 0x2_0048 0x2_0100 0x2_0108 L2CTL—L2 control register L2CEWAR0—L2 cache external write address register 0 L2CEWCR0—L2 cache external write control register 0 L2CEWAR1—L2 cache external write address register 1 L2CEWCR1—L2 cache external write control register 1 L2CEWAR2—L2 cache external write address register 2 L2CEWCR2—L2 cache external write control register 2 L2CEWAR3—L2 cache external write address register 3 L2CEWCR3—L2 cache external write control register 3 L2SRBAR0—L2 memory-mapped SRAM base address register 0 L2SRBAR1—L2 memory-mapped SRAM base address register 1 R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W 0x2000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 7.3.1.1/7-7 7.3.1.2/7-10 7.3.1.3/7-10 7.3.1.2/7-10 7.3.1.3/7-10 7.3.1.2/7-10 7.3.1.3/7-10 7.3.1.2/7-10 7.3.1.3/7-10 7.3.1.4/7-11 7.3.1.4/7-11 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 2-24 Freescale Semiconductor Complete CCSR Map Table 2-9. Memory Map (continued) Offset 0x2_0E00 0x2_0E04 0x2_0E08 0x2_0E20 0x2_0E24 0x2_0E28 0x2_0E40 0x2_0E44 0x2_0E48 0x2_0E4C 0x2_0E50 0x2_0E58 Register L2ERRINJHI—L2 error injection mask high register L2ERRINJLO—L2 error injection mask low register L2ERRINJCTL—L2 error injection tag/ECC control register L2CAPTDATAHI—L2 error data high capture register L2CAPTDATALO—L2 error data low capture register L2CAPTECC—L2 error syndrome register L2ERRDET—L2 error detect register L2ERRDIS—L2 error disable register L2ERRINTEN—L2 error interrupt enable register L2ERRATTR—L2 error attributes capture register L2ERRADDR—L2 error address capture register L2ERRCTL—L2 error control register DMA Registers General Registers 0x2_1100 0x2_1104 0x2_1108 0x2_110C 0x2_1110 0x2_1114 0x2_1118 0x2_111C 0x2_1120 0x2_1124 0x2_1128 0x2_1130 0x2_1134 0x2_1138 0x2_113C 0x2_1140 0x2_1144 MRn—DMA 0 mode register SRn—DMA 0 status register Reserved CLNDARn—DMA 0 current link descriptor address register SATRn—DMA 0 source attributes register SARn—DMA 0 source address register DATRn—DMA 0 destination attributes register DARn—DMA 0 destination address register BCRn—DMA 0 byte count register Reserved NLNDARn—DMA 0 next link descriptor address register Reserved CLSDAR n—DMA 0 current list alternate base descriptor address register Reserved NLSDAR n—DMA 0 next list descriptor address register SSRn—DMA 0 source stride register DSRn—DMA 0 destination stride register R/W Special — R/W R/W R/W R/W R/W R/W — R/W — R/W — R/W R/W R/W 0x0000_0000 0x0000_0000 — 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 — 0x0000_0000 — 0x0000_0000 — 0x0000_0000 0x0000_0000 0x0000_0000 15.3.2.1/15-10 15.3.2.2/15-13 — 15.3.2.3/15-14 15.3.2.4/15-16 15.3.2.5/15-18 15.3.2.6/15-19 15.3.2.7/15-21 15.3.2.8/15-23 — 15.3.2.9/15-23 — 15.3.2.10/15-24 — 15.3.2.11/15-25 15.3.2.12/15-25 15.3.2.13/15-26 Access R/W R/W R/W R R R Special R/W R/W R/W R R/W Reset 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 Section/Page 7.3.1.5.1/7-13 7.3.1.5.1/7-13 7.3.1.5.1/7-13 7.3.1.5.2/7-15 7.3.1.5.2/7-15 7.3.1.5.2/7-15 7.3.1.5.2/7-15 7.3.1.5.2/7-15 7.3.1.5.2/7-15 7.3.1.5.2/7-15 7.3.1.5.2/7-15 7.3.1.5.2/7-15 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 2-25 Memory Map Table 2-9. Memory Map (continued) Offset 0x2_1148– 0x2_117C 0x2_1180 0x2_1184 0x2_1188 0x2_118C 0x2_1190 0x2_1194 0x2_1198 0x2_119C 0x2_11A0 0x2_11A4 0x2_11A8 0x2_11AC– 0x2_11B0 0x2_11B4 0x2_11B8 0x2_11BC 0x2_11C0 0x2_11C4 0x2_11C8– 0x2_11FC 0x2_1200 0x2_1204 0x2_1208 0x2_120C 0x2_1210 0x2_1214 0x2_1218 0x2_121C 0x2_1220 0x2_1224 Reserved MRn—DMA 1 mode register SRn—DMA 1 status register Reserved CLNDARn—DMA 1 current link descriptor address register SATRn—DMA 1 source attributes register SARn—DMA 1 source address register DATRn—DMA 1 destination attributes register DARn—DMA 1 destination address register BCRn—DMA 1 byte count register Reserved NLNDARn—DMA 1 next link descriptor address register Reserved CLSDAR n—DMA 1 current list alternate base descriptor address register Reserved NLSDAR n—DMA 1 next list descriptor address register SSRn—DMA 1 source stride register DSRn—DMA 1 destination stride register Reserved MRn—DMA 2 mode register SRn—DMA 2 status register Reserved CLNDARn—DMA 2 current link descriptor address register SATRn—DMA 2 source attributes register SARn—DMA 2 source address register DATRn—DMA 2 destination attributes register DARn—DMA 2 destination address register BCRn—DMA 2 byte count register Reserved Register Access — R/W Special — R/W R/W R/W R/W R/W R/W — R/W — R/W — R/W R/W R/W — R/W Special — R/W R/W R/W R/W R/W R/W — Reset — 0x0000_0000 0x0000_0000 — 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 — 0x0000_0000 — 0x0000_0000 — 0x0000_0000 0x0000_0000 0x0000_0000 — 0x0000_0000 0x0000_0000 — 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 — Section/Page — 15.3.2.1/15-10 15.3.2.2/15-13 — 15.3.2.3/15-14 15.3.2.4/15-16 15.3.2.5/15-18 15.3.2.6/15-19 15.3.2.7/15-21 15.3.2.8/15-23 — 15.3.2.9/15-23 — 15.3.2.10/15-24 — 15.3.2.11/15-25 15.3.2.12/15-25 15.3.2.13/15-26 — 15.3.2.1/15-10 15.3.2.2/15-13 — 15.3.2.3/15-14 15.3.2.4/15-16 15.3.2.5/15-18 15.3.2.6/15-19 15.3.2.7/15-21 15.3.2.8/15-23 — MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 2-26 Freescale Semiconductor Complete CCSR Map Table 2-9. Memory Map (continued) Offset 0x2_1228 0x2_122C– 0x2_1230 0x2_1234 0x2_1238 0x2_123C 0x2_1240 0x2_1244 0x2_1248– 0x2_127C 0x2_1280 0x2_1284 0x2_1288 0x2_128C 0x2_1290 0x2_1294 0x2_1298 0x2_129C 0x2_12A0 0x2_12A4 0x2_12A8 0x2_12AC– 0x2_12B0 0x2_12B4 0x2_12B8 0x2_12BC 0x2_12C0 0x2_12C4 0x2_12C8– 0x2_12FC 0x2_1300 Register NLNDARn—DMA 2 next link descriptor address register Reserved CLSDAR n—DMA 2 current list alternate base descriptor address register Reserved NLSDAR n—DMA 2 next list descriptor address register SSRn—DMA 2 source stride register DSRn—DMA 2 destination stride register Reserved MRn—DMA 3 mode register SRn—DMA 3 status register Reserved CLNDARn—DMA 3 current link descriptor address register SATRn—DMA 3 source attributes register SARn—DMA 3 source address register DATRn—DMA 3 destination attributes register DARn—DMA 3 destination address register BCRn—DMA 3 byte count register Reserved NLNDARn—DMA 3 next link descriptor address register Reserved CLSDAR n—DMA 3 current list alternate base descriptor address register Reserved NLSDAR n—DMA 3 next list descriptor address register SSRn—DMA 3 source stride register DSRn—DMA 3 destination stride register Reserved DGSR—DMA general status register Access R/W — R/W — R/W R/W R/W — R/W Special — R/W R/W R/W R/W R/W R/W — R/W — R/W — R/W R/W R/W — Read Reset 0x0000_0000 — 0x0000_0000 — 0x0000_0000 0x0000_0000 0x0000_0000 — 0x0000_0000 0x0000_0000 — 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 — 0x0000_0000 — 0x0000_0000 — 0x0000_0000 0x0000_0000 0x0000_0000 — 0x0000_0000 Section/Page 15.3.2.9/15-23 — 15.3.2.10/15-24 — 15.3.2.11/15-25 15.3.2.12/15-25 15.3.2.13/15-26 — 15.3.2.1/15-10 15.3.2.2/15-13 — 15.3.2.3/15-14 15.3.2.4/15-16 15.3.2.5/15-18 15.3.2.6/15-19 15.3.2.7/15-21 15.3.2.8/15-23 — 15.3.2.9/15-23 — 15.3.2.10/15-24 — 15.3.2.11/15-25 15.3.2.12/15-25 15.3.2.13/15-26 — 15.3.2.14/15-26 TSEC1 General Control and Status Registers MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 2-27 Memory Map Table 2-9. Memory Map (continued) Offset 0x2_4000– 0x2_400C 0x2_4010 0x2_4014 0x2_4018 0x2_401C 0x2_4020 0x2_4024 0x2_4028 0x2_402C 0x2_4030 0x2_4034– 0x2_4088 Reserved IEVENT—Interrupt event register IMASK—Interrupt mask register EDIS—Error disabled register Reserved ECNTRL—Ethernet control register MINFLR—Minimum frame length register PTV—Pause time value register DMACTRL—DMA control register TBIPA—TBI PHY address register Reserved Register Access R R/W R/W R/W R R/W R/W R/W R/W R/W R TSEC1 FIFO Control and Status Registers 0x2_404C 0x2_4050– 0x2_4088 0x2_408C 0x2_4090– 0x2_4094 0x2_4098 0x2_409C 0x2_40A0– 0x2_40FC FIFO_PAUSE_CTRL—FIFO pause control register Reserved FIFO_TX_THR—FIFO transmit threshold register Reserved FIFO_TX_STARVE—FIFO transmit starve register FIFO_TX_STARVE_SHUTOFF—FIFO transmit starve shutoff register Reserved R/W R R/W R R/W R/W R TSEC1 Transmit Control and Status Registers 0x2_4100 0x2_4104 0x2_4108 0x2_410C 0x2_4110 0x2_4114– 0x2_4120 0x2_4124 0x2_4128– 0x2_4180 0x2_4184 TCTRL—Transmit control register TSTAT—Transmit status register Reserved TBDLEN—TxBD data length register TXIC—Transmit interrupt coalescing configuration register Reserved CTBPTR—Current TxBD pointer register Reserved TBPTR—TxBD pointer register R/W R/W R R R/W R R R R/W 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 14.5.3.3.1/14-33 14.5.3.3.2/14-34 — 14.5.3.3.3/14-35 14.5.3.3.4/14-35 — 14.5.3.3.5/14-36 — 14.5.3.3.6/14-37 0x0000_0000 0x0000_0000 0x0000_0100 0x0000_0000 0x0000_0080 0x0000_0100 0x0000_0000 14.5.3.2.1/14-30 — 14.5.3.2.2/14-31 — 14.5.3.2.3/14-32 14.5.3.2.4/14-32 — Reset 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0040 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 Section/Page — 14.5.3.1.1/14-20 14.5.3.1.2/14-23 14.5.3.1.3/14-24 — 14.5.3.1.4/14-25 14.5.3.1.5/14-26 14.5.3.1.6/14-27 14.5.3.1.7/14-28 14.5.3.1.8/14-29 — MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 2-28 Freescale Semiconductor Complete CCSR Map Table 2-9. Memory Map (continued) Offset 0x2_4188– 0x2_4200 0x2_4204 0x2_4208– 0x2_42AC 0x2_42B0 0x2_42B4 0x2_42B8– 0x2_42FC Reserved TBASE—TxBD base address register Reserved OSTBD—Out-of-sequence TxBD register OSTBDP—Out-of-sequence Tx data buffer pointer register Reserved Register Access R R/W R R/W R/W R TSEC1 Receive Control and Status Registers 0x2_4300 0x2_4304 0x2_4308 0x2_430C 0x2_4310 0x2_4314– 0x2_4320 0x2_4324 0x2_4328– 0x2_433C 0x2_4340 0x2_4344– 0x2_4380 0x2_4384 0x2_4388– 0x2_4400 0x2_4404 0x2_4408– 0x2_44FC RCTRL—Receive control register RSTAT—Receive status register Reserved RBDLEN—RxBD data length register RXIC—Receive interrupt coalescing configuration register Reserved CRBPTR—Current RxBD pointer register Reserved MRBLR—Maximum receive buffer length register Reserved RBPTR—RxBD pointer register Reserved RBASE—RxBD base address register Reserved TSEC1 MAC Registers 0x2_4500 0x2_4504 0x2_4508 0x2_450C 0x2_4510 0x2_4514– 0x2_451C MACCFG1—MAC configuration register 1 MACCFG2—MAC configuration register 2 IPGIFG—Inter-packet gap/inter-frame gap register HAFDUP—Half-duplex register MAXFRM—Maximum frame length register Reserved R/W R/W R/W R/W R/W R 0x0000_0000 0x0000_7000 0x4060_5060 0x00A1_F037 0x0000_0600 0x0000_0000 14.5.3.6.1/14-49 14.5.3.6.2/14-51 14.5.3.6.3/14-52 14.5.3.6.4/14-53 14.5.3.6.5/14-54 — R/W R/W R R R/W R R R R/W R R/W R R/W R 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 14.5.3.4.1/14-41 14.5.3.4.2/14-41 — 14.5.3.4.3/14-42 14.5.3.4.4/14-43 — 14.5.3.4.5/14-44 — 14.5.3.4.6/14-44 — 14.5.3.4.7/14-45 — 14.5.3.4.8/14-45 — Reset 0x0000_0000 0x0000_0000 0x0000_0000 0x0800_0000 0x0000_0000 0x0000_0000 Section/Page — 14.5.3.3.7/14-37 — 14.5.3.3.8/14-38 14.5.3.3.9/14-40 — MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 2-29 Memory Map Table 2-9. Memory Map (continued) Offset 0x2_4520 0x2_4524 0x2_4528 0x2_452C 0x2_4530 0x2_4534 0x2_4538 0x2_453C 0x2_4540 0x2_4544 0x2_4548– 0x2_467C Register MIIMCFG—MII management configuration register MIIMCOM—MII Management command register MIIMADD—MII Management address register MIIMCON—MII Management control register MIIMSTAT—MII Management status register MIIMIND—MII Management indicator register Reserved IFSTAT—Interface status register MACSTNADDR1—Station address register, part 1 MACSTNADDR2—Station address register, part 2 Reserved TSEC1 RMON MIB Registers TSEC1 Transmit and Receive Counters 0x2_4680 0x2_4684 0x2_4688 0x2_468C 0x2_4690 0x2_4694 0x2_4698 TR64—Transmit and receive 64-byte frame counter register TR127—Transmit and receive 65- to 127-byte frame counter register TR255—Transmit and receive 128- to 255-byte frame counter register TR511—Transmit and receive 256- to 511-byte frame counter register TR1K—Transmit and receive 512- to 1023-byte frame counter register TRMAX—Transmit and receive 1024- to 1518-byte frame counter register TRMGV—Transmit and receive 1519- to 1522-byte good VLAN frame count register TSEC1 Receive Counters 0x2_469C 0x2_46A0 0x2_46A4 0x2_46A8 0x2_46AC 0x2_46B0 0x2_46B4 0x2_46B8 RBYT—receive byte counter register RPKT—receive packet counter register RFCS—receive FCS error counter register RMCA—receive multicast packet counter register RBCA—receive broadcast packet counter register RXCF—receive control frame packet counter register RXPF—receive PAUSE frame packet counter register RXUO—receive unknown OP code counter register R/W R/W R/W R/W R/W R/W R/W R/W 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 14.5.3.7.8/14-64 14.5.3.7.9/14-65 14.5.3.7.10/14-65 14.5.3.7.11/14-66 14.5.3.7.12/14-66 14.5.3.7.13/14-67 14.5.3.7.14/14-67 14.5.3.7.15/14-68 R/W R/W R/W R/W R/W R/W R/W 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 14.5.3.7.1/14-61 14.5.3.7.2/14-61 14.5.3.7.3/14-62 14.5.3.7.4/14-62 14.5.3.7.5/14-63 14.5.3.7.6/14-63 14.5.3.7.7/14-64 Access R/W R/W R/W W R R R R/W R/W R/W R Reset 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 Section/Page 14.5.3.6.6/14-54 14.5.3.6.7/14-55 14.5.3.6.8/14-56 14.5.3.6.9/14-57 14.5.3.6.10/14-57 14.5.3.6.11/14-58 — 14.5.3.6.12/14-58 14.5.3.6.13/14-59 14.5.3.6.14/14-60 — MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 2-30 Freescale Semiconductor Complete CCSR Map Table 2-9. Memory Map (continued) Offset 0x2_46BC 0x2_46C0 0x2_46C4 0x2_46C8 0x2_46CC 0x2_46D0 0x2_46D4 0x2_46D8 0x2_46DC Register RALN—receive alignment error counter register RFLR—receive frame length error counter register RCDE—receive code error counter register RCSE—receive carrier sense error counter register RUND—receive undersize packet counter register ROVR—receive oversize packet counter register RFRG—receive fragments counter register RJBR—receive jabber counter register RDRP—receive drop register TSEC1 Transmit Counters 0x2_46E0 0x2_46E4 0x2_46E8 0x2_46EC 0x2_46F0 0x2_46F4 0x2_46F8 0x2_46FC 0x2_4700 0x2_4704 0x2_4708 0x2_470C 0x2_4710 0x2_4714 0x2_4718 0x2_471C 0x2_4720 0x2_4724 0x2_4728 0x2_472C TBYT—Transmit byte counter register TPKT—Transmit packet counter register TMCA—Transmit multicast packet counter register TBCA—Transmit broadcast packet counter register TXPF—Transmit PAUSE control frame counter register TDFR—Transmit deferral packet counter register TEDF—Transmit excessive deferral packet counter register TSCL—Transmit single collision packet counter register TMCL—Transmit multiple collision packet counter register TLCL—Transmit late collision packet counter register TXCL—Transmit excessive collision packet counter register TNCL—Transmit total collision counter register Reserved TDRP—Transmit drop frame counter register TJBR—Transmit jabber frame counter register TFCS—Transmit FCS error counter register TXCF—Transmit control frame counter register TOVR—Transmit oversize frame counter register TUND—Transmit undersize frame counter register TFRG—Transmit fragments frame counter register TSEC1 General Registers 0x2_4730 CAR1—Carry register one register R/W 0x0000_0000 14.5.3.7.44/14-82 R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R R/W R/W R/W R/W R/W R/W R/W 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 14.5.3.7.25/14-73 14.5.3.7.26/14-73 14.5.3.7.27/14-74 14.5.3.7.28/14-74 14.5.3.7.29/14-75 14.5.3.7.30/14-75 14.5.3.7.31/14-76 14.5.3.7.32/14-76 14.5.3.7.33/14-77 14.5.3.7.34/14-77 14.5.3.7.35/14-78 14.5.3.7.36/14-78 — 14.5.3.7.37/14-79 14.5.3.7.38/14-79 14.5.3.7.39/14-80 14.5.3.7.40/14-80 14.5.3.7.41/14-81 14.5.3.7.42/14-81 14.5.3.7.43/14-82 Access R/W R/W R/W R/W R/W R/W R/W R/W R/W Reset 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 Section/Page 14.5.3.7.16/14-68 14.5.3.7.17/14-69 14.5.3.7.18/14-69 14.5.3.7.19/14-70 14.5.3.7.20/14-70 14.5.3.7.21/14-71 14.5.3.7.22/14-71 14.5.3.7.23/14-72 14.5.3.7.24/14-72 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 2-31 Memory Map Table 2-9. Memory Map (continued) Offset 0x2_4734 0x2_4738 0x2_473C 0x2_4740– 0x2_47FC Register CAR2—Carry register two register CAM1—Carry register one mask register CAM2—Carry register two mask register Reserved Access R/W R/W R/W R TSEC1 Hash Function Registers 0x2_4800 0x2_4804 0x2_4808 0x2_480C 0x2_4810 0x2_4814 0x2_4818 0x2_481C 0x2_4820– 0x2_487C 0x2_4880 0x2_4884 0x2_4888 0x2_488C 0x2_4890 0x2_4894 0x2_4898 0x2_489C 0x2_48A0– 0x2_4BF4 IADDR0—Individual address register 0 IADDR1—Individual address register 1 IADDR2—Individual address register 2 IADDR3—Individual address register 3 IADDR4—Individual address register 4 IADDR5—Individual address register 5 IADDR6—Individual address register 6 IADDR7—Individual address register 7 Reserved GADDR0—Group address register 0 GADDR1—Group address register 1 GADDR2—Group address register 2 GADDR3—Group address register 3 GADDR4—Group address register 4 GADDR5—Group address register 5 GADDR6—Group address register 6 GADDR7—Group address register 7 Reserved TSEC1 Attribute Registers 0x2_4BF8 0x2_4BFC 0x2_48A0– 0x2_4AFC ATTR—Attribute register ATTRELI—Attribute EL & EI register Reserved R/W R/W R TSEC1 Future Expansion Space 0x2_4C00– 0x2_4FFC 0x2_5000– 0x2_5FFC Reserved TSEC2 registers2 FEC General Control and Status Registers R 0x0000_0000 — 0x0000_0000 0x0000_0000 0x0000_0000 14.5.3.9.1/14-89 14.5.3.9.2/14-90 — R/W R/W R/W R/W R/W R/W R/W R/W R R/W R/W R/W R/W R/W R/W R/W R/W R 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 — — 14.5.3.8.2/14-88 14.5.3.8.1/14-88 Reset 0x0000_0000 0xFE01_FFFF 0x000F_FFFF 0x0000_0000 Section/Page 14.5.3.7.45/14-84 14.5.3.7.46/14-85 14.5.3.7.47/14-86 — MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 2-32 Freescale Semiconductor Complete CCSR Map Table 2-9. Memory Map (continued) Offset 0x2_6000– 0x2_600C 0x2_6010 0x2_6014 0x2_6018 0x2_601C– 0x2_6020 0x2_6024 0x2_6028 0x2_602C 0x2_6030– 0x2_6088 Reserved IEVENT—Interrupt event register IMASK—Interrupt mask register EDIS—Error disabled register Reserved MINFLR—Minimum frame length register PTV—Pause time value register DMACTRL—DMA control register Reserved Register Access R R R/W R/W R R/W R/W R/W R FEC FIFO Control and Status Registers 0x2_604C 0x2_6050– 0x2_6088 0x2_608C 0x2_6090– 0x2_6094 0x2_6098 0x2_609C 0x2_60A0– 0x2_60FC FIFO_PAUSE_CTRL—FIFO pause control register Reserved FIFO_TX_THR—FIFO transmit threshold register Reserved FIFO_TX_STARVE—FIFO transmit starve register FIFO_TX_STARVE_SHUTOFF—FIFO transmit starve shutoff register Reserved R/W R R/W R R/W R/W R FEC Transmit Control and Status Registers 0x2_6100 0x2_6104 0x2_6108 0x2_610C 0x2_6110– 0x2_6120 0x2_6124 0x2_6128– 0x2_6180 0x2_6184 0x2_6188– 0x2_6200 0x2_6204 TCTRL—Transmit control register TSTAT—Transmit status register Reserved TBDLEN—TxBD data length Reserved CTBPTR—Current TxBD pointer register Reserved TBPTR—TxBD pointer register Reserved TBASE—TxBD base address register R/W R/W R R R R R R/W R R/W 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 21.5.3.3.1/21-25 21.5.3.3.2/21-26 — 21.5.3.3.3/21-27 — 21.5.3.3.4/21-27 — 21.5.3.3.5/21-28 — 21.5.3.3.6/21-28 0x0000_0000 0x0000_0000 0x0000_0080 0x0000_0000 0x0000_0020 0x0000_0080 0x0000_0000 21.5.3.2.1/21-22 — 21.5.3.2.2/21-23 — 21.5.3.3.4/21-27 21.5.3.3.5/21-28 — Reset 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0040 0x0000_0000 0x0000_0000 0x0000_0000 Section/Page — 21.5.3.1.1/21-13 21.5.3.1.2/21-16 21.5.3.1.3/21-18 — 21.5.3.1.4/21-18 21.5.3.1.5/21-19 21.5.3.1.6/21-20 — MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 2-33 Memory Map Table 2-9. Memory Map (continued) Offset 0x2_6208– 0x2_62AC 0x2_62B0 0x2_62B4 0x2_62B8– 0x2_62FC Reserved OSTBD—Out-of-sequence TxBD register OSTBDP—Out-of-sequence TxBD pointer register Reserved Register Access R R/W R/W R FEC Receive Control and Status Registers 0x2_6300 0x2_6304 0x2_6308 0x2_630C 0x2_6310– 0x2_6320 0x2_6324 0x2_6328– 0x2_633C 0x2_6340 0x2_6344– 0x2_6380 0x2_6384 0x2_6388– 0x2_6400 0x2_6404 0x2_6408– 0x2_64FC RCTRL—Receive control register RSTAT—Receive status register Reserved RBDLEN—RxBD data length register Reserved CRBPTR—Current RxBD pointer register Reserved MRBLR—Maximum receive buffer length register Reserved RBPTR—RxBD pointer register Reserved RBASE—RxBD base address register Reserved FEC MAC Registers 0x2_6500 0x2_6504 0x2_6508 0x2_650C 0x2_6510 0x2_6514– 0x2_6538 0x2_653C 0x2_6540 0x2_6544 0x2_6548– 0x2_67FC MACCFG1—MAC configuration register #1 MACCFG2—MAC configuration register #2 IPGIFG—Inter-packet gap/inter-frame gap register HAFDUP—Half-duplex register MAXFRM—Maximum frame length register Reserved IFSTAT—Interface status register MACSTNADDR1—Station address register, part 1 MACSTNADDR2—Station address register, part 2 Reserved R/W R/W R/W R/W R/W R R R/W R/W R 0x0000_0000 0x0000_7000 0x4060_5060 0x00A0_F037 0x0000_0600 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 21.5.3.6.1/21-38 21.5.3.6.2/21-40 21.5.3.6.3/21-41 21.5.3.6.4/21-41 21.5.3.6.5/21-42 — 21.5.3.6.6/21-43 21.5.3.6.7/21-43 21.5.3.6.8/21-44 — R/W R/W R R R R R R/W R R/W R R/W R 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 21.5.3.4.1/21-31 21.5.3.4.2/21-32 — 21.5.3.4.3/21-33 — 21.5.3.4.4/21-33 — 21.5.3.4.5/21-34 — 21.5.3.4.6/21-35 — 21.5.3.4.7/21-35 — Reset 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 Section/Page — 21.5.3.3.7/21-29 21.5.3.3.8/21-31 — MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 2-34 Freescale Semiconductor Complete CCSR Map Table 2-9. Memory Map (continued) Offset Register Access Reset Section/Page FEC Hash Function Registers 0x2_6800 0x2_6804 0x2_6808 0x2_680C 0x2_6810 0x2_6814 0x2_6818 0x2_681C 0x2_6820– 0x2_687C 0x2_6880 0x2_6884 0x2_6888 0x2_688C 0x2_6890 0x2_6894 0x2_6898 0x2_689C 0x2_68A0– 0x2_6BF4 IADDR0—Individual address register 0 IADDR1—Individual address register 1 IADDR2—Individual address register 2 IADDR3—Individual address register 3 IADDR4—Individual address register 4 IADDR5—Individual address register 5 IADDR6—Individual address register 6 IADDR7—Individual address register 7 Reserved GADDR0—Group address register 0 GADDR1—Group address register 1 GADDR2—Group address register 2 GADDR3—Group address register 3 GADDR4—Group address register 4 GADDR5—Group address register 5 GADDR6—Group address register 6 GADDR7—Group address register 7 Reserved FEC Attribute Registers 0x2_6BF8 0x2_6BFC ATTR—Attribute register ATTRELI—Attribute EL & EI register R/W R/W 0x0000_0000 0x0000_0000 21.5.3.8.1/21-46 21.5.3.8.2/21-47 R/W R/W R/W R/W R/W R/W R/W R/W R R/W R/W R/W R/W R/W R/W R/W R/W R 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 — — 21.5.3.7.2/21-46 21.5.3.7.1/21-45 FEC Future Expansion Space 0x2_6C00– 0x2_6FFC Reserved R PIC Register Address Map—Global Registers 0x4_0000– 0x4_0030 0x4_0040 0x4_0050 0x4_0060 0x4_0070 0x4_0080 Reserved IPIDR0—Interprocessor interrupt 0 (IPI 0) dispatch register IPIDR1—IPI 1 dispatch register IPIDR2—IPI 2 dispatch register IPIDR3—IPI 3 dispatch register CTPR—Current task priority register R/W 0x0000_000F 10.3.7.2/10-40 — W — 0x0000_0000 — 10.3.7.1/10-39 0x0000_0000 — MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 2-35 Memory Map Table 2-9. Memory Map (continued) Offset 0x4_0090 0x4_00A0 0x4_00B0 0x4_00C0– 0x4_0FF0 0x4_1000 0x4_1010 0x4_1020 0x4_1030 0x4_1040– 0x4_1070 0x4_1080 0x4_1090 0x4_10A0 0x4_10B0 0x4_10C0 0x4_10D0 0x4_10E0 0x4_10F0 0x4_1100 0x4_1110 0x4_1120 0x4_1130 0x4_1140 0x4_1150 0x4_1160 0x4_1170 0x4_1180 0x4_1190 0x4_11A0 0x4_11B0 0x4_11C0 0x4_11D0 0x4_11E0 0x4_11F0 Register WHOAMI—Who am I register IACK—Interrupt acknowledge register EOI—End of interrupt register Reserved FRR—Feature reporting register Reserved GCR—Global configuration register Reserved Vendor reserved VIR—Vendor identification register PIR—Processor initialization register IPIVPR0—IPI 0 vector/priority register IPIVPR1—IPI 1 vector/priority register IPIVPR2—IPI 2 vector/priority register IPIVPR3—IPI 3 vector/priority register SVR—Spurious vector register TFRR—Timer frequency reporting register GTCCR0—Global timer 0 current count register GTBCR0—Global timer 0 base count register GTVPR0—Global timer 0 vector/priority register GTDR0—Global timer 0 destination register GTCCR1—Global timer 1 current count register GTBCR1—Global timer 1 base count register GTVPR1—Global timer 1 vector/priority register GTDR1—Global timer 1 destination register GTCCR2—Global timer 2 current count register GTBCR2—Global timer 2 base count register GTVPR2—Global timer 2 vector/priority register GTDR2—Global timer 2 destination register GTCCR3—Global timer 3 current count register GTBCR3—Global timer 3 base count register GTVPR3—Global timer 3 vector/priority register GTDR3—Global timer 3 destination register R/W R/W R R/W R/W R/W R R/W R/W R/W R R/W R/W R/W R R/W R/W R/W 0x0000_FFFF 0x0000_0000 0x0000_0000 0x8000_0000 0x8000_0000 0x0000_0001 0x0000_0000 0x8000_0000 0x8000_0000 0x0000_0001 0x0000_0000 0x8000_0000 0x8000_0000 0x0000_0001 0x0000_0000 0x8000_0000 0x8000_0000 0x0000_0001 10.3.1.6/10-19 10.3.2.1/10-20 10.3.2.2/10-21 10.3.2.3/10-21 10.3.2.4/10-22 10.3.2.5/10-23 10.3.2.2/10-21 10.3.2.3/10-21 10.3.2.4/10-22 10.3.2.5/10-23 10.3.2.2/10-21 10.3.2.3/10-21 10.3.2.4/10-22 10.3.2.5/10-23 10.3.2.2/10-21 10.3.2.3/10-21 10.3.2.4/10-22 10.3.2.5/10-23 Access R R W — R — R/W — — R R/W R/W Reset 0x0000_0000 0x0000_0000 0x0000_0000 — 0x0037_0002 — 0x0000_0000 — — 0x0000_0000 0x0000_0000 0x8000_0000 Section/Page 10.3.7.3/10-41 10.3.7.4/10-41 10.3.7.5/10-42 — 10.3.1.1/10-16 — 10.3.1.2/10-17 — — 10.3.1.3/10-17 10.3.1.4/10-18 10.3.1.5/10-19 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 2-36 Freescale Semiconductor Complete CCSR Map Table 2-9. Memory Map (continued) Offset 0x4_1200– 0x4_12F0 0x4_1300 0x4_1310 0x4_1320 0x4_1330 0x4_1340 0x4_1350 0x4_1360 0x4_1370 0x4_1380 0x4_1390 0x4_13A0 0x4_13B0 0x4_13C0 0x4_13D0– 0x4_13F0 0x4_1400 0x4_1410 0x4_1420 0x4_1430 0x4_1440– 0x4_14F0 0x4_1500 0x4_1510 0x4_1520– 0x4_FFF0 Reserved TCR—Timer control register IRQSR0—IRQ_OUT summary register 0 IRQSR1—IRQ_OUT summary register 1 CISR0—Critical Interrupt summary register 0 CISR1—Critical Interrupt summary register 1 PM0MR0—Performance monitor 0 mask register 0 PM0MR1—Performance monitor 0 mask register 1 PM1MR0—Performance monitor 1 mask register 0 PM1MR1—Performance monitor 1 mask register 1 PM2MR0—Performance monitor 2 mask register 0 PM2MR1—Performance monitor 2 mask register 1 PM3MR0—Performance monitor 3 mask register 0 PM3MR1—Performance monitor 3 mask register 1 Reserved MSGR0—Message register 0 MSGR1—Message register 1 MSGR2—Message register 2 MSGR3—Message register 3 Reserved MER—Message enable register MSR—Message status register Reserved — R/W R/W — — 0x0000_0000 0x0000_0000 — — 10.3.5.2/10-31 10.3.5.3/10-31 — Register Access — R/W R R R R R/W R/W R/W R/W R/W R/W R/W R/W — R/W Reset — 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x00FF_FFFF 0xFFFF_FFFF 0x00FF_FFFF 0xFFFF_FFFF 0x00FF_FFFF 0xFFFF_FFFF 0x00FF_FFFF 0xFFFF_FFFF — 0x0000_0000 Section/Page — 10.3.2.6/10-24 10.3.3.1/10-26 10.3.3.2/10-27 10.3.3.3/10-27 10.3.3.4/10-28 10.3.4.1/10-29 10.3.4.2/10-30 10.3.4.1/10-29 10.3.4.2/10-30 10.3.4.1/10-29 10.3.4.2/10-30 10.3.4.1/10-29 10.3.4.2/10-30 — 10.3.5.1/10-30 PIC Register Address Map—Interrupt Source Configuration Registers 0x5_0000 0x5_0010 0x5_0020 0x5_0030 0x5_0040 EIVPR0—External interrupt 0 (IRQ0) vector/priority register EIDR0—External interrupt 0 (IRQ0) destination register EIVPR1—External interrupt 1 (IRQ1) vector/priority register EIDR1—External interrupt 1 (IRQ1) destination register EIVPR2—External interrupt 2 (IRQ2) vector/priority register R/W R/W R/W R/W R/W 0x8000_0000 0x0000_0001 0x8000_0000 0x0000_0001 0x8000_0000 10.3.6.1/10-32 10.3.6.2/10-33 10.3.6.1/10-32 10.3.6.2/10-33 10.3.6.1/10-32 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 2-37 Memory Map Table 2-9. Memory Map (continued) Offset 0x5_0050 0x5_0060 0x5_0070 0x5_0080 0x5_0090 0x5_00A0 0x5_00B0 0x5_00C0 0x5_00D0 0x5_00E0 0x5_00F0 0x5_0100 0x5_0110 0x5_0120 0x5_0130 0x5_0140 0x5_0150 0x5_0160 0x5_0170 0x5_0180– 0x5_01F0 0x5_0200 0x5_0210 Register EIDR2—External interrupt 2 (IRQ2) destination register EIVPR3—External interrupt 3 (IRQ3) vector/priority register EIDR3—External interrupt 3 (IRQ3) destination register EIVPR4—External interrupt 4 (IRQ4) vector/priority register EIDR4—External interrupt 4 (IRQ4) destination register EIVPR5—External interrupt 5 (IRQ5) vector/priority register EIDR5—External interrupt 5 (IRQ5) destination register EIVPR6—External interrupt 6 (IRQ6) vector/priority register EIDR6—External interrupt 6 (IRQ6) destination register EIVPR7—External interrupt 7 (IRQ7) vector/priority register EIDR7—External interrupt 7 (IRQ7) destination register EIVPR8—External interrupt 8 (IRQ8) vector/priority register EIDR8—External interrupt 8 (IRQ8) destination register EIVPR9—External interrupt 9 (IRQ9) vector/priority register EIDR9—External interrupt 9 (IRQ9) destination register EIVPR10—External interrupt 10 (IRQ10) vector/priority register EIDR10—External interrupt 10 (IRQ10) destination register EIVPR11—External interrupt 11 (IRQ11) vector/priority register EIDR11—External interrupt 11 (IRQ11) destination register Reserved IIVPR0—Internal interrupt 0 vector/priority register IIDR0—Internal interrupt 0 destination register Access R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W — R/W R/W Reset 0x0000_0001 0x8000_0000 0x0000_0001 0x8000_0000 0x0000_0001 0x8000_0000 0x0000_0001 0x8000_0000 0x0000_0001 0x8000_0000 0x0000_0001 0x8000_0000 0x0000_0001 0x8000_0000 0x0000_0001 0x8000_0000 0x0000_0001 0x8000_0000 0x0000_0001 — 0x8080_0000 0x0000_0001 Section/Page 10.3.6.2/10-33 10.3.6.1/10-32 10.3.6.2/10-33 10.3.6.1/10-32 10.3.6.2/10-33 10.3.6.1/10-32 10.3.6.2/10-33 10.3.6.1/10-32 10.3.6.2/10-33 10.3.6.1/10-32 10.3.6.2/10-33 10.3.6.1/10-32 10.3.6.2/10-33 10.3.6.1/10-32 10.3.6.2/10-33 10.3.6.1/10-32 10.3.6.2/10-33 10.3.6.1/10-32 10.3.6.2/10-33 — 10.3.6.3/10-34 10.3.6.4/10-35 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 2-38 Freescale Semiconductor Complete CCSR Map Table 2-9. Memory Map (continued) Offset 0x5_0220 0x5_0230 0x5_0240 0x5_0250 0x5_0260 0x5_0270 0x5_0280 0x5_0290 0x5_02A0 0x5_02B0 0x5_02C0 0x5_02D0 0x5_02E0 0x5_02F0 0x5_0300 0x5_0310 0x5_0320 0x5_0330 0x5_0340 0x5_0350 0x5_0360 0x5_0370 0x5_0380 0x5_0390 0x5_03A0 0x5_03B0 0x5_03C0 0x5_03D0 0x5_03E0 0x5_03F0 0x5_0400 0x5_0410 0x5_0420 0x5_0430 0x5_0440 Register IIVPR1—Internal interrupt 1 vector/priority register IIDR1—Internal interrupt 1 destination register IIVPR2—Internal interrupt 2 vector/priority register IIDR2—Internal interrupt 2 destination register IIVPR3—Internal interrupt 3 vector/priority register IIDR3—Internal interrupt 3 destination register IIVPR4—Internal interrupt 4 vector/priority register IIDR4—Internal interrupt 4 destination register IIVPR5—Internal interrupt 5 vector/priority register IIDR5—Internal interrupt 5 destination register IIVPR6—Internal interrupt 6 vector/priority register IIDR6—Internal interrupt 6 destination register IIVPR7—Internal interrupt 7 vector/priority register IIDR7—Internal interrupt 7 destination register IIVPR8—Internal interrupt 8 vector/priority register IIDR8—Internal interrupt 8 destination register IIVPR9—Internal interrupt 9 vector/priority register IIDR9—Internal interrupt 9 destination register IIVPR10—Internal interrupt 10 vector/priority register IIDR10—Internal interrupt 10 destination register IIVPR11—Internal interrupt 11 vector/priority register IIDR11—Internal interrupt 11 destination register IIVPR12—Internal interrupt 12 vector/priority register IIDR12—Internal interrupt 12 destination register IIVPR13—Internal interrupt 13 vector/priority register IIDR13—Internal interrupt 13 destination register IIVPR14—Internal interrupt 14 vector/priority register IIDR14—Internal interrupt 14 destination register IIVPR15—Internal interrupt 15 vector/priority register IIDR15—Internal interrupt 15 destination register IIVPR16—Internal interrupt 16 vector/priority register IIDR16—Internal interrupt 16 destination register IIVPR17—Internal interrupt 17 vector/priority register IIDR17—Internal interrupt 17 destination register IIVPR18—Internal interrupt 18 vector/priority register Access R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W Reset 0x8080_0000 0x0000_0001 0x8080_0000 0x0000_0001 0x8080_0000 0x0000_0001 0x8080_0000 0x0000_0001 0x8080_0000 0x0000_0001 0x8080_0000 0x0000_0001 0x8080_0000 0x0000_0001 0x8080_0000 0x0000_0001 0x8080_0000 0x0000_0001 0x8080_0000 0x0000_0001 0x8080_0000 0x0000_0001 0x8080_0000 0x0000_0001 0x8080_0000 0x0000_0001 0x8080_0000 0x0000_0001 0x8080_0000 0x0000_0001 0x8080_0000 0x0000_0001 0x8080_0000 0x0000_0001 0x8080_0000 Section/Page 10.3.6.3/10-34 10.3.6.4/10-35 10.3.6.3/10-34 10.3.6.4/10-35 10.3.6.3/10-34 10.3.6.4/10-35 10.3.6.3/10-34 10.3.6.4/10-35 10.3.6.3/10-34 10.3.6.4/10-35 10.3.6.3/10-34 10.3.6.4/10-35 10.3.6.3/10-34 10.3.6.4/10-35 10.3.6.3/10-34 10.3.6.4/10-35 10.3.6.3/10-34 10.3.6.4/10-35 10.3.6.3/10-34 10.3.6.4/10-35 10.3.6.3/10-34 10.3.6.4/10-35 10.3.6.3/10-34 10.3.6.4/10-35 10.3.6.3/10-34 10.3.6.4/10-35 10.3.6.3/10-34 10.3.6.4/10-35 10.3.6.3/10-34 10.3.6.4/10-35 10.3.6.3/10-34 10.3.6.4/10-35 10.3.6.3/10-34 10.3.6.4/10-35 10.3.6.3/10-34 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 2-39 Memory Map Table 2-9. Memory Map (continued) Offset 0x5_0450 0x5_0460 0x5_0470 0x5_0480 0x5_0490 0x5_04A0 0x5_04B0 0x5_04C0 0x5_04D0 0x5_04E0 0x5_04F0 0x5_0500 0x5_0510 0x5_0520 0x5_0530 0x5_0540 0x5_0550 0x5_0560 0x5_0570 0x5_0580 0x5_0590 0x5_05A0 0x5_05B0 0x5_05C0 0x5_05D0 0x5_05E0 0x5_05F0 0x5_0600– 0x5_15F0 0x5_1600 0x5_1610 0x5_1620 0x5_1630 Register IIDR18—Internal interrupt 18 destination register IIVPR19—Internal interrupt 19 vector/priority register IIDR19—Internal interrupt 19 destination register IIVPR20—Internal interrupt 20 vector/priority register IIDR20—Internal interrupt 20 destination register IIVPR21—Internal interrupt 21 vector/priority register IIDR21—Internal interrupt 21 destination register IIVPR22—Internal interrupt 22 vector/priority register IIDR22—Internal interrupt 22 destination register IIVPR23—Internal interrupt 23 vector/priority register IIDR23—Internal interrupt 23 destination register IIVPR24—Internal interrupt 24 vector/priority register IIDR24—Internal interrupt 24 destination register IIVPR25—Internal interrupt 25 vector/priority register IIDR25—Internal interrupt 25 destination register IIVPR26—Internal interrupt 26 vector/priority register IIDR126—Internal interrupt 26 destination register IIVPR27—Internal interrupt 27 vector/priority register IIDR27—Internal interrupt 27 destination register IIVPR28—Internal interrupt 28 vector/priority register IIDR28—Internal interrupt 28 destination register IIVPR29—Internal interrupt 29 vector/priority register IIDR29—Internal interrupt 29 destination register IIVPR30—Internal interrupt 30 vector/priority register IIDR30—Internal interrupt 30 destination register IIVPR31—Internal interrupt 31 vector/priority register IIDR31—Internal interrupt 31 destination register Reserved MIVPR0—Messaging interrupt 0 (MSG 0) vector/priority register MIDR0—Messaging interrupt 0 (MSG 0) destination register MIVPR1—Messaging interrupt 1 (MSG 1) vector/priority register MIDR1—Messaging interrupt 1 (MSG 1) destination register Access R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W — R/W R/W R/W R/W Reset 0x0000_0001 0x8080_0000 0x0000_0001 0x8080_0000 0x0000_0001 0x8080_0000 0x0000_0001 0x8080_0000 0x0000_0001 0x8080_0000 0x0000_0001 0x8080_0000 0x0000_0001 0x8080_0000 0x0000_0001 0x8080_0000 0x0000_0001 0x8080_0000 0x0000_0001 0x8080_0000 0x0000_0001 0x8080_0000 0x0000_0001 0x8080_0000 0x0000_0001 0x8080_0000 0x0000_0001 — 0x8000_0000 0x0000_0001 0x8000_0000 0x0000_0001 Section/Page 10.3.6.4/10-35 10.3.6.3/10-34 10.3.6.4/10-35 10.3.6.3/10-34 10.3.6.4/10-35 10.3.6.3/10-34 10.3.6.4/10-35 10.3.6.3/10-34 10.3.6.4/10-35 10.3.6.3/10-34 10.3.6.4/10-35 10.3.6.3/10-34 10.3.6.4/10-35 10.3.6.3/10-34 10.3.6.4/10-35 10.3.6.3/10-34 10.3.6.4/10-35 10.3.6.3/10-34 10.3.6.4/10-35 10.3.6.3/10-34 10.3.6.4/10-35 10.3.6.3/10-34 10.3.6.4/10-35 10.3.6.3/10-34 10.3.6.4/10-35 10.3.6.3/10-34 10.3.6.4/10-35 — 10.3.6.5/10-36 10.3.6.6/10-37 10.3.6.5/10-36 10.3.6.6/10-37 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 2-40 Freescale Semiconductor Complete CCSR Map Table 2-9. Memory Map (continued) Offset 0x5_1640 0x5_1650 0x5_1660 0x5_1670 0x5_1680– 0x5_FFF0 Register MIVPR2—Messaging interrupt 2 (MSG 2) vector/priority register MIDR2—Messaging interrupt 2 (MSG 2) destination register MIVPR3—Messaging interrupt 3 (MSG 3) vector/priority register MIDR3—Messaging interrupt 3 (MSG 3) destination register Reserved Access R/W R/W R/W R/W — PIC Register Address Map—Per-CPU Registers 0x6_0000– 0x6_0030 0x6_0040 0x6_0050 0x6_0060 0x6_0070 0x6_0080 0x6_0090 0x6_00A0 0x6_00B0 Reserved IPIDR0—P0 IPI 0 dispatch register IPIDR1—P0 IPI 1 dispatch register IPIDR2—P0 IPI 2 dispatch register IPIDR3—P0 IPI 3 dispatch register CTPR0—P0 current task priority register WHOAMI0—P0 who am I register IACK0—P0 interrupt acknowledge register EOI0—P0 end of interrupt register R/W R R W 0x0000_000F All zeros All zeros All zeros 10.3.7.2/10-40 10.3.7.3/10-41 10.3.7.4/10-41 10.3.7.5/10-42 — W — All zeros — 10.3.7.1/10-39 Reset 0x8000_0000 0x0000_0001 0x8000_0000 0x0000_0001 — Section/Page 10.3.6.5/10-36 10.3.6.6/10-37 10.3.6.5/10-36 10.3.6.6/10-37 — MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 2-41 Memory Map Table 2-9. Memory Map (continued) Offset Register RapidIO Registers RapidIO Architectural Registers 0xC_0000 0xC_0004 0xC_0008 0xC_000C 0xC_0010 0xC_0014 0xC_0018 0xC_001C 0xC_0040 0xC_0044 0xC_004C 0xC_0058 0xC_005C 0xC_0060 0xC_0068 0xC_006C 0xC_0100 0xC_0120 0xC_0124 0xC_013C 0xC_0140 0xC_0144 0xC_0148 DIDCAR—Device identity capability register DICAR—Device information capability register AIDCAR—Assembly identity capability register AICAR—Assembly information capability register PEFCAR—Processing element features capability register SPICAR—Switch port information capability register SOCAR—Source operations capability register DOCAR—Destination operations capability register MSR—Mailbox command and status register PWDCSR—Port-write and doorbell command and status register PELLCCSR—Processing element logical layer control command and status register Reserved LCSBA1CSR—Local configuration space base address 1 command and status register BDIDCSR—Base device ID command and status register HBDIDLCSR—Host base device ID lock command and status register CTCSR—Component tag command and status register PMBH0CSR—8/16 LP-LVDS port maintenance block header 0 command and status register PLTOCCSR—Port link time-out control command and status register PRTOCCSR—Port response time-out control command and status register PGCCSR—Port general control command and status register PLMREQCSR—Port link maintenance request command and status register PLMRESPCSR—Port link maintenance response command and status register PLASCSR—Port local ackID status command and status register R R R/W R/W R R R R R R R R R/W R/W Special R/W R R/W R/W R/W R/W R R/W 0x0002_0002 0x8030_0020 0x0000_0000 0x0000_0100 0xE088_0009 0x0000_0100 0x0600_FCF0 0x0600_FCF4 0x0000_0000 0x0000_0020 0x0000_0001 0x0000_0000 0x0000_0000 0x00[cfg]_0000 0x0000_FFFF 0x0000_0000 0x0000_0002 0xFFFF_FF00 0xFFFF_FF00 0x[cfg]000_0000 0x0000_0000 0x0000_0000 0x0000_0000 17.3.1.1/17-13 17.3.1.2/17-14 17.3.1.3/17-14 17.3.1.4/17-15 17.3.1.5/17-15 17.3.1.6/17-17 17.3.1.7/17-17 17.3.1.8/17-19 17.3.1.9/17-22 17.3.1.10/17-22 17.3.1.11/17-24 — 17.3.1.12/17-24 17.3.1.13/17-25 17.3.1.14/17-26 17.3.1.15/17-26 17.3.1.16/17-27 17.3.1.17/17-27 17.3.1.18/17-28 17.3.1.19/17-29 17.3.1.20/17-29 17.3.1.21/17-30 17.3.1.22/17-31 Access Reset Section/Page MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 2-42 Freescale Semiconductor Complete CCSR Map Table 2-9. Memory Map (continued) Offset 0xC_0158 0xC_015C Register PESCSR—Port error and status command and status register PCCSR—Port control command and status register Implementation Registers 0xD_0000 0xD_0010 0xD_0014 CR—Configuration register PCR—Port configuration register PEIR—Port error injection register ATMU Registers 0xD_0C00 0xD_0C10 0xD_0C20 0xD_0C28 0xD_0C30 0xD_0C40 0xD_0C48 0xD_0C50 0xD_0C60 0xD_0C68 0xD_0C70 0xD_0C80 0xD_0C88 0xD_0C90 0xD_0CA0 0xD_0CA8 ROWTAR0—RapidIO outbound window translation address register 0 ROWAR0—RapidIO outbound window attributes register 0 ROWTAR1—RapidIO outbound window translation address register 1 ROWBAR1—RapidIO outbound window base address register 1 ROWAR1—RapidIO outbound window attributes register 1 ROWTAR2—RapidIO outbound window translation address register 2 ROWBAR2—RapidIO outbound window base address register 2 ROWAR2—RapidIO outbound window attributes register 2 ROWTAR3—RapidIO outbound window translation address register 3 ROWBAR3—RapidIO outbound window base address register 3 ROWAR3—RapidIO outbound window attributes register 3 ROWTAR4—RapidIO outbound window translation address register 4 ROWBAR4—RapidIO outbound window base address register 4 ROWAR4—RapidIO outbound window attributes register 4 ROWTAR5—RapidIO outbound window translation address register 5 ROWBAR5—RapidIO outbound window base address register 5 R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W 0x0000_0000 0x8004_401F 0x0000_0000 0x0000_0000 0x0004_401F 0x0000_0000 0x0000_0000 0x0004_401F 0x0000_0000 0x0000_0000 0x0004_401F 0x0000_0000 0x0000_0000 0x0004_401F 0x0000_0000 0x0000_0000 17.3.2.2.1/17-37 17.3.2.2.3/17-39 17.3.2.2.1/17-37 17.3.2.2.2/17-38 17.3.2.2.3/17-39 17.3.2.2.1/17-37 17.3.2.2.2/17-38 17.3.2.2.3/17-39 17.3.2.2.1/17-37 17.3.2.2.2/17-38 17.3.2.2.3/17-39 17.3.2.2.1/17-37 17.3.2.2.2/17-38 17.3.2.2.3/17-39 17.3.2.2.1/17-37 17.3.2.2.2/17-38 R/W R/W R/W 0x0000_0001 0x0000_0010 0x0000_0000 17.3.2.1.1/17-34 17.3.2.1.2/17-35 17.3.2.1.3/17-35 Access R/W R/W Reset 0x0000_0000 0x4400_0000 Section/Page 17.3.1.23/17-31 17.3.1.24/17-33 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 2-43 Memory Map Table 2-9. Memory Map (continued) Offset 0xD_0CB0 0xD_0CC0 0xD_0CC8 0xD_0CD0 0xD_0CE0 0xD_0CE8 0xD_0CF0 0xD_0D00 0xD_0D08 0xD_0D10 0xD_0D60 0xD_0D68 0xD_0D70 0xD_0D80 0xD_0D88 0xD_0D90 0xD_0DA0 0xD_0DA8 0xD_0DB0 0xD_0DC0 0xD_0DC8 Register ROWAR5—RapidIO outbound window attributes register 5 ROWTAR6—RapidIO outbound window translation address register 6 ROWBAR6—RapidIO outbound window base address register 6 ROWAR6—RapidIO outbound window attributes register 6 ROWTAR7—RapidIO outbound window translation address register 7 ROWBAR7—RapidIO outbound window base address register 7 ROWAR7—RapidIO outbound window attributes register 7 ROWTAR8—RapidIO outbound window translation address register 8 ROWBAR8—RapidIO outbound window base address register 8 ROWAR8—RapidIO outbound window attributes register 8 RIWTAR4—RapidIO inbound window translation address register 4 RIWBAR4—RapidIO inbound window base address register 4 RIWAR4—RapidIO inbound window attributes register 4 RIWTAR3—RapidIO inbound window translation address register 3 RIWBAR3—RapidIO inbound window base address register 3 RIWAR3—RapidIO inbound window attributes register 3 RIWTAR2—RapidIO inbound window translation address register 2 RIWBAR2—RapidIO inbound window base address register 2 RIWAR2—RapidIO inbound window attributes register 2 RIWTAR1—RapidIO inbound window translation address register 1 RIWBAR1—RapidIO inbound window base address register 1 Access R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W Reset 0x0004_401F 0x0000_0000 0x0000_0000 0x0004_401F 0x0000_0000 0x0000_0000 0x0004_401F 0x0000_0000 0x0000_0000 0x0004_401F 0x0000_0000 0x0000_0000 0x0004_401F 0x0000_0000 0x0000_0000 0x0004_401F 0x0000_0000 0x0000_0000 0x0004_401F 0x0000_0000 0x0000_0000 Section/Page 17.3.2.2.3/17-39 17.3.2.2.1/17-37 17.3.2.2.2/17-38 17.3.2.2.3/17-39 17.3.2.2.1/17-37 17.3.2.2.2/17-38 17.3.2.2.3/17-39 17.3.2.2.1/17-37 17.3.2.2.2/17-38 17.3.2.2.3/17-39 17.3.2.2.4/17-40 17.3.2.2.5/17-41 17.3.2.2.6/17-42 17.3.2.2.4/17-40 17.3.2.2.5/17-41 17.3.2.2.6/17-42 17.3.2.2.4/17-40 17.3.2.2.5/17-41 17.3.2.2.6/17-42 17.3.2.2.4/17-40 17.3.2.2.5/17-41 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 2-44 Freescale Semiconductor Complete CCSR Map Table 2-9. Memory Map (continued) Offset 0xD_0DD0 0xD_0DE0 0xD_0DF0 Register RIWAR1—RapidIO inbound window attributes register 1 RIWTAR0—RapidIO inbound window translation address register 0 RIWAR0—RapidIO inbound window attributes register 0 Access R/W R/W R/W Reset 0x0004_401F 0x0000_0000 0x8004_401F Section/Page 17.3.2.2.6/17-42 17.3.2.2.4/17-40 17.3.2.2.6/17-42 Error Management Registers 0xD_0E00 0xD_0E04 0xD_0E08 0xD_0E0C 0xD_0E10 0xD_0E14 0xD_0E18 0xD_0E20 0xD_0E28 0xD_0E2C PNFEDR—Port notification/fatal error detect register PNFEDiR—Port notification/fatal error detect disable register PNFEIER—Port notification/fatal error interrupt enable register PECSR—Port error capture status register EPCR0—Error packet capture register 0 EPCR1—Error packet capture register 1 EPCR2—Error packet capture register 2 PREDR—Port recoverable error detect register PERTR—Port error recovery threshold register PRTR—Port retry threshold register RapidIO Message Unit RapidIO Outbound Message Registers 0xD_1000 0xD_1004 0xD_100C 0xD_1014 0xD_1018 0xD_101C 0xD_1020 0xD_1028 OMR—Outbound mode register OSR—Outbound status register ODQDPAR—Outbound descriptor queue dequeue pointer address register OSAR—Outbound source address register ODPR—Outbound destination port register ODATR—Outbound destination attributes register ODCR—Outbound double-word count register ODQEPAR—Outbound descriptor queue enqueue pointer address register R/W R/W R/W R/W R/W R/W R/W R/W 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0006_0000 0x0000_0000 0x0000_0000 17.3.3.1.1/17-65 17.3.3.1.2/17-67 17.3.3.1.3/17-68 17.3.3.1.4/17-69 17.3.3.1.5/17-70 17.3.3.1.6/17-70 17.3.3.1.7/17-71 17.3.3.1.8/17-72 R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x00FF_0000 0x00FF_0000 17.3.2.3.1/17-44 17.3.2.3.2/17-47 17.3.2.3.3/17-49 17.3.2.3.4/17-52 17.3.2.3.5/17-52 17.3.2.3.6/17-53 17.3.2.3.16/17-58 17.3.2.3.23/17-61 17.3.2.3.24/17-64 17.3.2.3.25/17-64 RapidIO Inbound Message Registers 0xD_1060 0xD_1064 0xD_106C 0xD_1074 IMR—Inbound mailbox mode register ISR—Inbound mailbox status register IFQDPAR—Inbound frame queue dequeue pointer address register IFQEPAR—Inbound frame queue enqueue pointer address register R/W R/W R/W R/W 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 17.3.3.2.1/17-72 17.3.3.2.2/17-74 17.3.3.2.3/17-75 17.3.3.2.4/17-76 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 2-45 Memory Map Table 2-9. Memory Map (continued) Offset Register Access Reset Section/Page RapidIO Doorbell Registers 0xD_1460 0xD_1464 0xD_146C 0xD_1474 DMR—Doorbell mode register DSR—Doorbell status register DQDPAR—Doorbell queue dequeue pointer address register DQEPAR—Doorbell queue enqueue pointer address register R/W R/W R/W R/W 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 17.3.3.3.1/17-77 17.3.3.3.2/17-78 17.3.3.3.3/17-79 17.3.3.3.4/17-80 RapidIO Port-Write Registers 0xD_14E0 0xD_14E4 0xD_14EC PWMR—Port-write mode register PWSR—Port-write status register PWQBAR—Port-write queue base address register Global Utilities Registers Power-On Reset Configuration Values 0xE_0000 0xE_0004 0xE_0008 0xE_000C 0xE_0010 0xE_0020 PORPLLSR—POR PLL ratio status register PORBMSR—POR boot mode status register PORIMPSCR—POR I/O impedance status and control register PORDEVSR—POR I/O device status register PORDBGMSR—POR debug mode status register GPPORCR—General-purpose POR configuration register R R R/W R R R 0x00nn_00nn 0xnnnn_0000 0x000n_007F See ref. See ref. See ref. 18.4.1.1/18-4 18.4.1.2/18-5 18.4.1.3/18-6 18.4.1.4/18-7 18.4.1.5/18-9 18.4.1.6/18-9 R/W R/W R/W 0x0000_0000 0x0000_0000 0x0000_0000 17.3.3.4.1/17-81 17.3.3.4.2/17-82 17.3.3.4.3/17-83 Signal Multiplexing and GPIO Controls 0xE_0030 0xE_0040 0xE_0050 0xE_0060 GPIOCR—GPIO control register GPOUTDR—General-purpose output data register GPINDR—General-purpose input data register PMUXCR—Alternate function signal multiplex control Device Disables 0xE_0070 DEVDISR—Device disable control R/W 0x0000_0000 18.4.1.11/18-13 R/W R/W R R/W 0x0000_0000 0x0000_0000 0xnnnn_0000 0x0000_0000 18.4.1.7/18-10 18.4.1.8/18-11 18.4.1.9/18-12 18.4.1.9/18-12 Power Management Registers 0xE_0080 POWMGTCSR—Power management status and control register Interrupt Reporting 0xE_0090 MCPSUMR—Machine check summary register Read/ Clear 0x0000_0000 18.4.1.13/18-16 R/W 0x0000_0000 18.4.1.12/18-15 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 2-46 Freescale Semiconductor Complete CCSR Map Table 2-9. Memory Map (continued) Offset Register Version Registers 0xE_00A0 0xE_00A4 PVR—Processor version register SVR—System version register Debug Control 0xE_0E00 0xE_0E10 0xE_0E20 CLKOCR—Clock out select register DDRDLLCR—DDR DLL control register LBDLLCR—LBC DLL control register R/W R/W R/W 0x0000_0000 0x0000_0000 0x0000_0000 18.4.1.16/18-18 18.4.1.17/18-19 18.4.1.18/18-20 R R e500 processor version MPC8540 system version 18.4.1.14/18-17 18.4.1.15/18-18 Access Reset Section/Page Performance Monitor Control Registers 0xE_1000 0xE_1010 0xE_1014 0xE_1018 0xE_101C 0xE_1020 0xE_1024 0xE_1028 0xE_1030 0xE_1034 0xE_1038 0xE_1040 0xE_1044 0xE_1048 0xE_1050 0xE_1054 0xE_1058 0xE_1060 PMGC0—Performance monitor global control register PMLCA0—Performance monitor local control register A0 PMLCB0—Performance monitor local control register B0 PMC0 (upper)—Performance monitor counter 0 upper PMC0 (lower)—Performance monitor counter 0 lower PMLCA1—Performance monitor local control register A1 PMLCB1—Performance monitor local control register B1 PMC1—Performance monitor counter 1 PMLCA2—Performance monitor local control register A2 PMLCB2—Performance monitor local control register B2 PMC2—Performance monitor counter 2 PMLCA3—Performance monitor local control register A3 PMLCB3—Performance monitor local control register B3 PMC3—Performance monitor counter 3 PMLCA4—Performance monitor local control register A4 PMLCB4—Performance monitor local control register B4 PMC4—Performance monitor counter 4 PMLCA5—Performance monitor local control register A5 R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 19.3.2.1/19-5 19.3.2.2/19-5 19.3.2.2/19-5 19.3.3.1/19-9 19.3.3.1/19-9 19.3.2.2/19-5 19.3.2.2/19-5 19.3.3.1/19-9 19.3.2.2/19-5 19.3.2.2/19-5 19.3.3.1/19-9 19.3.2.2/19-5 19.3.2.2/19-5 19.3.3.1/19-9 19.3.2.2/19-5 19.3.2.2/19-5 19.3.3.1/19-9 19.3.2.2/19-5 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 2-47 Memory Map Table 2-9. Memory Map (continued) Offset 0xE_1064 0xE_1068 0xE_1070 0xE_1074 0xE_1078 0xE_1080 0xE_1084 0xE_1088 0xE_1090 0xE_1094 0xE_1098 Register PMLCB5—Performance monitor local control register B5 PMC5—Performance monitor counter 5 PMLCA6—Performance monitor local control register A6 PMLCB6—Performance monitor local control register B6 PMC6—Performance monitor counter 6 PMLCA7—Performance monitor local control register A7 PMLCB7—Performance monitor local control register B7 PMC7—Performance monitor counter 7 PMLCA8—Performance monitor local control register A8 PMLCB8—Performance monitor local control register B8 PMC8—Performance monitor counter 8 Access R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W Reset 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 Section/Page 19.3.2.2/19-5 19.3.3.1/19-9 19.3.3.1/19-9 19.3.2.2/19-5 19.3.3.1/19-9 19.3.2.2/19-5 19.3.2.2/19-5 19.3.3.1/19-9 19.3.2.2/19-5 19.3.2.2/19-5 19.3.3.1/19-9 Debug and Watchpoint Monitor Registers Watchpoint Monitor Registers 0xE_2000 0xE_2004 0xE_200C 0xE_2014 0xE_2018 0xE_201C WMCR0—Watchpoint monitor control register 0 WMCR1—Watchpoint monitor control register 1 WMAR—Watchpoint monitor address register WMAMR—Watchpoint monitor address mask register WMTMR—Watchpoint monitor transaction mask register WMSR—Watchpoint monitor status register Trace Buffer Registers 0xE_2040 0xE_2044 0xE_204C 0xE_2054 0xE_2058 0xE_205C 0xE_2060 0xE_2064 0xE_2068 TBCR0—Trace buffer control register TBCR1—Trace buffer control register TBAR—Trace buffer address register TBAMR—Trace buffer address mask register TBTMR—Trace buffer transaction mask register TBSR—Trace buffer status register TBACR—Trace buffer access control register TBADHR—Trace buffer access data high register TBADR—Trace buffer access data register R/W R/W R/W R/W R/W R/W R/W R/W R/W 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 20.3.2.1/20-16 20.3.2.1/20-16 20.3.2.2/20-19 20.3.2.3/20-19 20.3.2.4/20-20 20.3.2.5/20-21 20.3.2.6/20-22 20.3.2.7/20-22 20.3.2.8/20-23 R/W R/W R/W R/W R/W R/W 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 20.3.1.1/20-11 20.3.1.1/20-11 20.3.1.2/20-13 20.3.1.3/20-14 20.3.1.4/20-14 20.3.1.5/20-16 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 2-48 Freescale Semiconductor Complete CCSR Map Table 2-9. Memory Map (continued) Offset Register Context ID Registers 0xE_20A0 0xE_20A4 PCIDR—Programmed context ID register CCIDR—Current context ID register Other Registers 0xE_20B0 1 2 Access Reset Section/Page R/W R/W 0x0000_0000 0x0000_0000 20.3.3.1/20-24 20.3.3.2/20-24 TOSR—Trigger output source register R/W 0x0000_0000 20.3.4.1/20-25 Port size for BR0 is configured from external pins during reset, hence ‘nn’ is either 0x08, 0x10, or 0x18. TSEC2 has the same memory-mapped registers that are described for TSEC1 from 0x 2_4000 to 0x2_4FFF except that the offsets are from 0x 2_5000 to 0x2_5FFF. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 2-49 Memory Map MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 2-50 Freescale Semiconductor Chapter 3 Signal Descriptions This chapter describes the MPC8540 external signals. It is organized into the following sections: • • • Overview of signals and cross-references for signals that serve multiple functions, including two lists: one by functional block and one alphabetical List of reset configuration signals List of output signal states at reset NOTE A bar over a signal name indicates that the signal is active low, such as IRQ_OUT (interrupt out). Active-low signals are referred to as asserted (active) when they are low and negated when they are high. Signals that are not active low, such as IRQ (interrupt input), are referred to as asserted when they are high and negated when they are low. Internal signals throughout this document are shown as lower case and in italics. For example, sys_logic_clk is an internal signal. These are referenced only as necessary for understanding of the external functionality of the device. 3.1 • • • • • • • • • • Signals Overview DDR memory interface signals RapidIO interface signals PCI/PCI-X interface signals Ethernet management interface signals TSEC1 interface signals TSEC2 interface signals FEC10/100 interface signals Local bus interface signals DMA interface signals PIC interface signals MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 The MPC8540 signals are grouped as follows: Freescale Semiconductor 3-1 Signal Descriptions • • • • • DUART interface signals I2C interface signals System control, power management, and debug signals Test, JTAG, and configuration signals Clock signals Figure 3-1 illustrates the external signals of the MPC8540, showing how the signals are grouped. Refer to the MPC8540 Integrated Processor Hardware Specifications for a pinout diagram showing pin numbers and a listing of all the electrical and mechanical specifications. Note that individual chapters of this document provide details for each signal, describing each signal’s behavior when the signal is asserted or negated and when the signal is an input or an output. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 3-2 Freescale Semiconductor Signals Overview Memory Interface 130 Signals MDQ[0:63] MECC[0:7] MDM[0:7] MDM[8] MDQS[0:7] MDQS[8] MBA[0:1] MA[0:14] MWE MRAS MCAS MCS[0:3] MCKE[0:1] MCK[0:5], MCK[0:5] MSYNC_IN, MSYNC_OUT RIO_RCLK, RIO_RCLK RIO_RD[0:7], RIO_RD[0:7] RIO_RFRAME, RIO_RFRAME RIO_TCLK, RIO_TCLK RIO_TD[0:7], RIO_TD[0:7] RIO_TFRAME, RIO_TFRAME RIO_TX_CLK_IN, RIO_TX_CLK_IN PCI_AD[63:0] PCI_C/BE [7:0] PCI_PAR, PCI_PAR64 PCI_FRAME PCI_TRDY PCI_IRDY PCI_STOP PCI_DEVSEL PCI_IDSEL PCI_REQ64 PCI_ACK64 PCI_PERR PCI_SERR PCI_REQ0 PCI_REQ [4:1] PCI_GNT0 PCI_GNT[4:1] EC_GTX_CLK125 EC_MDC EC_MDIO TSEC1_TXD[7:0] TSEC1_TX_EN TSEC1_TX_ER TSEC1_TX_CLK TSEC1_GTX_CLK TSEC1_CRS TSEC1_COL TSEC1_RXD[7:0] TSEC1_RX_DV TSEC1_RX_ER TSEC1_RX_CLK TSEC2_TXD[7:0] TSEC2_TX_EN TSEC2_TX_ER TSEC2_TX_CLK TSEC2_GTX_CLK TSEC2_CRS TSEC2_COL TSEC2_RXD[7:0] TSEC2_RX_DV TSEC2_RX_ER TSEC2_RX_CLK TCK TDI TDO TMS TRST 64 8 8 1 8 1 2 15 1 1 1 4 2 12 2 2 16 2 2 16 2 2 64 8 2 1 1 1 1 1 1 1 1 1 1 1 4 1 4 1 1 1 8 1 1 1 1 1 1 8 1 1 1 8 1 1 1 1 1 1 8 1 1 1 1 1 1 1 1 RapidIO Interface 42 Signals 32 4 1 4 4 4 4 1 1 1 1 1 1 1 1 1 3 2 4 4 4 1 1 9 3 1 2 2 2 2 1 1 1 1 1 1 1 1 1 1 5 1 LAD[0:31] LDP[0:3] LA27 LA[28:31] LCS [0:3] LCS [4:7] LWE [0:3]/LSDDQM[0:3]/LBS [0:3] LBCTL LALE LGPL0/LSDA10 LGPL1/LSDWE LGPL2/LOE/LSDRAS LGPL3/LSDCAS LGPL4/LGTA/LUPWAIT/LPBSE LGPL5 LCKE LCLK[0:2] LSYNC_IN, LSYNC_OUT DMA_DREQ[0:1],DMA_DREQ[2:3] DMA_DACK[0:1],DMA_DACK[2:3] DMA_DDONE[0:1],DMA_DDONE[2:3] MCP UDE IRQ[0:8] IRQ[9:11] IRQ_OUT UART_SIN[0:1] UART_SOUT[0:1] UART_CTS[0:1] UART_RTS [0:1] IIC_SDA IIC_SCL HRESET HRESET_REQ SRESET CKSTP_IN CKSTP_OUT READY TRIG_IN TRIG_OUT MSRCID[0:4] MDVAL Local Bus Interface 67 Signals DMA Interface 12 Signals PIC Interface 15 Signals PCI/PCI-X Interface 94 Signals Dual UART Interface 8 Signals I2C Interface 2 Signals System Control 6 Signals MPC8540 Ethernet Interface 3 Signals Debug 8 Signals TSEC1 Ethernet Interface 25 Signals 1 ASLEEP Power Management 1 Signal 1 1 1 SYSCLK RTC CLK_OUT Clock 3 Signals TSEC2 Ethernet Interface 25 Signals 1 1 1 1 2 4 1 1 1 1 1 4 1 1 1 LSSD_MODE L1_TSTCLK L2_TSTCLK TEST_SEL THERM[0:1] FEC_TXD[3:0] FEC_TX_EN FEC_TX_ER FEC_TX_CLK FEC_CRS FEC_COL FEC_RXD[3:0] FEC_RX_DV FEC_RX_ER FEC_RX_CLK Test 6 Signals JTAG 5 Signals 10/100 Ethernet Interface 16 Signals Figure 3-1. MPC8540 Signal Groupings MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 3-3 Signal Descriptions The following tables provide summaries of signal functions. Table 3-1 provides a summary of the signals grouped by function, and Table 3-2 provides the summary list of the signals grouped alphabetically. These tables detail the signal name, interface, alternate functions, number of signals, and whether the signal is an input, output, or bidirectional. The direction of the multiplexed signals applies for the primary signal function listed in the left-most column of the table for that row (and does not apply for the state of the reset configuration signals). Finally, the table provides a pointer to the table where the signal function is described. Table 3-1. MPC8540 Signal Reference by Functional Block Name MDQ[0:63] MECC[0:7] MDM[0:7] MDM8 MDQS[0:7] MDQS8 MBA[0:1] MA[0:14] MWE MRAS MCAS MCS[0:3] MCKE[0:1] Description DDR data DDR error correcting code DDR data mask DDR ECC data mask DDR data strobe DDR ECC data strobe DDR bank select DDR address DDR write enable DDR row address strobe DDR column address strobe DDR chip select (2/DIMM) DDR clock enable Functional Block DDR memory DDR memory DDR memory DDR memory DDR memory DDR memory DDR memory DDR memory DDR memory DDR memory DDR memory DDR memory DDR memory DDR memory DDR memory RapidIO RapidIO RapidIO RapidIO Alternate Function(s) — — — — — — — — — — — — — — — — — — — No. of Signals 64 8 8 1 8 1 2 15 1 1 1 4 2 12 2 2 16 2 2 I/O I/O I/O O O I/O I/O O O O O O O O O I/O I I I O Table/ Page 9-3/9-5 9-3/9-5 9-3/9-5 9-3/9-5 9-3/9-5 9-3/9-5 9-3/9-5 9-3/9-5 9-3/9-5 9-3/9-5 9-3/9-5 9-3/9-5 9-4/9-8 9-4/9-8 9-4/9-8 17-5/17-8 17-5/17-8 17-5/17-8 17-5/17-8 MCK[0:5], MCK [0:5] DDR differential clocks (3 pairs/DIMM) MSYNC_IN, MSYNC_OUT RIO_RCLK, RIO_RCLK RIO_RD[0:7], RIO_RD[0:7] RIO_RFRAME, RIO_RFRAME RIO_TCLK, RIO_TCLK DDR DLL synchronization in/out RapidIO receive clocks RapidIO receive data RapidIO receive frame RapidIO transmit clock MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 3-4 Freescale Semiconductor Signals Overview Table 3-1. MPC8540 Signal Reference by Functional Block (continued) Name RIO_TD[0:7], RIO_TD[0:7] RIO_TFRAME, RIO_TFRAME RIO_TX_CLK_IN RIO_TX_CLK_IN PCI_AD[63:0] PCI_C/BE[7:0] PCI_PAR PCI_PAR64 PCI_FRAME PCI_TRDY PCI_IRDY PCI_STOP PCI_DEVSEL PCI_IDSEL PCI_REQ64 PCI_ACK64 PCI_PERR PCI_SERR PCI_REQ0 PCI_REQ [4:1] PCI_GNT0 PCI_GNT[4:1] Description RapidIO transmit data RapidIO transmit frame RapidIO transmit clock in PCI address/data PCI command/byte enable PCI parity PCI parity 64 PCI frame PCI target ready PCI initiator ready PCI stop PCI device select PCI initial device select PCI request 64 PCI acknowledge 64 PCI parity error PCI system error PCI request 0 PCI request 4–1 PCI grant 0 PCI grant 4–1 Functional Block RapidIO RapidIO RapidIO PCI/X PCI/X PCI/X PCI/X PCI/X PCI/X PCI/X PCI/X PCI/X PCI/X PCI/X PCI/X PCI/X PCI/X PCI/X PCI/X PCI/X PCI/X Alternate Function(s) — — — — — — — — — — — — — cfg_pci_width — — — — — — cfg_pci_mode cfg_pci_debug cfg_pci_arbiter cfg_pci_impd — cfg_tsec_reduce — cfg_tsec1 cfg_rom_loc[0:2] No. of Signals 16 2 2 64 8 1 1 1 1 1 1 1 1 1 1 1 1 1 4 1 4 I/O O O I I/O I/O I/O I/O I/O I/O I/O I/O I/O I I/O I/O I/O I/O I/O I I/O O Table/ Page 17-5/17-8 17-5/17-8 17-5/17-8 16-2/16-8 16-2/16-8 16-2/16-8 16-2/16-8 16-2/16-8 16-2/16-8 16-2/16-8 16-2/16-8 16-2/16-8 16-2/16-8 16-2/16-8 16-2/16-8 16-2/16-8 16-2/16-8 16-2/16-8 16-2/16-8 16-2/16-8 16-2/16-8 EC_GTX_CLK125 EC_MDC EC_MDIO TSEC1_TXD7 TSEC1_TXD[6:4] Gigabit reference clock Ethernet management data clock Ethernet management data in/out TSEC1 transmit data 7 TSEC1 transmit data 6–4 Gigabit clock Ethernet management Ethernet management TSEC1 TSEC1 1 1 1 1 3 I O I/O O O 14-3/14-14 14-3/14-14 14-3/14-14 14-3/14-14 14-3/14-14 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 3-5 Signal Descriptions Table 3-1. MPC8540 Signal Reference by Functional Block (continued) Name TSEC1_TXD3 TSEC1_TXD2 TSEC1_TXD1 TSEC1_TXD0 TSEC1_TX_EN TSEC1_TX_ER TSEC1_TX_CLK Description TSEC1 transmit data 3 TSEC1 transmit data 2 TSEC1 transmit data 1 TSEC1 transmit data 0 TSEC1 transmit enable TSEC1 transmit error TSEC1 transmit clock in Functional Block TSEC1 TSEC1 TSEC1 TSEC1 TSEC1 TSEC1 TSEC1 TSEC1 TSEC1 TSEC1 TSEC1 TSEC1 TSEC1 TSEC1 TSEC2 TSEC2 TSEC2 TSEC2 TSEC2 TSEC2 TSEC2 TSEC2 TSEC2 TSEC2 TSEC2 TSEC2 TSEC2 TSEC2 TSEC2 10/100 10/100 10/100 Alternate Function(s) — — — — — — — — No. of Signals 1 1 1 1 1 1 1 1 1 1 8 1 1 1 1 2 3 1 1 1 1 1 1 1 1 8 1 1 1 4 1 1 I/O O O O O O O I O I I I I I I O O O O O O O I O I I I I I I O O O Table/ Page 14-3/14-14 14-3/14-14 14-3/14-14 14-3/14-14 14-3/14-14 14-3/14-14 14-3/14-14 14-3/14-14 14-3/14-14 14-3/14-14 14-3/14-14 14-3/14-14 14-3/14-14 14-3/14-14 14-3/14-14 14-3/14-14 14-3/14-14 14-3/14-14 14-3/14-14 14-3/14-14 14-3/14-14 14-3/14-14 14-3/14-14 14-3/14-14 14-3/14-14 14-3/14-14 14-3/14-14 14-3/14-14 14-3/14-14 21-1/21-8 21-1/21-8 21-1/21-8 TSEC1_GTX_CLK TSEC1 transmit clock out TSEC1_CRS TSEC1_COL TSEC1_RXD[7:0] TSEC1_RX_DV TSEC1_RX_ER TSEC1_RX_CLK TSEC2_TXD7 TSEC2_TXD[6:5] TSEC2_TXD[4:2] TSEC2_TXD1 TSEC2_TXD0 TSEC2_TX_EN TSEC2_TX_ER TSEC2_TX_CLK TSEC1 carrier sense TSEC1 collision detect TSEC1 receive data TSEC1 receive data valid TSEC1 receiver error TSEC1 receive clock TSEC2 transmit data 7 TSEC2 transmit data 6–5 TSEC2 transmit data 4–2 TSEC2 transmit data 1 TSEC2 transmit data 0 TSEC2 transmit enable TSEC2 transmit error TSEC2 transmit clock in — — — — — — cfg_tsec2 cfg_lb_hold[0:1] cfg_dev_ID[7:5] — — — — — — — — — — — — — — — TSEC2_GTX_CLK TSEC2 transmit clock out TSEC2_CRS TSEC2_COL TSEC2_RXD[0:7] TSEC2_RX_DV TSEC2_RX_ER TSEC2_RX_CLK FEC_TXD[0:3] FEC_TX_EN FEC_TX_ER TSEC2 carrier sense TSEC2 collision detect TSEC2 receive data TSEC2 receive data valid TSEC2 receiver error TSEC2 receive clock FEC transmit data FEC transmit enable FEC transmit error MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 3-6 Freescale Semiconductor Signals Overview Table 3-1. MPC8540 Signal Reference by Functional Block (continued) Name FEC_TX_CLK FEC_CRS FEC_COL FEC_RXD[0:3] FEC_RX_DV FEC_RX_ER FEC_RX_CLK LAD[0:31] LDP[0:3] LA27 LA[28:31] Description FEC transmit clock in FEC carrier sense FEC collision detect FEC receive data FEC receive data valid FEC receiver error FEC receive clock LBC address/data LBC data parity LBC burst address LBC port address Functional Block 10/100 10/100 10/100 10/100 10/100 10/100 10/100 LBC LBC LBC LBC Alternate Function(s) — — — — — — — cfg_gpporcr — cfg_cpu_boot cfg_sys_pll0 cfg_sys_pll1 cfg_sys_pll2 cfg_sys_pll3 — DMA_DREQ2 DMA_DACK2 DMA_DDONE2 cfg_pci_hold[0:1] No. of Signals 1 1 1 4 1 1 1 32 4 1 4 I/O I I I I I I I I/O I/O O O Table/ Page 21-1/21-8 21-1/21-8 21-1/21-8 21-1/21-8 21-1/21-8 21-1/21-8 21-1/21-8 13-1/13-5 13-1/13-5 13-1/13-5 13-1/13-5 LCS[0:4] LCS5 LCS6 LCS7 LWE [0:1]/ LSDDQM[0:1]/ LBS[0:1] LWE [2:3]/ LSDDQM[2:3]/ LBS[2:3] LBCTL LALE LGPL0/LSDA10 LBC chip select 0–4 LBC chip select 5 LBC chip select 6 LBC chip select 7 LBC write enable/byte lane data mask/byte select 0–1 LBC write enable/byte lane data mask/byte select 2–3 LBC data buffer control LBC address latch enable LBC UPM general purpose line 0/SDRAM address bit 10 LBC GP line 1/SDRAM write enable LBC GP line 2/output enable/SDRAM RAS LBC GP line 3/SDRAM CAS LBC LBC LBC LBC LBC 5 1 1 1 2 O O O O O 13-1/13-5 13-1/13-5 13-1/13-5 13-1/13-5 13-1/13-5 LBC cfg_host_agt0 cfg_host_agt1 — cfg_core_pll0 cfg_rio_clk0 2 O 13-1/13-5 LBC LBC LBC 1 1 1 O O O 13-1/13-5 13-1/13-5 13-1/13-5 LGPL1/LSDWE LGPL2/LOE/ LSDRAS LGPL3/LSDCAS LBC LBC LBC cfg_rio_clk1 cfg_core_pll1 cfg_boot_seq0 1 1 1 O O O 13-1/13-5 13-1/13-5 13-1/13-5 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 3-7 Signal Descriptions Table 3-1. MPC8540 Signal Reference by Functional Block (continued) Name LGPL4/LGTA/ LUPWAIT/LPBSE LGPL5 LCKE LCLK[0:2] LSYNC_IN, LSYNC_OUT DMA_DREQ [0:1] DMA_DREQ2 DMA_DREQ3 DMA_DACK [0:1] DMA_DACK2 DMA_DACK3 Description LBC GP line 4/GPCM terminate access/UPM wait/parity byte select LBC GP line 5 address LBC clock enable LBC clock LBC DLL synchronization DMA request 0–1 DMA request 2 DMA request 3 DMA acknowledge 0–1 DMA acknowledge 2 DMA acknowledge 3 Functional Block LBC Alternate Function(s) — No. of Signals 1 I/O I/O Table/ Page 13-1/13-5 LBC LBC LBC LBC DMA DMA DMA DMA DMA DMA DMA DMA DMA PIC PIC PIC PIC PIC PIC PIC Dual UART Dual UART Dual UART Dual UART I2C IC System control System control 2 cfg_boot_seq1 — — — — LCS5 IRQ9 — LCS6 IRQ10 — LCS7 IRQ11 — — — DMA_DREQ3 DMA_DACK3 DMA_DDONE3 — — — — — — — — — 1 1 3 2 2 1 1 2 1 1 2 1 1 1 1 9 1 1 1 1 2 2 2 2 1 1 1 1 O O O I/O I I I O O O O O O I I I I I I O I O I O I/O I/O I O 13-1/13-5 13-1/13-5 13-1/13-5 13-1/13-5 15-3/15-6 15-3/15-6 15-3/15-6 15-3/15-6 15-3/15-6 15-3/15-6 15-3/15-6 15-3/15-6 15-3/15-6 10-6/10-8 10-6/10-8 10-6/10-8 10-6/10-8 10-6/10-8 10-6/10-8 10-6/10-8 12-2/12-4 12-2/12-4 12-2/12-4 12-2/12-4 11-2/11-4 11-2/11-4 4-2/4-2 4-2/4-2 DMA_DDONE[0:1] DMA done 0–1 DMA_DDONE2 DMA_DDONE3 MCP UDE IRQ[0:8] IRQ9 IRQ10 IRQ11 IRQ_OUT UART_SIN[0:1] UART_SOUT[0:1] UART_CTS[0:1] UART_RTS[0:1] IIC_SDA IIC_SCL HRESET HRESET_REQ DMA done 2 DMA done 3 Machine check processor Unconditional debug event External interrupt 0–8 External interrupt 9 External interrupt 10 External interrupt 11 Interrupt output DUART serial data in DUART serial data out DUART clear to send DUART ready to send I2C serial data I C serial clock Hard reset Hard reset request 2 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 3-8 Freescale Semiconductor Signals Overview Table 3-1. MPC8540 Signal Reference by Functional Block (continued) Name SRESET CKSTP_IN CKSTP_OUT READY TRIG_IN TRIG_OUT MSRCID[0:1] MSRCID[2:4] MDVAL ASLEEP SYSCLK RTC CLK_OUT LSSD_MODE L1_TSTCLK L2_TSTCLK TEST_SEL THERM[0:1] TCK TDI TDO TMS TRST Description Soft reset Checkstop in Checkstop out Device ready Watchpoint trigger in Watchpoint trigger out Memory debug source port ID 0–1 Memory debug source port ID 2–4 Memory debug data valid Asleep System clock/PCI clock Real time clock Clock out LSSD mode L1 test clock L2 test clock Test select Thermal resistor access Test clock Test data in Test data out Test mode select Test reset Functional Block System control System control System control System control Debug Debug Debug Debug Debug Power mgmt Clock Clock Clock Test Test Test Test Test JTAG JTAG JTAG JTAG JTAG Alternate Function(s) — — — TRIG_OUT — READY cfg_mem_debug cfg_ddr_debug — — — — — — — — — — — — — — — — No. of Signals 1 1 1 1 1 1 2 3 1 1 1 1 1 1 1 1 1 2 1 1 1 1 1 I/O I I O O I O O O O O I I O I I I I I I I O I I Table/ Page 4-2/4-2 18.2/18-1 18.2/18-1 4-2/4-2 20-2/20-6 20-2/20-6 9-6/9-10 20-2/20-6 20-2/20-6 18.2/18-1 4-3/4-3 4-3/4-3 18.2/18-1 20-2/20-6 20-2/20-6 20-2/20-6 20-2/20-6 20-2/20-6 20-2/20-6 20-2/20-6 20-2/20-6 20-2/20-6 20-2/20-6 Table 3-2. MPC8540 Alphabetical Signal Reference Name ASLEEP CKSTP_IN CKSTP_OUT CLK_OUT Description Asleep Checkstop in Checkstop out Clock out Functional Block Power mgmt System control System control Clock Alternate Function(s) — — — — No. of Signals 1 1 1 1 I/O O I O O Table/ Page 18.2/18-1 18.2/18-1 18.2/18-1 18.2/18-1 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 3-9 Signal Descriptions Table 3-2. MPC8540 Alphabetical Signal Reference (continued) Name DMA_DACK2 DMA_DACK3 DMA_DACK [0:1] DMA_DDONE2 DMA_DDONE3 Description DMA acknowledge 2 DMA acknowledge 3 DMA acknowledge 0–1 DMA done 2 DMA done 3 Functional Block DMA DMA DMA DMA DMA DMA DMA DMA DMA Gigabit clock Ethernet management Ethernet management 10/100 10/100 10/100 10/100 10/100 10/100 10/100 10/100 10/100 10/100 System control System control I2C IC PIC PIC PIC PIC 2 Alternate Function(s) LCS6 IRQ10 — LCS7 IRQ11 — LCS5 IRQ9 — — cfg_tsec_reduce — — — — — — — — — — — — — — — — DMA_DACK3 DMA_DDONE3 DMA_DREQ3 No. of Signals 1 1 2 1 1 2 1 1 2 1 1 1 1 1 4 1 1 1 4 1 1 1 1 1 1 1 9 1 1 1 I/O O O O O O O I I I I O I/O I I I I I I O I O O I O I/O I/O I I I I Table/ Page 15-3/15-6 15-3/15-6 15-3/15-6 15-3/15-6 15-3/15-6 15-3/15-6 15-3/15-6 15-3/15-6 15-3/15-6 14-3/14-14 14-3/14-14 14-3/14-14 21-1/21-8 21-1/21-8 21-1/21-8 21-1/21-8 21-1/21-8 21-1/21-8 21-1/21-8 21-1/21-8 21-1/21-8 21-1/21-8 4-2/4-2 4-2/4-2 11-2/11-4 11-2/11-4 10-6/10-8 10-6/10-8 10-6/10-8 10-6/10-8 DMA_DDONE[0:1] DMA done 0–1 DMA_DREQ2 DMA_DREQ3 DMA_DREQ [0:1] EC_GTX_CLK125 EC_MDC EC_MDIO FEC_COL FEC_CRS FEC_RXD[0:3] FEC_RX_CLK FEC_RX_DV FEC_RX_ER FEC_TXD[0:3] FEC_TX_CLK FEC_TX_EN FEC_TX_ER HRESET HRESET_REQ IIC_SCL IIC_SDA IRQ[0:8] IRQ10 IRQ11 IRQ9 DMA request 2 DMA request 3 DMA request 0–1 Gigabit reference clock Ethernet management data clock Ethernet management data in/out FEC collision detect FEC carrier sense FEC receive data FEC receive clock FEC receive data valid FEC receiver error FEC transmit data FEC transmit clock in FEC transmit enable FEC transmit error Hard reset Hard reset request I2C serial clock I C serial data External interrupt 0–8 External interrupt 10 External interrupt 11 External interrupt 9 2 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 3-10 Freescale Semiconductor Signals Overview Table 3-2. MPC8540 Alphabetical Signal Reference (continued) Name IRQ_OUT L1_TSTCLK L2_TSTCLK LA27 LAD[0:31] LALE LA[28:31] Description Interrupt output L1 test clock L2 test clock LBC burst address LBC address/data LBC address latch enable LBC port address Functional Block PIC Test Test LBC LBC LBC LBC Alternate Function(s) — — — cfg_cpu_boot cfg_gpporcr cfg_core_pll0 cfg_sys_pll0 cfg_sys_pll1 cfg_sys_pll2 cfg_sys_pll3 — — — DMA_DREQ2 DMA_DACK2 DMA_DDONE2 — — cfg_rio_clk0 No. of Signals 1 1 1 1 32 1 4 I/O O I I O I/O O O Table/ Page 10-6/10-8 20-2/20-6 20-2/20-6 13-1/13-5 13-1/13-5 13-1/13-5 13-1/13-5 LBCTL LCKE LCLK[0:2] LCS5 LCS6 LCS7 LCS[0:4] LDP[0:3] LGPL0/LSDA10 LBC data buffer control LBC clock enable LBC clock LBC chip select 5 LBC chip select 6 LBC chip select 7 LBC Chip select 0–4 LBC data parity LBC UPM general purpose line 0/SDRAM address bit 10 LBC GP line 1/SDRAM write enable LBC GP line 2/output enable/SDRAM RAS LBC GP line 3/SDRAM CAS LBC GP line 4/GPCM terminate access/UPM wait/parity byte select LBC GP line 5 address LSSD mode LBC DLL synchronization LBC write enable/byte lane data mask/byte select 0–1 LBC LBC LBC LBC LBC LBC LBC LBC LBC 1 1 3 1 1 1 5 4 1 O O O O O O O I/O O 13-1/13-5 13-1/13-5 13-1/13-5 13-1/13-5 13-1/13-5 13-1/13-5 13-1/13-5 13-1/13-5 13-1/13-5 LGPL1/LSDWE LGPL2/LOE/ LSDRAS LGPL3/LSDCAS LGPL4/LGTA/ LUPWAIT/LPBSE LGPL5 LSSD_MODE LSYNC_IN, LSYNC_OUT LWE [0:1]/ LSDDQM[0:1]/ LBS[0:1] LBC LBC LBC LBC cfg_rio_clk1 cfg_core_pll1 cfg_boot_seq0 — 1 1 1 1 O O O I/O 13-1/13-5 13-1/13-5 13-1/13-5 13-1/13-5 LBC Test LBC LBC cfg_boot_seq1 — — cfg_pci_hold[0:1] 1 1 2 2 O I I/O O 13-1/13-5 20-2/20-6 13-1/13-5 13-1/13-5 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 3-11 Signal Descriptions Table 3-2. MPC8540 Alphabetical Signal Reference (continued) Name LWE [2:3]/ LSDDQM[2:3]/ LBS[2:3] MA[0:14] MBA[0:1] MCAS MCKE[0:1] Description LBC write enable/byte lane data mask/byte select 2–3 DDR address DDR bank select DDR column address strobe DDR clock enable Functional Block LBC Alternate Function(s) cfg_host_agt0 cfg_host_agt1 — — — — — — — — — — — — — — — cfg_mem_debug cfg_ddr_debug — — — — — — — — — No. of Signals 2 I/O O Table/ Page 13-1/13-5 DDR memory DDR memory DDR memory DDR memory DDR memory PIC DDR memory DDR memory DDR memory DDR memory DDR memory DDR memory Debug DDR memory DDR memory Debug Debug DDR memory DDR memory PCI/X PCI/X PCI/X PCI/X PCI/X PCI/X 15 2 1 2 12 1 4 1 8 1 8 64 1 8 1 2 3 2 1 1 64 8 1 1 1 O O O O O I O O O I/O I/O I/O O I/O O O O I/O O I/O I/O I/O I/O I/O I/O 9-3/9-5 9-3/9-5 9-3/9-5 9-4/9-8 9-4/9-8 10-6/10-8 9-3/9-5 9-3/9-5 9-3/9-5 9-3/9-5 9-3/9-5 9-3/9-5 20-2/20-6 9-3/9-5 9-3/9-5 9-6/9-10 20-2/20-6 9-4/9-8 9-3/9-5 16-2/16-8 16-2/16-8 16-2/16-8 16-2/16-8 16-2/16-8 16-2/16-8 MCK[0:5], MCK [0:5] DDR differential clocks (3 pairs/DIMM) MCP MCS[0:3] MDM8 MDM[0:7] MDQS8 MDQS[0:7] MDQ[0:63] MDVAL MECC[0:7] MRAS MSRCID[0:1] MSRCID[2:4] MSYNC_IN, MSYNC_OUT MWE PCI_ACK64 PCI_AD[63:0] PCI_C/BE[7:0] PCI_DEVSEL PCI_FRAME PCI_GNT0 Machine check processor DDR chip select (2/DIMM) DDR ECC data mask DDR data mask DDR ECC data strobe DDR data strobe DDR data Memory debug data valid DDR error correcting code DDR row address strobe Memory debug source port ID 0–1 Memory debug source port ID 2–4 DDR DLL synchronization in/out DDR write enable PCI acknowledge 64 PCI address/data PCI command/byte enable PCI device select PCI frame PCI grant 0 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 3-12 Freescale Semiconductor Signals Overview Table 3-2. MPC8540 Alphabetical Signal Reference (continued) Name PCI_GNT[4:1] Description PCI grant 4–1 Functional Block PCI/X Alternate Function(s) cfg_pci_mode cfg_pci_debug cfg_pci_arbiter cfg_pci_impd — — — — — — cfg_pci_width — — — — TRIG_OUT — — — — — — — — — — — — — No. of Signals 4 I/O O Table/ Page 16-2/16-8 PCI_IDSEL PCI_IRDY PCI_PAR PCI_PAR64 PCI_PERR PCI_REQ0 PCI_REQ64 PCI_REQ [4:1] PCI_SERR PCI_STOP PCI_TRDY READY RIO_RCLK, RIO_RCLK RIO_RD[0:7], RIO_RD[0:7] RIO_RFRAME, RIO_RFRAME RIO_TCLK, RIO_TCLK RIO_TD[0:7], RIO_TD[0:7] RIO_TFRAME, RIO_TFRAME RIO_TX_CLK_IN RIO_TX_CLK_IN RTC SRESET SYSCLK TCK TDI TDO PCI initial device select PCI initiator ready PCI parity PCI parity 64 PCI parity error PCI request 0 PCI request 64 PCI request 4–1 PCI system error PCI stop PCI target ready Device ready RapidIO receive clocks RapidIO receive data RapidIO receive frame RapidIO transmit clock RapidIO transmit data RapidIO transmit frame RapidIO transmit clock in Real time clock Soft reset System clock/PCI clock Test clock Test data in Test data out PCI/X PCI/X PCI/X PCI/X PCI/X PCI/X PCI/X PCI/X PCI/X PCI/X PCI/X System control RapidIO RapidIO RapidIO RapidIO RapidIO RapidIO RapidIO Clock System control Clock JTAG JTAG JTAG 1 1 1 1 1 1 1 4 1 1 1 1 2 16 2 2 16 2 2 1 1 1 1 1 1 I I/O I/O I/O I/O I/O I/O I I/O I/O I/O O I I I O O O I I I I I I O 16-2/16-8 16-2/16-8 16-2/16-8 16-2/16-8 16-2/16-8 16-2/16-8 16-2/16-8 16-2/16-8 16-2/16-8 16-2/16-8 16-2/16-8 4-2/4-2 17-5/17-8 17-5/17-8 17-5/17-8 17-5/17-8 17-5/17-8 17-5/17-8 17-5/17-8 4-3/4-3 4-2/4-2 4-3/4-3 20-2/20-6 20-2/20-6 20-2/20-6 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 3-13 Signal Descriptions Table 3-2. MPC8540 Alphabetical Signal Reference (continued) Name TEST_SEL THERM[0:1] TMS TRIG_IN TRIG_OUT TRST TSEC1_COL TSEC1_CRS Description Test select Thermal resistor access Test mode select Watchpoint trigger in Watchpoint trigger out Test reset TSEC1 collision detect TSEC1 carrier sense Functional Block Test Test JTAG Debug Debug JTAG TSEC1 TSEC1 TSEC1 TSEC1 TSEC1 TSEC1 TSEC1 TSEC1 TSEC1 TSEC1 TSEC1 TSEC1 TSEC1 TSEC1 TSEC1 TSEC1 TSEC2 TSEC2 TSEC2 TSEC2 TSEC2 TSEC2 TSEC2 TSEC2 TSEC2 TSEC2 Alternate Function(s) — — — — READY — — — — — — — — — — — — cfg_rom_loc[0:2] cfg_tsec1 — — — — — — — — — — — — cfg_dev_ID[7:5] No. of Signals 1 2 1 1 1 1 1 1 1 8 1 1 1 1 1 1 1 3 1 1 1 1 1 1 1 8 1 1 1 1 1 3 I/O I I I I O I I I O I I I I O O O O O O I O O I I O I I I I O O O Table/ Page 20-2/20-6 20-2/20-6 20-2/20-6 20-2/20-6 20-2/20-6 20-2/20-6 14-3/14-14 14-3/14-14 14-3/14-14 14-3/14-14 14-3/14-14 14-3/14-14 14-3/14-14 14-3/14-14 14-3/14-14 14-3/14-14 14-3/14-14 14-3/14-14 14-3/14-14 14-3/14-14 14-3/14-14 14-3/14-14 14-3/14-14 14-3/14-14 14-3/14-14 14-3/14-14 14-3/14-14 14-3/14-14 14-3/14-14 14-3/14-14 14-3/14-14 14-3/14-14 TSEC1_GTX_CLK TSEC1 transmit clock out TSEC1_RXD[7:0] TSEC1_RX_CLK TSEC1_RX_DV TSEC1_RX_ER TSEC1_TXD0 TSEC1_TXD1 TSEC1_TXD2 TSEC1_TXD3 TSEC1_TXD[6:4] TSEC1_TXD7 TSEC1_TX_CLK TSEC1_TX_EN TSEC1_TX_ER TSEC2_COL TSEC2_CRS TSEC1 receive data TSEC1 receive clock TSEC1 receive data valid TSEC1 receiver error TSEC1 transmit data 0 TSEC1 transmit data 1 TSEC1 transmit data 2 TSEC1 transmit data 3 TSEC1 transmit data 6–4 TSEC1 transmit data 7 TSEC1 transmit clock in TSEC1 transmit enable TSEC1 transmit error TSEC2 collision detect TSEC2 carrier sense TSEC2_GTX_CLK TSEC2 transmit clock out TSEC2_RXD[0:7] TSEC2_RX_CLK TSEC2_RX_DV TSEC2_RX_ER TSEC2_TXD0 TSEC2_TXD1 TSEC2_TXD[4:2] TSEC2 receive data TSEC2 receive clock TSEC2 receive data valid TSEC2 receiver error TSEC2 transmit data 0 TSEC2 transmit data 1 TSEC2 transmit data 4–2 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 3-14 Freescale Semiconductor Configuration Signals Sampled at Reset Table 3-2. MPC8540 Alphabetical Signal Reference (continued) Name TSEC2_TXD[6:5] TSEC2_TXD7 TSEC2_TX_CLK TSEC2_TX_EN TSEC2_TX_ER UART_CTS[0:1] UART_RTS[0:1] UART_SIN[0:1] UART_SOUT[0:1] UDE Description TSEC2 transmit data 6–5 TSEC2 transmit data 7 TSEC2 transmit clock in TSEC2 transmit enable TSEC2 transmit error DUART clear to send DUART ready to send DUART serial data in DUART serial data out Unconditional debug event Functional Block TSEC2 TSEC2 TSEC2 TSEC2 TSEC2 Dual UART Dual UART Dual UART Dual UART PIC Alternate Function(s) cfg_lb_hold[0:1] cfg_tsec2 — — — — — — — — No. of Signals 2 1 1 1 1 2 2 2 2 1 I/O O O I O O I O I O I Table/ Page 14-3/14-14 14-3/14-14 14-3/14-14 14-3/14-14 14-3/14-14 12-2/12-4 12-2/12-4 12-2/12-4 12-2/12-4 10-6/10-8 3.2 Configuration Signals Sampled at Reset The signals that serve alternate functions as configuration input signals during system reset are summarized in Table 3-3. The detailed interpretation of their voltage levels during reset is described in Chapter 4, “Reset, Clocking, and Initialization.” Note that throughout this document, the reset configuration signals are described as being sampled at the negation of HRESET. However, there is a setup and hold time for these signals relative to the rising edge of HRESET, as described in the MPC8540 Integrated Processor Hardware Specifications. Note that the PLL configuration signals have different setup and hold time requirements than the other reset configuration signals. The reset configuration signals are multiplexed with other functional signals. The values on these signals during reset are interpreted to be logic one or zero, regardless of whether the functional signal name is defined as active-low. Most of the reset configuration signals have internal pull-up resistors so that if the signals are not driven, the default value is high (a one), as shown in the table. Some signals do not have pull-up resistors and must be driven high or low during the reset period. For details about all the signals that require external pull-up resistors, see the MPC8540 Integrated Processor Hardware Specifications. Note that the multiplexing of various signals on the MPC8540 is controlled by the PMUXCR register described in Chapter 18, “Global Utilities.” MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 3-15 Signal Descriptions Table 3-3. MPC8540 Reset Configuration Signals Functional Interface PCI Functional Signal Name PCI_REQ64 PCI_GNT4 PCI_GNT3 PCI_GNT2 PCI_GNT1 Ethernet Management TSEC1 EC_MDC TSEC1_TXD7 TSEC1_TXD[6:4] TSEC2 TSEC2_TXD7 TSEC2_TXD[6:5] TSEC2_TXD[4:2] LBC LA27 LA[28:31] LWE[0:1]/LBS[0:1] LWE[2:3]/LBS[2:3] LALE LGPL0/LSDA10 LGPL1/LSDWE LGPL2/LOE/LSDRAS LGPL3/LSDCAS LGPL5/LSDAMUX LAD[0:31] Debug MSRCID0 MSRCID1 Reset Configuration Name cfg_pci_width cfg_pci_mode cfg_pci_debug cfg_pci_arbiter cfg_pci_impd cfg_tsec_reduce cfg_tsec1 cfg_rom_loc[0:2] cfg_tsec2 cfg_lb_hold[0:1] cfg_dev_ID[7:5] cfg_cpu_boot cfg_sys_pll[0:3] cfg_pci_hold[0:1] cfg_host_agt0 cfg_host_agt1 cfg_core_pll0 cfg_rio_clk0 cfg_rio_clk1 cfg_core_pll1 cfg_boot_seq0 cfg_boot_seq1 cfg_gpporcr cfg_mem_debug cfg_ddr_debug Default 1 1 1 1 1 1 1 111 1 11 111 1 Must be driven 11 11 Must be driven 1 1 Must be driven 1 1 Indeterminate if not driven (no default) 1 1 3.3 Output Signal States During Reset When a system reset is recognized (HRESET is asserted), the MPC8540 aborts all current internal and external transactions and releases all bidirectional I/O signals to a high-impedance state. See Chapter 4, “Reset, Clocking, and Initialization,” for a complete description of the reset functionality. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 3-16 Freescale Semiconductor Output Signal States During Reset During reset, the MPC8540 ignores most input signals (except for the reset configuration signals) and drives most of the output-only signals to an inactive state. Table 3-4 shows the states of the output-only signals (that is, the signals that are not multiplexed with other inputs, or are not used as reset configuration signals during system reset). Table 3-4. Output Signal States During System Reset Interface DDR Memory DDR Memory DDR Memory DDR Memory DDR Memory DDR Memory DDR Memory DDR Memory DDR Memory RapidIO RapidIO RapidIO TSEC1 TSEC1 TSEC1 TSEC1 TSEC2 TSEC2 TSEC2 TSEC2 FEC FEC FEC LBC LBC LBC MDM[0:8] MBA[0:1] MA[0:14] MWE MRAS MCAS MCS[0:3] MCKE[0:1] MCK[0:5], MCK [0:5] RIO_TCLK, RIO_TCLK RIO_TD[0:7], RIO_TD[0:7] RIO_TFRAME, RIO_TFRAME TSEC1_TXD[3:0] TSEC1_TX_EN TSEC1_TX_ER TSEC1_GTX_CLK TSEC2_TXD[1:0] TSEC2_TX_EN TSEC2_TX_ER TSEC2_GTX_CLK FEC_TXD[0:3] FEC_TX_EN FEC_TX_ER LCLK[0:2] LCS[0:5] DMA_DACK2/LCS6 DMA_DDONE2/LCS7 LBCTL DMA_DACK [0:1] Signal State During Reset High-Z High-Z High-Z High-Z High-Z High-Z High-Z Driven Low High-Z Driven (toggling) High-Z High-Z Input—reset config (test only) Driven Low High-Z High-Z High-Z Driven Low High-Z High-Z High-Z Driven Low High-Z High-Z High-Z High-Z LBC DMA Input—reset config (test only) High-Z MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 3-17 Signal Descriptions Table 3-4. Output Signal States During System Reset (continued) Interface DMA Signal DMA_DACK2/LCS6 DMA_DDONE2/LCS7 DMA_DDONE0 DMA_DDONE1 IRQ8 IRQ_OUT UART_SOUT[0:1] UART_RTS[0:1] State During Reset High-Z DMA DMA PIC PIC Dual UART Dual UART High-Z Driven (test only) Driven (test only) High-Z High-Z High-Z High-Z High-Z Input—reset config (test only) High-Z High-Z Input—reset config (test only) High-Z System Control HRESET_REQ System Control CKSTP_OUT Debug Debug Debug Power Mgmt Clock TRIG_OUT/READY MSRCID[2:4] MDVAL ASLEEP CLK_OUT MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 3-18 Freescale Semiconductor Chapter 4 Reset, Clocking, and Initialization This chapter describes the reset, clocking, and some overall initialization of the MPC8540, including a definition of the reset configurations signals and the options they select. Additionally, the configuration, control, and status registers are described. Note that chapters in this book describe specific initialization aspects for individual blocks. 4.1 Overview The reset, clocking, and control signals provide many options for the operation of the MPC8540. Additionally, many modes are selected with reset configuration signals during the assertion of a hard reset (assertion of HRESET). 4.2 External Signal Descriptions Table 4-1 summarizes the external signals described in this chapter. Table 4-2 and Table 4-3 have detailed signal descriptions, but Table 4-1 contains references to additional sections that contain more information. Table 4-1. Signal Summary Signal HRESET HRESET_REQ SRESET READY SYSCLK RTC I/O I O I O I I Description Hard reset input. Causes a power-on reset (POR) sequence Hard reset request output. An internal block requests that HRESET be asserted Soft reset input. Causes mcp assertion to the core The MPC8540 has completed the reset operation and is not in a power-down (nap, doze or sleep) or debug state. Primary clock input to the MPC8540 Real time clock input References (Section/Page) 4.4.1.2/4-9 — 4.4.1.1/4-9 4.4.2/4-9 4.4.4.1/4-23 4.4.4.4/4-25 The following sections describe the reset and clock signals in detail. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 4-1 Reset, Clocking, and Initialization 4.2.1 System Control Signals Table 4-2 describes some of the system control signals of the MPC8540. Section 4.4.3, “Power-On Reset Configuration,” describes the signals that also function as reset configuration signals. Note that the CKSTP_IN and CKSTP_OUT signals are described in Chapter 18, “Global Utilities.” Table 4-2. System Control Signals—Detailed Signal Descriptions Signal HRESET I/ O I Description Hard reset. Causes the MPC8540 to abort all current internal and external transactions and set all registers to their default values. HRESET m ay be asserted completely asynchronously with respect to all other signals. State Asserted/Negated—See Chapter 3, “Signal Descriptions,” and Section 4.4.3, “Power-On Reset Configuration,” for more information on the interpretation of the Meanin other MPC8540 signals during reset. g Timing Assertion/Negation—The MPC8540 Integrated Processor Hardware Specifications gives specific timing information for this signal and the reset configuration signals. HRESET_RE Q O Hard reset request. Indicates to the board (system in which the MPC8540 is embedded) that a condition requiring the assertion of HRESET has been detected. State Asserted—A watchdog timer, a RapidIO command, or a boot sequencer failure (see Section 11.4.5, “Boot Sequencer Mode”) has triggered a request for hard reset. Meanin g Negated—Indicates no reset request Timing Assertion/Negation—May occur anytime, synchronous to the core complex bus clock. Once asserted, HRESET_REQ does not negate until HRESET is asserted. SRESET I Soft reset. Causes a machine check interrupt to the e500 core. Note that if the e500 core is not configured to process machine check interrupts, the assertion of SRESET causes a core checkstop. SRESET need not be asserted during a hard reset. State Asserted—Asserting SRESET causes a machine check interrupt (edge sensitive) to Meanin the e500 core. SRESET has no effect while HRESET is asserted. However, the g POR sequence is paused if SRESET is asserted during POR. Timing Assertion—May occur at any time, asynchronous to any clock Negation—Must be asserted for at least two CCB_clk cycles READY O Ready. Multiplexed with TRIG_OUT. See Chapter 20, “Debug Features and Watchpoint Facility,” for more information on TOSR and TRIG_OUT. State Asserted—Indicates that the MPC8540 has completed the reset operation and is not in a power-down state (nap, doze or sleep) when TOSR[SEL] equals 0b000. See Meanin Section 4.4.2, “Power-On Reset Sequence,” for more information. g Timing Assertion/Negation—Initial assertion of READY after reset is synchronous with SYSCLK. Subsequent assertion/negation due to power down modes occurs asynchronously. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 4-2 Freescale Semiconductor Memory Map/Register Definition 4.2.2 Clock Signals Table 4-3 describes the overall clock signals of the MPC8540. Note that some clock signals are specific to blocks within the MPC8540, and although some of their functionality is described in Section 4.4.4, “Clocking,” they are defined in detail in their respective chapters. Note that there is also a CLK_OUT signal in the MPC8540; the signal driven on the CLK_OUT pin is selectable and described in Section 18.4.1.16, “Clock Out Control Register (CLKOCR).” Table 4-3. Clock Signals—Detailed Signal Descriptions Signal SYSCL K I/ O I Description System clock/PCI clock (SYSCLK/PCI_CLK). SYSCLK is the primary clock input to the MPC8540. It is the clock source for the e500 core and for all devices and interfaces that operate synchronously with the core. Multiplied up with a phased-lock loop (PLL) to create the core complex bus (CCB) clock (also called the platform clock) which is used by virtually all of the synchronous system logic, include the L2 cache, the DDR SDRAM and local bus memory controllers, and other internal blocks such as the DMA and interrupt controllers. The CCB clock, in turn, feeds the PLL in the e500 core and the DLLs that create the DDR SDRAM and local bus memory clocks. When the PCI/PCI-X interface is used, SYSCLK also functions as the PCI_CLK signal. Note that this is true whether the MPC8540 is in agent or host mode. The MPC8540 does not provide a separate PCI_CLK output in host mode. Timing Assertion/Negation—See the MPC8540 Integrated Processor Hardware Specifications for specific timing information for this signal. RTC I Real time clock. May be used (optionally) to clock the time base of the e500 core. The maximum input frequency should not exceed 1/8th the core frequency. See Section 4.4.4.4, “Real Time Clock.” This signal can also be used (optionally) to clock the global timers in the programmable interrupt controller (PIC). Timing Assertion/Negation—See the MPC8540 Integrated Processor Hardware Specifications for specific timing information for this signal. 4.3 Memory Map/Register Definition This section describes the configuration and control registers that control access to the configuration space and to the boot code as well as guidelines for accessing these regions. It also contains a brief description of the boot sequencer which may be used to initialize configuration registers or memory before the CPU is released to boot. 4.3.1 Local Configuration Control Table 4-4 shows the memory map for the configuration, control, and status registers. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 4-3 Reset, Clocking, and Initialization Table 4-4. Local Configuration Control Register Map Local Memory Offset (Hex) 0x0_0000 0x0_0008 0x0_0010 0x0_0020 Register CCSRBAR—Configuration, control, and status registers base address register ALTCBAR—Alternate configuration base address register ALTCAR—Alternate configuration attribute register BPTR—Boot page translation register Access R/W R/W R/W R/W Reset 0x000F_F700 0x0000_0000 0x0000_0000 0x0000_0000 Section/Page 4.3.1.1.2/4-5 4.3.1.2.1/4-6 4.3.1.2.2/4-6 4.3.1.3.1/4-8 4.3.1.1 Accessing Configuration, Control, and Status Registers The configuration, control, and status registers are memory mapped. The set of configuration, control, and status registers occupies a 1-Mbyte region of memory. Their location is programmable using the CCSR base address register (CCSRBAR). The default base address for the configuration, control, and status registers is 0xFF70_0000 (CCSRBAR = 0x000F_F700). CCSRBAR itself is part of the local access block of CCSR memory, which begins at offset 0x0 from CCSRBAR. Because CCSRBAR is at offset 0x0 from the beginning of the local access registers, CCSRBAR always points to itself. The contents of CCSRBAR are broadcast internally in the MPC8540 to all functional units that need to be able to identify or create configuration transactions. 4.3.1.1.1 Updating CCSRBAR Updates to CCSRBAR that relocate the entire 1-Mbyte region of configuration, control, and status registers, requires special treatment. The effect of the update must be guaranteed to be visible by the mapping logic before an access to the new location is seen. To make sure this happens, these guidelines should be followed: • CCSRBAR should be updated during initial configuration of the device when only one host or controller has access to the device. – If the boot sequencer is being used to initialize, it is recommended that the boot sequencer set CCSRBAR to its desired final location. – If an external host on PCI or RapidIO is configuring the device, it should set CCSRBAR to the desired final location before the e500 core is released to boot. – If the e500 core is initializing the device, it should set CCSRBAR to the desired final location before enabling other I/O devices to access the device. When the e500 core is writing to CCSRBAR, it should use the following sequence: – Read the current value of CCSRBAR using a load word instruction followed by an isync. This forces all accesses to configuration space to complete. – Write the new value to CCSRBAR. • MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 4-4 Freescale Semiconductor Memory Map/Register Definition – Perform a load of an address that does not access configuration space or the on-chip SRAM, but has an address mapping already in effect (for example, boot ROM). Follow this load with an isync. – Read the contents of CCSRBAR from its new location, followed by another isync instruction. 4.3.1.1.2 Configuration, Control, and Status Base Address Register (CCSRBAR) Figure 4-1 shows the fields of CCSRBAR. 0 1 2 3 4 5 6 7 8 9 10 11 12 23 24 25 26 27 28 29 30 31 R0 0 0 0 00 00 0 0 0 0 BASE_ADDR W Reset Offset 0000_0000_0000_1111_1111_0111_0000_0000 0x000 00000000 Figure 4-1. Configuration, Control, and Status Register Base Address Register (CCSRBAR) Table 4-5 defines the bit fields of CCSRBAR. Table 4-5. CCSRBAR Field Descriptions Bits 0–11 Name — Write reserved, read = 0 Description 12–23 BASE_ADDR Identifies the12 most-significant address bits of the window used for configuration accesses. The base address is aligned on a 1-Mbyte boundary. 24–31 — Write reserved, read = 0 4.3.1.2 Accessing Alternate Configuration Space An alternate configuration space can be accessed by configuring the ALTCBAR and ALTCAR registers. These are intended to be used with the boot sequencer to allow the boot sequencer to access an alternate 1-Mbyte region of configuration space. By loading the proper boot sequencer command in the serial ROM the base address in the ALTCBAR can be combined with the 20 bits of address offset supplied from the serial ROM to generate a 32-bit address that is mapped to the target specified in ALTCAR. Thus, by configuring these registers, the boot sequencer has access to the entire memory map, one 1-Mbyte block at a time. See Section 11.4.5, “Boot Sequencer Mode,” for more information. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 4-5 Reset, Clocking, and Initialization NOTE The enable bit in the ALTCAR register should be cleared either by the boot sequencer or by the boot code that executes after the boot sequencer has completed its configuration operations. This prevents problems with incorrect mappings if subsequent configuration of the local access windows uses a different target mapping for the address specified in ALTCBAR. 4.3.1.2.1 Alternate Configuration Base Address Register (ALTCBAR) Figure 4-2 shows the fields of ALTCBAR. 0 11 12 23 24 31 R0 W Reset Offset 0 0000000 0 00 BASE_ADDR 0 0 00 0 0 00 0000_0000_0000_0000_0000_0000_0000_0000 0x008 Figure 4-2. Alternate Configuration Base Address Register (ALTCBAR) Table 4-6 defines the bit fields of ALTCBAR. Table 4-6. ALTCBAR Field Descriptions Bits 0–11 Name — Write reserved, read = 0 Description 12–23 BASE_ADDR Identifies the12 most-significant address bits of an alternate window used for configuration accesses. 24–31 — Write reserved, read = 0 4.3.1.2.2 Alternate Configuration Attribute Register (ALTCAR) Figure 4-3 shows the fields of ALTCAR. 0 1 2 3 4 5 6 7 8 11 12 30 31 R EN W Reset Offset 0000000 TRGT_ID 00000000000000000000 0000_0000_0000_0000_0000_0000_0000_0000 0x010 Figure 4-3. Alternate Configuration Attribute Register (ALTCAR) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 4-6 Freescale Semiconductor Memory Map/Register Definition Table 4-7 defines ALTCAR fields. Table 4-7. ALTCAR Field Descriptions Bits 0 Name EN Description Enable for a second configuration window. Like CCSRBAR, it has a fixed size of 1 Mbyte. 0 Second configuration window is disabled 1 Second configuration window is enabled Write reserved, read = 0 1–7 8–11 — TRGT_ID Identifies the device ID to target when a transaction hits in the 1-Mbyte address range defined by the second configuration window. 0000 PCI/PCI-X interface 0001 Reserved 0010 Reserved 0011 Reserved 0100 Local bus controller 0101–0111Reserved 1000 Configuration, control, and status registers 1001–1011Reserved 1100 RapidIO 1101 Reserved 1110 Reserved 1111 Local memory —DDR SDRAM and on-chip SRAM — Write reserved, read = 0 12–31 4.3.1.3 Boot Page Translation When the e500 core comes out of reset, its MMU has one 4-Kbyte page defined at 0xFFFF_Fnnn. The first instruction executed by the e500 core is always address 0xFFFF_FFFC, which must be a branch to an address within the 4-Kbyte boot page. For systems in which the boot code resides at a different address, the MPC8540 provides boot page translation capability. Boot page translation is controlled by the boot page translation register (BPTR). If enabled, this translation affects CPU accesses to address range 0xFFFF_Fnnn. The boot sequencer can enable boot page translation, or the boot page translation can be enabled by an external host when the MPC8540 is configured to be in boot holdoff mode. If translation is performed to a page outside of the default boot ROM address range defined in the MPC8540 (8 Mbytes at 0xFF80_0000 to 0xFFFF_FFFF as defined in Section 4.4.3.3, “Boot ROM Location”), the external host or boot sequencer must also set up a local access window to define the routing of the boot code fetch to the target interface that contains the boot code because the BPTR defines only the address translation, not the target interface. See Section 2.1, “Local Memory Map Overview and Example,” and Section 11.4.5, “Boot Sequencer Mode,” for more information. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 4-7 Reset, Clocking, and Initialization 4.3.1.3.1 Boot Page Translation Register (BPTR) Figure 4-4 shows the fields of BPTR. 0 1 11 12 31 R EN W Reset Offset 00000000000 BOOT_PAGE 0000_0000_0000_0000_0000_0000_0000_0000 0x020 Figure 4-4. Boot Page Translation Register (BPTR) Table 4-8 describes BPTR field descriptions. Table 4-8. BPTR Field Descriptions Bits 0 Name EN Description Boot page translation enable 0 Boot page is not translated 1 Boot page is translated as defined in the BPTR[BOOT_PAGE] parameter Write reserved, read = 0 1–11 — 12–31 BOOT_PAGE Translation for boot page. If enabled, the high-order 20 bits of accesses to 0xFFFF_F nnn are replaced with this value. 4.3.2 Boot Sequencer The boot sequencer is a DMA engine that accesses a serial ROM on the I2C interface and writes data to CCSR memory or the memory space pointed to by the alternate configuration base address register (ALTCBAR). See Section 4.3.1.2, “Accessing Alternate Configuration Space.” The boot sequencer is enabled by reset configuration pins as described in Section 4.4.3.6, “Boot Sequencer Configuration.” If the boot sequencer is enabled, the e500 core is held in reset until the boot sequencer has completed its operation. For more details, see Section 11.4.5, “Boot Sequencer Mode,” in the I2C chapter. 4.4 Functional Description This section describes the various ways to reset the MPC8540 device, the POR configurations, and the clocking on the MPC8540. 4.4.1 Reset Operations The MPC8540 has reset input signals for hard and soft reset operation. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 4-8 Freescale Semiconductor Functional Description 4.4.1.1 Soft Reset Assertion of SRESET causes a machine check interrupt to the e500 core. When this occurs, the soft reset flag is recorded in the machine check summary register (MCPSUMR) in the global utilities block so that software can identify the machine check as a soft reset condition. See the PowerPC™ e500 Core Complex Reference Manual for more information on the machine check interrupt, and Section 18.4.1.13, “Machine Check Summary Register (MCPSUMR),” for more information on the setting of the soft reset flag. Note that if SRESET is asserted before the e500 core is configured to handle a machine check interrupt, a core checkstop condition occurs, which causes CKSTP_OUT to assert. 4.4.1.2 Hard Reset The MPC8540 can be completely reset by the assertion of the HRESET input. The assertion of this signal by external logic is the equivalent of a POR and causes the sequence of events described in Section 4.4.2, “Power-On Reset Sequence.” Refer to the MPC8540 Integrated Processor Hardware Specifications for the timing requirements for HRESET assertion and negation. The MPC8540 hard reset request output signal (HRESET_REQ) indicates to external logic that a hard reset is being requested by hardware or a RapidIO device. Hardware causes this signal to assert for a boot sequencer failure (see Section 11.4.5, “Boot Sequencer Mode,” and Section 11.4.5.2, “EEPROM Data Format,”) or when the e500 watchdog timer is configured to cause a reset request when it expires. A RapidIO device causes this signal to assert if it sends four consecutive RapidIO link maintenance reset commands without any other intervening packets or control symbols, except idle control symbols (see Section 17.4.1, “RapidIO Transaction, Packet, and Control Symbol Summary,” for details of RapidIO transaction types). 4.4.2 Power-On Reset Sequence The POR sequence for the MPC8540 is as follows: 1. Power is applied to meet the specifications in the MPC8540 Integrated Processor Hardware Specifications. 2. System asserts HRESET and TRST causing all registers to be initialized to their default states and most I/O drivers to be three-stated (some clock, clock enabled, and system control signals are active). 3. System applies a stable SYSCLK signal and stable PLL configuration inputs, and the device PLL begins locking to SYSCLK. 4. System negates HRESET after its required hold time and after POR configuration inputs have been valid for at least 4 SYSCLK cycles. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 4-9 Reset, Clocking, and Initialization NOTE If the JTAG signals are not used, TRST may be tied inactive; however it is recommended that TRST not remain asserted after the negation of HRESET. TRST may be connected directly to HRESET. There is no need to assert the SRESET signal when HRESET is asserted. If SRESET is asserted upon negation of HRESET, the POR sequence will be paused until SRESET is negated. 5. The MPC8540 enables I/O drivers. 6. The MPC8540 PCI/PCI-X interface can assert DEVSEL in response to configuration cycles. 7. The e500 PLL configuration inputs are applied, allowing the e500 PLL to begin locking to the device clock (the CCB clock). 8. The CCB clock is cycled for approximately 50 µs to lock the e500 PLL. 9. The internal hard reset to the e500 core is negated and soft resets are negated to the DLLs and other remaining I/O blocks. The DLLs begin to lock. 10. When DLL locking is completed, the boot sequencer is released, causing it to load configuration data from serial ROMs, if enabled, as described in Section 4.4.3.6, “Boot Sequencer Configuration.” 11. When the boot sequencer completes, the RapidIO interface begins training, the PCI interface is released to accept external requests, and the boot vector fetch by the e500 core is allowed to proceed unless processor booting is further held off by POR configuration inputs as described in Section 4.4.3.5, “CPU Boot Configuration.” The MPC8540 is now in its ready state. 12. The ASLEEP signal negates synchronized to a rising edge of SYSCLK, indicating the ready state. The ready state is also indicated by the assertion of READY/TRIG_OUT if TOSR[SEL] = 000. In this case, READY is asserted with the same rising edge of SYSCLK, to indicate that the device has reached its ready state. See Section 20.3.4.1, “Trigger Out Source Register (TOSR),” for more information on this register. Asserting READY allows external system monitors to know basic device status. For example, exactly when it emerges from reset, or if the device is in a low-power mode. For more information on the debug functions of TRIG_OUT, see Section 20.3.4, “Trigger Out Function.” For more information about power management states, see Section 18.4.1, “Register Descriptions.” MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 4-10 Freescale Semiconductor Functional Description Figure 4-5 shows a timing diagram of the POR sequence. HRESET HRESET_REQ (high impedance) SYSCLK TRESET SRESET PLL Configs POR Configs (High Impedance) ASLEEP READY1 1 (High Impedance) Multiplexed with TRIG_OUT. Figure 4-5. Power-On Reset Sequence 4.4.3 Power-On Reset Configuration Various device functions are initialized by sampling certain signals during the assertion of HRESET. The values of all these signals are sampled into registers while HRESET is asserted. These inputs are to be pulled high or low by external resistors. During HRESET, all other signal drivers connected to these signals must be in the high-impedance state. Most POR configuration signals have internal pull-up resistors so that if the desired setting is high, there is no need for a pull-up resistor on the board. Other POR configuration signals do not use pull-ups and, therefore, must be pulled high or low. Refer to the MPC8540Integrated Processor Hardware Specifications for proper resistor values to be used for pulling POR configuration signals high or low. This section describes the functions and modes configured by POR configuration signals. Note that many reset configuration settings are accessible to software through the following read-only memory-mapped registers described in Chapter 18, “Global Utilities.” MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 4-11 Reset, Clocking, and Initialization • • • • • • POR PLL status register (PORPLLSR) POR boot mode status register (PORBMSR) POR I/O impedance status and control register (PORIMPSCR) POR device status register (PORDEVSR) POR debug mode status register (PORDBGMSR) General-purpose POR configuration register (GPPORCR)—reports the value on LAD[0:31] during POR (can be used to external system configuration) NOTE In the following tables, the binary value 0b0 represents a signal pulled down to GND and a value of 0b1 represents a signal pulled up to VDD, regardless of the sense of the functional signal name on the signal. 4.4.3.1 System PLL Ratio The system PLL inputs, shown in Table 4-9, establish the clock ratio between the PCI_CLK/SYSCLK input and the platform clock used by the MPC8540. The platform clock, also called the CCB clock, drives the L2 cache, the DDR SDRAM data rate, and the e500 core complex bus (CCB). There is no default value for this PLL ratio; these signals must be pulled to the desired values. Note that the values latched on these signals during POR are accessible in the PORPLLSR (POR PLL status register), as described in Section 18.4.1.1, “POR PLL Status Register (PORPLLSR).” MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 4-12 Freescale Semiconductor Functional Description Table 4-9. CCB Clock PLL Ratio Functional Signals LA[28:31] Reset Configuration Name cfg_sys_pll[0:3] Value (Binary) 0000 0001 No default 0010 0011 0100 0101 0110 0111 1000 1001 1010 1011 1100 1101 1110 1111 CCB Clock : SYSCLK Ratio 16 : 1 Reserved 2:1 3:1 4:1 5:1 6:1 Reserved 8:1 9:1 10 : 1 Reserved 12 : 1 Reserved Reserved Reserved 4.4.3.2 e500 Core PLL Ratio Table 4-10 describes the e500 core clock PLL inputs that program the core PLL and establish the ratio between the e500 core clock and the e500 core complex bus (CCB) clock. There is no default value for this PLL ratio; these signals must be pulled to the desired values. Note that the values latched on these signals during POR are accessible through the memory-mapped PORPLLSR, as described in Section 18.4.1.1, “POR PLL Status Register (PORPLLSR),” and also in the e500 core HID1 register, as described in Section 6.10.2, “Hardware Implementation-Dependent Register 1 (HID1).” Table 4-10. e500 Core Clock PLL Ratios Functional Signals LALE, LGPL2 No default Reset Configuration Name cfg_core_pll[0:1] Value (Binary) 00 01 10 11 e500 Core: CCB ClockRatio 2:1 5 : 2 (2.5:1) 3:1 7 : 2 (3.5:1) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 4-13 Reset, Clocking, and Initialization 4.4.3.3 Boot ROM Location The first instruction executed by the e500 core is always address 0xFFFF_FFFC, which must be a branch to an address within the 4-Kbyte boot page. The MPC8540 defines the default boot ROM address range to be 8 Mbytes at address 0xFF80_0000 to 0xFFFF_FFFF. However, which on-chip peripheral handles these boot ROM accesses can be selected at power up. The boot ROM location inputs, shown in Table 4-11, establish the location of boot ROM. Accesses to the boot vector and the default boot ROM region of the local address map are directed to the interface specified by these inputs. Note that the values latched on these signals during POR are accessible through the memory-mapped PORBMSR (POR boot mode status register) described in Section 18.4.1.2, “POR Boot Mode Status Register (PORBMSR).” Table 4-11. Boot ROM Location Functional Signals TSEC1_TXD[6:4] Reset Configuration Name cfg_rom_loc[0:2] Value (Binary) 000 001 Default (111) 010 011 100 101 110 111 PCI/PCI-X DDR SDRAM Reserved RapidIO Reserved Local bus GPCM—8-bit ROM Local bus GPCM—16-bit ROM Local Bus GPCM—32-bit ROM (default) Meaning See Section 2.1, “Local Memory Map Overview and Example,” for an example memory map that relies on the default boot ROM values. Also, see Section 4.3.1.3.1, “Boot Page Translation Register (BPTR),” for information on translation of the boot page. If enabled, this translation only affects CPU accesses to 0xFFF_Fnnn. 4.4.3.4 Host/Agent Configuration The host/agent reset configuration inputs, shown in Table 4-12, configure the MPC8540 to act as a host or as an agent of a master on another interface. In host mode, the MPC8540 is immediately enabled to master transactions to the RapidIO and PCI interfaces. If the MPC8540 is an agent on the RapidIO or the PCI interfaces, then the MPC8540 is disabled from mastering transactions on that interface until the external host enables it to do so. The external host does this by setting the control registers of the MPC8540’s interfaces appropriately. See details in the PCI and RapidIO programming models described in Chapter 16, “PCI/PCI-X Bus Interface,” and Chapter 17, “RapidIO Interface,” respectively. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 4-14 Freescale Semiconductor Functional Description Note that the values latched on these signals during POR are accessible through the memory-mapped PORBMSR (POR boot mode status register) described in Section 18.4.1.2, “POR Boot Mode Status Register (PORBMSR).” S Table 4-12. Host/Agent Configuration Functional Signals LWE[2:3] Reset Configuration Name cfg_host_agt[0:1] Value (Binary) 00 01 10 11 Meaning MPC8540 acts as an agent of both a PCI/PCI-X and a RapidIO device. MPC8540 acts as an agent of a RapidIO host. MPC8540 acts as an agent of a PCI/PCI-X host. MPC8540 acts as the host processor (default). Default (11) NOTE If the MPC8540 is an agent on a particular interface, and the CPU is not in holdoff mode (as described in Section 4.4.3.5, “CPU Boot Configuration”) then the boot ROM should not be located on the interface where the external host exists, because the MPC8540 is not enabled to master reads onto that interface. 4.4.3.5 CPU Boot Configuration The CPU boot configuration input, shown in Table 4-13, specifies the boot configuration mode. If LA27 is sampled low at reset, the e500 core is prevented from fetching boot code until configuration by an external master is complete. The external master frees the CPU to boot by setting EEBPCR[CPU_EN] in the ECM CCB port configuration register (EEBPCR). See Section 8.2.1.2, “ECM CCB Port Configuration Register (EEBPCR),” for more information. Note that the values latched on these signals during POR are accessible through the memory-mapped PORBMSR (POR boot mode status register) described in Section 18.4.1.2, “POR Boot Mode Status Register (PORBMSR).” Table 4-13. CPU Boot Configuration Functiona Reset Configuration Value l Name (Binary) Signal LA27 Default (1) cfg_cpu_boot 0 1 Meaning CPU boot holdoff mode. The e500 core is prevented from booting until configured by an external master. The e500 core is allowed to boot without waiting for configuration by an external master (default). MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 4-15 Reset, Clocking, and Initialization 4.4.3.6 Boot Sequencer Configuration The boot sequencer configuration options, shown in Table 4-14, allow the boot sequencer to load configuration data from the serial ROM located on the I2C port before the host tries to configure the MPC8540. These options also specify normal or extended I2C addressing modes. See Section 11.4.5, “Boot Sequencer Mode,” for more information on the boot sequencer. Note that the values latched on these signals during POR are accessible through the memory-mapped PORBMSR (POR boot mode status register) described in Section 18.4.1.2, “POR Boot Mode Status Register (PORBMSR).” Table 4-14. Boot Sequencer Configuration Functiona l Signals LGPL3, LGPL5 Reset Configuration Name cfg_boot_seq[0:1] Value (Binary ) 00 01 Reserved Normal I2C addressing mode is used. Boot sequencer is enabled and loads configuration information from a ROM on the I2C interface. A valid ROM must be present. Extended I2C addressing mode is used. Boot sequencer is enabled and loads configuration information from a ROM on the I2C interface. A valid ROM must be present. Boot sequencer is disabled. No I2C ROM is accessed (default). Meaning Default (11) 10 11 NOTE When the boot sequencer is enabled, the processor core will be held in reset and thus prevented from fetching boot code until the boot sequencer has completed its task, regardless of the state of the CPU boot configuration signal described in Section 4.4.3.5, “CPU Boot Configuration.” 4.4.3.7 TSEC Width The TSEC width input, shown in Table 4-15, selects standard versus reduced width for both of the three-speed Ethernet controller interfaces. Note that the value latched on this signal during POR is accessible through the memory-mapped PORDEVSR (POR device status register) described in Section 18.4.1.4, “POR Device Status Register (PORDEVSR).” MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 4-16 Freescale Semiconductor Functional Description Table 4-15. TSEC Width Configuration Functiona l Signal EC_MDC Default (1) Reset Configuration Name cfg_tsec_reduce Value (Binary) 0 1 Meaning Ethernet interfaces operate in reduced mode, either RTBI or RGMII, using only four transmit data signals and four receive data signals. Ethernet interfaces operate in their standard TBI or GMII modes using eight transmit data signals and eight receive data signals (default). NOTE While the width of both interfaces is controlled by this one configuration input, the protocol (TBI or GMII) used by each is separately controlled with other configuration inputs described in Section 4.4.3.8, “TSEC1 Protocol,” and Section 4.4.3.9, “TSEC2 Protocol.” 4.4.3.8 TSEC1 Protocol The TSEC1 protocol input, shown in Table 4-16, selects the protocol (GMII or TBI) used by the TSEC1 controller. Note that the value latched on this signal during POR is accessible through the memory-mapped PORDEVSR (POR device status register) described in Section 18.4.1.4, “POR Device Status Register (PORDEVSR).” Table 4-16. TSEC1 Protocol Configuration Functional Signal TSEC1_TXD 7 Default (1) Reset Configuration Name cfg_tsec1 Value (Binary ) 0 Meaning The TSEC1 controller operates using the GMII protocol (or RGMII if configured in reduced mode as described in Section 4.4.3.7, “TSEC Width”). The TSEC1 controller operates using the TBI protocol (or RTBI if configured in reduced mode as described in Section 4.4.3.7, “TSEC Width”) (default). 1 4.4.3.9 TSEC2 Protocol The TSEC2 protocol input, shown in Table 4-17, selects the protocol (GMII or TBI) used by the TSEC2 controller. Note that the value latched on this signal during POR is accessible through the memory-mapped PORDEVSR (POR device status register) described in Section 18.4.1.4, “POR Device Status Register (PORDEVSR).” MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 4-17 Reset, Clocking, and Initialization Table 4-17. TSEC2 Protocol Configuration Functional Signal TSEC2_TXD 7 Default (1) Value Reset Configuration (Binary ) Name cfg_tsec2 0 Meaning The TSEC2 controller operates using the GMII protocol (or RGMII if configured in reduced mode as described in Section 4.4.3.7, “TSEC Width”). The TSEC2 controller operates using the TBI protocol (or RTBI if configured in reduced mode as described in Section 4.4.3.7, “TSEC Width”) (default). 1 4.4.3.10 RapidIO Transmit Clock Source The RapidIO transmit clock source inputs, shown in Table 4-18, specify the source for the RapidIO transmit clock. See Section 4.4.4.2, “RapidIO Clocks,” for more information. Note that the value latched on this signal during POR is accessible through the memory-mapped PORDEVSR (POR device status register) described in Section 18.4.1.4, “POR Device Status Register (PORDEVSR).” Table 4-18. RapidIO Transmit Clock Source Functional Signals LGPL0, LGPL1 Reset Configuration Name cfg_rio_clk[0:1] Value (Binary ) 00 01 10 11 Reserved The RapidIO receive clock is the source of the transmit clock. The RapidIO transmit clock inputs (RIO_TX_CLK_IN and RIO_TX_CLK) are the source of the transmit clock. The CCB clock is the source of the transmit clock (default). Meaning Default (11) 4.4.3.11 RapidIO Device ID The RapidIO device ID inputs, shown in Table 4-19, specify the 3 lower-order bits of the device ID of the MPC8540 as used by RapidIO hosts. Note that the 5 high-order RapidIO device ID bits cannot be set via POR configuration inputs. They may be initialized via the boot sequencer or by the processor from boot ROM or by the RapidIO discovery process. If configured as a RapidIO host, the upper order device ID bits default to zeros. If configured as a RapidIO agent, the upper order device ID bits default to ones. Unconnected cfg_dev_IDn inputs default to 1s regardless of the host/agent mode configuration. Note that the value latched on this signal at POR is accessible through the memory-mapped PORDEVSR described in Section 18.4.1.4, “POR Device Status Register (PORDEVSR).” MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 4-18 Freescale Semiconductor Functional Description Table 4-19. RapidIO Device ID Functional Signals TSEC2_TXD2 TSEC2_TXD3 TSEC2_TXD4 Reset Configuration Name cfg_dev_ID5 cfg_dev_ID6 cfg_dev_ID7 Meaning Device ID used for RapidIO hosts Device ID used for RapidIO hosts Device ID used for RapidIO hosts 4.4.3.12 PCI Width Configuration The PCI width configuration input, shown in Table 4-20, configures the PCI/PCI-X interface to 32- or 64-bit extended mode of operation. Note that the value latched on this signal during POR is accessible through the PORDEVSR described in Section 18.4.1.4, “POR Device Status Register (PORDEVSR).” Table 4-20. PCI-32 Configuration Value Functional Reset Configuration (Binary Signal Name ) PCI_REQ6 4 Default (1) cfg_pci_width 0 1 Meaning The PCI/PCI-X interface operates as a 64-bit interface. The PCI/PCI-X interface operates as a 32-bit interface (default). 4.4.3.13 PCI I/O Impedance The PCI I/O impedance input, shown in Table 4-21, selects the impedance of the PCI I/O drivers. Note that the values latched on these signals during POR are accessible through PORIMPSCR, described in Section 18.4.1.3, “POR I/O Impedance Status and Control Register (PORIMPSCR).” Table 4-21. PCI I/O Impedance Functional Signal PCI_GNT1 Default (1) Reset Configuration Name cfg_pci_impd Value (Binary) 0 1 Meaning 25-Ω I/O drivers are used on the PCI interface. 42-Ω I/O drivers are used on the PCI interface (default). 4.4.3.14 PCI Arbiter Configuration The PCI arbiter configuration input, shown in Table 4-22, enables the on-chip PCI/PCI-X arbiter. Note that the value latched on this signal during POR is accessible through the PORDEVSR described in Section 18.4.1.4, “POR Device Status Register (PORDEVSR).” MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 4-19 Reset, Clocking, and Initialization Table 4-22. PCI Arbiter Configuration Functiona l Signal PCI_GNT2 Default (1) Reset Configuration Name cfg_pci_arbiter Value (Binary ) 0 1 Meaning The on-chip PCI/PCI-X arbiter is disabled. External arbitration is required. The on-chip PCI/PCI-X arbiter is enabled (default). 4.4.3.15 PCI Debug Configuration The PCI debug configuration input, shown in Table 4-23, enables PCI/PCI-X debug mode. In this mode, source ID information is driven onto the highest order address bits PCI_AD[62:58] during the bus command phase (PCI) or attribute phase (PCI-X). Note that the value latched on this signal during POR is accessible through the PORDBGMSR described in Section 18.4.1.5, “POR Debug Mode Status Register (PORDBGMSR).” Table 4-23. PCI Debug Configuration Functiona Reset Configuration Value l Signal Name (Binary) PCI_GNT3 Default (1) 1 cfg_pci_debug 0 Meaning PCI debug is enabled. Source ID information is driven onto the highest order address bits, PCI_AD[62:58], during the bus command phase (PCI) or attribute phase (PCI-X). PCI operates in normal mode (default) 4.4.3.16 PCI-X Configuration The PCI-X configuration input, shown in Table 4-24, configures PCI or PCI-X mode on the PCI/PCI-X port. Note that this input does not support the three states of the PCIXCAP input defined in the PCI-X specification. It is sampled as either a 1 or a 0 during reset only. Note that the value latched on this signal during POR is accessible through the memory-mapped PORDEVSR, described in Section 18.4.1.4, “POR Device Status Register (PORDEVSR).” Table 4-24. PCI/PCI-X Configuration Functional Signal PCI_GNT4 Default (1) Reset Configuration Name cfg_pci_mode Value (Binary) 0 1 Meaning PCI-X mode PCI mode (default) 4.4.3.17 Memory Debug Configuration The memory debug configuration input, shown in Table 4-25, selects which debug outputs (DDR or LBC memory controller) are driven onto the MSRCID and MDVAL debug signals. Note that the value latched on this signal during POR is accessible through the memory-mapped MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 4-20 Freescale Semiconductor Functional Description PORDBGMSR (POR debug mode register) described in Section 18.4.1.5, “POR Debug Mode Status Register (PORDBGMSR).” Table 4-25. Memory Debug Configuration Functional Reset Configuration Value Signal Name (Binary) MSRCID0 Default (1) cfg_mem_debug 0 1 Meaning Debug information from the local bus controller (LBC) is driven on the MSRCID and MDVAL signals. Debug information from the DDR SDRAM controller is driven on the MSRCID and MDVAL signals (default). 4.4.3.18 DDR Debug Configuration The DDR debug configuration input, shown in Table 4-26, enables a DDR memory controller debug mode in which the DDR SDRAM source ID field and data valid strobe are driven onto the ECC pins. ECC checking and generation are disabled in this case. ECC signals driven from the SDRAMs must be electrically disconnected from the ECC I/O pins of the MPC8540 in this mode. Note that the value latched on this signal during POR is accessible through the memory-mapped PORDBGMSR (POR debug mode register) described in Section 18.4.1.5, “POR Debug Mode Status Register (PORDBGMSR).” Table 4-26. DDR Debug Configuration Functional Reset Configuration Value Signal Name (Binary) MSRCID1 Default (1) cfg_ddr_debug 0 1 Meaning Debug information is driven on the ECC pins instead of normal ECC I/O. ECC signals from memory devices must be disconnected. Debug information is not driven on ECC pins. ECC pins function in their normal mode (default). 4.4.3.19 PCI/PCI-X Output Hold Configuration The PCI output hold configuration inputs configure the output hold times for the PCI or PCI-X output drivers. The meanings are different for PCI and PCI-X because the required hold times are different in the two specifications. In either mode, the default value will meet the hold times required by the respective specification. Hold times are adjusted by adding or subtracting buffer delays to the intrinsic delay of the output driver. Refer to the MPC8540 Integrated Processor Hardware Specifications for specific timing information. Table 4-27 shows the hold time configurations for PCI mode, and Table 4-28 shows the hold time configurations for PCI-X mode. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 4-21 Reset, Clocking, and Initialization Table 4-27. PCI Output Hold Configuration (cfg_pci_mode = 1) Value Functional Reset Configuration (Binary Signals Name ) LWE[0:1] cfg_pci_hold[0:1] 11 10 Default (11) 01 00 Meaning Two added buffer delays—required to meet 2-ns hold time requirement Three added buffer delays (default + 1) Zero added buffer delays (default + 2 mod 4) One added buffer delay (default + 3 mod 4) Table 4-28. PCI-X Output Hold Configuration (cfg_pci_mode = 0) Functional Signals LWE[0:1] Reset Configuration Name cfg_pci_hold[0:1] Value (Binary ) 11 10 Default (11) 01 00 Meaning Zero added buffer delays—meets 0.7 ns hold time requirement One added buffer delay (default + 1) Two added buffer delays (default + 2) Three added buffer delays (default + 3) 4.4.3.20 Local Bus Output Hold Configuration The LBC output hold configuration inputs, shown in Table 4-29, configure the output hold times for the local bus interface output drivers. The default values are designed to meet the specified AC timing requirements for the local bus. Hold times are adjusted by adding buffer delays to the intrinsic delay of the output driver. Refer to the MPC8540 Integrated Processor Hardware Specifications for specific timing information. Note that for three of the POR configuration input settings, there is a minimum of one buffer of output hold delay between the LALE and LAD[0:31] signals. Table 4-29. Local Bus Output Hold Configuration Functional Signals TSEC2_TXD[6:5 ] Reset Configuration Name cfg_lb_hold[0:1] Value (Binary ) 11 Meaning One added buffer delay (default) (zero added buffer delays for LALE) Two added buffer delays (default + 1) (one added buffer delay for LALE) Three added buffer delays (default + 2) (one added buffer delay for LALE) Zero added buffer delays (zero added buffer delays for LALE) 10 Default (11) 01 00 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 4-22 Freescale Semiconductor Functional Description 4.4.3.21 General-Purpose POR Configuration The LBC address/data bus inputs, shown in Table 4-30, configure the value of the general-purpose POR configuration register defined in Section 18.4.1.6, “General-Purpose POR Configuration Register (GPPORCR).” This register is intended to facilitate POR configuration of user systems. A value placed on LAD[0:31] during POR is captured and stored (read only) in the GPPORCR. Software can then use this value to inform the operating system about initial system configuration. Typical interpretations include circuit board type, board ID number, or a list of available peripherals. Table 4-30. General-Purpose POR Configuration Functional Signals LAD[0:31] No default Reset Configuration Name cfg_gpporcr Value (Binary ) xx Meaning General-purpose POR configuration vector to be placed in GPPORCR 4.4.4 Clocking The following paragraphs describe the clocking within the MPC8540 device. 4.4.4.1 System Clock/PCI Clock The MPC8540 takes a single input clock, SYSCLK, as its primary clock source for the e500 core and all of the devices and interfaces that operate synchronously with the core. As shown in Figure 4-6, the SYSCLK input (frequency) is multiplied up using a phase lock loop (PLL) to create the core complex bus (CCB) clock (also called the platform clock). The CCB clock is used by virtually all of the synchronous system logic, including the L2 cache, and other internal blocks such as the DMA and interrupt controller. The CCB clock also feeds the PLL in the e500 core and the DLLs that create clocks for the DDR SDRAM and local bus memory controllers. Note that the divide-by-two CCB clock divider and the divide-by-n CCB clock divider, shown in Figure 4-6, are located in the DDR and local bus blocks, respectively. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 4-23 Reset, Clocking, and Initialization MPC8540 e500 Core cfg_core_pll[0:1] 2 Core PLL core_clk MSYNC_IN ÷2 cfg_sys_pll[0:3] 4 Device PLL CCB_clk DLL MSYNC_OUT 6 6 MCK[0:5] MCK[0:5] LSYNC_IN DDR Controller SYSCLK/ PCI_CLK ÷n CCB_clk to Rest of the Device LSYNC_OUT DLL LCLK0 LCLK1 LCLK2 LBC Figure 4-6. Clock Subsystem Block Diagram When the PCI/PCI-X interface is being used, SYSCLK also functions as the PCI_CLK signal. Note that this is true both when the MPC8540 is in agent mode and host mode. The MPC8540 does not provide a separate PCI_CLK output in host mode. 4.4.4.2 RapidIO Clocks As shown in Figure 4-7, the RapidIO transmit clocks (RIO_TCLK and RIO_TCLK) can be selected from one of three sources. If the desired RapidIO clock is sourced by the device to which the MPC8540 RapidIO port is connected, then the transmit clock can be derived from the receive clock (RIO_RCLK and RIO_RCLK). MPC8540 RIO_RCK RIO_RCK CCB_clk Clock Synthesis Chip RIO_TX_CLK_IN RIO_TCK RIO_TCK RIO_TX_CLK_IN cfg_rio_clk[0:1] 2 Figure 4-7. RapidIO Transmit Clock Options MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 4-24 Freescale Semiconductor Functional Description If the MPC8540 is the initial master of the RapidIO clock, then it can derive the RapidIO transmit clock from either its internal platform (CCB) clock or from a dedicated LVDS clock input, RIO_TX_CLK_IN / RIO_TX_CLK_IN. 4.4.4.3 Ethernet Clocks The Ethernet blocks operate asynchronously with respect to the rest of the device. These blocks use receive and transmit clocks supplied by their respective PHY chips, plus a 125-MHz clock input for gigabit protocols. Data transfers are synchronized to the CCB clock internally. 4.4.4.4 Real Time Clock As shown in Figure 4-8, the real time clock (RTC) input can optionally be used to clock the e500 core timer facilities. RTC can also be used (optionally) by the MPC8540 programmable interrupt controller (PIC) global timer facilities. The RTC is separate from the e500 core clock and is intended to support relatively low frequency timing applications. The RTC frequency range is specified in the MPC8540 Integrated Processor Hardware Specifications, but the maximum value should not exceed one-eighth of the core frequency. Before being distributed to the core time base, RTC is sampled and synchronized with the CCB clock. The clock source for the core time base is specified by two fields in HID0: time base enable (TBEN), and select time base clock (SEL_TBCLK). If the time base is enabled, (HID0[TBEN] is set), the clock source is determined as follows: • • HID0[SEL_TBCLK] = 0, the time base is updated every 8 CCB clocks HID0[SEL_TBCLK] = 1, the time base is updated on the rising edge of RTC The default source of the time base is the CCB clock divided by eight. For more details, see the section “Hardware Implementation-Dependent Register (HID0),” in the “PowerPC™ e500 Core Complex Reference Manual.” Section 10.3.2.6, “Timer Control Register (TCR),” provides additional information on the use of the RTC signal to clock the global timers in the PIC unit. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 4-25 Reset, Clocking, and Initialization e500 Core HID0 TBEN SEL_TBCLK Core Time Base (Incrementer) TBU 32 63 32 TBL 63 RTC (Sampled and Synchronized) Watchdog timer events based on one of the 64 TB bits selected by concatenating TCR[WPEXT] with the EIS-defined TCR[WP] (WPEXT||WP). Fixed-interval timer events based on one of the 64 TB bits selected by concatenating TCR[FPEXT] with the EIS-defined TCR[FP] (FPEXT||FP). • • • ÷8 Core Timer Facilities Clock • • • CCB Clock (Decrementer) DEC Decrementer Event (0 ⇒ 1 Detect) 32 Auto-Reload DECAR 63 Note: The logic circuits shown depict functional relationships only; they do not represent physical implementation details. Figure 4-8. RTC and Core Timer Facilities Clocking Options MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 4-26 Freescale Semiconductor Part II e500 Core Complex and L2 Cache This part describes the many features of the MPC8540 core processor at an overview level and the interaction between the core complex and the L2 cache. The following chapters are included: • • • Chapter 5, “Core Complex Overview,” provides an overview of the e500 core processor and the L1 caches and MMU that, together with the core, comprise the core complex. Chapter 6, “Core Register Summary,” provides a listing of the e500 registers in reference form. Chapter 7, “L2 Look-Aside Cache/SRAM,” describes the L2 cache of the MPC8540. Note that the L2 cache can also be addressed directly as memory-mapped SRAM. The e500 processor core is a low-power implementation of the family of reduced instruction set computing (RISC) embedded processors that implement the PowerPC architecture. This part provides additional information about the Book E architecture as it relates specifically to the e500 core complex and specific details on how its registers are accessed. The e500 core complex interacts with the L2 cache through the core complex bus (CCB). MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor II-1 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 II-2 Freescale Semiconductor Chapter 5 Core Complex Overview This chapter provides an overview of the PowerPC e500 microprocessor core as it is implemented on the MPC8540. This chapter includes the following: • • • • • • • • An overview of the Book E version of the PowerPC architecture features as implemented in this core and a summary of the core feature set A summary of the instruction pipeline and flow An overview of the programming model An overview of interrupts and exception handling A description of the memory management architecture High-level details of the e500 core memory and coherency model A brief description of the core complex bus (CCB) A summary of the Book E architecture compatibility and migration from the original version of the PowerPC architecture as it is defined by Apple, IBM, and Motorola (referred to as the AIM version of the PowerPC architecture) Specific details about the e500 are provided in the PowerPC e500 Core Complex Reference Manual (Freescale Semiconductor Document ID No. E500CORERM). The e500 core provides features that the integrated device may not implement or may implement in a more specific way. These differences are summarized in Section 5.14, “PowerQUICC III Implementation Details.” 5.1 Overview The e500 processor core is a low-power implementation of the family of reduced instruction set computing (RISC) embedded processors that implement the Book E definition of the PowerPC architecture. The e500 is a 32-bit implementation of the Book E architecture using the lower words in the 64-bit general-purpose registers (GPRs). Figure 5-1 is a block diagram of the processor core complex that shows how the functional units operate independently and in parallel. Note that this conceptual diagram does not attempt to show how these features are physically implemented. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 5-1 5-2 Instruction Unit Fetch Stages Instruction Queue (12 instructions) Branch Prediction Unit 128-Bit (4 Instructions) CTR LR BTB 512 Entry Core Complex Overview Memory Unit 32-Kbyte I Cache Tags L1 Instruction MMU 4-Entry I-L1VSP 64-Entry I-L1TLB4K Additional Features • Time Base Counter/Decrementer • Clock Multiplier • JTAG/COP Interface • Power Management • Performance Monitor Program order is maintained by passing instructions from the IQ to the CQ at dispatch. Branch Issue Queue (BIQ) Two Instruction Dispatch (1 BIQ, 2 GIQ) Two instruction issue to GIQ per clock Each execution unit can accept one instruction per cycle. General Issue Queue (GIQ) One instruction issue to BIQ per clock MAS Registers L2 MMUs Unified 16-Entry TLB Array (TLB1) 256-Entry TLB Array (TLB0) Reservation Station GPR File Rename Buffers (14) 32-/ 64-Bit GPR Operand Bus Completion Bus Reservation Station Simple Unit 1 (32/64 bit) Multiple Unit (64/32 bit) Simple Unit 2 (32 bit) Reservation Station Reservation Station Reservation Station Load/Store Unit (64/32 bit) L1 Data MMU 4-Entry D-L1VSP 64-Entry D-L1TLB4K Branch Unit Condition Register CR Field Rename Buffers (14) Figure 5-1. e500 Core Complex Block Diagram CRF Bus L1 Store Queue Load Miss Queue Tags 32-Kbyte D Cache MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Data Line Data Write Fill Buffer Buffer Completion Queue (14 Entry) Maximum Two Instructions Retire per Cycle Instruction Line Fill Buffer Core Interface Unit Core Complex Bus Freescale Semiconductor Overview Book E allows processors to provide auxiliary processing units (APUs), which are extensions to the architecture that can perform computational or system management functions. One of these on the e500 is the signal processing engine APU (SPE APU), which includes a suite of vector instructions that use the upper and lower halves of the GPRs as a single two-element operand. Most APUs implemented on the e500 are defined by the Freescale Semiconductor Book E implementation standards (EIS). 5.1.1 Upward Compatibility The e500 provides 32-bit effective addresses and integer data types of 8, 16, and 32 bits, and two-element 64-bit data types for the embedded vector floating-point APU which provides scalar and vector single-precision instructions that operate on operands comprised of two 32-bit elements and the SPE APU, which provides an extensive instruction set for 64-bit vector integer and fractional operations. The embedded scalar floating-point APU provides 32-bit single-precision instructions. NOTE The SPE APU and embedded floating-point APU functionality is implemented in all PowerQUICC III devices. However, these instructions will not be supported in devices subsequent to PowerQUICC III. Freescale Semiconductor strongly recommends that use of these instructions be confined to libraries and device drivers. Customer software that uses SPE or embedded floating-point APU instructions at the assembly level or that uses SPE intrinsics will require rewriting for upward compatibility with next-generation PowerQUICC devices. Freescale Semiconductor offers a libmoto_e500 library that uses SPE and embedded floating-point APU instructions. Freescale will also provide libraries to support next-generation PowerQUICC devices. 5.1.2 Core Complex Summary The core complex is a superscalar processor that can issue two instructions and complete two instructions per clock cycle. Instructions complete in order, but can execute out of order. Execution results are available to subsequent instructions through the rename buffers, but those results are recorded into architected registers in program order, maintaining a precise exception model. All arithmetic instructions that execute in the core operate on data in the GPRs. Although the GPRs are 64 bits wide, only SPE APU and embedded vector floating-point instructions operate on the upper word of the GPRs; the upper 32 bits are not affected by 32-bit instructions. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 5-3 Core Complex Overview The processor core integrates two simple instruction units (SU1, SU2), a multiple-cycle instruction unit (MU), a branch unit (BU), and a load/store unit (LSU). The LSU and SU2 support 64- and 32-bit instructions. The ability to execute five instructions in parallel and the use of simple instructions with short execution times yield high efficiency and throughput. Most integer instructions execute in one clock cycle. A series of independent vector floating-point add instructions can be issued and completed with a throughput of one instruction per cycle. The core complex includes independent, on-chip, 32-Kbyte, eight-way set-associative, physically addressed caches for instructions and data. It also includes on-chip first-level instruction and data memory management units (MMUs) and an on-chip second-level unified MMU. • The first-level MMUs contain two four-entry, fully-associative instruction and data translation lookaside buffer (TLB) arrays that provide support for demand-paged virtual memory address translation and variable-sized pages. They also contain two 64-entry, four-way set-associative instruction and data TLB arrays that support 4-Kbyte pages. These arrays are maintained completely by the hardware with a true least-recently-used (LRU) algorithm. The second-level MMU contains a 16-entry, fully-associative unified (instruction and data) TLB array that provides support for variable-sized pages and a 256-entry, two-way set-associative unified TLB for 4-Kbyte page size support. These TLBs are maintained completely by the software. • The core complex allows cache-line-based user-mode locks on the contents in either the instruction or data cache. This provides embedded applications with the capability for locking interrupt routines or other important (time-sensitive) instruction sequences into the instruction cache. It also allows data to be locked into the data cache, which supports deterministic execution time. The core complex supports a high-speed on-chip internal bus with data tagging called the core complex bus (CCB). The CCB has two general purpose read data buses, one write data bus, data parity bits, data tag bits, an address bus, and address attribute bits. The processor core complex supports out-of-order reads, in-order writes, and one level of pipelining for addresses with address-retry responses. It can also support single-beat and burst data transfers for memory accesses and memory-mapped I/O operations. 5.2 e500 Processor and System Version Numbers Table 5-1 matches the revision code in the processor version register (PVR) and the system version register (SVR) to the revision level marked on the device. These registers can be accessed as SPRs through the e500 core (see Chapter 6, “Core Register Summary”) or as memory-mapped registers defined by the integrated device (see Section 18.4.1, “Register Descriptions”). MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 5-4 Freescale Semiconductor Features Table 5-1. Revision Level-to-Device Marking Cross-Reference MPC8540 Revision 1.0 2.0 e500 Core Revision 1.0 2.0 Processor Version Register (PVR) 0x8020_0010 0x8021_0010 System Version Register (SVR) 0x8030_0010 0x8030_0020 5.3 • • Features Implements Book E 32-bit architecture Auxiliary processing units The branch target buffer (BTB) locking APU is specific to the e500. The BTB locking APU gives the user the ability to lock, unlock, and invalidate BTB entries. The EIS defines the following APUs: — Integer select. This APU consists of the Integer Select instruction, isel, which is a conditional register move that helps eliminate conditional branches, decreases latency, and reduces the code footprint. — Performance monitor. The performance monitor facility provides the ability to monitor and count predefined events such as processor clocks, misses in the instruction cache or data cache, types of instructions decoded, or mispredicted branches. The count of such events can be used to trigger the performance monitor exception. Additional performance monitor registers (PMRs) similar to SPRs are used to configure and track performance monitor operations. These registers are accessed with the Move to PMR and Move from PMR instructions (mtpmr and mfpmr). See Section 5.12, “Performance Monitoring.” — Cache line lock and unlock. This APU allows instructions and data to be locked into their respective caches on a cache line basis. Locking is performed by a set of touch-and-lock set instructions. This functionality can be enabled for user mode by setting MSR[UCLE]. The APU also provides resources for detecting and handling overlocking conditions. — Machine check. The machine check interrupt is treated as a separate level of interrupt. It uses its own save and restore registers (MCSRR0 and MCSRR1) and Return from Machine Check Interrupt (rfmci) instruction. See Section 5.8, “Interrupts and Exception Handling.” — Single-precision embedded scalar and vector floating-point APUs. These instructions are listed in the EREF: A Reference for Freescale Semiconductor Book E and the e500 Core. Key features of the e500 are summarized as follows: MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 5-5 Core Complex Overview • • — Signal processing engine APU (SPE APU). Note that the SPE is not a separate unit; SPE computational and logical instructions are executed in the simple and multiple-cycle units used by all other computational and logical instructions, and 64-bit loads and stores are executed in the common LSU. Figure 5-1 shows how execution logic for SU1, the MU, and the LSU is replicated to support operations on the upper halves of the GPRs. The e500 register set is modified as follows: – GPRs are widened to 64 bits to support 64-bit load, store, and merge operations. Note that the upper 32 bits are affected only by 64-bit instructions. – A 64-bit accumulator (ACC) has been added. – The signal processing and embedded floating-point status and control register (SPEFSCR) provides interrupt control and status for SPE and embedded floating-point instructions. These registers are shown in Figure 5-6. SPE instructions are grouped as follows: – Single-cycle integer add and subtract with the same latencies for SPE APU operations as for the 32-bit equivalent – Single-cycle logical operations – Single-cycle shift and rotate – Four-cycle integer pipelined multiplies – 4-, 11-, 19-, and 35-cycle integer divides – If rA or rB is zero, a floating-point divide takes 4 cycles. All other cases take 29 cycles. – 4-cycle SIMD pipelined multiply-accumulate (MAC) – 64-bit accumulator for no-stall MAC operations – 64-bit loads and stores – 64-bit merge instructions Cache structure—Separate 32-Kbyte, 32-byte line, 8-way set-associative level 1 instruction and data caches — 1.5-cycle cache array access, 3-cycle load-to-use latency — Pseudo-LRU (PLRU) replacement algorithm — Copy-back data cache that can function as a write-through cache on a page-by-page basis — Supports all Book E memory coherency modes — Supports EIS-defined cache-locking instructions, as listed in Table 5-3 Dual-issue superscalar control — Two-instructions-per-clock peak issue rate MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 5-6 Freescale Semiconductor Features • • • • • • • — Precise exception handling Decode unit — 12-entry instruction queue (IQ) — Full hardware detection of interlocks — Decodes as many as two instructions per cycle — Decode serialization control — Register dependency resolution and renaming Branch prediction unit (BPU) — Dynamic branch prediction using a 512-entry, four-way set-associative branch target buffer (BTB) supported by the e500 BTB instructions listed in Table 5-5. — Branch prediction is handled in the fetch stages. Completion unit — As many as 14 instructions allowed in 14-entry completion queue (CQ) — In-order retirement of as many as two instructions per cycle — Completion and refetch serialization control — Synchronization for all instruction flow changes—interrupts, mispredicted branches, context-synchronizing instructions Issue queues — Two-entry branch instruction issue queue (BIQ) — Four-entry general instruction issue queue (GIQ) Branch unit—The branch unit (BU) is an execution unit and is distinct from the BPU. It executes (resolves) all branch and CR logical instructions. Two simple units (SU1 and SU2) — Add and subtract — Shift and rotate — Logical operations — Support for 64-bit signal processing engine APU instructions in SU1 Multiple-cycle unit (MU)—The MU is shown in Figure 5-2. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 5-7 Core Complex Overview From GIQ0 or GIQ1 Reservation Station Upper Lower MU-1 Divide Bypass Path MU-2 MU-3 Divide Postdivide M U-4 Figure 5-2. Four-Stage MU Pipeline, Showing Divide Bypass • The MU has the following features: — Four-cycle latency for all multiplication, including SPE integer and fractional multiply instructions and embedded scalar and vector floating-point multiply instructions — Variable-latency divide: 4, 11, 19, and 35 cycles for all integer divide instructions. If rA or rB is zero, floating-point divide instructions take 4 cycles. All others take 29. Note that although most divide instructions take more than 4 cycles to execute, the MU allows subsequent multiply instructions to execute through all four MU stages in parallel with the divide. — Four-cycle floating-point add and subtract Load/store unit (LSU) is shown in Figure 5-3. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 5-8 Freescale Semiconductor Features Reservation Station Load/Store Unit (64-/32-Bit) Three-Stage Pipeline To completion queue To GPR operand bus To GPRs Queues and Buffers L1 Store Load Queue Miss Queue Data Line Fill Buffer Data Write Buffer To data cache To Core Interface Unit Figure 5-3. Three-Stage Load/Store Unit • • The LSU has the following features: — 3-cycle load latency — Fully pipelined — The load miss queue allows up to four load misses before stalling. — Load hits can continue to be serviced when the load miss queue is full. — Seven-entry L1 store queue allows full pipelining of stores — The three-entry data line fill buffer is used for loads and cacheable stores. Stores are allocated here so loads can access data from the store immediately. — The data write buffer contains three entries: one dedicated for snoop pushes, one dedicated for cast outs, and one that can be used for snoop pushes or cast outs. Cache coherency — Supports four-state cache coherency: modified-exclusive, exclusive, shared, and invalid (MESI) — Bus support for hardware-enforced coherency (bus snooping) Core complex bus (CCB)—internal bus — High-speed, on-chip local bus with data tagging — 32-bit address bus — Address protocol with address pipelining and retry/copyback derived from bus used by previous generations of PowerPC processors (referred to as the 60x bus) — Two general-purpose read data buses and one write data bus MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 5-9 Core Complex Overview • • • Extended exception handling — Supports Book E interrupt model – Less than 10-cycle interrupt latency – Interrupt vector prefix register (IVPR) – Vector offset registers (IVORs) 0–15 as defined in Book E, plus e500-defined IVORs 32–35 – Exception syndrome register (ESR) – Book E-defined preempting critical interrupt, including critical interrupt status registers (CSRR0 and CSRR1) and an rfci instruction — e500-specific interrupts not defined in Book E architecture – Machine-check APU – SPE APU unavailable exception – Floating-point data exception – Floating-point round exception – Performance monitor Memory management unit (MMU) — 32-bit effective address translated to 32-bit real address (using a 41-bit interim virtual address) — TLB entries for variable (4-Kbyte–256-Mbyte) and fixed-size (4-Kbyte) pages — Data L1 MMU – 4-entry, fully associative TLB array for variable-sized pages – 64-entry, 4-way set-associative TLB for 4-Kbyte pages — Instruction L1 MMU – 4-entry, fully associative TLB array for variable-sized pages – 64-entry, 4-way set-associative TLB for 4-Kbyte pages — Unified L2 MMU – 16-entry, fully associative TLB array for variable-sized pages – A 256-entry, 2-way set-associative unified (for instruction and data accesses) L2 TLB array (TLB0) supports only 4-Kbyte pages — Software reload for TLBs — Virtual memory support for as much as 4 Gbytes (232) of effective address space — Real memory support for as much as 4 Gbytes (232) of physical memory — Support for big-endian and true little-endian memory on a per-page basis Power management MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 5-10 Freescale Semiconductor Features • • — Low-power, 1.5-V design — Internal clock multipliers ranging from 1 to 8 times the bus clock, including integer and half-mode multipliers. The MPC8540 supports multipliers of 2, 2.5, 3, and 3.5. — Power-saving modes: core-halted and core-stopped — Dynamic power management of execution units, caches, and MMUs — NAP, DOZE, and SLEEP bits in HID0 can be used to assert nap, doze, and sleep output signals to initiate power-saving modes at the integrated device level. Testability — LSSD scan design — JTAG interface — ESP support — ABIST for arrays — LBIST Reliability and serviceability — Parity checking on caches — Parity checking on e500 local bus MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 5-11 Core Complex Overview 5.4 • Instruction Set The Book E instruction set for 32-bit implementations. This is composed primarily of the user-level instructions defined by the PowerPC user instruction set architecture (UISA). The e500 does not include Book E floating-point, load string, or store string instructions. The e500 supports the following implementation-specific instructions: — Integer select APU. This APU consists of the Integer Select instruction (isel), which functions as an if-then-else statement that selects between two source registers by comparison to a CR bit. This instruction eliminates conditional branches, decreases band latency, and reduces the code footprint. — Performance monitor APU. Table 5-2 lists performance monitor APU instructions. Table 5-2. Performance Monitor APU Instructions Name Move from Performance Monitor Register Move to Performance Monitor Register Mnemonic mfpmr mtpmr Syntax rD,PMRN PMRN,rS The e500 implements the following instructions: • — Cache line lock and unlock APU. The cache block lock and unlock APU consists of the instructions described in Table 5-3. Table 5-3. Cache Block Lock and Unlock APU Instructions Name Data Cache Block Lock Clear Data Cache Block Touch and Lock Set Data Cache Block Touch for Store and Lock Set Instruction Cache Block Lock Clear Instruction Cache Block Touch and Lock Set Mnemonic dcblc dcbtls dcbtstls icblc icbtls Syntax CT, rA, rB CT, rA, rB CT, rA, rB CT, rA, rB CT, rA, rB — Machine check APU. This APU defines the Return from Machine Check Interrupt instruction (rfmci). — SPE APU vector instructions. New vector instructions are defined that view the 64-bit GPRs as being composed of a vector of two 32-bit elements (some of the instructions also read or write 16-bit elements). Some scalar instructions are defined for DSP that produce a 64-bit scalar result. — The embedded floating-point APUs provide single-precision scalar and vector floating-point instructions. Scalar floating-point instructions use only the lower 32 bits of the GPRs for single-precision floating-point calculations. Table 5-4 lists embedded floating-point instructions. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 5-12 Freescale Semiconductor Instruction Set Table 5-4. Scalar and Vector Embedded Floating-Point APU Instructions Mnemonic Instruction Scalar Convert Floating-Point from Signed Fraction Convert Floating-Point from Signed Integer Convert Floating-Point from Unsigned Fraction Convert Floating-Point from Unsigned Integer Convert Floating-Point to Signed Fraction Convert Floating-Point to Signed Integer Convert Floating-Point to Signed Integer with Round toward Zero Convert Floating-Point to Unsigned Fraction Convert Floating-Point to Unsigned Integer Convert Floating-Point to Unsigned Integer with Round toward Zero Floating-Point Absolute Value Floating-Point Add Floating-Point Compare Equal Floating-Point Compare Greater Than Floating-Point Compare Less Than Floating-Point Divide Floating-Point Multiply Floating-Point Negate Floating-Point Negative Absolute Value Floating-Point Subtract Floating-Point Test Equal Floating-Point Test Greater Than Floating-Point Test Less Than efscfsf efscfsi efscfuf efscfui efsctsf efsctsi efsctsiz efsctuf efsctui efsctuiz efsabs efsadd efscmpeq efscmpgt efscmplt efsdiv efsmul efsneg efsnabs efssub efststeq efststgt efststlt Vector evfscfsf evfscfsi evfscfuf evfscfui evfsctsf evfsctsi evfsctsiz evfsctuf evfsctui evfsctuiz evfsabs evfsadd evfscmpeq evfscmpgt evfscmplt evfsdiv evfsmul evfsneg evfsnabs evfssub evfststeq evfststgt evfststlt rD,rB rD,rB rD,rB rD,rB rD,rB rD,rB rD,rB rD,rB rD,rB rD,rB rD,rA rD,rA,rB crD,rA,rB crD,rA,rB crD,rA,rB rD,rA,rB rD,rA,rB rD,rA rD,rA rD,rA,rB crD,rA,rB crD,rA,rB crD,rA,rB Syntax — BTB locking APU instructions. The core complex provides a 512-entry BTB for efficient processing of branch instructions. The BTB is a branch target address cache, organized as 128 rows with four-way set associativity, that holds the address and target instruction of the 512 most-recently taken branches. Table 5-5 lists BTB instructions. Table 5-5. BTB Locking APU Instructions Name Branch Buffer Load Entry and Lock Set Branch Buffer Entry Lock Reset Mnemonic bblels bbelr Syntax — — MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 5-13 Core Complex Overview 5.5 Instruction Flow The e500 core is a pipelined, superscalar processor with parallel execution units that allow instructions to execute out of order but record their results in order. Pipelining breaks instruction processing into discrete stages, so multiple instructions in an instruction sequence can occupy the successive stages: as an instruction completes one stage, it passes to the next, leaving the previous stage available to a subsequent instruction. So, even though it may take multiple cycles for an instruction to pass through all of the pipeline stages, once a pipeline is full, instruction throughput is much shorter than the latency. A superscalar processor is one that issues multiple independent instructions into separate execution units, allowing parallel execution. The e500 core has five execution units, one each for branch (BU), load/store (LSU), and multiple-cycle operations (MU), and two for simple arithmetic (SU1 and SU2). The MU and SU1 arithmetic execution units also execute 64-bit SPE vector instructions, using both the lower and upper halves of the 64-bit GPRs. The parallel execution units allow multiple instructions to execute in parallel and out of order. For example, a low-latency addition instruction that is issued to an SU after an integer divide is issued to the MU should finish executing before the higher latency divide instruction. The add instruction can make its results available to a subsequent instruction, but it cannot update the architected GPR specified as its target operand ahead of the multiple-cycle divide instruction. 5.5.1 Initial Instruction Fetch The e500 core begins execution at fixed virtual address 0xFFFF_FFFC. The MMU has a default page translation which maps this to the identical physical address. So, the instruction at physical address 0xFFFF_FFFC must be a branch to another address within the 4-Kbyte boot page. 5.5.2 Branch Detection and Prediction To improve branch performance, the e500 provides an implementation-specific dynamic branch prediction using the BTB to resolve branch instructions and improve the accuracy of branch predictions. Each of the 512 entries in the four-way set associative address cache of branch target addresses includes a 2-bit saturating branch history counter, whose value is incremented or decremented depending on whether the branch was taken. These bits can take four values indicating strongly taken, weakly taken, weakly not taken, and strongly not taken. The BTB is used not only to predict branches, but to detect branches during the fetch stage, offering an efficient way to access instruction streams for branches predicted as taken. In the e500, all branch instructions are assigned positions in the completion queue at dispatch. Speculative instructions in branch target streams are allowed to execute and proceed through the completion queue, although they can complete only after the branch prediction is resolved as correct and after the branch instruction itself completes. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 5-14 Freescale Semiconductor Instruction Flow If a branch resolves as correct, instructions in the target stream are marked nonspeculative and are allowed to complete. If the branch history bits in the BTB indicated weakly taken or weakly not taken, the prediction is upgraded to strongly taken or strongly not taken. If a branch resolves as incorrect, instructions in the target stream are flushed from the execution pipeline, the branch history bits are updated in the BTB entry, and nonspeculative fetching begins from the correct path. 5.5.3 e500 Execution Pipeline The seven stages of the e500 execution pipeline—fetch1, fetch2/predecode, decode/dispatch, issue, execute, complete, and write back—are highlighted in grey in Figure 5-4. Indicates stages Fetch Stage 1 Fetch Stage 2 At dispatch, instructions are deallocated from the IQ and assigned sequential positions in the CQ. Maximum four-instruction fetch per clock cycle Decode Stage Maximum two-instruction per cycle dispatch to the issue queues. BIQ can accept one per cycle; GIQ can accept at most two. Issue Stage Branch Issue Queue (BIQ) General Issue Queue (GIQ) Instruction Cache Execute Stage MU Stage 1 LSU Stage 1 Stage 2 Stage 3 Stage 2 Stage 3 Stage 4 Divide Bypass Divide Postdivide BU Execute BU Finish SU1 SU2 Completion Stage Maximum two-instruction completion per clock cycle Write-Back Stage Figure 5-4. Instruction Pipeline Flow MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 5-15 Core Complex Overview The common pipeline stages are as follows: • Instruction fetch—Includes the clock cycles necessary to request an instruction and the time the memory system takes to respond to the request. Instructions retrieved are latched into the instruction queue (IQ) for subsequent consideration by the dispatcher. Instruction fetch timing depends on many variables, such as whether an instruction is in the on-chip instruction cache or an L2 cache (if implemented). Those factors increase when it is necessary to fetch instructions from system memory and include the processor-to-bus clock ratio, the amount of bus traffic, and whether any cache coherency operations are required. Because there are so many variables, unless otherwise specified, the instruction timing examples in this chapter assume optimal performance and show the portion of the fetch stage in which the instruction is in the instruction queue. The fetch1 and fetch2 stages are primarily involved in retrieving instructions. The decode/dispatch stage fully decodes each instruction; most instructions are dispatched to the issue queues (however, isync, rfi, sc, nops, and some other instructions do not go to issue queues). The two issue queues, BIQ and GIQ, can accept as many as one and two instructions, respectively, in a cycle. The behavior of instruction dispatch is covered in significant detail in the e500 Software Optimization Guide. The following is a simplification that covers most cases: — Instructions dispatch only from the two lowest IQ entries—IQ0 and IQ1. — A total of two instructions can be dispatched to the issue queues per clock cycle. — Space must be available in the CQ for an instruction to decode and dispatch (this includes instructions that are assigned a space in the CQ but not in an issue queue). Dispatch is treated as an event at the end of the decode stage. The issue stage reads source operands from rename registers and register files and determines when instructions are latched into the execution unit reservation stations. Note that the e500 has 14 rename registers, one for each completion queue entry, so instructions cannot stall because of a shortage of rename registers. The general behavior of the two issue queues is described as follows: — The GIQ accepts as many as two instructions from the dispatch unit per cycle. SU1, SU2, MU, and all LSU instructions (including 64-bit loads and stores) are dispatched to the GIQ, shown in Figure 5-5. • • MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 5-16 Freescale Semiconductor Instruction Flow From IQ0/IQ1 GIQ3 GIQ2 GIQ1 GIQ0 To SU2, MU, or LSU To SU1, MU, or LSU Figure 5-5. GPR Issue Queue (GIQ) • • Instructions can be issued out-of-order from the bottom two GIQ entries (GIQ1–GIQ0). GIQ0 can issue to SU1, MU, and LSU. GIQ1 can issue to SU2, MU, and LSU. Note that SU2 executes a subset of the instructions that can be executed in SU1. The ability to identify and dispatch instructions to SU2 increases the availability of SU1 to execute more computational-intensive instructions. An instruction in GIQ1 destined for SU2 or the LSU need not wait for an MU instruction in GIQ0 that is stalled behind a long-latency divide. The execute stage accepts instructions from its issue queue when the appropriate reservation stations are not busy. In this stage, the operands assigned to the execution stage from the issue stage are latched. The execution unit executes the instruction (perhaps over multiple cycles), writes results on its result bus, and notifies the CQ when the instruction finishes. The execution unit reports any exceptions to the completion stage. Instruction-generated exceptions are not taken until the excepting instruction is next to retire. Most integer instructions have a 1-cycle latency, so results of these instructions are available 1 clock cycle after an instruction enters the execution unit. The MU and LSU are pipelined, as shown in Figure 5-4. Branches resolve in execute stage. If a branch is mispredicted, it takes 5 cycles for the next instruction to reach the execute stage. The complete and write-back stages maintain the correct architectural machine state and commit results to the architecture-defined registers in the proper order. If completion logic detects an instruction containing an exception status or a mispredicted branch, all following instructions are cancelled, their execution results in rename registers are discarded, and the correct instruction stream is fetched. The complete stage ends when the instruction is retired. Two instructions can be retired per clock cycle. If no dependencies exist, as many as two instructions are retired in program order. The write-back stage occurs in the clock cycle after the instruction is retired. The e500 core also provides new instructions that perform single-instruction, multiple-data (SIMD) operations. These signal processing instructions consist of parallel operations on both the MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 5-17 Core Complex Overview upper and lower 32 bits of two 64-bit GPR values and produce two 32-bit results written to a 64-bit GPR. As shown in Figure 5-4, the LSU, MU, and SU1 replicate logic to support 64-bit operations. Although a vector instruction generates separate, discrete results in the upper and lower halves of the target GPR, latency and throughput for vector instructions are the same as those for their scalar equivalents. 5.6 Programming Model The following section describes the e500 core registers defined in Book E, the Freescale Semiconductor Book E implementation standards (EIS), and registers that are specific to the e500. Table 5-6 shows the e500 register set. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 5-18 Freescale Semiconductor Programming Model User-Level Registers General-Purpose Registers 0 31 32 63 0 (upper) GPR02 (lower) GPR1 GPR2 GPR31 Instruction-Accessible Registers 31 32 CR 63 Condition register Count register Link register Integer exception register SPE FP status/control register Accumulator User General SPR (Read/Write) 32 spr 256 63 User SPR general 01 USPRG0 Generalpurpose registers spr 9 spr 8 spr 1 CTR LR XER General SPRs (Read-Only) spr 259 spr 260 spr 263 SPRG3 SPRG4 ••• SPR general registers 3–7 Performance Monitor Registers (Read-Only PMRs) pmr 384 UPMGC0 pmr 0–3 3 spr 512 SPEFSCR3 ACC 3 SPRG7 Time-Base Registers (Read-Only) spr 268 spr 269 TBL TBU Time base lower/upper Global control register Counter registers 0–3 Local control registers a0–a3 Local control registers b0–b3 Miscellaneous Registers spr 513 spr 514 BBEAR 3 BBTAR 3 Branch buffer entry address register Branch buffer target address register UPMCs 3 pmr 128–131 UPMLCas 3 pmr 256–259 UPMLCbs 3 L1 Cache (read-only) spr 515 spr 516 L1CFG0 3 L1CFG1 3 L1 cache configuration registers 0–1 Supervisor-Level Registers Interrupt Registers 32 spr 63 spr 26 spr 27 spr 58 spr 59 IVPR SRR0 SRR1 CSRR0 CSRR1 63 Interrupt vector prefix Save/restore registers 0/1 spr 415 Critical SRR 0/1 spr 528 spr 529 spr 530 spr 531 spr 400 spr 401 32 IVOR0 IVOR1 ••• Configuration Registers 63 Interrupt vector offset registers 0–15 32 MSR spr 1023 spr 286 spr 287 Interrupt vector offset registers 32–35 SCR PIR PVR 63 Machine state System version Processor ID Processor version IVOR15 IVOR32 IVOR33 IVOR34 IVOR35 3 3 3 3 spr 570 MCSRR0 3 spr 571 MCSRR1 3 spr 62 spr 572 spr 573 spr 61 ESR MCSR 3 MCAR DEAR Machine check SRR 0/1 Exception syndrome register Machine check syndrome register Machine check address register Data exception address register Timer/Decrementer Registers spr 22 spr 54 spr 284 spr 285 DEC DECAR TBL TBU TCR TSR Decrementer Decrementer auto-reload Time base lower/upper Timer control Timer status MMU Control and Status (Read/Write) spr 1012 MMUCSR0 spr 624 spr 625 spr 626 spr 627 spr 628 spr 630 MAS0 MAS1 MAS2 MAS3 MAS4 MAS6 PID0 PID1 3 PID2 3 Process ID registers 0–2 3 3 3 3 3 3 3 MMU control and status register 0 Debug Registers spr 308 spr 309 spr 310 spr 304 spr 312 spr 313 spr 316 spr 317 spr 318 spr 319 DBCR0 DBCR1 DBCR2 DBSR IAC1 IAC2 DAC1 DAC2 DVC1 DVC2 Debug control registers 0–2 Debug status register Instruction address compare registers 1–4 Data address compare registers 1 and 2 Data value compare registers 1 and 2 MMU assist registers 0–4 and 6 spr 340 spr 336 Miscellaneous Registers spr 1008 spr 1009 spr 1013 spr 272–279 HID0 3 HID1 3 BUCSR4 SPRG0–7 Hardware implementation dependent 0–1 Branch control and status register General SPRs 0–7 spr 48 spr 633 spr 634 MMU Control and Status (Read Only) spr 1015 MMUCFG 3 spr 688 TLB0CFG 3 spr 689 TLB1CFG 3 MMU configuration TLB configuration 0/1 Performance Monitor Registers pmr 400 PMGC0 3 3 Global control Counter registers 0–3 Local control a0–a3 Local control b0–b3 pmr 16–19 PMC0–3 L1 Cache (Read/Write) spr 1010 spr 1011 L1CSR0 3 L1CSR1 3 L1 cache control/ status 0/1 pmr 144–147 PMCa0–3 3 pmr 272–275 PMCb0–3 3 1 2 3 4 USPRG0 is a separate physical register from SPRG0. The 64-bit GPR registers are accessed by the SPE as separate 32-bit registers by SPE instructions. Only SPE vector instructions can access the upper word. These registers are defined by the EIS and are not part of the Book E architecture. These registers are e500-specific. Figure 5-6. e500 Core Programming Model MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 5-19 Core Complex Overview 5.7 On-Chip Cache Implementation The core complex contains separate 32-Kbyte, eight-way set-associative, level 1 (L1) instruction and data caches to give rapid access to instructions and data. The data cache supports four-state MESI memory coherency protocol with 3-bit status and 1-bit coherency valid fields. The core complex broadcasts all cache management functions based on the setting of the address broadcast enable bit, HID1[ABE], allowing management of other caches in the system. On the MPC8540 the ABE bit must be set to ensure that cache and TLB management instructions operate properly on the L2 cache. The caches implement a pseudo-least-recently-used (PLRU) replacement algorithm. Parity generation and checking may be enabled for both caches, and each cache can be independently invalidated through L1CSR1 and L1CSR0. Additionally, instructions are provided to perform cache locking and unlocking on both data and instruction caches on a cache-block granularity. These are listed in Section 5.10.3, “Cache Control Instructions.” Individual instruction cache lines and data cache lines can be invalidated using the icbi and dcbi instructions, respectively. The entire data cache can be invalidated by setting L1CSR0[CFI]; the entire instruction cache can be invalidated by setting L1CSR1[ICFI]. 5.8 Interrupts and Exception Handling The e500 core supports an extended exception handling model, with nested interrupt capability and extensive interrupt vector programmability. The following sections define the exception model, including an overview of exception handling as implemented on the e500 core, a brief description of the exception classes, and an overview of the registers involved in the processes. 5.8.1 Exception Handling In general, interrupt processing begins with an exception that occurs due to external conditions, errors, or program execution problems. When the exception occurs, the processor checks to verify interrupt processing is enabled for that particular exception. If enabled, the interrupt causes the state of the processor to be saved in the appropriate registers, and prepares to begin execution of the handler located at the associated vector address for that particular exception. Once the handler is executing, the implementation may need to check one or more bits in the exception syndrome register (ESR) or the SPEFSCR, depending on the exception, to verify the specific cause of the exception and take appropriate action. The core complex provides the interrupts described in Section 5.8.5, “Interrupt Registers.” MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 5-20 Freescale Semiconductor Interrupts and Exception Handling 5.8.2 • Interrupt Classes Asynchronous interrupts (such as machine check, critical input, and external interrupts) are caused by events that are independent of instruction execution. For asynchronous interrupts, the address reported in a save/restore register is the address of the instruction that would have executed next had the asynchronous interrupt not occurred. Synchronous interrupts are those that are caused directly by the execution or attempted execution of instructions. Synchronous inputs may be either precise or imprecise, which are described as follows: — Synchronous precise interrupts are those that precisely indicate the address of the instruction causing the exception that generated the interrupt or, in some cases, the address of the immediately following instruction. The interrupt type and status bits indicate which instruction is addressed in the appropriate save/restore register. — Synchronous imprecise interrupts are those that may indicate the address of the exception causing the exception that generated the interrupt, or some instruction after the instruction causing the interrupt. If the interrupt was caused by either the context synchronizing mechanism or the execution synchronizing mechanism, the address in the appropriate save/restore register is the address of the interrupt forcing instruction. If the interrupt was not caused by either of those mechanisms, the address in the save/restore register is the last instruction to start execution and may not have completed. No instruction following the instruction in the save/restore register has executed. All interrupts may be categorized as asynchronous/synchronous and critical/noncritical. • 5.8.3 Interrupt Types The e500 core processes all interrupts as either machine check, critical, or noncritical types. Separate control and status register sets are provided for each interrupt type. The core handles interrupts from these three types in the following priority order: 1. Machine check interrupt (highest priority)—The e500 defines a separate set of resources for the machine check interrupt. They use the machine check save and restore registers (MCSRR0/MCSRR1) to save state when they are taken, and they use the rfmci instruction to restore state. These interrupts can be masked by the machine check enable bit, MSR[ME]. 2. Noncritical interrupts—First-level interrupts that allow the processor to change program flow to handle conditions generated by external signals, errors, or unusual conditions arising from program execution or from programmable timer-related events. These interrupts are largely identical to those previously defined by the OEA portion of the MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 5-21 Core Complex Overview Power PC architecture. They use save and restore registers (SRR0/SRR1) to save state when they are taken and they use the rfi instruction to restore state. Asynchronous noncritical interrupts can be masked by the external interrupt enable bit, MSR[EE]. 3. Critical interrupts—Critical interrupts can be taken during a noncritical interrupt or during regular program flow. They use the critical save and restore registers (CSRR0/CSRR1) to save state when they are taken and they use the rfci instruction to restore state. These interrupts can be masked by the critical enable bit, MSR[CE]. Book E defines the critical input, watchdog timer, and machine check interrupts as critical interrupts, but the e500 defines a third set of resources for the machine check interrupt, as described in Table 5-6. All interrupts except machine check are ordered within the two categories of noncritical and critical, such that only one interrupt of each category is reported, and when it is processed (taken), no program state is lost. Because save/restore register pairs are serially reusable, program state may be lost when an unordered interrupt is taken. 5.8.4 Upper Bound on Interrupt Latencies Core complex interrupt latency is defined as the number of core clocks between the sampling of the interrupt signal as asserted and the initiation of the IVOR fetch (that is, the fetch of the first instruction in the handler). Core complex interrupt latency is determinate unless a guarded load or a cache-inhibited stwcx. is being executed, in which case the latency is indeterminate. The minimum latency is 3 core clocks and the maximum is 8, not including the 2 bus clock cycles required to synchronize the interrupt signal from the pad. When an interrupt is taken, all instructions in the IQ are thrown away except if the oldest instruction is a load/store instruction. That is, if an asynchronous interrupt is being serviced and the oldest instruction is not a load/store instruction, the core complex goes straight from sampling the interrupt to ensuring a recoverable state and issuing an exception. If a load/store instruction is oldest, the core complex waits 4 clocks before ensuring a recoverable state. During this time, any instruction finished by the LSU is deallocated. 5.8.5 Interrupt Registers Table 5-6. Interrupt Registers The registers associated with interrupt and exception handling are described in Table 5-6. Register Description Noncritical Interrupt Registers SRR0 SRR1 Save/restore register 0—Holds the address of the instruction causing the exception or the address of the instruction that will execute after the rfi instruction. Save/restore register 1—Holds machine state on noncritical interrupts and restores machine state after an rfi instruction is executed. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 5-22 Freescale Semiconductor Interrupts and Exception Handling Table 5-6. Interrupt Registers (continued) Register Description Critical Interrupt Registers CSRR0 CSRR1 Critical save/restore register 0—On critical interrupts, holds either the address of the instruction causing the exception or the address of the instruction that will execute after the rfci instruction. Critical save/restore register 1—Holds machine state on critical interrupts and restores machine state after an rfci instruction is executed. Machine Check Interrupt Registers MCSRR0 Machine check save/restore register 0—Used to store the address of the instruction that will execute after an rfmci instruction is executed. MCSRR1 Machine check save/restore register 1—Holds machine state on machine check interrupts and restores machine state (if recoverable) after an rfmci instruction is executed. MCAR Machine check address register—Holds the address of the data or instruction that caused the machine check interrupt. MCAR contents are not meaningful if a signal triggered the machine check interrupt. Syndrome Registers MCSR ESR Machine check syndrome register—Holds machine state information on machine check interrupts and restores machine state after an rfmci instruction is executed. Exception syndrome register—Provides a syndrome to differentiate between the different kinds of exceptions that generate the same interrupt type. Upon generation of a specific exception type, the associated bit is set and all other bits are cleared. SPE APU Interrupt Registers SPEFSC Signal processing and embedded floating-point status and control register—Provides interrupt control and R status as well as various condition bits associated with the operations performed by the SPE APU. Other Interrupt Registers DEAR IVPR IVORs Data exception address register—Holds the address that was referenced by a load, store, or cache management instruction that caused an alignment, data TLB miss, or data storage interrupt. Together, IVPR[32–47]||IVORn [48–59]||0b0000 define the address of an interrupt-processing routine. See Table 5-7 and the EREF for more information. Each interrupt has an associated interrupt vector address, obtained by concatenating the IVPR value with the address index in the associated IVOR (that is, IVPR[32–47]||IVORn[48–59]||0b0000). The resulting address is that of the instruction to be executed when that interrupt occurs. IVPR and IVOR values are indeterminate on reset, and must be initialized by the system software using mtspr. Table 5-7 lists IVOR registers implemented on the e500 and the associated interrupts. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 5-23 Core Complex Overview Table 5-7. Interrupt Vector Registers and Exception Conditions Register Interrupt Book E–Defined IVORs IVOR0 IVOR1 IVOR2 IVOR3 IVOR4 IVOR5 IVOR6 IVOR7 IVOR8 IVOR9 IVOR10 IVOR11 IVOR12 IVOR13 IVOR14 IVOR15 Critical input Machine check interrupt offset Data storage interrupt offset Instruction storage interrupt offset External input interrupt offset Alignment interrupt offset Program interrupt offset Floating-point unavailable interrupt offset System call interrupt offset Auxiliary processor unavailable interrupt offset Decrementer interrupt offset Fixed-interval timer interrupt offset Watchdog timer interrupt offset Data TLB error interrupt offset Instruction TLB error interrupt offset Debug interrupt offset e500-Specific IVORs IVOR32 IVOR33 IVOR34 IVOR35 SPE APU unavailable interrupt offset SPE floating-point data exception interrupt offset SPE floating-point round exception interrupt offset Performance monitor 5.9 Memory Management The e500 core complex supports demand-paged virtual memory as well other memory management schemes that depend on precise control of effective-to-physical address translation and flexible memory protection as defined by Book E. The mapping mechanism consists of software-managed TLBs that support variable-sized pages with per-page properties and permissions. The following properties can be configured for each TLB: • • • • User mode page execute access User mode page read access User mode page write access Supervisor mode page execute access MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 5-24 Freescale Semiconductor Memory Management • • • • • • • • Supervisor mode page read access Supervisor mode page write access Write-through required (W) Caching inhibited (I) Memory coherence required (M) (ignored on the MPC8540) Guarded (G) Endianess (E) User-definable (U0–U3), a 4-bit implementation-specific field The core complex employs a two-level memory management unit (MMU) architecture. There are separate instruction and data level-1 (L1) MMUs backed up by a unified level-2 (L2) MMU, as shown in Figure 5-7. Memory Unit To Instruction Unit 32-Kbyte I-Cache Tags Instruction Line FIll Buffer L1 Instruction MMU 4-Entry I-L1VSP 64-Entry I-L1TLB4K MAS Registers L2 MMUs Unified 16-Entry TLB Array (TLB1) 256-Entry TLB Array (TLB0) L1 Data MMU 4-Entry D-L1VSP 64-Entry D-L1TLB4K Data Line FIll Buffer To Load/Store Unit Tags 32-Kbyte D-Cache Core Interface Figure 5-7. MMU Structure Level-1 MMUs have the following features: • • Four-entry, fully associative TLB array that supports all nine page sizes 64-entry, 4-way set-associative TLB 4-Kbyte array that supports 4-Kbyte pages only MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 5-25 Core Complex Overview • • • • • • • Hardware partially managed by L2 MMU Supports snooping of TLBs by both internal and external tlbivax instructions A 16-entry, fully associative L2 TLB array (TLB1) that supports all nine page sizes 256-entry, 2-way set-associative TLB array (TLB0) that supports only 4-Kbyte pages Hardware assist for TLB miss exceptions Software managed by tlbre, tlbwe, tlbsx, tlbsync, tlbivax, and mtspr instructions Supports snooping of TLB by both internal and external tlbivax instructions The level-2 MMU has the following features: 5.9.1 Address Translation The core complex fetch and load/store units generate 32-bit effective addresses. The MMU translates these addresses to real 32-bit addresses (which are used for memory bus accesses) using an interim 41-bit virtual address. Figure 5-8 shows the translation flow. 8 bits PID0 AS PID1 PID2 15–20 Bits* Effective Page Number 32-bit Effective Address (EA) >12 Bits* Byte Address Three 41-bit Virtual Addresses (VAs) L1 MMUs L2 MMU (Unified) 16-Entry Fully-Assoc. VSP Array (TLB1) 256-Entry 2-Way Set Assoc. Array (TLB0) 15–20 Bits* 32-bit Real Address Real Page Number >12 Bits* Byte Address Instruction L1 MMU 2 TLBs Data L1 MMU 2 TLBs * Number of bits depends on page size (4 Kbytes–128 Mbytes) Figure 5-8. Effective-to-Real Address Translation Flow The appropriate L1 MMU (instruction or data) is checked for a matching address translation. The instruction L1 MMU and data L1 MMU operate independently and can be accessed in parallel, so that hits for instruction accesses and data accesses can occur in the same clock. If an L1 MMU misses, the request for translation is forwarded to the unified (instruction and data) L2 MMU. If found, the contents of the TLB entry are concatenated with the byte address to obtain the physical address of the requested access. On misses, the L1 TLB entries are replaced from their L2 TLB counterparts using a true LRU algorithm. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 5-26 Freescale Semiconductor Memory Management 5.9.2 MMU Assist Registers (MAS1–MAS4 and MAS6) Book E defines SPR numbers for the MMU assist registers, which are used to hold values either read from or to be written to the TLBs and information required to identify the TLB to be accessed. To ensure consistency among Freescale Semiconductor Book E processors, certain aspects of the implementation are defined by the Freescale Semiconductor Book E standard, whereas more specific details are left to individual implementations. MAS3 implements the real page number (RPN), the user attribute bits (U0–U3), and permission bits (UX, SX, UW, SW, UR, SR) that specify user and supervisor read, write, and execute permissions. The e500 does not implement MAS5. MAS registers are affected by the following instructions: • • MAS registers are accessed with the mtspr and mfspr instructions. The TLB Read Entry instruction (tlbre) causes the contents of a single TLB entry from the L2 MMU to be placed in defined locations in MAS0–MAS3. The TLB entry to be extracted is determined by information written to MAS0 and MAS2 before the tlbre instruction is executed. The TLB Write Entry instruction (tlbwe) causes the information stored in certain locations of MAS0–MAS3 to be written to the TLB specified in MAS0. The TLB Search Indexed instruction (tlbsx) updates MAS registers conditionally, based on success or failure of a lookup in the L2 MMU. The lookup is specified by the instruction encoding and specific search fields in MAS6. The values placed in the MAS registers may differ, depending on a successful or unsuccessful search. • • For TLB miss and certain MMU-related DSI/ISI exceptions, MAS4 provides default values for updating MAS0–MAS2. 5.9.3 Process ID Registers (PID0–PID2) The e500 core complex also implements three process ID (PID) registers that hold the values used to construct the three virtual addresses for each access. These process IDs provide an extended page sharing capability. Which of these three virtual addresses is used is controlled by the TID field of a matching TLB entry, and when TID = 0x00 (identifying a page as globally shared), the PID values are ignored. A hit to multiple TLB entries in the L1 MMU (even if they are in separate arrays) or a hit to multiple entries in the L2 MMU is considered to be a programming error. 5.9.4 TLB Coherency The core complex provides the ability to invalidate a TLB entry as defined in the Book E architecture. The tlbivax instruction invalidates a matching local TLB entry. Execution of this MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 5-27 Core Complex Overview instruction is also broadcast on the core complex bus (CCB) if HID1[ABE] is set. The core complex also snoops TLB invalidate transactions on the CCB from other bus masters. On the MPC8540 the ABE bit must be set to ensure that cache and TLB management instructions operate properly on the L2 cache. 5.10 Memory Coherency The core complex supports both three- and four-state memory coherency. Memory coherency is hardware-supported on the system bus through bus snooping and the retry/copyback bus protocol, and through broadcasting of cache management instructions. Translation coherency is also hardware-supported through broadcasting and bus snooping of TLB invalidate transactions. The four-state MESI protocol supports efficient large-scale real-time data sharing between multiple caching bus masters. 5.10.1 Atomic Update Memory References The e500 core supports atomic update memory references for both aligned word forms of data using the load and reserve and store conditional instruction pair, lwarx and stwcx.. Typically, a load and reserve instruction establishes a reservation and is paired with a store conditional instruction to achieve the atomic operation. However, there are restrictions and requirements for this functionality. Because the processor revokes reservations during context switches, the programmer must reacquire the reservation after a context switch to maintain reservations across context switches. 5.10.2 Memory Access Ordering The core complex supports weakly ordered references to memory. Thus the e500 manages the order and synchronization of instructions to ensure proper execution when memory is shared between multiple processes or programs. The cache and data memory control attributes along with msync and mbar provide the required access control. 5.10.3 Cache Control Instructions The core complex supports Book E instructions for performing a full range of cache control functions, including cache locking by line. The core complex supports broadcasting and snooping of these cache control instructions on the CCB. The e500 core also supports the following e500-specific cache locking instructions: • • • Data Cache Block Lock Clear (dcblc) Data Cache Block Touch-and-Lock Set (dcbtls) Data Cache Block Touch for Store and Lock Set (dcbtstls) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 5-28 Freescale Semiconductor Core Complex Bus (CCB) • • Instruction Cache Block Lock Clear (icblc) Instruction Cache Block Touch-and-Lock Set (icbtls) 5.10.4 Programmable Page Characteristics Cache and memory attributes are programmable on a per-page basis. In addition to the write-through, caching-inhibited, memory coherency enforce, and guarded characteristic defined by the WIMG bits, Book E defines an endianness bit, E, that allows selection of big- or little-endian byte ordering on a per-page basis. In addition to the WIMGE bits, the Book E MMU model defines user-definable page attribute bits (U0–U3). 5.11 Core Complex Bus (CCB) The core complex defines a versatile local bus interface that allows a wide range of system performance and system-complexity trade-offs. The interface defines the following buses. • • • An address-out bus for mastering bus transactions An address-in bus for snooping internal resources Three tagged data buses Two of the data buses are general-purpose data-in buses for reads, and the third is a data-out bus for writes. The two data-in buses feature support for out-of-order read transactions from two different sources simultaneously, and all three data buses may be operated concurrently. The address-in bus supports snooping for external management of the L1 caches and TLBs by other bus masters. The core complex broadcasts and snoops the cache and TLB management instructions accordingly. It is envisioned that a wide range of system implementations can be constructed from the defined interface. 5.12 Performance Monitoring The e500 core provides a performance monitoring capability that allows counting of events such as processor clocks, instruction cache misses, data cache misses, mispredicted branches, and others. The count of these events may be configured to trigger a performance monitor exception following the e500 interrupt model. This interrupt is assigned to vector offset register IVOR35. The register set associated with the performance monitoring function consists of counter registers, a global control register, and local control registers. These registers are read/write from supervisor mode, and each register is reflected to a corresponding read-only register for user mode. Two instructions, mtpmr and mfpmr, are provided for moving data to and from these registers. An overview of the performance monitoring registers is provided in the following sections. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 5-29 Core Complex Overview 5.12.1 Global Control Register The PMGC0 register provides global control of the performance monitoring facility from supervisor mode. From this register all counters may be frozen, unfrozen, or configured to freeze on an enabled condition or event. Additionally, the performance monitoring facility may be disabled or enabled from this register. The contents of PMGC0 are reflected to UPMGC0 which may be read from user mode using the mfpmr instruction. 5.12.2 Performance Monitor Counter Registers There are four counter registers (PCM0–PCM3) provided in the performance monitoring facility. These 32-bit registers hold the current count for software selectable events and can be programmed to generate an exception on overflow. These registers may be written or read from supervisor mode using the mtpmr and mfpmr instructions. The contents of these registers are reflected to UPCM0–UPCM3, which can be read from user mode with mfpmr. Performance monitor exceptions occur only if all of the following conditions are met: • • • • A counter is in the overflow state The counter's overflow signalling is enabled Overflow exception generation is enabled in PMGC0 MSR[EE] is set 5.12.3 Local Control Registers For each of the counter registers, there are two corresponding local control registers. These two registers specify which of the 128 available events is to be counted, what specific action is to be taken on overflow, and various options for freezing a counter value under given modes or conditions. • PMLCa0–PMLCa3 provide fields that allow freezing of the corresponding counter in user mode, supervisor mode, or under software control. Additionally, the overflow condition may be enabled or disabled from this register. The contents of these registers are reflected to UPMCLa0–UPMLCa3, which can be read from user mode with mfpmr. PMLCb0–PMLCb3 provide count scaling for each counter register using configurable threshold and multiplier values. The threshold is a 6-bit value and the multiplier is a 3-bit encoded value, allowing eight multiplier values in the range of 1 to 128. Any counter may be configured to increment only when an event occurs more than [threshold × multiplier] times. The contents of these registers are reflected to UPMCLb0–UPMLCb3, which can be read from user mode with mfpmr. • MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 5-30 Freescale Semiconductor Legacy Support of PowerPC Architecture 5.13 Legacy Support of PowerPC Architecture This section provides an overview of the architectural differences and compatibilities of the e500 core compared with the AIM PowerPC architecture. The two levels of the e500 programming environment are as follows: • User level—This defines the base user-level instruction set, user-level registers, data types, memory conventions, and the memory and programming models seen by application programmers. Supervisor level—This defines supervisor-level resources typically required by an operating system, the memory management model, supervisor level registers, and the exception model. • In general, the e500 core supports the user-level architecture from the existing AIM architecture. The following subsections are intended to highlight the main differences. For specific implementation details refer to the relevant chapter. 5.13.1 Instruction Set Compatibility The following sections generally describe the user and supervisor instruction sets. 5.13.1.1 User Instruction Set The e500 core executes legacy user-mode binaries and object files except for the following: • The e500 supports vector and scalar single-precision floating-point operations as APUs. These instructions have different encoding than the AIM definition of the PowerPC architecture. Additionally, the e500 core uses GPRs for floating-point operations, rather than the FPRs defined by the UISA. Most porting of floating-point operations can be handled by recompiling. String instructions are not implemented on the e500; therefore, trap emulation must be provided to ensure backward compatibility. • 5.13.1.2 Supervisor Instruction Set The supervisor mode instruction set defined by the AIM version of the PowerPC architecture is compatible with the e500 with the following exceptions: • • The MMU architecture is different, so some TLB manipulation instructions have different semantics. Instructions that support the BATs and segment registers are not implemented. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 5-31 Core Complex Overview 5.13.2 Memory Subsystem Both Book E and the AIM version of the PowerPC architecture provide separate instruction and data memory resources. The e500 provides additional cache control features, including cache locking. 5.13.3 Exception Handling Exception handling is generally the same as that defined in the AIM version of the PowerPC architecture for the e500, with the following differences: • • Book E defines a new critical interrupt, providing an extra level of interrupt nesting. The critical interrupt includes external critical and watchdog timer time-out inputs. The machine check exception differs from the Book E and from the AIM definition. It defines the Return from Machine Check Interrupt instruction, rfmci, and two machine check save/restore registers, MCSRR0 and MCSRR1. Book E processors can use IVPR and IVORs to set exception vectors individually, but they can be set to the address offsets defined in the OEA to provide compatibility. Unlike the AIM version of the PowerPC architecture, Book E does not define a reset vector; execution begins at a fixed virtual address, 0xFFFF_FFFC. Some Book E and e500-specific SPRs are different from those defined in the AIM version of the PowerPC architecture, particularly those related to the MMU functions. Much of this information has been moved to a new exception syndrome register (ESR). Timer services are generally compatible, although Book E defines a new decrementer auto reload feature, the fixed-interval timer critical interrupt, and the watchdog timer interrupt, which are implemented in the e500 core. • • • • An overview of the interrupt and exception handling capabilities of the e500 core can be found in Section 5.8, “Interrupts and Exception Handling.” 5.13.4 Memory Management The e500 core implements a straightforward virtual address space that complies with the Book E MMU definition, which eliminates segment registers and block address translation resources. Book E defines resources for fixed 4-Kbyte pages and multiple, variable page sizes that can be configured in a single implementation. TLB management is provided with new instructions and SPRs. 5.13.5 Reset Book E–compliant cores do not share a common reset vector with the AIM version of the PowerPC architecture. Instead, at reset fetching begins at address 0xFFFF_FFFC. In addition to MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 5-32 Freescale Semiconductor PowerQUICC III Implementation Details the Book E reset definition, the EIS and the e500 define specific aspects of the MMU page translation and protection mechanisms. Unlike the AIM version of the PowerPC core, as soon as instruction fetching begins, the e500 core is in virtual mode with a hardware-initialized TLB entry. 5.13.6 Little-Endian Mode Unlike the AIM version of the PowerPC architecture, where little-endian mode is controlled on a system basis, Book E allows control of byte ordering on a memory page basis. In addition, the little-endian mode used in Book E is true little endian. 5.14 PowerQUICC III Implementation Details Table 5-8 summarizes e500 core functionality that is not implemented by PowerQUICC III devices. Table 5-8. Differences Between the e500 Core and the PowerQUICC III Core Implementation Feature Cache protocol Multiprocessor functionality PowerQUICC Implementation The L2 cache does not support MESI cache protocol. Because PowerQUICC III is designed for a uniprocessor environment, the following e500 functionality is not implemented: • The memory coherence bit, M, which is controlled through MAS2[M] and MAS4[MD] has no effect. • HID1[ABE] has meaning only in that it must be set to ensure that cache and TLB management instructions operate properly with respect to the L2 cache. • Dynamic snooping does not occur in power-stopped state (see the note below in the entry for dynamic bus snooping). Nexus is not supported. The Nexus processor ID register (NPIDR) and the Nexus bus enable bit (HID1[NEXEN]) are not supported. R1 and R2 data bus parity are disabled on PowerQUICC III devices. HID1[R1DPE,R2DPE] are reserved. The PowerQUICC III devices do not perform dynamic bus snooping as described here. That is, when the e500 core is in core-stopped state (which is the state of the core when the PowerQUICC III devices is in either the nap or sleep state), the core is not awakened to perform snoops on global transactions. Therefore, before entering nap or sleep modes, L1 caches should be flushed if coherency is required during these power-down modes. For more information, see Section 18.5.1.9, “Snooping in Power-Down Modes.” PowerQUICC III devices define values for 01, 10, and 11, as follows: 00 No watchdog timer reset can occur. 01 Force processor checkstop on second timeout of watchdog timer 10 Assert processor reset output (core_hreset_req) on second timeout of watchdog timer 11 Reserved The SPE and SPFP APU functionality will not be implemented in next generation of PowerQUICC devices. Freescale Semiconductor strongly recommends that use of these instructions be confined to libraries and device drivers. Customer software that uses SPE or SPFP instructions at the assembly level or that uses SPE or SPFP intrinsics will require rewriting for upward compatibility with next generation PowerQUICC devices. Freescale Semiconductor offers a libmoto_e500 library that uses SPE and SPFP APU instructions. Freescale Semiconductor will also provide future libraries to support next generation PowerQUICC devices. Nexus support R1 and R2 data bus parity Dynamic bus snooping Supported TCR[WRC] SPE and SPFP APUs MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 5-33 Core Complex Overview Table 5-8. Differences Between the e500 Core and the PowerQUICC III Core Implementation (continued) Feature HID0 implementation HID1 Implementation PowerQUICC Implementation SEL_TBCLK bit. Selects time base clock. If this bit is set and the time base is enabled, the time base is based on the TBCLK input, which on the PowerQUICC III devices is RTC. PLL_MODE Set to 01 PLL_CFG. PowerQUICC III devices support the following: 00010 0 2:1 00010 1 5:2 (2.5:1) 00011 0 3:1 00011 1 7:2 (3.5:1) NEXEN, R1DPE, R2DPE, MPXTT, MSHARS, SSHAR, ATS, and MID are not implemented On PowerQUICC III devices, ABE must be set to ensure that cache and TLB management instructions operate properly on the L2 cache. HID1[RFXE] controls whether assertion of core_fault_in causes a machine check interrupt. Assertion of core_fault_in can result from uncorrectable data error, such as an L2 multibit ECC error. It can also occur for a system error if logic on the integrated device signals a fault for nonfatal errors (read data is corrupt (or zero), but the transaction can complete without corrupting other system state, making it unnecessary to take a machine check (for example, a master abort of a PCI transaction). If RFXE is 0, and core_fault_in is asserted, any data on the CCB is dropped, either stalling the load/store unit and causing the e500 pipeline to stall until an interrupt occurs (typically generated by the programmable interrupt controller (PIC) in response to the fault) or allowing processingþto continue with the bad data until the interrupt occurs. Because core_fault_in cannot cause a machine check if RFXE is 0, it is critical that the system be configured to generate the appropriate interrupt, as described below. It is also possible to hang the processor (requiring a hard reset to recover) if a guardedþload hits in the L2 cache and gets an uncorrectable ECC error. Because of this, avoid defining memory as cacheable but guarded. If this combination is required, RFXE must be enabled, in which case an error causes both a machine check interrupt and an external interrupt when a bus fault condition is detected unless interrupts are masked for all sources of bus faults. If RFXE is 0, conditions that cause the assertion of core_fault_in cannot directly cause the e500 to generate a machine check; however, PowerQUICC III devices must be configured to detect and enable such conditions. The following describes how error bits should be configured: • ECM mapping errors: EEER[LAEE] must be set. See Section 8.2.1.4, “ECM Error Enable Register (EEER).” • L2 multiple-bit ECC errors: L2ERRDIS[MBECCDIS] must be cleared to ensure that error can be detected. L2ERRINTEN[MBECCINTEN] must be set. SeeSection 7.3.1.5, “L2 Error Registers.” • DDR multiple-bit ECC errors. ERR_DISABLE[MBED] and ERR_INT_EN[MBEE] must be zero and DDR_SDRAM_CFG[ECC_EN] must be one to ensure that an interrupt is generated. See Section 9.4.1, “Register Descriptions.” • PCI. The appropriate parity detect and master-abort bits in ERR_DR must be cleared and the corresponding enable bits in ERR_EN must be set to ensure that an interrupt is generated. Section 16.3.1.4, “PCI/X Error Management Registers.” • Local bus controller parity errors. LTEDR[PARD] must be cleared and LTEIR[PARI] must be set to ensure that an parity errors can generate an interrupt. See Section 13.3.1.11, “Transfer Error Check Disable Register (LTEDR),” and Section 13.3.1.12, “Transfer Error Interrupt Enable Register (LTEIR).” • RapidIO. PCR[CCE] must be set to ensure that an interrupt is generated due to a CRC error. See Section 17.3.2.1.2, “Port Configuration Register (PCR).” RFXE must also be set if software requires that code execution stop immediately when a bus fault occurs rather than continuing with the bad data until the interrupt arrives. Again, this results in both a machine check interrupt and an external interrupt when a bus fault is detected, unless all possible sources for bus fault have their interrupts masked.þThe machine check interrupt can then reenable normal interrupts and wait for the interrupt due to the fault to be received before returning from the machine check. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 5-34 Freescale Semiconductor PowerQUICC III Implementation Details Table 5-8. Differences Between the e500 Core and the PowerQUICC III Core Implementation (continued) Feature PIR value PVR value PowerQUICC Implementation The PIR value is all zeros on PowerQUICC III devices. The PVR reset value is 0x8020_nnnn. See Table 5-1. PVR[VERSION] = 0x8020 PVR[REVISION] = 0xnnnn (revision specific value) The SVR reset value is0x8030_nnnn. See Table 5-1. SVR[VERSION] = 0x8030 SVR[REVISION] = 0xnnnn (revision specific value) SVR value MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 5-35 Core Complex Overview MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 5-36 Freescale Semiconductor Chapter 6 Core Register Summary This chapter describes the e500 register model and indicates whether each register is defined by Book E, by the Freescale Semiconductor Book E implementation standards (EIS), or by the implementation. For the programmer, drawing this distinction indicates the degree to which code is portable among Freescale Semiconductor Book E processors. This chapter provides reference material—figures for each register and complete descriptions of register fields, including how the registers are accessed, reset values, and whether they can be accessed by user- and supervisor-level software. Detailed discussions of how these registers are used are provided in individual chapters. Note that all registers described here are implemented in the hardware as part of the e500 core. 6.1 Overview As shown in Figure 6-1, most of the registers implemented on a Freescale Semiconductor Book E processor are defined by the Book E architecture, and most of those were defined by the AIM definition of the PowerPC architecture and have changed very little. Additional registers and fields within Book E–defined registers are defined by the EIS and by the implementation. Book E defines some register fields in a very general way, leaving some details as implementation specific. In some cases, this more specific functionality is defined by the EIS; in others it is left up to the processor. This chapter identifies the level at which each features is defined. 6.1.1 Register Set Table 6-1 shows the e500 register set, grouped by whether they can be accessed by user- or supervisor-level software. Unless otherwise indicated, these registers are defined by Book E. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 6-1 Core Register Summary User-Level Registers General-Purpose Registers 0 31 32 63 (upper) GPR0 2 (lower) GPR1 GPR2 ••• Instruction-Accessible Registers 0 31 32 CR 63 Condition register Count register Link register Integer exception register SPE FP status/control register Accumulator User General SPR (read/write) spr 256 32 63 USPRG0 User SPR general 0 1 Generalpurpose registers spr 9 spr 8 spr 1 CTR LR XER General SPRs (read only) spr 259 spr 260 spr 263 SPRG3 SPRG4 ••• GPR31 SPR general registers 3–7 Performance Monitor Registers (read-only PMRs) pmr 384 UPMGC03 pmr 0–3 UPMCs 3 Global control register Counter registers 0–3 Local control registers a0–a3 Local control registers b0–b3 spr 512 SPEFSCR3 ACC 3 SPRG7 Time-Base Registers (read only) spr 268 spr 269 TBL TBU Time base lower/upper Miscellaneous Registers spr 513 spr 514 BBEAR3 BBTAR 3 pmr 128–131 UPMLCas3 pmr 256–259 UPMLCbs3 Branch buffer entry address register Branch buffer target address register L1 Cache (read only) spr 515 spr 516 L1CFG0 3 L1CFG1 3 L1 cache configuration registers 0–1 Supervisor-Level Registers Interrupt Registers 32 spr 63 spr 26 spr 27 spr 58 spr 59 IVPR SRR0 SRR1 CSRR0 CSRR1 63 Interrupt vector prefix Save/restore registers 0/1 spr 415 Critical SRR 0/1 spr 528 spr 529 spr 530 spr 531 spr 400 spr 401 32 IVOR0 IVOR1 ••• Configuration Registers 63 Interrupt vector offset registers 0–15 32 MSR spr 1023 spr 286 spr 287 Interrupt vector offset registers 32–35 SCR PIR PVR 63 Machine state System version Processor ID Processor version IVOR15 IVOR323 IVOR333 IVOR343 IVOR353 spr 570 MCSRR03 spr 571 MCSRR13 spr 62 spr 572 spr 573 spr 61 ESR MCSR3 MCAR DEAR Machine check SRR 0/1 Exception syndrome register Machine check syndrome register Machine check address register Data exception address register Timer/Decrementer Registers spr 22 spr 54 spr 284 spr 285 DEC DECAR TBL TBU TCR TSR Decrementer Decrementer auto-reload Time base lower/upper Timer control Timer status MMU Control and Status (read/write) spr 1012 MMUCSR03 spr 624 spr 625 spr 626 spr 627 spr 628 spr 630 MAS0 MAS13 MAS23 MAS3 MAS43 MAS63 PID0 PID13 PID23 3 3 MMU control and status register 0 Debug Registers spr 308 spr 309 spr 310 spr 304 spr 312 spr 313 spr 316 spr 317 spr 318 spr 319 DBCR0 DBCR1 DBCR2 DBSR IAC1 IAC2 DAC1 DAC2 DVC1 DVC2 Debug control registers 0–2 Debug status register Instruction address compare registers 1–4 Data address compare registers 1 and 2 Data value compare registers 1 and 2 MMU assist registers 0–4 and 6 spr 340 spr 336 Miscellaneous Registers spr 1008 Process ID registers 0–2 spr 1009 spr 1013 spr 272–279 HID0 3 HID13 BUCSR4 SPRG0–7 Hardware implementation dependent 0–1 Branch control and status register General SPRs 0–7 spr 48 spr 633 spr 634 MMU Control and Status (read only) spr 1015 MMUCFG 3 3 MMU configuration TLB configuration 0/1 Performance Monitor Registers pmr 400 pmr 16–19 PMGC03 PMC0–33 Global control Counter registers 0–3 Local control a0–a3 Local control b0–b3 spr 688 TLB0CFG spr 689 TLB1CFG3 L1 Cache (read/write) spr 1010 spr 1011 L1CSR03 L1CSR13 L1 cache control/status 0/1 pmr 144–147 PMCa0–33 pmr 272–275 PMCb0–33 1 2 3 4 USPRG0 is a separate physical register from SPRG0. The 64-bit GPR registers are accessed by the SPE as separate 32-bit registers by SPE instructions. Only SPE vector instructions can access the upper word. These registers are defined by the EIS and are not part of the Book E architecture. These registers are e500-specific. Figure 6-1. Core Register Model MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 6-2 Freescale Semiconductor Register Model for 32-Bit Implementations 6.2 • Register Model for 32-Bit Implementations Book E–defined registers that are accessed as part of instruction execution. These include the following: — Registers used for computation. These include the following: – General-purpose registers (GPRs)—Book E defines a set of 32 GPRs used to hold source and destination operands for load, store, arithmetic, and computational instructions, and to read and write to other registers. The e500 implements these as 64-bit registers for use with 64-bit load, store, and merge instructions, as described in Section 6.3.1, “General-Purpose Registers (GPRs).” – Integer exception register (XER)—Bits in this register are set based on the operation of an instruction considered as a whole, not on intermediate results. (For example, the Subtract from Carrying instruction (subfc), the result of which is specified as the sum of three values, sets bits in the XER based on the entire operation, not on an intermediate sum.) These registers are described in Section 6.3, “Registers for Computational Operations.” — Condition register (CR)—Used to record conditions such as overflows and carries that occur as a result of executing arithmetic instructions (including those implemented by the SPE APU). The CR is described in Section 6.4, “Registers for Branch Operations.” — Machine state register (MSR)—Used by the operating system to configure parameters such as user/supervisor mode, address space, and enabling of asynchronous interrupts. This register is described in Section 6.5.1, “Machine State Register (MSR),” grouped with processor control SPRs. Book E–defined special-purpose registers (SPRs) that are accessed explicitly using mtspr and mfspr instructions. These registers are listed in Table 6-1 in Section 6.2.1, “Special-Purpose Registers (SPRs).” Freescale Semiconductor EIS– and e500-defined SPRs that are accessed explicitly using mtspr and mfspr are listed in Table 6-2 in Section 6.2.1, “Special-Purpose Registers (SPRs).” Freescale Semiconductor EIS–defined performance monitor registers (PMRs). These registers are similar to SPRs, but are accessed with Freescale Semiconductor EIS–defined move to and move from PMR instructions (mtpmr and mfpmr). Section 6.4, “Registers for Branch Operations,” describes the count register (CTR) and the link register (LR) defined by Book E. Section 6.5, “Processor Control Registers” Section 6.6, “Timer Registers” MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Embedded 32-bit processors implement the following types of software-accessible registers: • • • In this chapter, SPRs are grouped by function as follows: • • • Freescale Semiconductor 6-3 Core Register Summary • • • • • • • • Section 6.7, “Interrupt Registers” Section 6.8, “Software-Use SPRs (SPRG0–SPRG7 and USPRG0),” describes Book E–defined SPRs defined for software use. Section 6.9, “Branch Target Buffer (BTB) Registers,” describes e500-specific registers defined to support the e500 BTBs. Section 6.10, “Hardware Implementation-Dependent Registers,” describes HID0 and HID1. Section 6.11, “L1 Cache Configuration Registers” Section 6.12, “MMU Registers” Section 6.13, “Debug Registers” Section 6.14, “Signal Processing and Embedded Floating-Point Status and Control Register (SPEFSCR)” The e500 core implements 64-bit GPRs, the upper 32 bits of which are used only with 64-bit load, store, and merge instructions. 6.2.1 Special-Purpose Registers (SPRs) Table 6-1 summarizes SPRs defined in Book E. The SPR numbers are used in the instruction mnemonics. Bit 5 in an SPR number indicates whether an SPR is accessible from user- or supervisor-level software. An mtspr or mfspr instruction that specifies an unsupported SPR number is considered an invalid instruction. Table 6-1. Book E Special-Purpose Registers (by SPR Abbreviation) SPR Abbreviatio n CSRR0 CSRR1 CTR DAC1 DAC2 DBCR0 DBCR1 DBCR2 DBSR DEAR DEC DECAR ESR Defined SPR Number Name Decimal Critical save/restore register 0 Critical save/restore register 1 Count register Data address compare 1 Data address compare 2 Debug control register 0 Debug control register 1 Debug control register 2 Debug status register Data exception address register Decrementer Decrementer auto-reload Exception syndrome register 58 59 9 316 317 308 309 310 304 61 22 54 62 Binary 00001 11010 00001 11011 00000 01001 01001 11100 01001 11101 01001 10100 01001 10101 01001 10110 01001 10000 00001 11101 00000 10110 00001 10110 00001 11110 Read/Write Read/Write Read/Write Read/Clear1 Read/Write Read/Write Write-only Read/Write Yes 6.7.1.8/6-20 Yes Yes Yes Yes Yes Yes 6.13.2/6-46 6.7.1.5/6-19 6.6.4/6-17 6.13.1/6-42 Read/Write Read/Write Read/Write Read/Write Access Superviso r Only Yes Yes No Yes Section/ Page 6.7.1.1/6-18 6.7.1.2/6-18 6.4.3/6-11 6.13.4/6-48 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 6-4 Freescale Semiconductor Register Model for 32-Bit Implementations Table 6-1. Book E Special-Purpose Registers (by SPR Abbreviation) (continued) SPR Abbreviatio n IAC1 IAC2 IVOR0 IVOR1 IVOR2 IVOR3 IVOR4 IVOR5 IVOR6 IVOR8 IVOR10 IVOR11 IVOR12 IVOR13 IVOR14 IVOR15 IVPR LR PID PIR PVR SPRG0 SPRG1 SPRG2 SPRG3 Defined SPR Number Name Decimal Instruction address compare 1 Instruction address compare 2 Critical input Machine check interrupt offset Data storage interrupt offset Instruction storage interrupt offset External input interrupt offset Alignment interrupt offset Program interrupt offset System call interrupt offset Decrementer interrupt offset Fixed-interval timer interrupt offset Watchdog timer interrupt offset Data TLB error interrupt offset Instruction TLB error interrupt offset Debug interrupt offset Interrupt vector Link register Process ID register 2 Access Binary 01001 11000 01001 11001 01100 10000 01100 10001 01100 10010 01100 10011 01100 10100 01100 10101 01100 10110 01100 11000 01100 11010 01100 11011 01100 11100 01100 11101 01100 11110 01100 11111 00001 11111 00000 01000 00001 10000 01000 11110 01000 11111 01000 10000 01000 10001 01000 10010 01000 00011 01000 10011 01000 00100 01000 10100 Read/Write Read/Write Read/Write Read-only Read-only Read/Write Read/Write Read/Write Read-only Read/Write Read-only Read/Write Read/Write Read/Write 312 313 400 401 402 403 404 405 406 408 410 411 412 413 414 415 63 8 48 286 287 272 273 274 259 275 Superviso r Only Yes Section/ Page 6.13.3/6-47 Yes 6.7.1.7/6-19 Yes No Yes Yes Yes Yes Yes Yes No3 Yes No Yes 6.7.1.6/6-19 6.4.2/6-11 6.12.1/6-34 6.5.2/6-14 6.5.3/6-14 6.8/6-24 Processor ID register Processor version register SPR general 0 SPR general 1 SPR general 2 SPR general 3 SPRG4 SPR general 4 260 276 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 6-5 Core Register Summary Table 6-1. Book E Special-Purpose Registers (by SPR Abbreviation) (continued) SPR Abbreviatio n SPRG5 Defined SPR Number Name Decimal SPR general 5 261 277 SPRG6 SPR general 6 262 278 SPRG7 SPR general 7 263 279 SRR0 SRR1 TBL Save/restore register 0 Save/restore register 1 Time base lower 26 27 268 284 TBU Time base upper 269 285 TCR TSR USPRG0 XER 1 2 3 4 5 Access Binary 01000 00101 01000 10101 01000 00110 01000 10110 01000 00111 01000 10111 00000 11010 00000 11011 01000 01100 01000 11100 01000 01101 01000 11101 01010 10100 01010 10000 01000 00000 00000 00001 Read-only Read/Write Read-only Read/Write Read-only Read/Write Read/Write Read/Write Read-only Write-only Read-only Write-only Read/Write Read/Clear 4 Superviso r Only No Yes No Yes No Yes Yes Yes No Yes No Yes Yes Yes No No Section/ Page 6.8/6-24 6.7.1.1/6-18 6.7.1.2/6-18 6.6.3/6-17 Timer control register Timer status register User SPR general 0 5 340 336 256 1 6.6.1/6-15 6.6.2/6-16 6.8/6-24 6.3.2/6-8 Read/Write Read/Write Integer exception register The DBSR is read using mfspr. It cannot be directly written to. Instead, DBSR bits corresponding to 1 bits in the GPR can be cleared using mtspr. Implementations may support more than one PID. For implementations with multiple PIDs, the EIS implements the Book E–defined PID as PID0. User-mode read access to SPRG3 is implementation dependent. The TSR is read using mfspr. It cannot be directly written to. Instead, TSR bits corresponding to 1 bits in the GPR can be cleared using mtspr. USPRG0 is a separate physical register from SPRG0. Table 6-2 describes the implementation-specific SPRs. Compilers should recognize the mnemonic names given in Table 6-2 when parsing instructions. Table 6-2. Implementation-Specific SPRs (by SPR Abbreviation) SPR Abbreviatio n BBEAR BBTAR BUCSR HID0 HID1 Name SPR Number Access Superviso r Only No No Yes Yes Yes Section/ Page 6.9.1/6-25 6.9.2/6-25 6.9.3/6-26 6.10.1/6-27 6.10.1/6-27 Branch buffer entry address register Branch buffer target address register Branch unit control and status register Hardware implementation dependent reg 0 Hardware implementation dependent reg 1 513 514 1013 1008 1009 Read/Write Read/Write Read/Write Read/Write Read/Write MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 6-6 Freescale Semiconductor Register Model for 32-Bit Implementations Table 6-2. Implementation-Specific SPRs (by SPR Abbreviation) (continued) SPR Abbreviatio n IVOR32 IVOR33 IVOR34 IVOR35 L1CFG0 L1CFG1 L1CSR0 L1CSR1 MAS0 MAS1 MAS2 MAS3 MAS4 MAS6 MCAR MCSR MCSRR0 MCSRR1 MMUCFG MMUCSR0 PID0 PID1 PID2 SPEFSCR SVR Name SPR Number Access Superviso r Only Yes Yes Yes Yes No No Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes No Yes 6.14/6-48 6.5.4/6-14 6.11.3/6-32 6.11.4/6-33 6.11.1/6-30 6.11.2/6-31 6.12.5.1/6-3 7 6.12.5.2/6-3 8 6.12.5.3/6-3 9 6.12.5.4/6-4 0 6.12.5.5/6-4 0 6.12.5.6/6-4 1 6.7.2.3/6-22 6.7.2.4/6-23 6.7.2.1/6-21 6.7.2.2/6-22 6.12.3/6-35 6.12.2/6-34 6.12.1/6-34 Section/ Page 6.7.1.7/6-19 SPE APU unavailable interrupt offset Floating-point data exception interrupt offset Floating-point round exception interrupt offset Performance monitor L1 cache configuration register 0 L1 cache configuration register 1 L1 cache control and status register 0 L1 cache control and status register 1 MMU assist register 0 MMU assist register 1 MMU assist register 2 MMU assist register 3 MMU assist register 4 MMU assist register 6 Machine check address register Machine check syndrome register Machine check save/restore register 0 Machine check save/restore register 1 MMU configuration register MMU control and status register 0 Process ID register 0. Book E refers to this as PID instead of PID0. Process ID register 1 Process ID register 2 Signal processing and embedded floating-point status and control register System version register 528 529 530 531 515 516 1010 1011 624 625 626 627 628 630 573 572 570 571 1015 1012 48 633 634 512 1023 Read/Write Read/Write Read/Write Read/Write Read-only Read-only Read/Write Read/Write Read/Write Read/Write Read/Write Read/Write Read/Write Read/Write Read-only Read/Write Read/Write Read/Write Read-only Read/Write Read/Write Read/Write Read/Write Read/Write Read-only MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 6-7 Core Register Summary Table 6-2. Implementation-Specific SPRs (by SPR Abbreviation) (continued) SPR Abbreviatio n TLB0CFG TLB1CFG Name SPR Number Access Superviso r Only Yes Yes Section/ Page 6.12.4/6-35 6.12.4.2/6-3 6 TLB configuration register 0 TLB configuration register 1 688 689 Read-only Read-only 6.3 Registers for Computational Operations NOTE Register fields designated as write-one-to-clear are cleared only by writing ones to them. Writing zeros to them has no effect. The following sections describe general-purpose and integer exception registers. 6.3.1 General-Purpose Registers (GPRs) Book E implementations provide 32 GPRs (GPR0–GPR31) for integer operations. The instruction formats provide 5-bit fields for specifying the GPRs to be used in the execution of the instruction. Each GPR is a 64-bit register, although only 64-bit load, store, and merge instructions use GPR bits 0–31. 6.3.2 Field Reset R/W SPR Integer Exception Register (XER) 32 33 34 35 56 57 63 SO OV CA — All zeros R/W SPR 1 Number of bytes Figure 6-2. Integer Exception Register (XER) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 6-8 Freescale Semiconductor Registers for Branch Operations Table 6-3. XER Field Description Bits 32 Name SO Description Summary overflow. Set when an instruction (except mtspr) sets the overflow bit. Once set, SO remains set until it is cleared by mtspr[XER] or mcrxr. SO is not altered by compare instructions or by other instructions (except mtspr[XER] and mcrxr) that cannot overflow. Executing mtspr[XER], supplying the values 0 for SO and 1 for OV, causes SO to be cleared and OV to be set. Overflow. X-form add, subtract from, and negate instructions having OE = 1 set OV if the carry out of bit 32 is not equal to the carry out of bit 33, and clear OV otherwise to indicate a signed overflow. X-form multiply low word and divide word instructions having OE = 1 set OV if the result cannot be represented in 32 bits (mullwo, divwo, and divwuo) and clear OV otherwise. OV is not altered by compare instructions or by other instructions (except mtspr[XER] and mcrxr) that cannot overflow. Carry. Add carrying, subtract from carrying, add extended, and subtract from extended instructions set CA if there is a carry out of bit 32 and clear it otherwise. CA can be used to indicate unsigned overflow for add and subtract operations that set CA. Shift right algebraic word instructions set CA if any 1 bits are shifted out of a negative operand and clear CA otherwise. Compare instructions and instructions that cannot carry (except Shift Right Algebraic Word, mtspr[XER], and mcrxr) do not affect CA. Reserved, should be cleared. 33 OV 34 CA 35–56 — 57–63 No. of Supports emulation of load and store string instructions. Specifies the number of bytes to be bytes transferred by a load string indexed or store string indexed instruction. 6.4 6.4.1 Registers for Branch Operations Condition Register (CR) 32 35 36 39 40 43 44 47 48 51 52 55 56 59 60 63 This section describes registers that support Book E branch and CR operations. Field Reset R/W CR0 CR1 CR2 CR3 CR4 CR5 CR6 CR7 All zeros R/W Figure 6-3. Condition Register (CR) Table 6-4. BI Operand Settings for CR Fields CRn Bits CR Bits CR0[0] 32 BI Description 00000 Negative (LT)—Set when the result is negative. For SPE APU vector compare and vector test instructions: Set if the high-order element of rA is equal to the high-order element of rB; cleared otherwise. 00001 Positive (GT)—Set when the result is positive (and not zero). For SPE APU vector compare and vector test instructions: Set if the low-order element of rA is equal to the low-order element of rB; cleared otherwise. CR0[1] 33 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 6-9 Core Register Summary Table 6-4. BI Operand Settings for CR Fields (continued) CRn Bits CR Bits CR0[2] 34 BI Description 00010 Zero (EQ)—Set when the result is zero. For SPE APU vector compare and vector test instructions: Set to the OR of the result of the compare of the high and low elements. 00011 Summary overflow (SO). Copy of XER[SO] at the instruction’s completion. For SPE APU vector compare and vector test instructions: Set to the AND of the result of the compare of the high and low elements. 00100 Negative (LT) For SPE APU vector compare and vector test instructions: Set if the high-order element of rA is equal to the high-order element of rB; cleared otherwise. 00101 Positive (GT) For SPE APU vector compare and vector test instructions: Set if the low-order element of rA is equal to the low-order element of rB; cleared otherwise. 00110 Zero (EQ) For SPE APU vector compare and vector test instructions: Set to the OR of the result of the compare of the high and low elements. 00111 Summary overflow (SO) For SPE APU vector compare and vector test instructions: Set to the AND of the result of the compare of the high and low elements. 01000 01100 10000 10100 11000 11100 01001 01101 10001 10101 11001 11101 01010 01110 10010 10110 11010 11110 01011 01111 10011 10111 11011 11111 Less than (LT) For integer compare instructions: rA < SIMM or rB (signed comparison) or rA < UIMM or rB (unsigned comparison). For SPE APU vector compare and vector test instructions: Set if the high-order element of rA is equal to the high-order element of rB; cleared otherwise. Greater than (GT) For integer compare instructions: rA > SIMM or rB (signed comparison) or rA > UIMM or rB (unsigned comparison). For SPE APU vector compare and vector test instructions: Set if the low-order element of rA is equal to the low-order element of rB; cleared otherwise. Equal (EQ) For integer compare instructions: rA = SIMM, UIMM, or rB. For SPE APU vector compare and vector test instructions: Set to the OR of the result of the compare of the high and low elements. CR0[3] 35 CR1[0] 36 CR1[1] 37 CR1[2] 38 CR1[3] 39 CRn[0] 40 44 48 52 56 60 41 45 49 53 57 61 42 46 50 54 58 62 43 47 51 55 59 63 CRn[1] CRn[2] CRn[3] Summary overflow (SO) For integer compare instructions, this is a copy of XER[SO] at the completion of the instruction. For SPE APU vector compare and vector test instructions: Set to the AND of the result of the compare of the high and low elements. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 6-10 Freescale Semiconductor Processor Control Registers The bits of CR0 are interpreted as described in Table 6-5. Table 6-5. CR0 Bit Descriptions CR Bit 32 33 34 35 Name Negative (LT) Positive (GT) Zero (EQ) Bit 32 of the result is equal to 1. Bit 32 of the result is equal to 0 and at least one bit from 33–63 of the result is non-zero. Bits 32–63 of the result are equal to 0. Description Summary overflow (SO) This is a copy of the final state of XER[SO] at the completion of the instruction. 6.4.2 Link Register (LR) 32 63 Field Reset R/W SPR Link address All zeros R/W SPR 8 Figure 6-4. Link Register (LR) 6.4.3 Count Register (CTR) 32 63 Field Reset R/W SPR Count value All zeros R/W SPR 9 Figure 6-5. Count Register (CTR) 6.5 Processor Control Registers This section addresses machine state, processor ID, and processor version registers. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 6-11 Core Register Summary 6.5.1 32 Machine State Register (MSR) 36 37 38 39 44 45 46 47 48 49 50 51 52 53 54 55 57 58 59 60 61 62 63 Field Reset R/W — UCLE SPE — WE CE — EE PR — ME — UBLE DE All zeros R/W — IS DS — PMM — Figure 6-6. Machine State Register (MSR) Table 6-6. MSR Field Descriptions Bits 32–36 37 Name — Reserved, should be cleared.1 Description UCLE User-mode cache lock enable (defined by the EIS). Used to restrict user-mode cache-line locking by the operating system 0 Any cache lock instruction executed in user-mode takes a cache-locking DSI exception and sets either ESR[DLK] or ESR[ILK]. This allows the operating system to manage and track the locking/unlocking of cache lines by user-mode tasks. 1 Cache-locking instructions can be executed in user-mode and they do not take a DSI for cache-locking (they may still take a DSI for access violations though). SPE SPE enable (defined by the EIS) 0 If software attempts to execute an SPE APU or SPFP instruction, an SPE APU unavailable exception is taken. 1 Software can execute supported SPE and SPFP APU instructions. Note: The SPE APU and SPFP APU functionality will be implemented in the all PowerQUICC III devices. However, these instructions will not be supported in devices subsequent to PowerQUICC III. Freescale Semiconductor strongly recommends that use of these instructions be confined to libraries and device drivers. Customer software that uses SPE or SPFP APU instructions at the assembly level or that uses SPE intrinsics will require rewriting for upward compatibility with next generation PowerQUICC devices. Freescale Semiconductor offers a libmoto_e500 library that uses SPE and SPFP APU instructions. Freescale will also provide future libraries to support next-generation PowerQUICC devices. Reserved, should be cleared. 1 Wait state enable. Allows the core complex to signal a request for power management, according to the states of HID0[DOZE], HID0[NAP], and HID0[SLEEP]. 0 The processor is not in wait state and continues processing. No power management request is signaled to external logic. 1 The processor enters wait state by ceasing to execute instructions and entering low-power mode. Details of how wait state is entered and exited and how the processor behaves in the wait state are implementation-dependent. On the e500, MSR[WE] gates the DOZE, NAP, and SLEEP outputs from the core complex; as a result, these outputs negate to the external power management logic on entry to the interrupt and then return to their previous state on return from the interrupt. WE is cleared on entry to any interrupt and restored to its previous state upon return. Critical enable 0 Critical input and watchdog timer interrupts are disabled. 1 Critical input and watchdog timer interrupts are enabled. Preserved for Book III ILE 38 39–44 45 — WE 46 CE 47 — MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 6-12 Freescale Semiconductor Processor Control Registers Table 6-6. MSR Field Descriptions (continued) Bits 48 Name EE Description External enable 0 External input, decrementer, fixed-interval timer, and performance monitor interrupts are disabled. 1 External input, decrementer, fixed-interval timer, and performance monitor interrupts are enabled. User mode (problem state) 0 The processor is in supervisor mode, can execute any instruction, and can access any resource (for example, GPRs, SPRs, and the MSR). 1 The processor is in user mode, cannot execute any privileged instruction, and cannot access any privileged resource. PR also affects memory access control Reserved, should be cleared. 1 Machine check enable 0 Machine check interrupts are disabled. 1 Machine check interrupts are enabled. Reserved, should be cleared. 1 49 PR 50 51 — ME 52 53 — UBLE In the e500, it is the user BTB lock enable bit. 0 Execution of the BTB lock instructions for user mode is disabled; privileged instruction exception taken instead. 1 Execution of the BTB lock instructions for user mode is enabled. DE Debug interrupt enable 0 Debug interrupts are disabled. 1 Debug interrupts are enabled if DBCR0[IDM] = 1. See the description of the DBSR[UDE] in Section 6.13.2, “Debug Status Register (DBSR).” Reserved, should be cleared. 1 Instruction address space 0 The processor directs all instruction fetches to address space 0 (TS = 0 in the relevant TLB entry). 1 The processor directs all instruction fetches to address space 1 (TS = 1 in the relevant TLB entry). Data address space 0 The processor directs data memory accesses to address space 0 (TS = 0 in the relevant TLB entry). 1 The processor directs data memory accesses to address space 1 (TS = 1 in the relevant TLB entry). Reserved, should be cleared. 1 54 55–57 58 — IS 59 DS 60 61 — PMM Performance monitor mark bit (defined by the EIS). System software can set PMM when a marked process is running to enable statistics to be gathered only during execution of the marked process. MSR[PR] and MSR[PMM] together define a state that the processor (supervisor or user) and the process (marked or unmarked) may be in at any time. If this state matches an individual state specified in the PMLCax, the state for which monitoring is enabled, counting is enabled. — Preserved for OEA-defined RI and LE, respectively 62–63 1 An MSR bit that is reserved may be altered by a return from interrupt instruction. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 6-13 Core Register Summary 6.5.2 Processor ID Register (PIR) 32 63 Field Reset R/W SPR Processor ID All zeros Read only SPR 286 Figure 6-7. Processor ID Register (PIR) 6.5.3 Processor Version Register (PVR) 32 47 48 63 Field Reset R/W SPR Version 0x8020_nnnn Read only SPR 287 Revision Figure 6-8. Processor Version Register (PVR) Table 6-7. PVR Field Descriptions Bits 32–47 Name Version Description A 16-bit number that identifies the version of the processor. Different version numbers indicate major differences between processors, such as which optional facilities and instructions are supported. 48–63 Revision A 16-bit number that distinguishes between implementations of the version. Different revision numbers indicate minor differences between processors having the same version number, such as clock rate and engineering change level. 6.5.4 System Version Register (SVR) 32 63 Field Reset R/W SPR System version SoC-dependent value: 0x8030_nnnn Read only SPR 1023 Figure 6-9. System Version Register (SVR) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 6-14 Freescale Semiconductor Timer Registers 6.6 6.6.1 Timer Registers Timer Control Register (TCR) 32 33 34 35 36 37 38 39 40 41 42 43 46 47 50 51 63 Field WP WRC WIE DIE Reset R/W SPR FP FIE ARE — WPEXT FPEXT — All zeros R/W SPR 340 Figure 6-10. Timer Control Register (TCR) Table 6-8. TCR Field Descriptions Bits 32–33 Name WP Description Watchdog timer period. When concatenated with WPEXT, specifies one of 64-bit locations of the time base used to signal a watchdog timer exception on a transition from 0 to 1. WPEXT[0–3] || WP[0–1] = 0b00_0000 selects TBU[32] (the msb of the TB) WPEXT[0–3] || WP[0–1] = 0b11_1111 selects TBL[63] (the lsb of the TB) Watchdog timer reset control. This value is written into TSR[WRS] when a watchdog event occurs. WRC may be set by software but cannot be cleared by software, except by a software-induced reset. Once written to a non-zero value, WRC may no longer be altered by software. 00 No watchdog timer reset will occur. 01 A second timeout is ignored, regardless of the value of MSR[ME]. 10 Assert processor reset output (core_hreset_req) on second timeout of watchdog timer 11 Reserved 36 WIE Watchdog timer interrupt enable 0 Watchdog timer interrupts disabled 1 Watchdog timer interrupts enabled Decrementer interrupt enable 0 Decrementer interrupts disabled 1 Decrementer interrupts enabled Fixed interval timer period. When concatenated with FPEXT, FP specifies one of 64 bit locations of the time base used to signal a fixed-interval timer exception on a transition from 0 to 1. FPEXT[0–3] || FP[0–1] = 0b00_0000 selects TBU[32] (the msb of the TB) FPEXT[0–3] || FP[0–1] = 0b11_1111 selects TBL[63] (the lsb of the TB) Fixed interval interrupt enable 0 Fixed interval interrupts disabled 1 Fixed interval interrupts enabled Auto-reload enable. Controls whether the DECAR value is reloaded into the DEC when the DEC value reaches 0000_0001. See EREF: A Reference for Freescale Semiconductor Book E and the e500 Core. 0 Auto-reload disabled 1 Auto-reload enabled Reserved, should be cleared. 34–35 WRC 37 DIE 38–39 FP 40 FIE 41 ARE 42 — 43–46 WPEXT Watchdog timer period extension (see the description for WP) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 6-15 Core Register Summary Table 6-8. TCR Field Descriptions (continued) Bits 47–50 51–63 Name Description FPEXT Fixed-interval timer period extension (see the description for FP) — Reserved, should be cleared. 6.6.2 Timer Status Register (TSR) 32 33 34 35 36 37 38 63 Field ENW WIS Reset WRS DIS FIS All zeros — R/W Set by hardware. Read with mfspr and cleared with mtspr by writing ones to any TSR bit positions to be cleared and zeros in all other bit positions. SPR SPR 336 Figure 6-11. Timer Status Register (TSR) Table 6-9. TSR Field Descriptions Bits 32 Name Description ENW Enable next watchdog time. Functions as write-one-to-clear. 0 Action on next watchdog timer time-out is to set TSR[ENW] 1 Action on next watchdog timer time-out is governed by TSR[WIS] When a watchdog timer time-out occurs while WIS = 0 and the next watchdog time-out is enabled (ENW = 1), a watchdog timer exception is generated and logged by setting WIS. This is referred to as a watchdog timer first time out. A watchdog timer interrupt occurs if enabled by TCR[WIE] and MSR[CE]. To avoid another watchdog timer interrupt once MSR[CE] is reenabled, (assuming TCR[WIE] is not cleared instead), the interrupt handler must reset TSR[WIS]. WIS Watchdog timer interrupt status. Functions as write-one-to-clear. 0 A watchdog timer event has not occurred. 1 A watchdog timer event occurred. When MSR[CE] = 1 and TCR[WIE] = 1, a watchdog timer interrupt is taken. See the description of ENW for more information about how WIS is used. 33 34–35 WRS Watchdog timer reset status. Functions as write-one-to-clear. Defined at reset (value = 00). Set to TCR[WRC] when a reset is caused by the watchdog timer. 36 DIS Decrementer interrupt status. Functions as write-one-to-clear. 0 A decrementer event has not occurred. 1 A decrementer event occurred. When MSR[EE] = TCR[DIE] = 1, a decrementer interrupt is taken. Fixed-interval timer interrupt status. Functions as write-one-to-clear. 0 A fixed-interval timer event has not occurred. 1 A fixed-interval timer event occurred. When MSR[EE] = 1 and TCR[FIE] = 1, a fixed-interval timer interrupt is taken. Reserved, should be cleared. 37 FIS 38–63 — MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 6-16 Freescale Semiconductor Timer Registers 6.6.3 Field Reset R/W SPR Time Base Registers 32 63 32 63 TBU All zeros User read/Supervisor write 269 read/285 write TBL All zeros User read/Supervisor write 268 read/284 write Figure 6-12. Time Base Upper/Lower Registers (TBU/TBL) 6.6.4 Field Reset R/W SPR Decrementer Register 32 63 Decrementer value All zeros R/W SPR 22 Figure 6-13. Decrementer Register (DEC) 6.6.5 Field Reset R/W SPR Decrementer Auto-Reload Register (DECAR) 32 63 Decrementer auto-reload value All zeros Write only SPR 54 Figure 6-14. Decrementer Auto-Reload Register (DECAR) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 6-17 Core Register Summary 6.7 6.7.1 6.7.1.1 Interrupt Registers Interrupt Registers Defined by Book E Save/Restore Register 0 (SRR0) 32 63 Field Reset R/W SPR Next instruction address All zeros R/W SPR 26 Figure 6-15. Save/Restore Register 0 (SRR0) 6.7.1.2 Save/Restore Register 1 (SRR1) 32 63 Field Reset R/W SPR MSR state information All zeros R/W SPR 27 Figure 6-16. Save/Restore Register 1 (SRR1) 6.7.1.3 Critical Save/Restore Register 0 (CSRR0) 32 63 Field Reset R/W SPR Next instruction address All zeros R/W SPR 58 Figure 6-17. Critical Save/Restore Register 0 (CSRR0) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 6-18 Freescale Semiconductor Interrupt Registers 6.7.1.4 Critical Save/Restore Register 1 (CSRR1) 32 63 Field Reset R/W SPR MSR state information All zeros R/W SPR 59 Figure 6-18. Critical Save/Restore Register 1 (CSRR1) 6.7.1.5 Data Exception Address Register (DEAR) 32 63 Field Reset R/W SPR Exception address All zeros R/W SPR 61 Figure 6-19. Data Exception Address Register (DEAR) 6.7.1.6 Field Reset R/W SPR Interrupt Vector Prefix Register (IVPR) 32 47 48 63 Interrupt vector prefix All zeros R/W SPR 63 — Figure 6-20. Interrupt Vector Prefix Register (IVPR) 6.7.1.7 32 Interrupt Vector Offset Registers (IVORn) 47 48 59 60 63 Field Reset R/W SPR — All zeros R/W (See Table 6-10.) Interrupt vector offset — Figure 6-21. Interrupt Vector Offset Registers (IVORn) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 6-19 Core Register Summary Table 6-10. IVOR Assignments IVOR Number IVOR0 IVOR1 IVOR2 IVOR3 IVOR4 IVOR5 IVOR6 IVOR8 IVOR10 IVOR11 IVOR12 IVOR13 IVOR14 IVOR15 IVOR16–IVOR31 IVOR32 IVOR33 IVOR34 IVOR35 IVOR36–IVOR63 SPR 400 401 402 403 404 405 406 408 410 411 412 413 414 415 — 528 529 530 531 — Critical input Machine check Data storage Instruction storage External input Alignment Program System call Decrementer Fixed-interval timer interrupt Watchdog timer interrupt Data TLB error Instruction TLB error Debug Reserved for future architectural use SPE APU unavailable (EIS defined) Floating-point data exception (EIS defined) Floating-point round exception (EIS defined) Performance monitor (EIS defined) Allocated for implementation-dependent use Interrupt Type 6.7.1.8 32 Exception Syndrome Register (ESR) 35 36 37 38 39 40 41 42 43 44 45 46 47 55 56 57 63 Field Reset R/W SPR — PIL PPR PTR — ST — DLK ILK All zeros R/W SPR 62 — BO — SPE — Figure 6-22. Exception Syndrome Register (ESR) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 6-20 Freescale Semiconductor Interrupt Registers Table 6-11. ESR Field Descriptions Bits 32–35 36 37 38 39 Name — PIL PPR PTR — Syndrome Reserved, should be cleared. (Defined by Book E as allocated.) Illegal instruction exception Privileged instruction exception Trap exception Interrupt Types — Program Program Program Defined by Book E for floating-point operations. Reserved and permanently — cleared because the e500 does not implement a Book E FPU. Setting it has no effect. Store operation Alignment, data storage, data TLB error — 40 ST 41 42 — DLK Reserved, should be cleared. Cache locking. Settings are implementation-dependent. Data storage 0 Default 1 On the e500, DLK is set when a DSI occurs because dcbtls, dcbtstls, or dcblc is executed in user mode while MSR[UCLE] = 0. Set when a DSI occurs because icbtl or icblc is executed in user mode (MSR[PR] = 1) and MSR[UCLE] = 0 Reserved, should be cleared. Byte-ordering exception Reserved, should be cleared. SPE exception bit (e500-specific) 0 Default 1 Any exception caused by an SPE or SPFP instruction Reserved, should be cleared. (Defined by Book E as allocated.) Data storage, instruction storage — SPE unavailable Data storage 43 44–45 46 47–55 56 ILK — BO — SPE 57–63 — — 6.7.2 6.7.2.1 EIS-Defined Interrupt Registers Machine Check Save/Restore Register 0 (MCSRR0) 32 63 Field Reset R/W SPR Next instruction address All zeros R/W SPR 570 Figure 6-23. Machine Check Save/Restore Register 0 (MCSRR0) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 6-21 Core Register Summary 6.7.2.2 Machine Check Save/Restore Register 1 (MCSRR1) 32 63 Field Reset R/W SPR MSR state information All zeros R/W SPR 571 Figure 6-24. Machine Check Save/Restore Register 1 (MCSRR1) 6.7.2.3 Machine Check Address Register (MCAR) 32 63 Field Reset R/W SPR Machine check address All zeros Read only SPR 573 Figure 6-25. Machine Check Address Register (MCAR) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 6-22 Freescale Semiconductor Interrupt Registers 6.7.2.4 Field Reset R/W Machine Check Syndrome Register (MCSR) 32 33 34 35 36 37 38 39 MCP ICPERR DCP_PERR DCPERR All zeros R/W — — — — 40 41 42 43 44 45 46 47 Field Reset R/W — — — — All zeros R/W — — — GL_CI 48 49 50 51 52 53 54 55 Field Reset R/W — — — — All zeros R/W — — — — 56 57 58 59 60 61 62 63 Field BUS_IAERR BUS_RAERR BUS_WAERR BUS_IBERR BUS_RBERR BUS_WBERR BUS_IPERR BUS_RPERR Reset R/W SPR All zeros R/W SPR 572 Figure 6-26. Machine Check Syndrome Register (MCSR) Table 6-12. MCSR Field Descriptions Bit 32 33 34 35 Name MCP ICPERR DCP_PERR DCPERR Description Machine check input pin Instruction cache parity error Data cache push parity error Data cache parity error Recoverable Maybe Precise Maybe Precise. (Cache-inhibited stwcx. instructions and guarded loads that initiate this error are deallocated before a machine check is taken. MCSRR0 holds the address of the next instruction, not the instruction that initiated the error.) — 36–46 47 — GL_CI Reserved, should be cleared. Set when a guarded load or cache-inhibited stwcx. causes a cache parity error (ICPERR, DCPERR)—MCSRR0 holds the address for the instruction after the instruction that caused the error. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 6-23 Core Register Summary Table 6-12. MCSR Field Descriptions (continued) Bit 48–55 56 57 58 59 60 61 62 63 Name — BUS_IAERR BUS_RAERR BUS_WAERR BUS_IBERR BUS_RBERR BUS_WBERR BUS_IPERR BUS_RPERR Description Reserved, should be cleared. Bus instruction address error Bus read address error Bus write address error Bus instruction data bus error Bus read data bus error Bus write bus error Bus instruction parity error Bus read parity error Recoverable — Maybe Maybe Maybe Maybe Maybe Maybe Maybe Maybe 6.8 Field Reset SPR R/W Software-Use SPRs (SPRG0–SPRG7 and USPRG0) 32 63 Software-determined information All zeros SPRG0 SPRG1 SPRG2 SPRG3 272 273 274 259 275 SPRG4 260 276 SPRG5 261 277 SPRG6 262 278 SPRG7 263 279 USPRG0 256 Read/Write Read/Write Read/Write Read-only Read/Write Read-only Read/Write Read-only Read/Write Read-only Read/Write Read-only Read/Write Read/Write Supervisor Supervisor Supervisor User/Supervisor Supervisor User/Supervisor Supervisor User/Supervisor Supervisor User/Supervisor Supervisor User/Supervisor Supervisor User/Supervisor Figure 6-27. Software-Use SPRs (SPRG0–SPRG7 and USPRG0) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 6-24 Freescale Semiconductor Branch Target Buffer (BTB) Registers 6.9 6.9.1 Field Reset R/W SPR Branch Target Buffer (BTB) Registers Branch Buffer Entry Address Register (BBEAR) 32 61 62 63 Branch buffer entry address All zeros R/W SPR 513 IAB[0–1] Figure 6-28. Branch Buffer Entry Address Register (BBEAR) Table 6-13. BBEAR Field Descriptions Bits 32–61 62–63 Name Branch buffer entry address IAB[0–1] Description Branch buffer entry effective address bits 0–29 Instruction after branch (with BBTAR[62]). 3-bit pointer that points to the instruction in the cache line after the branch. See the description in the bblels instruction in the EREF. If the branch is the last instruction in the cache block, IAB = 000, to indicate the next sequential instruction, which resides in the zeroth position of the next cache block. 6.9.2 Field Reset R/W SPR Branch Buffer Target Address Register (BBTAR) 32 61 62 63 Branch buffer target address All zeros R/W SPR 514 IAB2 BDIRPR Figure 6-29. Branch Buffer Target Address Register (BBTAR) Table 6-14. BBTAR Field Descriptions Bits 32–61 Name Description Branch buffer target effective address bits 0–29 Branch buffer target address MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 6-25 Core Register Summary Table 6-14. BBTAR Field Descriptions (continued) Bits 62 Name IAB2 Description Instruction after branch bit 2 (with BBEAR[62–63]). IAB is a 3-bit pointer that points to the instruction in the cache line after the branch. See the description for bblels in the EREF. If the branch is the last instruction in the cache block, IAB = 000, to indicate the next sequential instruction, which resides in the zeroth position of the next cache block. Branch direction prediction. The user can pick the direction of the predicted branch. 0 The locked address is always predicted as not taken. 1 The locked address is always predicted as taken. 63 BDIRPR 6.9.3 Field Reset R/W SPR Branch Unit Control and Status Register (BUCSR) 32 53 54 55 56 57 58 62 63 — All zeros R/W BBFI BBLO BBUL BBLFC — BPEN SPR 1013 Figure 6-30. Branch Unit Control and Status Register (BUCSR) Table 6-15. BUCSR Field Descriptions Bits 32–5 3 54 Name — BBFI Reserved, should be cleared. Branch buffer flash invalidate. Clearing and then setting BBFI flash clears the valid bit of all entries in the branch buffer; clearing occurs independently from the value of the enable bit (BPEN). BBFI is always read as 0. Description 55 BBLO Branch buffer lock overflow status 0 Indicates a lock overflow condition was not encountered in the branch buffer 1 Indicates a lock overflow condition was encountered in the branch buffer This sticky bit is set by hardware and is cleared by writing 0 to this bit location. BBUL Branch buffer unable to lock 0 Indicates a lock overflow condition in the branch buffer 1 Indicates a lock set instruction failed in the branch buffer, for example, if the BTB is disabled. This sticky bit is set by hardware and is cleared by writing 0 to this bit location. BBLF C — Branch buffer lock bits flash clear. Clearing and then setting BBLFC flash clears the lock bit of all entries in the branch buffer; clearing occurs independently from the value of the enable bit (BPEN). BBLFC is always read as 0. Reserved, should be cleared. 56 57 58–6 2 63 BPEN Branch prediction enable 0 Branch prediction disabled 1 Branch prediction enabled (enables BTB to predict branches) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 6-26 Freescale Semiconductor Hardware Implementation-Dependent Registers 6.10 Hardware Implementation-Dependent Registers 6.10.1 Hardware Implementation-Dependent Register 0 (HID0) 32 33 39 40 41 42 43 48 49 50 51 62 63 Field EMCP Reset R/W SPR — DOZE NAP SLEEP — All zeros R/W SPR 1008 TBEN SEL_TBCLK — NOPTI Figure 6-31. Hardware Implementation-Dependent Register 0 (HID0) Table 6-16. HID0 Field Descriptions Bits 32 Name EMCP Description Enable machine check pin, MCP. Used to mask out further machine check exceptions caused by assertion of MCP. 0 MCP is disabled. 1 MCP is enabled. If MSE[ME] = 0, asserting MCP causes checkstop. If MSR[ME] = 1, asserting MCP causes a machine check exception. Reserved, should be cleared. Doze power management mode. If MSR[WE] is set, this bit controls DOZE mode. 0 Core not in doze mode 1 Core in doze mode Nap power management mode. If MSR[WE] is set, this bit controls NAP mode. 0 Core not in nap mode 1 Core in nap mode Configure for sleep power management mode. Controls SLEEP mode if MSR[WE] is set. 0 Core not in sleep mode 1 Core in sleep mode Reserved, should be cleared. Time base enable 0 Time base disabled (no counting) 1 Time base enabled • If HID0[TBEN] = 1 and HID0[SEL_TBCLK] = 0, the time base is updated every 8 bus clocks • If HID0[TBEN] = 1 and HID0[SEL_TBCLK] = 1, the time base is updated on the rising edge of core_TBCLK (sampled at bus rate). The maximum supported frequency can be found in the electrical specifications, but this value is approximately 25% of the bus clock frequency. 33–39 40 — DOZE 41 NAP 42 SLEEP 43–48 49 — TBEN 50 SEL_TBCL Select time base clock. If the time base is enabled, this field functions as follows: K 0 Time base is based on the processor clock 1 Time base is based on the TBCLK (RTC) input MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 6-27 Core Register Summary Table 6-16. HID0 Field Descriptions (continued) Bits 51–62 63 Name — NOPTI Reserved, should be cleared. No-op the data and instruction cache touch instructions. 0 dcbt, dcbtst, and icbt are enabled, as defined by the EIS. On the e500, if CT = 0, icbt is always a no-op, regardless of the value of NOPTI. If CT = 1, icbt does a touch load to an L2 cache, if one is present. 1 dcbt, dcbtst, and icbt are treated as no-ops; dcblc and dcbtls are not. Description 6.10.2 Hardware Implementation-Dependent Register 1 (HID1) 32 33 34 39 40 45 46 47 49 50 51 52 63 Field PLL_MODE PLL_CFG Reset R/W SPR — RFXE — ASTME ABE — 01[PLL_CFG]_0000_0000_0000_0000_0000_0000 R/W SPR 1009 Figure 6-32. Hardware Implementation-Dependent Register 1 (HID1) Table 6-17. HID1 Field Descriptions Bits 32–33 Name PLL_MOD Read-only for integrated devices. E 01 Fixed value for MPC8540 PLL_CFG Reflected directly from configuration input pins (read-only). PLL_CFG[0–4] corresponds to the integer divide ratio and PLL_CFG5 is the half-mode bit. MPC8540 supports the following: 00010 0 ratio of 2:1 00010 1 ratio of 5:2 (2.5:1) 00011 0 ratio of 3:1 00011 1 ratio of 7:2 (3.5:1) Note that this value is also reflected to PORPLLSR[e500_Ratio]. See Section 18.4.1.1, “POR PLL Status Register (PORPLLSR).” Reserved, should be cleared. Description 34–39 40–45 — MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 6-28 Freescale Semiconductor Hardware Implementation-Dependent Registers Table 6-17. HID1 Field Descriptions (continued) Bits 46 Name RFXE Description Read fault exception enable. Controls whether assertion of core_fault_in causes a machine check interrupt. The assertion of core_fault_in can result from an L2 multibit ECC error. It can also occur for a system error if logic on the integrated device signals a fault for nonfatal errors (read data is corrupt (or zero), but the bus transaction can complete without corrupting other system state, making it unnecessary to take a machine check (for example, a master abort of a PCI transaction). 0 Assertion of core_fault_in cannot cause a machine check. RFXE should be left clear if an interrupt is to be reported by the integrated device through int or c_int for this condition. If RFXE = 0, it is important that the integrated device generates an interrupt if core_fault_in is asserted. If core_fault_in is asserted, any data on the CCB is dropped, stalling the load/store unit and eventually causing the e500 pipeline either to stall until an interrupt occurs (typically generated by the programmable interrupt controller (PIC) in response to the fault) or to continue processing with bad data untill the interrupt occurs. Because core_fault_in cannot cause a machine check, if RFXE is 0, it is critical that the system be configured to generate the appropriate interrupt. It is also possible to hang the processor (requiring a hard reset to recover) if a guarded load hits in the L2 cache and gets an uncorrectable ECC error. Because of this, avoid defining memory as cacheable but guarded. If this combination is required, RFXE must be enabled, in which case an error causes both a machine check interrupt and an external interrupt when a bus fault condition is detected unless interrupts are masked for all sources of bus faults, such as DRAM ECC errors, PCI parity errors, local bus parity errors, and others. RFXE must also be set if software requires that code execution stop immediately when a bus fault occurs rather than continuing with the bad data until the interrupt arrives. Again, this results in both a machine check interrupt and an external interrupt when a bus fault is detected, unless all possible sources for bus fault have their interrupts masked.þThe machine check interrupt can then reenable normal interrupts and wait for the interrupt due to the fault to be received before returning from the machine check. 1 A machine check can occur due to assertion of core_fault_in. If MSR[ME] = 1 and a fault is signaled, a machine check interrupt occurs. If MSR[ME] = 0 and a fault is signaled, a checkstop occurs. Note that if RFXE is set and another mechanism is configured to generate an interrupt in response to assertion of core_fault_in, the same event causes two interrupts, the machine check enabled by setting RFXE and the interrupt triggered by the on-chip peripheral or other block; therefore, RFXE should be set only if no other mechanism is configured to generate an interrupt for this case. Note that the L2 cache detects any assertion of core_fault_in and ensures that the L2 cache is not corrupted when data is dropped for this type of transaction. Reserved, should be cleared. Address bus streaming mode enable. This bit, along with the ECM stream control bits in the EEBACR, enables address bus streaming on the CCB. See Section 8.2.1.1, “ECM CCB Address Configuration Register (EEBACR).” 0 Address bus streaming mode disabled 1 Address bus streaming mode enabled Address broadcast enable. The e500 broadcasts cache management instructions (dcbst, dcblc (CT = 1), icblc (CT = 1), dcbf, dcbi, mbar, msync, tlbsync, icbi) based on ABE. On the MPC8540, ABE must be set to allow management of external L2 caches. 0 Address broadcasting disabled 1 Address broadcasting enabled Reserved, should be cleared. 47–49 50 — ASTME 51 ABE 52–63 — MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 6-29 Core Register Summary 6.11 L1 Cache Configuration Registers 6.11.1 L1 Cache Control and Status Register 0 (L1CSR0) Line Locking APU Bits 32 46 47 48 49 51 52 53 54 55 56 61 62 63 Field Reset R/W SPR — CPE CPI — CSLC CUL CLO CLFR — CFI CE All zeros R/W SPR 1010 Figure 6-33. L1 Cache Control and Status Register 0 (L1CSR0) Table 6-18. L1CSR0 Field Descriptions Bits 32–46 47 Nam e — Reserved, should be cleared. Description CPE (Data) Cache parity enable 0 Parity checking of the cache disabled 1 Parity checking of the cache enabled CPI (Data) Parity error injection enable 0 Parity error injection disabled 1 Parity error injection enabled. Cache parity must also be enabled (CPE = 1) when this bit is set. Reserved, should be cleared. (Data) Cache snoop lock clear. Sticky bit set by hardware if a dcbi snoop (either internally or externally generated) invalidated a locked cache line. Note that the lock bit for that line is cleared whenever the line is invalidated. This bit can be cleared only by software. 0 The cache has not encountered an dcbi snoop that invalidated a locked line. 1 The cache has encountered an dcbi snoop that invalidated a locked line. 48 49–51 52 — CSL C 53 CUL (Data) Cache unable to lock. Sticky bit set by hardware and cleared by writing 0 to this bit location. 0 Indicates a lock set instruction was effective in the cache 1 Indicates a lock set instruction was not effective in the cache CLO (Data) Cache lock overflow. Sticky bit set by hardware and cleared by writing 0 to this bit location. 0 Indicates a lock overflow condition was not encountered in the cache 1 Indicates a lock overflow condition was encountered in the cache CLF R (Data) Cache lock bits flash reset. Writing a 1 during a flash clear operation causes an undefined operation. Writing a 0 during a flash clear operation is ignored. Clearing occurs regardless of the enable (CE) value. 0 Default. 1 Hardware initiates a cache lock bits flash clear operation. This bit is cleared when the operation is complete. Reserved, should be cleared. 54 55 56–61 — MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 6-30 Freescale Semiconductor L1 Cache Configuration Registers Table 6-18. L1CSR0 Field Descriptions (continued) Bits 62 Nam e CFI Description (Data) Cache flash invalidate. 0 No cache invalidate. Writing a 0 to CFI during an invalidation operation is ignored. 1 Cache invalidation operation. A cache invalidation operation is initiated by hardware. Once complete, this bit is cleared. Writing a 1 during an invalidation operation causes an undefined operation. Invalidation occurs regardless of the enable (CE) value. (Data) Cache enable 0 The cache is neither accessed or updated. 1 Enables cache operation 63 CE 6.11.2 L1 Cache Control and Status Register 1 (L1CSR1) Line Locking APU Bits 32 46 47 48 49 51 52 53 54 55 56 60 61 62 63 Field Reset R/W SPR — ICPE ICPI — ICSLC ICUL ICLO ICLFR — ICFI ICE All zeros R/W SPR 1011 Figure 6-34. L1 Cache Control and Status Register 1 (L1CSR1) Table 6-19. L1CSR1 Field Descriptions Bits 32–46 47 Name — Reserved, should be cleared. Description ICPE Instruction cache parity enable 0 Parity checking of the instruction cache disabled 1 Parity checking of the instruction cache enabled ICPI Instruction parity error injection enable 0 Parity error injection disabled 1 Parity error injection enabled. Note that instruction cache parity must also be enabled (ICPE = 1) when this bit is set. Reserved, should be cleared. 48 49–51 52 — ICSL Instruction cache snoop lock clear. Sticky bit set by hardware if an icbi snoop (either internally or externally C generated) invalidated a locked line in the instruction cache. Note that the lock bit for that line is cleared whenever the line is invalidated. This bit can only be cleared by software. 0 The instruction cache has not encountered an icbi snoop that invalidated a locked line. 1 The instruction cache has encountered an icbi snoop that invalidated a locked line. ICUL Instruction cache unable to lock. Sticky bit set by hardware and cleared by writing 0 to this bit location. 0 Indicates a lock set instruction was effective in the instruction cache 1 Indicates a lock set instruction was not effective in the instruction cache 53 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 6-31 Core Register Summary Table 6-19. L1CSR1 Field Descriptions (continued) Bits 54 Name Description ICLO Instruction cache lock overflow. Sticky bit set by hardware and cleared by writing 0 to this bit location. 0 Indicates a lock overflow condition was not encountered in the instruction cache 1 Indicates a lock overflow condition was encountered in the instruction cache ICLF Instruction cache lock bits flash reset. Writing 0 and then 1 flash clears the lock bit of all entries in the R instruction cache; clearing occurs independently from the value of the enable bit (ICE). ICLFR is always read as 0. — ICFI ICE Reserved, should be cleared. Instruction cache flash invalidate. Written to 0 and then 1 to flash clear the valid bit of all entries in the instruction cache; operates independently from the value of the enable bit (ICE). ICI is always read as 0. Instruction cache enable 0 The instruction cache is neither accessed or updated. 1 Enables instruction cache operation 55 56–61 62 63 6.11.3 L1 Cache Configuration Register 0 (L1CFG0) 32 33 34 38 39 40 41 42 43 44 45 49 50 52 53 55 56 63 Field CARCH Reset R/W SPR 00 — 00_000 CBSIZE CREPL CLA CPA 00 01 Read only SPR 515 1 1 — 000_00 CNWAY 11_1 — 000 CSIZE 0010_0000 Figure 6-35. L1 Cache Configuration Register 0 (L1CFG0) Table 6-20. L1CFG0 Field Descriptions Bits 32–33 Name CARCH Cache architecture 00 Harvard 01 Unified Reserved, should be cleared. Cache line size 0 32 bytes 1 64 bytes Cache replacement policy 0 True LRU 1 Pseudo LRU Cache locking APU available 0 Unavailable 1 Available Description 34–38 39–40 — CBSIZE 41–42 CREPL 43 CLA MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 6-32 Freescale Semiconductor L1 Cache Configuration Registers Table 6-20. L1CFG0 Field Descriptions (continued) Bits 44 Name CPA Cache parity available 0 Unavailable 1 Available Reserved, should be cleared. Cache number of ways 111 Indicates 8 ways Reserved, should be cleared. Cache size 0x20 indicates 32 Kbytes Description 45–49 50–52 53–55 56–63 — CNWAY — CSIZE 6.11.4 L1 Cache Configuration Register 1 (L1CFG1) 32 38 39 40 41 42 43 44 45 52 53 63 Field Reset R/W SPR — 0000_000 ICBSIZ 0_0 ICREPL ICLA ICP A 01 1 1 ICNWAY 000_0011_1 ICSIZE 000_0010_0000 Read only SPR 516 Figure 6-36. L1 Cache Configuration Register 1 (L1CFG1) Table 6-21. L1CFG1 Field Descriptions Bits 32–38 39–40 41–42 Name — ICBSIZ Reserved, should be cleared. Instruction cache block size 00 Indicates block size of 32 bytes Description ICREPL Instruction cache replacement policy 01 Indicates pseudo-LRU policy ICLA Instruction cache locking available 1 Indicates available Instruction cache parity available 1 Indicates available 43 44 ICPA 45–52 ICNWAY Instruction cache number of ways 111 Indicates 8 ways 53–63 ICSIZE Instruction cache size 0x20 indicates 32 Kbytes MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 6-33 Core Register Summary 6.12 MMU Registers 6.12.1 Process ID Registers (PID0–PID2) 32 55 56 63 Field Reset R/W SPR — All zeros R/W Process ID Book E defined: SPR 48(PID0); EIS defined: SPR 633 (PID1); SPR 634 (PID2) Figure 6-37. Process ID Registers (PID0–PID2) 6.12.2 MMU Control and Status Register 0 (MMUCSR0) 32 60 61 62 63 Field Reset R/W SPR — All zeros R/W SPR 1012 L2TLB0_FI L2TLB1_FI — Figure 6-38. MMU Control and Status Register 0 (MMUCSR0) Table 6-22. MMUCSR0 Field Descriptions Bits 32–60 61 62 Name — L2TLB0_FI L2TLB1_FI Reserved, should be cleared. TLB0 flash invalidate (write to 1 to invalidate) TLB1 flash invalidate (write 1 to invalidate) 0 No flash invalidate. Writing a 0 to this bit during an invalidation operation is ignored. 1 TLB1 invalidation operation. Hardware initiates a TLB1 invalidation operation. When this operation is complete, this bit is cleared. Writing a 1 during an invalidation operation causes an undefined operation. Reserved, should be cleared. Description 63 — MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 6-34 Freescale Semiconductor M MU Registers 6.12.3 MMU Configuration Register (MMUCFG) 32 48 49 52 53 57 58 59 60 61 62 63 Field Reset R/W SPR — 0000_0000_0000_0000_0 Read only SPR 1015 NPIDS 001_1 PIDSIZE 001_11 — 00 NTLBS MAVN 01 00 Figure 6-39. MMU Configuration Register (MMUCFG) Table 6-23. MMUCFG Field Descriptions Bits 32–48 49–52 53–57 Name — NPIDS Reserved, should be cleared. Number of PID registers, a 4-bit field that indicates the number of PID registers provided by the processor. The e500 implements three PIDs. Description PIDSIZE PID register size. The 5-bit value of PIDSIZE is one less than the number of bits in each of the PID registers implemented by the processor. The processor implements only the least significant PIDSIZE+1 bits in the PID registers. 00111 indicates 8-bit registers. This is the value presented by the e500. — NTLBS Reserved, should be cleared. Number of TLBs. The value of NTLBS is one less than the number of software-accessible TLB structures that are implemented by the processor. NTLBS is set to one less than the number of TLB structures so that its value matches the maximum value of MAS0[TLBSEL]. 00 1 TLB 01 2 TLBs. This is the value presented by the e500. 10 3 TLBs 11 4 TLBs MMU architecture version number. Indicates the version number of the architecture of the MMU implemented by the processor. 0b00 indicates version 1.0. 58–59 60–61 62–63 MAVN 6.12.4 TLB Configuration Registers (TLBnCFG) 6.12.4.1 TLB0 Configuration Register 0 (TLB0CFG) 32 39 40 43 44 47 48 49 50 51 52 63 Field Reset R/W SPR ASSOC 0000_0010 MINSIZE MAXSIZE IPROT AVAIL 0001 0001 0 0 — 00 NENTRY 0001_0000_0000 Read only SPR 688 Figure 6-40. TLB Configuration Register 0 (TLB0CFG) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 6-35 Core Register Summary Table 6-24. TLB0CFG Field Descriptions Bits 32–39 Name ASSOC Description Associativity of TLB0 0x02 indicates associativity is 2-way set associative Minimum page size of TLB0 0x1 indicates smallest page size is 4K 40–43 MINSIZE 44–47 MAXSIZE Maximum page size of TLB0 0x1 indicates maximum page size is 4K 48 IPROT Invalidate protect capability of TLB0 0 Indicates invalidate protection capability not supported Page size availability of TLB0 0 No variable-sized pages available (MINSIZE = MAXSIZE) Reserved, should be cleared. 49 AVAIL 50–51 52–63 — NENTRY Number of entries in TLB0 0x100: TLB0 contains 256 entries 6.12.4.2 TLB1 Configuration Register 1 (TLB1CFG) 32 39 40 43 44 47 48 49 50 51 52 63 Field e500 Reset Values ASSOC 0001_0000 MINSIZE MAXSIZE IPROT AVAIL 0001 1001 1 Read only SPR 689 1 — 00 NENTRY 0000_0001_0000 R/W SPR Figure 6-41. TLB Configuration Register 1 (TLB1CFG) Table 6-25. TLB1CFG Field Descriptions Bits 32–39 40–43 Name ASSOC Description Associativity of TLB1 0x10 indicates associativity is 16 MINSIZE Minimum page size of TLB1 0x1 indicates smallest page size is 4K MAXSIZE Maximum page size of TLB1 0x9 Indicates maximum page size is 256 Mbyte IPROT Invalidate protect capability of TLB1 1 Indicates that TLB1 supports invalidate protection capability Page size availability of TLB1 1 Indicates all page sizes between MINSIZE and MAXSIZE supported 44–47 48 49 AVAIL MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 6-36 Freescale Semiconductor M MU Registers Table 6-25. TLB1CFG Field Descriptions (continued) Bits 50–51 52–63 Name — Description Reserved, should be cleared. NENTRY Number of entries in TLB1 0x010: TLB1 contains 16 entries 6.12.5 MMU Assist Registers 6.12.5.1 MAS Register 0 (MAS0) 32 34 35 36 43 44 47 48 62 63 Field Reset R/W SPR — TLBSEL — ESEL All zeros R/W SPR 624 — NV Figure 6-42. MAS Register 0 (MAS0) Table 6-26. MAS0 Field Descriptions—MMU Read/Write and Replacement Control Bits 32–34 35 Name — Reserved, should be cleared. Descriptions TLBSEL Selects TLB for access 0 TLB0 1 TLB1 — ESEL Reserved, should be cleared. Entry select. Number of entry in selected array to be used for tlbwe. This field is also updated on TLB error exceptions (misses), and tlbsx hit and miss cases. For the e500, ESEL serves as the way select for the corresponding TLB as follows: When TLBSEL = 00 (TLB0 selected), only bit 47 is used (and bits 44–46 should be cleared). This bit selects between way 0 and way 1 of TLB0. EA bits 45–51 from MAS2[EPN] are used to index into the TLB to further select the entry for the operation. When TLBSEL = 01 (TLB1 selected), all four bits are used to select one of 16 entries in the array. Reserved, should be cleared. Next victim. Next victim bit value to be written to TLB0[NV] on execution of tlbwe. This field is also updated on TLB error exceptions (misses), tlbsx hit and miss cases and on execution of tlbre. This field is updated based on the calculated next victim bit for TLB0 (based on the round-robin replacement algorithm.) Note that this field is not defined for operations that specify TLB1 (when TLBSEL = 01). 36–43 44–47 48–62 63 — NV MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 6-37 Core Register Summary 6.12.5.2 MAS Register 1 (MAS1) 32 33 34 39 40 47 48 50 51 52 55 56 63 Field V IPROT Reset R/W SPR — TID All zeros R/W SPR 625 — TS TSIZE — Figure 6-43. MAS Register 1 (MAS1) Table 6-27. MAS1 Field Descriptions—Descriptor Context and Configuration Control Bits 32 Name V TLB valid bit 0 This TLB entry is invalid. 1 This TLB entry is valid. Descriptions 33 IPRO Invalidate protect. Set to protect this TLB entry from invalidate operations due the execution of tlbiva[x] T (TLB1 only). Note that not all TLB arrays are necessarily protected from invalidation with IPROT. Arrays that support invalidate protection are denoted as such in the TLB configuration registers. 0 Entry is not protected from invalidation 1 Entry is protected from invalidation. — TID Reserved, should be cleared. Translation identity. An 8-bit field that defines the process ID for this TLB entry. TID is compared with the current process IDs of the three virtual address to be translated. A TID value of 0 defines an entry as global and matches with all process IDs. Reserved, should be cleared. Translation space. This bit is compared with the IS or DS fields of the MSR (depending on the type of access) to determine if this TLB entry may be used for translation. 34–39 40–47 48–50 51 — TS 52–55 TSIZE Translation size. Defines the TLB entry page size. For arrays that contain fixed-size TLB entries, TSIZE is ignored. For variable page size arrays, the page size is 4TSIZE Kbytes. Note that although the Freescale Semiconductor Book E standard supports all 16 page sizes defined in Book E, the e500 supports only the following: 0001 4 Kbyte 0010 16 Kbyte 0011 64 Kbyte 0100 256 Kbyte 0101 1 Mbyte 0110 4 Mbyte 0111 16 Mbyte 1000 64 Mbyte 1001 256 Mbyte 56–63 — Reserved, should be cleared. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 6-38 Freescale Semiconductor M MU Registers 6.12.5.3 MAS Register 2 (MAS2) 32 51 52 56 57 58 59 60 61 62 63 Field Reset R/W SPR EPN All zeros R/W SPR 626 — X0 X1 W I M G E Figure 6-44. MAS Register 2 (MAS2) Table 6-28. MAS2 Field Descriptions—EPN and Page Attributes Bits 32–5 1 52–5 6 57 58 59 Name EPN — X0 X1 W Description Effective page number. Depending on page size, only the bits associated with a page boundary are valid. Bits that represent offsets within a page are ignored and should be cleared. Reserved for implementation-specific use Implementation-dependent page attribute Implementation-dependent page attribute Write-through 0 This page is considered write-back with respect to the caches in the system. 1 All stores performed to this page are written through the caches to main memory. Caching-inhibited 0 Accesses to this page are considered cacheable. 1 The page is considered caching-inhibited. All loads and stores to the page bypass the caches and are performed directly to main memory. Memory coherence required 0 Memory coherence is not required. 1 Memory coherence is required. This allows loads and stores to this page to be coherent with loads and stores from other processors (and devices) in the system, assuming all such devices are participating in the coherence protocol. Guarded 0 Accesses to this page are not guarded and can be performed before it is known if they are required by the sequential execution model. 1 All loads and stores to this page that miss in the L1 cache are performed without speculation (that is, they are known to be required). Speculative loads can be performed if they hit in the L1 cache. In addition, accesses to caching-inhibited pages are performed using only the memory element that is explicitly specified. Endianness. Determines endianness for the corresponding page. Little-endian operation is true little endian, which differs from the modified little-endian byte-ordering model optionally available in previous devices that implement the original PowerPC architecture. See the e500 Reference Manual for more information on the Book E definition of endianness. 0 The page is accessed in big-endian byte order. 1 The page is accessed in true little-endian byte order. 60 I 61 M 62 G 63 E MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 6-39 Core Register Summary 6.12.5.4 MAS Register 3 (MAS3) 32 51 52 53 54 57 58 59 60 61 62 63 Field Reset R/W SPR RPN All zeros R/W SPR 627 — U0–U3 UX SX UW SW UR SR Figure 6-45. MAS Register 3 (MAS3) Table 6-29. MAS3 Field Descriptions–RPN and Access Control Bits 32–51 52–53 54–57 Name RPN — Description Real page number. Depending on page size, only the bits associated with a page boundary are valid. Bits that represent offsets within a page are ignored and should be cleared. Reserved, should be cleared. U0–U3 User attribute bits. Associated with a TLB entry and can be used by system software. For example, they can hold information useful to a page-scanning algorithm or mark more abstract page attributes. PERMI Permission bits (UX, SX, UW, SW, UR, SR). User and supervisor read, write, and execute S permission bits. 58–63 6.12.5.5 MAS Register 4 (MAS4) 32 34 35 36 45 46 47 48 51 52 55 56 57 58 59 60 61 62 63 Field Reset R/W SPR — TLBSELD — TIDSELD — All zeros R/W SPR 628 TSIZED — X0D X1D WD ID MD GD ED Figure 6-46. MAS Register 4 (MAS4) Table 6-30. MAS4 Field Descriptions—Hardware Replacement Assist Configuration Bits 32–34 35 Name — Reserved, should be cleared. Description TLBSELD TLBSEL default value. The default value to be loaded in MAS0[TLBSEL] on a TLB miss exception. 0 TLB0 1 TLB1 — Reserved, should be cleared. 36–45 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 6-40 Freescale Semiconductor M MU Registers Table 6-30. MAS4 Field Descriptions—Hardware Replacement Assist Configuration (continued) Bits 46–47 Name Description TIDSELD TID default selection value. A 2-bit field that specifies which of the current PID registers should be used to load the MAS1[TID] field on a TLB miss exception. The e500 implementation defines this field as follows: 00 PID0 01 PID1 10 PID2 11 TIDZ (0x00) (all zeros) 48–51 52–55 56 57 58 59 60 61 62 63 — TSIZED — X0D X1D WD ID MD GD ED Reserved, should be cleared. Default TSIZE value. Specifies the default value to be loaded into MAS1[TSIZE] on a TLB miss exception. Reserved, should be cleared. Default X0 value. Specifies the default value to be loaded into MAS2[X0] on a TLB miss exception. Default X1 value. Specifies the default value to be loaded into MAS2[X1] on a TLB miss exception. Default W value. Specifies the default value to be loaded into MAS2[W] on a TLB miss exception. Default I value. Specifies the default value to be loaded into MAS2[I] on a TLB miss exception. Default M value. Specifies the default value to be loaded into MAS2[M] on a TLB miss exception. Default G value. Specifies the default value to be loaded into MAS2[G] on a TLB miss exception. Default E value. Specifies the default value to be loaded into MAS2[E] on a TLB miss exception. 6.12.5.6 MAS Register 6 (MAS6) 32 39 40 47 48 62 63 Field Reset R/W SPR — SPID0 All zeros R/W SPR 630 — SAS Figure 6-47. MAS Register 6 (MAS6) Table 6-31. MAS6—TLB Search Context Register 0 Bits 32–39 40–47 48–62 63 Name — SPID0 — SAS Comments, or Function when Set Reserved, should be cleared. Specifies the PID value (recent value of PID0) used when searching the TLB during execution of tlbsx. Reserved, should be cleared. Address space (AS) value for searches. Specifies the value of AS used when searching the TLB (during execution of tlbsx). MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 6-41 Core Register Summary 6.13 Debug Registers 6.13.1 Debug Control Registers (DBCR0–DBCR2) 6.13.1.1 Debug Control Register 0 (DBCR0) 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 62 63 Field — IDM RST ICMP BRT IRPT TRAP IAC1 IAC2 Reset R/W SPR — DAC1 DAC2 RET — FT All zeros R/W SPR 308 Figure 6-48. Debug Control Register 0 (DBCR0) Table 6-32. DBCR0 Field Descriptions Bits 32 33 Name — IDM Reserved, should be cleared. Internal debug mode 0 Debug interrupts are disabled. No debug interrupts are taken and debug events are not logged. 1 If MSR[DE] = 1, the occurrence of a debug event or the recording of an earlier debug event in the DBSR when MSR[DE] = 0 or DBCR0[IDM] = 0 causes a debug interrupt. Programming note: Software must clear debug event status in the DBSR in the debug interrupt handler when a debug interrupt is taken before re-enabling interrupts through MSR[DE]. Otherwise, redundant debug interrupts are taken for the same debug event. Reset. Book E defines this field such that 00 is always no action and all other settings are implementation specific. The e500 implements these bits as follows: 0x Default (No action) 1x Causes a hard reset if MSR[DE] and DBCR0[IDM] are set. Always cleared on subsequent cycle. Instruction completion debug event enable 0 ICMP debug events are disabled 1 ICMP debug events are enabled Note: Instruction completion does not cause an ICMP debug event if MSR[DE] = 0. Branch taken debug event enable 0 BRT debug events are disabled 1 BRT debug events are enabled Note: Taken branches do not cause a BRT debug event if MSR[DE] = 0. Interrupt taken debug event enable. This bit affects only noncritical interrupts. 0 IRPT debug events are disabled 1 IRPT debug events are enabled Trap debug event enable 0 TRAP debug events cannot occur 1 TRAP debug events can occur Description 34–35 RST 36 ICMP 37 BRT 38 IRPT 39 TRAP MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 6-42 Freescale Semiconductor Debug Registers Table 6-32. DBCR0 Field Descriptions (continued) Bits 40 Name IAC1 Description Instruction address compare 1 debug event enable 0 IAC1 debug events cannot occur 1 IAC1 debug events can occur Instruction address compare 2 debug event enable 0 IAC2 debug events cannot occur 1 IAC2 debug events can occur Reserved, should be cleared. Data address compare 1 debug event enable 00 DAC1 debug events cannot occur 01 DAC1 debug events can occur only if a store-type data storage access 10 DAC1 debug events can occur only if a load-type data storage access 11 DAC1 debug events can occur on any data storage access Data address compare 2 debug event enable 00 DAC2 debug events cannot occur 01 DAC2 debug events can occur only if a store-type data storage access 10 DAC2 debug events can occur only if a load-type data storage access 11 DAC2 debug events can occur on any data storage access Return debug event enable 0 RET debug events cannot occur 1 RET debug events can occur Note: An rfci does not cause an RET debug event if MSR[DE] = 0 at the time that rfci executes. Reserved, should be cleared. Freeze timers on debug event 0 Enable clocking of timers 1 Disable clocking of timers if any DBSR bit is set (except MRR) 41 IAC2 42–43 44–45 — DAC1 46–47 DAC2 48 RET 49–62 63 — FT 6.13.1.2 Debug Control Register 1 (DBCR1) 32 33 34 35 36 37 38 39 40 41 42 63 Field IAC1US IAC1ER IAC2US IAC2ER IAC12M Reset R/W SPR All zeros R/W SPR 309 — Figure 6-49. Debug Control Register 1 (DBCR1) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 6-43 Core Register Summary Table 6-33. DBCR1 Field Descriptions Bits Name Description 32–33 IAC1US Instruction address compare 1 user/supervisor mode 00 IAC1 debug events can occur 01 Reserved 10 IAC1 debug events can occur only if MSR[PR] = 0 11 IAC1 debug events can occur only if MSR[PR] = 1 34–35 IAC1ER Instruction address compare 1 effective/real mode 00 IAC1 debug events are based on effective addresses 01 Reserved on the e500 10 IAC1 debug events are based on effective addresses and can occur only if MSR[IS] = 0 11 IAC1 debug events are based on effective addresses and can occur only if MSR[IS] = 1 36–37 IAC2US Instruction address compare 2 user/supervisor mode 00 IAC2 debug events can occur 01 Reserved 10 IAC2 debug events can occur only if MSR[PR] = 0 11 IAC2 debug events can occur only if MSR[PR] = 1 38–39 IAC2ER Instruction address compare 2 effective/real mode 00 IAC2 debug events are based on effective addresses 01 Reserved on the e500 10 IAC2 debug events are based on effective addresses and can occur only if MSR[IS] = 0 11 IAC2 debug events are based on effective addresses and can occur only if MSR[IS] = 1 40–41 IAC12M Instruction address compare 1/2 mode 00 Exact address compare. IAC1 debug events can occur only if the address of the instruction fetch is equal to the value specified in IAC1. IAC2 debug events can occur only if the address of the instruction fetch is equal to the value specified in IAC2. 01 Address bit match. IAC1 and IAC2 debug events can occur only if the address of the instruction fetch, ANDed with the contents of IAC2 are equal to the contents of IAC1,plus ANDed with the contents of IAC2. If IAC1US≠IAC2US or IAC1ER≠IAC2ER, results are boundedly undefined. 10 Inclusive address range compare. IAC1 and IAC2 debug events can occur only if the address of the instruction fetch is greater than or equal to the value specified in IAC1 and less than the value specified in IAC2. If IAC1US≠IAC2US or IAC1ER≠IAC2ER, results are boundedly undefined. 11 Exclusive address range compare. IAC1 and IAC2 debug events can occur only if the address of the instruction fetch is less than the value specified in IAC1 or is greater than or equal to the value specified in IAC2. If IAC1US≠IAC2US or IAC1ER≠IAC2ER, results are boundedly undefined. 42–63 — Reserved, should be cleared. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 6-44 Freescale Semiconductor Debug Registers 6.13.1.3 Debug Control Register 2 (DBCR2) 32 33 34 35 36 37 38 39 40 41 42 63 Field DAC1US DAC1ER DAC2US DAC2ER DAC12M Reset R/W SPR All zeros R/W SPR 310 — Figure 6-50. Debug Control Register 2 (DBCR2) Table 6-34. DBCR2 Field Descriptions Bits 32–33 Name Description DAC1US Data address compare 1 user/supervisor mode 00 DAC1 debug events can occur 01 Reserved 10 DAC1 debug events can occur only if MSR[PR] = 0. 11 DAC1 debug events can occur only if MSR[PR] = 1. DAC1ER Data address compare 1 effective/real mode 00 DAC1 debug events are based on effective addresses. 01 Reserved on the e500 10 DAC1 debug events are based on effective addresses and can occur only if MSR[DS] = 0. 11 DAC1 debug events are based on effective addresses and can occur only if MSR[DS] = 1. DAC2US Data address compare 2 user/supervisor mode 00 DAC2 debug events can occur. 01 Reserved 10 DAC2 debug events can occur only if MSR[PR] = 0. 11 DAC2 debug events can occur only if MSR[PR] = 1. DAC2ER Data address compare 2 effective/real mode 00 DAC2 debug events are based on effective addresses. 01 Reserved on the e500 10 DAC2 debug events are based on effective addresses and can occur only if MSR[DS] = 0. 11 DAC2 debug events are based on effective addresses and can occur only if MSR[DS] = 1. 34–35 36–37 38–39 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 6-45 Core Register Summary Table 6-34. DBCR2 Field Descriptions (continued) Bits 40–41 Name Description DAC12M Data address compare 1/2 mode 00 Exact address compare. DAC1 debug events can occur only if the address of the data storage access is equal to the value specified in DAC1. DAC2 debug events can occur only if the address of the data storage access is equal to the value specified in DAC2. 01 Address bit match. DAC1 and DAC2 debug events can occur only if the address of the data storage access, ANDed with the contents of DAC2 are equal to the contents of DAC1, also ANDed with the contents of DAC2. If DAC1US ≠ DAC2US or DAC1ER ≠ DAC2ER, results are boundedly undefined. 10 Inclusive address range compare. DAC1 and DAC2 debug events can occur only if the address of the data storage access is greater than or equal to the value specified in DAC1 and less than the value specified in DAC2. If DAC1US ≠ DAC2US or DAC1ER ≠ DAC2ER, results are boundedly undefined. 11 Exclusive address range compare. DAC1 and DAC2 debug events can occur only if the address of the data storage access is less than the value specified in DAC1 or is greater than or equal to the value specified in DAC2. If DAC1US ≠ DAC2US or DAC1ER ≠ DAC2ER, results are boundedly undefined. — Reserved, should be cleared. 42–63 6.13.2 Debug Status Register (DBSR) 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 63 Field IDE UDE MRR ICMP BRT IRPT TRAP IAC1 IAC2 Reset R/W SPR 00 Undef. — DAC1R DAC1W DAC2R DAC2W RET — 0000_0000_0000 R/W SPR 304 Figure 6-51. Debug Status Register (DBSR) Table 6-35. DBSR Field Descriptions Bits 32 33 Name IDE UDE Description Imprecise debug event. Set if MSR[DE] = 0 and a debug event causes its respective DBSR bit to be set. Functions as write-one-to-clear. Unconditional debug event. Set if an unconditional debug event occurred. Functions as write-one-to-clear. If UDE (level sensitive, active low) is asserted, DBSR[UDE] is affected as follows: MSR[DE]DBCR0[IDM]Action X 0 No action. 0 1 UDE is set. 1 1 UDE is set and a debug interrupt is taken. Most recent reset. Functions as write-one-to-clear. Undefined at power-up. The e500 implements HRESET as follows: 0x No hard reset occurred since this bit was last cleared by software. 1x The previous reset was a hard reset. Instruction complete debug event. Set if an instruction completion debug event occurred and DBCR0[ICMP] = 1. Functions as write-one-to-clear. 34–35 MRR 36 ICMP MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 6-46 Freescale Semiconductor Debug Registers Table 6-35. DBSR Field Descriptions (continued) Bits 37 38 39 40 41 42–43 44 45 46 47 48 49–63 Name BRT IRPT TRAP IAC1 IAC2 — Description Branch taken debug event. Set if a branch taken debug event occurred (DBCR0[BRT] = 1). Functions as write-one-to-clear. Interrupt taken debug event. Set if an interrupt taken debug event occurred (DBCR0[IRPT] = 1). Functions as write-one-to-clear. Trap instruction debug event. Set if a trap Instruction debug event occurred (DBCR0[TRAP] = 1). Functions as write-one-to-clear. Instruction address compare 1 debug event. Set if an IAC1 debug event occurred (DBCR0[IAC1] = 1). Functions as write-one-to-clear. Instruction address compare 2 debug event. Set if an IAC2 debug event occurred (DBCR0[IAC2] = 1). Functions as write-one-to-clear. Reserved, should be cleared DAC1R Data address compare 1 read debug event. Set if a read-type DAC1 debug event occurred (DBCR0[DAC1] = 10 or 11). Functions as write-one-to-clear. DAC1 W Data address compare 1 write debug event. Set if a write-type DAC1 debug event occurred (DBCR0[DAC1] = 01 or 11). Functions as write-one-to-clear. DAC2R Data address compare 2 read debug event.Set if a read-type DAC2 debug event occurred (DBCR0[DAC2] = 10 or 11). Functions as write-one-to-clear. DAC2 W RET — Data address compare 2 write debug event. Set if a write-type DAC2 debug event occurred (DBCR0[DAC2] = 01 or 11). Functions as write-one-to-clear. Return debug event. Set if a return debug event occurred (DBCR0[RET] = 1). Functions as write-one-to-clear. Reserved, should be cleared. 6.13.3 Instruction Address Compare Registers (IAC1–IAC2) 32 61 62 63 Field Reset R/W SPR Instruction address All zeros R/W SPR 312 (IAC1); SPR 313 (IAC2) — Figure 6-52. Instruction Address Compare Registers (IAC1–IAC2) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 6-47 Core Register Summary 6.13.4 Data Address Compare Registers (DAC1–DAC2) 32 63 Field Reset R/W SPR Data address All zeros R/W SPR 316 (DAC1); SPR 317 (DAC2) Figure 6-53. Data Address Compare Registers (DAC1–DAC2) 6.14 Signal Processing and Embedded Floating-Point Status and Control Register (SPEFSCR) H igh-Word Error Bits 32 33 34 35 36 37 38 39 40 41 42 43 Status Bits 44 45 46 47 Field SOVH OVH FGH FXH FINVH FDBZH FUNFH FOVFH Reset R/W — FINXS FINVS FDBZS FUNFS FOVFS MODE 0000_0000_0000_0000 R/W Enable Bits n 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 Field SOV Reset R/W SPR OV FG FX FINV FDBZ FUNF FOVF — FINXE FINVE FDBZE FUNFE FOVFE FRMC 0000_0000_0000_0000 R/W SPR 512 Figure 6-54. Signal Processing and Embedded Floating-Point Status and Control Register (SPEFSCR) Table 6-36. SPEFSCR Field Descriptions Bits 32 33 34 Name SOVH OVH FGH Function Summary integer overflow high. Set whenever an instruction (except mtspr) sets OVH. SOVH remains set until it is cleared by an mtspr[SPEFSCR]. Integer overflow high. An overflow occurred in the upper half of the register while executing a SPE integer instruction Embedded floating-point guard bit high. Floating-point guard bit from the upper half. The value is undefined if the processor takes a floating-point exception due to input error, floating-point overflow, or floating-point underflow. Embedded floating-point sticky bit high. Floating bit from the upper half. The value is undefined if the processor takes a floating-point exception due to input error, floating-point overflow, or floating-point underflow. 35 FXH MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 6-48 Freescale Semiconductor Signal Processing and Embedded Floating-Point Status and Control Register (SPEFSCR) Table 6-36. SPEFSCR Field Descriptions (continued) Bits 36 37 38 39 40–41 42 43 44 45 46 47 48 49 50 Name FINVH Function Embedded floating-point invalid operation error high. Set when an input value on the high side is a NaN, Inf, or Denorm. Also set on a divide if both the dividend and divisor are zero. FDBZH Embedded floating-point divide by zero error high. Set if the dividend is non-zero and the divisor is zero. FUNFH Embedded floating-point underflow error high FOVFH Embedded floating-point overflow error high — FINXS FINVS Reserved, should be cleared. Embedded floating-point inexact sticky. FINXS = FINXS | FGH | FXH | FG | FX Embedded floating-point invalid operation sticky. Location for software to use when implementing true IEEE floating point. FDBZS Embedded floating-point divide by zero sticky. FDBZS = FDBZS | FDBZH | FDBZ FUNFS Embedded floating-point underflow sticky. Storage location for software to use when implementing true IEEE floating point. FOVFS Embedded floating-point overflow sticky. Storage location for software to use when implementing true IEEE floating point. MODE SOV OV FG Embedded floating-point mode (read only on e500) Integer summary overflow. Set whenever an SPE instruction (except mtspr) sets OV. SOV remains set until it is cleared by mtspr[SPEFSCR]. Integer overflow. An overflow occurred in the lower half of the register while a SPE integer instruction is being executed. Embedded floating-point guard bit. Floating-point guard bit from the lower half. The value is undefined if the processor takes a floating-point exception due to input error, floating-point overflow, or floating-point underflow. Embedded floating-point sticky bit. Floating bit from the lower half. The value is undefined if the processor takes a floating-point exception due to input error, floating-point overflow, or floating-point underflow. Embedded floating-point invalid operation error. Set when an input value on the high side is a NaN, Inf, or Denorm. Also set on a divide if both the dividend and divisor are zero. Embedded floating-point divide by zero error. Set of the dividend is non-zero and the divisor is zero. Embedded floating-point underflow error Embedded floating-point overflow error Reserved, should be cleared. Embedded floating-point inexact enable Embedded floating-point invalid operation/input error exception enable 0 Exception disabled 1 Exception enabled If the exception is enabled, a floating-point data exception is taken if FINV or FINVH is set by a floating-point instruction. 51 FX 52 53 54 55 56 57 58 FINV FDBZ FUNF FOVF — FINXE FINVE MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 6-49 Core Register Summary Table 6-36. SPEFSCR Field Descriptions (continued) Bits 59 Name Function FDBZE Embedded floating-point divide-by-zero exception enable 0 Exception disabled 1 Exception enabled If the exception is enabled, a floating-point data exception is taken if FDBZ or FDBZH is set by a floating-point instruction. FUNFE Embedded floating-point underflow exception enable 0 Exception disabled 1 Exception enabled If the exception is enabled, a floating-point data exception is taken if FUNF or FUNFH is set by a floating-point instruction. FOVFE Embedded floating-point overflow exception enable 0 Exception disabled 1 Exception enabled If the exception is enabled, a floating-point data exception is taken if FOVF or FOVFH is set by a floating-point instruction. FRMC Embedded floating-point rounding mode control 00 Round to nearest 01 Round toward zero 10 Round toward +infinity 11 Round toward –infinity 60 61 62–63 6.14.1 Accumulator (ACC) 0 31 32 63 Field Reset R/W Upper word All zeros R/W Lower word Figure 6-55. Accumulator (ACC) Table 6-37. ACC FIeld Descriptions Bits 0–31 32–63 Name Upper word Lower word Function Holds the upper-word accumulate value for SPE multiply with accumulate instructions Holds the lower-word accumulate value for SPE multiply with accumulate instructions MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 6-50 Freescale Semiconductor Performance Monitor Registers (PMRs) 6.15 Performance Monitor Registers (PMRs) Table 6-38. Supervisor-Level PMRs (PMR[5] = 1) Abbreviation PMC0 PMC1 PMC2 PMC3 PMGC0 PMLCa0 PMLCa1 PMLCa2 PMLCa3 PMLCb0 PMLCb1 PMLCb2 PMLCb3 Register Name Performance monitor counter 0 Performance monitor counter 1 Performance monitor counter 2 Performance monitor counter 3 Performance monitor global control register 0 Performance monitor local control a0 Performance monitor local control a1 Performance monitor local control a2 Performance monitor local control a3 Performance monitor local control b0 Performance monitor local control b1 Performance monitor local control b2 Performance monitor local control b3 PMR Number 16 17 18 19 400 144 145 146 147 272 273 274 275 pmr[0–4] pmr[5–9] 00000 00000 00000 00000 01100 00100 00100 00100 00100 01000 01000 01000 01000 10000 10001 10010 10011 10000 10000 10001 10010 10011 10000 10001 10010 10011 6.15.3/6-53 6.15.1/6-52 6.15.2/6-52 Section/ Page 6.15.4/6-54 Table 6-39. User-Level PMRs (PMR[5] = 0) (Read Only) Abbreviation UPMC0 UPMC1 UPMC2 UPMC3 UPMLCa0 UPMLCa1 UPMLCa2 UPMLCa3 UPMLCb0 UPMLCb1 UPMLCb2 UPMLCb3 UPMGC0 Register Name User performance monitor counter 0 User performance monitor counter 1 User performance monitor counter 2 User performance monitor counter 3 User performance monitor local control a0 User performance monitor local control a1 User performance monitor local control a2 User performance monitor local control a3 User performance monitor local control b0 User performance monitor local control b1 User performance monitor local control b2 User performance monitor local control b3 User performance monitor global control register 0 PMR Number 0 1 2 3 128 129 130 131 256 257 258 259 384 pmr[0–4] pmr[5–9] 00000 00000 00000 00000 00100 00100 00100 00100 01000 01000 01000 01000 01100 00000 00001 00010 00011 00000 00001 00010 00011 00000 00001 00010 00011 00000 6.15.2/6-52 6.15.3/6-53 6.15.3/6-53 Section/ Page 6.15.4/6-54 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 6-51 Core Register Summary 6.15.1 Global Control Register 0 (PMGC0, UPMGC0) 32 33 34 35 63 Field FAC PMIE FCECE Reset R/W PMR All zeros — PMGC0: Supervisor R/W; UPMGC0: User read-only PMGC0: PMR400; UPMGC0: PMR384 Figure 6-56. Performance Monitor Global Control Register 0 (PMGC0), User Performance Monitor Global Control Register 0 (UPMGC0) Table 6-40. PMGC0 Field Descriptions Bits 32 Name FAC Description Freeze all counters. When FAC is set by hardware or software, PMLCx[FC] maintains its current value until it is changed by software. 0 The PMCs are incremented (if permitted by other PM control bits). 1 The PMCs are not incremented. Performance monitor interrupt enable 0 Performance monitor interrupts are disabled. 1 Performance monitor interrupts are enabled and occur when an enabled condition or event occurs. 33 PMIE 34 FCECE Freeze counters on enabled condition or event 0 The PMCs can be incremented (if permitted by other PM control bits). 1 The PMCs can be incremented (if permitted by other PM control bits) only until an enabled condition or event occurs. When an enabled condition or event occurs, PMGC0[FAC] is set. It is up to software to clear FAC. — Reserved, should be cleared. 35–63 6.15.2 Local Control A Registers (PMLCa0–PMLCa3, UPMLCa0–UPMLCa3) 32 33 34 35 36 37 38 40 41 47 48 63 Field FC FCS FCU FCM1 FCM0 CE Reset R/W PMR — EVENT All zeros PMLCa0–PMLCa3: Supervisor R/W UPMLCa0–UPMLCa3: User read-only PMLCa0: PMR144, PMLCa1: PMR145, PMLCa2: PMR146, PMLCa3: PMR147 UPMLCa0: PMR128, UPMLCa1: PMR129, UPMLCa2: PMR130, UPMLCa3: PMR131 Figure 6-57. Local Control A Registers (PMLCa0–PMLCa3), User Local Control A Registers (UPMLCa0–UPMLCa3) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 6-52 Freescale Semiconductor Performance Monitor Registers (PMRs) Table 6-41. PMLCa0–PMLCa3 Field Descriptions Bits 32 Name FC Description Freeze counter 0 The PMC is incremented (if permitted by other PM control bits). 1 The PMC is not incremented. Freeze counter in supervisor state 0 The PMC is incremented (if permitted by other PM control bits). 1 The PMC is not incremented if MSR[PR] = 0. Freeze counter in user state 0 The PMC is incremented (if permitted by other PM control bits). 1 The PMC is not incremented if MSR[PR] = 1. 33 FCS 34 FCU 35 FCM1 Freeze counter while mark = 1 0 The PMC is incremented (if permitted by other PM control bits). 1 The PMC is not incremented if MSR[PMM] = 1. FCM0 Freeze counter while mark = 0 0 The PMC is incremented (if permitted by other PM control bits). 1 The PMC is not incremented if MSR[PMM] = 0. CE Condition enable 0 PMC x overflow conditions cannot occur (PMC x cannot cause interrupts, cannot freeze counters) 1 Overflow conditions occur when the most-significant bit of PMCx is equal to 1. It is recommended that CE be cleared when counter PMCx is selected for chaining. Reserved, should be cleared. 36 37 38–40 — 41–47 EVENT Event selector. Up to 128 events selectable. These events are described in the PowerPC e500 Core Reference Manual. 48–63 — Reserved, should be cleared. 6.15.3 Local Control B Registers (PMLCb0–PMLCb3, UPMLCb0–UPMLCb3) 32 52 53 55 56 57 58 63 Field Reset R/W PMR — All zeros THRESHMUL — THRESHOLD PMLCb0–PMLCb3: Supervisor R/W UPMLCb0–UPMLCb3: User read-only PMLCb0: PMR272, PMLCb1: PMR273, PMLCb2: PMR274, PMLCb3: PMR275 UPMLCb0: PMR256, UPMLCb1: PMR257, UPMLCb2: PMR258, UPMLCb3: PMR259 Figure 6-58. Local Control B Registers (PMLCb0–PMLCb3), User Local Control B Registers (UPMLCb0–UPMLCb3) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 6-53 Core Register Summary Table 6-42. PMLCb0–PMLCb3 Field Descriptions Bits 32–52 53–55 Name — THRESHMUL Reserved, should be cleared. Threshold multiple 000 Threshold field is multiplied by 1 (PMLCbn[THRESHOLD] × 1) 001 Threshold field is multiplied by 2 (PMLCbn[THRESHOLD] × 2) 010 Threshold field is multiplied by 4 (PMLCbn[THRESHOLD] × 4) 011 Threshold field is multiplied by 8 (PMLCbn[THRESHOLD] × 8) 100 Threshold field is multiplied by 16 (PMLCbn[THRESHOLD] × 16) 101 Threshold field is multiplied by 32 (PMLCbn[THRESHOLD] × 32) 110 Threshold field is multiplied by 64 (PMLCbn[THRESHOLD] × 64) 111 Threshold field is multiplied by 128 (PMLCbn[THRESHOLD] × 128) Reserved, should be cleared. Threshold. Only events that exceed the threshold value are counted. Such events are implementation-dependent as are the dimension (for example duration in cycles) and granularity with which the value is interpreted. By varying the value, software can obtain a profile of the event characteristics subject to thresholding. For example, if PMC1 is configured to count cache misses that last longer than the threshold value, software can obtain the distribution of cache miss durations for a given program by monitoring the program repeatedly using a different threshold each time. Description 56–57 58–63 — THRESHOLD 6.15.4 Performance Monitor Counter Registers (PMC0–PMC3, UPMC0–UPMC3) 32 33 63 Field OV Reset R/W PMR Counter value All zeros PMC0–PMC3: Supervisor R/W UPMC0–UPMC3: User read only PMC0: PMR16, PMC1: PMR17, PMC2,: PMR18, PMC3: PMR19 UPMC0: PMR0, UPMC1: PMR1, UPMC2,: PMR2, UPMC3: PMR3 Figure 6-59. Performance Monitor Counter Registers (PMC0–PMC3), User Performance Monitor Counter Registers (UPMC0–UPMC3) Table 6-43. PMC0–PMC3 Field Descriptions Bits 32 33–63 Name OV Counter value Description Overflow. When this bit is set, it indicates this counter reaches its maximum value. Indicates the number of occurrences of the specified event MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 6-54 Freescale Semiconductor Chapter 7 L2 Look-Aside Cache/SRAM This chapter describes the organization of the on-chip L2/SRAM, cache coherency rules, cache line replacement algorithm, cache control instructions, and various cache operations. It also describes the interaction between the L2/SRAM and the e500 core complex. 7.1 L2 Cache Overview The integrated 256-Kbyte L2 cache is organized as 1024 eight-way sets of 32-byte cache lines based on 32-bit physical addresses, as shown in Figure 7-1. 256-Kbyte L2 Cache/SRAM e500 Core Complex e500 Core Two 128-Kbyte banks (8-way) Independently programmable as L2 cache or SRAM WR IN DOUT RD IN 64 128 128 32-Kbyte L1 Data Cache 32-Kbyte L1 Instruction Cache Core Complex Bus RD1 RD2 WR Coherency Module Figure 7-1. L2 Cache/SRAM Configuration The SRAM can be configured with memory-mapped registers as externally accessible memory-mapped SRAM in addition to or instead of cache. The L2 cache can operate in the following modes, described in Section 7.2, “Cache Organization”: • • • Full cache mode (256-Kbyte cache) Full memory-mapped SRAM mode (256-Kbyte SRAM mapped as a single 256-Kbyte block or two 128-Kbyte blocks) Half SRAM and half cache mode (128-Kbyte cache and 128-Kbyte memory-mapped SRAM) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 7-1 L2 Look-Aside Cache/SRAM 7.1.1 L2 Cache and SRAM Features Two 128-Kbyte arrays can be designated independently as externally-accessible memory-mapped SRAM or cache. The L2 cache has the following characteristics: • Write-through, front-side cache — Front-side design provides easier cache access for I/O masters such as Ethernet and RapidIO — Write-through design is more efficient on the processor bus for front-side caches Valid, locked, and stale states (no modified state) Two input data buses (64 and 128 bits) and one output data bus (128 bits wide) All accesses are fully pipelined and non-blocking (allows hits under misses) 256-Kbyte array organized as 1024 eight-way sets of 32-byte cache lines Eight-way set associativity (high level of associativity yields good performance even with many locked lines) Tag arrays contain 17 tag bits and 1 tag parity bit per line to support 256-Kbyte cache (1024 sets), or 18 tag bits to support 128-Kbyte cache (512 sets). Configurable to allocate processor instructions, data, or both — Allows external writes (stashing) to allocate and optionally lock a line using one of the two following methods: – Attributes attached to the transactions by initiator or ATMU – I/O devices can force memory writes to be allocated using programmed memory ranges Pseudo-LRU (7-bit replacement algorithm) Data ECC on 64-bit boundaries (single-error correction, double-error detection) Tag parity for 256-Kbyte mode (1 tag bit per line covering cache tags) Cache locking methods — Individual line locks are set and cleared by using e500 cache locking APU instructions— Data Cache Block Touch and Lock Set (dcbtls), Data Cache Block Touch for Store and Lock Set (dcbtstls), and Instruction Cache Block Touch and Lock Set (icbtls). — A lock attribute can be attached to write operations. — Individual line locks are set and cleared through core-initiated instructions, by external reads or writes, or by accesses to programmed memory ranges defined in L2 cache external write address registers (L2CEWARn). — The entire cache can be locked by setting a configuration register appropriately. Lock clearing methods — Individual locks cleared by cache locking APU instructions (Instruction Cache Block Lock Clear (icblc) and Data Cache Block Lock Clear (dcblc)) or by snooped flush unless entire cache is locked MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 7-2 Freescale Semiconductor • • • • • • • • • • • • Cache Organization • • • • — Flash clearing of all instruction and/or data locks is done by writes to configuration registers — An unlock attribute attached to a read instruction. Error injection modes for testing I/O devices access SRAM regions by marking transactions as snoopable (global) Regions can reside at any aligned location in the memory map For accesses of less than a cache-line, byte-accessible ECC is protected using read-modify-write transactions. SRAM features include the following: 7.2 Cache Organization When the entire 256-Kbyte array is used as a cache, it has two banks each containing 512 sets of eight cache blocks (eight ways), as shown in Figure 7-2. Each block consists of 32 bytes of data and an address tag. 512 Sets Bank 0 Way 0 Address Tag 0 Way 1 Address Tag 1 Way 2 Address Tag 2 Way 3 Address Tag 3 Way 4 Address Tag 4 Bank 1 512 Sets Block 5 Address Tag 5 Address Tag 6 Block 6 Block 7Address Tag 7 Way 0 Address Tag 0 Way 1 Address Tag 1 Way 2 Address Tag 2 Way 3 Address Tag 3 Way 4 Address Tag 4 Way 5 Address Tag 5 Way 6 Address Tag 6 Way 7 Address Tag 7 Words [0–7] Words [0–7] Words [0–7] Words [0–7] Words [0–7] Words [0–7] Words [0–7] Words [0–7] 8 Words/Block Words [0–7] Words [0–7] Words [0–7] Words [0–7] Words [0–7] Words [0–7] Words [0–7] Words [0–7] 8 Words/Block Figure 7-2. Cache Organization MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 7-3 L2 Look-Aside Cache/SRAM The tag size depends on whether both 128-Kbyte arrays are configured as cache. Figure 7-3 shows how physical address bits are used to access the L2 in full cache mode (256-Kbyte cache). Tag (17 bits) 0 16 17 18 26 Tag Lookup Bank select Set index (9 bits) 17 18 Data Read/Write 26 27 31 Bank select Set index (9 bits) Byte select (within a cache line) (5 bits) Figure 7-3. 256-Kbyte L2 Cache Address Configuration—Full Cache Mode Physical address bits 17–26 identify the bank and set of the tag and data. Physical address bits 0–16 are compared against the tags of all 8 ways. A match of a valid tag selects a 32-byte block of data within the set. Physical address bits 27–31 identify the byte or bytes of data within the block. If the L2 is programmed as one block of 256-Kbyte SRAM, physical address bits 14–16 are used as the way select, without a tag lookup. In half SRAM, half cache mode (128-Kbyte cache), there is no bank select, as shown in Figure 7-4, so the tag is 18 bits (with no parity bit) for cache accesses. The cache resides in bank 1, and the SRAM in bank 0. For SRAM accesses, physical address bits 15–17 are used as the way select. If the L2 is programmed as two blocks of 128-Kbyte SRAM, bank 0 is block 0 (defined by L2SRBAR0, see Table 7-5) and bank 1 is block 1. As shown in Figure 7-4, the tag is 18 bits (with no parity bit) for cache accesses. For SRAM accesses, physical address bits 15–17 are used as the way select. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 7-4 Freescale Semiconductor Cache Organization Tag (18 bits) 0 17 18 26 Tag Lookup Set Index (9 bits) 15 Data Read/Write 17 18 26 27 31 Way Select (SRAM access only) (3 bits) Set Index (9 bits) Byte Select (within a cache line) (5 bits) Figure 7-4. 128-Kbyte L2 Cache Address Configuration—Half SRAM, Half Cache Mode The e500 core connects to the L2 cache and the system interface through the high-speed core complex bus (CCB). The e500 core and the L2 cache connect to the rest of the integrated device through the e500 coherency module (ECM). Figure 7-5 shows the data connections of the e500 core and L2/SRAM. The e500 core can simultaneously read 128 bits of data from the L2/SRAM, read 64 bits of data from the system interface, and write 128 bits of data to the L2/SRAM and/or system interface. The L2/SRAM can be accessed by the e500 core or the system interface through the ECM. The L2 cache does not initiate transactions. Figure 7-5 shows the data bus connections of the e500 core and L2/SRAM. . e500 Core RD1 RD2 WR L2/SRAM RD IN DOUT WR IN 64 128 128 e500 Coherency Module (ECM) Figure 7-5. Data Bus Connection of CCB MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 7-5 L2 Look-Aside Cache/SRAM Figure 7-6 shows address connections of the e500 core and L2/SRAM. e500 Core SNOOP MSTR L2/SRAM ADDR_IN e500 Coherency Module (ECM) Figure 7-6. Address Bus Connection of CCB In SRAM mode, if a non–cache-line read or write transaction is not preceded by a cache-line write, an ECC error occurs; such a non–cache-line write transaction cannot be allocated in the L2. 7.3 Memory Map/Register Definition Table 7-1. L2/SRAM Memory-Mapped Registers Offset Register L2CTL—L2 control register L2CEWAR0—L2 cache external write address register 0 L2CEWCR0—L2 cache external write control register 0 L2CEWAR1—L2 cache external write address register 1 L2CEWCR1—L2 cache external write control register 1 L2CEWAR2—L2 cache external write address register 2 L2CEWCR2—L2 cache external write control register 2 L2CEWAR3—L2 cache external write address register 3 L2CEWCR3—L2 cache external write control register 3 L2SRBAR0—L2 memory-mapped SRAM base address register 0 L2SRBAR1—L2 memory-mapped SRAM base address register 1 L2ERRINJHI—L2 error injection mask high register L2ERRINJLO—L2 error injection mask low register L2ERRINJCTL—L2 error injection tag/ECC control register L2CAPTDATAHI—L2 error data high capture register L2CAPTDATALO—L2 error data low capture register Access R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R R Reset 0x2000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 Section/Page 7.3.1.1/7-7 7.3.1.2/7-10 7.3.1.3/7-10 7.3.1.2/7-10 7.3.1.3/7-10 7.3.1.2/7-10 7.3.1.3/7-10 7.3.1.2/7-10 7.3.1.3/7-10 7.3.1.4/7-11 7.3.1.4/7-11 7.3.1.5.1/7-13 7.3.1.5.1/7-13 7.3.1.5.1/7-13 7.3.1.5.2/7-15 7.3.1.5.2/7-15 Table 7-1 shows the memory map for the L2/SRAM registers. Undefined 4-byte address spaces within offset 0x000–0xFFF are reserved. 0x2_0000 0x2_0010 0x2_0018 0x2_0020 0x2_0028 0x2_0030 0x2_0038 0x2_0040 0x2_0048 0x2_0100 0x2_0108 0x2_0E00 0x2_0E04 0x2_0E08 0x2_0E20 0x2_0E24 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 7-6 Freescale Semiconductor Memory Map/Register Definition Table 7-1. L2/SRAM Memory-Mapped Registers (continued) Offset 0x2_0E28 0x2_0E40 0x2_0E44 0x2_0E48 Register L2CAPTECC—L2 error syndrome register L2ERRDET—L2 error detect register L2ERRDIS—L2 error disable register L2ERRINTEN—L2 error interrupt enable register Access R Special R/W R/W R/W R R/W Reset 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 Section/Page 7.3.1.5.2/7-15 7.3.1.5.2/7-15 7.3.1.5.2/7-15 7.3.1.5.2/7-15 7.3.1.5.2/7-15 7.3.1.5.2/7-15 7.3.1.5.2/7-15 0x2_0E4C L2ERRATTR—L2 error attributes capture register 0x2_0E50 0x2_0E58 L2ERRADDR—L2 error address capture register L2ERRCTL—L2 error control register 7.3.1 L2/SRAM Register Descriptions The following sections describe registers that control and configure the L2/SRAM array. 7.3.1.1 L2 Control Register (L2CTL) The L2 control register (L2CTL), shown in Figure 7-7, controls configuration and operation of the L2/SRAM array. The sequence for modifying L2CTL is as follows: 1. 2. 3. 4. 5. mbar isync stw (WIMG = 01xx) CCSRBAR+0x2_0000 lwz (WIMG = 01xx) CCSRBAR+0x2_0000 mbar 0 1 2 3 4 5 6 8 9 10 11 12 13 15 R L2E W Reset 16 17 18 L2SIZ L2I L2BLKSZ 0 0 0 L2DO L2IO 0 L2INTDIS L2SRAM 0010_0000_0000_0000 19 20 21 22 23 24 31 R W Reset Offset 0 0 L2LO L2SLC 0 L2LFR L2LFRID 0 0 0 0 0 0 0 0 0000_0000_0000_0000 0x2_0000 Figure 7-7. L2 Control Register (L2CTL) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 7-7 L2 Look-Aside Cache/SRAM Table 7-2 describes L2CTL fields. Table 7-2. L2CTL Field Descriptions Bits 0 Name L2E Description L2 enable. Used to enable the L2 array (cache or memory-mapped SRAM). 0 The L2 SRAM (cache and memory-mapped SRAM) is disabled and is not accessed for reads, snoops, or writes. Setting the L2 flash invalidate bit (L2I) is allowed. 1 The L2 SRAM (cache or memory-mapped SRAM) is enabled. Note that L2I can be set regardless of the value of L2E. L2 flash invalidate 0 The L2 status and LRU bits are not being cleared. 1 Setting L2I invalidates the L2 cache globally by clearing the all the L2 status bits, as well as the LRU algorithm. Memory-mapped SRAM is unaffected. Data in memory-mapped SRAM regions is unaffected by the flash invalidate. The hardware automatically clears L2I when the invalidate is complete. L2 SRAM size (read only). Indicates the total available size of L2 SRAM (to be configured as cache or memory-mapped SRAM). 00 Reserved 01 Reserved 10 256 Kbyte 11 Reserved L2 cache/memory-mapped SRAM block size. Determines the L2 cache/memory-mapped SRAM block size. To change these bits, the L2 must be disabled (L2CTL[L2E] = 0). L2BLKSZ must be ≤ L2SIZ when L2E = 1. See the description of L2CTL[L2SRAM] for information on configuring the total SRAM as cache or memory-mapped SRAM. 00 Reserved 01 128 Kbyte 10 256 Kbyte 11 Reserved Reserved L2 data-only. Reserved in full memory-mapped SRAM mode. L2DO may be changed while the L2 is enabled or disabled. 0 The L2 cache allocates entries for instruction fetches that miss in the L2. 1 The L2 cache allocates entries for processor data loads that miss in the L2 and for processor L1 castouts but does not allocate entries for instruction fetches that miss in the L2. Instruction accesses that hit in the L2, data accesses, and accesses from the system (including I/O stash writes) are unaffected. Note that if L2DO and L2IO are both set, no new lines are allocated into the L2 cache for any processor transactions, and processor writes and castouts that hit existing data in the cache invalidate those lines rather than updating them. L2 instruction-only. Reserved in full memory-mapped SRAM mode. Causes the L2 cache to allocate lines for instruction cache transactions only. L2IO may be changed while the L2 is enabled or disabled. 0 The L2 cache entries are allocated for data loads that miss in the L2 and for processor L1 castouts. 1 The L2 cache allocates entries for instruction fetch misses, but does not allocate entries for processor data transactions. Data accesses that hit in the L2, instruction accesses, and accesses from the system (including I/O stash writes) are unaffected. Note that if L2DO and L2IO are both set, no new lines are allocated into the L2 cache for any processor transactions, and processor writes and castouts that hit existing data in the cache invalidate those lines rather than updating them. Reserved MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 7-8 Freescale Semiconductor 1 L2I 2–3 L2SIZ 4–5 L2BLKSZ 6–8 9 — L2DO 10 L2IO 11 — Memory Map/Register Definition Table 7-2. L2CTL Field Descriptions (continued) Bits 12 Name L2INTDIS Description Cache read intervention disable. Reserved for full memory-mapped SRAM mode. Used to disable cache read intervention. May be changed while the L2 is enabled or disabled. 0 Cache intervention is enabled. The ECM ensures that if a data read from another device hits in the L2 cache, it is serviced from the L2 cache. 1 Cache intervention is disabled L2 block assignment. Determines the L2 cache/memory-mapped SRAM block assignment. L2SIZ = L2BLKSIZ (1 block): 000 Block 0 = Cache 001 Block 0 = SRAM0 010–111 Reserved L2SIZ = L2BLKSIZ × 2 (2 blocks): Block 0 Block 1 000 Unused Cache 001 SRAM0 Unused 010 SRAM0 Cache 011 SRAM0 SRAM1 1xx Reserved To change these bits, the L2 must be disabled (L2CTL[L2E] = 0). Reserved L2 cache lock overflow. Reserved in full memory-mapped SRAM mode. This sticky bit is set if an overlock condition is detected in the L2 cache. A lock overflow is triggered either by executing instruction or data cache block touch and lock set instructions or by performing L2 cache external writes with lock set. If all ways are locked and an attempt to stash is made, the stash is not allocated. 0 The L2 cache did not encounter a lock overflow. L2LO is cleared only by software. 1 The L2 cache encountered a lock overflow condition. L2 snoop lock clear. This sticky bit is set if a snoop invalidated a locked data cache line. Note that the lock bit for that line is cleared whenever the line is invalidated. L2SLC is reserved in full memory-mapped SRAM mode. 0 A snoop did not invalidate a locked L2 cache line. L2SLC is cleared only by software. 1 The L2 cache encountered a snoop that invalidated a locked line. Reserved L2 cache lock bits flash reset. The L2 cache must be enabled (L2CTL[L2E] = 1) for reset to occur. This field is reserved in full memory-mapped SRAM mode. 0 The L2 cache lock bits are not cleared or the clear operation completed. 1 A reset operation is issued that clears each L2 cache line’s lock bits. Depending on the L2LFRID value, data or instruction locks, or both can be reset. Cache access is blocked during this time. After L2LFR is set, the L2 cache unit automatically clears L2LFR when the reset operation is complete (if L2CTL[L2E] is set). L2 cache lock bits flash reset select instruction or data. Indicates whether data or instruction lock bits or both are reset. 00 Not used 01 Reset data locks if L2LFR = 1 10 Reset instruction locks if L2LFR = 1 11 Reset both data and instruction locks if L2LFR = 1 Reserved 13–15 L2SRAM 16–17 18 — L2LO 19 L2SLC 20 21 — L2LFR 22–23 L2LFRID 24–31 — MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 7-9 L2 Look-Aside Cache/SRAM 7.3.1.2 L2 Cache External Write Address Registers 0–3 (L2CEWARn) The MPC8540 supports allocating and locking of L2 cache lines from external agents, such as PCI. This functionality is called stashing. The L2 cache external write address registers 0–3 (L2CEWARn) are paired with the L2 cache external write control registers 0–3 (L2CEWCRn) to control the cache external write functionality. Each register pair (for example, L2CEWAR0 and L2CEWCR0) specifies a programmed memory range that can be locked with a snoop write transaction. The address register must be naturally aligned to the window size in the corresponding control register. L2CEWARn registers contain identical fields, as shown in Figure 7-8. 0 23 24 31 R ADDR W Reset Offset All zeros 0x2_0010, 0x2_0020, 0x2_0030, 0x2_0040 00000000 Figure 7-8. L2 Cache External Write Address Registers (L2CEWARn) Table 7-3 describes L2CEWARn fields. Table 7-3. L2CEWARn Field Descriptions Bits 0–23 24–3 1 Nam e Description ADD L2 cache external write base address R — Reserved 7.3.1.3 L2 Cache External Write Control Registers 0–3 (L2CEWCRn) The L2CEWARn registers are paired with the L2 cache external write control registers 0–3 (L2CEWCRn), shown in Figure 7-9, to control cache external write functionality. 0 1 2 7 8 31 R E LOCK W Reset Offset 000000 SIZMASK All zeros 0x2_0018, 0x2_0028, 0x2_0038, 0x2_0048 Figure 7-9. L2 Cache External Write Control Registers (L2CEWCR0–L2CEWCR3) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 7-10 Freescale Semiconductor Memory Map/Register Definition The L2CEWCRn registers contain identical fields, which are described in Table 7-4. Table 7-4. L2CEWCRn Field Descriptions Bits 0 Name E Description External write enable. An external write matching the address window defined by L2CEWAR n/L2CEWCRn is allocated or updated in the L2 cache. 0 External writes for the L2CEWARn/L2CEWCR n pair are disabled. 1 External writes are enabled for the L2CEWARn/L2CEWCRn pair. Lock lines in the targeted cache. An external write matching the address window defined by L2CEWAR n/L2CEWCRn is locked in the L2 cache when it is allocated or updated. 0 The locked bit is not set when a line is allocated unless explicitly specified by transaction attributes. 1 Cache lines are allocated as locked. A hit to a valid, unlocked lines sets the lock. Reserved 1 LOCK 2–7 — 8–31 SIZMASK Mask size. Defines the size of the naturally aligned address region for cache external writes. The address region must be aligned to a boundary that is a multiple of the mask size. Any value not listed below is illegal and produces boundedly undefined results. 1111 1111 1111 1111 1111 1111 1111 1111 1111 1111 1111 1111 1111 1111 1111 1111 1111 1111 1111 1111 1111 1111 1111 1111 1111 1111 1111 1111 1111 1111 256 bytes 1111 1111 1111 1110 512 bytes 1111 1111 1111 1100 1 Kbyte 1111 1111 1111 1000 2 Kbytes 1111 1111 1111 0000 4 Kbytes 1111 1111 1110 0000 8 Kbytes 1111 1111 1100 0000 16 Kbytes 1111 1111 1000 0000 32 Kbytes 1111 1111 0000 0000 64 Kbytes 1111 1110 0000 0000 128 Kbytes 1111 1100 0000 0000 256 Kbytes 1111 1000 0000 0000 512 Kbytes 1111 0000 0000 0000 1 Mbyte 1111 1111 1110 0000 0000 0000 2 Mbytes 1111 1111 1100 0000 0000 0000 4 Mbytes 1111 1111 1000 0000 0000 0000 8 Mbytes 1111 1111 0000 0000 0000 0000 16 Mbytes 1111 1110 0000 0000 0000 0000 32 Mbytes 1111 1100 0000 0000 0000 0000 64 Mbytes 1111 1000 0000 0000 0000 0000 128 Mbytes 1111 0000 0000 0000 0000 0000 256 Mbytes 1110 0000 0000 0000 0000 0000 512 Mbytes 1100 0000 0000 0000 0000 0000 1 Gbyte 1000 0000 0000 0000 0000 0000 2 Gbytes 0000 0000 0000 0000 0000 0000 4 Gbytes 7.3.1.4 L2 Memory-Mapped SRAM Base Address Registers 0–1 (L2SRBARn) The L2 memory-mapped SRAM base address registers (L2SRBARn), shown in Figure 7-10, control the memory-mapped SRAM mode functionality. Specified addresses must be aligned to the value in L2CTL[L2BLKSZ]. If L2CTL[L2SRAM] specifies one memory-mapped SRAM block, its base address must be written to L2SRBAR0; if it specifies two memory-mapped SRAM blocks, L2SRBAR0 and L2SRBAR1 are used. 0 17 18 31 R ADDR W Reset Offset All zeros 0x2_0100, 0x2_0108 0000000000000 Figure 7-10. L2 Memory-Mapped SRAM Base Address Registers (L2SRBARn) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 7-11 L2 Look-Aside Cache/SRAM L2SRBAR bits are described in Table 7-5. Table 7-5. L2SRBARn Field Descriptions Bits 0–17 18–31 Name Description ADDR L2 memory-mapped SRAM base address — Reserved When enabled, the windows defined in L2SRBARn supersede all other mappings of these addresses for processor and global (snoopable) I/O transactions. Therefore, SRAM windows must never overlap configuration space as defined by CCSRBAR (see Section 4.3.1.1.2, “Configuration, Control, and Status Base Address Register (CCSRBAR)”). Overlapping SRAM and local access windows is discouraged because processor and snoopable I/O transactions would map to the SRAM while non-snooped I/O transactions would be mapped by the local access windows. Only if all accesses to the SRAM address range are snoopable can results be consistent if SRAM and local access windows overlap. Furthermore, L2SRBARn should not overlap a DDR DRAM space for which a valid chip select is defined. This could result in spurious ECC errors if ECC and speculative reads are enabled. See Section 2.2.3.7, “Illegal Interaction Between Local Access Windows and DDR SDRAM Chip Selects.” 7.3.1.5 L2 Error Registers L2 error detection, reporting, and injection allow flexible handling of ECC and parity errors in the L2 data and tag arrays. When the MPC8540 detects an L2 error, the appropriate bit in the error detect register (L2ERRDET) is set. Error detection is disabled by setting the corresponding bit in the error disable register (L2ERRDIS). The address and attributes of the first detected error are also saved in the error capture registers (L2ERRADDR, L2ERRATTR, L2CAPTDATAHI, L2CAPTDATALO, and L2CAPTACC). Subsequent errors set error bits in the error detection registers, but information is saved only for the first one. Error reporting (by generating an interrupt) is enabled by setting the corresponding bit in the error interrupt enable register (L2ERRINTEN). Note that the error detect bit is set regardless of the state of the interrupt enable bit. When an error is detected, if error detection is enabled the L2 cache/SRAM always asserts an internal error signal with read data to prevent the L1 caches and architectural registers from being loaded with corrupt data. If error detection is disabled, the detected error bit is not set and no internal signal is asserted. The L2 error detect register (L2ERRDET) is implemented as a bit-reset type register. Reading from this register occurs normally; however, write operations can clear but not set bits. A bit is cleared whenever the register is written, and the data in the corresponding bit location is a 1. For example, to clear bit 6 and not affect any other bits in the register, the value 0x0200_0000 is written to the register. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 7-12 Freescale Semiconductor Memory Map/Register Definition Note that in SRAM mode, if a non–cache-line read or write transaction is not preceded by a cache-line write, an ECC error occurs; such a non–cache-line write transaction cannot be allocated in the L2. 7.3.1.5.1 Error Injection Registers The L2 cache includes support for injecting errors into the L2 data, data ECC, or tag. This may be used to test error recovery software by deterministically creating error scenarios. The preferred method for error injection is to set all data pages to cache-inhibited (MMU TLB entry I = 1) except a scratch page, set L2CTL[L2DO] to prevent allocation of instruction accesses, and invalidate the L2 by setting L2CTL[L2I] = 1. The following code sequence triggers an error, then detects it (A is an address in the scratch page): dcbz A dcbtls_L2 A lwz A | allocates the line in the L1 in the modified state | forces the line from the L1 and allocates the line in the L2 Data or tag errors are injected into the line, according to the error injection settings in L2ERRINJHI, L2ERRINJLO, and L2ERRINJCTL, at allocation. The final load detects and reports the error (if enabled) and allows software to examine the offending data, address, and attributes. Note that error injection enable bits in L2ERRINJCTL must be cleared by software and the L2 must be invalidated (by setting L2CTL[L2I]) before resuming L2 normal operation. Figure 7-11 shows the L2 error injection mask high register (L2ERRINJHI). 0 31 R EIMASKHI W Reset Offset All zeros 0x2_0E00 Figure 7-11. L2 Error Injection Mask High Register (L2ERRINJHI) Table 7-6 describes L2ERRINJHI[EIMASKHI]. Table 7-6. L2ERRINJHI Field Description Bits 0–31 Name Description EIMASKH Error injection mask/high word. A set bit corresponding to a data path bit causes that bit on the data I path to be inverted on cache/SRAM writes if L2ERRINJCTL[DERRIEN] = 1. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 7-13 L2 Look-Aside Cache/SRAM Figure 7-12 shows the L2 error injection mask low register (L2ERRINJLO) fields. 0 31 R EIMASKLO W Reset Offset All zeros 0x2_0E04 Figure 7-12. L2 Error Injection Mask Low Register (L2ERRINJLO) Table 7-7 describes L2ERRINJLO[EIMASKLO]. Table 7-7. L2ERRINJLO Field Description Bits 0–3 1 Name EIMASKL O Description Error injection mask/low word. A set bit corresponding to a data path bit causes that bit on the data path to be inverted on SRAM writes if L2ERRINJCTL[DERRIEN] = 1. Figure 7-13 shows the L2 error injection mask control register (L2ERRINJCTL). 0 14 15 16 21 22 23 24 31 R00000000000000 0 TERRIEN W Reset Offset 000000 ECCMB DERRIEN All zeros 0x2_0E08 ECCERRIM Figure 7-13. L2 Error Injection Mask Control Register (L2ERRINJCTL) Table 7-8 describes L2ERRINJCTL fields. Table 7-8. L2ERRINJCTL Field Descriptions Bits 0–14 15 Name — TERRIEN Reserved L2 tag array error injection enable 0 No tag errors are injected. 1 All subsequent entries written to the L2 tag array have the parity bit inverted. Reserved ECC mirror byte enable 0 ECC byte mirroring is disabled 1 The most significant data path byte is mirrored onto the ECC byte if DERRIEN = 1. Description 16–21 22 — ECCMB MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 7-14 Freescale Semiconductor Memory Map/Register Definition Table 7-8. L2ERRINJCTL Field Descriptions (continued) Bits 23 Name DERRIEN Description L2 data array error injection enable 0 No data errors are injected. 1 Subsequent entries written to the L2 data array have data or ECC bits inverted as specified in the data and ECC error injection masks and/or data path byte mirrored onto ECC as specified by ECC mirror byte enable. Note: if both ECC mirror byte and data error injection are enabled, ECC mask error injection is performed on the mirrored ECC. 24–31 ECCERRI Error injection mask for the ECC bits. A set bit corresponding to an ECC bit causes that bit to be M inverted on SRAM writes if DERRIEN = 1. 7.3.1.5.2 Error Control and Capture Registers The error control and capture registers control detection and reporting of tag parity, ECC and L2 configuration errors. L2 configuration errors are illegal combinations of L2 size and block size and are detected when the L2 is enabled (L2CTL[L2E] = 1). Figure 7-14 shows the L2 error capture data high register (L2CAPTDATAHI). 0 31 R W Reset Offset L2DATA All zeros 0x2_0E20 Figure 7-14. L2 Error Capture Data High Register (L2CAPTDATAHI) Table 7-9 describes L2CAPTDATAHI[L2DATA] fields. Table 7-9. L2CAPTDATAHI Field Description Bits 0–31 Name L2DATA Description L2 data high word Figure 7-15 shows the L2 error capture data low register (L2CAPTDATALO). 0 31 R W Reset Offset L2DATA All zeros 0x2_0E24 Figure 7-15. L2 Error Capture Data Low Register (L2CAPTDATALO) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 7-15 L2 Look-Aside Cache/SRAM Table 7-10 describes L2CAPTDATALO[L2DATA] fields. Table 7-10. L2CAPTDATALO Field Description Bits 0–31 Name L2DATA Description L2 data low word Figure 7-16 shows the L2 error syndrome register (L2CAPTECC). 0 7 8 23 24 31 R W Reset Offset ECCSYND 0000000000000000 ECCCHKSUM All zeros 0x2_0E28 Figure 7-16. L2 Error Syndrome Register (L2CAPTECC) Table 7-11 describes L2CAPTECC fields. Table 7-11. L2CAPTECC Field Descriptions Bits 0–7 8–23 24–3 1 Name ECCSYND — Description The calculated ECC syndrome of the failing double word Reserved ECCCHKSU The data path ECC of the failing double word M Figure 7-17 shows the L2 error detect register (L2ERRDET). 0 1 26 27 28 29 30 31 R MULL2ERR 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 TPARERR MBECCERR SBECCERR 0 L2CFGDIS W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x2_0E40 Figure 7-17. L2 Error Detect Register (L2ERRDET) Table 7-12 describes L2ERRDET fields. Table 7-12. L2ERRDET Field Descriptions Bits 0 Name MULL2ERR Description Multiple L2 errors (bit reset, write 1 to clear) 0 Multiple L2 errors of the same type were not detected. 1 Multiple L2 errors of the same type were detected. Reserved 1–26 — MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 7-16 Freescale Semiconductor Memory Map/Register Definition Table 7-12. L2ERRDET Field Descriptions (continued) Bits 27 Name TPARERR Description Tag parity error (bit reset, write 1 to clear) 0 Tag parity error was not detected 1 Tag parity error was detected Note that if an L2 cache tag parity error occurs on an attempt to write a new line, the L2 cache must be flash invalidated. L2 functionality is not guaranteed if flash invalidation is not performed after a tag parity error. 28 MBECCERR Multiple-bit ECC error (bit reset, write 1 to clear) 0 Multiple-bit ECC errors were not detected. 1 Multiple-bit ECC errors were detected. SBECCERR Single-bit ECC error (bit reset, write 1 to clear) 0 Single-bit ECC error was not detected 1 Single-bit ECC error was detected — L2CFGERR Reserved L2 configuration error (bit reset, write 1 to clear). Reports inconsistencies between the L2SIZ, L2BLKSZ, and L2SRAM settings of the L2 control register (L2CTL). 0 L2 configuration errors were not detected 1 L2 illegal configuration error detected 29 30 31 Figure 7-18 shows the L2 error disable register (L2ERRDIS). 0 26 27 28 29 30 31 R00000000000000000000000000 0 TPARDIS MBECCDIS SBECCDIS W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x2_0E44 0 L2CFGDIS Figure 7-18. L2 Error Disable Register (L2ERRDIS) Table 7-13 describes L2ERRDIS fields. Table 7-13. L2ERRDIS Field Descriptions Bits 0–26 27 Name — TPARDIS Reserved Tag parity error disable 0 Tag parity error detection enabled 1 Tag parity error detection disabled Description 28 MBECCDIS Multiple-bit ECC error disable. Note that uncorrectable read errors may cause the assertion of core_fault_in, which causes the core to generate a machine check interrupt, unless it is disabled (by clearing HID1[RFXE]). If RFXE is zero and this error occurs, MBECCDIS must be cleared and L2ERRINTEN[MBECCINTEN] must be set to ensure that an interrupt is generated. For more information, see Section 6.10.2, “Hardware Implementation-Dependent Register 1 (HID1).” 0 Multiple-bit ECC error detection enabled 1 Multiple-bit ECC error detection disabled MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 7-17 L2 Look-Aside Cache/SRAM Table 7-13. L2ERRDIS Field Descriptions (continued) Bits 29 Name SBECCDIS Single-bit ECC error disable 0 Single-bit ECC error detection enabled 1 Single-bit ECC error detection disabled — Reserved Description 30 31 L2CFGDIS L2 configuration error disable 0 L2 configuration error detection enabled 1 L2 configuration error detection disabled Figure 7-19 shows the L2 error interrupt enable register (L2ERRINTEN). When an enabled error condition exists, the L2 signals an interrupt to the core through the internal int signal. 0 26 27 28 29 30 31 R00000000000000000000000000 0 TPARINTEN MBECCINTEN SBECCINTEN 0 L2CFGINTEN W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x2_0E48 Figure 7-19. L2 Error Interrupt Enable Register (L2ERRINTEN) Table 7-14 describes L2ERRINTEN fields. Table 7-14. L2ERRINTEN Field Descriptions Bits 0–26 27 Name — TPARINTEN Reserved Tag parity error reporting enable 0 Tag parity error reporting disabled 1 Tag parity error reporting enabled Multiple-bit ECC error reporting enable. Note that uncorrectable read errors may cause the assertion of core_fault_in, which causes the core to generate a machine check interrupt, unless it is disabled (by clearing HID1[RFXE]). If RFXE is zero and this error occurs, L2ERRDIS[MBECCDIS] must be cleared and MBECCINTEN must be set to ensure that an interrupt is generated. For more information, see Section 6.10.2, “Hardware Implementation-Dependent Register 1 (HID1).” 0 Multiple-bit ECC error reporting disabled 1 Multiple-bit ECC error reporting enabled Single-bit ECC error reporting enable 0 Single-bit ECC error reporting disabled 1 Single-bit ECC error reporting enabled Reserved L2 configuration error reporting enable 0 L2 configuration error reporting disabled 1 L2 configuration error reporting enabled Description 28 MBECCINTEN 29 SBECCINTEN 30 31 — L2CFGINTEN MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 7-18 Freescale Semiconductor Memory Map/Register Definition Figure 7-20 shows the L2 error attributes capture register (L2ERRATTR). 01 2 3 4 5 7 8 9 10 11 15 16 17 18 19 20 30 31 R00 DWNUM W Reset Offset 0 TRANSSIZ BURST 00 TRANSSRC 00 TRANSTYPE 0 0000000000 VALINFO 0000_0000_0000_0000_0000_0000_0000_0000 0x2_0E4C Figure 7-20. L2 Error Attributes Capture Register (L2ERRATTR) Table 7-15 describes L2ERRATTR fields. Table 7-15. L2ERRATTR Field Descriptions Bits 0–1 2–3 4 5–7 Name — DWNUM — TRANSSIZ Reserved Double-word number of the detected error (data ECC errors only) Reserved Transaction size for detected error Single-beat Burst 000 8 bytes Reserved 001 1 byte 16 bytes 010 2 bytes 32 bytes 011 3 bytes Reserved Burst transaction for detected error 0 Single-beat (≤ 64 bits) transaction 1 Burst transaction Reserved Single-beat 100 4 bytes 101 5 bytes 110 6 bytes 111 7 bytes Burst Reserved Reserved Reserved Reserved Description 8 BURST 9–10 — 11–15 TRANSSRC Transaction source for detected error 00000External (system logic) 10000Processor (instruction) 10001Processor (data) 16–17 18–19 — Reserved TRANSTYP Transaction type for detected error E 00 Snoop (tag/status read) 01 Write 10 Read 11 Read-modify-write — VALINFO Reserved L2 capture registers valid 0 L2 capture registers contain no valid information or no enabled errors were detected. 1 L2 capture registers contain information of the first detected error which has reporting enabled. Software must clear this bit to unfreeze error capture so error detection hardware can overwrite the capture address/data/attributes for a newly detected error. 20–30 31 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 7-19 L2 Look-Aside Cache/SRAM Figure 7-21 shows the L2 error address capture register (L2ERRADDR). 0 31 R W Reset Offset L2ADDR All zeros 0x2_0E50 Figure 7-21. L2 Error Address Capture Register (L2ERRADDR) Table 7-16 describes L2ERRADDR[L2ADDR] fields. Table 7-16. L2ERRADDR Field Description Bits 0–31 Name L2ADDR Description L2 address corresponding to detected error Figure 7-22 shows the L2 error control register (L2ERRCTL). 0 7 8 15 16 23 24 31 R0 0 0 0 W Reset Offset 0000 L2CTHRESH 0000 0000 L2CCOUNT All zeros 0x2_0E58 Figure 7-22. L2 Error Control Register (L2ERRCTL) Table 7-17 describes L2ERRCTL fields. Table 7-17. L2ERRCTL Field Descriptions Bits 0–7 8–15 16–23 24–31 Name — Reserved Description L2CTHRESH L2 cache threshold. Threshold value for the number of ECC single-bit errors that are detected before reporting an error condition — L2CCOUNT Reserved L2 count. Counts ECC single-bit errors detected. If L2CCOUNT equals the ECC single-bit error trigger threshold, an error is reported if single-bit error reporting is enabled MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 7-20 Freescale Semiconductor External Writes to the L2 Cache (Cache Stashing) 7.4 External Writes to the L2 Cache (Cache Stashing) Data from an I/O master can be allocated into the L2 cache while simultaneously being written to memory. External writes can be performed from any I/O master: • • • • Ethernet RapidIO PCI DMA New cache lines are allocated for full cache line writes, unless the line is already valid or locked. Sub–cache-line write data is stashed only if the line is valid in the cache. A read-modify-write process is used to merge the stashed data with the valid line data. The L2 cache external write address registers 0–3 (L2CEWARn) are paired with the L2 cache external write control registers 0–3 (L2CEWCRn) to control the cache stashing functionality. Each register pair (for example, L2CEWAR0 and L2CEWCR0) specifies a programmed memory range that can be locked with a snoop write transaction. The address register must be naturally aligned to the window size in the corresponding control register. For more information, see Section 7.3.1.2, “L2 Cache External Write Address Registers 0–3 (L2CEWARn),” and Section 7.3.1.3, “L2 Cache External Write Control Registers 0–3 (L2CEWCRn).” Note that stashing can occur regardless of whether the L1 cache is enabled. Note that stashing can also occur even if the cache-inhibited bit in the MMU is set for the page. When the core looks for (and does not find) the stashed data in the L1 cache, a transaction is generated in which the L2 cache responds by updating the L1 cache with the requested data. For information on how to initiate cache stashing, see the respective chapters for the I/O masters, listed above, that support stashing. 7.5 L2 Cache Timing Table 7-18 shows the timing of back-to-back loads that miss in the L1 data cache and hit in the L2 cache, assuming the core is running at 2 1/2 times the L2 cache frequency. The L2 returns the 128 bits containing the requested data (critical quad word) first. This data is forwarded to the result register before the full cache line reloads the L1. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 7-21 L2 Look-Aside Cache/SRAM . Table 7-18. Fastest Read Timing—Hit in L2 Core Clocks e500 core load 1 D-cache miss To D-cache 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 LSU reads command LSU reads out data LSU DLFB LSU reads command LSU reads out data e500 core load 2 D-cache miss to D-cache LSU DLFB Result bus To CIU To CIU CIU Q CCB clocks CCB address bus load 1 CCB address bus load 2 CCB data bus load 1 CCB data bus load 2 To CIU to CIU CIU Q L2 Cache and SRAM Coherency • • • • • • • • • • • • The L2 allocates entries for data cast-out or pushed (non-global, non-write-through write with kill) from the L1 caches. Lines for e500 core-initiated burst read transactions are allocated as exclusive in the L2. The L2 supports I/O devices reading data from valid lines in the L2 cache (data intervention) if L2CTL[L2INTDIS] = 0. An optional unlock attribute causes I/O reads to clear a lock when the read is performed. The L2 cache does not respond to cache-inhibited read transactions. e500 core-initiated, cache-inhibited store transactions invalidate the line when they hit on a valid L2 line. If the line is locked, it goes to the stale state. For other write transactions the cache-inhibited bit is ignored. Non-burst cacheable write transactions from the e500 core (generated by write-through cacheable stores) update a valid L2 cache line through a read-modify-write operation. e500 core cast out transactions that hit on a stale line in the L2 cache cause a data update of the line and a change to the valid locked state for that line. An e500 core-initiated, cacheable, non-write-through store that misses in L1 and hits on a line in the L2 invalidates that line in the L2. If the line is marked exclusive locked, the L2 marks the line as stale. Transactions that hit a stale L2 cache line that would cause an allocate if they miss cause a data update of the line (when data arrives from memory) and a change to the line’s valid locked state. Data is not supplied by the L2 cache for the read in this case. The following transactions kill the data and the respective locks when they hit a valid L2 line: — dcbf — dcbi The L2 cache supports mixed cache external writes and core-initiated writes to the same addresses if the core-initiated writes are marked coherency-required, caching allowed, not write-through (WIMG = 001x) and the external writes are marked coherency-required, caching-allowed. The L2 cache supports writes to the L2 cache from peripheral devices or from I/O controllers through snoop write transactions with addresses that hit in a programmed memory range. Full-cache-line (32-byte) write transactions update the data for a valid line in the L2, and if the line is not valid in the L2, a line is allocated. Sub-cache-line write transactions update the data only for valid L2 cache lines through read-modify-write operations. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 7-23 L2 Look-Aside Cache/SRAM • • The L2 cache supports burst writes that lock an L2 cache line from peripheral devices or from I/O controllers through write transactions with addresses that hit in a programmed memory range that has the lock attribute set. The L2 cache supports burst writes that allocate and/or lock an L2 cache line from peripheral devices or I/O controllers through a write allocate transaction. See the system logic programming model (for example, that of the DMA controller) for details on how to set the transaction type for cache external writes to the L2. 7.6.2 • Memory-Mapped SRAM Coherency Rules External (non–core-initiated) accesses to memory-mapped SRAM must be marked coherency-required. External accesses marked coherency-not-required to memory-mapped SRAM may cause an address unavailable error. Accesses to memory-mapped SRAM are cacheable only in the corresponding e500 L1 caches. External accesses must be marked cache-inhibited or be performed with non-caching transactions. Memory-mapped SRAM coherency rules are as follows: • 7.7 • L2 Cache Locking Cache locking methods — Individual line locks are set and cleared by using instructions defined by the e500 cache locking APU, which is part of the Freescale Book E Implementation Standards (EIS). These instructions include Data Cache Block Touch and Lock Set (dcbtls), Data Cache Block Touch for Store and Lock Set (dcbtstls), and Instruction Cache Block Touch and Lock Set (icbtls). For detailed information about these instructions, see the PowerPC e500 Core Reference Manual. — A lock attribute can be attached to write operations. — Individual line locks are set and cleared through core-initiated instructions, by external reads or writes, or by accesses to programmed memory ranges defined in L2 cache external write address registers (L2CEWARn). — The entire cache can be locked by setting configuration registers appropriately. Methods for clearing locks — Individual locks are cleared by cache locking APU instructions (Instruction Cache Block Lock Clear (icblc) and Data Cache Block Lock Clear (dcblc)) or by snooped flush unless the entire cache is locked. The MPC8540 caches can be locked and cleared using the following methods: • MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 7-24 Freescale Semiconductor L2 Cache Locking — Flash clearing of all instruction and/or data locks can be done by writes to configuration registers. — An unlock attribute can be attached to I/O read operations. 7.7.1 Locking the Entire L2 Cache The entire L2 cache can be locked by setting L2CTL[L2DO] = 1 and L2CTL[L2IO] = 1. This has the effect of preventing any further allocation of new lines in the cache by core requests. If there are lines in the cache that are not valid, they cannot be used by core requests until the cache is unlocked. While the cache is locked, read requests are serviced as normal, and snooping continues as normal to maintain coherency. Lines invalidated to satisfy coherency requirements cannot be reallocated by core requests while the cache remains locked. The L2 cache can be unlocked by clearing L2CTL[L2IO] and/or L2CTL[L2DO]. Note that L2CTL[L2DO] and L2CTL[L2IO] have no effect on cache external write allocations or memory-mapped SRAM. Note that this form of cache locking does not use the lock bits of the cache and cannot be cleared by resetting the cache or lock bits. 7.7.2 Locking Programmed Memory Ranges A programmed memory range can be locked with a snoop write transaction that matches a cache external write address range (specified by L2CEWARn and L2CEWCRn). There is no clearing of locks through the programmed address ranges. Locks can be cleared using clear lock instructions, flushes, or read-and-clear-lock snoop (RWNITC with clear lock attribute), or flash clear locks. 7.7.3 • • • • • Locking Selected Lines icbtls (CT = 1)—Instruction Cache Block Touch and Lock Set instruction dcbtls (CT = 1)—Data Cache Block Touch and Lock Set instruction dcbtstls (CT = 1)—Data Cache Block Touch for Store and Lock Set instruction Snoop burst write—If the address hits on a programmed cache external write space with the lock attribute set, or if the write allocate transaction type is used Snoop non-burst write—If the address hits on a programmed cache external write space with the lock attribute set Individual lines are locked when the L2 receives one of the following burst transactions: Note that the core complex broadcasts these instructions to the L2 if the CT field in the instruction specifies the L2 cache (CT = 1). When the L2 cache is specified, data is not placed in the L1, only the L2. If the L1 cache is specified (CT = 0), the L2 does not lock the line, and the data is placed in the L1 (and locked). MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 7-25 L2 Look-Aside Cache/SRAM When the touch lock set L2 instruction (dcbtls or dcbtstls) hits are modified in the L1 cache, the modified data is allocated into the L2 cache (and written back to main memory) and a data lock is set. The L1 line state transitions to invalid. Note that if the L2 receives a request to allocate and lock a line, but all lines in the selected way are locked, the requested L2 line is not allocated and the L2 cache lock overflow bit (L2CTL[L2LO]) is set. Lines invalidated to satisfy coherency requirements cannot be reallocated while the cache remains locked. 7.7.4 Clearing Locks on Selected Lines Individual locks in the L2 are cleared by a lock clear (icblc or dcblc, CT = 1) instruction. This directs the L2 cache to clear a lock on that line if it hits in the L2 cache. Both data and instruction locks are cleared by the icblc and dcblc instructions. Note that the lock on a line is cleared if the line is invalidated by a snooped flush transaction, and the line in the cache is available for allocation of a new line of instruction or data unless the entire cache is locked. NOTE There is a scenario in which a lock clear operation appears to fail to clear a lock in the L2 cache. This occurs only when the attempt to set the lock results in a bus error (for example, PCI returns an error condition). Assume the following scenario: 1. The e500 attempts to set a lock in the L2 cache (by executing a dcbtls or icbtls instruction with CT = 1). The line is not already present in the cache, so it must be read from external memory. This read encounters an error which, depending on the chip configuration, will be reported to the core (probably as an interrupt). 2. At (or near) the same time, a cache external write to the same cache line is being mastered by the ECM. 3. Very soon after the cache external write, a transaction to clear the lock occurs. This can be caused by the processor executing a dcblc or icblc instruction with CT = 1, or by the ECM mastering a lock clear transaction. If this scenario occurs within a tight timing window, the cache line may unexpectedly remain locked at the end of the sequence. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 7-26 Freescale Semiconductor PLRU L2 Replacement Policy The interrupt handler may want to clear the erroneously remaining lock in this case. 7.7.5 Flash Clearing of Instruction and Data Locks Locks for instructions and data are recorded separately in the L2 cache, and they can be flash cleared separately by writing the appropriate value to the L2 cache control register (L2CTL[L2LFR] and L2CTL[L2LFRID]). Flash invalidating of the L2 (setting L2CTL[L2I]) clears all locks on both instructions and data. Note that flash clearing is the only way to clear data locks without clearing instruction locks, or to clear instruction locks without clearing data locks. All instructions and snoop transactions that clear locks clear both data and instruction locks. 7.7.6 Locks with Stale Data If data is locked in the L2 and either the e500 core performs a cacheable copyback store or a dcbtst misses in the L1, the L2 invalidates the line; however, the L2 clears the valid bit for the data, the lock remains, and the line cannot be victimized. If the e500 core casts out modified data or pushes it in response to a non-flush snoop, the L2 updates the data and sets the valid bit again, maintaining the lock and keeping the data in the cache hierarchy. 7.8 PLRU L2 Replacement Policy Line replacement is determined using a pseudo least-recently-used (PLRU) algorithm. There is a valid bit (V0–V7) for each line. To determine the replacement victim (the line to be cast out), there are seven PLRU bits (P0–P6) for each set. PLRU bits are updated every time a new line is allocated or replaced and every time a line is modified or invalidated. There are two sets of lock bits, one for instructions (I0–I7) and one for data (D0–D7) for every line. The lock bits act as a mask over the PLRU bits to determine victim selection. The PLRU bits are updated regardless of line locking. Figure 7-23 shows the binary decision tree used to generate the victim line. The eight ways of the L2 cache are labeled W0–W7; the seven PLRU bits are labeled P0–P6. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 7-27 L2 Look-Aside Cache/SRAM P0 = 0? Yes P1 = 0? Yes P3 = 0? Yes No No P4 = 0? Yes No Yes P5 = 0? Yes No No P2 = 0? No P6 = 0? Yes No Allocate Allocate Allocate Allocate Allocate Allocate Allocate Allocate W0 W1 W2 W3 W4 W5 W6 W7 P0 P1 P2 P3 P4 P5 P6 0 0 x 0 x x x 0 0 x 1 x x x 0 1 x x 0 x x 0 1 x x 1 x x 1 x 0 x x 0 x 1 x 0 x x 1 x 1 x 1 x x x 0 1 x 1 x x x 1 Figure 7-23. L2 Cache Line Replacement Algorithm 7.8.1 PLRU Bit Update Considerations PLRU bits are updated when a way is modified, either by a write, an invalidate, or an allocation of a new line. PLRU bit updates depend on which cache way was last accessed, as summarized in Table 7-19. Table 7-19. PLRU Bit Update Algorithm PLRU Bits Last Way Accessed P 0 1 1 1 1 0 0 0 0 P 1 1 1 0 0 — — — — P 2 — — P 3 1 0 P 4 P 5 P 6 0 1 2 3 4 5 6 7 ——— ——— 1 0 —— —— 1 0 — — 1 0 —— —— 1 1 0 0 —— —— ——— ——— When an L2 line is invalidated, the PLRU bits are updated, marking the corresponding way as least-recently-used. This causes the invalidated way to be selected as the next victim. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 7-28 Freescale Semiconductor L2 Cache Operation 7.8.2 Allocation of Lines L2 cache lines are locked through the status array lock bits. There are two lock bits for each way of each set (1024 sets by eight ways). These bits are set or cleared through special L2 controller commands. Lock bits are used at allocate time to steer the PLRU algorithm away from selecting locked victims. In the following discussion, the eight lock bits for a particular set are called L0–L7. Where Lock Way i: Li = Di | Ii, i=0...7 (Di = data lock, Ii = instruction lock) An effective value of each PLRU bit is calculated as follows: P0_eff P1_eff P2_eff P3_eff P4_eff P5_eff P6_eff = = = = = = = f(P0,L0,L1,L2,L3,L4,L5,L6,L7) = (L0 & L1 & L2 & L3) | (P0 & ~(L4 & L5 & L6 & L7)) f(P1,L0,L1,L2,L3) = (L0 & L1) | (P1 & ~(L2 & L3)) f(P2,L4,L5,L6,L7) = (L4 & L5) | (P2 & ~(L6 & L7)) f(P3,L0,L1) = L0 | (P3 & ~L1) f(P4,L2,L3) = L2 | (P4 & ~L3) f(P5,L4,L5) = L4 | (P5 & ~L5) f(P6,L6,L7) = L6 | (P6 & ~L7) These effective PLRU bits are used to select a victim, as indicated in Table 7-20. Table 7-20. PLRU-Based Victim Selection Mechanism Way Selected W0 W1 W2 W3 W4 W5 W6 W7 PLRU State (Binary) 00x0xxx 00x1xxx 01xx0xx 01xx1xx 1x0xx0x 1x0xx1x 1x1xxx0 1x1xxx1 Reduced Logic Equation ~P0 & ~P1 & ~P3 ~P0 & ~P1 & P3 ~P0 & P1 & ~P4 ~P0 & P1 & P4 P0 & ~P2 & ~P5 P0 & ~P2 & P5 P0 & P2 & ~P6 P0 & P2 & P6 7.9 L2 Cache Operation This section describes the behavior of the L1 and L2 cache in response to various operations and in various configurations. 7.9.1 L2 Cache States The L2 status array uses four bits for each line to determine the status of the line. Different combinations of these bits result in different L2 states. The status bits are as follows: • • • • Valid (V) Instruction locked (IL) Data locked (DL) Stale (T) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 7-29 L2 Look-Aside Cache/SRAM Table 7-21 shows L2 cache states. Note that these conventions are also used in Table 7-22. Table 7-21. L2 Cache States V 0 1 1 1 1 1 1 1 1 T x 0 0 0 0 1 1 1 1 IL x 0 0 1 1 0 0 1 1 DL x 0 1 0 1 0 1 0 1 Invalid (I) Exclusive (E) Exclusive data locked (EDL) Exclusive instruction locked (EIL) Exclusive instruction and data locked (EL) Stale (data invalid, locks invalid) (T) Stale (data invalid, dlock valid) (TDL) Stale (data invalid, ilock valid) (TIL) Stale (data invalid, locks valid) (TL) L2 states 7.9.2 Flash Invalidation of the L2 Cache The L2 cache may be completely invalidated by setting the L2I bit of the L2 control register (L2CTL). Note that no data is lost in this process because the L2 cache is a write-through cache and contains no modified data. Flash invalidation of the cache is necessary when the cache is initially enabled and may be necessary to recover from some error conditions such as a tag parity error. The invalidation process requires several cycles to complete. The L2I bit remains set during this procedure and is then cleared automatically when the procedure is complete. The L2 cache controller issues retries for all transactions on the e500 core complex bus (CCB) while the flash invalidation process is in progress. Note that the contents of memory-mapped SRAM regions of the data array are unaffected by a flash invalidation of the L2 cache regions fo the array. 7.9.3 L2 State Transitions Table 7-22 lists state transitions for all e500 core-initiated transactions that change the L2 cache state. Core-initiated transactions caused when the core executes msync, mbar, tlbivax, or tlbsync do not change the L2 cache state. The table does not list initial L1 states for transactions that hit in the L1 (iL1 or dL1) and are not sent to the L2. In the table, the heading ‘L2 hit’ indicates that the L2 provides (on a read) or captures (on a write) data for an existing line. Some entries list two final L1 states. L2 touch instructions never allocate into iL1 or dL1. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 7-30 Freescale Semiconductor L2 Cache Operation Note that if the L2 SRAM is disabled, the L2 initial and final states are always I and the L2 never hits. Similarly, if the L2 SRAM is in full memory-mapped SRAM mode, the L2 initial and final states are always I and the L2 never hits for addresses not in the memory-mapped SRAM address range. The L2 always hits for addresses in the enabled memory-mapped SRAM address ranges. Table 7-22. State Transitions Due to Core-Initiated Transactions Initial States Source of Transaction L1 Cacheable instruction fetch icbtls_L1 icbt_L2 iL1 I L2 I/T I dL1 I,E E/EL T I/T E EL I E EL T Cache-inhibited instruction fetch Cacheable load (4-state) Cacheable lwarx (4-state) dcbt_L1 (4-state) dcbtls_L1 (4-state) N/A dL1 I N/A I/T E EL I E/EL T Cache-inhibited load Cache-inhibited lwarx Writeback Store N/A N/A dL1 I N/A N/A I/T E EL Writeback stwcx. dL1 I I/T E EL L2 Hit No No Yes No No Yes Yes No Yes Yes No No No Yes Yes No Yes No No No No Yes Yes No Yes Yes Final States Comments L1 I/V I/V I/V I/V I I I I I I I N/A E E E E E EL N/A N/A M M M M M M L2 Same L2CTL[L2DO] = 1. L2 touch instructions not allocated in L1 E Same EL L2CTL[L2DO] = 0. Restore locked line in L2 with valid data from bus L2CTL[L2DO] = 0 icbtls_L2 dL1 I,E Same L2CTL[L2DO] = 1 I T EL EL L2CTL[L2DO] = 1 L2CTL[L2DO] = 1 L2CTL[L2DO] = 0 L2CTL[L2DO] = 0 Same L2CTL[L2DO] = 0 EL N/A L2CTL[L2DO] = 0. Restore locked line in L2 with valid data from bus No L1/L2 effect Same L2CTL[L2IO] = 1 I T E L2CTL[L2IO] = 1 L2CTL[L2IO] = 1 L2CTL[L2IO] = 0 Same L2CTL[L2IO] = 0 EL N/A N/A L2CTL[L2IO] = 0. Restore locked line in L2 with valid data from bus No L1/L2 effect No L2 effect Same L2 allocates when a line is cast out of L1. I T Same I T MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 7-31 L2 Look-Aside Cache/SRAM Table 7-22. State Transitions Due to Core-Initiated Transactions (continued) Initial States Source of Transaction L1 Cacheable load (3-state) Cacheable lwarx (3-state) dcbt_L1 (3-state) dcbtls_L1 (3-state) dL1 I L2 I T E EL I dcbt_L2 dcbtst_L2 dcbtst_L1 dcbtstls_L1 dL1 I,E E/EL T I/T E EL dcbtls_L2 dcbtstls_L2 dL1 I,E I T E EL I E/EL T Write-through store dL1 I,E,M N/A I/T E/EL I/E EL/T Cache-inhibited stwcx. N/A I/E EL/T dcblc_L2 icblc_L2 dL1 I,E,M I/E EL T Victim castout dcbt_L2 icbt_L2 dcbtst_L2 dL1 M I/T L2 Hit No No Yes Yes No Yes No No Yes Yes No No Yes Yes No Yes No No Yes No No No No No No No No Final States Comments L1 E/I E/I E/I E/I E/I E/I E/I E E E I I I I I I I Same Same N/A N/A N/A N/A Same Same Same I L2 I T I T E L2CTL[L2IO] = 1 L2CTL[L2IO] = 1 L2CTL[L2IO] = 1 L2CTL[L2IO] = 1 L2CTL[L2IO] = 0 Same L2CTL[L2IO] = 0 EL Same I T I T I T EL EL EL I Same Read-modify-write I T I T Same E I Same L2CTL[L2IO] = 1. If software sharing cache lines between instructions and data wishes to capture instruction lines in L2 with L2CTL[L2IO] = 1, it must perform dcbst to flush the line out of the dL1 before fetching it into L2. E I/T L2CTL[L2IO] = 0 L2CTL[L2IO] = 1 Invalidate line Invalidate data, keep lock Invalidate line Invalidate data, keep lock L2CTL[L2IO] = 1 L2CTL[L2IO] = 1 L2CTL[L2IO] = 1 L2CTL[L2IO] = 1 L2CTL[L2IO] = 0 L2CTL[L2IO] = 0 L2CTL[L2IO] = 0. Restore locked line with valid data from bus L2CTL[L2IO] = 0. Restore locked line with valid data from bus dL1 I Cache-inhibited store I E/EL No No Yes I I I I Same L2CTL[L2IO] = 0 EL L2CTL[L2IO = 0 T Yes MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 7-32 Freescale Semiconductor L2 Cache Operation Table 7-22. State Transitions Due to Core-Initiated Transactions (continued) Initial States Source of Transaction L1 dcbtls_L2 icbtls_L2 dcbtstls_L2 dL1 M L2 I L2 Hit No Final States Comments L1 I L2 EL An icbtls_L2 that hits modified in L1 cannot be distinguished from dcbtls_L2 and sets the L2 dlock bit. If software shares cache lines between instructions and data and wishes to set ilocks in L2, it must perform dcbst to flush the line out of the dL1 before locking it in L2. E/EL/T Yes Snoop push dL1 M dL1 M dL1 I dL1 I,E,M dL1 I,E iL1 I,V I/E EL/T I/E/EL I/E EL I/ E/EL/T I/ E/EL/T I/ E/EL/T No No No No No No No No I I/E I/E I M M I I I EL I T I I T I I I Invalidate data, keep lock dcbf dcbst dcbz dcba dcbi dcbf dcbst icbi Table 7-23 lists L2 cache state transitions for all system-initiated (non-core) transactions that change the L2. The transaction types and attributes listed follow MPX bus nomenclature, with the addition of write allocate (burst write with L2 cache allocation). Table 7-23 accounts for changes caused by L1 snoop pushes triggered by snoops, listed in Table 7-22. Table 7-23. State Transitions Due to System-Initiated Transactions Transaction Type Clean IKill Flush Write allocate wt x x 0 1 ci gbl x x 1 1 0 0 0 0 Initial L2 State I/E/EL/T I/E/EL/T I/E/EL/T I/E EL/T x 0 0 I/E EL/T Final L2 State Same I EL E EL I T No allocate if cache-inhibited Invalidate data, keep lock Allocate and lock regardless of cache external write (CEW) window Allocate regardless of CEW window Comments MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 7-33 L2 Look-Aside Cache/SRAM Table 7-23. State Transitions Due to System-Initiated Transactions (continued) Transaction Type WWK 32-byte WWF 32-byte WWF atomic wt x ci gbl 1 0 Initial L2 State I/E/EL/T I EL E T x 0 0 I/E EL/T < 32-byte WWF < 32-byte WWF atomic x 1 0 I/E EL/T I/T EL E x 0 0 I/E EL/T Read Read atomic 1 1 0 I/T E EL x RWNITC 1 0 1 0 0 N/A I/L/T E EL 0 1 0 I EL T x Kill RWITM RWITM atomic RClaim x 0 1 0 0 N/A I/E EL/T x 0 0 I/E/EL/T Final L2 State I E/EL Same E/EL EL I T I T Same Same E/EL I T Same E EL N/A Same E EL Same E I N/A I T Same Invalidate data, keep lock No L1/L2 effect Read-and-clear-lock No L1/L2 effect Invalidate line Invalidate data, keep lock Miss in cache external write windows Miss in cache external write windows Hit in CEW window but need burst data Hit in cache external write window Hit in cache external write window. Set lock if CEW lock attribute set Invalidate line Invalidate data, keep lock Comments Miss in cache external write windows Hit in cache external write window. Set lock if CEW lock attribute set 7.10 Initialization/Application Information This section describes some required steps for initializing the L2 SRAM both in L2 cache mode, and in memory-mapped SRAM mode. Also, it includes some guidlines for error managment. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 7-34 Freescale Semiconductor Initialization/Application Information 7.10.1 Initialization This section describes L1 cache and memory-mapped SRAM initialization. 7.10.1.1 L2 Cache Initialization After power-on reset, the valid bits in the L2 cache status array are in random states. Therefore, it is necessary to perform a flash invalidate operation before using the array as an L2 cache. This is done by writing a 1 to the L2I bit of the L2 control register (L2CTL). This can be done before or simultaneously with the write that enables the L2 cache. That is, the L2E and L2I bits of L2CTL can be set simultaneously. The L2I bit clears automatically, so no further writes are necessary. 7.10.1.2 Memory-Mapped SRAM Initialization After power-on reset, the contents of the data and ECC arrays and are random, so all SRAM data must be initialized before it is read. If the cache is initialized by the core or any other device that uses sub-cacheline transactions, ECC error checking should be disabled during the initialization process to avoid false ECC errors generated during the read-modify-write process used for sub-cacheline writes to the SRAM array. This is done by setting the multi- and single-bit ECC error disable bits of the L2 error disable register (L2ERRDIS[MBECCDIS, SBECCDIS]). See Section 7.3.1.5.2, “Error Control and Capture Registers”. If the array is initialized by a DMA engine using cache-line writes, then ECC checking can remain enabled during the initialization process. 7.10.2 Managing Errors This section describes recommended handling for ECC and tag parity errors. 7.10.2.1 ECC Errors An individual soft error that causes a single- or mulit-bit ECC error can be cleared from the L2 array simply by executing a dcbf instruction for the address captured in the L2ERRADDR register. This will invalidate the line in the L2 cache. When the load that caused the ECC error is performed again, the data will be re-allocated into the L2 with ECC bits set properly again. If the threshold for single bit errors set in the L2ERRCTL register is exceeded, then the L2 cache should be flash invalidated to clear out all single-bit errors. Note that no data is lost by executing dcbf instructions or flash invalidate operations because the L2 cache is write-through and contains no modified data. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 7-35 L2 Look-Aside Cache/SRAM 7.10.2.2 Tag Parity Errors A tag parity error must be fixed by flash invalidating the L2 cache. Note that executing a dcbf instruction for the address that caused the error to be reported is not sufficient because a tag parity error is seen as an L2 miss and does not cause invalidation of the bad tag. Proper L2 operation cannot be guaranteed if an L2 tag parity error is not repaired by a flash invalidation of the entire array. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 7-36 Freescale Semiconductor Part III Memory and I/O Interfaces P Part III defines the memory and I/O interfaces of the MPC8540 and it describes how these blocks interact with one another and with other blocks on the device. The following chapters are included: • Chapter 8, “e500 Coherency Module,” defines the e500 coherency module and how it facilitates communication between the e500 core complex, the L2 cache, and the other blocks that comprise the coherent memory domain of the MPC8540. The ECM provides a mechanism for I/O-initiated transactions to snoop the core complex bus (CCB) of the e500 core in order to maintain coherency across cacheable local memory. It also provides a flexible, easily expandable switch-type structure for e500- and I/O-initiated transactions to be routed (dispatched) to target modules on the MPC8540. Chapter 9, “DDR Memory Controller,” describes the DDR SDRAM memory controller of the MPC8540. This fully programmable controller supports most DDR memories available today, including both buffered and unbuffered devices. The built-in error checking and correction (ECC) ensures very low bit error rates for reliable high-frequency operation. Dynamic power management and auto-precharge modes simplify memory system design. A large set of special features like DLL software override, crawl mode, and ECC error injection support rapid system debug. Chapter 10, “Programmable Interrupt Controller,” describes the embedded programmable interrupt controller (PIC) of the MPC8540. This controller is an OpenPIC-compliant interrupt controller that provides interrupt management, and is responsible for receiving hardware-generated interrupts from different sources (both internal and external), prioritizing them, and delivering them to the CPU for servicing. Chapter 11, “I2C Interface,” describes the inter-IC (IIC or I2C) bus controller of the MPC8540. This synchronous, serial, bidirectional, multi-master bus allows two-wire connection of devices such as microcontrollers, EEPROMs, real-time clock devices, A/D converters, and LCDs. The MPC8540 powers up in boot sequencer mode, which allows the I2C controller to initialize configuration registers. Chapter 12, “DUART,” describes the (dual) universal asynchronous receiver/transmitters (UARTs) which feature a PC16552D-compatible programming model. These independent UARTs are provided specifically to support system debugging. Chapter 13, “Local Bus Controller,” describes the local bus controller of the MPC8540. The main component of the local bus controller (LBC) is its memory controller which provides a seamless interface to many types of memory devices and peripherals. The MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor III-1 • • • • • • • • • memory controller is responsible for controlling eight memory banks shared by a high performance SDRAM machine, a general-purpose chip-select machine (GPCM), and up to three user-programmable machines (UPMs). As such, it supports a minimal glue logic interface to synchronous DRAM (SDRAM), SRAM, EPROM, flash EPROM, burstable RAM, regular DRAM devices, extended data output DRAM devices, and other peripherals. Chapter 14, “Three-Speed Ethernet Controllers,” describes the two three-speed Ethernet controllers on the MPC8540. These controllers provide 10/100/1Gb Ethernet support with a complete set of media-independent interface options including GMII, RGMII, TBI, and RTBI. Each controller provides very high throughput using a captive DMA channel and direct connection to the MPC8540 memory coherency module. Chapter 15, “DMA Controller,” describes the four-channel general-purpose DMA controller of the MPC8540. The DMA controller transfers blocks of data, independent of the e500 core or external hosts. Data movement occurs among RapidIO and the local address space. The DMA controller has four high-speed channels. Both the e500 core and external masters can initiate a DMA transfer. All channels are capable of complex data movement and advanced transaction chaining. Chapter 16, “PCI/PCI-X Bus Interface,” describes the PCI/PCI-X controller of the MPC8540. Chapter 17, “RapidIO Interface,” describes the RapidIO controller of the MPC8540. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 III-2 Freescale Semiconductor Chapter 8 e500 Coherency Module 8.1 Introduction The e500 coherency module (ECM) provides a flexible, easily expandable switching structure for routing e500- and I/O-initiated transactions to target modules on the device. Figure 8-1 shows a high-level block diagram of the ECM. Core Complex Bus (to/from e500) CCB Interface Configuration Bus Config Reg’s Port Data Bus (internal) OCeaN TSEC1 Request Buses TSEC2 10/100 CCB Arbiter Transaction Queue Dispatch Bus (to OCeaN, DDR, LBC…) Global Data Bus Arbiter Global Data Mux I/O Arbiter Global Data Bus (to OCeaN, DDR, LBC,…) Data Bus Internal Global • Port Data Buses • (from OCeaN, • DDR, LBC…) • Figure 8-1. e500 Coherency Module Block Diagram 8.1.1 Overview The ECM routes transactions initiated by the e500 core to the appropriate target interface on the device. In a manner analogous to a bridging router in a local area network, the ECM forwards I/O-initiated transactions that are tagged with the global attribute onto the core complex bus MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 8-1 e500 Coherency Module (CCB). This allows on-chip caches to snoop these transactions as if they were locally initiated and to take actions to maintain coherency across cacheable memory. 8.1.2 • Features Support for e500 core and an L2/SRAM on the CCB, including a CCB arbiter. It sources a 64-bit data bus for returning read data from the ECM to the e500 core and routing write data from the ECM to the L2/SRAM. It sinks a 128-bit data bus for receiving data from the L2/SRAM and a 128-bit write data bus from the e500 core. Four connection points for I/O initiating (mastering into the device) interfaces. One of those connection points services OCeaN initiators, one services TSEC1, one services TSEC2, and the last services other I/O initiators. The ECM supports five connection points for I/O targets. The DDR memory controller, local bus, OCeaN targets, programmable interrupt controller (PIC), and configuration register access block all have a target port connection to the ECM. Split transaction support—separate address and data tenures allow for pipelining of transactions and out-of-order data tenures between initiators and targets. Proper ordering of I/O-initiated transactions. Speculative read bus for low-latency dispatch of reads to the DDR controller. Low-latency path for returning read data from DDR to the e500. Error registers trap transactions with invalid addresses. Errors can be programmed to generate interrupts to the e500 core, as described in the following sections: — Section 8.2.1.3, “ECM Error Detect Register (EEDR)” — Section 8.2.1.4, “ECM Error Enable Register (EEER)” — Section 8.2.1.5, “ECM Error Attributes Capture Register (EEATR)” — Section 8.2.1.6, “ECM Error Address Capture Register (EEADR)” Errors from I/O devices on reads (for example, a master-aborted read transaction on the PCI interface) terminate with corrupt data sent to the requester. If the requester is the e500 core, the ECM also asserts core_fault_in to the core, which causes the core to generate a machine check interrupt, unless it is disabled (by clearing HID1[RFXE]). If RFXE is zero and one of these errors occurs, appropriate interrupts must be enabled to ensure that an interrupt is generated. See Section 6.10.2, “Hardware Implementation-Dependent Register 1 (HID1).” The ECM includes these distinctive features: • • • • • • • 8.2 Memory Map/Register Definition Table 8-1 shows the ECM’s memory map. Undefined 4-byte address spaces within offset 0x000–0xFFF are reserved. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 8-2 Freescale Semiconductor Memory Map/Register Definition Table 8-1. ECM Memory Map Local Memory Offset 0x0_1000 0x0_1010 0x0_1E00 0x0_1E08 0x0_1E0C 0x0_1E10 Register EEBACR—ECM CCB address configuration register EEBPCR—ECM CCB port configuration register EEDR—ECM error detect register EEER—ECM error enable register EEATR—ECM error attributes capture register EEADR—ECM error address capture register Acces s R/W R/W Reset 0x0000_000 3 0x0n00_000 0 Section/Pag e 8.2.1.1/8-3 8.2.1.2/8-4 8.2.1.3/8-5 8.2.1.4/8-6 8.2.1.5/8-7 8.2.1.6/8-8 Special 0x0000_000 0 R/W R R 0x0000_000 0 0x0000_000 0 0x0000_000 0 8.2.1 Register Descriptions This section consists of detailed descriptions of those registers summarized in Table 8-1. Note that these registers are shown in big-endian format. 8.2.1.1 ECM CCB Address Configuration Register (EEBACR) The ECM CCB address configuration register, shown in Figure 8-2, controls arbitration and streaming policies for the CCB. 0 27 28 29 30 31 R A_STRM_DIS CORE_STRM_DIS A_STRM_CNT W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0011 0x0_1000 Figure 8-2. ECM CCB Address Configuration Register (EEBACR) Table 8-2 describes the EEBACR fields. Table 8-2. EEBACR Field Descriptions Bits 0–27 28 Name — A_STRM_DIS Reserved Controls whether the ECM allows any streaming to occur. 0 Streaming is enabled. 1 Streaming is disabled. Description MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 8-3 e500 Coherency Module Table 8-2. EEBACR Field Descriptions (continued) Bits 29 Name Description CORE_STRM_DIS With A_STRM_DIS, controls whether the e500 core can stream commands onto the CCB. A_STRM_DIS and CORE_STRM_DIS must both be cleared for the e500 core to be enabled to stream address tenures that it masters. 0 Stream address tenures initiated by the e500 core, provided A_STRM_DIS is cleared. 1 Streaming of address tenures initiated by the e500 core not allowed. A_STRM_CNT Stream count. Specifies the maximum number of transactions that any master can stream (issue sequentially without preemption) on the CCB following an initial transaction. 00 Reserved 01 One transaction can be streamed with the initial transaction. 10 Two transactions can be streamed with the initial transaction. 11 Three transactions can be streamed with the initial transaction. Default. 30–31 8.2.1.2 ECM CCB Port Configuration Register (EEBPCR) The ECM CCB port configuration register (EEBPCR) is shown in Figure 8-3. 0 6 7 8 29 30 31 R CPU_EN W Reset Offset 0000_000n_0000_0000_0000_0000_0000_0000 0x0_1010 CPU_PRI Figure 8-3. ECM CCB Port Configuration Register (EEBPCR) Table 8-3 describes EEBPCR fields. Table 8-3. EEBPCR Field Descriptions Bits 0–6 7 Name — Reserved Description CPU_EN CPU port enable. Controls boot holdoff mode when the device is an agent of an external host. Specifies whether the e500 core (CPU) port is enabled to run transactions on the CCB. The CPU boot configuration power-on reset pin (cfg_cpu_boot) determines the initial value of this bit. If the pin is sampled as a logic 1 at the negation of reset, the CPU is enabled to boot at the end of the POR sequence. Otherwise, the CPU cannot fetch its boot vector until an external host sets the CPU_EN bit. 0 Boot holdoff mode. CPU arbitration is disabled on the CCB and no bus grants are issued. 1 CPU is enabled and receives bus grants in response to bus requests for the boot vector. After this bit is set, it should not be cleared by software. It is not intended to dynamically enable and disable CPU operation. It is only intended to end boot holdoff mode. See Section 4.4.3.5, “CPU Boot Configuration,” for more information. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 8-4 Freescale Semiconductor Memory Map/Register Definition Table 8-3. EEBPCR Field Descriptions (continued) Bits 8–29 Name — Reserved Description 30–31 CPU_PRI Specifies the priority level of the e500 core (CPU) port. This priority level is used to determine whether a particular port’s bus request can cause the CCB arbiter to terminate another port’s streaming of address tenures. 00 Lowest priority level 01 Second lowest priority level 10 Highest priority level 11 Reserved 8.2.1.3 ECM Error Detect Register (EEDR) The ECM error detect register (EEDR) is shown in Figure 8-4. 0 1 30 31 R MULT_ERR W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x0_1E00 LAE Figure 8-4. ECM Error Detect Register (EEDR) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 8-5 e500 Coherency Module Table 8-4 describes EEDR fields. Table 8-4. EEDR Field Descriptions Bits 0 Name Description MULT_ERR Multiple error. Indicates the occurrence of multiple errors of the same type. Write one to clear. 0 Multiple errors of the same type were not detected. 1 Multiple errors of the same type were detected. — LAE Reserved Local access error. Write one to clear. Three cases can generate LAEs: • Transaction does not map to any target. In this case the ECM injects read responses (with the corrupt attribute set) and write data is dropped. Note that a read that attempts to access an unmapped target causes the assertion of core_fault_in , which causes the core to generate a machine check interrupt, unless it is disabled (by clearing HID1[RFXE]). If RFXE is zero and this error occurs, EEER[LAEE] must be set to ensure that an interrupt is generated. For more information, see Section 6.10.2, “Hardware Implementation-Dependent Register 1 (HID1).” • Source and target IDs indicate that an OCN port initiated a transaction that targets an OCN port. This loopback behavior can result from programming errors where inbound ATMU window targets are inconsistent with targets configured in the local access windows for a given address range. For this type of LAE, the dispatch (to OCN target in this case) is not screened off; the LAE error is reported, but the transaction is still sent to its OCN target. • Inbound read atomic set, clear, inc, or dec transactions from RapidIO target an interface other than DDR (only the DDR supports these transactions). This condition screens off the dispatch to the target (if one is mapped); the ECM injects read responses (with the corrupt attribute set) and write data is dropped. 0 Local access error has not occurred. 1 Local access error occurred. 1–30 31 8.2.1.4 ECM Error Enable Register (EEER) The ECM error enable register (EEER) shown in Figure 8-5 enables the reporting of error conditions to the e500 core through the internal int interrupt signal. 0 30 31 R LAEE W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x0_1E08 Figure 8-5. ECM Error Enable Register (EEER) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 8-6 Freescale Semiconductor Memory Map/Register Definition Table 8-5 describes EEER fields. Table 8-5. EEER Field Descriptions Bits Name 0–30 31 — Reserved Description LAEE Local access error enable. Note that a read that attempts to access an unmapped target causes the assertion of core_fault_in, which causes the core to generate a machine check interrupt, unless it is disabled (by clearing HID1[RFXE]). If RFXE is zero and this error occurs, LAEE must be set to ensure that an interrupt is generated. For more information, see Section 6.10.2, “Hardware Implementation-Dependent Register 1 (HID1).” 0 Disable reporting local access errors as interrupts. 1 Enable reporting local access errors as interrupts. 8.2.1.5 ECM Error Attributes Capture Register (EEATR) The ECM error attributes capture register (EEATR) is shown in Figure 8-6. 0 2 3 7 8 10 11 15 16 17 20 21 30 31 R W Reset Offset BYTE_CNT SRC_ID TTYPE VAL 0000_0000_0000_0000_0000_0000_0000_0000 0x0_1E0C Figure 8-6. ECM Error Attributes Capture Register (EEATR) Table 8-6 describes EEATR fields. Table 8-6. EEATR Field Descriptions Bits 0–2 3–7 Name — Reserved Description BYTE_CNT Byte count. Specifies the transaction byte count. 00100 4 bytes 00000 32 bytes 01000 8 bytes 00001 1 byte 10000 16 bytes 00010 2 bytes 00011 3 bytes — Reserved 8–10 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 8-7 e500 Coherency Module Table 8-6. EEATR Field Descriptions (continued) Bits 11–15 Name SRC_ID Description Source ID. Specifies the source device mastering the transaction. 10101 DMA 00000 PCI/PCI-X interface 10110 Reserved 00001–01001 Reserved 10111 SAP 01010 Boot sequencer 11000 TSEC1 01011 Reserved 11001 TSEC2 01100 RapidIO 11010 10/100 FEC 01101–01111 Reserved 11011 Reserved 10000 Processor (instruction) 11100 RapidIO message units 10001 Processor (data) 11101 RapidIO doorbell units 10010 Reserved 11110 RapidIO port-write units 10011 Reserved 11111 Reserved 10100 Reserved Reserved Transaction type. Defined as follows: 0000 Write 0001 Reserved 0010 Write with allocate 0011 Write with allocate with lock 0100 Address only transaction 0101–0111 Reserved 1000 Read Reserved Register data valid. 0 ECM error attribute capture register does not contain valid information. 1 ECM error attribute capture register contains valid information. 1001 101x 1100 1101 1110 1111 Read with Reserved Read with Read with Read with Read with unlock clear atomic set atomic decrement atomic increment atomic 16 17–20 — TTYPE 21–30 31 — VAL 8.2.1.6 ECM Error Address Capture Register (EEADR) The ECM error address capture register (EEADR) is shown in Figure 8-7. 0 31 R W Reset Offset ADDR 0000_0000_0000_0000_0000_0000_0000_0000 0x0_1E10 Figure 8-7. ECM Error Address Capture Register (EEADR) Table 8-7 describes EEADR fields. Table 8-7. EEADR Field Descriptions Bits 0–31 Name ADDR Description Address. Specifies the 32-bit address of the transaction. Qualified by EEATR[VAL]. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 8-8 Freescale Semiconductor Functional Description 8.3 8.3.1 Functional Description I/O Arbiter The following is a very general discussion of ECM operation. Figure 8-1 shows the I/O arbiter block that manages I/O-initiated address tenure requests arriving on the request buses. Four request buses compete for access to the ECM, which can only process one request at a time. The ECM uses two factors to select the winning request bus: the primary factor is request priority and the secondary factor is longest waiting/least recently granted status. By default all requesters use the lowest priority (priority 0) for requests. The TSEC controllers dynamically raise and lower their priority levels based on FIFO depth. A starvation avoidance algorithm prevents high priority requests from indefinitely starving low priority requesters. The transaction from the winning request bus competes with e500 core requests for the CCB and entry into the transaction queue. 8.3.2 CCB Arbiter Figure 8-1 shows the CCB arbiter block coordinating the entry of new transactions into the ECM’s transaction queue. It handles arbitration for requests to use the CCB from the e500 core and the winning request bus and consequently controls when these new transactions can enter the transaction queue. Because the CCB bus operates most efficiently when it streams commands from one initiator, the CCB arbiter alternates grants between streams of transactions from the processor and from the winner of the I/O arbiter. The length of a stream (number of back-to-back transactions) is limited by the A_STRM_CNT field in the EEBACR register. However, the arbiter also uses the priority of the requests to limit streaming. If the priority of a new request is higher than that of a stream in progress, then the higher priority transaction will interrupt the other stream. The priority of e500 transactions is set by the CPU_PRI field in EEBPCR register. 8.3.3 Transaction Queue The ECM’s transaction queue performs three basic functions: target mapping and dispatching, enforcement of ordering, and enforcement of coherency. The address of each transaction is compared against each local access window, and the transaction is then routed to the appropriate target interface associated with the local access window that the address hits within. Even though the CCB and ECM allow the pipelining of transactions, the address tenures of all transactions issued from masters other than the e500 core (all I/O masters) are strictly ordered and are dispatched to their target interfaces in the same order they are submitted. For those transactions accessing address space marked as snoopable, or space that may be cached by the e500 core, the MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 8-9 e500 Coherency Module ECM enforces coherency, snooping those transactions on the CCB, and taking castouts from the e500 core as is necessary. 8.3.4 Global Data Multiplexor Figure 8-1 shows how the global data multiplexor takes data bus connections and multiplexes them onto one 128-bit global data bus. The global data mux allows initiators of write transactions to route data to their targets and read targets to return data to the initiators. 8.3.5 CCB Interface Figure 8-1 shows the CCB interface for both CCB address and data tenures. This interface formats CCB address tenures for the ECM transaction queue. It also contains the queueing and buffering needed to manage outstanding CCB data tenures. The buffers receive e500 core-initiated write and I/O-initiated read data (that hit in the L2/SRAM module) from the e500 write (128-bit wide) and read (128-bit wide) data buses and route them through the global data mux to the global data bus. The buffers also receive e500 core-initiated read and I/O-initiated write data (that hit in the L2/SRAM module) from the global data bus and forward them onto the CCB data bus (64 bits). 8.4 Initialization/Application Information If the e500 core is used to initialize the MPC8540, the CPU boot configuration power-on reset pin should be pulled high to initially set EEBPCR[CPU_EN]. See Chapter 4, “Reset, Clocking, and Initialization,” for more information on power-up reset initialization. If any device other than the e500 core, such as the boot sequencer or PCI, is used to initialize the MPC8540, the CPU boot configuration power-on reset pin should be pulled low to initially clear EEBPCR[CPU_EN]. This prevents the e500 core from accessing any configuration registers or local memory space during initialization. However, in any such system, one step near the end of the initialization routine must set EEBPCR[CPU_EN] to re-enable the e500 core. Note that for basic functionality, EEBPCR[CPU_EN] is the only field that must be written (provided a device other than the e500 core is used to initialize the device) in the ECM. EEBPCR[CPU_PRI] specifies the priority level associated with all e500 core initiated transactions. This value allows users running time-critical applications to adjust the average response latency of transactions initiated by the core compared to those initiated by I/O masters. This priority level affects whether the e500 core requests can interrupt the streaming of address tenures initiated by (the ECM on behalf of) I/O masters. Only transactions with a priority greater than the current CCB transaction can interrupt streaming. The higher the core’s priority, the lower the average latency needed for it to obtain bus grants from the ECM, because it can interrupt lower priority streaming. The default value of zero gives all core-initiated transactions the lowest priority, which prevents the core from interrupting I/O master transaction streams. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 8-10 Freescale Semiconductor Initialization/Application Information EEBACR[A_STRM_CNT] allows users to balance response latency with throughput and should prove useful in tuning systems with multiple time-critical tasks. The default value of 0b11 causes the ECM to attempt to stream as many as four transactions initiated from the same CCB master. Increasing this value increases the maximum number of transactions that may be streamed together from any one CCB master. Raising this value can increase throughput for high priority transactions, but may increase latency for lower priority transactions from another CCB master. Note that the e500 core must also have streaming enabled (through HID1[ASTME]) for the CCB to stream. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 8-11 e500 Coherency Module MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 8-12 Freescale Semiconductor Chapter 9 DDR Memory Controller 9.1 Introduction The fully programmable DDR SDRAM controller supports most first-generation JEDEC standard ×8 or ×16 DDR memories available, including buffered and unbuffered DIMMs. However, mixing unbuffered and registered DIMMs in the same system is not supported. Built-in error checking and correction (ECC) ensures very low bit-error rates for reliable high-frequency operation. Dynamic power management and auto-precharge modes simplify memory system design. A large set of special features, including DLL software override, crawl mode, and ECC error injection support rapid system debug. NOTE In this chapter, the word ‘bank’ refers to a physical bank specified by a chip select; ‘logical bank’ refers to one of the four sub-banks in each SDRAM chip. A sub-bank is specified by the two least significant bits (lsbs) of a bank address. Figure 9-1 is a high-level block diagram of the DDR memory controller with its associated interfaces. Section 9.5, “Functional Description,” contains detailed figures of the controller. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 9-1 DDR Memory Controller Request from Master Address from Master Address Decode Input Staging Queue Row Open Address Control DDR SDRAM Memory Array MA[0:14] MBA[0:1] DDR SDRAM Memory Control MCS[0:3] MCAS MRAS MWE MDM[0:8] MCKE[0:1] Data Qualifiers MDQS[0:8] Physical Bank, Logical Bank, Row ECM Open Row Table SDRAM Control EN To Error Management Error Signals ECC Delay chain Data from SDRAM FIFO Data Signals MDQ[0:63] MECC[0:7] EN Clocks MCK[0:5] MCK[0:5] MSYNC_OUT MSYNC_IN RMW Data from Master ECC SDRAM Control DLL Figure 9-1. DDR Memory Controller Simplified Block Diagram 9.2 • • • • Features Support for DDR SDRAM 64-/72-bit SDRAM data bus Programmable settings for meeting all SDRAM timing parameters The following SDRAM configurations are supported — As many as four physical banks (chip selects), each bank independently addressable — 64-Mbit to 1-Gbit devices with ×8/×16 data ports (no direct ×4 support) — Unbuffered and registered DIMMs Support for data mask signals and read-modify-write for sub-double word writes. Note that read-modify-write sequence is only necessary when ECC is enabled. Support for double-bit error detection and single-bit error correction ECC (8-bit check word across 64-bit data) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 The DDR memory controller includes these distinctive features: • • 9-2 Freescale Semiconductor External Signal Descriptions • • • • Two-entry input request queue Open page management (dedicated entry for each logical bank) Memory controller clock frequency of two times the SDRAM clock with support for sleep power management Support for error injection 9.2.1 • Modes of Operation Dynamic power management mode. The DDR memory controller can reduce power consumption by negating the SDRAM CKE signal when no transactions are pending to the SDRAM. Auto-precharge mode. Clearing DDR_SDRAM_INTERVAL[BSTOPRE] causes the memory controller to issue an auto precharge command with every read or write transaction. Auto precharge mode can be enabled for separate chip selects by setting CSn_CONFIG[AP_n_EN]. The DDR memory controller supports the following modes: • 9.3 External Signal Descriptions This section provides descriptions of the DDR memory controller’s external signals. It describes each signal’s behavior when the signal is asserted or negated and when the signal is an input or an output. NOTE A bar over a signal name indicates that the signal is active low, such as AS (address strobe). Active-low signals are referred to as asserted (active) when they are low and negated when they are high. Signals that are not active low, such as NMI (nonmaskable interrupt), are referred to as asserted when they are high and negated when they are low. 9.3.1 • • • Signals Overview Memory interface signals Clock signals Debug signals Memory controller signals are grouped as follows: MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 9-3 DDR Memory Controller Table 9-1 shows how DDR memory controller external signals are grouped. The MPC8540 Integrated Host Processor Hardware Specifications has a pinout diagram showing pin numbers. It also lists all electrical and mechanical specifications. Table 9-1. DDR Memory Interface Signal Summary Name MDQ[0:63] MDQS[0:8] MECC[0:7] MCAS MA[0:14] MBA[0:1] MCS[0:3] MWE MRAS MDM[0:8] MCK[0:5] MCK[0:5] MCKE[0:1] MSYNC_IN Function/Description Data bus Data strobes Error checking and correcting Column address strobe Address bus Logical bank address Chip select Write enable Row address strobe Data mask DRAM clock outputs DRAM clock outputs (complement) DRAM clock enable DRAM DLL synchronization in Reset All zeros All zeros All zeros One All zeros All zeros All ones One One All zeros All zeros All ones All zeros Zero One Zero All zeros Pin s 64 9 8 1 15 2 4 1 1 9 6 6 2 1 1 1 5 I/O I/O I/O I/O O O O O O O O O O O I O O O MSYNC_OUT DRAM DLL synchronization out MDVAL Memory debug data valid MSRCID[0:4] Memory debug source ID Table 9-2 shows the memory address signal mappings. Table 9-2. Memory Address Signal Mappings Signal Name (Outputs) msb MA14 MA13 MA12 MA11 MA10 MA9 MA8 MA7 MA6 JEDEC DDR DIMM Signals (Inputs) SDRAM 168-Pin DIMM — A13 A12 A11 A10(AP) A9 A8 A7 A6 lsb Signal Name (Outputs) MA5 MA4 MA3 MA2 MA1 MA0 MBA1 MBA0 JEDEC DDR DIMM Signals (Inputs) SDRAM 168-Pin DIMM A5 A4 A3 A2 A1 A0 BA1 BA0 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 9-4 Freescale Semiconductor External Signal Descriptions 9.3.2 Detailed Signal Descriptions The following sections describe the DDR SDRAM controller input and output signals, the meaning of their different states, and relative timing information for assertion and negation. 9.3.2.1 Memory Interface Signals Table 9-3. Memory Interface Signals—Detailed Signal Descriptions Table 9-3 describes the DDR controller memory interface signals. Signal(s) MDQ[0:63] I/O Description I/O Data bus. Both input and output signals on the DDR memory controller. O As outputs for the bidirectional data bus, these signals operate as described below. State Asserted/Negated—Represent the value of data being driven by the DDR memory Meaning controller. Timing Assertion/Negation—Driven coincident with corresponding data strobes (MDQS) signal. High-impedance—No READ or WRITE command is in progress; data is not being driven by the memory controller or the DRAM. I As inputs for the bidirectional data bus, these signals operate as described below. State Asserted/Negated—Represents the state of data being driven by the external DDR Meaning SDRAMs. Timing Assertion/Negation—The DDR DIMM drives data during a READ transaction. High-impedance—No READ or WRITE command in progress; data is not being driven by the memory controller or the DRAM. MDQS[0:8] I/O Data strobes. Inputs with read data and as outputs with write data. O As outputs, the data strobes are driven by the DDR memory controller during a write transaction. The memory controller always drives these signals low unless a read has been issued and incoming data strobes are expected. This keeps the data strobes from floating high when there are no transactions on the DRAM interface. State Asserted/Negated—Driven high when positive capture data is transmitted/received and Meaning driven low when negative capture data is transmitted/received. Centered in the data “eye” for writes; coincident with the data eye for reads. Treated as a clock. Data is valid when signals toggle. See Table 9-25 for byte lane assignments. Timing Assertion/Negation—If a WRITE command is registered at clock edge n, data strobes at the DRAM assert centered in the data “eye” on clock edge n + 1. See the JEDEC DDR SDRAM specification for more information. I As inputs, the data strobes are driven by the external DDR SDRAMs during a read transaction. The data strobes are used by the memory controller to synchronize data latching. State Asserted/Negated—Driven high when positive capture data is transmitted/received and Meaning driven low when negative capture data is transmitted/received. Centered in the data “eye” for writes; coincident with the data eye for reads. Treated as a clock. Data is valid when signals toggle. See Table 9-25 for byte lane assignments. Timing Assertion/Negation—If a READ command is registered at clock edge n, and the latency is programmed in TIMING_CFG_1[CASLAT] to be m clocks, data strobes at the DRAM assert coincident with the data on clock edge n + m. See the JEDEC DDR SDRAM specification for more information. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 9-5 DDR Memory Controller Table 9-3. Memory Interface Signals—Detailed Signal Descriptions (continued) Signal(s) MECC[0:7] I/O Description I/O Error checking and correcting codes. Input and output signals for the DDR controller’s bidirectional ECC bus. MECC[0:5] function in both normal and debug modes. O As normal mode outputs the ECC signals represent the state of ECC driven by the DDR controller on writes. As debug mode outputs MECC[0:5] provide source ID and data-valid information. See Section 20.4.3.2, “Debug Information on ECC Pins,” for more details. State Asserted/Negated—Represents the state of ECC being driven by the DDR controller on Meaning writes. Timing Assertion/Negation—Same timing as MDQ High-impedance—Same timing as MDQ I As inputs, the ECC signals represent the state of ECC driven by the SDRAM devices on reads. State Asserted/Negated—Represents the state of ECC being driven by the DDR SDRAMs on Meaning reads. Timing Assertion/Negation—Same timing as MDQ High-impedance—Same timing as MDQ MA[0:14] O Address bus. Memory controller outputs for the address to the DRAM. MA[0:14] carry 15 of the address bits for the DDR memory interface corresponding to the row and column address bits. MA[0] is the lsb of the address output from the memory controller. State Asserted/Negated—Represents the address driven by the DDR memory controller. Meaning Contains different portions of the address depending on the memory size and the DRAM command being issued by the memory controller. SeeTable 9-27 for a complete description of the mapping of these signals. Timing Assertion/Negation—The address is always driven when the memory controller is active. It is valid when a transaction is driven to DRAM (when MCSn is negated). High-impedance—When the memory controller is idle. MBA[0:1] O Logical bank address. Outputs that drive the logical (or internal) bank address pins of the SDRAM. Each SDRAM supports four addressable logical sub-banks. Bit zero of the memory controller’s output bank address must be connected to bit zero of the SDRAM’s input bank address. MBA[0] is asserted during the mode register set command to specify the extended mode register. State Asserted/Negated—Selects the DDR SDRAM logical (or internal) bank to be activated Meaning during the row address phase and selects the SDRAM internal bank for the read or write operation during the column address phase of the memory access. Table 9-27 describes the mapping of these signals in all cases. Timing Assertion/Negation—Same timing as MAn High-impedance—Same timing as MA n MCAS O Column address strobe. Active-low SDRAM address multiplexing signal. MCAS is asserted for read or write transactions and for mode register set, refresh, and precharge commands. State Asserted—Indicates that a valid SDRAM column address is on the address bus for read Meaning and write transactions. See Table 9-28 for more information on the states required on MCAS for various other SDRAM commands. Negated—The column address is not guaranteed to be valid. Timing Assertion/Negation—Assertion and negation timing is directed by the values described in Section 9.4.1.3, “DDR SDRAM Timing Configuration 1 (TIMING_CFG_1).” High-impedance—MCAS is always driven unless the memory controller is idle. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 9-6 Freescale Semiconductor External Signal Descriptions Table 9-3. Memory Interface Signals—Detailed Signal Descriptions (continued) Signal(s) MRAS I/O O Description Row address strobe. Active-low SDRAM address multiplexing signal. Asserted for activate commands. In addition; used for mode register set commands and refresh commands. State Asserted—Indicates that a valid SDRAM row address is on the address bus for read and Meaning write transactions. See Table 9-28 for more information on the states required on MRAS for various other SDRAM commands. Negated—The row address is not guaranteed to be valid. Timing Assertion/Negation—Timing is directed by the values described in Section 9.4.1.3, “DDR SDRAM Timing Configuration 1 (TIMING_CFG_1).” High impedance—MRAS is always driven unless the memory controller is idle. MCS[0:3] O Chip select. Four chip selects supported by the memory controller. State Asserted—Selects a physical SDRAM bank to perform a memory operation as described Meaning in Section 9.4.1.1, “Chip Select Memory Bounds (CSn_BNDS),” and Section 9.4.1.2, “Chip Select Configuration (CSn_CONFIG).” The DDR controller asserts one of the MCS[0:3] signals to begin a memory cycle. Negated—Indicates no SDRAM action during the current cycle Timing Assertion/Negation—Asserted to signal any new transaction to the SDRAM. The transaction must adhere to the timing constraints set in TIMING_CFG_1. High impedance—Always driven unless the memory controller is disabled MWE O Write enable. Asserted when a write transaction is issued to the SDRAM. This is also used for mode registers set commands and precharge commands. State Asserted—Indicates a memory write operation. See Table 9-28 for more information on the Meaning states required on MWE for various other SDRAM commands. Negated—Indicates a memory read operation Timing Assertion/Negation—Similar timing as MRAS and MCAS. Used for write commands. High impedance—MWE is always driven unless the memory controller is idle. MDM[0:8] O DDR SDRAM data output mask. Masks unwanted bytes of data transferred during a burst write. They are needed to support sub-burst-size transactions (such as single-byte writes) on SDRAM where all I/O occurs in multi-byte bursts. MDM0 corresponds to the most significant byte (MSB); MDM7 corresponds to the LSB. MDM8 corresponds to the ECC byte. Table 9-25 shows byte lane encodings. State Asserted—Prevents writing to DDR SDRAM. Asserted when data is written to DRAM if the Meaning corresponding byte(s) should be masked for the write. Note that the MDMn signals are active-high for the DDR controller. MDMn is part of the DDR command encoding. Negated—Allows the corresponding byte to be read from or written to the SDRAM Timing Assertion/Negation—Same timing as MDQx as outputs High impedance—Always driven unless the memory controller is disabled MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 9-7 DDR Memory Controller 9.3.2.2 Clock Interface Signals Table 9-4. Clock Signals—Detailed Signal Descriptions Table 9-4 contains the detailed descriptions of the clock signals of the DDR controller. Signal(s) MCK[0:5], MCK[0:5] I/ O Description O DRAM clock outputs and their complements. See Section 9.5.4.1, “Clock Distribution.” State Asserted/Negated—The JEDEC DDR SDRAM specifications require true and complement clocks. A clock edge is seen by the SDRAM when the true and Meanin complement cross. g Timing Assertion/Negation—Should be synchronized at the SDRAM with the internal clocks of the MPC8540. This is done with the DLL. MCKE[0:1] O Clock enable. Two identical output signals (each hereafter referred to simply as MCKE) used as the clock enable to one or more SDRAMs. MCKE can be negated to stop clocking the DDR SDRAM. While this results in system power savings, the user is cautioned to disable SDRAM clocking only when there are no transactions on the interface. State Asserted—Clocking to the SDRAM is enabled. Meanin Negated—Clocking to the SDRAM is disabled and the SDRAM should ignore signal transitions on MCK or MCK. MCK/MCK are don’t cares while MCKE is negated. g Timing Assertion/Negation—Similar timing to MAn High impedance—Always driven MSYNC_IN I DRAM DLL synchronization input. The DLL uses this input from the feedback loop to synchronize SDRAM clocks with the internal clocks of the MPC8540. See Section 9.5.4.1, “Clock Distribution.” State Asserted/Negated—Toggles at the same frequency as the applied SDRAM clocks (MCKx), but is not in phase with MCK n Meanin g Timing Assertion/Negation—This is used for locking the DLL. MSYNC_OU T O DRAM DLL synchronization output. Output for the DLL feedback loop. See Section 9.5.4.1, “Clock Distribution.” State Asserted/Negated—See MSYNC_IN description Meanin g 9.3.2.3 Debug Signals The debug signals MSRCID[0:4] and MDVAL have no function in normal DDR controller operation. A detailed description of these signals can be found in Section 20.4.3, “DDR SDRAM Interface Debug.” MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 9-8 Freescale Semiconductor Memory Map/Register Definition 9.4 Memory Map/Register Definition Table 9-5 shows the register memory map for the DDR memory controller. Undefined 4-byte address spaces within offset 0x000–0xFFF are reserved. Table 9-5. DDR Memory Controller Memory Map Offset Register Access R/W Reset 0x0000_0000 Section/Page 9.4.1.1/9-10 0x0_2000 CS0_BNDS—Chip select 0 memory bounds 0x0_2008 CS1_BNDS—Chip select 1 memory bounds 0x0_2010 CS2_BNDS—Chip select 2 memory bounds 0x0_2018 CS3_BNDS—Chip select 3 memory bounds 0x0_2080 CS0_CONFIG—Chip select 0 configuration 0x0_2084 CS1_CONFIG—Chip select 1 configuration 0x0_2088 CS2_CONFIG—Chip select 2 configuration 0x0_208C CS3_CONFIG—Chip select 3 configuration 0x0_2108 TIMING_CFG_1—DDR SDRAM timing configuration 1 0x0_210C TIMING_CFG_2—DDR SDRAM timing configuration 2 0x0_2110 0x0_2118 DDR_SDRAM_CFG—DDR SDRAM control configuration DDR_SDRAM_MODE—DDR SDRAM mode configuration R/W 0x0000_0000 9.4.1.2/9-10 R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W Special R/W R/W R/W R/W R/W 0x0000_0000 0x0000_0000 0x0200_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 9.4.1.3/9-11 9.4.1.4/9-13 9.4.1.5/9-14 9.4.1.6/9-16 9.4.1.7/9-16 9.4.1.8/9-17 9.4.1.9/9-18 9.4.1.10/9-18 9.4.1.11/9-19 9.4.1.12/9-20 9.4.1.13/9-20 9.4.1.14/9-21 9.4.1.15/9-21 9.4.1.16/9-22 9.4.1.17/9-23 9.4.1.18/9-24 9.4.1.19/9-25 0x0_2124 DDR_SDRAM_INTERVAL—DDR SDRAM interval configuration 0x0_2E00 DATA_ERR_INJECT_HI—Memory data path error injection mask high 0x0_2E04 DATA_ERR_INJECT_LO—Memory data path error injection mask low 0x0_2E08 ECC_ERR_INJECT—Memory data path error injection mask ECC 0x0_2E20 CAPTURE_DATA_HI—Memory data path read capture high 0x0_2E24 CAPTURE_DATA_LO—Memory data path read capture low 0x0_2E28 CAPTURE_ECC—Memory data path read capture ECC 0x0_2E40 ERR_DETECT—Memory error detect 0x0_2E44 ERR_DISABLE—Memory error disable 0x0_2E48 ERR_INT_EN—Memory error interrupt enable 0x0_2E4C CAPTURE_ATTRIBUTES—Memory error attributes capture 0x0_2E50 CAPTURE_ADDRESS—Memory error address capture 0x0_2E58 ERR_SBE—Single-Bit ECC memory error management 9.4.1 Register Descriptions This section describes the DDR memory controller registers. Shading indicates reserved fields that should not be written. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 9-9 DDR Memory Controller 9.4.1.1 Chip Select Memory Bounds (CSn_BNDS) The chip select bounds registers (CSn_BNDS) shown in Figure 9-2 define the starting and ending address of the memory space that corresponds to the individual chip selects. Note that the size specified in CSn_BNDS should equal the size of physical DRAM. Also, note that EAn must be greater than or equal to SAn. If the high-order 8 bits of an address are greater than or equal to SAn, and they are less than or equal to EAn, then chip select n will be used. 0 7 8 15 16 23 24 31 R0 0 0 0 0 0 0 0 SAn W Reset Offset 00000000 EAn 0000_0000_0000_0000_0000_0000_0000_0000 0x0_2000, 0x0_2008, 0x0_2010, 0x0_2018 Figure 9-2. Chip Select Bounds Registers (CSn_BNDS) Table 9-6 describes the CSn_BNDS register fields. Table 9-6. CSn_BNDS Field Descriptions Bits 0–7 8–15 16–23 24–31 Name — SAn — EAn Reserved Starting address for chip select (bank) n. This value is compared against the 8 msbs of the address. Reserved Ending address for chip select (bank) n. This value is compared against the 8 msbs of the address. Description 9.4.1.2 Chip Select Configuration (CSn_CONFIG) The chip select configuration (CSn_CONFIG) registers shown in Figure 9-3 enable the DDR chip selects and set the number of row and column bits used for each chip select. These registers should be loaded with the correct number of row and column bits for each SDRAM. Because CSn_CONFIG[ROW_BITS_CS_n,COL_BITS_CS_n] establish address multiplexing, the user should take great care to set these values correctly. 0 1 7 8 9 20 21 23 24 28 29 31 R CS_n _EN W Reset Offset 0000000 AP_n_EN 00000000000 0 ROW_BITS_CS_n 0 000 0 COL_BITS_CS_n 0000_0000_0000_0000_0000_0000_0000_0000 0x0_2080, 0x0_2084, 0x0_2088, 0x0_208C Figure 9-3. Chip Select Configuration Register (CSn_CONFIG) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 9-10 Freescale Semiconductor Memory Map/Register Definition Table 9-7 describes the CSn_CONFIG register fields. Table 9-7. CSn_CONFIG Field Descriptions Bits 0 Name CS_n_EN Description Chip select n enable 0 Chip select n is not active. 1 Chip select n is active and assumes the state set in CSn_BNDS. Reserved Chip select n auto precharge enable 0 Chip select n is auto precharged only if global auto precharge mode is enabled (DDR_SDRAM_INTERVAL[BSTOPRE] = 0). 1 Chip select n always issues an auto precharge for read and write transactions. Reserved Number of row bits for SDRAM on chip select n 000 12 row bits 001 13 row bits 010 14 row bits 011–111 Reserved Reserved Number of column bits for SDRAM on chip select n 000 8 column bits 001 9 column bits 010 10 column bits 011 11 column bits 011–111 Reserved 1–7 8 — AP_n_EN 9–20 21–23 — ROW_BITS_CS_n 24–28 29–31 — COL_BITS_CS_n 9.4.1.3 DDR SDRAM Timing Configuration 1 (TIMING_CFG_1) DDR SDRAM timing configuration register 1, shown in Figure 9-4, sets the number of clock cycles between various SDRAM control commands. 0 1 3 4 7 8 9 11 12 13 15 16 19 20 21 23 24 25 27 28 29 31 R0 PRETOACT ACTTOPRE W Reset Offset 0 ACTTORW 0 CASLAT REFREC 0 WRREC 0 ACTTOACT 0 WRTORD 0000_0000_0000_0000_0000_0000_0000_0000 0x0_2108 Figure 9-4. DDR SDRAM Timing Configuration Register 1 (TIMING_CFG_1) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 9-11 DDR Memory Controller Table 9-8 describes TIMING_CFG_1 fields. Table 9-8. TIMING_CFG_1 Field Descriptions Bits 0 1–3 Name — Reserved, should be cleared. Description PRETOAC Precharge-to-activate interval (trp ). Determines the number of clock cycles from a precharge T command until an activate or refresh command is allowed. 100 4 clocks 000 Reserved 101 5 clocks 001 1 clock 110 6 clocks 010 2 clocks 111 7 clocks 011 3 clocks ACTTOPR Activate to precharge interval (tr as ). Determines the number of clock cycles from an activate E command until a precharge command is allowed. 0010 2 clocks 0000 Reserved … 0001 1 clock 1111 15 clocks 0010 2 clocks — Reserved, should be cleared. 4–7 8 9–11 ACTTORW Activate to read/write interval for SDRAM (trc d ). Controls the number of clock cycles from an activate command until a read or write command is allowed. 100 4 clocks 000 Reserved 101 5 clocks 001 1 clock 110 6 clocks 010 2 clocks 111 7 clocks 011 3 clocks — CASLAT Reserved, should be cleared. MCAS latency from READ command. Number of clock cycles between registration of a READ command by the SDRAM and the availability of the first output data. If a READ command is registered at clock edge n and the latency is m clocks, data is available nominally coincident with clock edge n + m. This value must be programmed at initialization as described in Section 9.4.1.6, “DDR SDRAM Mode Configuration (DDR_SDRAM_MODE).”) 100 2.5 clocks 000 Reserved 101 3 clocks 001 1 clock 110 3.5 clocks 010 1.5 clocks 111 4 clocks 011 2 clocks Refresh recovery time (t rfc ). Controls the number of clock cycles from a refresh command until an activate command is allowed. Refresh recovery time is equal to eight plus the REFREC value. 0011 11 clocks 0000 Reserved … 0001 9 clocks 1111 23 clocks 0010 10 clocks Reserved, should be cleared. Last data to precharge minimum interval (twr ). Determines the number of clock cycles from the last data associated with a write command until a precharge command is allowed. 100 4 clocks 000 0 clock 101 5 clocks 001 1 clock 110 6 clocks 010 2 clocks 111 7 clocks 011 3 clocks Reserved, should be cleared. 12 13–15 16–19 REFREC 20 21–23 — WRREC 24 — MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 9-12 Freescale Semiconductor Memory Map/Register Definition Table 9-8. TIMING_CFG_1 Field Descriptions (continued) Bits 25–27 Name Description ACTTOAC Activate-to-activate interval (t rrd ). Number of clock cycles from an activate command until another T activate command is allowed for a different logical bank in the same physical bank (chip select). 011 3 clocks 000 Reserved 100 4 clocks 001 1 clock 101–111 Reserved 010 2 clocks — WRTORD Reserved, should be cleared. Last write data pair to read command issue interval (twtr ). Number of clock cycles between the last write data pair and the subsequent read command to the same physical bank. 100 4 clocks 000 Reserved 101 5 clocks 001 1 clock 110 6 clocks 010 2 clocks 111 7 clocks 011 3 clocks 28 29–31 9.4.1.4 DDR SDRAM Timing Configuration 2 (TIMING_CFG_2) DDR SDRAM timing configuration 2, shown in Figure 9-5, sets the clock delay to data for writes. 0 3 4 7 8 11 12 13 18 19 21 22 31 R0000 CPO W Reset Offset 0000 ACSM 000000 WR_DATA_DELAY 0000000000 0000_0000_0000_0000_0000_0000_0000_0000 0x0_210C Figure 9-5. DDR SDRAM Timing Configuration Register 2 (TIMING_CFG_2) Table 9-9 describes the TIMING_CFG_2 fields. Table 9-9. TIMING_CFG_2 Register Field Descriptions Bits 0–3 4–7 Name — CPO Reserved MCAS-to-preamble override. Defines the number of DRAM cycles between when a read is issued and when the corresponding DQS preamble is valid for the memory controller. 0110 CASLAT + 5/2 0000 Default. MCAS to preamble is 0111 CASLAT + 3 defined as CASLAT + 1 1000 CASLAT + 7/2 0001 CASLAT 1001 CASLAT + 4 0010 CASLAT + 1/2 1010 CASLAT + 9/2 0011 CASLAT + 1 1011 CASLAT + 5 0100 CASLAT + 3/2 1100–1111 Reserved 0101 CASLAT + 2 Reserved Description 8–11 — MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 9-13 DDR Memory Controller Table 9-9. TIMING_CFG_2 Register Field Descriptions (continued) Bits 12 Name ACSM Description Address and control shift mode 0 The DRAM address and control buses are output in the default mode. 1 The DRAM address and control buses are delayed by 1/2 DRAM cycle before being driven onto the pins. Reserved 13–1 8 19–2 1 — WR_DATA_DELA Write command to write data strobe timing adjustment. Controls the amount of delay Y applied to the data and data strobes for writes. 011 6/8 clock delay 000 0 clock delay 100 1 clock delay 001 2/8 clock delay (recommended) 101–111 Reserved 010 4/8 clock delay — Reserved 22–3 1 9.4.1.5 DDR SDRAM Control Configuration (DDR_SDRAM_CFG) The DDR SDRAM control configuration register, shown in Figure 9-6, enables the interface logic and specifies certain operating features such as self refreshing, error checking and correcting, registered DIMMS, and dynamic power management. 0 1 2 3 4 5 6 7 8 9 10 11 13 14 15 R MEM_EN W Reset 16 17 00 SREN ECC_EN RD_EN SDRAM_TYPE 00 DYN_PWR 000 NCAP 0 0000_0010_0000_0000 31 0 2T_EN 0 0 00 0 0 00 0 000 0 0 0000_0000_0000_0000 Offset 0x0_2110 Figure 9-6. DDR SDRAM Control Configuration Register (DDR_SDRAM_CFG) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 9-14 Freescale Semiconductor Memory Map/Register Definition Table 9-10 describes the DDR_SDRAM_CFG fields. Table 9-10. DDR_SDRAM_CFG Field Descriptions Bits 0 Name MEM_EN Description DDR SDRAM interface logic enable 0 SDRAM interface logic is disabled 1 SDRAM interface logic is enabled. Must not be set until all other memory configuration parameters have been appropriately configured by initialization code. Self refresh enable (during sleep) 0 SDRAM self refresh is disabled during sleep or soft-stop. Whenever self-refresh is disabled, the system is responsible for preserving the integrity of SDRAM during sleep or soft-stop. 1 SDRAM self refresh is enabled during sleep or soft-stop ECC enable. Note that uncorrectable read errors may cause the assertion of core_fault_in, which causes the core to generate a machine check interrupt unless it is disabled (by clearing HID1[RFXE]). If RFXE is zero and this error occurs, ERR_DISABLE[MBED] must be zero and ERR_INT_EN[MBEE] and ECC_EN must be one to ensure an interrupt is generated. See Section 6.10.2, “Hardware Implementation-Dependent Register 1 (HID1).” 0 No ECC errors are reported. No ECC interrupts are generated. 1 ECC is enabled. Registered DIMM enable. Specifies the type of DIMM used in the system. 0 Indicates unbuffered DIMMs. 1 Indicates registered DIMMs. Reserved 1 SREN 2 ECC_EN 3 RD_EN 4–5 6–7 — SDRAM_TYPE Type of SDRAM device to be used 00–01 Reserved 10 DDR SDRAM 11 Reserved — DYN_PWR Reserved Dynamic power management mode 0 Dynamic power management mode is disabled. 1 Dynamic power management mode is enabled. If there is no ongoing memory activity, the SDRAM CKE signal is negated. Reserved Non-concurrent auto precharge 0 SDRAMs in system support concurrent auto precharge. 1 SDRAMs in system do not support concurrent auto precharge. Reserved 2T timing enable 0 1T timing is used. The SDRAM command/address are held for only 1 cycle on the SDRAM bus. 1 2T timing is enabled. The SDRAM command/address are held for 2 full cycles on the SDRAM bus for every SDRAM transaction. However, the chip select is only held for the second cycle. Reserved 8–9 10 11–13 14 — NCAP 15 16 — 2T_EN 17–31 — MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 9-15 DDR Memory Controller 9.4.1.6 DDR SDRAM Mode Configuration (DDR_SDRAM_MODE) The DDR SDRAM mode configuration register, shown in Figure 9-7, sets the values loaded into the DDR’s mode registers. 0 15 16 31 R ESDMODE W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x0_2118 SDMODE Figure 9-7. DDR SDRAM Mode Configuration Register (DDR_SDRAM_MODE) Table 9-11 describes the DDR_SDRAM_MODE fields. Table 9-11. DDR_SDRAM_MODE Field Descriptions Bits 0–15 Name Description ESDMOD Extended SDRAM mode. Specifies the initial value loaded into the DDR SDRAM extended mode E[0:15] register. The range and meaning of legal values is specified by the DDR SDRAM manufacturer. The value of ESDMODE[1:15] is driven onto MA[14:0] during the EXTENDED MODE REGISTER SET operation of the intialization sequence. The lsb of ESDMODE (ESDMODE[15]) is driven onto MA[0] and ESDMODE[1] is driven onto MA[14]. SDMODE SDRAM mode. Specifies the initial value loaded into the DDR SDRAM mode register. The range [0:15] of legal values of legal values is specified by the DDR SDRAM manufacturer. The value of SDMODE[1:15] is driven onto MA[14:0] during the MODE REGISTER SET operation of the intialization sequence. The lsb of SDMODE (SDMODE[15]) is driven onto MA[0] and SDMODE[1] is driven onto MA[14]. Because the memory controller forces SDMODE[3–8] to certain values depending upon the state of the initialization sequence (for resetting the SDRAM’s DLL) the corresponding bits of this field is ignored by the memory controller. 16–3 1 9.4.1.7 DDR SDRAM Interval Configuration (DDR_SDRAM_INTERVAL) The DDR SDRAM interval configuration register, shown in Figure 9-8, sets the number of DRAM clock cycles between bank refreshes issued to the DDR SDRAMs. In addition, the number of DRAM cycles that a page is maintained after it is accessed is provided here. 0 1 2 15 16 17 18 31 R0 0 REFINT W Reset Offset 00 BSTOPRE 0000_0000_0000_0000_0000_0000_0000_0000 0x0_2124 Figure 9-8. DDR SDRAM Interval Configuration Register (DDR_SDRAM_INTERVAL) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 9-16 Freescale Semiconductor Memory Map/Register Definition Table 9-12 describes the DDR_SDRAM_INTERVAL fields. Table 9-12. DDR_SDRAM_INTERVAL Field Descriptions Bits 0–1 2–15 Name — REFINT Reserved Refresh interval. Represents the number of memory bus clock cycles between refresh cycles. One row is refreshed in each DDR SDRAM physical bank during each refresh cycle. The value for REFINT depends on the specific SDRAMs used and the interface clock frequency. Note that REFINT must be set to a non-zero value in order for the DDR to enter sleep mode. See Section 18.5.1.5.3, “Sleep Mode,” for additional details. Reserved Description 16–17 18–31 — BSTOPRE Precharge interval. Sets the duration (in memory bus clocks) that a page is retained after a DDR SDRAM access. If BSTOPRE is zero, the DDR memory controller uses auto precharge read and write commands rather than operating in page mode. This is called global auto precharge mode. 9.4.1.8 Memory Data Path Error Injection Mask High (DATA_ERR_INJECT_HI) The memory data path error injection mask high register is shown in Figure 9-9. 0 31 R EIMH W Reset Offset 0000_0000_0000_0000 0x0_2E00 Figure 9-9. Memory Data Path Error Injection Mask High Register (DATA_ERR_INJECT_HI) Table 9-13 describes the DATA_ERR_INJECT_HI fields. Table 9-13. DATA_ERR_INJECT_HI Field Descriptions Bits Nam e Description 0–31 EIMH Error injection mask high data path. Used to test ECC by forcing errors on the high word of the data path. Setting a bit causes the corresponding data path bit to be inverted on memory bus writes. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 9-17 DDR Memory Controller 9.4.1.9 Memory Data Path Error Injection Mask Low (DATA_ERR_INJECT_LO) The memory data path error injection mask low register is shown in Figure 9-10. 0 31 R EIML W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x0_2E04 Figure 9-10. Memory Data Path Error Injection Mask Low Register (DATA_ERR_INJECT_LO) Table 9-14 describes the DATA_ERR_INJECT_LO fields. Table 9-14. DATA_ERR_INJECT_LO Field Descriptions Bits 0–31 Name Description EIML Error injection mask low data path. Used to test ECC by forcing errors on the low word of the data path. Setting a bit causes the corresponding data path bit to be inverted on memory bus writes. 9.4.1.10 Memory Data Path Error Injection Mask ECC (ECC_ERR_INJECT) The memory data path error injection mask ECC register, shown in Figure 9-11, sets the ECC mask, enables errors to be written to ECC memory, and allows the ECC byte to mirror the most significant data byte. 0 21 22 23 24 31 R0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 EMB EIEN W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x0_2E08 EEIM Figure 9-11. Memory Data Path Error Injection Mask ECC Register (ECC_ERR_INJECT) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 9-18 Freescale Semiconductor Memory Map/Register Definition Table 9-15 describes the ECC_ERR_INJECT fields. Table 9-15. ECC_ERR_INJECT Field Descriptions Bits 0–21 22 Name — EMB Reserved ECC mirror byte 0 Mirror byte functionality disabled 1 Mirror the most significant data path byte onto the ECC byte. Description 23 EIEN Error injection enable 0 Error injection disabled 1 Error injection enabled. This applies to the data mask bits, the ECC mask bits, and the ECC mirror bit. 24–31 EEIM ECC error injection mask. Setting a mask bit causes the corresponding ECC bit to be inverted on memory bus writes. 9.4.1.11 Memory Data Path Read Capture High (CAPTURE_DATA_HI) The memory data path read capture high register, shown in Figure 9-12, stores the high word of the read data path during error capture. 0 31 R ECHD W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x0_2E20 Figure 9-12. Memory Data Path Read Capture High Register (CAPTURE_DATA_HI) Table 9-16 describes the CAPTURE_DATA_HI fields. Table 9-16. CAPTURE_DATA_HI Field Descriptions Bits 0–31 Name Description ECHD Error capture high data path. Captures the high word of the data path when errors are detected. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 9-19 DDR Memory Controller 9.4.1.12 Memory Data Path Read Capture Low (CAPTURE_DATA_LO) The memory data path read capture low register, shown in Figure 9-13, stores the low word of the read data path during error capture. 0 31 R ECLD W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x0_2E24 Figure 9-13. Memory Data Path Read Capture Low Register (CAPTURE_DATA_LO) Table 9-17 describes the CAPTURE_DATA_LO fields. Table 9-17. CAPTURE_DATA_LO Field Descriptions Bits 0–31 Name Description ECLD Error capture low data path. Captures the low word of the data path when errors are detected. 9.4.1.13 Memory Data Path Read Capture ECC (CAPTURE_ECC) The memory data path read capture ECC register, shown in Figure 9-14, stores the ECC syndrome bits that were on the data bus when an error was detected. 0 23 24 31 R0 W Reset Offset 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 ECE 0000_0000_0000_0000_0000_0000_0000_0000 0x0_2E28 Figure 9-14. Memory Data Path Read Capture ECC Register (CAPTURE_ECC) Table 9-18 describes the CAPTURE_ECC fields. Table 9-18. CAPTURE_ECC Field Descriptions Bits 0–23 24–31 Name — ECE Reserved Error capture ECC. Captures the ECC bits on the data path whenever errors are detected. Description MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 9-20 Freescale Semiconductor Memory Map/Register Definition 9.4.1.14 Memory Error Detect (ERR_DETECT) The memory error detect register, shown in Figure 9-15, stores the detection bits for multiple memory errors, single- and multiple-bit ECC errors, and memory select errors. It is a read/write register. A bit can be cleared by writing a high value to the bit. System software can determine the type of memory error by examining the contents of this register. If an error is disabled with ERR_DISABLE, the corresponding error is never detected or captured in ERR_DETECT. 0 1 27 28 29 30 31 R MME W Reset Offset 000000000000000000000000000 MBE SBE 0000_0000_0000_0000_0000_0000_0000_0000 0x0_2E40 0 MSE Figure 9-15. Memory Error Detect Register (ERR_DETECT) Table 9-19 describes the ERR_DETECT fields. Table 9-19. ERR_DETECT Field Descriptions Bits 0 Name Description MME Multiple memory errors. This bit is cleared by software writing a 1. 0 Multiple memory errors of the same type were not detected. 1 Multiple memory errors of the same type were detected. — MBE Reserved Multiple-bit error. This bit is cleared by software writing a 1. 0 A multiple-bit error has not been detected. 1 A multiple-bit error has been detected. Single-bit ECC error. This bit is cleared by software writing a 1. 0 The number of single-bit ECC errors detected has not crossed the threshold set in ERR_SBE[SBET]. 1 The number of single-bit ECC errors detected crossed the threshold set in ERR_SBE[SBET]. Reserved Memory select error. This bit is cleared by software writing a 1. 0 A memory select error has not been detected. 1 A memory select error has been detected. 1–27 28 29 SBE 30 31 — MSE 9.4.1.15 Memory Error Disable (ERR_DISABLE) The memory error disable register, shown in Figure 9-16, allows selective disabling of the DDR controller’s error detection circuitry. Disabled errors are not detected or reported. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 9-21 DDR Memory Controller 0 27 28 29 30 31 R0000000000000000000000000000 MBED SBED W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x0_2E44 0 MSED Figure 9-16. Memory Error Disable Register (ERR_DISABLE) Table 9-20 describes the ERR_DISABLE fields. Table 9-20. ERR_DISABLE Field Descriptions Bits 0–27 28 Name — Reserved Description MBED Multiple-bit ECC error disable 0 Multiple-bit ECC errors are detected if DDR_SDRAM_CFG[ECC_EN] is set. They are reported if ERR_INT_EN[MBEE] is set. Note that uncorrectable read errors cause the assertion of core_fault_in, which causes the core to generate a machine check interrupt, unless it is disabled (by clearing HID1[RFXE]). If RFXE is zero and this error occurs, MBED must be zero and ERR_INT_EN[MBEE] and ECC_EN must be one to ensure that an interrupt is generated. 1 Multiple-bit ECC errors are not detected or reported. SBED Single-bit ECC error disable 0 Single-bit ECC errors are enabled. 1 Single-bit ECC errors are disabled. — Reserved 29 30 31 MSED Memory select error disable 0 Memory select errors are enabled. 1 Memory select errors are disabled. 9.4.1.16 Memory Error Interrupt Enable (ERR_INT_EN) The memory error interrupt enable register, shown in Figure 9-17, enables ECC interrupts or memory select error interrupts. When an enabled interrupt condition occurs, the internal int signal is asserted to the programmable interrupt controller (PIC). 0 27 28 29 30 31 R0000000000000000000000000000 MBEE SBEE W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x0_2E48 0 MSEE Figure 9-17. Memory Error Interrupt Enable Register (ERR_INT_EN) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 9-22 Freescale Semiconductor Memory Map/Register Definition Table 9-21 describes the ERR_INT_EN fields. Table 9-21. ERR_INT_EN Field Descriptions Bits 0–27 28 Name — Reserved Description MBEE Multiple-bit ECC error interrupt enable. Note that uncorrectable read errors may cause the assertion of core_fault_in, which causes the core to generate a machine check interrupt, unless it is disabled (by clearing HID1[RFXE]). If RFXE is zero and this error occurs, ERR_DISABLE[MBED] must be zero and MBEE and DDR_SDRAM_CFG[ECC_EN] must be set to ensure that an interrupt is generated. For more information, see Section 6.10.2, “Hardware Implementation-Dependent Register 1 (HID1).” 0 Multiple-bit ECC errors cannot generate interrupts. 1 Multiple-bit ECC errors generate interrupts. SBEE Single-bit ECC error interrupt enable 0 Single-bit ECC errors cannot generate interrupts. 1 Single-bit ECC errors generate interrupts. — Reserved 29 30 31 MSEE Memory select error interrupt enable 0 Memory select errors do not cause interrupts. 1 Memory select errors generate interrupts. 9.4.1.17 Memory Error Attributes Capture (CAPTURE_ATTRIBUTES) The memory error attributes capture register, shown in Figure 9-18, sets attributes for errors including type, size, source, and others. 0 1 3 4 5 7 8 10 11 15 16 17 18 19 20 30 31 R0 BNUM W Reset Offset 0 TSIZ 000 TSRC 00 TTYP 00000000000 VLD 0000_0000_0000_0000_0000_0000_0000_0000 0x0_2E4C Figure 9-18. Memory Error Attributes Capture Register (CAPTURE_ATTRIBUTES) Table 9-22 describes the CAPTURE_ATTRIBUTES fields. Table 9-22. CAPTURE_ATTRIBUTES Field Descriptions Bits 0 1–3 4 5–7 8–10 Name — Reserved Description BNUM Data beat number. Captures the data beat number for the detected error. Relevant only for ECC errors. — TSIZ — Reserved Transaction size for the error. Captures the transaction size in double words. Reserved MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 9-23 DDR Memory Controller Table 9-22. CAPTURE_ATTRIBUTES Field Descriptions (continued) Bits 11–15 Name TSRC Transaction source for the error 00000 PCI 00001–00011 Reserved 00100 Local bus 00101–00111 Reserved 01000 Configuration space 01001 Reserved 01010 Boot sequencer 01011 Reserved 01100 RapidIO 01101–01111 Reserved 10000 Processor (instruction) 10001 Processor (data) 10010–10011 Reserved Reserved Transaction type for the error 00 Reserved 01 Write 10 Read 11 Read-modify-write Reserved Valid. Set as soon as valid information is captured in the error capture registers. Description 10100 10101 10110 10111 11000 11001 11010 11011 11100 11101 11110 11111 Reserved DMA Reserved SAP TSEC1 TSEC2 FEC Reserved RapidIO message units RapidIO doorbell units RapidIO port—write units Reserved 16–17 18–19 — TTYP 20–30 31 — VLD 9.4.1.18 Memory Error Address Capture (CAPTURE_ADDRESS) The memory error address capture register, shown in Figure 9-19, holds the 32 lsbs of a transaction when a DDR ECC error is detected. 0 31 R CADDR W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x0_2E50 Figure 9-19. Memory Error Address Capture Register (CAPTURE_ADDRESS) Table 9-23 describes the CAPTURE_ADDRESS fields. Table 9-23. CAPTURE_ADDRESS Field Descriptions Bits 0–31 Name Description CADDR Captured address. Captures the 32 lsbs of the transaction address when an error is detected. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 9-24 Freescale Semiconductor Functional Description 9.4.1.19 Single-Bit ECC Memory Error Management (ERR_SBE) The single-bit ECC memory error management register, shown in Figure 9-20, stores the threshold value for reporting single-bit errors and the number of single-bit errors counted since the last error report. When the counter field reaches the threshold, it wraps back to the reset value (0). If necessary, software must clear the counter after it has managed the error. 0 7 8 15 16 23 24 31 R00 0 0 0 0 0 0 SBET W Reset Offset 00000000 SBEC 0000_0000_0000_0000_0000_0000_0000_0000 0x0_2E58 Figure 9-20. Single-Bit ECC Memory Error Management Register (ERR_SBE) Table 9-24 describes the ERR_SBE fields. Table 9-24. ERR_SBE Field Descriptions Bits 0–7 8–15 16–23 Name — Reserved Description SBET Single-bit error threshold. Establishes the number of single-bit errors that must be detected before an error condition is reported. — Reserved 24–31 SBEC Single-bit error counter. Indicates the number of single-bit errors detected and corrected since the last error report. If single-bit error reporting is enabled, an error is reported and an interrupt is generated when this value equals SBET. SBEC is automatically cleared when the threshold value is reached. 9.5 Functional Description The DDR SDRAM controller controls processor and I/O interactions with system memory. It provides support for JEDEC-compliant DDR SDRAMs (first generation dual data rate). The memory system allows a wide range of memory devices to be mapped to any arbitrary chip select. However, registered DIMMs cannot be mixed with unbuffered DIMMs. Figure 9-21 is a high-level block diagram of the DDR memory controller. Requests are received from the internal mastering device and the address is decoded to generate the physical bank, logical bank, row, and column addresses. The transaction is then loaded into the input staging queue with the decoded information. The lower two entries of the input queue are compared with values in the row open table to determine if the address maps to an open page. If the address from either entry does not map to an open page, an activate command is issued for the entry that did not hit an open page, with the lowest entry having priority over the next lowest. Commands are always issued from the lowest input queue entry. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 9-25 DDR Memory Controller Request from Master Address from Master Address Decode Input Staging Queue Address Control DDR SDRAM Memory Array: MA[0:14] MBAA[0:1] Debug Signals: MSRCID[0:4] MDVAL DDR SDRAM Memory Control: MCS[0:3] MCAS MRAS MWE MDM[0:8] MCKE[0:1] Data Strobes: MDQS[0:8] Physical Bank, Logical Bank, Row ECM Row Open Table Row Open SDRAM Control EN To Error Management Error Signals ECC DQ Delay Chain Data from SDRAM DQ POS FIFO NEG FIFO RMW ECC DQ EN Data Signals: MDQ[0:63] MECC[0:7] Data from Master DLL Clocks: MCK[0:5] MCK[0:5] MSYNC_OUT MSYNC_IN Figure 9-21. DDR Memory Controller Block Diagram The DDR SDRAM interface supports as many as four physical banks of 64/72 bit-wide memory. A bank size up to 1 Gbyte is supported, providing up to a maximum of 4 Gbytes of DDR main memory. However, 4 Gbytes would span the entire 32-bit address space, and some space must be reserved for boot ROM, configuration registers, and other important addressable locations. Programmable parameters allow for a variety of memory organizations and timings. Optional error checking and correcting (ECC) protection is provided for the DDR SDRAM data bus. Using ECC, the DDR memory controller detects and corrects all single-bit errors within the 64-bit data bus, detects all double-bit errors within the 64-bit data bus, and detects all errors within a nibble. The controller allows as many as 16 pages to be open simultaneously. The amount of time (in clock cycles) the pages remain open is programmable with DDR_SDRAM_INTERVAL[BSTOPRE]. Read and write accesses to the DDR SDRAM are burst oriented; accesses start at a selected location and continue for a programmed number of higher locations (2, 4, or 8) in a programmed sequence, (sequential or interleaved). Accesses to closed pages start with the registration of an MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 9-26 Freescale Semiconductor Functional Description ACTIVE command followed by a READ or WRITE. (Accessing open pages does not require an ACTIVE command.) The address bits registered coincident with the activate command specifies the logical bank and row to be accessed. The address coincident with the READ or WRITE command specify the logical bank and starting column for the burst access. The data interface is source synchronous, meaning whatever sources the data also provides a clocking signal to synchronize data reception. These bidirectional data strobes (MDQS[0:8]) are inputs to the controller during reads, and outputs during writes. The DDR SDRAM specification requires the data strobe signals to be centered within the data tenure during writes and to be offset by the controller to the center of the data tenure during reads. This is implemented in the controller with delay chains for the data strobe signals during reads and a delay chain on the data multiplexer select during writes. When ECC is enabled, one clock cycle is added to the read path to check ECC and correct single-bit errors. ECC generation does not add a cycle to the write path. Figure 9-22 shows how the on-chip DLL can be used with the memory controller. Clock PCB Trace Length “L” DDR Controller DDR SDRAM Internal Clock DLL CK CK DLL DLL_Sync_Out DLL timing loop DLL_Sync_In L -2 Figure 9-22. Controller DLL Timing Loop MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 9-27 DDR Memory Controller Figure 9-23 shows an example DDR SDRAM configuration with four physical banks. Data Bus Data-out Registers SDRAM Data-in Registers MUX, MASK, Read Data Latch ADDR COMMAND: MCS, MRAS, MCAS, MWE DQM BA1,BA0 CKE, MCK, MCK Control Logical Bank 0 Logical Bank 1 Logical Bank 2 Logical Bank 3 Figure 9-23. Typical Dual Data Rate SDRAM Internal Organization Figure 9-24 shows some typical signal connections. 64M × 1 byte 512 Mbit ADDR ‘SUB’ BANK ADDR MRAS MCAS MWE MCS DM CKE MCK MCK CK Write Enable Command Bus 13 2 A[12:0] BA[1:0] DQS DATA STROBE DQ[7:0] 8 DATA Figure 9-24. Typical DDR SDRAM Interface Signals Figure 9-25 shows an example DDR SDRAM configuration with four physical banks each comprised of nine 8M × 8 DDR modules for a total of 256 Mbytes of system memory. One of the nine modules is used for the memory’s ECC checking function. Certain address and control lines may require buffering. Analysis of the MPC8540 device’s AC timing specifications, desired memory operating frequency, capacitive loads, and board routing loads, can assist the system designer in deciding signal buffering requirements. The MPC8540 DDR memory controller drives 15 address pins, but in this example the DDR SDRAM devices use only 12 bits. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 9-28 Freescale Semiconductor Functional Description MDQS[0:8] MDQ[0:63] MECC[0:7] MA[0:14] MBA[0:1] MRAS MCAS MWE MCKE MCK[0:5] MCK[0:5] MCS[0:3] MCS0 8M×8 SDRAM 2Mx8 A[11:0] SDRAM 2Mx8 BA[1:0] SDRAM A[0-11] RAS2 Mx8 SDRAM BA[0-1] MDQS A(11-0) CAS 2Mx8 SDRAM RAS BA(1-0) A(11-0) WE 2Mx8 SDRAM CAS DQ[7:0] RAS BA(1-0) A[0-11] CKE CAS 2Mx8 SDRAM RAS DQ[0-7] BA[0-1] A[0-11] CKCAS DQ(7-0) CKE RAS2 Mx8 SDRAM CS BA[0-1] CLK A[0-11] CKE CAS DQ(7-0) DM RAS CS BA[0-1] CLK A[0-11] CKE CAS DQ[0-7] DM RAS CS BA[0-1] CLK CKE CAS DQ[0-7] RAS DMCKE CS CLK CAS DQ[0-7] DMCKE CS CLK DQ[0-7] DMCKE CS CLK DM CS CLK DM CS DM Memory Data Bus and Strobes To all SDRAM Devices in Common MCS1 8M×8 SDRAM 2Mx8 A[11:0] SDRAM 2 Mx8 BA[1:0] SDRAM A[0-11] RAS2Mx8 SDRAM BA[0-1] MDQS A(11-0) CAS 2Mx8 SDRAM RAS BA(1-0) A(11-0) WE 2Mx8 SDRAM CAS DQ[7:0] RAS BA(1-0) A[0-11] CKE CAS 2 Mx8 SDRAM RAS DQ[0-7] BA[0-1] CK A[0-11] CKE CAS 2Mx8 SDRAM RAS DQ(7-0) CS BA[0-1] CLK A[0-11] CKE CAS DQ(7-0) RAS DMCKE CS BA[0-1] CLK A[0-11] CAS DQ[0-7] RAS DMCKE CS BA[0-1] CLK CAS DQ[0-7] RAS DMCKE CS CLK CAS DQ[0-7] DM CS CLK CKE DQ[0-7] DM CS CLK CKE DM CS CLK DM CS DM MDQS[0] MDQ[0:7] MDQ[8:15] MDQ[16:23] MDQ[24:31] MDQ[32:39] MDQ[40:47] MDQ[48:55] MDQ[56:63] MDQS[7] MDQS[0] MDQ[0:7] MDQ[8:15] MDQ[16:23] MDQ[24:31] MDQ[32:39] MDQ[40:47] MDQ[48:55] MDQ[56:31] MDQS[7] DDR Controller 0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7 MDM[0:8] 8M×8 SDRAM A[11:0] BA[1:0] RAS MDQS CAS DQ[7:0] WE CKE CK CS 8 DM 8M×8 SDRAM A[11:0] BA[1:0] RAS MDQS CAS DQ[7:0] WE CKE CK CS 8 DM MDQS[8] ECC[0:7] MDQS[8] ECC[0:7] Bank 0 Bank 1 Banks 2–3 8M × 72 (64 Mbytes) 8M ×72 (64 Mbytes) 1. All signals are connected in common (in parallel) except for MCS[0:3],MCK[0:5], MDM[0:8], and the data bus signals. 2. Each of the MCS[0:3] signals correspond with a separate physical bank of memory. 3. Buffering may be needed if large memory arrays are used. 4. MCK[0:5] may be apportioned among all memory devices. Complementary bus is not shown. Figure 9-25. Example 256-Mbyte DDR SDRAM Configuration with ECC MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 9-29 DDR Memory Controller Section 9.5.12, “Error Management,” explains how the DDR memory controller handles errors. 9.5.1 DDR SDRAM Interface Operation The DDR memory controller supports many different DDR SDRAM configurations. SDRAMs with different sizes can be used in the same system. Fourteen multiplexed address signals and two logical bank select signals support device densities from 64 Mbit to 1 Gbit. Four chip select (CS) signals support up to two DIMMs of memory. The DDR SDRAM physical banks can be built from standard memory modules or directly-attached memory devices. The data path to individual physical banks is 64 bits wide, 72 bits with ECC. The MPC8540 DDR memory controller supports physical bank sizes from 32 Mbytes to 1 Gbytes. The physical banks can be constructed using × 8, or ×16 memory devices. The memory technologies supported are 64, 128, 256, and 512 Mbit, and 1 Gbit. Nine data qualifier (DQM) signals provide byte selection for memory accesses. NOTE An 8-bit DDR SDRAM device has a DQM signal and 8 data signals (DQ[0:7]). A 16-bit DDR SDRAM device has 2 DQM signals associated with specific halves of the 16 data signals (DQ[0:7] and DQ[8:15]). When ECC is enabled, all memory accesses are performed on double-word boundaries (that is, all DQM signals are set simultaneously). However, when ECC is disabled, the memory system uses the DQM signals for byte lane selection. Table 9-25 shows the DDR memory controller’s relationships between data byte lane 0–7, MDM[0:7], MDQS[0:7], and MDQ[0:63]. Table 9-25. Byte Lane to Data Relationship Data Byte Lane Data Bus Mask 0 (MSB) 1 2 3 4 5 6 7 (LSB) MDM[0] MDM[1] MDM[2] MDM[3] MDM[4] MDM[5] MDM[6] MDM[7] Data Bus Strobe MDQS[0] MDQS[1] MDQS[2] MDQS[3] MDQS[4] MDQS[5] MDQS[6] MDQS[7] Data Bus 64-Bit Mode MDQ[0:7] MDQ[8:15] MDQ[16:23] MDQ[24:31] MDQ[32:39] MDQ[40:47] MDQ[48:55] MDQ[56:63] MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 9-30 Freescale Semiconductor Functional Description 9.5.1.1 Supported DDR SDRAM Organizations Although the DDR memory controller multiplexes row and column address bits onto 14 memory address signals and 2 logical bank select signals, individual physical banks may be implemented with memory devices requiring fewer than 28 address bits. Each physical bank may be individually configured to provide from 12 to 14 row address bits, plus 2 logical bank-select bits and from 8–11 column address bits. Table 9-26 describes DDR SDRAM device configurations supported by the DDR memory controller. NOTE DDR SDRAM is limited to 27 total address bits. Table 9-26. Supported DDR SDRAM Device Configurations SDRAM Device 64 Mbits 64 Mbits 128 Mbits 128 Mbits 256 Mbits 256 Mbits 512 Mbits 512 Mbits 1 Gbits 1 Gbits Device Configuration Row × Column Bits 8 Mbits × 8 4 Mbits × 16 16 Mbits × 8 8 Mbits × 16 32 Mbits × 8 16 Mbits × 16 64 Mbits × 8 32 Mbits × 16 128 Mbits × 8 64 Mbits × 16 12 × 9 12 × 8 12 × 10 12 × 9 13 × 10 13 × 9 13 × 11 13 × 10 14 × 11 14 × 10 64-Bit Bank Size 64 Mbytes 32 Mbytes 128 Mbytes 64 Mbytes 256 Mbytes 128 Mbytes 512 Mbytes 256 Mbytes 1 Gbytes 512 Mbytes Four Banks of Memory 256 Mbytes 128 Mbytes 512 Mbytes 256 Mbytes 1 Gbytes 512 Mbytes 2 Gbytes 1 Gbytes 4 Gbytes 2 Gbytes If a transaction request is issued to the DDR memory controller and the address does not lie within any of the programmed address ranges for an enabled chip select, a memory select error is flagged. Errors are described in detail in Section 9.5.12, “Error Management.” By using a memory-polling algorithm at power-on reset or by querying the JEDEC serial presence detect capability of memory modules, system firmware uses the memory-boundary registers to configure the DDR memory controller to map the size of each bank in memory. The memory controller uses its bank map to assert the appropriate MCSn signal for memory accesses according to the provided bank starting and ending addresses. The memory banks are not required to be mapped to a contiguous address space. 9.5.2 DDR SDRAM Address Multiplexing Table 9-27 shows the address bit encodings for each DDR SDRAM configuration. The address presented at the memory controller signals MA[14:0] use MA[14] as the msb and MA[0] as the lsb. Also, MA[10] is used as the auto precharge bit for reads and writes, so the column address can never use MA[10]. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 9-31 DDR Memory Controller Table 9-27. DDR SDRAM Address Multiplexing Row x Col 14 MRAS × MBA 11 MCAS 14 MRAS × MBA 10 MCAS 13 MRAS × MBA 11 MCAS 13 MRAS × MBA 10 MCAS 13 MRAS × MBA 9 MCAS 12 MRAS × MBA 10 MCAS 12 MRAS × MBA 9 MCAS 12 MRAS × MBA 8 MCAS msb 0 Address from Core Master lsb 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29–31 13 12 11 10 9 8 7 6 5 4 3 2 1 0 1 0 11 9 13 12 11 10 9 8 7 6 5 4 3 2 1 0 1 0 9 12 11 10 9 8 7 6 5 4 3 2 1 0 1 0 11 9 12 11 10 9 8 7 6 5 4 3 2 1 0 1 0 9 12 11 10 9 8 7 6 5 4 3 2 1 0 1 0 8 11 10 9 8 7 6 5 4 3 2 1 0 1 0 9 11 10 9 8 7 6 5 4 3 2 1 0 1 0 8 11 10 9 8 7 6 5 4 3 2 1 0 1 0 7 6 5 4 3 2 1 0 7 6 5 4 3 2 1 0 8 7 6 5 4 3 2 1 0 7 6 5 4 3 2 1 0 8 7 6 5 4 3 2 1 0 8 7 6 5 4 3 2 1 0 8 7 6 5 4 3 2 1 0 8 7 6 5 4 3 2 1 0 9.5.3 JEDEC Standard DDR SDRAM Interface Commands All read or write accesses to DDR SDRAM are performed by the DDR memory controller using JEDEC standard DDR SDRAM interface commands. The SDRAM device samples command and address inputs on rising edges of the memory clock; data is sampled using both the rising and falling edges of DQS. Data read from the DDR SDRAM is also sampled on both edges of DQS. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 9-32 Freescale Semiconductor Functional Description The following DDR SDRAM interface commands (summarized in Table 9-28) are provided by the DDR controller. All actions for these commands are described from the perspective of the SDRAM device. • Row activate—Latches row address and initiates memory read of that row. Row data is latched in SDRAM sense amplifiers and must be restored by a precharge command before another row activate occurs. Precharge—Restores data from the sense amplifiers to the appropriate row. Also initializes the sense amplifiers in preparation for reading another row in the memory array, (performing another activate command). Precharge must occur after read or write, if the row address changes on the next open page mode access. Read—Latches column address and transfers data from the selected sense amplifier to the output buffer as determined by the column address. During each succeeding clock edge, additional data is driven without additional read commands. The amount of data transferred is determined by the burst size which defaults to 4. Write—Latches column address and transfers data from the data pins to the selected sense amplifier as determined by the column address. During each succeeding clock edge, additional data is transferred to the sense amplifiers from the data pins without additional write commands. The amount of data transferred is determined by the burst size, which is set to four by the DDR memory controller. Refresh (similar to MCAS before MRAS)—Causes a row to be read in all logical banks (JEDEC SDRAM) as determined by the refresh, row address counter. This refresh row address counter is internal to the SDRAM. After being read, the row is automatically rewritten in the memory array. All logical banks must be in a precharged state before executing a refresh. Mode register set (for configuration)—Allows setting of DDR SDRAM options. These options are: MCAS latency, burst type, and burst length. MCAS latency may be chosen as provided by the preferred SDRAM (some SDRAMs provide MCAS latency {1,2,3}, some provide MCAS latency {1,2,3,4}, and so on). Burst type is always sequential. Although some SDRAMs provide burst lengths of 1, 2, 4, 8, and page size, this memory controller supports a burst length of 4. The mode register set command is performed by the DDR memory controller during system initialization. Parameters such as mode register data, MCAS latency, burst length, and burst type, are set by software in DDR_SDRAM_MODE[SDMODE] and transferred to the SDRAM array by the DDR memory controller after DDR_SDRAM_CFG[MEM_EN] is set. Self refresh (for long periods of standby)—Used when the device is in standby for very long periods of time. Automatically generates internal refresh cycles to keep the data in all memory banks refreshed. Before execution of this command, all logical banks are in a precharged state. • • • • • • MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 9-33 DDR Memory Controller Table 9-28. SDRAM Command Table Operation Activate Precharge select logical bank Precharge all logical banks Read Read with auto precharge Write Write with auto precharge Mode register set Auto refresh Self refresh CKE CKE MCS MRAS MCAS MWE Prev. Current H H H H H H H H H H H H H H H H H H H L L L L L L L L L L L L L L H H H H L L L H H H L L L L L L L H L L H H L L L H H MBA Logical bank select Logical bank select X Logical bank select Logical bank select Logical bank select Logical bank select Opcode X X MA10 Row L H L H L H Opcode X X MA Row X X Column Column Column Column Opcode and mode X X 9.5.4 SDRAM Interface Timing The DDR memory controller supports four-beat bursts to SDRAM. For single-beat reads, the DDR memory controller performs a four-beat burst read, but ignores the last three beats. Single-beat writes are performed by masking the last three beats of the four-beat burst using the data mask MDM[0:8]. If ECC is disabled, writes smaller than double words are performed by appropriately activating the data mask. If ECC is enabled, the controller performs a read-modify write. NOTE If a second read or write is pending, reads shorter than four beats are not terminated early even if some data is irrelevant. To accommodate available memory technologies across a wide spectrum of operating frequencies, the DDR memory controller allows the setting of the intervals defined in Table 9-29 with granularity of one memory clock cycle, except for CASLAT, which can be programmed with 1/2 clock granularity. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 9-34 Freescale Semiconductor Functional Description Table 9-29. DDR SDRAM Interface Timing Intervals Timing Intervals ACTTOACT ACTOPRE ACTORW BSTOPRE Definition The number of clock cycles from a bank-activate command until another bank-activate command within a physical bank. This interval is listed in the AC specifications of the SDRAM. The number of clock cycles from an activate command until a precharge command is allowed. This interval is listed in the AC specifications of the SDRAM. The number of clock cycles from an activate command until a read or write command is allowed. This interval is listed in the AC specifications of the SDRAM. The number of clock cycles to maintain a page open after an access. The page open duration counter is reloaded with BSTOPRE each time the page is accessed (including page hits). When the counter expires, the open page is closed with a SDRAM precharge bank command as soon as possible. READ latency. The number of clock cycles between the registration of a READ command by the SDRAM and the availability of the first piece of output data. If a READ command is registered at clock edge n, and the latency is m clocks, the data is available nominally coincident with clock edge n + m. The number of clock cycles from a precharge command until an activate or a refresh command is allowed. This interval is listed in the AC specifications of the SDRAM. Refresh interval. Represents the number of memory bus clock cycles between refresh cycles. One row is refreshed in each SDRAM bank during each refresh cycle. The value of REFINT depends on the specific SDRAMs used and the frequency of the interface. The number of clock cycles from the refresh command until an activate command is allowed. This can be calculated by referring to the AC specification of the SDRAM device. The AC specification indicates a maximum refresh to activate interval in nanoseconds. Provides different options for the timing between a write command and the write data strobe. This allows write data to be sent later than the nominal time to meet the SDRAM timing requirement between the registration of a write command and the reception of a data strobe associated with the write command. The specification dictates that the data strobe may not be received earlier than 75% of a cycle, or later than 125% of a cycle, from the registration of a write command. This parameter is not defined in the SDRAM specification. It is implementation-specific, defined for the DDR memory controller in TIMING_CFG_2. The number of clock cycles from the last beat of a write until a precharge command is allowed. This interval, write recovery time, is listed in the AC specifications of the SDRAM. Last write pair to read command. Controls the number of clock cycles from the last write data pair to the subsequent read command to the same bank. CASLAT PRETOACT REFINT REFREC WR_DATA_ DELAY WRREC WRTORD The value of the above parameters (in whole clock cycles) must be set by boot code at system start-up (in the TIMING_CFG_1 and TIMING_CFG_2 registers as described in Section 9.4.1.3, “DDR SDRAM Timing Configuration 1 (TIMING_CFG_1),” and Section 9.4.1.4, “DDR SDRAM Timing Configuration 2 (TIMING_CFG_2)”) and be kept in the DDR memory controller configuration register space. The following figures show SDRAM timing for various types of accesses. System software is responsible (at reset) for optimally configuring SDRAM timing parameters. The programmable timing parameters apply to both read and write timing configuration. The configuration process MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 9-35 DDR Memory Controller must be completed and the DDR SDRAM initialized before any accesses to SDRAM are attempted. Figure 9-26 through Figure 9-28 show DDR SDRAM timing for various types of accesses; see Figure 9-26 for a back-to-back burst read operation, Figure 9-27 for a single-beat write operation, and Figure 9-28 for a burst-write operation. Note that all signal transitions occur on the rising edge of the memory bus clock and that single-beat read operations are identical to burst-reads. 0 1 2 3 4 5 6 7 8 9 10 11 12 SDRAM clock MCS ACTTORW MRAS MCAS MA[13:0] ROW COL COL MWE MDQ[0:63] MDQS CASLAT D0 D1 D2 D3 D0 D1 D2 D3 Figure 9-26. DDR SDRAM Burst Read Timing—ACTTORW = 3, MCAS Latency = 2 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 9-36 Freescale Semiconductor Functional Description 0 1 2 3 4 5 6 7 8 9 10 11 12 SDRAM clock MCS ACTTORW PRECHARGE MRAS MCAS WRREC PRETOACT ROW MA[0:13] ROW COL A10=0 MWE MDQ[0:63] MDQS MDM[0:7] 00 FF FF FF D0 D1 D2 D3 Figure 9-27. DDR SDRAM Single-Beat (Double-Word) Write Timing—ACTTORW = 3 0 1 2 3 4 5 6 7 8 9 10 11 12 SDRAM Clock MCS0 MCS1 ACTTORW MRAS MCAS MA[0:13] ROW ROW’ COL COL COL’ COL’ MWE MDQ[0:63] MDQS MDM[0:7] 00 D0 D1 D2 D3 D0 D1 D2 D3 D0 D1 D2 D3 D0 D1 D2 D3 Figure 9-28. DDR SDRAM Burst Write Timing—ACTTORW = 4 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 9-37 DDR Memory Controller 9.5.4.1 • • Clock Distribution • • • If running with many devices, zero-delay PLL clock buffers, JEDEC-JESD82 standard, should be used. These buffers were designed for DDR applications. The DLL timing loop PCB route should mimic the MCK route as closely as possible, although clock edges can be intentionally advanced or retarded by modifying this loop. With all DDR clocks as nearly identical as possible, the single DLL timing loop can be finely adjusted to advance or retard all of them. A 72 bit × 64 Mbytes DDR bank has 9-byte-wide DDR chips, resulting in 18 DDR chips in a two-bank system. In this case, each MCK/MCK signal pair should drive exactly three devices. PCB traces for DDR clock signals should be short, all on the same layer, and of equal length and loading as the DLL timing loop. DDR SDRAM manufacturers provide detailed information on PCB layout and termination issues. A[13:0], BA[1:0], MRAS , MCAS , MWE, CKE CS[0] CS[1] DDR DQ[0:7], DQS[0], DM[0] DQ[8:15], DQS[1], DM[1] DQ[16:23], DQS[2], DM[2] MCK[3], MCK[3] MCK[0], MCK[0] MCK[1], MCK[1] DQ[24:31], DQS[3], DM[3] DQ[32:39], DQS[4], DM[4] DQ[40:47], DQS[5], DM[5] MCK[4], MCK[4] MCK[2], MCK[2] DQ[48:55], DQS[6], DM[6] DQ[56:63], DQS[7], DM[7] ECC[0:7], DQS[8], DM[8] MCK[5], MCK[5] Figure 9-29. DDR SDRAM Clock Distribution Example 9.5.5 DDR SDRAM Mode-Set Command Timing The DDR memory controller transfers the extended mode and base mode register data DDR_SDRAM_MODE[ESDMODE,SDMODE] to the SDRAM array by issuing two mode-set commands separated by two SDRAM clock periods. Figure 9-30 shows the timing of the mode-set command. The first transfer corresponds to the ESDMODE code; the second corresponds to SDMODE. Following commands must wait two SDRAM cycles. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 9-38 Freescale Semiconductor Functional Description 0 1 2 3 4 5 6 7 8 9 10 11 12 SDRAM Clock MCS MRAS MCAS MA[0:14] MBA[0:1] MWE MDQ[0:63] MDQS Code 0x1 Code 0x0 Figure 9-30. DDR SDRAM Mode-Set Command Timing 9.5.6 DDR SDRAM Registered DIMM Mode To reduce loading, registered DIMMs latch the DDR SDRAM control signals internally before using them to access the array. Setting DDR_SDRAM_CFG[RD_EN] compensates for this delay on the DIMMs’ control bus by delaying the data and data mask writes (on SDRAM buses) by an extra SDRAM clock cycle. Enabling registered DIMM mode does not affect bus timing for DDR reads. However to compensate for latch delay on the registered DIMMs’ control signals, the programmed SDRAM read latency value (TIMING_CFG_1[CASLAT]) must be one greater than the value needed for non-registered DIMMs. Figure 9-31 shows the registered DDR SDRAM DIMM back-to-back burst write timing. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 9-39 DDR Memory Controller 0 1 2 3 4 5 6 7 8 9 10 11 12 SDRAM Clock MCS ACTTORW MRAS MCAS MA[13:0] ROW COL COL MWE MDQ[0:63] MDQS MDM[0:7] 00 D0 D1 D2 D3 D0 D1 D2 D3 Figure 9-31. Registered DDR SDRAM DIMM Burst Write Timing 9.5.7 DDR SDRAM Write Timing Adjustments The DDR memory controller facilitates system design flexibility by providing a write timing adjustment parameter, write data delay, (TIMING_CFG_2[WR_DATA_DELAY]) for data and DQS. The DDR SDRAM specification requires DQS be received no sooner than 75% of an SDRAM clock period—and no later than 125% of a clock period—from the capturing clock edge of the command/address at the SDRAM. The WR_DATA_DELAY parameter may be used to meet this timing requirement for a variety of system configurations, ranging from a system with one DIMM to a fully populated system with two DIMMS. TIMING_CFG_2[WR_DATA_DELAY] specifies how much to delay the launching of DQS and data from the first clock edge occurring one SDRAM clock cycle after the command is launched. The delay increment step sizes are in 1/4th SDRAM clock periods starting with the default value of 1/4th period delay. Figure 9-32 shows the use of the WR_DATA_DELAY parameter. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 9-40 Freescale Semiconductor Functional Description 0 1 2 3 4 5 6 7 8 9 10 11 12 SDRAM Clock MCS ACTTORW MRAS MCAS MA[13:0] ROW COL COL MWE MDQ[0:63] MDQS MDM[0:7] MDQ[0:63] MDQS MDM[0:7] 00 00 D0 D1 D2 D3 D0 D1 D2 D3 D0 D1 D2 D3 D0 D1 D2 D3 1/4 Delay 1/2 Delay Figure 9-32. Write Timing Adjustments Example 9.5.8 DDR SDRAM Refresh The DDR memory controller supports auto-refresh and self-refresh. Auto refresh is used during normal operation and is controlled by the DDR_SDRAM_INTERVAL[REFINT] value; self-refresh is used only when the DDR memory controller is set to enter a sleep power management state or a soft-stop state. The REFINT value, which represents the number of memory bus clock cycles between refresh cycles, must allow for possible outstanding transactions to complete before a refresh request is sent to the memory after the REFINT value is reached. If a memory transaction is in progress when the refresh interval is reached, the refresh cycle waits for the transaction to complete. In the worst case, the refresh cycle must wait the number of bus clock cycles required by the longest programmed access. To ensure that the latency caused by a memory transaction does not violate the device refresh period, it is recommended that the programmed value of REFINT be less than that required by the SDRAM. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 9-41 DDR Memory Controller When a refresh cycle is required, the DDR memory controller does the following: 1. Completes all current memory requests 2. Closes all open pages with a PRECHARGE-ALL command to each DDR SDRAM bank with an open page (as indicated by the row open table) 3. Issues an auto-refresh command to each DDR SDRAM bank (as identified by its chip select) to refresh one row in each logical bank of the selected physical bank The auto-refresh commands are staggered across the four possible banks to reduce the system’s instantaneous power requirements. Three sets of auto refresh commands must be issued on consecutive cycles when the memory is fully populated with two DIMMs. The initial PRECHARGE-ALL commands are also staggered in three groups for convenience. It is important to note that when entering self-refresh mode, only one refresh command is issued simultaneously to all physical banks. CKE is negated at this time, so it would not be possible to stagger two more refresh commands. For this entire refresh sequence, no cycle optimization occurs for the usual case where fewer than four banks are installed. After the refresh sequence completes, any pending memory request is initiated after an inactive period specified by TIMING_CFG_1 [REFREC]. 9.5.8.1 DDR SDRAM Refresh Timing Refresh timing for the DDR SDRAM is controlled by the programmable timing parameter TIMING_CFG_1 [REFREC], which specifies the number of memory bus clock cycles from the refresh command until a logical bank activate command is allowed. The DDR memory controller implements bank staggering for refreshes, as shown in Figure 9-33 (TIMING_CFG_1 [REFREC] = 10 in this example). 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 SDRAM Clock CKE MCS(0,3) MCS(1) MCS(2) MRAS MCAS MA[13:0] ROW REFREC Figure 9-33. DDR SDRAM Bank-Staggered Auto-Refresh Timing MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 9-42 Freescale Semiconductor Functional Description System software is responsible for optimal configuration of TIMING_CFG_1 [REFREC] at reset. Configuration must be completed before DDR SDRAM accesses are attempted. 9.5.8.2 DDR SDRAM Refresh and Power-Saving Modes In full-on mode, the DDR memory controller supplies the normal auto refresh to SDRAM. In sleep mode, the DDR memory controller can be configured to take advantage of self-refreshing SDRAMs or to provide no refresh support. Self-refresh support is enabled with the SREN memory control parameter of the DDR_SDRAM_CFG register. Table 9-30 summarizes the refresh types available in each power-saving mode. Table 9-30. DDR SDRAM Power-Saving Modes Refresh Configuration Power Saving Mode Sleep Refresh Type Self None SREN 1 0 Note that in the absence of refresh support, system software must preserve DDR SDRAM data (such as by copying the data to disk) before entering the power-saving mode. All open pages are precharged before self refresh mode is entered. The dynamic power-saving mode uses the CKE DDR SDRAM pin to dynamically power down when there is no system memory activity. The CKE pin is negated when both of the following conditions are met: • • No memory refreshes are scheduled. No memory accesses are scheduled. CKE is reasserted when a new access or refresh is scheduled or the dynamic power mode is disabled. This mode is controlled with DDR_SDRAM_CFG[DYN_PWR]. Dynamic power management mode offers tight control of the memory system’s power consumption by trading power for performance through the use of DDR_SDRAM_INTERVAL[BSTOPRE]. Powering up the DDR SDRAM when a new memory reference is scheduled causes a one-clock access latency penalty, as shown in Figure 9-34. Mem Bus Clock CKE COMMAND NOP NOP ACT Figure 9-34. DDR SDRAM Power-Down Mode MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 9-43 DDR Memory Controller 9.5.8.2.1 Self-Refresh in Sleep Mode The entry and exit timing for self-refreshing SDRAMs is shown in Figure 9-35 and Figure 9-36. 0 1 2 3 4 5 6 7 8 9 10 11 12 SDRAM Clock CKE MCS MRAS MCAS MA[13:0] MWE MDQ[0:63] MDQS (High Impedance) Figure 9-35. DDR SDRAM Self-Refresh Entry Timing 0 1 2 3 4 5 6 7 202 203 204 205 206 SDRAM Clock MCS MRAS MCAS MA[13:0] MWE MDQ[0:63] MDQS (High Impedance) 200 Cycles Figure 9-36. DDR SDRAM Self-Refresh Exit Timing MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 9-44 Freescale Semiconductor Functional Description 9.5.9 DDR Data Beat Ordering Transfers to and from memory are always performed in four-beat bursts (four beats = 32 bytes). For transfer sizes other than four beats, the data transfers are still operated as four-beat bursts. If ECC is enabled for a sub-doubleword write transaction, a full read-modify-write is performed to properly update ECC bits. If ECC is disabled then no read-modify-write is required for sub-doubleword writes, and the data masks (MDM[0:8]) are used to prevent writing unwanted data to SDRAM. The DDR memory controller also uses data masks to prevent all unintended full double words from writing to SDRAM. For example, if a write transaction is desired with a size of one double word (8 bytes), then the 2nd, 3rd, and 4th beats of data are not written to DRAM. Table 9-31 lists the data beat sequencing to and from the DDR SDRAM and the data queues for each of the possible transfer sizes with each of the possible starting double-word offsets. All underlined double-word offsets are valid for the transaction. Table 9-31. Memory Controller–Data Beat Ordering Transfer Size 1 double word Starting Double-Word Offset 0 1 2 3 0 1 2 0 1 0 1 2 3 Double-Word Sequence1 to/from DRAM and Queues 0 1 2 3 -1-2-3 -2-3-0 -3-0-1 -0-1-2 2 double words 0-1-2-3 1-2-3-0 2-3-0-1 0-1-2-3 1-2-3-0 0-1-2-3 1-2-3-0 2-3-0-1 3-0-1-2 3 double words 4 double words 1 All underlined double-word offsets are valid for the transaction. 9.5.10 Page Mode and Logical Bank Retention The DDR memory controller supports an open/closed page mode with an allowable open page for each logical bank of DRAM used. In closed page mode, the DDR memory controller uses the SDRAM AUTO PRECHARGE feature, which allows the controller to indicate that the page must be automatically closed by the DDR SDRAM after the READ or WRITE access. This is performed by using MA[10] of the address during the COMMAND phase of the access to enable AUTO PRECHARGE. AUTO PRECHARGE is non-persistent in that it is either enabled or disabled for each individual READ or WRITE command. It can however, be enabled or disabled separately for each chip select. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 9-45 DDR Memory Controller When the DDR memory controller operates in open page mode, it retains the currently active SDRAM page by not issuing a precharge command. The page remains opens until one of the following conditions occurs: • • • Refresh interval is met The user-programmable DDR_SDRAM_INTERVAL[BSTOPRE] value is exceeded. There is a logical bank row collision with another transaction that must be issued. Page mode can dramatically reduce access latencies for page hits. Depending on the memory system design and timing parameters, using page mode can save two to three clock cycles for subsequent burst accesses that hit in an active page. Also, better performance can be obtained by using more banks, especially in systems which use many different channels. Page mode is disabled by clearing DDR_SDRAM_INTERVAL[BSTOPRE] and CS_n_CONFIG[AP_n_EN] 9.5.11 Error Checking and Correcting (ECC) The DDR memory controller supports error checking and correcting (ECC) for the data path between the core master and system memory. The memory detects all double-bit errors, detects all multi-bit errors within a nibble, and corrects all single-bit errors. Other errors may be detected, but are not guaranteed to be corrected or detected. Multiple-bit errors are always reported when error reporting is enabled. When a single-bit error occurs, the single-bit error counter register is incremented, and its value compared to the single-bit error trigger register. An error is reported when these values are equal. The single-bit error registers can be programmed such that minor memory faults are corrected and ignored, but a catastrophic memory failure generates an interrupt. For writes that are smaller than 64 bits, the DDR memory controller performs a double-word read from system memory of the address for the write (checking for errors), and merges the write data with the data read from memory. Then, a new ECC code is generated for the merged double word. The data and ECC code is then written to memory. If a multi-bit error is detected on the read, the transaction completes the read-modify-write to keep the DDR memory controller from hanging. This read-modify-write operation is performed as an atomic transaction in the DDR controller. The write command is then issued 3–5 memory clocks after the completion of the read, depending on various system parameters. However, the corrupt data is masked on the write, so the original contents in SDRAM remain unchanged. The syndrome encodings for the ECC code are shown in Table 9-32 and Table 9-33. Table 9-32. DDR SDRAM ECC Syndrome Encoding Data Bit 0 1 2 Syndrome Bit 0 1 2 3 4 5 6 7 Data Bit 32 33 34 Syndrome Bit 0 1 2 3 4 5 6 7 • • • • • • • • • • • • • • • • • • MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 9-46 Freescale Semiconductor Functional Description Table 9-32. DDR SDRAM ECC Syndrome Encoding (continued) Data Bit 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 Syndrome Bit 0 1 2 3 4 5 6 7 Data Bit 35 36 37 38 39 Syndrome Bit 0 1 2 3 4 5 6 7 • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • 40 41 42 43 • • • • • • • • • • • • 44 45 46 47 48 49 50 51 • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • 52 53 54 • • • • 55 56 57 58 • • • • 59 60 61 62 63 • • MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 9-47 DDR Memory Controller Table 9-33. DDR SDRAM ECC Syndrome Encoding (Check Bits) Check Bit 0 1 2 3 4 5 6 7 Syndrome Bit 0 1 2 3 4 5 6 7 • • • • • • • • 9.5.12 Error Management The DDR memory controller detects three different kinds of errors: single-bit, multi-bit, and memory select errors. The following discussion assumes all the relevant error detection, correction, and reporting functions are enabled as described in Section 9.4.1.16, “Memory Error Interrupt Enable (ERR_INT_EN),” Section 9.4.1.15, “Memory Error Disable (ERR_DISABLE),” and Section 9.4.1.14, “Memory Error Detect (ERR_DETECT).” Single-bit errors are counted and reported based on the ERR_SBE value. When a single-bit error is detected, the DDR memory controller does the following: • • • • Corrects the data Increments the single-bit error counter ERR_SBE[SBEC] Generates an interrupt if the counter value ERR_SBE[SBEC] equals the programmable threshold ERR_SBE[SBET] Completes the transaction normally If a multi-bit error is detected for a read, the DDR memory controller logs the error and generates an interrupt (if enabled, as described in Section 9.4.1.15, “Memory Error Disable (ERR_DISABLE)”). The final error the DDR memory controller detects is a memory select error, which causes the DDR memory controller to log the error and generate an interrupt (if enabled, as described in Section 9.4.1.14, “Memory Error Detect (ERR_DETECT)”). This error is detected if the address from the memory request does not fall into any of the enabled, programmed chip select address ranges. Table 9-34 shows the errors with their descriptions. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 9-48 Freescale Semiconductor Initialization/Application Information Table 9-34. Memory Controller Errors Category Notification Error Single-bit ECC threshold Multi-bit ECC error Memory select error Descriptions The number of ECC errors has reached the threshold specified in the ERR_SBE. A multi-bit ECC error is detected during a read, or read-modify-write memory operation. Read, or write, address does not fall within the address range of any of the memory banks. Action The error is reported through an interrupt if enabled Detect Register The error control register only logs read versus write, not full type. Access error 9.6 Initialization/Application Information System software must configure the DDR memory controller, using a memory polling algorithm at system start-up, to correctly map the size of each bank in memory. Then, the DDR memory controller uses its bank map to assert the appropriate MCS[0:3] signal for memory accesses according to the provided bank depths. System software must also configure the DDR memory controller at system start-up to multiplex appropriately the row and column address bits for each bank. Refer to row-address configuration in Section 9.4.1.2, “Chip Select Configuration (CSn_CONFIG).” Address multiplexing occurs according to these configuration bits. At system reset, initialization software (bootcode) must set up the programmable parameters in the memory interface configuration registers (MICRs). See Section 9.4.1, “Register Descriptions,” for more detailed descriptions of the configuration registers. These parameters are shown in Table 9-35. Table 9-35. Memory Interface Configuration Register Initialization Parameters Name CSn_BNDS CSn_CONFIG Description Chip select memory bounds Chip select configuration Parameter SAn EAn CS_n_EN ROW_BITS_CS_n COL_BITS_CS_n PRETOACT ACTTOPRE ACTTORW CASLAT REFREC WRREC ACTTOACT WRTORD WR_DATA_DELAY Section/Pag e 9.4.1.1/9-10 9.4.1.2/9-10 TIMING_CFG_1 DDR SDRAM timing configuration 9.4.1.3/9-11 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 9-49 DDR Memory Controller Table 9-35. Memory Interface Configuration Register Initialization Parameters (continued) Name DDR_SDRAM_CFG Description DDR SDRAM control configuration Parameter SREN ECC_EN RD_EN SDRAM_TYPE DYN_PWR ESDMODE SDMODE REFINT BSTOPRE Section/Pag e 9.4.1.5/9-14 DDR_SDRAM_MODE DDR SDRAM mode configuration 9.4.1.6/9-16 9.4.1.7/9-16 DDR_SDRAM_INTERVAL DDR SDRAM interval configuration 9.6.1 DDR SDRAM Initialization Sequence After configuration of all parameters is complete, system software must set DDR_SDRAM_CFG[MEM_EN] to enable the memory interface. Note that 200 µs must elapse after the memory clocks are stable (that is, the DLL has locked or initialization is complete of all clock related configuration registers) before MEM_EN can be set, so a delay loop in the initialization code may be necessary if software is enabling the memory controller. After MEM_EN has been set, the DDR memory controller automatically performs the JEDEC-compliant initialization sequence to initialize memories according to the information in the SDMODE and ESDMODE fields of the DDR_SDRAM_MODE register. The initialization sequence is as follows: 1. 2. 3. 4. 5. 6. PRECHARGE ALL MODE REGISTER SET for extended mode register MODE REGISTER SET for mode register PRECHARGE ALL Two AUTO REFRESH commands MODE REGISTER SET for mode register with reset DLL bit deactivated Note that the BA0 and BA1 bits are automatically driven appropriately during the MODE REGISTER SET commands. After this automatic initialization is complete the memory array is ready for access and the memory controller begins processing memory transactions as they arrive. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 9-50 Freescale Semiconductor Chapter 10 Programmable Interrupt Controller This chapter describes the programmable interrupt controller (PIC) interrupt protocol, various types of interrupt sources controlled by the PIC unit, and the PIC registers with some programming guidelines. 10.1 Introduction A block diagram of portions of the MPC8540 showing the relationship of the various functional blocks and external signals to the PIC unit is shown in Figure 10-1. The PIC unit receives interrupt signals from the following sources: • • External interrupts. Triggered by signals external to the integrated device, namely IRQ[0:11] Internal interrupts. Triggered by signal internal to the integrated device, typically representing major blocks by from the L2 cache, the ECM, the DDR controller, the local bus controller (LBC), the 4-channel DMA controller, the PCI/PCI-X block, the RapidIO interface, the dual three-speed Ethernet controllers (TSECs), the 10/100 Ethernet controller, the DUART, the performance monitor, and the I2C controller. Interrupts generated from within the PIC itself. Names the messaging, global timer, and interprocessor interrupts defined by the OpenPIC specification To the core through the external interrupt signal, int. For interrupts signalled through the int signal, the PIC unit prioritizes and manages interrupts such that the highest priority interrupt is always recognized and taken, and that any lower priority interrupt that is deferred when a higher priority interrupt is taken, resumes execution as soon as all higher priority interrupts have been handled. To the core through the critical interrupt signal, cint To an external interrupt controller, through the IRQ_OUT signal • The PIC can be configured such that interrupts can be directed as follows: • • • MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 10-1 Programmable Interrupt Controller Table 10-1 shows which signalling mechanisms are available to each interrupt source. Table 10-1. Interrupt Source Signalling Options Source Internal External Message in cin IRQ_OU t t T √ √ √ √ √ √ x x √ √ √ x x Interprocessor √ Global timer √ Note that the PIC can be disabled, in which case internal interrupts are passed through IRQ_OUT. All other interrupts are disabled. 10.1.1 Overview The PIC is compliant with the OpenPIC architecture. The interrupt controller provides interrupt management, and is responsible for receiving hardware-generated interrupts from different sources (both internal and external), prioritizing them, and delivering them to the CPU for servicing. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 10-2 Freescale Semiconductor Introduction Internal Interrupts L2 cache e500 coherency model (ECM) DDR DRAM Local bus controller (LBC) DMA channels 0–3 PCI/PCI-X RapidIO inbound port write/ error interrupt RapidIO doorbell inbound interrupt RapidIO outbound message interrupt RapidIO inbound message interrupt TSEC 1/2 transmit interrupts TSEC 1/2 receive interrupts TSEC 1/2 receive/transmit error interrupts PIC e500 Core Timer Facility Watchdog Timer PIC-defined interrupts core_mcp core_ude core_fault core_tbint core_halt core_hreset_req core_hreset core_ckstop_out core_wrs_[0:1] int cint core_reset PIR[P0] SRESET MCP UDE CKSTP_IN CKSTP_OUT HRESET HRESET_REQ Global Utilities Wake up core/ service interrupt 10-3 Global Timer (4) Messaging (4) InterProcessor (4) Fast Ethernet controller (FEC) DUART I2C controller Performance monitor interrupt IRQ[0:8] DMA_DREQ3/IRQ9 DMA_DACK3/IRQ10 DMA_DDONE3/IRQ11 IRQ_OUT External Interrupts Figure 10-1. MPC8540 Interrupt Sources Block Diagram 10.1.2 Features • • • Programming model compliant with the OpenPIC architecture Support for 12 external and 23 internal interrupt sources. Serial interrupts are not supported. Four interprocessor interrupt channels MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor Programmable Interrupt Controller • • • • • • • • • Four 32-bit messaging interrupt channels Four global high resolution timers that can be clocked with the CCB (platform) clock or the RTC input. Fully-nested interrupt delivery Processor initialization control Programmable resetting of the PIC unit through the global configuration register 16 programmable interrupt priority levels Support for connection of external interrupt controller device such as an 8259 programmable interrupt controller In 8259 mode, it generates a local (internal) interrupt output signal, IRQ_OUT. Recovery from spurious interrupts 10.1.3 Interrupts to the Processor Core The int signal, which causes the external interrupt exception, is the main interrupt output from the PIC unit to the processor core. External, internal, and message interrupts can alternately be configured as critical interrupts (in the destination registers); these are reported to the core through the cint signal and do not use the PIC logic described in Section 10.4.1, “Flow of Interrupt Control.” The Book E architecture implemented by the e500 core defines a separate critical interrupt type with its own save and restore registers (CSRR0 and CSRR1) and return instruction (Return from Critical Interrupt, rfci). In addition to the external and critical interrupts that are generated by the PIC, other MPC8540 conditions (shown in Table 10-2) cause interrupts to the core (and wake up the core when it is in a low-power state). Table 10-2. Processor Interrupts Generated Outside the Core—Types and Sources Core Interrupt Type Signaled by (Input to Core) Sources PIC-Programmable Interrupts External interrupt Critical interrupt int cint Generated by the PIC, as described in Section 10.1.5, “Interrupt Sources.” Generated by the PIC, as described in Section 10.1.5, “Interrupt Sources.” Other Interrupts Generated Outside the Core M achine check core_mcp (MPC8540 causes) • MCP • SRESET • Assertion of core_mcp by global utilities block MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 10-4 Freescale Semiconductor Introduction Table 10-2. Processor Interrupts Generated Outside the Core—Types and Sources (continued) Core Interrupt Type Unconditional debug event Reset Signaled by (Input to Core) core_ude Sources UDE. Asserting UDE generates an unconditional debug exception type debug interrupt and sets a bit in the debug status register, DBSR[UDE], as described in Section 6.13.2, “Debug Status Register (DBSR).” • HRESET assertion (and negation) • core_hreset_req. Output from core—caused by writing to the core DBCR0[RST]. This condition is additionally qualified with MSR[DE] and DBCR0[IDM] bits. Note that assertion of this signal causes a hard reset of the core only. • core_hreset_req can also be caused by a second timer timeout condition as described in Section 6.6.1, “Timer Control Register (TCR).” • core_reset. Output from PIC. See Section 10.3.1.4, “Processor Initialization Register (PIR).” core_hreset The global utilities block monitors two additional interrupt conditions generated by the e500 core (core_tbint and core_fault_out signals), as shown in Figure 10-1. Assertion of either of these signals causes the processor to exit a low-power state. These cases are caused by core conditions, and after the global utilities logic wakes up the core, they are handled by the core as shown in Table 10-3. Table 10-3. e500 Core-Generated Interrupts that Cause a Wake-up Core Interrupt Type Fixed interval timer Decrementer Signaled by (Output from Core) core_tbint Sources The source of both of these interrupts is the time base facility within the e500 core. The MPC8540 monitors this core output signal and considers it a processor interrupt for the purposes of power management (causes the core to exit a low-power state). For more information about the interaction between core-generated signals and power management, see Chapter 18, “Global Utilities.” O ccurs when the L1 cache has a parity error on a snoop push operation, which can occur while the core is halted. The MPC8540 monitors this signal and considers it a processor interrupt for the purposes of power management (causes the core to exit a low-power state). Machine check core_fault_out 10.1.4 Modes of Operation Mixed or pass-through mode can be chosen by setting or clearing GCR[M] as described in Section 10.3.1.2, “Global Configuration Register (GCR).” 10.1.4.1 Mixed Mode (GCR[M] = 1) In mixed mode, the external and internal interrupts are delivered using the normal priority and delivery mechanisms detailed in Section 10.3.6.1, “External Interrupt Vector/Priority Registers (EIVPR0–EIVPR11),” through Section 10.3.6.4, “Internal Interrupt Destination Registers (IIDR0–IIDR31).” MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 10-5 Programmable Interrupt Controller 10.1.4.2 Pass-Through Mode (GCR[M] = 0) The PIC unit provides a mechanism to support alternate external interrupt controllers such as the PC/AT compatible 8259 interrupt controller architecture. After a hard reset, the PIC unit defaults to pass-through mode, in which active-high interrupts from external source IRQ0 are passed directly to the e500 core, as shown in Table 10-2, all other external interrupt signals are ignored. Thus, the interrupt signal from an external interrupt controller can be connected to IRQ0 and cause direct interrupts to the processor. The PIC does not perform a vector fetch from an 8259 interrupt controller. IRQ[1:11] (Disabled) PIC GCR[M] = 0 External PC/AT-Compatible 8259 Interrupt Controller IRQ0 int e500 Core MPC8540 IRQ_OUT MPC8540 Internal interrupts (see Table 10-4) (not including PIC-generated interrupts) Figure 10-2. Pass-Through Mode Example When pass-through mode is enabled, the internally-generated interrupts shown in Table 10-4 are not forwarded to the e500 core. Instead, the PIC passes the raw interrupts from the internal sources to IRQ_OUT. Note that in pass-through mode, interrupts generated by the PIC itself (global timers, interprocessor, and message register interrupts) cannot be used. If internal or PIC-generated interrupts must be reported internally to the processor, pass-through mode must be disabled. 10.1.5 Interrupt Sources Aside from the sources of machine check, unconditional debug event, and reset interrupts to the core described in Table 10-2, the PIC unit can receive 47 separate interrupts from five different sources as follows: • • • 12 external—Off-chip signals, IRQ[0:11] 23 internal—On-chip. Sources are L2, ECM, DDR, local bus controller, DMA, PCI/PCI-X, RapidIO, TSEC, FEC, DUART, performance monitor, and I2C 4 global timers—From inside the PIC MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 10-6 Freescale Semiconductor Introduction • • 4 inter-processor (IPI)—Intended for communication between different processor cores on the same device. Used only for self-interrupt in a single-core device such as the MPC8540 4 message registers—From inside the PIC. Triggered on register write, cleared on read. Used for inter-process communication 10.1.5.1 Interrupt Routing—Mixed Mode When the PIC receives an internal or external interrupt, its destination register is checked to determine if it should be routed off-chip to the external IRQ_OUT signal (if the incoming interrupt has its xIDR[EP] bit set). Alternatively, if the incoming interrupt has been configured in xIDR[CI] as a critical interrupt, the PIC completes its processing of the interrupt by asserting the cint input to the core, causing it to be serviced as a critical interrupt. As a third alternative (if neither xIDR[CI] or xIDR[EP] are set), the interrupt can be serviced as a normal external interrupt by the processor core (through the int signal). In this case, the interrupt is latched by the interrupt pending register (IPR) and the interrupt flow described in Section 10.4.1, “Flow of Interrupt Control,” is followed. 10.1.5.2 Internal Interrupt Sources Table 10-4 shows the assignments of the 23 internal interrupt sources for the MPC8540. Note that this list does not include the interrupts generated by the PIC unit. Table 10-4. Internal Interrupt Assignments Internal Interrupt Number 0 1 2 3 4 5 6 7 8 9 10 11 12 13 L2 cache ECM DDR DRAM LBC DMA channel 0 DMA channel 1 DMA channel 2 DMA channel 3 PCI/PCI-X RapidIO inbound port write/ error interrupt RapidIO doorbell inbound interrupt RapidIO outbound message interrupt RapidIO inbound message interrupt TSEC 1 transmit interrupt Interrupt Source Internal Interrupt Number 14 15–17 18 19 20 21–23 24 25 26 27 28 29 30 31 Interrupt Source TSEC 1 receive interrupt Reserved TSEC 1 receive/transmit error interrupt TSEC 2 transmit interrupt TSEC 2 receive interrupt Reserved TSEC 2 receive/transmit error interrupt Fast Ethernet controller (FEC) DUART I2C controller Performance monitor interrupt Unused Unused Unused MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 10-7 Programmable Interrupt Controller 10.2 External Signal Descriptions The following sections provide an overview and detailed descriptions of the PIC signals. 10.2.1 Signal Overview PIC interface signals are described in Table 10-5. There are 12 distinct external interrupt request input signals (IRQ[0:11]) and 1 interrupt request output signal (IRQ_OUT). As Table 10-5 shows, three IRQ inputs are multiplexed with DMA signals for DMA channel 3. Table 10-5. PIC Interface Signals Signal Name IRQ[0:8] IRQ9/DMA_DREQ31 IRQ10/DMA_DACK31 IRQ11/DMA_DDONE31 IRQ_OUT MCP UDE 1 I/O I I I or O I or O O I I External interrupts Description External interrupt/DMA channel 3 request External interrupt (input)/DMA channel 3 acknowledge (output) External interrupt (input)/DMA channel 3 done (output) Interrupt request out Processor machine check Unconditional debug event IRQ9–IRQ11 are multiplexed with DMA3 signals. These functions are mutually exclusive; the active function is specified in PMUXCR of the global utilities block as described in Section 18.4.1.10, “Alternate Function Signal Multiplex Control Register (PMUXCR).” 10.2.2 Detailed Signal Descriptions Table 10-6 provides detailed descriptions of the external PIC signals. Table 10-6. Interrupt Signals—Detailed Signal Descriptions Signal IRQ[0:11] I/ O I Description Interrupt request 0–11. The polarity and sense of each of these signals is programmable. All of these inputs can be driven completely asynchronously. State Asserted—When an external interrupt signal is asserted (according to the programmed Meaning polarity), the priority is checked by the PIC unit, and the interrupt is conditionally passed to the processor. In pass-through mode, only interrupts detected on IRQ0 are passed directly to the processor core. Negated—There is no incoming interrupt from that source. Timing Assertion—All of these inputs can be asserted completely asynchronously. Negation—Interrupts programmed as level-sensitive must remain asserted until serviced. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 10-8 Freescale Semiconductor Memory Map/Register Definition Table 10-6. Interrupt Signals—Detailed Signal Descriptions (continued) Signal I/ O Description IRQ_OUT O Interrupt request out. Active-low, open drain. When the PIC is programmed in pass-through mode, this output reflects the raw interrupts generated by on-chip sources. See Section 10.1.4, “Modes of Operation,” for more details. State Asserted—At least one interrupt is currently being signalled to the external system. Meaning Negated—Indicates no interrupt source currently routed to IRQ_OUT. Timing Because external interrupts are asynchronous with respect to the system clock, both assertion and negation of IRQ_OUT occurs asynchronously with respect to the interrupt source. All timing given here is approximate. Assertion—Internal interrupt source: 2 CCB clock cycles after interrupt occurs. External interrupt source: 4 cycles after interrupt occurs. Message interrupts: 2 cycles after write to message register. Negation—Follows interrupt source negation with the following delay: Internal interrupt: 2 CCB clock cycles External interrupt: 4 cycles. Message interrupts: 2 cycles after message register cleared. MCP I Machine check processor. Assertion causes a machine check interrupt to the e500 core. Note that if the e500 core is not configured to process machine check interrupts (MSR[ME] = 0), assertion of MCP causes a checkstop condition. Note that internal sources for the internal core_mcp signal can also cause a machine check interrupt to the processor core, as described in Section 18.4.1.13, “Machine Check Summary Register (MCPSUMR),” Table 10-2 and Table 10-3. State Asserted—The MPC8540 should initiate a machine check interrupt or enter the checkstop Meaning state as directed by the MSR. Negated—Machine check handling is not being requested by the external system. Timing Assertion—May occur at any time, asynchronous to any clock. Negation—Because MCP is edge-triggered, it can be negated one clock after its assertion. UDE I Unconditional debug event. Assertion signal causes an unconditional debug exception to the e500 core. State Asserted—Indicates that the MPC8540 should initiate an unconditional debug event interrupt Meaning to the processor core. Negated—Indicates that unconditional debug event handling is not being requested by UDE. Timing Assertion—May occur at any time, asynchronous to any clock. Negation—Should remain asserted until software in the unconditional debug event interrupt handler causes the external device asserting the UDE signal to negate it. 10.3 Memory Map/Register Definition The PIC programmable register map occupies 256 Kbytes of memory-mapped space. Reading undefined portions of the memory map returns all zeros; writing has no effect. All PIC registers are 32 bits wide and, although located on 128-bit address boundaries, should only be accessed as 32-bit quantities. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 10-9 Programmable Interrupt Controller The PIC address offset map, shown in Table 10-7, is divided into three areas: • • • 0xnn4_0000–0xnn4_FFF0—Global registers 0xnn5_0000–0xnn5_FFF0—Interrupt source configuration registers 0xnn6_0000–0xnn7_FFF0—Per-CPU registers Table 10-7. PIC Register Address Map Offset Register Access Reset Section/Page PIC Register Address Map—Global Registers 0x4_0000– Reserved 0x4_0030 0x4_0040 IPIDR0—Interprocessor interrupt 0 (IPI 0) dispatch register 1 0x4_0050 IPIDR1—IPI 1 dispatch register 0x4_0060 IPIDR2—IPI 2 dispatch register 0x4_0070 IPIDR3—IPI 3 dispatch register 0x4_0080 CTPR—Current task priority register 0x4_0090 WHOAMI—Who am I register 0x4_00A0 IACK—Interrupt acknowledge register 0x4_00B0 EOI—End of interrupt register 0x4_00C0– Reserved 0x4_0FF0 0x4_1000 FRR—Feature reporting register 0x4_1010 Reserved 0x4_1020 GCR—Global configuration register 0x4_1030 Reserved 0x4_1040– Vendor reserved 0x4_1070 0x4_1080 VIR—Vendor identification register 0x4_1090 PIR—Processor initialization register 0x4_10A0 IPIVPR0—IPI 0 vector/priority register 0x4_10B0 IPIVPR1—IPI 1 vector/priority register 0x4_10C0 IPIVPR2—IPI 2 vector/priority register 0x4_10D0 IPIVPR3—IPI 3 vector/priority register 0x4_10E0 SVR—Spurious vector register 0x4_10F0 TFRR—Timer frequency reporting register 0x4_1100 GTCCR0—Global timer 0 current count register 0x4_1110 GTBCR0—Global timer 0 base count register R/W R/W R R/W 0x0000_FFFF 0x0000_0000 0x0000_0000 0x8000_0000 10.3.1.6/10-19 10.3.2.1/10-20 10.3.2.2/10-21 10.3.2.3/10-21 R/W R R W — R — R/W — — R R/W R/W 0x0000_000F 0x0000_0000 0x0000_0000 0x0000_0000 — 0x0037_0002 — 0x0000_0000 — — 0x0000_0000 0x0000_0000 0x8000_0000 10.3.7.2/10-40 10.3.7.3/10-41 10.3.7.4/10-41 10.3.7.5/10-42 — 10.3.1.1/10-16 — 10.3.1.2/10-17 — — 10.3.1.3/10-17 10.3.1.4/10-18 10.3.1.5/10-19 — W — 0x0000_0000 — 10.3.7.1/10-39 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 10-10 Freescale Semiconductor Memory Map/Register Definition Table 10-7. PIC Register Address Map (continued) Offset Register Access R/W R/W R R/W R/W R/W R R/W R/W R/W R R/W R/W R/W — R/W R R R R R/W R/W R/W R/W R/W R/W R/W R/W — Reset 0x8000_0000 0x0000_0001 0x0000_0000 0x8000_0000 0x8000_0000 0x0000_0001 0x0000_0000 0x8000_0000 0x8000_0000 0x0000_0001 0x0000_0000 0x8000_0000 0x8000_0000 0x0000_0001 — 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x00FF_FFFF 0xFFFF_FFFF 0x00FF_FFFF 0xFFFF_FFFF 0x00FF_FFFF 0xFFFF_FFFF 0x00FF_FFFF 0xFFFF_FFFF — Section/Page 10.3.2.4/10-22 10.3.2.5/10-23 10.3.2.2/10-21 10.3.2.3/10-21 10.3.2.4/10-22 10.3.2.5/10-23 10.3.2.2/10-21 10.3.2.3/10-21 10.3.2.4/10-22 10.3.2.5/10-23 10.3.2.2/10-21 10.3.2.3/10-21 10.3.2.4/10-22 10.3.2.5/10-23 — 10.3.2.6/10-24 10.3.3.1/10-26 10.3.3.2/10-27 10.3.3.3/10-27 10.3.3.4/10-28 10.3.4.1/10-29 10.3.4.2/10-30 10.3.4.1/10-29 10.3.4.2/10-30 10.3.4.1/10-29 10.3.4.2/10-30 10.3.4.1/10-29 10.3.4.2/10-30 — 0x4_1120 GTVPR0—Global timer 0 vector/priority register 0x4_1130 GTDR0—Global timer 0 destination register 0x4_1140 GTCCR1—Global timer 1 current count register 0x4_1150 GTBCR1—Global timer 1 base count register 0x4_1160 GTVPR1—Global timer 1 vector/priority register 0x4_1170 GTDR1—Global timer 1 destination register 0x4_1180 GTCCR2—Global timer 2 current count register 0x4_1190 GTBCR2—Global timer 2 base count register 0x4_11A0 GTVPR2—Global timer 2 vector/priority register 0x4_11B0 GTDR2—Global timer 2 destination register 0x4_11C0 GTCCR3—Global timer 3 current count register 0x4_11D0 GTBCR3—Global timer 3 base count register 0x4_11E0 GTVPR3—Global timer 3 vector/priority register 0x4_11F0 GTDR3—Global timer 3 destination register 0x4_1200– Reserved 0x4_12F0 0x4_1300 TCR—Timer control register 0x4_1310 IRQSR0—IRQ_OUT summary register 0 0x4_1320 IRQSR1—IRQ_OUT summary register 1 0x4_1330 CISR0—Critical Interrupt summary register 0 0x4_1340 CISR1—Critical Interrupt summary register 1 0x4_1350 PM0MR0—Performance monitor 0 mask register 0 0x4_1360 PM0MR1—Performance monitor 0 mask register 1 0x4_1370 PM1MR0—Performance monitor 1 mask register 0 0x4_1380 PM1MR1—Performance monitor 1 mask register 1 0x4_1390 PM2MR0—Performance monitor 2 mask register 0 0x4_13A0 PM2MR1—Performance monitor 2 mask register 1 0x4_13B0 PM3MR0—Performance monitor 3 mask register 0 0x4_13C0 PM3MR1—Performance monitor 3 mask register 1 0x4_13D0– Reserved 0x4_13F0 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 10-11 Programmable Interrupt Controller Table 10-7. PIC Register Address Map (continued) Offset Register Access R/W Reset 0x0000_0000 Section/Page 10.3.5.1/10-30 0x4_1400 MSGR0—Message register 0 0x4_1410 MSGR1—Message register 1 0x4_1420 MSGR2—Message register 2 0x4_1430 MSGR3—Message register 3 0x4_1440– Reserved 0x4_14F0 0x4_1500 MER—Message enable register 0x4_1510 MSR—Message status register 0x4_1520– Reserved 0x4_FFF0 — R/W R/W — — 0x0000_0000 0x0000_0000 — — 10.3.5.2/10-31 10.3.5.3/10-31 — PIC Register Address Map—Interrupt Source Configuration Registers 0x5_0000 EIVPR0—External interrupt 0 (IRQ0) vector/priority register 0x5_0010 EIDR0—External interrupt 0 (IRQ0) destination register 0x5_0020 EIVPR1—External interrupt 1 (IRQ1) vector/priority register 0x5_0030 EIDR1—External interrupt 1 (IRQ1) destination register 0x5_0040 EIVPR2—External interrupt 2 (IRQ2) vector/priority register 0x5_0050 EIDR2—External interrupt 2 (IRQ2) destination register 0x5_0060 EIVPR3—External interrupt 3 (IRQ3) vector/priority register 0x5_0070 EIDR3—External interrupt 3 (IRQ3) destination register 0x5_0080 EIVPR4—External interrupt 4 (IRQ4) vector/priority register 0x5_0090 EIDR4—External interrupt 4 (IRQ4) destination register 0x5_00A0 EIVPR5—External interrupt 5 (IRQ5) vector/priority register 0x5_00B0 EIDR5—External interrupt 5 (IRQ5) destination register 0x5_00C0 EIVPR6—External interrupt 6 (IRQ6) vector/priority register 0x5_00D0 EIDR6—External interrupt 6 (IRQ6) destination register 0x5_00E0 EIVPR7—External interrupt 7 (IRQ7) vector/priority register 0x5_00F0 EIDR7—External interrupt 7 (IRQ7) destination register 0x5_0100 EIVPR8—External interrupt 8 (IRQ8) vector/priority register 0x5_0110 EIDR8—External interrupt 8 (IRQ8) destination register 0x5_0120 EIVPR9—External interrupt 9 (IRQ9) vector/priority register 0x5_0130 EIDR9—External interrupt 9 (IRQ9) destination register 0x5_0140 EIVPR10—External interrupt 10 (IRQ10) vector/priority register R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W 0x8000_0000 0x0000_0001 0x8000_0000 0x0000_0001 0x8000_0000 0x0000_0001 0x8000_0000 0x0000_0001 0x8000_0000 0x0000_0001 0x8000_0000 0x0000_0001 0x8000_0000 0x0000_0001 0x8000_0000 0x0000_0001 0x8000_0000 0x0000_0001 0x8000_0000 0x0000_0001 0x8000_0000 10.3.6.1/10-32 10.3.6.2/10-33 10.3.6.1/10-32 10.3.6.2/10-33 10.3.6.1/10-32 10.3.6.2/10-33 10.3.6.1/10-32 10.3.6.2/10-33 10.3.6.1/10-32 10.3.6.2/10-33 10.3.6.1/10-32 10.3.6.2/10-33 10.3.6.1/10-32 10.3.6.2/10-33 10.3.6.1/10-32 10.3.6.2/10-33 10.3.6.1/10-32 10.3.6.2/10-33 10.3.6.1/10-32 10.3.6.2/10-33 10.3.6.1/10-32 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 10-12 Freescale Semiconductor Memory Map/Register Definition Table 10-7. PIC Register Address Map (continued) Offset Register Access R/W R/W R/W — R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W Reset 0x0000_0001 0x8000_0000 0x0000_0001 — 0x8080_0000 0x0000_0001 0x8080_0000 0x0000_0001 0x8080_0000 0x0000_0001 0x8080_0000 0x0000_0001 0x8080_0000 0x0000_0001 0x8080_0000 0x0000_0001 0x8080_0000 0x0000_0001 0x8080_0000 0x0000_0001 0x8080_0000 0x0000_0001 0x8080_0000 0x0000_0001 0x8080_0000 0x0000_0001 0x8080_0000 0x0000_0001 0x8080_0000 0x0000_0001 0x8080_0000 Section/Page 10.3.6.2/10-33 10.3.6.1/10-32 10.3.6.2/10-33 — 10.3.6.3/10-34 10.3.6.4/10-35 10.3.6.3/10-34 10.3.6.4/10-35 10.3.6.3/10-34 10.3.6.4/10-35 10.3.6.3/10-34 10.3.6.4/10-35 10.3.6.3/10-34 10.3.6.4/10-35 10.3.6.3/10-34 10.3.6.4/10-35 10.3.6.3/10-34 10.3.6.4/10-35 10.3.6.3/10-34 10.3.6.4/10-35 10.3.6.3/10-34 10.3.6.4/10-35 10.3.6.3/10-34 10.3.6.4/10-35 10.3.6.3/10-34 10.3.6.4/10-35 10.3.6.3/10-34 10.3.6.4/10-35 10.3.6.3/10-34 10.3.6.4/10-35 10.3.6.3/10-34 0x5_0150 EIDR10—External interrupt 10 (IRQ10) destination register 0x5_0160 EIVPR11—External interrupt 11 (IRQ11) vector/priority register 0x5_0170 EIDR11—External interrupt 11 (IRQ11) destination register 0x5_0180– Reserved 0x5_01F0 0x5_0200 IIVPR0—Internal interrupt 0 vector/priority register 0x5_0210 IIDR0—Internal interrupt 0 destination register 0x5_0220 IIVPR1—Internal interrupt 1 vector/priority register 0x5_0230 IIDR1—Internal interrupt 1 destination register 0x5_0240 IIVPR2—Internal interrupt 2 vector/priority register 0x5_0250 IIDR2—Internal interrupt 2 destination register 0x5_0260 IIVPR3—Internal interrupt 3 vector/priority register 0x5_0270 IIDR3—Internal interrupt 3 destination register 0x5_0280 IIVPR4—Internal interrupt 4 vector/priority register 0x5_0290 IIDR4—Internal interrupt 4 destination register 0x5_02A0 IIVPR5—Internal interrupt 5 vector/priority register 0x5_02B0 IIDR5—Internal interrupt 5 destination register 0x5_02C0 IIVPR6—Internal interrupt 6 vector/priority register 0x5_02D0 IIDR6—Internal interrupt 6 destination register 0x5_02E0 IIVPR7—Internal interrupt 7 vector/priority register 0x5_02F0 IIDR7—Internal interrupt 7 destination register 0x5_0300 IIVPR8—Internal interrupt 8 vector/priority register 0x5_0310 IIDR8—Internal interrupt 8 destination register 0x5_0320 IIVPR9—Internal interrupt 9 vector/priority register 0x5_0330 IIDR9—Internal interrupt 9 destination register 0x5_0340 IIVPR10—Internal interrupt 10 vector/priority register 0x5_0350 IIDR10—Internal interrupt 10 destination register 0x5_0360 IIVPR11—Internal interrupt 11 vector/priority register 0x5_0370 IIDR11—Internal interrupt 11 destination register 0x5_0380 IIVPR12—Internal interrupt 12 vector/priority register 0x5_0390 IIDR12—Internal interrupt 12 destination register 0x5_03A0 IIVPR13—Internal interrupt 13 vector/priority register MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 10-13 Programmable Interrupt Controller Table 10-7. PIC Register Address Map (continued) Offset Register Access R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W Reset 0x0000_0001 0x8080_0000 0x0000_0001 0x8080_0000 0x0000_0001 0x8080_0000 0x0000_0001 0x8080_0000 0x0000_0001 0x8080_0000 0x0000_0001 0x8080_0000 0x0000_0001 0x8080_0000 0x0000_0001 0x8080_0000 0x0000_0001 0x8080_0000 0x0000_0001 0x8080_0000 0x0000_0001 0x8080_0000 0x0000_0001 0x8080_0000 0x0000_0001 0x8080_0000 0x0000_0001 0x8080_0000 0x0000_0001 0x8080_0000 0x0000_0001 0x8080_0000 Section/Page 10.3.6.4/10-35 10.3.6.3/10-34 10.3.6.4/10-35 10.3.6.3/10-34 10.3.6.4/10-35 10.3.6.3/10-34 10.3.6.4/10-35 10.3.6.3/10-34 10.3.6.4/10-35 10.3.6.3/10-34 10.3.6.4/10-35 10.3.6.3/10-34 10.3.6.4/10-35 10.3.6.3/10-34 10.3.6.4/10-35 10.3.6.3/10-34 10.3.6.4/10-35 10.3.6.3/10-34 10.3.6.4/10-35 10.3.6.3/10-34 10.3.6.4/10-35 10.3.6.3/10-34 10.3.6.4/10-35 10.3.6.3/10-34 10.3.6.4/10-35 10.3.6.3/10-34 10.3.6.4/10-35 10.3.6.3/10-34 10.3.6.4/10-35 10.3.6.3/10-34 10.3.6.4/10-35 10.3.6.3/10-34 0x5_03B0 IIDR13—Internal interrupt 13 destination register 0x5_03C0 IIVPR14—Internal interrupt 14 vector/priority register 0x5_03D0 IIDR14—Internal interrupt 14 destination register 0x5_03E0 IIVPR15—Internal interrupt 15 vector/priority register 0x5_03F0 IIDR15—Internal interrupt 15 destination register 0x5_0400 IIVPR16—Internal interrupt 16 vector/priority register 0x5_0410 IIDR16—Internal interrupt 16 destination register 0x5_0420 IIVPR17—Internal interrupt 17 vector/priority register 0x5_0430 IIDR17—Internal interrupt 17 destination register 0x5_0440 IIVPR18—Internal interrupt 18 vector/priority register 0x5_0450 IIDR18—Internal interrupt 18 destination register 0x5_0460 IIVPR19—Internal interrupt 19 vector/priority register 0x5_0470 IIDR19—Internal interrupt 19 destination register 0x5_0480 IIVPR20—Internal interrupt 20 vector/priority register 0x5_0490 IIDR20—Internal interrupt 20 destination register 0x5_04A0 IIVPR21—Internal interrupt 21 vector/priority register 0x5_04B0 IIDR21—Internal interrupt 21 destination register 0x5_04C0 IIVPR22—Internal interrupt 22 vector/priority register 0x5_04D0 IIDR22—Internal interrupt 22 destination register 0x5_04E0 IIVPR23—Internal interrupt 23 vector/priority register 0x5_04F0 IIDR23—Internal interrupt 23 destination register 0x5_0500 IIVPR24—Internal interrupt 24 vector/priority register 0x5_0510 IIDR24—Internal interrupt 24 destination register 0x5_0520 IIVPR25—Internal interrupt 25 vector/priority register 0x5_0530 IIDR25—Internal interrupt 25 destination register 0x5_0540 IIVPR26—Internal interrupt 26 vector/priority register 0x5_0550 IIDR126—Internal interrupt 26 destination register 0x5_0560 IIVPR27—Internal interrupt 27 vector/priority register 0x5_0570 IIDR27—Internal interrupt 27 destination register 0x5_0580 IIVPR28—Internal interrupt 28 vector/priority register 0x5_0590 IIDR28—Internal interrupt 28 destination register 0x5_05A0 IIVPR29—Internal interrupt 29 vector/priority register MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 10-14 Freescale Semiconductor Memory Map/Register Definition Table 10-7. PIC Register Address Map (continued) Offset Register Access R/W R/W R/W R/W R/W — R/W R/W R/W R/W R/W R/W R/W R/W — Reset 0x0000_0001 0x8080_0000 0x0000_0001 0x8080_0000 0x0000_0001 — 0x8000_0000 0x0000_0001 0x8000_0000 0x0000_0001 0x8000_0000 0x0000_0001 0x8000_0000 0x0000_0001 — Section/Page 10.3.6.4/10-35 10.3.6.3/10-34 10.3.6.4/10-35 10.3.6.3/10-34 10.3.6.4/10-35 — 10.3.6.5/10-36 10.3.6.6/10-37 10.3.6.5/10-36 10.3.6.6/10-37 10.3.6.5/10-36 10.3.6.6/10-37 10.3.6.5/10-36 10.3.6.6/10-37 — 0x5_05B0 IIDR29—Internal interrupt 29 destination register 0x5_05C0 IIVPR30—Internal interrupt 30 vector/priority register 0x5_05D0 IIDR30—Internal interrupt 30 destination register 0x5_05E0 IIVPR31—Internal interrupt 31 vector/priority register 0x5_05F0 IIDR31—Internal interrupt 31 destination register 0x5_0600– Reserved 0x5_15F0 0x5_1600 MIVPR0—Messaging interrupt 0 (MSG 0) vector/priority register 0x5_1610 MIDR0—Messaging interrupt 0 (MSG 0) destination register 0x5_1620 MIVPR1—Messaging interrupt 1 (MSG 1) vector/priority register 0x5_1630 MIDR1—Messaging interrupt 1 (MSG 1) destination register 0x5_1640 MIVPR2—Messaging interrupt 2 (MSG 2) vector/priority register 0x5_1650 MIDR2—Messaging interrupt 2 (MSG 2) destination register 0x5_1660 MIVPR3—Messaging interrupt 3 (MSG 3) vector/priority register 0x5_1670 MIDR3—Messaging interrupt 3 (MSG 3) destination register 0x5_1680– Reserved 0x5_FFF0 PIC Register Address Map—Per-CPU Registers 0x6_0000– Reserved 0x6_0030 0x6_0040 IPIDR0—P0 IPI 0 dispatch register 0x6_0050 IPIDR1—P0 IPI 1 dispatch register 0x6_0060 IPIDR2—P0 IPI 2 dispatch register 0x6_0070 IPIDR3—P0 IPI 3 dispatch register 0x6_0080 CTPR0—P0 current task priority register 0x6_0090 WHOAMI0—P0 who am I register 0x6_00A0 IACK0—P0 interrupt acknowledge register 0x6_00B0 EOI0—P0 end of interrupt register 1 — W — All zeros — 10.3.7.1/10-39 R/W R R W 0x0000_000F All zeros All zeros All zeros 10.3.7.2/10-40 10.3.7.3/10-41 10.3.7.4/10-41 10.3.7.5/10-42 Note that these registers provider private access space to the same set of registers at offset 0x6_xxxx. This private access is described in Section 10.3.7, “Per-CPU Registers.” MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 10-15 Programmable Interrupt Controller 10.3.1 Global Registers Most registers have one address. Some registers are replicated for each processor in a multiprocessor device. In this case, each processor accesses its separate registers using the same address, the address decoding being sensitive to the processor ID. A copy of the per-CPU registers is available to each processor core at the same physical address, that is, the private access address space. The private access address space acts like an alias to a processor’s own copy of the per-CPU registers. As shown in Figure 10-31, the ID of the processor initiating the read/write transaction is used to determine which processor’s per-CPU registers to access. For more information on per-CPU registers, see Section 10.3.7, “Per-CPU Registers.” NOTE Register fields designated as write-one-to-clear are cleared only by writing ones to them. Writing zeros to them has no effect. 10.3.1.1 Feature Reporting Register (FRR) The feature reporting register (FRR) shown in Figure 10-3 provides information about interrupt and processor configurations. It also informs the programming environment of the controller version. 0 4 5 15 16 18 19 23 24 31 R0 0 0 0 0 W Reset Offset NIRQ 000 NCPU VID 0000_0000_0011_0111_0000_0000_0000_0010 0x4_1000 Figure 10-3. Feature Reporting Register (FRR) Table 10-8 describes the FRR fields. Table 10-8. FRR Field Descriptions Bits 0–4 5–15 Name — NIRQ Reserved. Number of interrupts. Contains the binary value of the maximum number of interrupt sources supported minus one. The value is 55 (0x37) because this device supports 56 interrupts: 12 external sources, 32 internal sources (only 23 used), four timer sources, four IPI sources and four messaging sources. A zero in this field corresponds to one source. Reserved. Description 16–18 19–23 24–31 — NCPU Number of CPUs. The number of the highest physical CPU supported. The MPC8540 implements one CPU (the processor core), referenced as P0. VID Version ID. Version ID for this interrupt controller. Reports the OpenPIC specification revision level supported by this implementation. A value of two corresponds to revision 1.2 which is the revision level currently supported. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 10-16 Freescale Semiconductor Memory Map/Register Definition 10.3.1.2 Global Configuration Register (GCR) The global configuration register (GCR) shown in Figure 10-4 controls the PIC’s operating mode, and allows software to reset the PIC. 0 1 2 3 31 R RST W Reset Offset 0 M 00000000000000000000000000000 0000_0000_0000_0000_0000_0000_0000_0000 0x4_1020 Figure 10-4. Global Configuration Register (GCR) Table 10-9 describes the GCR fields. Table 10-9. GCR Field Descriptions Bits 0 1 2 Name RST — M Description Reset. Setting this field forces the PIC to be reset. Cleared automatically when the reset sequence is complete. SeeSection 10.4.7, “Reset of the PIC,” for more information on resetting the PIC. Reserved Mode. PIC operating mode. 0 Pass-through mode. On-chip PIC is disabled and interrupts detected on IRQ0 are passed directly to the processor core. See Section 10.1.4, “Modes of Operation,” for more details. 1 Mixed mode. Interrupts are handled by the normal priority and delivery mechanisms of the PIC. See Section 10.1.4, “Modes of Operation,” for more details. Reserved 3–31 — 10.3.1.3 Vendor Identification Register (VIR) The vendor identification register (VIR) shown in Figure 10-5 has specific read-only information about the vendor and the device revision. 0 7 8 15 16 23 24 31 R0 0 0 0 0 0 0 0 W Reset Offset STEP DEVICE ID VENDOR ID 0000_0000_0000_0000_0000_0000_0000_0000 0x4_1080 Figure 10-5. Vendor Identification Register (VIR) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 10-17 Programmable Interrupt Controller Table 10-10 describes the VIR fields. Table 10-10. VIR Field Descriptions Bits 0–7 8–15 16–23 24–31 Name — STEP Reserved Stepping. Indicates the silicon revision for this device. Has no meaning when the VENDOR ID value is zero. Description DEVICE ID Device identification. Vendor-specified identifier for this device. Has no meaning when the VENDOR ID value is zero. VENDOR ID Vendor identification. Specifies the manufacturer of this part. A value of zero implies a generic PIC-compliant device. 10.3.1.4 Processor Initialization Register (PIR) The processor initialization register in the PIC provides a mechanism for software to reset the processor. The core_reset signal to the processor is held active until a zero is written to the processor initialization register field. Thus, PIR should only be written by an external host and the external host should wait at least 100 µsec to clear this field after writing it. The processor should not attempt to write to PIR because writing to it resets the core, and the core will not be able to clear the reset field, and so it would remain in reset indefinitely. Note that although the architecture was designed to support multiple processing cores, only fields corresponding to processors on the device are implemented, as shown in Figure 10-6. In a uniprocessor system, only PIR[P0] is implemented. 0 30 31 R0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x4_1090 00 P0 Figure 10-6. Processor Initialization Register (PIR) Table 10-11 describes the PIR fields. Table 10-11. PIR Field Descriptions Bits 0–30 31 Name — P0 Reserved Processor 0 core reset. Setting this bit causes the PIC unit to assert the core_reset signal to processor 0 (the e500 core). Description MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 10-18 Freescale Semiconductor Memory Map/Register Definition 10.3.1.5 IPI Vector/Priority Registers (IPIVPRn) The interprocessor interrupt (IPI) vector/priority registers contain the interrupt vector and priority fields for the four interprocessor interrupt channels as shown in Figure 10-7. There is one IPI vector/priority register per channel. The VECTOR and PRIORITY values should not be changed while IPIVPRn[A] is set. 0 1 2 11 12 15 16 31 R MSK W Reset Offset A0 0 00 0 0 0 0 0 0 PRIORITY VECTOR 1000_0000_0000_0000_0000_0000_0000_0000 0x4_10A0, 0x4_10B0, 0x4_10C0, 0x4_10D0 Figure 10-7. IPI Vector/Priority Registers (IPIVPRn) Table 10-12 describes the IPIVPRn fields. Table 10-12. IPIVPRn Field Descriptions Bits 0 Name MSK Description Mask. Mask interrupts from this source. Always set following reset. 0 An interrupt request is generated if the corresponding IPR bit is set. 1 Further interrupts from this source are disabled. Activity. Indicates an interrupt has been reported or is in-service. Note that this field is read only. The VECTOR and PRIORITY values should not be changed while IPIVPR n[A] is set. 0 No current interrupt activity associated with this source. 1 The interrupt bit for this source in the IPR or ISR is set. Reserved 1 A 2–11 — 12–15 PRIORITY Priority. Specifies the interrupt priority. The lowest priority is 0 and the highest priority is 15. A priority level of 0 disables interrupt reporting from this source. 16–31 VECTOR Vector. The vector value in this field is returned when the interrupt acknowledge (IACK) register is read and this interrupt resides in the interrupt request register (IRR) shown in Figure 10-37. 10.3.1.6 Spurious Vector Register (SVR) The spurious vector register (SVR) shown in Figure 10-8 contains the 16-bit vector returned to the processor when the IACK register is read for a spurious interrupt. See Section 10.4.4, “Spurious Vector Generation,” for more information about the events and conditions that may cause a spurious vector fetch. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 10-19 Programmable Interrupt Controller 0 15 16 31 R0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 VECTOR W Reset Offset 0000_0000_0000_0000_1111_1111_1111_1111 0x4_10E0 Figure 10-8. Spurious Vector Register (SVR) Table 10-13 describes the SVR fields. Table 10-13. SVR Field Descriptions Bits 0–15 16–3 1 Name — Reserved Description VECTO Spurious interrupt vector. Value returned when IACK is read during a spurious vector fetch. R 10.3.2 Global Timer Registers This section describes the global timer registers. Note that each of the four timers have four individual configuration registers (GTCCRn, GTBCRn, GTVPRn, GTDRn), but they are only shown once in this section. 10.3.2.1 Timer Frequency Reporting Register (TFRR) The timer frequency reporting register (TFRR), shown in Figure 10-9, is written by software to report the clocking frequency of the PIC timers. Note that although TFRR is read/write, the value of this register is ignored by the PIC unit. 0 31 R FREQ W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x4_10F0 Figure 10-9. Timer Frequency Reporting Register (TFRR) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 10-20 Freescale Semiconductor Memory Map/Register Definition Table 10-14 describes the TFRR register. Table 10-14. TFRR Field Descriptions Bits Name Description 0–31 FREQ Timer frequency (in ticks/second (Hz)). Used to communicate the frequency of the global timers’ clock source (the core complex bus (CCB) clock) to user software. See Section 4.4.4, “Clocking,” for more details. TFRR is set only by software for later use by other applications and its value in no way affects the operating frequency of the global timers. The timers operate at a ratio of this clock frequency set by TCR[CLKR]. See Section 10.3.2.6, “Timer Control Register (TCR).” 10.3.2.2 Global Timer Current Count Registers (GTCCRn) The global timer current count registers (GTCCRs), shown in Figure 10-10, contain the current count for each of the four PIC timers. 0 1 31 R TOG W Reset Offset COUNT 0000_0000_0000_0000_0000_0000_0000_0000 0x4_1100, 0x4_1140, 0x4_1180, 0x4_11C0 Figure 10-10. Global Timer Current Count Registers (GTCCRn) Table 10-15 describes the GTCCRn fields. Table 10-15. GTCCRn Field Descriptions Bits 0 1–31 Name TOG Description Toggle. Toggles when the current count decrements to zero. Cleared when GTBCR n[CI] goes from 1 to 0. COUN Current count. Decremented while GTBCRn[CI] is zero. When the timer count reaches zero, an interrupt T is generated (provided it is not masked), the toggle bit is inverted, and the count is reloaded. For non-cascaded timers, the reload value is the contents of the corresponding base count register. Cascaded timers are reloaded with either all ones, or the contents of the base count register, depending on the value of TCR[ROVR]. See Section 10.3.2.6, “Timer Control Register (TCR),” for more details. 10.3.2.3 Global Timer Base Count Registers (GTBCRn) The global timer base count registers (GTBCRn) contain the base counts for each of the four PIC timers as shown in Figure 10-11. This value is reloaded into the corresponding GTCCRn when the current count reaches zero. Note that when zero is written to the base count field, (and GTCCRn[CI] = 0), the timer generates an interrupt on every timer cycle. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 10-21 Programmable Interrupt Controller 0 1 31 R CI W Reset Offset 1000_0000_0000_0000_0000_0000_0000_0000 0x4_1110, 0x4_1150, 0x4_1190, 0x4_11D0 BASE CNT Figure 10-11. Global Timer Base Count Registers (GTBCRn) Table 10-16 describes the GTBCRn fields. Table 10-16. GTBCRn Field Descriptions Bits 0 Name CI Count inhibit. Always set following reset 0 Counting enabled 1 Counting inhibited Base count. When CI transitions from 1 to 0, this value is copied into the corresponding current count register and the toggle bit is cleared. If CI is already cleared (counting is in progress), the base count is copied to the current count register at the next zero crossing of the current count. Description 1–31 BASE CNT 10.3.2.4 Global Timer Vector/Priority Registers (GTVPRn) The global timer vector/priority registers (GTVPRn) contain the interrupt vector and the interrupt priority as shown in Figure 10-12. They also contain the mask and activity fields. 0 1 2 11 12 15 16 31 R MSK W Reset Offset A0000000000 PRIORITY VECTOR 1000_0000_0000_0000_0000_0000_0000_0000 0x4_1120, 0x4_1160, 0x4_11A0, 0x4_11E0 Figure 10-12. Global Timer Vector/Priority Registers (GTVPRn) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 10-22 Freescale Semiconductor Memory Map/Register Definition Table 10-17 describes the GTVPRn fields. Table 10-17. GTVPRn Field Descriptions Bits 0 Name MSK Description Mask. Inhibits interrupts from this source. 0 Interrupt requests are generated when the corresponding IPR bit is set. 1 Further interrupts from this source are disabled. Activity. This field is read-only. 0 No current interrupt activity associated with this source. 1 An interrupt has been requested by the corresponding source or is currently being serviced. The interrupt bit for this source is set in the IPR or ISR. The VECTOR and PRIORITY values should not be changed while the A bit is set. Reserved 1 A 2–11 12–15 16–31 — PRIORIT Priority. Specifies the interrupt priority. The lowest priority is 0 and the highest priority is 15. A Y priority level of 0 disables interrupts from this source. VECTOR Vector. The vector value in this field is returned when the interrupt acknowledge (IACK) register is read and this interrupt resides in the interrupt request register shown in Figure 10-37. 10.3.2.5 Global Timer Destination Registers (GTDRn) The global timer destination register controls the destination processor for this timer’s interrupt, as shown in Figure 10-13. 0 30 31 R 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 P0 W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0001 0x4_1130, 0x4_1170, 0x4_11B0, 0x4_11F0 Figure 10-13. Global Timer Destination Registers (GTDRn) Table 10-18 describes the GTDRn fields. Table 10-18. GTDRn Field Descriptions Bits 0–30 31 Name — P0 Reserved Processor 0. Indicates that processor 0 handles any interrupt. This bit is meaningful only in a multi-core device. Because the MPC8540 is a single-core device, internally serviced interrupts are always directed to processor 0. Permanently set and read only. 1 Interrupt directed to processor 0. Description MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 10-23 Programmable Interrupt Controller 10.3.2.6 Timer Control Register (TCR) The timer control register (TCR) shown in Figure 10-15 provides various configuration options such as count frequency and roll-over behavior. There are two choices for the clock source for the timers: a selectable frequency ratio from the CCB clock, or the RTC signal. The TCR also provides the ability to create timers larger than the default 31-bit global timers. Timer cascade fields allow configuration of up to two 63-bit timers, one 95-bit timer or one 127-bit timer. With one exception mentioned below, the value reloaded into a timer is determined by its roll-over control field TCR[ROVR]. Setting a timer’s roll-over field causes its current count register to roll over to all ones when the count reaches zero. This is equivalent to reloading the count register with 0xFFFF_FFFF instead of its base count value. Clearing a timer’s associated ROVR bit ensures the timer always reloads with its base count value. When timers are cascaded the last (most significant) counter in the cascade also affects their roll-over behavior. Cascaded timers always reload their base count when the most significant counter has decremented to zero, regardless of the settings in TCR[ROVR]. For example, timers 0, 1, and 2 can be cascaded to generate one interrupt every hour. As shown in Table 10-19, given a CCB clock of 333 MHz, letting the timer clock frequency default to 1/8 the system clock, (TCR[CLKR] = 0 sets a clock ratio of 8), provides a basic input of 41.625 MHz to timer 0. Setting timer 0 to count 41,625,000 (0x27B_25A8) timer clock cycles will generate one output per second. Setting both timers 1 and 2 to 59, and cascading all three timers, generates one interrupt every hour from timer 2. Table 10-19. Parameters for Hourly Interrupt Timer Cascade Example System Clock 333 MHz 1 Clock Ratio 1/8 Timer Clock 41.625 MHz Timer 0 Count 41.625 x 10 (0x027B_25A8) 6 Timer 1 Count 59 (0x0000_0036) 1 Timer 2 Count 59 (0x0000_0036) Counting down from 59 through 0 requires 60 ticks. (41.625 x 106 ticks/sec) * (60 sec/min) * (60 min/hr) = total ticks/hr generating 1 interrupt/hr Figure 10-14. Example Calculation for Cascaded Timers 0 4 5 7 8 14 15 16 21 22 23 24 28 29 31 R0 0 0 0 0 ROVR W Reset Offset 0000000 RTM 000000 CLKR 00000 CASC 0000_0000_0000_0000_0000_0000_0000_0000 0x4_1300 Figure 10-15. Timer Control Register (TCR) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 10-24 Freescale Semiconductor Memory Map/Register Definition Table 10-20 describes the TCR fields. Table 10-20. TCR Field Descriptions Bits 0–4 5–7 Name — Reserved Description ROVR Roll-over control for cascaded timers only. Specifies behavior when count reaches zero by identifying the source of the reload value. Cascaded timers are always reloaded with their base count value when the more significant timer in the cascade (the upstream timer) is zero. Bits 5–7 correspond to timers 2–0. Note that global timer 3 always reloads with its base count register. 0 Timer does not roll over. When the count reaches zero, current count register is reloaded with the base count register value. 1 Timer rolls over at zero to all ones. (When the count reaches zero, current count register is reloaded with 0xFFFF_FFFF.) 000 All timers reload with base count. 001 Timers 1 and 2 reload with base count, timer 0 rolls over (reloads with 0xFFFF_FFFF). 010 Timers 0 and 2 reload with base count, timer 1 rolls over (reloads with 0xFFFF_FFFF). 011 Timer 2 reloads with base count, timers 0 and 1 roll over (reload with 0xFFFF_FFFF). 100 Timers 0 and 1 reload with base count, timer 2 rolls over (reloads with 0xFFFF_FFFF). 101 Timer 1 reloads with base count, timers 0 and 2 roll over (reload with 0xFFFF_FFFF). 110 Timer 0 reloads with base count, timers 1 and 2 roll over (reload with 0xFFFF_FFFF). 111 Timers 0, 1, and 2 roll over (reload with 0xFFFF_FFFF). — RTM Reserved Real time mode. Specifies the clock source for the PIC timers. 0 Timer clock frequency is a ratio of the frequency of the platform (CCB) clock as determined by the CLKR field. This is the default value. 1 The RTC signal is used to clock the PIC timers. If this bit is set, the CLKR field has no meaning. Reserved 8–14 15 16–21 22–23 — CLKR Clock ratio. Specifies the ratio of the timer frequency to the platform (CCB) clock. The following clock ratios are supported: 00 Default. Divide by 8 01 Divide by 16 10 Divide by 32 11 Divide by 64 — Reserved 24–28 29–31 CASC Cascade timers. Specifies the output of particular global timers as input to others. 000 Default. Timers not cascaded 001 Cascade timers 0 and 1 010 Cascade timers 1 and 2 011 Cascade timers 0, 1, and 2 100 Cascade timers 2 and 3 101 Cascade timers 0 and 1; timers 2 and 3 110 Cascade timers 1, 2, and 3 111 Cascade timers 0, 1, 2, and 3 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 10-25 Programmable Interrupt Controller 10.3.3 IRQ_OUT and Critical Interrupt Summary Registers The summary registers indicate the interrupt sources directed to IRQ_OUT or cint. Summary register bits are cleared when the corresponding interrupt that caused a bit to be setþ is negated. Note that only level-sensitive interrupts can be directed to IRQ_OUT or cint. 10.3.3.1 IRQ_OUT Summary Register 0 (IRQSR0) The IRQ_OUT summary registers contain one bit for each interrupt source. The corresponding bit is set if the interrupt is active and is directed to IRQ_OUT (that is, if the corresponding xIDR[EP] is set). Figure 10-16 shows IRQSR0. 0 15 16 17 18 19 20 31 R0 0 W Reset Offset 0 0 0 0 0 0 0 0 0 0 0 0 0 0 MSG0 MSG1 MSG2 MSG3 EXTn 0000_0000_0000_0000_0000_0000_0000_0000 0x4_1310 Figure 10-16. IRQ_OUT Summary Register 0 (IRQSR0) Table 10-21 describes the IRQSR0 fields. Table 10-21. IRQSR0 Field Descriptions Bits 0–15 16 Name — Reserved Description MSG0 Message interrupt 0 status 0 Interrupt is not active or not directed to IRQ_OUT. 1 Interrupt is active and is directed to the IRQ_OUT pin (that is, if the corresponding xIDR[EP] is set. MSG1 Message interrupt 1 status 0 Interrupt is not active or not directed to IRQ_OUT. 1 Interrupt is active and is directed to the IRQ_OUT pin (that is, if the corresponding xIDR[EP] is set. MSG2 Message interrupt 2 status 0 Interrupt is not active or not directed to IRQ_OUT. 1 Interrupt is active and is directed to IRQ_OUT (that is, if the EP corresponding xIDR[EP] is set. MSG3 Message interrupt 3 status 0 Interrupt is not active or not directed to IRQ_OUT. 1 Interrupt is active and is directed to IRQ_OUT (that is, if corresponding xIDR[EP] is set. EXTn External interrupts 0–11. Each bit corresponds to a different interrupt according to the following: Bit Interrupt 20IRQ0 21IRQ1 ·· ·· 31IRQ11 0 The corresponding interrupt is not active or not directed to IRQ_OUT. 1 The corresponding interrupt is active and directed to IRQ_OUT (if the corresponding xIDR[EP] is set. 17 18 19 20–31 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 10-26 Freescale Semiconductor Memory Map/Register Definition 10.3.3.2 IRQ_OUT Summary Register 1 (IRQSR1) Figure 10-17 shows IRQSR1. 0 31 R W Reset Offset INTn 0000_0000_0000_0000_0000_0000_0000_0000 0x4_1320 Figure 10-17. IRQ_OUT Summary Register 1 (IRQSR1) Table 10-22 describes the IRQSR1 fields. Table 10-22. IRQSR1 Field Descriptions Bits 0–31 Name INTn Description Internal interrupts 0–31 status. Bit 0 represents INT0. Bit 31 represents INT31. 0 The corresponding interrupt is not active or not directed to IRQ_OUT. 1 The corresponding interrupt is active and is directed to IRQ_OUT (that is, if the corresponding xIDR[EP] is set). 10.3.3.3 Critical Interrupt Summary Register 0 (CISR0) The critical interrupt summary registers contain one bit for each interrupt source. The corresponding bit is set if the interrupt is active and is directed to the processor’s critical interrupt signal cint (if the CI field in its corresponding destination register is set). Figure 10-18 shows CISR0. 0 15 16 19 20 31 R0 0 0 0 W Reset Offset 000000 000000 MSGn EXTn 0000_0000_0000_0000_0000_0000_0000_0000 0x4_1330 Figure 10-18. Critical Interrupt Summary Register 0 (CISR0) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 10-27 Programmable Interrupt Controller Table 10-23 describes the CISR0 fields. Table 10-23. CISR0 Field Descriptions Bits 0–15 16–19 Name — Reserved Description MSGn Message interrupts 0–3. Bit 16 represents MSG0; bit 19 represents MSG3. 0 The corresponding interrupt is not active or not directed to cint. 1 The corresponding interrupt is active and is directed to the cint (if the corresponding xIDR[CI] is set). EXTn External interrupts 0–11. Bit 20 represents IRQ0. Bit 31 represents IRQ11. 0 The corresponding interrupt is not active or not directed to cint. 1 The corresponding interrupt is active and is directed to the cint (if the corresponding xIDR[CI] is set). 20–31 10.3.3.4 Critical Interrupt Summary Register 1 (CISR1) Figure 10-19 shows CISR1. 0 31 R W Reset Offset INTn 0000_0000_0000_0000_0000_0000_0000_0000 0x4_1340 Figure 10-19. Critical Interrupt Summary Register 1 (CISR1) Table 10-24 describes CISR1 register. Table 10-24. CISR1 Field Descriptions Bits Name 0–3 1 INTn Description Internal interrupts 0–31. Bit 0 represents INT0. Bit 31 represents INT31. 0 Corresponding interrupt is not active or not directed to cint. 1 The corresponding interrupt is active and is directed to the cint (if the corresponding xIDR[CI] is set). 10.3.4 Performance Monitor Mask Registers (PMMRs) There are four pairs of performance monitor mask registers, PMnMR0 and PMnMR1. Each pair can be configured to select one interrupt source (interprocessor, timer, message, external, or internal) to generate a performance monitor event. The performance monitor can be configured to track this event in the performance monitor local control registers. See Section 19.3.2.2, “Performance Monitor Local Control Registers (PMLCAn and PMLCBn).” MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 10-28 Freescale Semiconductor Memory Map/Register Definition 10.3.4.1 Performance Monitor Mask Register (Lower) (PMnMR0) Figure 10-20 shows the PMnMR0 registers. Each register is paired with a PMnMR1 register. Because each unreserved bit in the 64-bit pair (PMnMR0/1) specifies a different interrupt, only one bit in each pair can be unmasked at a time. Unmasking more than one bit per pair is considered a programming error and results in unpredictable behavior. 0 7 8 11 12 15 16 19 20 31 R0 0 0 0 0 0 0 0 IPI W Reset Offset 0000_0000_1111_1111_1111_1111_1111_1111 PM0MR0 0x4_1350, PM1MR0 0x4_1370, PM2MR0 0x4_1390, PM3MR0 0x4_13B0 TIMER MSG EXT Figure 10-20. Performance Monitor Mask Registers (PMnMR0) Table 10-25 describes the PMnMR0 fields. Table 10-25. PMnMR0 Field Descriptions Bits 0–7 8–11 Name — IPI Reserved IPI interrupts 0–3 0 The corresponding interrupt source generates a performance monitor event when the interrupt occurs. 1 The corresponding interrupt does not generate a performance monitor event. Description 12–15 TIMER Timer interrupts 0–3 0 The corresponding interrupt source generates a performance monitor event when the interrupt occurs. 1 The corresponding interrupt does not generate a performance monitor event. 16–19 MSG Message interrupts 0–3 0 The corresponding interrupt source generates a performance monitor event when the interrupt occurs. 1 The corresponding interrupt does not generate a performance monitor event. External interrupts IRQ[0:11] 0 The corresponding interrupt source generates a performance monitor event when the interrupt occurs. 1 The corresponding interrupt does not generate a performance monitor event. 20–31 EXT MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 10-29 Programmable Interrupt Controller 10.3.4.2 Performance Monitor Mask Registers (Upper) (PMnMR1) Figure 10-21 shows the PM0MR1–PM3MR1 fields. 0 31 R W Reset Offset INT 1111_1111_1111_1111_1111_1111_1111_1111 PM0MR1 0x4_1360, PM1MR1 0x4_1380, PM2MR1 0x4_13A0, PM3MR1 0x4_13C0 Figure 10-21. Performance Monitor Mask Registers (PMnMR1) Table 10-26 describes the PMnMR1 registers. Table 10-26. PMnMR1 Field Descriptions Bits 0–31 Nam e INT Description Internal interrupts 0–31 0 The corresponding interrupt source generates a performance monitor event when the interrupt occurs. 1 The corresponding interrupt does not generate a performance monitor event. 10.3.5 Message Registers Writing to one of the four message registers (MSGR0–MSGR3) causes a messaging interrupt to the processor. Reading this register clears the messaging interrupt. Note that a messaging interrupt can also be cleared by writing a one to the corresponding status field of the PIC message status register (MSR), shown in Figure 10-24. 10.3.5.1 Message Registers (MSGR0–MSGR3) The message registers (MSGR0–MSGR3) are shown in Table 10-22. 0 31 R W Reset Offset MSG 0000_0000_0000_0000_0000_0000_0000_0000 MSGR0 0x4_1400, MSGR1 0x4_1410, MSGR2 0x4_1420, MSGR3 0x4_1430 Figure 10-22. Message Registers (MSGRs) Table 10-27 describes the MSGR registers. Table 10-27. MSGRn Field Descriptions Bits 0–31 Name MSG Description Message. Contains the 32-bit message data. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 10-30 Freescale Semiconductor Memory Map/Register Definition 10.3.5.2 Message Enable Register (MER) The message enable register (MER) shown in Figure 10-23 contains the enable bits for each message register. The enable bit must be set to enable interrupt generation when the corresponding message register is written. 0 27 28 29 30 31 R0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x4_1500 E3 E2 E1 E0 Figure 10-23. Message Enable Register (MER) Table 10-28 describes the MER fields. Table 10-28. MER Field Descriptions Bits 0–27 28 Name — E3 Reserved Enable 3. Used to enable interrupt generation for MSGR3. 0 Interrupt generation for MSGR3 disabled. 1 Interrupt generation for MSGR3 enabled. Enable 2. Used to enable interrupt generation for MSGR2. 0 Interrupt generation for MSGR2 disabled. 1 Interrupt generation for MSGR2 enabled. Enable 1. Used to enable interrupt generation for MSGR1. 0 Interrupt generation for MSGR1 disabled. 1 Interrupt generation for MSGR1 enabled. Enable 0. Used to enable interrupt generation for MSGR0. 0 Interrupt generation for MSGR0 disabled. 1 Interrupt generation for MSGR0 enabled. Description 29 E2 30 E1 31 E0 10.3.5.3 Message Status Register (MSR) The PIC message status register (MSR) shown in Figure 10-24 contains status bits for each message register. The status bit is set when the corresponding messaging interrupt is active. Reading the corresponding message register or writing a 1 to a status bit clears the corresponding message interrupt and the status bit. 0 27 28 29 30 31 R0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x4_1510 S3 S2 S1 S0 Figure 10-24. Message Status Register (MSR) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 10-31 Programmable Interrupt Controller Table 10-29 describes the MSR fields. Table 10-29. MSR Field Descriptions Bits 0–27 28 Name — S3 Reserved Status 3. Reports status of messaging interrupt 3. Writing a 1 clears this field. 0 Messaging interrupt 3 is not active. 1 Messaging interrupt 3 is active. Status 2. Reports status of messaging interrupt 2. Writing a 1 clears this field. 0 Messaging interrupt 2 is not active. 1 Messaging interrupt 2 is active. Status 1. Reports status of messaging interrupt 1. Writing a 1 clears this field. 0 Messaging interrupt 1 is not active. 1 Messaging interrupt 1 is active. Status 0. Reports status of messaging interrupt 0. Writing a 1 clears this field. 0 Messaging interrupt 0 is not active. 1 Messaging interrupt 0 is active. Description 29 S2 30 S1 31 S0 10.3.6 Interrupt Source Configuration Registers The interrupt source configuration registers control the source of each interrupt, specifying parameters such as the interrupting event, signal polarity, and relative priority. 10.3.6.1 External Interrupt Vector/Priority Registers (EIVPR0–EIVPR11) The external interrupt vector/priority registers (EIVPRs) contain polarity and sense fields for the external interrupts caused by the assertion of any of IRQ[0:11]. The format of the EIVPRs is shown in Figure 10-25. 0 1 2 7 8 9 10 11 12 15 16 31 R MSK W Reset Offset A000000 PS 00 PRIORITY VECTOR 1000_0000_0000_0000_0000_0000_0000_0000 EIVPR0 0x5_0000, EIVPR1 0x5_0020, EIVPR2 0x5_0040, EIVPR3 0x5_0060, EIVPR4 0x5_0080, EIVPR5 0x5_00A0, EIVPR6 0x5_00C0, EIVPR7 0x5_00E0, EIVPR8 0x5_0100, EIVPR9 0x5_0120, EIVPR10 0x5_0140, EIVPR11 0x5_0160 Figure 10-25. External Interrupt Vector/Priority Registers (EIVPR0–EIVPR11) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 10-32 Freescale Semiconductor Memory Map/Register Definition Table 10-30 describes the EIVPR fields. Table 10-30. EIVPRn Field Descriptions Bits 0 Name MSK Description Mask. Masks interrupts from this source. 0 If the corresponding IPR bit is set an interrupt request is generated. 1 Interrupts from this source are disabled. Activity. A read-only field indicating that an interrupt has been requested or is in-service. Note: The P (polarity), S (sense), VECTOR and PRIORITY values should not be changed while the corresponding interrupt is active, that is, while EIVPR n [A] is set. 0 No current interrupt activity associated with this source. 1 The interrupt bit for this source is set in the IPR or ISR. Reserved. Polarity. Specifies the polarity for the external interrupt. 0 Polarity is active-low or negative edge triggered. 1 Polarity is active-high or positive edge-triggered. Sense. Specifies the sense for external interrupts. 0 The external interrupt is edge sensitive. 1 The external interrupt is level-sensitive. Reserved. 1 A 2–7 8 — P 9 S 10–11 12–15 16–31 — PRIORIT Priority. Specifies the interrupt priority. The lowest priority is 0 and the highest priority is 15. A Y priority level of 0 disables interrupts from this source. VECTOR Vector. Contains the value returned when the interrupt acknowledge (IACK) register is read and this interrupt resides in the interrupt request register (IRR) shown in Figure 10-37. 10.3.6.2 External Interrupt Destination Registers (EIDR0–EIDR11) The external interrupt destination registers (EIDRn), shown in Figure 10-26, control the destination of external interrupts caused by the assertion of any of IRQ[0:11]. All external interrupts are directed to the processor 0 interrupt, int, by default. External interrupts can be selectively directed to IRQ_OUT or to the processor 0 critical interrupt signal, cint, instead of int by writing to the appropriate EIDRn fields. NOTE The behavior of the PIC unit is not defined if both the EP and CI bits of the same interrupt destination register are set. 0 1 2 30 31 R EP CI W Reset 0 00 000 00 000 000 00 000 00 000 00 0 0 0 P0 0000_0000_0000_0000_0000_0000_0000_0001 Offset EIDR0 0x5_0010, EIDR1 0x5_0030, EIDR2 0x5_0050, EIDR3 0x5_0070, EIDR4 0x5_0090, EIDR5 0x5_00B0, EIDR6 0x5_00D0, EIDR7 0x5_00F0, EIDR8 0x5_0110, EIDR9 0x5_0130, EIDR10 0x5_0150, EIDR11 0x5_0170 Figure 10-26. External Interrupt Destination Registers (EIDRn) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 10-33 Programmable Interrupt Controller Table 10-31 describes the EIDR fields. Because external interrupts can be channeled only to processor 0, the P0 bit is permanently set. As shown in Figure 10-37, if either the CI or EP bits are set, the interrupt is not sent to the processor’s interrupt input. The EP or CI fields must be set only for level-sensitive interrupts. Setting these fields for edge-sensitive interrupts does not provide reliable interrupt response. Table 10-31. EIDRn Field Descriptions Bits 0 Name EP Description External pin. Allows external interrupt to be serviced externally. 0 External interrupt is serviced internally with int signal to the processor core. 1 External interrupt is directed to IRQ_OUT for external service. Critical interrupt. 0 External interrupt is serviced internally with int signal to the processor core. 1 External interrupt is directed to the processor 0 as a critical interrupt with the cint signal. Reserved Processor 0. Indicates that processor 0 handles any interrupt. This bit is meaningful only in a multi-core device. Because the MPC8540 is a single-core device, all interrupts that are serviced internally are always directed to processor 0. Permanently set and read only. 1 Interrupt directed to processor 0. 1 CI 2–30 31 — P0 10.3.6.3 Internal Interrupt Vector/Priority Registers (IIVPR0–IIVPR31) The internal interrupt vector/priority registers (IIVPRn), shown in Figure 10-27, have the same fields and format as the GTVPRn, except that they apply to the internal interrupt sources listed in Table 10-4. These interrupts are all level-sensitive. NOTE Because all internal interrupts are active-high, clearing any IIVPRn polarity field disables that interrupt. Care should be taken to ensure this field is not inadvertently corrupted when loading or reloading IIVPRs with priority, mask, or vector data. 0 1 2 7 8 9 11 12 15 16 31 R MSK W Reset Offset A000000 P 000 PRIORITY VECTOR 1000_0000_1000_0000_0000_0000_0000_0000 IIVPR0–7: 0x5_0200, 0x5_0220, 0x5_0240, 0x5_0260, 0x5_0280, 0x5_02A0, 0x5_02C0, 0x5_02E0 IIVPR8–15: 0x5_0300, 0x5_0320, 0x5_0340, 0x5_0360, 0x5_0380, 0x5_03A0, 0x5_03C0, 0x5_03E0 IIVPR16–23: 0x5_0400, 0x5_0420, 0x5_0440, 0x5_0460, 0x5_0480, 0x5_04A0, 0x5_04C0, 0x5_04E0 IIVPR24–31: 0x5_0500, 0x5_0520, 0x5_0540, 0x5_0560, 0x5_0580, 0x5_05A0, 0x5_05C0, 0x5_05E0 Figure 10-27. Internal Interrupt Vector/Priority Registers (IIVPRn) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 10-34 Freescale Semiconductor Memory Map/Register Definition Table 10-32 describes the IIVPRn fields. Table 10-32. IIVPRn Field Descriptions Bits 0 Name MSK Description Mask. Mask interrupts from this source. 0 An interrupt request is generated when the corresponding IPR field is set. 1 Further interrupts from this source are disabled. Activity. Indicates an interrupt has been requested or is in-service. Note this field is read only. The VECTOR and PRIORITY values should not be changed while IIVPRn [A] is set. 0 No current interrupt activity associated with this source. 1 The interrupt field for this source is set in the IPR or ISR. Reserved Polarity. Specifies the polarity for the internal interrupt. Note: Because all internal interrupts are active-high, clearing this bit disables the interrupt. 0 Interrupt polarity is active-low. This value disables the interrupt. 1 Interrupt polarity is active-high. This is the reset value should not be changed. Reserved. 1 A 2–7 8 — P 9–11 12–15 16–31 — PRIORIT Priority. Specifies the interrupt priority. The lowest priority is 0 and the highest priority is 15. A Y priority level of 0 disables interrupts from this source. VECTOR Vector. The vector value in this field is returned when the interrupt acknowledge (IACK) register is read and this interrupt resides in the interrupt request register (IRR) shown in Figure 10-37. 10.3.6.4 Internal Interrupt Destination Registers (IIDR0–IIDR31) Internal interrupt destination registers (IIDRn), shown in Figure 10-28, have the same fields and format as EIVDRn, except that they apply to the internal interrupt sources listed in Table 10-4. Like external interrupts, internal interrupts can be directed only to processor 0. NOTE The behavior of the PIC unit is not defined if both the EP and CI bits of the same interrupt destination register are set. 0 1 2 30 31 R EP CI W Reset Offset 0000000000000000000000000000 0 P0 0000_0000_0000_0000_0000_0000_0000_0001 IIDR0–7: 0x5_0210, 0x5_0230, 0x5_0250, 0x5_0270, 0x5_0290, 0x5_02B0, 0x5_02D0, 0x5_02F0 IIDR8–15: 0x5_0310, 0x5_0330, 0x5_0350, 0x5_0370, 0x5_0390, 0x5_03B0, 0x5_03D0, 0x5_03F0 IIDR16–23: 0x5_0410, 0x5_0430, 0x5_0450, 0x5_0470, 0x5_0490, 0x5_04B0, 0x5_04D0, 0x5_04F0 IIDR24–31: 0x5_0510, 0x5_0530, 0x5_0550, 0x5_0570, 0x5_0590, 0x5_05B0, 0x5_05D0, 0x5_05F0 Figure 10-28. Internal Interrupt Destination Registers (IIDRn) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 10-35 Programmable Interrupt Controller Table 10-33 describes the IIDRn fields. Table 10-33. IIDRn Field Descriptions Bits 0 Name EP Description External pin. Allows internal interrupt to be serviced externally. 0 Internal interrupt is serviced internally with int signal to the processor core. 1 Internal interrupt is directed to IRQ_OUT for external service. The behavior of the PIC is not defined if both EP and CI are set on the same interrupt destination register. Critical interrupt. 0 Internal interrupt is serviced internally with int signal to the processor core. 1 Internal interrupt is directed to the processor 0 as a critical interrupt with the cint signal. The behavior of the PIC is not defined if both EP and CI are set on the same interrupt destination register. Reserved Processor 0. Indicates that processor 0 handles any interrupt. This bit is meaningful only in a multi-core device. Because the MPC8540 is a single-core device, all interrupts that are serviced internally are always directed to processor 0. Permanently set and read only. 1 Interrupt directed to processor 0. 1 CI 2–30 31 — P0 10.3.6.5 Messaging Interrupt Vector/Priority Registers (MIVPR0–MIVPR3) The messaging interrupt vector/priority registers (MIVPRn), shown in Figure 10-29, have the same fields and format as the GTVPRn, except they apply to messaging interrupts. 0 1 2 11 12 15 16 31 R MSK W Reset Offset A0000000000 PRIORITY VECTOR 1000_0000_0000_0000_0000_0000_0000_0000 0x5_1600, 0x5_1620, 0x5_1640, 0x5_1660, Figure 10-29. Messaging Interrupt Vector/Priority Registers (MIVPRn) Table 10-34 describes the MIVPRn fields. Table 10-34. MIVPRn Field Descriptions Bits 0 Name MSK Description Mask. Mask interrupts from this source. 0 An interrupt request is generated when the corresponding IPR field is set. 1 Further interrupts from this source are disabled. Activity. Indicates an interrupt has been requested or is in-service. Note this field is read only. The VECTOR and PRIORITY values should not be changed while MIVPR n[A] is set. 0 No current interrupt activity associated with this source. 1 The interrupt field for this source is set in the IPR or ISR. Reserved 1 A 2–11 — MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 10-36 Freescale Semiconductor Memory Map/Register Definition Table 10-34. MIVPRn Field Descriptions (continued) Bits Name Description 12–15 PRIORITY Priority. Specifies the interrupt priority. The lowest priority is 0 and the highest priority is 15. A priority level of 0 disables interrupts from this source. 16–31 VECTOR Vector. The vector value in this field is returned when the interrupt acknowledge (IACK) register is read and this interrupt resides in the interrupt request register (IRR) shown in Figure 10-37. 10.3.6.6 Messaging Interrupt Destination Registers (MIDR0–MIDR3) The messaging interrupt destination registers (MIDRn), shown in Figure 10-30, control the destination for the messaging interrupts. MIDR enables the user to direct the interrupt to either the external interrupt output pin (IRQ_OUT), the core’s critical interrupt input (cint), or to its normal interrupt input (int). NOTE The behavior of the PIC unit is not defined if both the EP and CI bits of the same interrupt destination register are set. 0 1 2 30 31 R EP CI W Reset Offset 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 P0 0000_0000_0000_0000_0000_0000_0000_0001 0x5_1610, 0x5_1630, 0x5_1650, 0x5_1670 Figure 10-30. Messaging Interrupt Destination Registers (MIDRn) Table 10-35 describes the MIDRn fields. Table 10-35. MIDRn Field Descriptions Bits 0 Name EP Description External pin. Allows message interrupt to be serviced externally. 0 Message interrupt is serviced internally with int signal to the processor core. 1 Message interrupt is directed to IRQ_OUT for external service. The behavior of the PIC is not defined if both EP and CI are set on the same interrupt destination register. Critical interrupt 0 Message interrupt is serviced internally with int signal to the processor core. 1 Message interrupt is directed to the processor 0 as a critical interrupt with the cint signal. The behavior of the PIC is not defined if both EP and CI are set on the same interrupt destination register. Reserved Processor 0. Indicates that processor 0 handles any interrupt. This bit is meaningful only in a multi-core device. Because the MPC8540 is a single-core device, all interrupts that are serviced internally are always directed to processor 0. Permanently set and read only. 1 Interrupt directed to processor 0. 1 CI 2–30 31 — P0 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 10-37 Programmable Interrupt Controller 10.3.7 Per-CPU Registers The OpenPIC programming model supports multiprocessor systems of up to 32 separate processors. As such, the OpenPIC interface specification provides for coordinating both the requesting and servicing of interrupts among several processor cores within a single integrated device. To fully comply with the OpenPIC specification, the PIC incorporates several of these multiprocessor capabilities. Because the value of features such as private address space for per-CPU registers and interprocessor interrupts is fully realized only in a multi-core environment, their utility in this single-core device is not intuitive. The registers in Table 10-36 are called per-CPU registers, because they would be duplicated for each core in a multi-core device. The OpenPIC interface specifies that a copy of these registers be available to each core at the same physical address by using the ID of the processor that initiates the transaction to determine the set of per-CPU registers to access. Table 10-36. Per-CPU Registers—Private Access Address Offsets Register Name IPI 0 dispatch register (IPIDR0) IPI 1 dispatch register (IPIDR1) IPI 2 dispatch register (IPIDR2) IPI 3 dispatch register (IPIDR3) Current task priority register (CTPR) Who am I register (WHOAMI) Interrupt acknowledge register (IACK) End of interrupt register (EOI) Offset 0x4_0040 0x4_0050 0x4_0060 0x4_0070 0x4_0080 0x4_0090 0x4_00A0 0x4_00B0 These addresses, shown in Table 10-36, appear in the memory map at the same offset for every processor, and are called the private access space. Because the MPC8540 has only one core, there is only one set of per-CPU registers, each register having two addresses. For example, the CTPR is located normally at 0x6_0080 and at the private access address of 0x4_0080. While this double mapping seems superfluous on a single-core device, the purpose of this feature is to enable user code to execute correctly in an multiprocessor environment without needing to know which CPU it is running on. It is included on this device to simplify the porting of such code. An example of how the different registers are addressed in a four-core device is illustrated in Figure 10-31. Note that when accessing a register normally, each core sources a different address. However, when accessing the same register using the per-CPU address space, each core sources the same address. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 10-38 Freescale Semiconductor Memory Map/Register Definition IACK Registers (one per core) Reg 3 Normal Register Address Decode Reg 2 0x 63 Reg 1 62 0x 61 0x Reg 0 0x 60 0x60 0x61 0x62 0x63 0x 40 + 0 0x40 4 0x 0 + 1 + 2 4 0x 0 4 0x 0 + 3 ID0 Address 0x40 ID1 Uses Core ID Per-CPU Register Address Decode 0x40 ID2 Processor Cores 0x40 Address ID3 Figure 10-31. Per-CPU Register Address Decoding in a Four-Core Device 10.3.7.1 Interprocessor Interrupt Dispatch Register (IPIDR0–IPIDR3) There are four interprocessor interrupt dispatch registers (IPIDRs), one for each IPI channel, as shown in Figure 10-32. Writing to an IPIDR with a bit set causes a self interrupt. Interprocessor interrupts are of limited usefulness in a single-core device; however, this feature can serve as a doorbell type interrupt because external bus masters can write to these registers. 0 30 31 R0 W Reset Offset 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 P0 0000_0000_0000_0000_0000_0000_0000_0000 IPIDR0 0x6_0040, IPIDR1 0x6_0050 IPIDR2 0x6_0060, IPIDR3 0x6_0070 (see Section 10.3.7, “Per-CPU Registers.”) Figure 10-32. Interprocessor Interrupt Dispatch Registers (IPIDR0–IPIDR3) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 10-39 Programmable Interrupt Controller Table 10-37 describes the IPIDRn fields. Table 10-37. IPIDRn Field Descriptions Bits 0–30 31 Name — P0 Reserved Processor 0. Determines if processor 0 receives the interrupt. Write only. 0 Processor 0 does not receive the interrupt. 1 Directs the interrupt to processor 0. Description 10.3.7.2 Processor Current Task Priority Register (CTPR) There is one CTPR on this device (MPC8540) shown in Figure 10-33. Software must write the priority of the current processor task in the CTPR. The PIC uses this value for comparison with the priority of incoming interrupts. Given several concurrent incoming interrupts, the highest priority interrupt is asserted to the core if the following apply: • • The interrupt is not masked The priority of the interrupt is higher than the values in the CTPR and ISR Priority levels from 0 (lowest) to 15 (highest) are supported. Setting the task priority to 15 masks all external interrupts signaled through the int signal. Hardware sets CTPR to 0x0000_000F during reset or when the Px field of the processor initialization register is set. This effectively prevents the PIC from signaling any interrupts through the int signal. 0 27 28 31 R0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 TASKP W Reset Offset 0000_0000_0000_0000_0000_0000_0000_1111 0x6_0080 (see Section 10.3.7, “Per-CPU Registers.”) Figure 10-33. Processor Current Task Priority Register (CTPR) Table 10-38 describes the CTPR. Table 10-38. CTPR Field Descriptions Bits 0–27 28–31 Name — Reserved Description TASKP Task priority. This field is set from 0 to 15, where 15 corresponds to the highest priority for processor tasks. If CTPR[TASKP] = 0xF, no interrupts are signaled to the processor. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 10-40 Freescale Semiconductor Memory Map/Register Definition 10.3.7.3 Who Am I Register (WHOAMI) The processor who am I register (WHOAMI), shown in Figure 10-34, can be read by the processor to determine its physical connection to the PIC. The value returned when reading this register may be used to determine the value for the destination masks used for dispatching interrupts. 0 26 27 31 R00000000000 W Reset Offset 00000 0 0 0 0 0 0 0 0 0 0 0 ID 0000_0000_0000_0000_0000_0000_0000_0000 0x6_0090 (see Section 10.3.7, “Per-CPU Registers.”) Figure 10-34. Processor Who Am I Register (WHOAMI) Table 10-39 describes the WHOAMI fields. Table 10-39. WHOAMI Field Descriptions Bits 0–26 27–3 1 Name — ID Reserved ID. Returns the ID of the processor reading this register. Always returns zero. 00000Processor core Description 10.3.7.4 Processor Interrupt Acknowledge Register (IACK) In systems based on processors that implement the PowerPC architecture, the interrupt acknowledge function occurs as an explicit read operation to a memory-mapped interrupt acknowledge register (IACK), shown in Figure 10-35. Reading the interrupt acknowledge register returns the interrupt vector corresponding to the highest priority pending interrupt. Reading IACK also has the following side effects: • • • The associated field in the interrupt pending register (IPR) is cleared for edge-sensitive interrupts to indicated that the interrupt is no longer pending. The corresponding field in the in-service register (ISR) is set to indicated that the interrupt has been signaled to the core. The interrupt signal (int) to the processor is negated. Reading this register when there is no pending interrupt returns the spurious vector value as described in Section 10.3.1.6, “Spurious Vector Register (SVR).” MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 10-41 Programmable Interrupt Controller 0 15 16 31 R0000000000000000 W Reset Offset VECTOR 0000_0000_0000_0000_0000_0000_0000_0000 0x6_00A0 (see Section 10.3.7, “Per-CPU Registers.”) Figure 10-35. Processor Interrupt Acknowledge Register (IACK) Table 10-40 describes the IACK fields. Table 10-40. IACK Field Descriptions Bits 0–15 Name — Reserved Description 16–31 VECTOR Interrupt vector. Vector of the highest pending interrupt (read only) 10.3.7.5 Processor End of Interrupt Register (EOI) Writing to the end of interrupt (EOI) register shown in Figure 10-36 signals the end of processing for the highest-priority interrupt currently in-service by the processor. The write to the EOI updates the ISR by retiring the highest priority interrupt. Data values written to this register are ignored, and zero is assumed. 0 27 28 31 R000000000000000000000000000 W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x6_00B0 (see Section 10.3.7, “Per-CPU Registers.”) 0 EOI CODE Figure 10-36. End of Interrupt Register (EOI) Table 10-41 describes the EOI fields. Table 10-41. EOI Field Descriptions Bits 0–27 Name — Reserved Description 28–31 EOI CODE 0000 (write only) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 10-42 Freescale Semiconductor Functional Description 10.4 Functional Description This section is a functional description of the PIC. 10.4.1 Flow of Interrupt Control Figure 10-37 is a block diagram of the PIC unit showing the flow of interrupt processing. This figure is intended to aid in understanding and does not fully represent all internal circuitry of the actual implementation. The PIC receives interrupt signals from both external and internal sources. These signals are qualified and latched in the interrupt pending register (IPR). The IPR feeds the interrupt selector (IS). The interrupt router of the PIC monitors the outputs of its internal interrupt request register (IRR) and other configuration registers. When the priority of the interrupt latched in the IRR is higher than the value in the processor’s task priority register, the interrupt router asserts the internal interrupt signal (int) to indicate an interrupt request to the processor. This causes the processor to vector to the external interrupt handler. Note that the IPR, IS, and IRR are internal registers that are not accessible to the programmer. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 10-43 Programmable Interrupt Controller int To Processor Core Interrupt Router Processor’s Task Priority Register Pipeline Latch Interrupt Request Register (IRR) Highest Priority Interrupt IACK Interrupt Selector (IS) PIC Configuration and Status Registers EOI End of Interrupt In-Service Register (ISR) Activity Bit Mask Register Interrupt Pending Register (IPR) Global Timers (4) Critical Interrupt xIDR[CI] Internal, External, and Message Interrupts Inter-Processor cint To Processor Core External Pin xIDR[EP] IRQ_OUT Note: All signal lines represent buses except for int, cint, and IRQ_OUT. The behavior of the PIC unit is not defined if both the EP and CI bits of the same interrupt destination register are set. Figure 10-37. PIC Interrupt Processing Flow Diagram The interrupt handler executing on the processor should then acknowledge the interrupt by explicitly reading IACK (at which point the interrupt is considered to be in-service). The PIC unit interprets this read as an interrupt acknowledge (IACK) cycle; in response, the PIC unit returns the MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 10-44 Freescale Semiconductor Functional Description vector associated with the interrupt source to the interrupt handler routine. The handler can further vector to different branches of interrupt handling accordingly. Note that reading IACK also negates the interrupt signal to the processor. See Section 10.3.7.4, “Processor Interrupt Acknowledge Register (IACK),” for more details. The internal in-service register (ISR) tracks all interrupts that have caused int to be asserted and that have received an IACK cycle. An interrupt is considered in-service from the time its vector is read (through an IACK cycle) until the end of interrupt (EOI) register is written, generating what the PIC considers an EOI signal. 10.4.1.1 Interrupt Source Priority Each interrupt source is assigned a priority value of 0–15 through its corresponding vector/priority register, where 15 is the highest priority. Interrupts are delivered only when the priority in the vector/priority is greater than the priority programmed into the current task priority register, CTPR. Setting a source priority to zero inhibits that interrupt. Likewise, setting the CTPR value to 15 disables all interrupts directed to the int signal. The PIC unit services simultaneous interrupts occurring with the same priority according to the following order: 1. 2. 3. 4. 5. MSG0–MSG3 IPI0–IPI3 Timer0–timer3 IRQ[0:11] Internal0–internal31 For example, if MSG0, MSG2, and IPI0 have the same priority and generate simultaneous interrupts, they are serviced in the following order: 1. MSG0 2. MSG2 3. IPI0 10.4.1.2 Processor Current Task Priority The CTPR is set by system software to indicate the relative importance of the task running on the processor. The processor does not receive interrupts with a priority level equal to or lower than its current task priority. Therefore setting the current task priority to 15 for a particular processor prevents the delivery of any interrupt to the processor. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 10-45 Programmable Interrupt Controller 10.4.1.3 Interrupt Acknowledge The PIC unit notifies the processor core of an interrupt by asserting the int signal. When the processor core acknowledges the interrupt request by reading the interrupt acknowledge register (IACK) in the PIC unit, the PIC returns the 16-bit vector associated with the interrupt source to the processor. The interrupt is then considered to be in-service, and remains in-service until the processor performs a write to the PIC unit end of interrupt register (EOI). Writing to the EOI is referred to as an EOI cycle. 10.4.2 Nesting of Interrupts If the processor is servicing an interrupt, it can only be interrupted again if the PIC receives an interrupt request from a source with higher priority than the one currently being serviced. This is true even if software, as part of its interrupt service routine, writes a new and lower value into the CTPR. Thus, although several interrupts may be in-service simultaneously, the code currently executing is always handling the highest priority of all the interrupts that are in service. When the processor performs an EOI cycle, this highest priority interrupt is taken out-of-service. The next EOI cycle takes the next-highest priority interrupt out-of-service, and so on. An interrupt with lower priority than those currently in-service is not started until all higher priority interrupts complete even if its priority is greater than the CTPR value. 10.4.3 Processor Initialization The processor can reset itself by writing to the processor initialization register (PIR). This causes the assertion of the core_reset output signal. When this occurs, the processor also gets written to 0x000F to disable the delivery of any interrupts. 10.4.4 Spurious Vector Generation Under certain circumstances, the PIC has no valid vector to return to the processor during an interrupt acknowledge cycle. In these cases, the spurious vector from the spurious vector register is returned. The following cases cause a spurious vector fetch: • • • • int is asserted in response to an externally sourced interrupt that is activated with level-sensitive logic and the source negates before the interrupt is acknowledged. int is asserted for an interrupt source that is later masked (using the mask bit in the corresponding vector/priority register) before the interrupt is acknowledged. int is asserted for an interrupt source that is later masked by an increase in the task priority level before the interrupt is acknowledged. An interrupt acknowledge cycle is performed by the processor in spite of the fact that the PIC did not assert the int signal. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 10-46 Freescale Semiconductor Functional Description In all cases, a spurious vector is not returned if another pending interrupt has sufficient priority to interrupt the processor. If such an interrupt is available, the vector for that source is returned. The EOI register should not be written in response to the spurious vector. Otherwise, a previously-accepted interrupt might be cleared unintentionally. 10.4.5 Messaging Interrupts There are four 32-bit message registers that can be used to send 32-bit messages to the processor. A messaging interrupt is generated by writing a message register if the corresponding enable bit in the message enable/status register is set, and the interrupt is not masked. Reading the message register or writing a 1 to the status bit clears the interrupt. 10.4.6 Global Timers There are appropriate clock prescalers and synchronizers to provide a time base for the four internal timers of the PIC unit. The timers can be individually programmed to generate a processor interrupt when they count down to zero and can be used to generate regular periodic interrupts. Each timer has the following four configuration and control registers: • • • • Global timer current count register (GTCCRn) Global timer base count register (GTBCRn) Global timer vector-priority register (GTVPRn) Global timer destination register (GTDRn) The timer frequency should be written to the TFRR. (All of the timers operate at this frequency.) Refer to Section 10.3.2.1, “Timer Frequency Reporting Register (TFRR),” for a description of this register. Timer interrupts are all edge-triggered interrupts. If a timer period expires while a previous interrupt from the same source is pending or in-service, the subsequent interrupt is lost. The timer control register (TCR) provides users with the ability to create timers larger than the 31-bit global timers. The option also exists to change the timer frequency by setting the appropriate fields of the TCR. See Section 10.3.2.6, “Timer Control Register (TCR).” 10.4.7 Reset of the PIC The PIC unit is reset by a device power-on reset (POR) or by software that sets the GCR[RST] bit. Both of these actions cause the following: • • All pending and in-service interrupts are cleared. All interrupt mask bits are set. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 10-47 Programmable Interrupt Controller • • • • • • • • • • Polarity, sense, external pin, critical interrupt, and activity fields are reset to their default values. PIR, TFRR, TCR, MER, MSR, and MSGR0–3 are cleared. MSG and timer destination fields are set. The IPI dispatch registers are cleared. All timer base count values are reset to zero and count inhibited. The CTPR[TASKP] is reset to 0xF, thus disabling interrupt delivery to the processor. The spurious interrupt vector resets to 0xFFFF. The PMMRs are reset to 0xFFFF. The PIC defaults to the pass-through mode (GCR[M] = 0). All other registers remain at their pre-reset programmed values. The GCR[RST] bit is automatically cleared when the reset sequence is complete. 10.5 Initialization/Application Information This section contains initialization and application information for the PIC. 10.5.1 Programming Guidelines The following subsections contain information about programming PIC registers. 10.5.1.1 PIC Registers Most PIC control and status registers are readable and return the last value written. The exceptions to this rule are as follows: • • • • IPI dispatch registers and the EOI register, which return zeros on reads. Activity bit (A) of the vector/priority registers, which returns the value according to the status of the current interrupt source. IACK register, which returns the vector of highest priority which is currently pending, or the spurious vector. Reserved fields always return 0. The following guidelines are recommended when the PIC unit is programmed in mixed mode (GCR[M] = 1): • • All PIC registers must be located in a cache-inhibited and guarded area (through the processor MMU). The PIC portion of the address map must be set-up appropriately. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 10-48 Freescale Semiconductor Initialization/Application Information In addition, the following initialization sequence is recommended: 1. Write the vector, priority, and polarity values in each interrupt’s vector/priority register, leaving their MSK (mask) bit set. This is required only if interrupts are used. 2. Clear the CTPR (CTPR = 0x0000_0000). 3. Program the PIC to mixed mode by setting GCR[M]. 4. Clear the MSK bit in the vector/priority registers to be used. 5. Perform a software loop to clear all pending interrupts: — Load counter with FPR[NIRQ]. — While counter > 0, perform IACK and EOIs to guarantee all the interrupt pending and in-service registers are cleared. 6. Set the processor CTPR value to the desired value. Depending on the interrupt system configuration, the PIC may generate spurious interrupts to clear interrupts latched during power-up. A spurious or non-spurious vector is returned for an interrupt acknowledge cycle in this case. See the programming note below for the non-spurious case. NOTE: Because the default polarity/sense for external interrupts is edge-sensitive, and edge-sensitive interrupts are not cleared until they are acknowledged, it is possible for the PIC to store spurious edges detected during power-up as pending external interrupts. If software permanently configures an external interrupt source to be edge-sensitive, it may receive the vector for the interrupt source and not a spurious interrupt vector when software clears the mask bit. This can occur once for any edge-sensitive interrupt when its mask bit is first cleared and the PIC is in mixed mode. To avoid having to handle a false interrupt for this case, software can clear the PIC interrupt pending register of these spurious edge detections by first configuring the polarity/sense of external interrupt sources to be level-sensitive; high-level if the input is a positive-edge source, low-level if it’s a negative-edge source (while the mask bit remains set). After this is complete, configuring the external interrupt source as edge-sensitive will not cause a false interrupt. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 10-49 Programmable Interrupt Controller 10.5.1.2 Changing Interrupt Source Configuration To change the vector, priority, polarity, sense or destination of an active (unmasked) interrupt source, the following sequence should be performed: 1. 2. 3. 4. Mask the source using the mask (MSK) bit in the vector/priority register. Wait for the activity (A) bit for that source to be cleared. Make the desired changes. Unmask the source. Internal Interrupts L2 cache e500 coherency model (ECM) DDR DRAM Local bus controller (LBC) DMA channels 0–3 PCI/PCI-X RapidIO inbound port write/ error interrupt RapidIO doorbell inbound interrupt RapidIO outbound message interrupt RapidIO inbound message interrupt TSEC 1/2 transmit interrupts TSEC 1/2 receive interrupts TSEC 1/2 receive/transmit error interrupts int Fast Ethernet controller (FEC) DUART PIR[P0] I2C controller Performance monitor interrupt IRQ[0:8] DMA_DREQ3/IRQ9 DMA_DACK3/IRQ10 DMA_DDONE3/IRQ11 IRQ_OUT SRESET MCP UDE CKSTP_IN CKSTP_OUT HRESET HRESET_REQ Global Utilities cint core_reset Wake up core/ service interrupt InterProcessor (4) Messaging (4) Global Timer (4) core_mcp core_ude core_fault core_tbint core_halt core_hreset_req core_hreset core_ckstop_out core_wrs_[0:1] Freescale Semiconductor Timer Facility Watchdog Timer PIC e500 Core PIC-defined interrupts External Interrupts MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 10-50 Chapter 11 I2C Interface This chapter describes the inter-IC (IIC or I2C) bus interface implemented on this device. 11.1 Introduction The I2C bus is a two-wire—serial data (SDA) and serial clock (SCL)— bidirectional serial bus that provides a simple efficient method of data exchange between this device and other devices, such as microcontrollers, EEPROMs, real-time clock devices, A/D converters, and LCDs. Figure 11-1 shows a block diagram of the I2C interface. Address and Control Interrupt Data Addr Decode Data Mux I2CDSRR I2CADR I2CFDR I2CCR I2CSR I2CDR Clock Divide & Control START/ STOP/ Restart and Arbitration Control In/Out Data Shift Register Input Sync and Digital Filter Address Compare arb_lost SDA SCL Figure 11-1. I2C Block Diagram MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 11-1 I2C Interface 11.1.1 Overview The two-wire I2C bus minimizes interconnections between devices. The synchronous, multiple-master I2C bus allows the connection of additional devices to the bus for expansion and system development. The bus includes collision detection and arbitration that prevent data corruption if two or more masters attempt to control the bus simultaneously. 11.1.2 Features The I2C interface includes the following features: • • • • • • • • • • Two-wire interface Multiple-master operation Arbitration lost interrupt with automatic mode switching from master to slave Calling address identification interrupt START and STOP signal generation/detection Acknowledge bit generation/detection Bus busy detection Software-programmable clock frequency Software-selectable acknowledge bit On-chip filtering for spikes on the bus 11.1.3 Modes of Operation The I2C unit on this device can operate in one of the following modes: • • • Master mode—The I2C is the driver of the SDA line. It cannot use its own slave address as a calling address. The I2C cannot be a master and a slave simultaneously. Slave mode—The I2C is not the driver of the SDA line. The module must be enabled before a START condition from a non-I2C master is detected. Interrupt-driven byte-to-byte data transfer—When successful slave addressing is achieved (and SCL returns to zero), the data transfer can proceed on a byte-to-byte basis in the direction specified by the R/W bit sent by the calling master. Each byte of data must be followed by an acknowledge bit, which is signalled from the receiving device. Several bytes can be transferred during a data transfer session. Boot sequencer mode—This mode can be used to initialize the configuration registers in the device after the I2C module is initialized. Note that the MPC8540 powers up with boot sequencer mode disabled as a default, but this mode can be selected with the cfg_boot_seq[0:1] power-on reset (POR) configuration signals that are located on the LGPL3 and LGPL5 signals. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 11-2 Freescale Semiconductor • External Signal Descriptions Additionally, the following three I2C-specific states are defined for the I2C interface: • • • START condition—This condition denotes the beginning of a new data transfer (each data transfer contains several bytes of data) and awakens all slaves. Repeated START condition—A START condition that is generated without a STOP condition to terminate the previous transfer. STOP condition—The master can terminate the transfer by generating a STOP condition to free the bus. 11.2 External Signal Descriptions The following sections give an overview of signals and provide detailed signal descriptions. 11.2.1 Signal Overview The I2C interface uses the SDA and SCL signals, described in Table 11-1, for data transfer. All devices connected to these two signals must have open-drain or open-collector outputs. A logical AND function is performed on both signals with external pull-up resistors. Note that the signal patterns driven on SDA represent address, data, or read/write information at different stages of the protocol. Table 11-1. I2C Interface Signal Description Signal Name Idle State I/O Serial Clock (SCL) HIGH I State Meaning When the I2C module is idle or acts as a slave, SCL defaults as an input. The unit uses SCL to synchronize incoming data on SDA. The bus is assumed to be busy when SCL is detected low. As a master, the I2C module drives SCL along with SDA when transmitting. As a slave, the I2C module drives SCL low for data pacing. When the I2C module is idle or in a receiving mode, SDA defaults as an input. The unit receives data from other I2C devices on SDA. The bus is assumed to be busy when SDA is detected low. When writing as a master or slave, the I2C module drives data on SDA synchronous to SCL. O Serial Data (SDA) HIGH I O 11.2.2 Detailed Signal Descriptions SDA and SCL, described in Table 11-2, serve as a communication interconnect with other devices. All devices connected to these two signals must have open-drain or open-collector outputs. The logic AND function is performed on both of these signals with external pull-up resistors. Refer to the MPC8540 Integrated Host Processor Hardware Specifications for the electrical characteristics of these signals. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 11-3 I2C Interface Table 11-2. I2C Interface Signal—Detailed Signal Descriptions Signal SCL I/O I/O O Description Serial clock. Performs as an input when the MPC8540 is programmed as an I2C slave. SCL also performs as an output when the MPC8540 is programmed as an I2C master. As outputs for the bidirectional serial clock, these signals operate as described below. State Asserted/Negated—Driven along with SDA as the clock for the data. Meaning I As inputs for the bidirectional serial clock, these signals operate as described below. State Asserted/Negated—The I2C unit uses this signal to synchronize incoming data on SDA. The bus is assumed to be busy when this signal is detected low. Meaning SDA I/O O Serial data. Performs as an input when the MPC8540 is in a receiving mode. SDA also performs as an output signal when the MPC8540 is transmitting (as an I2C master or a slave). As outputs for the bidirectional serial data, these signals operate as described below. State Asserted/Negated— Data is driven. Meaning I As inputs for the bidirectional serial data, these signals operate as described below. State Asserted/Negated—Used to receive data from other devices. The bus is assumed to Meaning be busy when SDA is detected low. 11.3 Memory Map/Register Definition Table 11-3 lists the I2C–specific registers and their addresses. Table 11-3. I2C Memory Map Address 0x0_3000 0x0_3004 0x0_3008 0x0_300C 0x0_3010 0x0_3014 I2C Register I2CADR—I2C address register I2CFDR—I2C frequency divider register I2CCR—I2C control register I2CSR—I C status register I2CDR—I C data register I2CDFSRR—I C digital filter sampling rate register 2 2 2 Access R/W R/W R/W R/W R/W R/W Reset 0x00 0x00 0x00 0x81 0x00 0x10 Section/Page 11.3.1.1/11-5 11.3.1.2/11-5 11.3.1.3/11-6 11.3.1.4/11-7 11.3.1.5/11-9 11.3.1.6/11-10 11.3.1 Register Descriptions This section describes the I2C registers in detail. NOTE Reserved bits should always be written with the value they returned when read. That is, the register should be programmed by reading the value, modifying appropriate fields, and writing back the value. The return value of the reserved fields should not be assumed, even though the reserved fields return zero. This note does not apply to the I2C data register (I2CDR). MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 11-4 Freescale Semiconductor Memory Map/Register Definition 11.3.1.1 I2C Address Register (I2CADR) Figure 11-2 shows the I2CADR register, which contains the address to which the I2C interface responds when addressed as a slave. Note that this is not the address that is sent on the bus during the address-calling cycle when the I2C module is in master mode. 0 6 7 R W Reset Offset ADDR 0000_0000 0x0_3000 Figure 11-2. I2C Address Register (I2CADR) Table 11-4 describes the bit settings of I2CADR. Table 11-4. I2CADR Field Descriptions Bits Name Description 0–6 ADDR Slave address. Contains the specific slave address that is used by the I2C interface. Note that the default mode of the I2C interface is slave mode for an address match. Note that an address match is one of the conditions that can cause I2CSR[MIF] to be set, signaling an interrupt pending condition. 7 — Reserved 11.3.1.2 I2C Frequency Divider Register (I2CFDR) Figure 11-3 shows the bits of the I2C frequency divider register. 0 1 2 7 R W Reset Offset FDR 0000_0000 0x0_3004 Figure 11-3. I2C Frequency Divider Register (I2CFDR) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 11-5 I2C Interface Table 11-5 describes the bit settings of I2CFDR. It also maps the I2CFDR[FDR] field to the clock divider values. Table 11-5. I2CFDR Field Descriptions Bits Name 0–1 2–7 — FDR Reserved Frequency divider ratio. Used to prescale the clock for bit rate selection. The serial bit clock frequency of SCL is equal to the CCB clock divided by the divider. Note that the frequency divider value can be changed at any point in a program. The serial bit clock frequency divider selections are described as follows: FDR Divider (Decimal) FDR Divider (Decimal) FDR Divider (Decimal) 0x00 288 0x16 12288 0x2B 1024 0x01 320 0x17 15360 0x2C 1280 0x02 384 0x18 18432 0x2D 1536 0x03 480 0x19 20480 0x2E 1792 0x04 576 0x1A 24576 0x2F 2048 0x05 640 0x1B 30720 0x30 2560 0x06 768 0x1C 36864 0x31 3072 0x07 960 0x1D 40960 0x32 3584 0x08 1152 0x1E 49152 0x33 4096 0x09 1280 0x1F 61440 0x34 5120 0x0A 1536 0x20 160 0x35 6144 0x0B 1920 0x21 192 0x36 7168 0x0C 2304 0x22 224 0x37 8192 0x0D 2560 0x23 256 0x38 10240 0x0E 3072 0x24 320 0x39 12288 0x0F 3840 0x25 384 0x3A 14336 0x10 4608 0x26 448 0x3B 16384 0x11 5120 0x27 512 0x3C 20480 0x12 6144 0x28 640 0x3D 24576 0x13 7680 0x29 768 0x3E 28672 0x14 9216 0x2A 896 0x3F 32768 0x15 10240 Description 11.3.1.3 I2C Control Register (I2CCR) Figure 11-4 shows the I2C control register. 0 1 2 3 4 5 6 7 R W Reset Offset MEN MIEN MSTA MTX TXAK RSTA BCST 0000_0000 0x0_3008 Figure 11-4. I2C Control Register (I2CCR) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 11-6 Freescale Semiconductor Memory Map/Register Definition Table 11-6 describes the bit settings of the I2CCR . Table 11-6. I2CCR Field Descriptions Bits Name 0 Description MEN Module enable. This bit controls the software reset of the I2C module. 0 The module is reset and disabled. When low, the interface is held in reset but the registers can still be accessed. 1 The I2C module is enabled. This bit must be set before any other control register bits have any effect. All I2C registers for slave receive or master START can be initialized before setting this bit. MIEN Module interrupt enable 0 Interrupts from the I2C module are disabled. This does not clear any pending interrupt conditions. 1 Interrupts from the I2C module are enabled. An interrupt occurs provided I2CSR[MIF] is also set. MSTA Master/slave mode START 0 When this bit is changed from one to zero, a STOP condition is generated and the mode changes from master to slave. 1 Cleared without generating a STOP condition when the master loses arbitration. When this bit is changed from zero to one, a START condition is generated on the bus, and master mode is selected. MTX Transmit/receive mode select. This bit selects the direction of the master and slave transfers. When configured as a slave, this bit should be set by software according to I2CSR[SRW]. In master mode, the bit should be set according to the type of transfer required. Therefore, for address cycles, this bit will always be high. The MTX bit is cleared when the master loses arbitration. 0 Receive mode 1 Transmit mode 1 2 3 4 TXAK Transfer acknowledge. This bit specifies the value driven onto the SDA line during acknowledge cycles for both master and slave receivers. The value of this bit only applies when the I2C module is configured as a receiver, not a transmitter. It also does not apply to address cycles; when the device is addressed as a slave, an acknowledge is always sent. 0 An acknowledge signal (low value on SDA) is sent out to the bus at the 9th clock bit after receiving one byte of data. 1 No acknowledge signal response (high value on SDA) is sent. RSTA Repeat START. Setting this bit always generates a repeated START condition on the bus, provides the device with the current bus master. Attempting a repeated START at the wrong time (or if the bus is owned by another master), results in loss of arbitration. Note that this bit is not readable, which means if a read is performed to I2CCR[RSTA], a zero value will be returned. 0 No START condition is generated 1 Generates repeat START condition — Reserved 5 6 7 BCST Broadcast 0 Disables the broadcast accept capability 1 Enables the I2C to accept broadcast messages at address zero 11.3.1.4 I2C Status Register (I2CSR) The status register, shown in Figure 11-5, is read only with the exception of the MIF and MAL bits, which can be cleared by software. The MCF and RXAK bits are set at reset; all other I2CSR bits are cleared on reset. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 11-7 I2C Interface 0 1 2 3 4 5 6 7 R W Reset Offset MCF MAAS MBB MAL BCSTM SRW MIF RXAK 1000_0001 0x0_300C Figure 11-5. I2C Status Register (I2CSR) Table 11-7 describes the bit settings of the I2CSR. Table 11-7. I2CSR Field Descriptions Bits 0 Name MCF Description Data transfer. When one byte of data is transferred, the bit is cleared. It is set by the falling edge of the 9th clock of a byte transfer. 0 Byte transfer in progress. MCF is cleared under the following conditions: •When I2CSR is read in receive mode or •When I2CDR is written in transmit mode 1 Byte transfer is completed Addressed as a slave. When the value in I2CDR matches with the calling address, this bit is set. The processor is interrupted, if I2CCR[MIEN] is set. Next, the processor must check the SRW bit and set I2CCR[MTX] accordingly. Writing to the I2CCR automatically clears this bit. 0 Not addressed as a slave 1 Addressed as a slave Bus busy. Indicates the status of the bus. When a START condition is detected, MBB is set. If a STOP condition is detected, it is cleared. 0 I2C bus is idle 1 I2C bus is busy Arbitration lost. Automatically set when the arbitration procedure is lost. Note that the device does not automatically retry a failed transfer attempt. 0 Arbitration is not lost. Can only be cleared by software 1 Arbitration is lost 1 MAAS 2 MBB 3 MAL 4 BCSTM Broadcast match 0 There has not been a broadcast match. 1 The calling address matches with the broadcast address instead of the programmed slave address. This will also be set if this I2C drives an address of all 0s and broadcast mode is enabled. SRW Slave read/write. When MAAS is set, SRW indicates the value of the R/W command bit of the calling address, which is sent from the master. 0 Slave receive, master writing to slave 1 Slave transmit, master reading from slave. This bit is valid only when both of the following conditions are true: •A complete transfer occurred and no other transfers have been initiated. •The I2C interface is configured as a slave and has an address match. By checking this bit, the processor can select slave transmit/receive mode according to the command of the master. 5 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 11-8 Freescale Semiconductor Memory Map/Register Definition Table 11-7. I2CSR Field Descriptions (continued) Bits 6 Name MIF Description Module interrupt. The MIF bit is set when an interrupt is pending, causing a processor interrupt request (provided I2CCR[MIEN] is set). 0 No interrupt is pending. Can be cleared only by software. 1 Interrupt is pending. MIF is set when one of the following events occurs: •One byte of data is transferred (set at the falling edge of the 9th clock). •The value in I2CADR matches with the calling address in slave-receive mode. •Arbitration is lost. Received acknowledge. The value of SDA during the reception of acknowledge bit of a bus cycle. If the received acknowledge bit (RXAK) is low, it indicates that an acknowledge signal has been received after the completion of eight bits of data transmission on the bus. If RXAK is high, it means no acknowledge signal has been detected at the 9th clock. 0 Acknowledge received 1 No acknowledge received 7 RXAK 11.3.1.5 I2C Data Register (I2CDR) The I2C data register is shown in Figure 11-6. 0 7 R DATA W Reset Offset 0000_0000 0x0_3010 Figure 11-6. I2C Data Register (I2CDR) Table 11-8 shows the bit descriptions for I2CDR. Table 11-8. I2CDR Field Description Bits 0–7 Name DATA Description Transmission starts when an address and the R/W bit are written to the data register and the I2C interface performs as the master. A data transfer is initiated when data is written to the I2CDR. The most significant bit is sent first in both cases. In the master receive mode, reading the data register allows the read to occur, but also allows the I2C module to receive the next byte of data on the I2C interface. In slave mode, the same function is available after it is addressed. Note that the very first read is always a dummy read. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 11-9 I2C Interface 11.3.1.6 Digital Filter Sampling Rate Register (I2CDFSRR) The digital filter sampling rate register (I2CDFSRR) is shown in Figure 11-7. 0 1 2 7 R W Reset Offset DFSR 0001_0000 0x0_3014 Figure 11-7. I2C Digital Filter Sampling Rate Register (I2CDFSRR) Table 11-9 shows the field descriptions for I2CDFSRR. Table 11-9. I2CDFSRR Field Descriptions Bits Name 0–1 — Reserved Description 2–7 DFSR Digital filter sampling rate. To assist in filtering out signal noise, the sample rate is programmed. This field is used to prescale the frequency at which the digital filter takes samples from the I2C bus. The resulting sampling rate is calculated by dividing the platform (CCB clock) frequency by the non-zero value of DFSR. If I2CDFSRR is set to zero, the I2C bus sample points default to the reset divisor 0x10. 11.4 Functional Description The I2C unit always performs as a slave receiver as a default, unless explicitly programmed to be a master or slave transmitter. After the boot sequencer has completed (when powered up in boot sequencer mode), the I2C interface will perform as a slave receiver. 11.4.1 Transaction Protocol A standard I2C transfer consists of the following: • • • • START condition Slave target address transmission Data transfer STOP condition Figure 11-8 shows the interaction of these four parts with the calling address, data byte, and new calling address components of the I2C protocol. The details of the protocol are described in the following subsections. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 11-10 Freescale Semiconductor Functional Description SCL 1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9 SDA A0 A1 A2 A3 A4 A5 A6 R/W Calling address XX XX D0 D1 D2 D3 D4 D5 D6 D7 Data byte START Read/ write Ack No ack STOP SCL 1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9 SDA A0 A1 A2 A3 A4 Calling address A5 A6 R/W A0 A1 A2 A3 A4 A5 A6 R/W New calling address START Read/ write Ack Repeat START No ack STOP Figure 11-8. I2C Interface Transaction Protocol 11.4.1.1 START Condition When the I2C bus is not engaged (both SDA and SCL lines are at logic high), a master can initiate a transfer by sending a START condition. As shown in Figure 11-8, a START condition is defined as a high-to-low transition of SDA while SCL is high. This condition denotes the beginning of a new data transfer. Each data transfer can contain several bytes and awakens all slaves. The START condition is initiated by a software write that sets I2CCR[MSTA]. 11.4.1.2 Slave Address Transmission The first byte of data is transferred by the master immediately after the START condition is the slave address. This is a seven-bit calling address followed by a R/W bit, which indicates the direction of the data being transferred to the slave. Each slave in the system has a unique address. In addition, when the I2C module is operating as a master, it must not transmit an address that is the same as its slave address. An I2C device cannot be master and slave at the same time. Only the slave with a calling address that matches the one transmitted by the master responds by returning an acknowledge bit (pulling the SDA signal low at the 9th clock) as shown in Figure 11-8. If no slave acknowledges the address, the master should generate a STOP condition or a repeated START condition. When slave addressing is successful (and SCL returns to zero), the data transfer can proceed on a byte-to-byte basis in the direction specified by the R/W bit sent by the calling master. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 11-11 I2C Interface The I2C module responds to a general call (broadcast) command when I2CCR[BCST] is set. A broadcast address is always zero; however the I2C module will not check the R/W bit. The second byte of the broadcast message is the master address. Because the second byte is automatically acknowledged by hardware, the receiver device software must verify that the broadcast message is intended for itself by reading the second byte of the message. If the master address is for another receiver device and the third byte is a write command, software can ignore the third byte during the broadcast. If the master address is for another receiver device and the third byte is a read command, software must write 0xFF to I2CDR with I2CCR[TXAK] = 1, so that it does not interfere with the data written from the addressed device. Each data byte is 8 bits long. Data bits can be changed only while SCL is low and must be held stable while SCL is high, as shown in Figure 11-8. There is one clock pulse on SCL for each data bit, and the most significant bit (msb) is transmitted first. Each byte of data must be followed by an acknowledge bit, which is signaled from the receiving device by pulling the SDA line low at the 9th clock. Therefore, one complete data byte transfer takes 9 clock pulses. Several bytes can be transferred during a data transfer session. If the slave receiver does not acknowledge the master, the SDA line must be left high by the slave. The master can then generate a stop condition to abort the data transfer or a START condition (repeated START) to begin a new calling. If the master receiver does not acknowledge the slave transmitter after a byte of transmission, the slave interprets that the end-of-data has been reached. Then the slave releases the SDA line for the master to generate a STOP or a START condition. 11.4.1.3 Repeated START Condition Figure 11-8 shows a repeated START condition, which is generated without a STOP condition that can terminate the previous transfer. The master uses this method to communicate with another slave or with the same slave in a different mode (transmit/receive mode) without releasing the bus. 11.4.1.4 STOP Condition The master can terminate the transfer by generating a STOP condition to free the bus. A STOP condition is defined as a low-to-high transition of the SDA signal while SCL is high. For more information, see Figure 11-8. Note that a master can generate a STOP even if the slave has transmitted an acknowledge bit, at which point the slave must release the bus. The STOP condition is initiated by a software write that clears I2CCR[MSTA]. As described in Section 11.4.1.3, “Repeated START Condition,” the master can generate a START condition followed by a calling address without generating a STOP condition for the previous transfer. This is called a repeated START condition. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 11-12 Freescale Semiconductor Functional Description 11.4.1.5 Protocol Implementation Details The following sections give details of how aspects of the protocol are implemented in this I2C module. 11.4.1.5.1 Transaction Monitoring—Implementation Details The different conditions of the I2C data transfers are monitored as follows: • START conditions are detected when an SDA fall occurs while SCL is high. • STOP conditions are detected when an SDA rise occurs while SCL is high. • Data transfers in progress are canceled when a STOP condition is detected or if there is a slave address mismatch. Cancellation of data transactions resets the clock module. • The bus is detected to be busy upon the detection of a START condition, and idle upon the detection of a STOP condition. 11.4.1.5.2 Control Transfer—Implementation Details The I2C module contains logic that controls the output to the serial data (SDA) and serial clock (SCL) lines of the I2C. The SCL output is pulled low as determined by the internal clock generated in the clock module. The SDA output can only change at the midpoint of a low cycle of the SCL, unless it is performing a START, STOP, or restart condition. Otherwise, the SDA output is held constant. The SDA signal is pulled low when one or more of the following conditions are true in either master or slave mode: • Master mode — Data bit (transmit) — Ack bit (receive) — START condition — STOP condition — Restart condition Slave mode — Acknowledging address match — Data bit (transmit) — Ack bit (receive) • MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 11-13 I2C Interface The SCL signal corresponds to the internal SCL signal when one or more of the following conditions are true in either master or slave mode: • Master mode — Bus owner — Lost arbitration — START condition — STOP condition — Restart condition begin — Restart condition end Slave mode — Address cycle — Transmit cycle — Ack cycle • 11.4.1.6 Address Compare—Implementation Details Address compare block determines if a slave has been properly addressed, either by its slave address or by the general broadcast address (which addresses all slaves). The three performed address comparisons are described as follows: • • • Whether a broadcast message has been received, to update the I2CSR Whether the module has been addressed as a slave, to update the I2CSR and to generate an interrupt If the address transmitted by the current master matches the general broadcast address 11.4.2 Arbitration Procedure The I2C interface is a true multiple-master bus that allows more than one master device to be connected on it. If two or more masters simultaneously try to control the bus, each master’s clock synchronization procedure (including the I2C module) determines the bus clock—the low period is equal to the longest clock low period and the high is equal to the shortest one among the masters. A bus master loses arbitration if it transmits a logic 1 on SDA while another master transmits a logic 0. The losing masters immediately switch to slave-receive mode and stop driving the SDA line. In this case, the transition from master to slave mode does not generate a STOP condition. Meanwhile, the I2C unit sets the I2CSR[MAL] status bit to indicate the loss of arbitration and, as a slave, services the transaction if it is directed to itself. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 11-14 Freescale Semiconductor Functional Description If the I2C module is enabled in the middle of an ongoing byte transfer, the interface behaves as follows: • • Slave mode—The I2C module ignores the current transfer on the bus and starts operating whenever a subsequent START condition is detected. Master mode—The I2C module cannot tell whether the bus is busy; therefore, if a START condition is initiated, the current bus cycle can be corrupted. This ultimately results in the current bus master of the I2C interface losing arbitration, after which bus operations return to normal. 11.4.2.1 Arbitration Control The arbitration control block controls the arbitration procedure of the master mode. A loss of arbitration occurs whenever the master detects a 0 on the external SDA line while attempting to drive a 1, tries to generate a START or restart at an inappropriate time, or detects an unexpected STOP request on the line. In master mode, arbitration by the master is lost (and I2CSR[MAL] is set) under the following conditions: • • • • • • SDA samples low when the master drives high during an address or data-transmit cycle (transmit). SDA samples low when the master drives high during a data-receive cycle of the acknowledge (Ack) bit (receive). A START condition is attempted when the bus is busy. A repeated START condition is requested in slave mode. A start condition is attempted when the requesting device is not the bus owner Unexpected STOP condition detected Note that the I2C module does not automatically retry a failed transfer attempt. 11.4.3 Handshaking The clock synchronization mechanism can be used as a handshake in data transfer. Slave devices can hold SCL low after completion of a 1-byte transfer (9 bits). In such cases, it halts the bus clock and forces the master clock into wait states until the slave releases the SCL line. 11.4.4 Clock Control The clock control block handles requests from the clock signal for transferring and controlling data for multiple tasks. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 11-15 I2C Interface A 9-cycle data transfer clock is requested for the following conditions: • Master mode — Transmit slave address after START condition — Transmit slave address after restart condition — Transmit data — Receive data • Slave mode — Transmit data — Receive data — Receive slave address after START or restart condition 11.4.4.1 Clock Synchronization Due to the wire AND logic on the SCL line, a high-to-low transition on the SCL line affects all devices connected on the bus. The devices begin counting their low period when the master drives the SCL line low. After a device has driven SCL low, it holds the SCL line low until the clock high state is reached. However, the change of low-to-high in a device clock may not change the state of the SCL line if another device is still within its low period. Therefore, the synchronized clock signal, SCL, is held low by the device with the longest low period. Devices with shorter low periods enter a high wait state during this time. When all devices concerned have counted off their low period, the synchronized SCL line is released and pulled high. Then there is no difference between the devices’ clocks and the state of the SCL line, and all the devices begin counting their high periods. The first device to complete its high period pulls the SCL line low again. 11.4.4.2 Input Synchronization and Digital Filter The following sections describes the synchronizing of the input signals, and the filtering of the SCL and SDA lines in detail. 11.4.4.2.1 Input Signal Synchronization The input synchronization block synchronizes the input SCL and SDA signals to the system clock and detects transitions of these signals. 11.4.4.2.2 Filtering of SCL and SDA Lines The SCL and SDA inputs are filtered to eliminate noise. Three consecutive samples of the SCL and SDA lines are compared to a pre-determined sampling rate. If they are all high, the output of the filter is high. If they are all low, the output is low. If they are any combination of highs and lows, the output is whatever the value of the line was in the previous clock cycle. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 11-16 Freescale Semiconductor Functional Description The sampling rate is equal to a binary value stored in the frequency register I2CDFSRR. The duration of the sampling cycle is controlled by a down counter. This allows a software write to the frequency register to control the filtered sampling rate. 11.4.4.3 Clock Stretching Slaves can use the clock synchronization mechanism to slow down the transfer bit rate. After the master has driven the SCL line low, the slave can drive SCL low for the required period and then release it. If the slave SCL low period is greater than the master SCL low period, then the resulting SCL bus signal low period is stretched. 11.4.5 Boot Sequencer Mode If boot sequencer mode is selected on POR (by the settings on the LGPL3 and LGPL5 reset configuration signals, as described in Section 4.4.3.6, “Boot Sequencer Configuration”), the I2C module communicates with one or more EEPROMs through the I2C interface. The EEPROM(s) can be programmed to initialize one or more configuration registers of this integrated device. The boot sequencer mode also supports an extension of the standard I2C interface that uses more address bits to allow for EEPROM devices that have more than 256 bytes, and this extended addressing mode is selectable during POR with a different encoding on the LGPL3 and LGPL5 reset configuration signals (see Section 4.4.3.6, “Boot Sequencer Configuration”). In this mode, only one EEPROM device may be used, and the maximum number of registers is limited by the size of the EEPROM. If the standard I2C interface is used, the I2C module addresses the first EEPROM, and reads 256 bytes. Then it issues a repeated start and addresses the next EEPROM address. This sequence continues until the CONT bit is cleared. If the last register is not detected before wrapping back to the first address, an error condition is detected. In other words, if the CONT bit for not cleared on the final 7 bytes, an error condition is detected, causing the device to hang and the HRESET_REQ signal to assert externally. The I2C module continues to read from the EEPROM as long as the continue (CONT) bit is set in the EEPROM. The CONT bit resides in the address/attributes field that is transferred from the EEPROM, as described in Section 11.4.5.1, “EEPROM Calling Address.” There should be no other I2C traffic when the boot sequencer is active. Note that as described in Section 4.4.3.6, “Boot Sequencer Configuration,” the default value for the LGPL3 and LGPL5 reset configuration pins is 0b11, which corresponds to the I2C boot sequencer being disabled at power-up. 11.4.5.1 EEPROM Calling Address The MPC8540 uses 0b101_0000 for the EEPROM calling address. The first EEPROM to be addressed must be programmed to respond to this address, or an error is generated. If more EEPROMs are used, they are addressed in sequential order. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 11-17 I2C Interface 11.4.5.2 EEPROM Data Format The I2C module expects that a particular data format be used for data in the EEPROM. A preamble should be the first 3 bytes programmed into the EEPROM. It should have a value of 0xAA55AA. The I2C module checks to ensure that this preamble is correctly detected before proceeding further. Following the preamble, there should be a series of configuration registers (known as register preloads) programmed into the EEPROM. Each configuration register should be programmed according to a particular format, as shown in Figure 11-9. The first 3 bytes hold the attributes and address offset, as follows. The attributes contained are alternate configuration space (ACS), byte enables, and continue (CONT). The boot sequencer expects the address offset to be a 32-bit (word) offset, that is, the 2 low-order bits are not included in the boot sequencer command. For example, to access LAWBAR0 (byte offset of 0x00C08), the boot sequencer ADDR[0:17] should be set to 0x00302. After the first 3 bytes, 4 bytes of data should hold the desired value of the configuration register, regardless of the size of the transaction. Byte enables should be asserted for any byte that will be written to the configuration register, and they should be asserted contiguously, creating a 1-, 2-, or 4-byte write to a register. The boot sequencer assumes that a big-endian address is stored in the EEPROM. In addition, byte enable bit 0 (bit 1 of the byte) corresponds to the most-significant byte of data (data[0:7]), and byte enable bit 3 (bit 4 of the byte) corresponds to the LSB of data (data[24:31]). By setting ACS, an alternate configuration space address is prepended to the write request from the boot sequencer. Otherwise, CCSRBAR is prepended to the EEPROM address. If CONT is cleared, the first 3 bytes, including ACS, the byte enables, and the address, must also be cleared. Also, the data contains the final cyclic redundancy check (CRC). A CRC-32 algorithm is used to check the integrity of the data. The polynomial used is: 1 + x1 + x2 + x4 + x5 + x7 + x8 + x10 + x11 + x12 + x16 + x22 + x23 + x26 + x32 CRC values are calculated using the above polynomial with a start value of 0xFFFF_FFFF and an XOR with 0x0000_0000. The CRC should cover all bytes stored in the EEPROM prior to the CRC. This includes the preamble, all register preloads, and the first 3 bytes of the last 7-byte preload (which should be all zeros). If a preamble or CRC fail is detected, the device hangs and the external HRESET_REQ signal asserts. If there is a preamble fail, the boot sequencer may continue to pull I2C pins low until a hard reset occurs. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 11-18 Freescale Semiconductor Functional Description 0 1 4 5 6 7 ACS BYTE_EN ADDR[2–9] CONT ADDR[0–1] ADDR[10–17] DATA[0–7] DATA[8–15] DATA[16–23] DATA[24–31] Figure 11-9. EEPROM Data Format for One Register Preload Command Figure 11-10 shows an example of the EEPROM contents, including the preamble, data format, and CRC. 0 1 2 3 4 5 6 7 1 0 1 ACS 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 1 0 1 0 1 0 Preamble BYTE_EN ADDR[2–9] ADDR[10–17] DATA[0–7] DATA[8–15] DATA[16–23] DATA[24–31] ADDR[0–1] First Configuration Preload Command ACS BYTE_EN ADDR[2–9] ADDR[10–17] DATA[0–7] DATA[8–15] DATA[16–23] DATA[24–31] 1 ADDR[0–1] Second Configuration Preload Command . . . Figure 11-10. EEPROM Contents MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 11-19 I2C Interface ACS BYTE_EN ADDR[2–9] ADDR[10–17] DATA[0–7] DATA[8–15] DATA[16–23] DATA[24–31] 1 ADDR[0–1] Last Configuration Preload Command 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 End Command CRC[0–7] CRC[8–15] CRC[16–23] CRC[24–31] Cyclic Redundancy Check Figure 11-10. EEPROM Contents (continued) 11.5 Initialization/Application Information This section describes some programming guidelines recommended for the I2C interface. Figure 11-11 is a recommended flowchart for I2C interrupt service routines. The I2C registers in this chapter are shown in big-endian format. If the system is in little-endian mode, software must swap the bytes appropriately. This appropriate byte swapping is needed as I2C registers are byte registers. Also, an msync assembly instruction must be executed after each I2C register read/write access to guarantee in-order execution. The I2C controller does not guarantee its recovery from all illegal I2C bus activity. In addition, a malfunctioning device may hold the bus captive. A good programming practice is for software to rely on a watchdog timer to help recover from I2C bus hangs. The recovery routine should also handle the case when the status bits returned after an interrupt are not consistent with what was expected due to illegal I2C bus protocol behavior. 11.5.1 Initialization Sequence A hard reset initializes all the I2C registers to their default states. The following initialization sequence initializes the I2C unit: 1. All I2C registers must be located in a cache-inhibited page. 2. Update I2CFDR[FDR] and select the required division ratio to obtain the SCL frequency from the CCB (platform) clock. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 11-20 Freescale Semiconductor Initialization/Application Information 3. Update I2CADR to define the slave address for this device. 4. Modify I2CCR to select master/slave mode, transmit/receive mode, and interrupt-enable or disable. 5. Set the I2CCR[MEN] to enable the I2C interface. 11.5.2 Generation of START After initialization, the following sequence can be used to generate START: 1. If the device is connected to a multimaster I2C system, test the state of I2CSR[MBB] to check whether the serial bus is free (I2CSR[MBB] = 0) before switching to master mode. 2. Select master mode (set I2CCR[MSTA]) to transmit serial data and select transmit mode (set I2CCR[MTX]) for the address cycle. 3. Write the slave address being called into I2CDR. The data written to I2CDR[0–6] comprises the slave calling address. I2CCR[MTX] indicates the direction of transfer (transmit/receive) required from the slave. The scenario above assumes that the I2C interrupt bit (I2CSR[MIF]) is cleared. If MIF is set at any time, an I2C interrupt is generated (provided interrupt reporting is enabled with I2CCR[MIEN] =1) so that the I2C interrupt handler can handle the interrupt. 11.5.3 Post-Transfer Software Response Transmission or reception of a byte automatically sets the data transferring bit (I2CSR[MCF]), which indicates that one byte has been transferred. The I2C interrupt bit (I2CSR[MIF]) is also set and an interrupt is generated to the processor if the interrupt function is enabled during the initialization sequence (I2CCR[MIEN] is set). In the interrupt handler, software must take the following steps: 1. Clear I2CSR[MIF] 2. Read the contents of the I2C data register (I2CDR) in receive mode or write to I2CDR in transmit mode. Note that this causes I2CSR[MCF] to be cleared. See Section 11.5.8, “Interrupt Service Routine Flowchart.” When an interrupt occurs at the end of the address cycle, the master remains in transmit mode. If master receive mode is required, I2CCR[MTX] must be toggled at this stage. See Section 11.5.8, “Interrupt Service Routine Flowchart.” If the interrupt function is disabled, software can service the I2CDR in the main program by monitoring I2CSR[MIF]. In this case, I2CSR[MIF] must be polled rather than I2CSR[MCF] because MCF behaves differently when arbitration is lost. Note that interrupt or other bus conditions may be detected before the I2C signals have time to settle. Thus, when polling MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 11-21 I2C Interface I2CSR[MIF] (or any other I2CSR bits), software delays may be needed in order to give the I2C signals sufficient time to settle. During slave-mode address cycles (I2CSR[MAAS] is set), I2CSR[SRW] should be read to determine the direction of the subsequent transfer and I2CCR[MTX] should be programmed accordingly. For slave-mode data cycles (MAAS is cleared), I2CSR[SRW] is not valid and I2CCR[MTX] must be read to determine the direction of the current transfer. See Section 11.5.8, “Interrupt Service Routine Flowchart,” for more details. 11.5.4 Generation of STOP A data transfer ends with a STOP condition generated by the master device. A master transmitter can generate a STOP condition after all the data has been transmitted. If a master receiver wants to terminate a data transfer, it must inform the slave transmitter by not acknowledging the last byte of data (by setting the transmit acknowledge bit (I2CCR[TXAK])) before reading the next-to-last byte of data. At this time, the next-to-last byte of data has already been transferred on the I2C interface, so the last byte will not receive the data acknowledge (because I2CCR[TXAK] is set). For 1-byte transfers, a dummy read should be performed by the interrupt service routine (see Section 11.5.8, “Interrupt Service Routine Flowchart”). Before the interrupt service routine reads the last byte of data, a STOP condition must first be generated. The I2C controller automatically generates a STOP if I2CCR[TXAK] is set. Therefore, I2CCR[TXAK] must be set before allowing the I2C module to receive the last data byte on the I2C bus. Eventually, I2CCR[TXAK] needs to be cleared again for subsequent I2C transactions. This can be accomplished when setting up the I2CCR for the next transfer. 11.5.5 Generation of Repeated START At the end of a data transfer, if the master still wants to communicate on the bus, it can generate another START condition followed by another slave address without first generating a STOP condition. This is accomplished by setting I2CCR[RSTA]. 11.5.6 Generation of SCL When SDA Low It is sometimes necessary to force the I2C module to become the I2C bus master out of reset and drive SCL (even though SDA may already be driven, which indicates that the bus is busy). This can occur when a system reset does not cause all I2C devices to be reset. Thus, SDA can be driven low by another I2C device while this I2C module is coming out of reset and will stay low indefinitely. The following procedure can be used to force this I2C module to generate SCL so that the device driving SDA can finish its transaction: 1. Disable the I2C module and set the master bit by setting I2CCR to 0x20 2. Enable the I2C module by setting I2CCR to 0xA0 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 11-22 Freescale Semiconductor Initialization/Application Information 3. Read the I2CDR 4. Return the I2C module to slave mode by setting I2CCR to 0x80 11.5.7 Slave Mode Interrupt Service Routine In the slave interrupt service routine, the module addressed as a slave should be tested to check if a calling of its own address has been received. If I2CSR[MAAS] is set, software should set the transmit/receive mode select bit (I2CCR[MTX]) according to the R/W command bit (I2CSR[SRW]). Writing to I2CCR clears MAAS automatically. MAAS is read as set only in the interrupt handler at the end of that address cycle where an address match occurred; interrupts resulting from subsequent data transfers clear MAAS. A data transfer can then be initiated by writing to I2CDR for slave transmits or dummy reading from I2CDR in slave-receive mode. The slave drives SCL low between byte transfers. SCL is released when the I2CDR is accessed in the required mode. 11.5.7.1 Slave Transmitter and Received Acknowledge In the slave transmitter routine, the received acknowledge bit (I2CSR[RXAK]) must be tested before sending the next byte of data. The master signals an end-of-data by not acknowledging the data transfer from the slave. When no acknowledge is received (I2CSR[RXAK] is set), the slave transmitter interrupt routine must clear I2CCR[MTX] to switch the slave from transmitter to receiver mode. A dummy read of I2CDR then releases SCL so that the master can generate a STOP condition. See Section 11.5.8, “Interrupt Service Routine Flowchart.” 11.5.7.2 • • • Loss of Arbitration and Forcing of Slave Mode When a master loses arbitration the following conditions all occur: I2CSR[MAL] is set I2CCR[MSTA] is cleared (changing the master to slave mode) An interrupt occurs (if enabled) at the falling edge of the 9th clock of this transfer Thus, the slave interrupt service routine should first test I2CSR[MAL] and software should clear it if it is set. See Section 11.4.2.1, “Arbitration Control,” for more information. 11.5.8 Interrupt Service Routine Flowchart Figure 11-11 shows an example algorithm for an I2C interrupt service routine. Deviation from the flowchart may result in unpredictable I2C bus behavior. However, in the slave receive mode (not shown), the interrupt service routine may need to set I2CCR[TXAK] when the next-to-last byte is to be accepted. It is recommended that an msync instruction follow each I2C register read or write to guarantee in-order instruction execution. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 11-23 I2C Interface Clear I2CSR[MIF] == 1 I2CCR[MTX] Master Xmit == 1 Y Last byte == 1 N Clear I2CSR[MAL] I2CSR[RXAK] == 0 Generate STOP N End of address phase for master receive mode? Y Y Only one byte to receive? N Slave Xmit == 1 I2CSR[RXAK] == 0 Clear I2CCR[MTX] Clear I2CCR[MTX] Read I2CDR (dummy read) Read I2CDR (dummy read) Write next byte to I2CDR N All done Y I2CSR[MAAS] == 0 EOI Slave Data Cycle == 1 B == 0 A == 1 I2CSR[MAL] == 0 == 1 I2CCR[MSTA] == 0 I2CSR[MAAS] == 0 B Write next byte to I2CDR == 1 I2CCR[MTX] == 0 Slave Received Read I2CDR and store Set I2CCR[TXAK] Set I2CCR[TXAK] EOI A Master Rcv N Last byte N Generate STOP Y EOI B Slave Addr. Phase == 1 I2CSR[SRW] == 0 Y Next-to-last byte Set I2CCR[TXAK] Set I2CCR[MTX] Clear I2CCR[MTX] Dummy read Write I2CDR Read I2CDR and store EOI EOI Figure 11-11. Example I2C Interrupt Service Routine Flowchart MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 11-24 Freescale Semiconductor Chapter 12 DUART This chapter describes the dual universal asynchronous receiver/transmitters (DUART) of the MPC8540. It describes the functional operation, the DUART initialization sequence, and the programming details for the DUART registers and features. 12.1 Overview The DUART of the MPC8540 consists of two universal asynchronous receiver/transmitters (UARTs). The UARTs act independently; all references to UART refer to one of these receiver/transmitters. Each UART is clocked by the core complex bus (CCB) clock. The DUART programming model is compatible with the PC16552D. The UART interface is point to point, meaning that only two UART devices are attached to the connecting signals. As shown in Figure 12-1, each UART module consists of the following: • • • • Receive and transmit buffers Clear to send (CTS) input port and request to send (RTS) output port for data flow control 16-bit counter for baud rate generation Interrupt control logic MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 12-1 DUART Data SIN Receive Buffer Address Bus Control Logic Control UART Module Internal Bus SOUT Transmit Buffer int Interrupt Control Input Port CTS Output Port RTS HRESET CCB_clk 16-Bit Counter/ Baud Rate Generator Figure 12-1. UART Block Diagram 12.1.1 Features The DUART includes these distinctive features: • • • • • • • • • • • Full-duplex operation Programming model compatible with the original PC16450 UART and the PC16550D (an improved version of the PC16450 that also operates in FIFO mode) PC16450 register reset values FIFO mode for both transmitter and receiver, providing 16-byte FIFOs Serial data encapsulation and decapsulation with standard asynchronous communication bits (START, STOP, and parity) Maskable transmit, receive, line status, and MODEM status interrupts Software-programmable baud generators that divide the CCB clock by 1 to (216 – 1) and generate a 16x clock for the transmitter and receiver engines Clear to send (CTS) and ready to send (RTS) MODEM control functions Software-selectable serial interface data format (data length, parity, 1/1.5/2 STOP bit, baud rate) Line and MODEM status registers Line-break detection and generation MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 12-2 Freescale Semiconductor External Signal Descriptions • • • Internal diagnostic support, local loopback, and break functions Prioritized interrupt reporting Overrun, parity, and framing error detection 12.1.2 Modes of Operation The communication channel provides a full-duplex asynchronous receiver and transmitter using an operating frequency derived from the CCB clock. The transmitter accepts parallel data from a write to the transmitter holding register (UTHR). In FIFO mode, the data is placed directly into an internal transmitter shift register of the transmitter FIFO. The transmitter converts the data to a serial bit stream inserting the appropriate start, stop, and optional parity bits. Finally, it outputs a composite serial data stream on the channel transmitter serial data output signal (SOUT). The transmitter status may be polled or interrupt driven. The receiver accepts serial data bits on the channel receiver serial data input signal (SIN), converts it to parallel format, checks for a start bit, parity (if any), stop bits, and transfers the assembled character (with start, stop, parity bits removed) from the receiver buffer (or FIFO) in response to a read of the UART’s receiver buffer register (URBR). The receiver status may be polled or interrupt driven. 12.2 External Signal Descriptions This section contains a signal overview and detailed signal descriptions. 12.2.1 Signal Overview Table 12-1 summarizes the DUART signals. Note that although the actual device signal names are prepended with the UART_ prefix as shown in the table, the functional (abbreviated) signal names are often used throughout this chapter. Table 12-1. DUART Signal Overview Signal Name UART_SIN[0:1] UART_SOUT[0:1] UART_CTS[0:1] UART_RTS[0:1] I/O I O I O Pins 2 2 2 2 Reset Value 1 1 1 1 State Meaning Serial in data UART0 and UART1 Serial out data UART0 and UART1 Clear to send UART0 and UART1 Request to send UART0 and UART1 12.2.2 Detailed Signal Descriptions The eight DUART signals are described in detail in Table 12-2. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 12-3 DUART Table 12-2. DUART Signals—Detailed Signal Descriptions Signal UART_SIN[0:1] I/O I Description Serial data in. Data is received on the receivers of UART0 and UART1 through its respective serial data input signal, with the least-significant bit received first. State Asserted/Negated—Represents the data being received on the UART interface. Meaning Timing Assertion/Negation—An internal logic sample signal, rxcnt, uses the frequency of the baud-rate generator to sample the data on SIN. UART_SOUT[0:1] O Serial data out. The serial data output signals for the UART0 and UART1 are set ('mark' condition) when the transmitter is disabled, idle, or operating in the local loop-back mode. Data is shifted out on these signals, with the least significant bit transmitted first. State Asserted/Negated—Represents the data being transmitted on the respective UART Meaning interface. Timing Assertion/Negation— An internal logic sample signal, rxcnt, uses the frequency of the baud-rate generator to update and drive the data on SOUT. UART_CTS[0:1] I Clear to send. These active-low inputs are the clear-to-send inputs. They are connected to the respective RTS outputs of the other UART devices on the bus. They can be programmed to generate an interrupt on change-of-state of the signal. State Asserted/Negated—Represent the clear to send condition for their respective UART. Meaning Timing Assertion/Negation—Sampled at the rising edge of every CCB clock. UART_RTS[0:1] O Request to send. UART_RTS x are active-low output signals that can be programmed to be automatically negated and asserted by either the receiver or transmitter. When connected to the clear-to-send (CTS ) input of a transmitter, this signal can be used to control serial data flow. State Asserted/Negated—Represents the data being transmitted on the respective UART Meaning interface. Timing Assertion/Negation—Updated and driven at the rising edge of every CCB clock. 12.3 Memory Map/Register Definition There are two complete sets of DUART registers (one for UART0 and one for UART1). The two UARTs on the device are identical, except that the registers for UART0 are located at offsets 0x0_4500 (local), and the registers for UART1 are located at offsets 0x0_4600 (local). Throughout this chapter, the registers are described by a singular acronym: for example, LCR represents the line control register for either UART0 or UART1. The 14 registers in each UART interface are used for configuration, control, and status. The divisor latch access bit, ULCR[DLAB], is used to access the divisor latch least- and most-significant bit registers and the alternate function register. Refer to Section 12.3.1.7, “Line Control Registers (ULCR0, ULCR1),” for more information on ULCR[DLAB]. All the DUART registers are one byte wide. Reads and writes to these registers must be byte-wide operations. Table 12-3 provides a register summary with references to the section and page that MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 12-4 Freescale Semiconductor Memory Map/Register Definition contains detailed information about each register. Undefined byte address spaces within offset 0x000–0xFFF are reserved. Table 12-3. DUART Register Summary Offset 0x0_4500 0x0_4500 0x0_4500 0x0_4501 0x0_4501 0x0_4502 0x0_4502 0x0_4502 0x0_4503 0x0_4504 0x0_4505 0x0_4506 0x0_4507 0x0_4510 0x0_4600 0x0_4600 0x0_4600 0x0_4601 0x0_4601 0x0_4602 0x0_4602 0x0_4602 0x0_4603 0x0_4604 0x0_4605 0x0_4606 0x0_4607 0x0_4610 Register URBR—ULCR[DLAB] = 0 UART0 receiver buffer register UTHR—ULCR[DLAB] = 0 UART0 transmitter holding register UDLB—ULCR[DLAB] = 1 UART0 divisor least significant byte register UIER—ULCR[DLAB] = 0 UART0 interrupt enable register UDMB—ULCR[DLAB] = 1 UART0 divisor most significant byte register UIIR—ULCR[DLAB] = 0 UART0 interrupt ID register UFCR—ULCR[DLAB] = 0 UART0 FIFO control register UAFR—ULCR[DLAB] = 1 UART0 alternate function register ULCR—ULCR[DLAB] = x UART0 line control register UMCR—ULCR[DLAB] = x UART0 MODEM control register ULSR—ULCR[DLAB] = x UART0 line status register UMSR—ULCR[DLAB] = x UART0 MODEM status register USCR—ULCR[DLAB] = x UART0 scratch register UDSR—ULCR[DLAB] = x UART0 DMA status register URBR—ULCR[DLAB] = 0 UART1 receiver buffer register UTHR—ULCR[DLAB] = 0 UART1 transmitter holding register UDLB—ULCR[DLAB] = 1 UART1 divisor least significant byte register UIER—ULCR[DLAB] = 0 UART1 interrupt enable register UDMB_ULCR[DLAB] = 1 UART1 divisor most significant byte register UIIR—ULCR[DLAB] = 0 UART1 interrupt ID register UFCR—ULCR[DLAB] = 0 UART1 FIFO control register UAFR—ULCR[DLAB] = 1 UART1 alternate function register ULCR—ULCR[DLAB] = x UART1 line control register UMCR—ULCR[DLAB] = x UART1 MODEM control register ULSR—ULCR[DLAB] = x UART1 line status register UMSR—ULCR[DLAB] = x UART1 MODEM status register USCR—ULCR[DLAB] = x UART1 scratch register UDSR—ULCR[DLAB] = x UART1 DMA status register Access Reset R W R/W R/W R/W R W R/W R/W R/W R R R/W R R W R/W R/W R/W R W R/W R/W R/W R R R/W R 0x00 0x00 0x00 0x00 0x00 0x01 0x00 0x00 0x00 0x00 0x60 0x00 0x00 0x01 0x00 0x00 0x00 0x00 0x00 0x01 0x00 0x00 0x00 0x00 0x60 0x00 0x00 0x01 Section/Page 12.3.1.1/12-6 12.3.1.2/12-6 12.3.1.3/12-7 12.3.1.4/12-9 12.3.1.3/12-7 12.3.1.5/12-10 12.3.1.6/12-11 12.3.1.12/12-17 12.3.1.7/12-12 12.3.1.8/12-14 12.3.1.9/12-15 12.3.1.10/12-16 12.3.1.11/12-17 12.3.1.13/12-18 12.3.1.1/12-6 12.3.1.2/12-6 12.3.1.3/12-7 12.3.1.4/12-9 12.3.1.3/12-7 12.3.1.5/12-10 12.3.1.6/12-11 12.3.1.12/12-17 12.3.1.7/12-12 12.3.1.8/12-14 12.3.1.9/12-15 12.3.1.10/12-16 12.3.1.11/12-17 12.3.1.13/12-18 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 12-5 DUART 12.3.1 Register Descriptions The following sections describe the MPC8540 UART0 and UART1 registers. 12.3.1.1 Receiver Buffer Registers (URBR0, URBR1) These registers contain the data received from the transmitter on the UART buses. In FIFO mode, when read, they return the first byte received. For FIFO status information, refer to the UDSR[RXRDY] description. Except for the case when there is an overrun, URBR returns the data in the order it was received from the transmitter. Refer to the ULSR[OE] description, Section 12.3.1.9, “Line Status Registers (ULSR0, ULSR1).” Figure 12-3 shows the receiver buffer registers. Note that these registers have same offset as the UTHRs. Figure 12-2 shows the bits in the URBRs. 0 7 R W Reset Offset DATA 0x00 0x0_4500, 0x0_4600 Figure 12-2. Receiver Buffer Registers (URBR0, URBR1) Table 12-4 describes the fields of URBR. Table 12-4. URBR Field Definition Bits 0–7 Name DATA Description Data received from the transmitter on the UART bus (read only) 12.3.1.2 Transmitter Holding Registers (UTHR0, UTHR1) A write to these 8-bit registers causes the UART devices to transfer 5–8 data bits on the UART bus in the format set up in the ULCR (line control register). In FIFO mode, data written to UTHR is placed into the FIFO. The data written to UTHR is the data sent onto the UART bus, and the first byte written to UTHR will be the first byte onto the bus. UDSR[TXRDY] indicates when the FIFO is full. Refer to the Table 12-21 and the Table 12-22 for more details. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 12-6 Freescale Semiconductor Memory Map/Register Definition Figure 12-3 shows the bits in the UTHRs. 0 7 R W Reset Offset DATA 0x00 0x0_4500, 0x0_4600 Figure 12-3. Transmitter Holding Registers (UTHR0, UTHR1) Table 12-5 describes the fields of UTHR. Table 12-5. UTHR Field Definition Bits 0–7 Name DATA Data that is written to UTHR (write only) Description 12.3.1.3 Divisor Most and Least Significant Byte Registers (UDMB and UDLB) The divisor least significant byte register (UDLB) is concatenated with the divisor most significant byte register (UDMB) to create the divisor used to divide the input clock into the DUART. The output frequency of the baud generator is 16 times the baud rate; therefore the desired baud rate = platform clock frequency / (16 × [UDMB||UDLB]). Equivalently, þþ [UDMB||UDLB:0b0000]þ = platform clock frequency / desired baud rate. Baud rates that can be generated by specific input clock frequencies are shown in Table 12-8. Figure 12-4 shows the bits in the UDMBs. 0 7 R UDMB W Reset Offset 0x00 0x0_4501, 0x0_4601 Figure 12-4. Divisor Most Significant Byte Registers (UDMB0, UDMB1) Table 12-6 describes the fields of UDMB registers. Table 12-6. UDMB Field Definition Bits 0–7 Name Description UDMB Divisor most-significant byte MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 12-7 DUART Figure 12-5 shows the bits in the UDLBs. 0 7 R UDLB W Reset Offset 0x00 0x0_4500, 0x0_4600 Figure 12-5. Divisor Least Significant Byte Registers (UDLB0, UDLB1) Table 12-7 describes the fields of UDLB registers. Table 12-7. UDLB Field Definition Bits 0–7 Name UDLB Description Divisor least-significant byte. This is concatenated with UDMB. Table 12-8 shows baud rate when the input clock is at certain frequencies. Table 12-8. Baud Rate Examples Divisor Baud Rate (Decimal) 9,600 19,200 38,400 56,000 128,000 256,000 9,600 19,200 38,400 56,000 128,000 256,000 Decimal 1736 868 434 298 130 65 2170 1085 543 372 163 81 Hex 6C8 364 1B2 12A 82 41 87A 43D 21F 174 A3 51 Input Clock(CCB) Frequency (MHz) 266 266 266 266 266 266 333 333 333 333 333 333 Percent Error (Decimal) 0.0064 0.0064 0.0064 0.1280 0.1600 0.1600 0.0064 0.0064 0.0858 0.0064 0.1472 0.4672 To get the percent error value, the following three steps are taken: 1. The input clock frequency (ICF) is divided by the actual frequency input (AFI) to get the correct divisor value (ICF/AFI where AFI = baud rate × 16). 2. The divisor value is subtracted from 1. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 12-8 Freescale Semiconductor Memory Map/Register Definition 3. The result from step two is multiplied by 100 to calculate the final percent error. The result is calculated in absolute value (no negative numbers). These steps can be described with the following equation: Percent error value = (1 – AFI/ICF) × 100 12.3.1.4 Interrupt Enable Register (UIER) The UIER gives the user the ability to mask specific UART interrupts to the MPC8540 programmable interrupt controller (PIC). Figure 12-6 shows the bits in the UIER. 0 3 4 5 6 7 R EMSI W Reset Offset 0x00 0x0_4501, 0x0_4601 ERLSI ETHREI ERDAI Figure 12-6. Interrupt Enable Register (UIER) Table 12-9 describes the fields of UIER. Table 12-9. UIER Field Definitions Bits 0–3 4 Name — EMSI Reserved Enable MODEM status interrupt 0 Mask interrupts caused by UMSR[DCTS] being set 1 Enable and assert interrupts when the clear-to-send bit in the UART MODEM status register (UMSR) changes state Enable receiver line status interrupt 0 Mask interrupts when ULSR’s overrun, parity error, framing error or break interrupt bits are set 1 Enable and assert interrupts when ULSR’s overrun, parity error, framing error or break interrupt bits are set Description 5 ERLSI 6 ETHREI Enable transmitter holding register empty interrupt 0 Mask interrupt when ULSR[THRE] is set 1 Enable and assert interrupts when ULSR[THRE] is set ERDAI Enable received data available interrupt 0 Mask interrupt when new receive data is available or receive data time out has occurred 1 Enable and assert interrupts when a new data character is received from the external device and/or a time-out interrupt occurs in the FIFO mode 7 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 12-9 DUART 12.3.1.5 Interrupt ID Registers (UIIR0, UIIR1) The UIIRs indicate when an interrupt is pending from the corresponding UART and what type of interrupt is active. They also indicate if the FIFOs are enabled. The DUART prioritizes interrupts into four levels and records these in the corresponding UIIR. The four levels of interrupt conditions in order of priority are: 1. 2. 3. 4. Receiver line status Received data ready/character time-out Transmitter holding register empty MODEM status See Table 12-11 for more details. When the UIIR is read, the associated DUART serial channel freezes all interrupts and indicates the highest priority pending interrupt. While this read transaction is occurring, the associated DUART serial channel records new interrupts, but does not change the contents of UIIR until the read access is complete. Figure 12-7 shows the bits in the UIIR. 0 1 2 3 4 5 6 7 R W Reset Offset FE — IID3 IID2 IID1 IID0 0x01 0x0_4502, 0x0_4602 Figure 12-7. Interrupt ID Registers (UIIR) Table 12-10 describes the fields of the UIIR. Table 12-10. UIIR Field Definitions Bits 0–1 2–3 4 5–6 7 Name FE — IID3 Description FIFOs enabled. Reflects the setting of UFCR[FEN] Reserved Interrupt ID bits identify the highest priority interrupt that is pending as indicated in Table 12-11. IID3 is set along with IID2 only when a timeout interrupt is pending for FIFO mode. IID2–1 Interrupt ID bits identify the highest priority interrupt that is pending as indicated in Table 12-11. IID0 IID0 indicates when an interrupt is pending. 0 The UART has an active interrupt ready to be serviced. 1 No interrupt is pending. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 12-10 Freescale Semiconductor Memory Map/Register Definition The bits contained in the UIIR registers are described in Table 12-11. Table 12-11. UIIR IID Bits Summary IID Bits Priority IID[3–0] Level 0b0001 0b0110 0b0100 — Interrupt Type — Interrupt Description — Overrun error, parity error, framing error, or break interrupt How To Reset Interrupt — Read the line status register. Highest Receiver line status Second Received data available Receiver data available or trigger level Read the receiver buffer register or reached in FIFO mode interrupt is automatically reset if the number of bytes in the receiver FIFO drops below the trigger level. Second Character time-out No characters have been removed from or input to the receiver FIFO during the last 4 character times and there is at least one character in the receiver FIFO during this time. Transmitter holding register is empty CTS input value changed since last read of UMSR Read the receiver buffer register. 0b1100 0b0010 0b0000 Third Fourth UTHR empty MODEM status Read the UIIR or write to the UTHR. Read the UMSR. 12.3.1.6 FIFO Control Registers (UFCR0, UFCR1) The UFCR, a write-only register, is used to enable and clear the receiver and transmitter FIFOs, set a receiver FIFO trigger level to control the received data available interrupt, and select the type of DMA signaling. When the UFCR bits are written, the FIFO enable bit must also be set or else the UFCR bits are not programmed. When changing from FIFO mode to 16450 mode (non-FIFO mode) and vice versa, data is automatically cleared from the FIFOs. After all the bytes in the receiver FIFO are cleared, the receiver internal shift register is not cleared. Similarly, the bytes are cleared in the transmitter FIFO, but the transmitter internal shift register is not cleared. Both TFR and RFR are self-clearing bits. Figure 12-8 shows the bits in the UFCRs. 0 1 2 3 4 5 6 7 R W Reset Offset RTL — 0x00 0x0_4502, 0x0_4602 DMS TFR RFR FEN Figure 12-8. FIFO Control Registers (UFCR0, UFCR1) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 12-11 DUART Table 12-12 describes the fields of the UFCRs. Table 12-12. UFCR Field Definitions Bits Name 0–1 RTL Description Receiver trigger level. A received data available interrupt occurs when UIER[ERDAI] is set and the number of bytes in the receiver FIFO equals the designated interrupt trigger level as follows: 00 1 byte 01 4 bytes 10 8 bytes 11 14 bytes Reserved 2–3 4 — DMS DMA mode select. See Section 12.4.5.2, “DMA Mode Select,” for more information. 0 UDSR[RXRDY] and UDSR[TXRDY] bits are in mode 0. 1 UDSR[RXRDY] and UDSR[TXRDY] bits are in mode 1 if UFCR[FEN] = 1. TFR Transmitter FIFO reset 0 No action 1 Clears all bytes in the transmitter FIFO and resets the FIFO counter/pointer to 0 Receiver FIFO reset 0 No action 1 Clears all bytes in the receiver FIFO and resets the FIFO counter/pointer to 0 FIFO enable 0 FIFOs are disabled and cleared 1 Enables the transmitter and receiver FIFOs 5 6 RFR 7 FEN 12.3.1.7 Line Control Registers (ULCR0, ULCR1) The ULCRs specify the data format for the UART bus and set the divisor latch access bit ULCR[DLAB], which controls the ability to access the divisor latch least and most significant bit registers and the alternate function register. After initializing the ULCR, the software should not re-write the ULCR when valid transfers on the UART bus are active. The software should not re-write the ULCR until the last STOP bit has been received and there are no new characters being transferred on the bus. The stick parity bit, ULCR[SP], assigns a set parity value for the parity bit time slot sent on the UART bus. The set value is defined as mark parity (logic 1) or space parity (logic 0). ULCR[PEN] and ULCR[EPS] help determine the set parity value. See Table 12-14 for more information. ULCR[NSTB], defines the number of STOP bits to be sent at the end of the data transfer. The receiver only checks the first STOP bit, regardless of the number of STOP bits selected. The word length select bits (1 and 0) define the number of data bits that are transmitted or received as a serial character. The word length does not include START, parity, and STOP bits. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 12-12 Freescale Semiconductor Memory Map/Register Definition Figure 12-9 shows the bits in the ULCRs. 0 1 2 3 4 5 6 7 R DLAB W Reset Offset 0x00 0x0_4503, 0x0_4603 SB SP EPS PEN NSTB WLS Figure 12-9. Line Control Register (ULCR) Table 12-13 describes the fields of the ULCRs. Table 12-13. ULCR Field Definitions Bits Name 0 Description DLAB Divisor latch access bit 0 Access to all registers except UDLB, UAFR, and UDMB 1 Ability to access divisor latch least and most significant byte registers and alternate function register (UAFR) SB Set break 0 Send normal UTHR data onto the serial output (SOUT) signal 1 Force logic 0 to be on the SOUT signal. Data in the UTHR is not affected Stick parity 0 Stick parity is disabled. 1 If PEN = 1 and EPS = 1, space parity is selected. And if PEN = 1 and EPS = 0, mark parity is selected. Even parity select. See Table 12-14 for more information. 0 If PEN = 1 and SP = 0, odd parity is selected. 1 If PEN = 1 and SP = 0, even parity is selected. Parity enable 0 No parity generation and checking 1 Generate parity bit as a transmitter, and check parity as a receiver 1 2 SP 3 EPS 4 PEN 5 NTSB Number of STOP bits 0 One STOP bit is generated in the transmitted data. 1 When a 5-bit data length is selected, 1 ½ STOP bits are generated. When either a 6-, 7-, or 8-bit word length is selected, two STOP bits are generated. WLS Word length select. Number of bits that comprise the character length. The word length select values are as follows: 00 5 bits 01 6 bits 10 7 bits 11 8 bits 6–7 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 12-13 DUART Table 12-14. Parity Selection Using ULCR[PEN], ULCR[SP], and ULCR[EPS] PEN 0 0 0 0 1 1 1 1 SP 0 0 1 1 0 0 1 1 EPS 0 1 0 1 0 1 0 1 Parity Selected No parity No parity No parity No parity Odd parity Even parity Mark parity Space parity 12.3.1.8 MODEM Control Registers (UMCR0, UMCR1) The UMCRs control the interface with the external peripheral device on the UART bus. Figure 12-10 shows the bits in the UMCRs. 0 2 3 4 5 6 7 R LOOP W Reset Offset 0x00 0x0_4504, 0x0_4604 RTS Figure 12-10. Modem Control Register (UMCR) Table 12-15 describes the fields of UMCRs. . Table 12-15. UMCR Field Definitions Bits 0–2 3 Name — Reserved Description LOOP Local loop-back mode 0 Normal operation 1 Functionally, the data written to UTHR can be read from URBR of the same UART, and UMCR[RTS] is tied to UMSR[CTS]. — RTS Reserved Ready to send 0 Negates corresponding UART_RTS output 1 Assert corresponding UART_RTS output. Informs external MODEM or peripheral that the UART is ready for sending/receiving data Reserved 4–5 6 7 — MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 12-14 Freescale Semiconductor Memory Map/Register Definition 12.3.1.9 Line Status Registers (ULSR0, ULSR1) The ULSRs are read-only registers that monitor the status of the data transfer on the UART buses. To isolate the status bits from the proper character received through the UART bus, software should read the ULSR and then the URBR. Figure 12-11 shows the bits in the ULSRs. 0 1 2 3 4 5 6 7 R W Reset Offset RFE TEMT THRE BI FE PE OE DR 0x60 0x0_4505, 0x0_4605 Figure 12-11. Line Status Register (ULSR) Table 12-16 describes the fields of the ULSRs. Table 12-16. ULSR Field Definitions Bits Name 0 RFE Description Receiver FIFO error 0 This bit is cleared when there are no errors in the receiver FIFO or on a read of the ULSR with no remaining receiver FIFO errors. 1 Set to one when one of the characters in the receiver FIFO encounters an error (framing, parity, or break interrupt) 1 TEMT Transmitter empty 0 Either or both the UTHR or the internal transmitter shift register has a data character. In FIFO mode, a data character is in the transmitter FIFO or the internal transmitter shift register. 1 Both the UTHR and the internal transmitter shift register are empty. In FIFO mode, both the transmitter FIFO and the internal transmitter shift register are empty. THRE Transmitter holding register empty 0 The UTHR is not empty. 1 A data character has transferred from the UTHR into the internal transmitter shift register. In FIFO mode, the transmitter FIFO contains no data character. BI Break interrupt 0 This bit is cleared when the ULSR is read or when a valid data transfer is detected (that is, STOP bit is received). 1 Received data of logic 0 for more than START bit + Data bits + Parity bit + one STOP bits length of time. A break condition is expected to last at least two character lengths and a new character is not loaded until SIN returns to the mark state (logic 1) and a valid START is detected. In FIFO mode, a zero character is encountered in the FIFO (the zero character is at the top of the FIFO). In FIFO mode, only one zero character is stored. Note that the ULSR[BI] is set immediately after ULSR is read if bus remains zero and no mark state followed by a valid new character has been detected. 2 3 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 12-15 DUART Table 12-16. ULSR Field Definitions (continued) Bits Name 4 FE Description Framing error 0 This bit is cleared when ULSR is read or when a new character is loaded into the URBR from the receiver shift register. 1 Invalid STOP bit for receive data (only the first STOP bit is checked). In FIFO mode, this bit is set when the character that detected a framing error is encountered in the FIFO (that is the character at the top of the FIFO). An attempt to resynchronize occurs after a framing error. The UART assumes that the framing error (due to a logic 0 being read when a logic 1 (STOP) was expected) was due to a STOP bit overlapping with the next START bit, so it assumes this logic 0 sample is a true START bit and then will receive the following new data. Parity error 0 This bit is cleared when ULSR is read or when a new character is loaded into the URBR. 1 Unexpected parity value encountered when receiving data. In FIFO mode, the character with the error is at the top of the FIFO. Overrun error 0 This bit is cleared when ULSR is read. 1 Before the URBR is read, the URBR was overwritten with a new character. The old character is loss. In FIFO mode, the receiver FIFO is full (regardless of the receiver FIFO trigger level setting) and a new character has been received into the internal receiver shift register. The old character was overwritten by the new character. Data in the receiver FIFO was not overwritten. Data ready 0 This bit is cleared when URBR is read or when all of the data in the receiver FIFO is read. 1 A character has been received in the URBR or the receiver FIFO. 5 PE 6 OE 7 DR 12.3.1.10 MODEM Status Registers (UMSR0, UMSR1) The UMSRs track the status of the MODEM (or external peripheral device) clear to send (CTS) signal for the corresponding UART. Figure 12-12 shows the bits in the UMSRs. 0 2 3 4 6 7 R W Reset Offset — CTS — DCTS 0x00 0x0_4506, 0x0_4606 Figure 12-12. Modem Status Register (UMSR) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 12-16 Freescale Semiconductor Memory Map/Register Definition Table 12-17 describes the fields of the UMSRs. Table 12-17. UMSR Field Definitions Bits Name 0–2 3 — CTS Reserved Clear to send. Represents the inverted value of the CTS input pin from the external peripheral device 0 Corresponding CTSn is negated 1 Corresponding CTSn is asserted. The MODEM or peripheral device is ready for data transfers. Reserved Description 4–6 7 — DCTS Delta clear to send 0 No change on the corresponding CTSn signal since the last read of UMSR[CTS] 1 The CTSn value has changed, since the last read of UMSR[CTS]. Causes an interrupt if UIER[EMSI] is set to detect this condition 12.3.1.11 Scratch Registers (USCR0, USCR1) The USCR registers are for debugging software or the DUART hardware. The USCRs do not affect the operation of the DUART. Figure 12-13 shows the bits in USCRs. 0 7 R DATA W Reset Offset 0x00 0x0_4507, 0x0_4607 Figure 12-13. Scratch Register (USCR) Table 12-18 describes the fields of the USCRs. Table 12-18. USCR Field Definitions Bits 0–7 Name DATA Data Description 12.3.1.12 Alternate Function Registers (UAFR0, UAFR1) The UAFRs give software the ability to write to both UART0 and UART1 registers simultaneously with the same write operation. The UAFRs also provide a means for the device’s performance monitor to track the baud clock. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 12-17 DUART Figure 12-14 shows the bits in the UAFRs. 0 5 6 7 R BO W Reset Offset 0x00 0x0_4502, 0x0_4602 CW Figure 12-14. Alternate Function Register (UAFR) Table 12-19 describes the fields of the UAFRs. Table 12-19. UAFR Field Definitions Bits Name 0–5 6 — BO Reserved Baud clock select 0 The baud clock is not gated off. 1 The baud clock is gated off. Concurrent write enable 0 Disables writing to both UART0 and UART1 1 Enables concurrent writes to corresponding UART registers. A write to a register in UART0 is also a write to the corresponding register in UART1 and vice versa. Description 7 CW 12.3.1.13 DMA Status Registers (UDSR0, UDSR1) The DMA status registers (UDSRs) are read-only registers that return transmitter and receiver FIFO status. UDSRs also provide the ability to assist DMA data operations to and from the FIFOs. Figure 12-15 shows the bits in UDSRs. 0 5 6 7 R W Reset Offset 0x01 0x0_4510, 0x0_4610 TXRDY RXRDY Figure 12-15. DMA Status Register (UDSR) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 12-18 Freescale Semiconductor Memory Map/Register Definition Table 12-20 describes the fields of the UDSRs. Table 12-20. UDSR Field Definitions Bits 0–5 6 Name — Reserved Description TXRDY Transmitter ready. This read-only bit reflects the status of the transmitter FIFO or the UTHR. The status depends on the DMA mode selected, which is determined by the DMS and FEN bits in the UFCR. 0 The bit is cleared, as shown in Table 12-22. 1 This bit is set, as shown in Table 12-21. RXRDY Receiver ready. This read-only bit reflects the status of the receiver FIFO or URBR. The status depends on the DMA mode selected, which is determined by the DMS and FEN bits in the UFCR. 0 The bit is cleared, as shown in Table 12-24. 1 This bit is set, as shown in Table 12-23. 7 Table 12-21. UDSR[TXRDY] Set Conditions DMS FEN DMA Mode 0 0 1 1 0 1 0 1 0 0 0 1 TXRDY is set when the transmitter FIFO is full. Meaning TXRDY is set after the first character is loaded into the transmitter FIFO or UTHR. Table 12-22. UDSR[TXRDY] Cleared Conditions DMS FEN DMA Mode 0 0 1 1 0 1 0 1 0 0 0 1 TXRDY is cleared when there are no characters in the transmitter FIFO or UTHR. TXRDY remains clear when the transmitter FIFO is not yet full. Meaning TXRDY is cleared when there are no characters in the transmitter FIFO or UTHR. Table 12-23. UDSR[RXRDY] Set Conditions DMS FEN DMA Mode 0 0 1 1 0 1 0 1 0 0 0 1 RXRDY is set when the trigger level has not been reached and there has been no time out. Meaning RXRDY is set when there are no characters in the receiver FIFO or URBR. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 12-19 DUART Table 12-24. UDSR[RXRDY] Cleared DMS FEN DMA Mode 0 0 1 1 0 1 0 1 0 0 0 1 RXRDY is cleared when the trigger level or a time-out has been reached. RXRDY remains cleared until the receiver FIFO is empty. Meaning RXRDY is cleared when there is at least one character in the receiver FIFO or URBR. 12.4 Functional Description The communication channel provides a full-duplex asynchronous receiver and transmitter using an operating frequency derived from the CCB clock signal. The transmitter accepts parallel data with a write access to the transmitter holding register (UTHR). In FIFO mode, the data is placed directly into an internal transmitter shift register, or into the transmitter FIFO—see Section 12.4.5, “FIFO Mode.” The transmitting registers convert the data to a serial bit stream, by inserting the appropriate START, STOP, and optional parity bits. Finally, the registers output a composite serial data stream on the channel transmitter serial data output (SOUT). The transmitter status may be polled or interrupt-driven. The receiver accepts serial data on the channel receiver serial data input (SIN), converts the data into parallel format, and checks for START, STOP, and parity bits. In FIFO mode, the receiver removes the START, STOP, and parity bits and then transfers the assembled character from the receiver buffer, or receiver FIFO. This transfer occurs in response to a read of the UART receiver buffer register (URBR). The receiver status may be polled or interrupt driven. 12.4.1 Serial Interface The UART bus is a serial, full-duplex, point-to-point bus as shown in Figure 12-16. Therefore, only two devices are attached to the same signals and there is no need for address or arbitration bus cycles. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 12-20 Freescale Semiconductor Functional Description rxcnt 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 SOUT1 D6 D5 D4 D3 D2 D1 D0 PTY Data bits D6 D5 D4 D3 D2 D1 D0 PTY Data bits START STOP bits START Optional even/odd parity Two 7-bit data transmissions with parity and 2-bit STOP transactions STOP bits Optional even/odd parity Figure 12-16. UART Bus Interface Transaction Protocol Example A standard UART bus transfer is composed of either three or four parts: • • • • START bit Data transfer bits (least-significant bit is first data bit on the bus) Parity bit (optional) STOP bits An internal logic sample signal, rxcnt, uses the frequency of the baud-rate generator to drive the bits on SOUT. The following sections describe the four components of the serial interface, the baud-rate generator, local loop-back mode, different errors, and FIFO mode. 12.4.1.1 START Bit A write to the transmitter holding register (UTHR) generates a START bit on the SOUT signal. Figure 12-16 shows that the START bit is defined as a logic 0. The START bit denotes the beginning of a new data transfer which is limited to the bit length programmed in the UART line control register (ULCR).When the bus is idle, SOUT is high. 12.4.1.2 Data Transfer Each data transfer contains 5–8 bits of data. The ULCR data bit length for the transmitter and receiver UART devices must agree before a transfer begins; otherwise, a parity or framing error may occur. A transfer begins when UTHR is written. At that time a START bit is generated followed by 5–8 of the data bits previously written to the UTHR. The data bits are driven from the least significant to the most significant bits. After the parity and STOP bits, a new data transfer can begin if new data is written to the UTHR. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 12-21 DUART 12.4.1.3 Parity Bit The user has the option of using even, odd, no parity, or stick parity (see Section 12.3.1.7, “Line Control Registers (ULCR0, ULCR1).” Both the receiver and transmitter parity definition must agree before attempting to transfer data. When receiving data a parity error can occur if an unexpected parity value is detected. (See Section 12.3.1.9, “Line Status Registers (ULSR0, ULSR1).”) 12.4.1.4 STOP Bit The transmitter device ends the write transfer by generating a STOP bit. The STOP bit is always high. The user can program the length of the STOP bit(s) in the ULCR. Both the receiver and transmitter STOP bit length must agree before attempting to transfer data. A framing error can occur if an invalid STOP bit is detected. 12.4.2 Baud-Rate Generator Logic Each UART contains an independent programmable baud-rate generator, that is capable of taking the CCB clock input and dividing the input by any divisor from 1 to 216 – 1. The baud rate is defined as the number of bits per second that can be sent over the UART bus. The formula for calculating baud rate is as follows: Baud rate = (1/16) × (CCB clock frequency/divisor value) Therefore, the output frequency of the baud-rate generator is 16 times the baud rate. The divisor value is determined by the following two 8-bit registers to form a 16-bit binary number: • • UART divisor most significant byte register (UDMB) UART divisor least significant byte register (UDLB) Upon loading either of the divisor latches, a 16-bit baud-rate counter is loaded. The divisor latches must be loaded during initialization to ensure proper operation of the baud-rate generator. Both UART devices on the same bus must be programmed for the same baud-rate before starting a transfer. The baud clock can be passed to the performance monitor by enabling the UAFR[BO] bit. This can be used to determine baud rate errors. 12.4.3 Local Loop-Back Mode Local loop-back mode is provided for diagnostic testing. The data written to UTHR can be read from the receiver buffer register (URBR) of the same UART. In this mode, the MODEM control MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 12-22 Freescale Semiconductor Functional Description register UMCR[RTS] is internally tied to the MODEM status register UMSR[CTS]. The transmitter SOUT is set to a logic 1 and the receiver SIN is disconnected. The output of the transmitter shift register is looped back into the receiver shift register input. The CTS (input signal) is disconnected, RTS is internally connected to CTS, and the RTS (output signal) becomes inactive. In this diagnostic mode, data that is transmitted is immediately received. In local loop-back mode the transmit and receive data paths of the DUART can be verified. Note that in local loop-back mode, the transmit/receive interrupts are fully operational and can be controlled by the interrupt enable register (UIER). 12.4.4 Errors The following sections describe framing, parity, and overrun errors which may occur while data is transferred on the UART bus. Each of the error bits are usually cleared, as described below, when the line status register (ULSR) is read. 12.4.4.1 Framing Error When an invalid STOP bit is detected, a framing error occurs and ULSR[FE] is set. Note that only the first STOP bit is checked. In FIFO mode, ULSR[FE] is set when the character at the top of the FIFO detects a framing error. An attempt to re-synchronize occurs after a framing error. The UART assumes that the framing error (due to a logic 0 being read when a logic 1 (STOP) was expected) was due to a STOP bit overlapping with the next START bit. ULSR[FE] is cleared when ULSR is read or when a new character is loaded into the URBR from the receiver shift register. 12.4.4.2 Parity Error A parity error occurs, and ULSR[PE] is set, when unexpected parity values are encountered while receiving data. In FIFO mode, ULSR[PE] is set when the character with the error is at the top of the FIFO. ULSR[PE] is cleared when ULSR is read or when a new character is loaded into the URBR. 12.4.4.3 Overrun Error When a new (overwriting character) STOP bit is detected and the old character is lost, an overrun error occurs and ULSR[OE] is set. In FIFO mode, ULSR[OE] is set after the receiver FIFO is full (despite the receiver FIFO trigger level setting) and a new character has been received into the internal receiver shift register. Data in the FIFO is not overwritten; only the shift register data is overwritten. Therefore, the interrupt occurs immediately. ULSR[OE] is cleared when ULSR is read. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 12-23 DUART 12.4.5 FIFO Mode The UARTs use an alternate mode (FIFO mode) to relieve the processor core from excessive software overhead. The FIFO control register (UFCR) is used to enable and clear the receiver and transmitter FIFOs and set the FIFO receiver trigger level UFCR[RTL] to control the received data available interrupt UIER[ERDAI]. The UFCR also selects the type of DMA signaling. The UDSR[RXRDY] indicates the status of the receiver FIFO. The DMA status registers (UDSR[TXRDY]) indicate when the transmitter FIFO is full. When in FIFO mode, data written to UTHR is placed into the transmitter FIFO. The first byte written to UTHR is the first byte onto the UART bus. 12.4.5.1 FIFO Interrupts In FIFO mode, the UIER[ERDAI] is set when a time-out interrupt occurs. When a receive data time-out occurs there is a maskable interrupt condition (through UIER[ERDAI]). See Section 12.3.1.4, “Interrupt Enable Register (UIER),” for more details on interrupt enables. The interrupt ID register (UIIR) indicates if the FIFOs are enabled. Interrupt ID3 UIIR[IID3] bit is only set for FIFO mode interrupts. The character time-out interrupt occurs when no characters have been removed from or input to the receiver FIFO during the last four character times and there is at least one character in the receiver FIFO during this time. The character time-out interrupt (controlled by UIIR[IID]) is cleared when the URBR is read. See Section 12.3.1.5, “Interrupt ID Registers (UIIR0, UIIR1),” for more information. The UIIR[FE] bits indicate if FIFO mode is enabled. 12.4.5.2 DMA Mode Select The UDSR[RXRDY] bit reflects the status of the receiver FIFO or URBR. In mode 0 (UFCR[DMS] is cleared), UDSR[RXRDY] is cleared when there is at least one character in the receiver FIFO or URBR and it is set when there are no more characters in the receiver FIFO or URBR. This occurs regardless of the setting of the UFCR[FEN] bit. In mode 1 (UFCR[DMS] and UFCR[FEN] are set), UDSR[RXRDY] is cleared when the trigger level or a time-out has been reached and it is set when there are no more characters in the receiver FIFO. The UDSR[TXRDY] bit reflects the status of the transmitter FIFO or UTHR. In mode 0 (UFCR[DMS] is cleared), UDSR[TXRDY] is cleared when there are no characters in the transmitter FIFO or UTHR and it is set after the first character is loaded into the transmitter FIFO or UTHR. This occurs regardless of the setting of the UFCR[FEN] bit. In mode 1 (UFCR[DMS] and UFCR[FEN] are set), UDSR[TXRDY] is cleared when there are no characters in the transmitter FIFO or UTHR and it is set when the transmitter FIFO is full. See Section 12.3.1.13, “DMA Status Registers (UDSR0, UDSR1),” for a complete description of the USDR[RXRDY] and USDR[TXRDY] bits. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 12-24 Freescale Semiconductor DUART Initialization/Application Information 12.4.5.3 Interrupt Control Logic An interrupt is active when DUART interrupt ID register bit 0 (UIIR[0]), is cleared. The interrupt enable register (UIER) is used to mask specific interrupt types. For more details refer to the description of UIER in Section 12.3.1.4, “Interrupt Enable Register (UIER).” When the interrupts are disabled in UIER, polling software can not use UIIR[0] to determine whether the UART is ready for service. The software must monitor the appropriate bits in the line status (ULSR) and/or the MODEM status (UMSR) registers. UIIR[0] can be used for polling if the interrupts are enabled in UIER. 12.5 DUART Initialization/Application Information The following requirements must be met for DUART accesses: • • All DUART registers must be mapped to a cache-inhibited area. All DUART registers are 1 byte wide. Reads and writes to these registers must be byte-wide operations. A system reset puts the DUART registers to a default state. Before the interface can transfer serial data, the following initialization steps are recommended: 1. Update the programmable interrupt controller (PIC) DUART channel interrupt vector source registers. 2. Set data attributes and control bits in the ULCR, UFCR, UAFR, UMCR, UDLB, and UDMB. 3. Set the data attributes and control bits of the external MODEM or peripheral device. 4. Set the interrupt enable register (UIER). 5. To start a write transfer, write to the UTHR. 6. Poll UIIR if the interrupts generated by the DUART are masked. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 12-25 DUART MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 12-26 Freescale Semiconductor Chapter 13 Local Bus Controller This chapter describes the local bus controller (LBC) block. It describes the external signals and the memory-mapped registers as well as a functional description of the general-purpose chip-select machine (GPCM), SDRAM machine, and user-programmable machines (UPMs) of the LBC. Finally, it includes an initialization and applications information section with many specific examples of its use. 13.1 Introduction Figure 13-1 is a functional block diagram of the LBC, which supports three interfaces: GPCM, UPM, and SDRAM controller. LCS[0:7] LBCTL Timers/ Counters Config Regs Refresh SDRAM Memory Banks and Compare Local Memory Controller UPM CTRL Control Signal Timing Generator UPM CTRL UPM RAM LGPL[0:3] LGPL4/LUPWAIT LGPL5 LBS[0:3] SDRAM CTRL LSDA10 LSDWE LSDRAS LSDCAS LSDDQM[0:3] Clock Divider Local Address Gasket and Mux Address and Data Machine GPCM CTRL LOE LGTA LWE[0:3] LA[27:31] LALE LPBSE LDP[0:3] MDVAL MSRCID[0:4] LAD[0:31] LCLK LCKE LSYNC_IN LSYNC_OUT GPCM Local Data Transfer Acknowledge Figure 13-1. Local Bus Controller Block Diagram MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 13-1 Addr/Data Mux Local Bus Controller 13.1.1 Overview The main component of the LBC is its memory controller, which provides a seamless interface to many types of memory devices and peripherals. The memory controller is responsible for controlling eight memory banks shared by a high performance SDRAM machine, a GPCM, and up to three UPMs. As such, it supports a minimal glue logic interface to synchronous DRAM (SDRAM), SRAM, EPROM, flash EPROM, burstable RAM, regular DRAM devices, extended data output DRAM devices, and other peripherals. The external address latch signal (LALE) allows multiplexing of addresses with data signals to reduce the device pin count. The LBC also includes a number of data checking and protection features such as data parity generation and checking, write protection and a bus monitor to ensure that each bus cycle is terminated within a user-specified period. 13.1.2 Features The LBC main features are as follows: • Memory controller with eight memory banks — 34-bit address decoding with mask — Variable memory block sizes (32 Kbytes to 4 Gbytes) — Selection of control signal generation on a per-bank basis — Data buffer controls activated on a per-bank basis — Up to 256-byte bursts, arbitrarily aligned — Automatic segmentation of large transactions — Odd/even parity checking including read-modify-write (RMW) parity for single accesses — Write-protection capability — Atomic operation — Parity byte-select SDRAM machine — Provides the control functions and signals for glueless connection to JEDEC-compliant SDRAM devices — Supports up to four concurrent open pages per device — Supports SDRAM port size of 32, 16, and 8 bits — Supports external address and/or command lines buffering • MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 13-2 Freescale Semiconductor Introduction • • • • General-purpose chip-select machine (GPCM) — Compatible with SRAM, EPROM, FEPROM, and peripherals — Global (boot) chip-select available at system reset — Boot chip-select support for 8-, 16-, 32-bit devices — Minimum 3-clock access to external devices — Four byte-write-enable signals (LWE[0:3]) — Output enable signal (LOE) — External access termination signal (LGTA) Three user-programmable machines (UPMs) — Programmable-array-based machine controls external signal timing with a granularity of up to one-quarter of an external bus clock period — User-specified control-signal patterns run when an internal master requests a single-beat or burst read or write access. — UPM refresh timer runs a user-specified control signal pattern to support refresh — User-specified control-signal patterns can be initiated by software — Each UPM can be defined to support DRAM devices with depths of 64, 128, 256, and 512 Kbytes, and 1, 2, 4, 8, 16, 32, 64, 128, and 256 Mbytes — Support for 8-, 16-, 32-bit devices — Page mode support for successive transfers within a burst — Internal address multiplexing supporting 64-, 128-, 256-, and 512-Kbyte, and 1-, 2-, 4-, 8-, 16-, 32-, 64-, 128-, and 256-Mbyte page banks Optional monitoring of transfers between local bus internal masters and local bus slaves (local bus error reporting) Support for delay-locked loop (DLL) with software-configurable bypass for low frequency bus clocks 13.1.3 Modes of Operation The LBC provides one GPCM, one SDRAM machine, and three UPMs for the local bus, with no restriction on how many of the eight banks (chip selects) can be programmed to operate with any given machine. When a memory transaction is dispatched to the LBC, the memory address is compared with the address information of each bank (chip select). The corresponding machine assigned to that bank (GPCM, SDRAM, or UPM) then takes ownership of the external signals that control the access and maintains control until the transaction ends. Thus, with the LBC in GPCM, SDRAM, or UPM mode, only one of the eight chip selects is active at any time for the duration of the transaction. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 13-3 Local Bus Controller 13.1.3.1 LBC Bus Clock and Clock Ratios The LBC supports ratios of 2, 4, and 8 between the faster system (CCB) clock and the slower external bus clock (LCLK[0:2]). This ratio is software programmable through the clock ratio register (LCRR[CLKDIV]). In addition to establishing the frequency of the external local bus clock, CLKDIV also affects the resolution of signal timing shifts in GPCM mode and the interpretation of UPM array words in UPM mode. The bus clock is driven identically onto pins, LCLK[0:2], to allow the clock load to be shared equally across a pair of signal nets, thereby enhancing the edge rates of the bus clock. 13.1.3.2 Source ID Debug Mode In debug mode, the LBC provides the ID of a transaction source on external device pins. This mode is enabled on power-on reset, as described in Section 20.4.4, “Local Bus Interface Debug.” When placed in this mode, the 5-bit internal ID of the current transaction source appears on MSRCID[0:4] whenever valid address or data is available on the LBC external pins. The reserved value of 0x1F, which indicates invalid address or data, appears on the source ID pins at all other times. The combination of a valid source ID (any value except 0x1F) and the value of external address latch enable (LALE) and data valid (MDVAL) facilitate capturing useful debug data as follows: • If a valid source ID is detected on MSRCID[0:4] and LALE is asserted, a valid full 32-bit address may be latched from LAD[0:31]. Note that in SDRAM mode the address vector contains the full address as {row, bank, column, lsbs} where row corresponds to the same row address for the given column address and lsbs are the unconnected lsbs of the address for a given port size. If a valid source ID is detected on MSRCID[0:4] and MDVAL is asserted, valid data may be latched from LAD[0:31]. • 13.1.4 Power-Down Mode The LBC can enter a power-down mode when the system stops the internal (system) clock to the block by using a handshake protocol initiated by the DEVDISR[LBC] setting in the global utilities block. On entering power-down mode, the LBC places any SDRAM devices, if used, in self-refresh mode before the bus clock is stopped. The LBC also allows the DLL sufficient time to recover following the reapplication of the system clock. 13.1.5 References • MPC8260 PowerQUICC II Family Reference Manual, Chapters 4, 6, and 10 (order no. MPC8260UM). MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 13-4 Freescale Semiconductor External Signal Descriptions 13.2 External Signal Descriptions Table 13-1 contains a list of external signals related to the LBC and summarizes their function. Table 13-1. Signal Properties—Summary Name LALE LCS LWE/ LSDDQM/ LBS LSDA10/ LGPL0 Number of Signals 1 8 4 Direction Output Output Output Output Output Output Output Function External address latch enable Chip selects GPCM mode: write enable SDRAM mode: byte lane data mask UPM mode: byte (lane) select SDRAM mode: row address bit/command bit UPM mode: general-purpose line 0 SDRAM mode: write enable UPM mode: general-purpose line 1 GPCM mode: output enable SDRAM mode: row address strobe UPM mode: general-purpose line 2 SDRAM mode: column address strobe UPM mode: general-purpose line 3 GPCM mode: transaction termination UPM mode: general-purpose line 4 UPM mode: external device wait Local bus parity byte select UPM mode: general-purpose line 5 Data buffer control Local bus non-multiplexed address lsb’s 1 LSDWE / LGPL1 LOE/ LSDRAS/ LGPL2 LSDCAS/ LGPL3 LGTA / LGPL4/ LUPWAIT/ LPBSE LGPL5 LBCTL LA[27:31] LAD[0:31] LDP LCKE LCLK[0:2] LSYNC_IN LSYNC_OUT MDVAL MSRCID 1 Output Output Output Output Output Output Output Input Output Input Output Output Output Output 1 1 1 1 1 5 32 4 1 3 1 1 1 5 Input/Output Multiplexed address/data bus Input/Output Local bus data parity Output Output Input Output Output Output Local bus clock enable Local bus clocks DLL synchronize input DLL synchronize output In LBC debug mode: local bus data valid In LBC debug mode: local bus source ID MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 13-5 Local Bus Controller Table 13-2 shows the detailed external signal descriptions for the LBC. Table 13-2. Local Bus Controller Detailed Signal Descriptions Signal LALE I/O Description O External address latch enable. The local bus memory controller provides control for an external address latch, which allows address and data to be multiplexed on the device pins. State Asserted/Negated—LALE is asserted with the address at the beginning of each memory Meaning controller transaction. The number of cycles for which it is asserted is governed by the ORn[EAD] and LCRR[EADC] fields. The exact timing of the negation of LALE is controlled by the LBCR[AHD] field. Note that no other control signals are asserted during the assertion of LALE. LCS[0:7] O Chip selects. Eight chip selects are provided which are mutually exclusive. State Asserted/Negated—Used to enable specific memory devices or peripherals connected to the Meaning LBC. LCS [0:7] are provided on a per-bank basis with LCS0 corresponding to the chip select for memory bank 0, which has the memory type and attributes defined by BR0 and OR0. LWE[0:3]/ O GPCM write enable/SDRAM data mask/UPM byte select. These signals select or validate each byte lane of the data bus. For banks with port sizes of 32 bits (as set by BRn[PS]), all four signals are defined. For LSDDQM[0:3]/ a 16-bit port size, only bits 0–1 are defined; and for an 8-bit port size, bit 0 is the only defined signal. The LBS[0:3] least significant address bits of each access also determine which byte lanes are considered valid for a given data transfer. State Asserted/Negated—For GPCM operation, LWE[0:3] assert for each byte lane enabled for Meaning writing. For SDRAM operation, LSDDQM[0:3] function as the DQM or data mask signals provided by JEDEC-compliant SDRAM devices, with one DQM provided per byte lane. LSDDQM[0:3] are driven high when the LBC wishes to mask a write or disable read data output from the SDRAM. LBS[0:3] are programmable byte-select signals in UPM mode. See Section 13.4.4.4, “RAM Array,” for programming details about LBS [0:3]. Timing Assertion/Negation—See Section 13.4.2, “General-Purpose Chip-Select Machine (GPCM),” for details regarding the timing of LWE[0:3]. LSDA10/ LGPL0 O SDRAM A10/General-purpose line 0 State Asserted/Negated—For SDRAM accesses, represents address bit 10. When the row address Meaning is driven, it drives the value of address bit 10. When the column address is driven, it forms part of the SDRAM command. One of six general-purpose signals when in UPM mode; it drives a value programmed in the UPM array. O SDRAM write enable/General-purpose line 1 State Asserted/Negated—Should be connected to the SDRAM device WE input. Acts as the SDRAM Meaning write enable when accessing SDRAM. One of six general-purpose signals when in UPM mode, and drives a value programmed in the UPM array. O GPCM output enable/SDRAM RAS/General-purpose line 2 State Asserted/Negated—Controls the output buffer of memory when accessing memory/devices in Meaning GPCM mode. For SDRAM accesses, it is the row address strobe (RAS). One of six general-purpose lines when in UPM mode; it drives a value programmed in the UPM array. LSDWE / LGPL1 LOE/ LSDRAS/ LGPL2 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 13-6 Freescale Semiconductor External Signal Descriptions Table 13-2. Local Bus Controller Detailed Signal Descriptions (continued) Signal LSDCAS/ LGPL3 I/O O SDRAM CAS/General-purpose line 3 State Asserted/Negated—In SDRAM mode, drives the column address strobe (CAS). Meaning One of six general-purpose signals when in UPM mode, and drives a value programmed in the UPM array. I/O GPCM transfer acknowledge/General-purpose line 4/UPM wait/parity byte select State Asserted/Negated—Input in GPCM mode used for transaction termination. It may also be Meaning configured as one of six general-purpose output signals when in UPM mode or as an input to force the UPM controller to wait for the memory/device. When configured as LPBSE, it disables any use in GPCM or UPM modes. Because systems that use read-modify-write parity require an additional memory device, they must generate a byte-select like a normal data device. ANDing LBS[0:3] through external logic to achieve the logical function of this byte-select adds a delay to the byte-select path that can affect memory access timing. The LBC provides this optional byte-select pin that is an internal AND of the four (active low) byte selects, allowing glueless, faster connection to RMW-parity devices. O General-purpose line 5 State Asserted/Negated—One of six general-purpose signals when in UPM mode, and drives a Meaning value programmed in the UPM array. LBCTL O Data buffer control. The memory controller activates LBCTL for the local bus when a GPCM- or UPM-controlled bank is accessed. Access to an SDRAM machine-controlled bank does not activate the buffer control. Buffer control is disabled by setting OR n[BCTLD]. State Asserted/Negated—The LBCTL pin normally functions as a write/read control for a bus Meaning transceiver connected to the LAD lines. Note that an external data buffer must not drive the LAD lines in conflict with the LBC when LBCTL is high, because LBCTL remains high after reset and during address phases. LA[27:31] O Local bus nonmultiplexed address lsbs. All bits driven on LA[27:31] are defined for 8-bit port sizes. For 32-bit port sizes, LA[30:31] are don’t cares; for 16-bit port sizes LA31 is a don’t care. State Asserted/Negated—Although the LBC shares an address and data bus, up to five lsbs of the Meaning RAM address always appear on the dedicated address pins, LA[27:31]. These may be used, unlatched, in place of LAD[27:31] to connect the five lsbs of the address for address phases. For some RAM devices, such as fast-page DRAM, LA[27:31] serve as the column address offset during a burst access. Description LGTA/ LGPL4/ LUPWAIT/ LPBSE LGPL5 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 13-7 Local Bus Controller Table 13-2. Local Bus Controller Detailed Signal Descriptions (continued) Signal LAD[0:31] I/O Description I/O Multiplexed address/data bus. For configuration of a port size in BRn[PS] as 32 bits, all of LAD[0:31] must be connected to the external RAM data bus, with LAD[0:7] occupying the most significant byte lane (at address offset 0). For a port size of 16 bits, LAD[0:7] connect to the most significant byte lane (at address offset 0), while LAD[8:15] connect to the least-significant byte lane (at address offset 1); LAD[16:31] are unused for 16-bit port sizes. For a port size of 8 bits, only LAD[0:7] are connected to the external RAM. State Asserted/Negated—LAD[0:31] is the shared 32-bit address/data bus through which external Meaning RAM devices transfer data and receive addresses. Timing Assertion/Negation—During assertion of LALE, LAD[0:31] are driven with the RAM address for the access to follow. External logic should propagate the address on LAD[0:31] while LALE is asserted, and latch the address upon negation of LALE. After LALE is negated, LAD[0:31] are either driven by write data or are made high impedance by the LBC in order to sample read data driven by an external device. Following the last data transfer of a write access, LAD[0:31] are again taken into a high-impedance state. LDP[0:3] I/O Local bus data parity. Drives and receives the data parity corresponding with the data phases on LAD[0:31]. State Asserted/Negated—During write accesses, a parity bit is generated for each 8 bits of Meaning LAD[0:31], such that LDP0 is even/odd parity for LAD[0:7], while LDP3 is even/odd parity for LAD[24:31]. Unused byte lanes for port sizes less than 32 bits have undefined parity. Timing Assertion/Negation—Drive and receive the data parity corresponding with the data phases on LAD[0:31]. For read accesses, the parity bits for each byte lane are sampled on LDP[0:3] with the same timing that read data is sampled on LAD[0:31]. LDP[0:3] change impedance in concert with LAD[0:31]. LCKE O Local bus clock enable State Asserted/Negated—LCKE is the bus clock enable signal (CKE) for JEDEC-standard SDRAM Meaning devices. Asserted during normal SDRAM operation. LCLK[0:2] O Local bus clocks State Asserted/Negated—LCLK[0:2] drive an identical bus clock signal for distributed loads. If the Meaning LBC DLL is enabled (see LCRR[DBYP], Figure 13-19 on page 13-32), the bus clock phase is shifted earlier than transitions on other LBC signals (such as LAD[0:31] and LCSn) by a time delay matching the delay of the DLL timing loop set up between LSYNC_OUT and LSYNC_IN. LSYNC_OUT O DLL synchronization out State Asserted/Negated—A replica of the bus clock, appearing on LSYNC_OUT, should be Meaning propagated through a passive timing loop and returned to LSYNC_IN for achieving correct DLL lock. Timing Assertion/Negation—The time delay of the timing loop should be such that it compensates for the round-trip flight time of LCLK[0:2] and clocked drivers in the system. No load other than a timing loop should be placed on LSYNC_OUT. LSYNC_IN I DLL synchronization in State Asserted/Negated—See description of LSYNC_OUT. Meaning MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 13-8 Freescale Semiconductor Memory Map/Register Definition Table 13-2. Local Bus Controller Detailed Signal Descriptions (continued) Signal MDVAL I/O O Local bus data valid (LBC debug mode only) State Asserted/Negated—For a read, MDVAL asserts for one bus cycle in the cycle immediately Meaning preceding the sampling of read data on LAD[0:31]. For a write, MDVAL asserts for one bus cycle during the final cycle for which the current write data on LAD[0:31] is valid. During burst transfers, MDVAL asserts for each data beat. Timing Assertion/Negation—Valid only while the LBC is in system debug mode. In debug mode, MDVAL asserts when the LBC generates a data transfer acknowledge. MSRCID[0:4] O Local bus source ID (LBC debug mode only). In debug mode, all MSRCID[0:4] pins are driven high unless MSRCID[0:4] is driving a debug source ID for identifying the internal system device controlling the LBC. State Asserted/Negated—Remain high until the last bus cycle of the assertion of LALE, in which case Meaning the source ID of the address is indicated, or until MDVAL is asserted, in which case the source ID relating to the data transfer is indicated. In case of address debug, MSRCID[0:4] is valid only when the address on LAD[0:31] consists of all physical address bits—with optional padding—for reconstructing the system address presented to the LBC. For example, MSRCID[0:4] is valid only during CAS phases of SDRAM accesses, because the column, bank select, and (normally unused) row address bits are all present on LAD[0:31] during a CAS cycle Description 13.3 Memory Map/Register Definition Table 13-3 shows the memory mapped registers of the LBC. Undefined 4-byte address spaces within offset 0x000–0xFFF are reserved. Table 13-3. Local Bus Controller Memory Map Address Offset 0x0_5000 0x0_5008 0x0_5010 0x0_5018 0x0_5020 0x0_5028 0x0_5030 0x0_5038 BR0—Base register 0 BR1—Base register 1 BR2—Base register 2 BR3—Base register 3 BR4—Base register 4 BR5—Base register 5 BR6—Base register 6 BR7—Base register 7 Use Access R/W Reset 0x0000_nn011 0x0000_0000 Section/Page 13.3.1.1/13-11 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 13-9 Local Bus Controller Table 13-3. Local Bus Controller Memory Map (continued) Address Offset 0x0_5004 0x0_500C 0x0_5014 0x0_501C 0x0_5024 0x0_502C 0x0_5034 0x0_503C 0x0_5068 0x0_5070 0x0_5074 0x0_5078 0x0_5084 0x0_5088 0x0_5094 0x0_50A0 0x0_50A4 0x0_50B0 0x0_50B4 0x0_50B8 0x0_50BC 0x0_50C0 0x0_50D0 0x0_50D4 1 Use OR0—Options register 0 OR1—Options register 1 OR2—Options register 2 OR3—Options register 3 OR4—Options register 4 OR5—Options register 5 OR6—Options register 6 OR7—Options register 7 MAR—UPM address register MAMR—UPMA mode register MBMR—UPMB mode register MCMR—UPMC mode register MRTPR—Memory refresh timer prescaler register MDR—UPM data register LSDMR—SDRAM mode register LURT—UPM refresh timer LSRT—SDRAM refresh timer LTESR—Transfer error status register LTEDR—Transfer error disable register LTEIR—Transfer error interrupt register LTEATR—Transfer error attributes register LTEAR—Transfer error address register LBCR—Configuration register LCRR—Clock ratio register Access R/W Reset 0x0000_0FF7 0x0000_0000 Section/Page 13.3.1.2/13-12 R/W R/W R/W R/W R/W R/W R/W R/W R/W 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 13.3.1.3/13-18 13.3.1.4/13-19 13.3.1.4/13-19 13.3.1.4/13-19 13.3.1.5/13-21 13.3.1.6/13-22 13.3.1.7/13-22 13.3.1.8/13-24 13.3.1.9/13-25 Read/Bit reset 0x0000_0000 13.3.1.10/13-26 R/W R/W R/W R/W R/W R/W 0x0000_0000 13.3.1.11/13-27 0x0000_0000 13.3.1.12/13-28 0x0000_0000 13.3.1.13/13-29 0x0000_0000 13.3.1.14/13-30 0x0000_0000 13.3.1.15/13-31 0x8000_0008 13.3.1.16/13-32 Port size for BR0 is configured from external pins during reset, hence ‘nn’ is either 0x08, 0x10, or 0x18. 13.3.1 Register Descriptions This section provides a detailed description of the LBC configuration, status, and control registers with detailed bit and field descriptions. Address offsets in the LBC address range that are not defined in Table 13-3 should not be accessed for reading or writing. Similarly, only zero should be written to reserved bits of defined registers, as writing ones can have unpredictable results in some cases. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 13-10 Freescale Semiconductor Memory Map/Register Definition Bits designated as write-one-to-clear are cleared only by writing ones to them. Writing zeros to them has no effect. 13.3.1.1 Base Registers (BR0–BR7) The base registers (BRn), shown in Figure 13-2, contain the base address and address types for each memory bank. The memory controller uses this information to compare the address bus value with the current address accessed. Each register (bank) includes a memory attribute and selects the machine for memory operation handling. Note that after system reset, BR0[V] is set, BR1[V]–BR7[V] are cleared, and the value of BR0[PS] reflects the initial port size configured by the boot ROM location pins. 0 16 17 18 19 20 21 22 23 24 26 27 28 29 30 31 R BA W Reset Offset 1 — PS DECC WP MSEL ATOM V 0000_0000_0000_0000_000P_S000_0000_0000/1 1 0x0_5000 (BR0); 0x0_5008 (BR1); 0x0_5010 (BR2); 0x0_5018 (BR3); 0x0_5020 (BR4); 0x0_5028 (BR5); 0x0_5030 (BR6); 0x0_5038 (BR7); BR0 has its valid bit set during reset. Thus bank 0 is valid with the port size (PS) configured from external boot ROM configuration pins during reset. All other base registers have all bits cleared to zero during reset. Figure 13-2. Base Registers (BRn) Table 13-4 describes BRn fields. Table 13-4. BRn Field Descriptions Bits 0–16 Name BA Description Base address. The upper 17 bits of each base register are compared to the address on the address bus to determine if the bus master is accessing a memory bank controlled by the memory controller. Used with the address mask bits ORn[AM]. Reserved Port size. Specifies the port size of this memory region. For BR0, PS is configured from the boot ROM location pins during reset. For all other banks the value is reset to 00 (port size not defined). 00 Reserved 01 8-bit 10 16-bit 11 32-bit 17–18 19–20 — PS 21–22 DECC Specifies the method for data error checking. 00 Data error checking disabled, but normal parity generation 01 Normal parity generation and checking 10 Read-modify-write parity generation and normal parity checking (32-bit port size only) 11 Reserved MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 13-11 Local Bus Controller Table 13-4. BRn Field Descriptions (continued) Bits 23 Name WP Description Write protect 0 Read and write accesses are allowed. 1 Only read accesses are allowed. The memory controller does not assert LCS n on write cycles to this memory bank. WP is set (if enabled) if a write to this memory bank is attempted, and a local bus error interrupt is generated (if enabled), terminating the cycle. 24–26 MSEL Machine select. Specifies the machine to use for handling memory operations. 000 GPCM (reset value) 001 Reserved 010 Reserved 011 SDRAM 100 UPMA 101 UPMB 110 UPMC 111 Reserved 27 — Reserved 28–29 ATOM Atomic operation. Writes (reads) to the address space handled by the memory controller bank reserve the selected memory bank for the exclusive use of the accessing device. The reservation is released when the device performs a read (write) operation to this memory controller bank. If a subsequent read (write) request to this memory controller bank is not detected within 256 bus clock cycles of the last write (read), the reservation is released and an atomic error is reported (if enabled). 00 The address space controlled by this bank is not used for atomic operations. 01 Read-after-write-atomic (RAWA) 10 Write-after-read-atomic (WARA) 11 Reserved 30 31 — V Reserved Valid bit. Indicates that the contents of the BRn and ORn pair are valid. LCSn does not assert unless V is set (an access to a region that has no valid bit set may cause a bus time-out). After a system reset, only BR0[V] is set. 0 This bank is invalid. 1 This bank is valid. 13.3.1.2 Option Registers (OR0–OR7) The ORn registers define the sizes of memory banks and access attributes. The ORn attribute bits support the following three modes of operation as defined by BRn[MSEL]. • • • GPCM mode UPM mode SDRAM mode The ORn registers are interpreted differently depending on which of the three machine types is selected for that bank. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 13-12 Freescale Semiconductor Memory Map/Register Definition 13.3.1.2.1 Address Mask The address mask fields of the option registers (ORn[XAM,AM]) mask up to 19 corresponding BRn[BA,XBA] fields. The 15 lsbs of the 34-bit internal address do not participate in bank address matching in selecting a bank for access. Masking address bits independently allows external devices of different size address ranges to be used. Address mask bits can be set or cleared in any order in the field, allowing a resource to reside in more than one area of the address map. Table 13-5 shows memory bank sizes from 256 Kbytes to 4 Gbytes. Table 13-5. Memory Bank Sizes in Relation to Address Mask Bits 17–18 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 11 AM 0000_0000_0000_0000_0 1000_0000_0000_0000_0 1100_0000_0000_0000_0 1110_0000_0000_0000_0 1111_0000_0000_0000_0 1111_1000_0000_0000_0 1111_1100_0000_0000_0 1111_1110_0000_0000_0 1111_1111_0000_0000_0 1111_1111_1000_0000_0 1111_1111_1100_0000_0 1111_1111_1110_0000_0 1111_1111_1111_0000_0 1111_1111_1111_1000_0 1111_1111_1111_1100_0 1111_1111_1111_1110_0 1111_1111_1111_1111_0 1111_1111_1111_1111_1 Memory Bank Size 4 Gbytes 2 Gbytes 1 Gbyte 512 Mbytes 256 Mbytes 128 Mbytes 64 Mbytes 32 Mbytes 16 Mbytes 8 Mbytes 4 Mbytes 2 Mbytes 1 Mbyte 512 Kbytes 256 Kbytes 128 Kbytes 64 Kbytes 32 Kbytes MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 13-13 Local Bus Controller 13.3.1.2.2 Option Registers (ORn)—GPCM Mode Figure 13-3 shows the bit fields for ORn when the corresponding BRn[MSEL] selects the GPCM machine. 0 16 17 18 19 20 21 22 23 24 27 28 29 30 31 R AM W Reset Offset 1 — BCTLD CSNT ACS XACS SCY SETA TRLX EHTR EAD 0000_0000_0000_0000_0000_1111_1111_0111 1 (OR0); 0000_0000_0000_0000_0000_0000_0000_0000 (all others) 0x0_5004 (OR0); 0x0_500C (OR1); 0x0_5014 (OR2); 0x0_501C (OR3); 0x0_5024 (OR4); 0x0_502C (OR5); 0x0_5034 (OR6); 0x0_503C (OR7); OR0 has this value set during reset (GPCM is the default control machine for all banks coming out of reset). All other option registers have all bits cleared. Figure 13-3. Option Registers (ORn) in GPCM Mode Table 13-6 describes ORn fields for GPCM mode. Table 13-6. ORn—GPCM Field Descriptions Bits 0–16 Name AM Description GPCM address mask. Masks corresponding BRn bits. Masking address bits independently allows external devices of different size address ranges to be used. Address mask bits can be set or cleared in any order in the field, allowing a resource to reside in more than one area of the address map. 0 Corresponding address bits are masked. 1 Corresponding address bits are used in the comparison between base and transaction addresses. Reserved 17–18 19 — BCTLD Buffer control disable. Disables assertion of LBCTL during access to the current memory bank. 0 LBCTL is asserted upon access to the current memory bank. 1 LBCTL is not asserted upon access to the current memory bank. CSNT Chip select negation time. Determines when LCSn and LWE are negated during an external memory write access handled by the GPCM, provided that ACS ≠ 00 (when ACS = 00, only LWE is affected by the setting of CSNT). This helps meet address/data hold times for slow memories and peripherals. 0 LCSn and LWE are negated normally. 1 LCSn and LWE are negated earlier depending on the value of LCRR[CLKDIV]. 20 LCRR[CLKDIV] x 2 4 or 8 CSNT 0 1 1 Meaning LCSn and LWE are negated normally. LCSn and LWE are negated normally. LCSn and LWE are negated quarter of a bus clock cycle earlier. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 13-14 Freescale Semiconductor Memory Map/Register Definition Table 13-6. ORn—GPCM Field Descriptions (continued) Bits 21–22 Name ACS Description Address to chip-select setup. Determines the delay of the LCS n assertion relative to the address change when the external memory access is handled by the GPCM. At system reset, OR0[ACS] = 11. LCRR[CLKDIV] x Value 00 01 Meaning LCSn is output at the same time as the address lines. Note that this overrides the value of CSNT such that CSNT=0. Reserved. LCSn is output a half bus clock cycle after the address lines. LCSn is output a half bus clock cycle after the address lines. LCSn is output a quarter bus clock cycle after the address lines. LCSn is output a half bus clock cycle after the address lines. 2 10 11 4 or 8 10 11 23 XACS Extra address to chip-select setup. Setting this bit increases the delay of the LCSn assertion relative to the address change when the external memory access is handled by the GPCM. After a system reset, OR0[XACS] = 1. 0 Address to chip-select setup is determined by ORx[ACS] and LCRR[CLKDIV]. 1 Address to chip-select setup is extended (see Table 13-23 and Table 13-24 for LCRR[CLKDIV] = 4 or 8, Table 13-25 and Table 13-26 for LCRR[CLKDIV] = 2). SCY Cycle length in bus clocks. Determines the number of wait states inserted in the bus cycle, when the GPCM handles the external memory access. Thus it is the main parameter for determining cycle length. The total cycle length depends on other timing attribute settings. After a system reset, OR0[SCY] = 1111. 0000 No wait states 0001 1 bus clock cycle wait state ... 1111 15 bus clock cycle wait states External address termination 0 Access is terminated internally by the memory controller unless the external device asserts LGTA earlier to terminate the access. 1 Access is terminated externally by asserting the LGTA external pin. (Only LGTA can terminate the access). Timing relaxed. Modifies the settings of timing parameters for slow memories or peripherals. 0 Normal timing is generated by the GPCM. 1 Relaxed timing on the following parameters: •Adds an additional cycle between the address and control signals (only if ACS ≠ 00) •Doubles the number of wait states specified by SCY, providing up to 30 wait states •Works in conjunction with EHTR to extend hold time on read accesses •LCSn (only if ACS ≠ 00) and LWE signals are negated one cycle earlier during writes. 24–27 28 SETA 29 TRLX MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 13-15 Local Bus Controller Table 13-6. ORn—GPCM Field Descriptions (continued) Bits 30 Name EHTR Description Extended hold time on read accesses. Indicates with TRLX how many cycles are inserted between a read access from the current bank and the next access. TRLX 0 0 1 1 31 EAD EHTR 0 1 0 1 Meaning The memory controller generates normal timing. No additional cycles are inserted. 1 idle clock cycle is inserted. 4 idle clock cycles are inserted. 8 idle clock cycles are inserted. External address latch delay. Allow extra bus clock cycles when using external address latch (LALE). 0 No additional bus clock cycles (LALE asserted for one bus clock cycle only) 1 Extra bus clock cycles are added (LALE is asserted for the number of bus clock cycles specified by LCRR[EADC]). 13.3.1.2.3 Option Registers (ORn)—UPM Mode Figure 13-4 shows the bit fields for ORn when the corresponding BRn[MSEL] selects a UPM machine. 0 18 19 20 22 23 24 28 29 30 31 R AM W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x0_5004 (OR0); 0x0_500c (OR1); 0x0_5014 (OR2); 0x0_501c (OR3); 0x0_5024 (OR4); 0x0_502c (OR5); 0x0_5034 (OR6); 0x0_503c (OR7); BCTLD BI TRLX EHTR EAD Figure 13-4. Option Registers (ORn) in UPM Mode Table 13-7 describes BRn fields for UPM mode. Table 13-7. ORn—UPM Field Descriptions Bits 0–16 Name AM Description UPM address mask. Masks corresponding BRn bits. Masking address bits independently allows external devices of different size address ranges to be used. Address mask bits can be set or cleared in any order in the field, allowing a resource to reside in more than one area of the address map. 0 Corresponding address bits are masked. 1 The corresponding address bits are used in the comparison with address pins. Reserved 17–18 — MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 13-16 Freescale Semiconductor Memory Map/Register Definition Table 13-7. ORn—UPM Field Descriptions (continued) Bits 19 Name Description BCTLD Buffer control disable. Disables assertion of LBCTL during access to the current memory bank. 0 LBCTL is asserted upon access to the current memory bank. 1 LBCTL is not asserted upon access to the current memory bank. — BI Reserved Burst inhibit. Indicates if this memory bank supports burst accesses. 0 The bank supports burst accesses. 1 The bank does not support burst accesses. The selected UPM executes burst accesses as a series of single accesses. Reserved Timing relaxed. Works in conjunction with EHTR to extend hold time on read accesses. Extended hold time on read accesses. Indicates with TRLX how many cycles are inserted between a read access from the current bank and the next access. TRLX EHTR 0 0 1 1 0 1 0 1 Meaning The memory controller generates normal timing. No additional cycles are inserted. 1 idle clock cycle is inserted. 4 idle clock cycles are inserted. 8 idle clock cycles are inserted. 20–22 23 24–28 29 30 — TRLX EHTR 31 EAD External address latch delay. Allow extra bus clock cycles when using external address latch (LALE). 0 No additional bus clock cycles (LALE asserted for one bus clock cycle only) 1 Extra bus clock cycles are added (LALE is asserted for the number of bus clock cycles specified by LCRR[EADC]). 13.3.1.2.4 Option Registers (ORn)—SDRAM Mode Figure 13-5 shows the bit fields for ORn when the corresponding BRn[MSEL] selects the SDRAM machine. 0 16 17 18 19 21 22 23 25 26 27 28 30 31 R AM W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x0_5004 (OR0); 0x0_500c (OR1); 0x0_5014 (OR2); 0x0_501c (OR3); 0x0_5024 (OR4); 0x0_502c (OR5); 0x0_5034 (OR6); 0x0_503c (OR7); COLS ROWS PMSEL EAD Figure 13-5. Option Registers (ORn) in SDRAM Mode MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 13-17 Local Bus Controller Table 13-8 describes BRn fields for SDRAM mode. Table 13-8. ORn—SDRAM Field Descriptions Bits 0–16 Name AM Description SDRAM address mask. Masks corresponding BR n bits. Masking address bits independently allows external devices of different size address ranges to be used. Address mask bits can be set or cleared in any order in the field, allowing a resource to reside in more than one area of the address map. AM can be read or written at any time. 0 Corresponding address bits are masked. 1 The corresponding address bits are used in the comparison with address pins. Reserved Number of column address 100 000 7 101 001 8 110 010 9 111 011 10 Reserved lines. Sets the number of column address lines in the SDRAM device. 11 12 13 14 17–18 19–21 — COLS 22 23–25 — ROWS Number of row address lines. Sets the number of row address lines in the SDRAM device. 100 13 000 9 101 14 001 10 110 15 010 11 111 Reserved 011 12 PMSEL Page mode select. Selects page mode for the SDRAM connected to the memory controller bank. 0 Back-to-back page mode (normal operation). Page is closed when the bus becomes idle. 1 Page is kept open until a page miss or refresh occurs. — EAD Reserved External address latch delay. Allow extra bus clock cycles when using external address latch (LALE). 0 No additional bus clock cycles (LALE asserted for one bus clock cycle only) 1 Extra bus clock cycles are added (LALE is asserted for the number of bus clock cycles specified by LCRR[EADC]). 26 27–30 31 13.3.1.3 UPM Memory Address Register (MAR) Figure 13-6 shows the fields of the UPM memory address register (MAR). 0 31 R A W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x0_5068 Figure 13-6. UPM Memory Address Register (MAR) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 13-18 Freescale Semiconductor Memory Map/Register Definition Table 13-9 describes the MAR field. Table 13-9. MAR Field Description Bits 0–31 Name A Description Address that can be output to the address signals under control of the AMX bits in the UPM RAM word. 13.3.1.4 UPM Mode Registers (MxMR) The UPM machine mode registers (MAMR, MBMR and MCMR), shown in Figure 13-7, contain the configuration for the three UPMs. 0 1 2 3 4 5 7 8 9 10 12 13 14 17 18 21 22 25 26 31 R RFEN W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x0_5070 (MAMR); 0x0_5074 (MBMR); 0x0_5078 (MCMR) OP UWPL AM DS G0CL GPL4 RLF WLF TLF MAD Figure 13-7. UPM Mode Registers (MxMR) Table 13-10 describes UPM mode fields. Table 13-10. MxMR Field Descriptions Bits 0 1 Name — Reserved Description RFEN Refresh enable. Indicates that the UPM needs refresh services. This bit must be set for UPMA (refresh executor) if refresh services are required on any UPM assigned chip selects. If MAMR[RFEN] = 0, no refresh services can be provided, even if UPMB and/or UPMC have their RFEN bit set. 0 Refresh services are not required 1 Refresh services are required OP Command opcode. Determines the command executed by the UPM n when a memory access hits a UPM assigned bank. See Section 13.4.4.2, “Programming the UPMs,” for important programming considerations. 00 Normal operation 01 Write to UPM array. On the next memory access that hits a UPM assigned bank, write the contents of the MDR into the RAM location pointed to by MAD. After the access, MAD is automatically incremented. 10 Read from UPM array. On the next memory access that hits a UPM assigned bank, read the contents of the RAM location pointed to by MAD into the MDR. After the access, MAD is automatically incremented. 11 Run pattern. On the next memory access that hits a UPM assigned bank, run the pattern written in the RAM array. The pattern run starts at the location pointed to by MAD and continues until the LAST bit is set in the RAM word. 2–3 4 UWPL LUPWAIT polarity active low. Sets the polarity of the LUPWAIT pin when in UPM mode. 0 LUPWAIT is active high. 1 LUPWAIT is active low. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 13-19 Local Bus Controller Table 13-10. MxMR Field Descriptions (continued) Bits 5–7 Name AM Description Address multiplex size. Determines how the address of the current memory cycle can be output on the address pins. This field is needed when interfacing with devices requiring row and column addresses multiplexed on the same pins. Value LA0–LA15 000 001 010 011 100 101 110 111 8–9 DS 0 0 0 0 0 0 LA16 A8 A7 A6 A5 A4 A3 LA17 A9 A8 A7 A6 A5 A4 LA18 A10 A9 A8 A7 A6 A5 LA19–LA28 A11–A20 A10–A19 A9–A18 A8–A17 A7–A16 A6–A15 LA29 A21 A20 A19 A18 A17 A16 LA30 A22 A21 A20 A19 A18 A17 LA31 A23 A22 A21 A20 A19 A18 Reserved Reserved Disable timer period. Guarantees a minimum time between accesses to the same memory bank controlled by UPMn. The disable timer is turned on by the TODT bit in the RAM array word, and when expired, the UPMn allows the machine access to handle a memory pattern to the same bank. Accesses to a different bank by the same UPM n is also allowed. To avoid conflicts between successive accesses to different banks, the minimum pattern in the RAM array for a request serviced, should not be shorter than the period established by DS. 00 1-bus clock cycle disable period 01 2-bus clock cycle disable period 10 3-bus clock cycle disable period 11 4-bus clock cycle disable period 10–12 G0CL General line 0 control. Determines which logical address line can be output to the LGPL0 pin when the UPMn is selected to control the memory access. 000 A12 001 A11 010 A10 011 A9 100 A8 101 A7 110 A6 111 A5 GPL4 LGPL4 output line disable. Determines how the LGPL4/LUPWAIT pin is controlled by the corresponding bits in the UPMn array. See Table 13-30 on page 13-69. Value LGPL4/LUPWAIT Pin Function LGPL4 (output) LUPWAIT (input) Interpretation of UPM Word Bits G4T1/DLT3 G4T1 DLT3 G4T3/WAEN G4T3 WAEN 13 0 1 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 13-20 Freescale Semiconductor Memory Map/Register Definition Table 13-10. MxMR Field Descriptions (continued) Bits 14–17 Name RLF Description Read loop field. Determines the number of times a loop defined in the UPMn will be executed for a burst- or single-beat read pattern or when MxMR[OP] = 11 (RUN command) 0000 16 0001 1 0010 2 0011 3 ... 1110 14 1111 15 Write loop field. Determines the number of times a loop defined in the UPMn will be executed for a burst- or single-beat write pattern. 0000 16 0001 1 0010 2 0011 3 ... 1110 14 1111 15 Refresh loop field. Determines the number of times a loop defined in the UPMn will be executed for a refresh service pattern. 0000 16 0001 1 0010 2 0011 3 ... 1110 14 1111 15 Machine address. RAM address pointer for the command executed. This field is incremented by 1 each time the UPM is accessed, and the OP field is set to WRITE or READ. Address range is 64 words per UPMn. 18–21 WLF 22–25 TLF 26–31 MAD 13.3.1.5 Memory Refresh Timer Prescaler Register (MRTPR) The refresh timer prescaler register (MRTPR), shown in Figure 13-8, is used to divide the system clock to provide the SDRAM and UPM refresh timers clock. 0 7 8 31 R PTP W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x0_5084 Figure 13-8. Memory Refresh Timer Prescaler Register (MRTPR) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 13-21 Local Bus Controller Table 13-11 describes MRTPR fields. Table 13-11. MRTPR Field Descriptions Bits 0–7 8–31 Name PTP — Description Refresh timers prescaler. Determines the period of the refresh timers input clock. The system clock is divided by PTP except when the value is 0000_0000, which represents the maximum divider of 256. Reserved 13.3.1.6 UPM Data Register (MDR) The memory data register (MDR), shown in Figure 13-9, contains data written to or read from the RAM array for UPM read or write commands. MDR must be set up before issuing a write command to the UPM. 0 31 R D W Reset Offset 0000_0000_0000_0000 0x0_5088 Figure 13-9. UPM Data Register (MDR) Table 13-12 describes MDR[D]. Table 13-12. MDR Field Description Bits 0–31 Name D Description The data to be read or written into the RAM array when a write or read command is supplied to the UPM (MxMR[OP] = 01 or MxMR[OP] = 10). 13.3.1.7 SDRAM Machine Mode Register (LSDMR) The local bus SDRAM mode register (LSDMR), shown in Figure 13-10, is used to configure operations pertaining to SDRAM. 0 1 2 45 78 10 11 13 14 15 16 17 19 20 22 23 24 25 26 27 28 29 30 31 R RFEN OP W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x0_5094 BSMA RFCR RFCR PRETOACT ACTTORW BL WRC BUFCMD CL Figure 13-10. SDRAM Machine Mode Register (LSDMR) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 13-22 Freescale Semiconductor Memory Map/Register Definition Table 13-12 describes LSDMR fields. Table 13-13. LSDMR Field Descriptions Bits 0 1 Name — RFEN Reserved Refresh enable. Indicates that the SDRAM requires refresh services. 0 Refresh services are not required. 1 Refresh services are required. SDRAM operation. Selects the operation that occurs when the SDRAM device is accessed. Value 000 001 010 011 100 101 110 111 5–7 8–10 — BSMA Reserved Bank select multiplexed address line. Selects which address pins serve as the 2-bit bank-select address for SDRAM. Note that only 4-bank SDRAMs are supported. 100 LA16:LA17 000 LA12:LA13 101 LA17:LA18 001 LA13:LA14 110 LA18:LA19 010 LA14:LA15 111 LA19:LA20 011 LA15:LA16 Reserved Refresh recovery. Sets the refresh recovery interval in bus clock cycles. Defines the earliest timing for an ACTIVATE or REFRESH command after a REFRESH command. 100 6 clocks 000 Reserved 101 7 clocks 001 3 clocks 110 8 clocks 010 4 clocks 111 16 clocks 011 5 clocks Meaning Normal operation Auto refresh Self refresh Mode Register write Precharge bank Precharge all banks Activate bank Use Normal operation Initialization Power-down mode or debug Initialization Debug Initialization Debug Description 2–4 OP Read/write without valid data transfer Debug 11–13 14–16 — RFCR 17–19 PRETOACT Defines the earliest timing for ACTIVATE or REFRESH command after a PRECHARGE command (number of bus clock cycle wait states). 100 4 000 8 101 5 001 1 110 6 010 2 111 7 011 3 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 13-23 Local Bus Controller Table 13-13. LSDMR Field Descriptions (continued) Bits 20–22 Name Description ACTTORW Defines the earliest timing for READ/WRITE command after an ACTIVATE command (number of bus clock cycle wait states). 100 4 000 8 101 5 001 Reserved 110 6 010 2 111 7 011 3 BL Sets the burst length for SDRAM accesses. 0 SDRAM burst length is 4. Use this value if the device port size is16. 1 SDRAM burst length is 8. Use this value if the device port size is 32 or 8. Reserved Write recovery time. Defines the earliest timing for PRECHARGE command after the last data is written to the SDRAM. 00 4 01 Reserved 10 2 11 3 Reserved Control line assertion timing. If external buffers are placed on the control lines going to both the SDRAM and address lines, setting BUFCMD causes all SDRAM control lines except LCSn, LCKE, LALE and LSDDQM [0:3] to be asserted for LCRR[BUFCMDC] cycles, instead of one. 0 Normal timing for the control lines 1 All control lines except LCS n are asserted for the number of cycles specified by LCRR[BUFCMDC]. CAS latency. Defines the timing for first read data after SDRAM samples a column address. 00 Extended CAS latency. According to LCRR[ECL]. See Table 13-22 on page 13-32. 01 1 10 2 11 3 23 24–25 26–27 — WRC 28 29 — BUFCMD 30–31 CL 13.3.1.8 UPM Refresh Timer (LURT) The UPM refresh timer (LURT), shown in Figure 13-11, generates a refresh request for all valid banks that selected a UPM machine and are refresh-enabled (MxMR[RFEN] = 1). Each time the timer expires, a qualified bank generates a refresh request using the selected UPM. The qualified banks rotate their requests. 0 7 8 31 R LURT W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x0_50A0 Figure 13-11. UPM Refresh Timer (LURT) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 13-24 Freescale Semiconductor Memory Map/Register Definition Table 13-14 describes LURT fields. Table 13-14. LURT Field Descriptions Bits 0–7 Name Description LURT UPM refresh timer period. Determines, along with the timer prescaler (MRTPR), the timer period according to the following equation: LURT TimerPeriod = --------------------------------------------Fsystemclock ⎛ --------------------------------------- ⎞ ⎝ MRTPR [ PTP ]⎠ Example: For a 266-MHz system clock and a required service rate of 15.6 µs, given MRTPR[PTP] = 32, the LURT value should be 128 decimal. 128/(266 MHz/32) = 15.4 µs, which is less than the required service period of 15.6 µs. Note that the reset value (0x00) sets the maximum period to 256 x MRTPR[PTP] system clock cycles. 8–31 — Reserved 13.3.1.9 SDRAM Refresh Timer (LSRT) The SDRAM refresh timer (LSRT), shown in Figure 13-12, generates a refresh request for all valid banks that selected a SDRAM machine and are refresh-enabled (LSDMR[RFEN] = 1). Each time the timer expires, all qualifying banks generate a bank staggering auto-refresh request using the SDRAM machine. 0 7 8 31 R LSRT W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x0_50A4 Figure 13-12. LSRT SDRAM Refresh Timer (LSRT) Table 13-15 describes LSRT fields. Table 13-15. LSRT Field Descriptions Bits 0–7 Name LSRT Description SDRAM refresh timer period. Determines, along with the timer prescaler (MRTPR), the timer period according to the following equation: LSRT TimerPeriod = --------------------------------------------⎛ Fsystemclock⎞ --------------------------------------⎝ MRTPR [ PTP ]⎠ Example: For a 266-MHz system clock and a required service rate of 15.6 µs, given PTP = 32, the LSRT value should be 128 decimal. 128/(266 MHz/32) = 15.4 µs, which is less than the required service period of 15.6 µs. Note that the reset value, 0x00, sets the maximum period to 256 × M RTPR[PTP] system clock cycles. 8–31 — Reserved MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 13-25 Local Bus Controller 13.3.1.10 Transfer Error Status Register (LTESR) The LBC has five registers for error management. • • • • • The transfer error status register (LTESR) indicates the cause of an error. The transfer error check disable register (LTEDR) is used to enable (and disable) error checking. The transfer error check interrupt register (LTEIR) enables reporting of errors through an interrupt. The transfer error attributes register (LTEATR) captures source attributes of an error. The transfer error address register (LTEAR) captures the address of a transaction that caused an error. LTESR, shown in Figure 13-13, is a write-one-to-clear register. Reading LTESR occurs normally; however, write operations can clear but not set bits. A bit is cleared whenever the register is written and the data in the corresponding bit location is a 1. For example, to clear only the write protect error bit (LTESR[WP]) without affecting other LTESR bits, 0x0400_0000 should be written to the register. 0 1 2 3 4 5 6 7 8 9 10 11 12 13 31 R BM W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x0_50B0 PAR WP ATMW ATMR CS Figure 13-13. Transfer Error Status Register (LTESR) Table 13-16 describes LTESR fields. Table 13-16. LTESR Field Descriptions Bits 0 Name BM Description Bus monitor time-out 0 No local bus monitor time-out occurred. 1 Local bus monitor time-out occurred. No data beat was acknowledged on the bus within LBCR[BMT] × 8 bus clock cycles from the start of a transaction. Reserved Parity 0 No local bus parity error 1 Local bus parity error. LTEATR[PB] indicates the byte lane that caused the error and LTEATR[BNK] indicates which memory controller bank was accessed. Reserved 1 2 — PAR 3–4 — MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 13-26 Freescale Semiconductor Memory Map/Register Definition Table 13-16. LTESR Field Descriptions (continued) Bits 5 Name WP Description Write protect error 0 No write protect error occurred. 1 A write was attempted to a local bus memory region that was defined as read-only in the memory controller. Usually, in this case, a bus monitor time-out will occur (as the cycle is not automatically terminated). Reserved 6–7 8 — ATMW Atomic error write 0 No atomic write error occurred. 1 The subsequent write (WARA) to a memory bank did not occur within 256 bus clock cycles. ATMR Atomic error read 0 No atomic read error occurred. 1 The subsequent read (RAWA) to a memory bank did not occur within 256 bus clock cycles. — CS Reserved Chip select error 0 No chip select error occurred. 1 A transaction was sent to the LBC that did not hit any memory bank. Reserved 9 10–11 12 13–31 — 13.3.1.11 Transfer Error Check Disable Register (LTEDR) The transfer error check disable register (LTEDR), shown in Figure 13-14, is used to disable error checking. Note that control of error checking is independent of control of reporting of errors (LTEIR) through the interrupt mechanism. 0 1 2 3 4 5 6 7 8 9 10 11 12 13 31 R BMD W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x0_50B4 PARD WPD WARA RAWA CSD Figure 13-14. Transfer Error Check Disable Register (LTEDR) Table 13-17 describes LTEDR fields. Table 13-17. LTEDR Field Descriptions Bits 0 Name BMD Bus monitor disable 0 Bus monitor is enabled 1 Bus monitor is disabled Reserved Description 1 — MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 13-27 Local Bus Controller Table 13-17. LTEDR Field Descriptions (continued) Bits 2 Name Description PARD Parity error checking disabled. Note that uncorrectable read errors may cause the assertion of core_fault_in, which causes the core to generate a machine check interrupt, unless it is disabled (by clearing HID1[RFXE]). If RFXE is zero and this error occurs, PARD must be cleared and LTEIR[PARI] must be set to ensure that an interrupt is generated. For more information, see Section 6.10.2, “Hardware Implementation-Dependent Register 1 (HID1).” 0 Parity error checking is enabled. 1 Parity error checking is disabled. — Reserved 3–4 5 WPD Write protect error checking disable 0 Write protect error checking is enabled. 1 Write protect error checking is disabled. — Reserved 6–7 8 WARA Write-after-read atomic (WARA) error checking disable 0 WARA error checking is enabled. 1 WARA error checking is disabled. RAWA Read-after-write atomic (RAWA) error checking disable 0 RAWA error checking is enabled. 1 RAWA error checking is disabled. — CSD Reserved Chip select error checking disable 0 Chip select error checking is enabled. 1 Chip select error checking is disabled. Reserved 9 10–11 12 13–31 — 13.3.1.12 Transfer Error Interrupt Enable Register (LTEIR) The transfer error interrupt enable register (LTEIR), shown in Figure 13-15, is used to send or block error reporting through the LBC internal interrupt mechanism. Software should clear pending errors in LTESR before enabling interrupts. After an interrupt has occurred, clearing relevant LTESR error bits negates the interrupt. 0 1 2 3 4 5 6 7 8 9 10 11 12 13 31 R BMI W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x0_50B8 PARI WPI WARA RAWA CSI Figure 13-15. Transfer Error Interrupt Enable Register (LTEIR) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 13-28 Freescale Semiconductor Memory Map/Register Definition Table 13-18 describes LTEIR fields. Table 13-18. LTEIR Field Descriptions Bits 0 Name BMI Bus monitor error interrupt enable 0 Bus monitor error reporting is disabled. 1 Bus monitor error reporting is enabled. Reserved Parity error interrupt enable. Note that uncorrectable read errors may cause the assertion of core_fault_in, which causes the core to generate a machine check interrupt, unless it is disabled (by clearing HID1[RFXE]). If RFXE is zero and this error occurs, LTEDR[PARD] must be cleared and PARI must be set to ensure that an interrupt is generated. For more information, see Section 6.10.2, “Hardware Implementation-Dependent Register 1 (HID1).” 0 Parity error reporting is disabled. 1 Parity error reporting is enabled. Reserved Write protect error interrupt enable 0 Write protect error reporting is disabled. 1 Write protect error reporting is enabled. Reserved Description 1 2 — PARI 3–4 5 — WPI 6–7 8 — WARA Write-after-read atomic (WARA) error interrupt enable 0 WARA error reporting is disabled. 1 WARA error reporting is enabled. RAWA Read-after-write atomic (RAWA) error interrupt enable 0 RAWA error reporting is disabled. 1 RAWA error reporting is enabled. — CSI Reserved Chip select error interrupt enable 0 Chip select error reporting is disabled. 1 Chip select error reporting is enabled. Reserved 9 10–11 12 13–31 — 13.3.1.13 Transfer Error Attributes Register (LTEATR) Figure 13-16 shows the LTEATR. After LTEATR[V] has been set, software must clear this bit to allow LTEATR and LTEAR to update following any subsequent errors. 0 2 3 4 10 11 15 16 19 20 27 28 29 30 31 R RWB W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x0_50BC SRCID PB BNK XA V Figure 13-16. Transfer Error Attributes Register (LTEATR) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 13-29 Local Bus Controller Table 13-19 describes LTEATR fields. Table 13-19. LTEATR Field Descriptions Bits 0–2 3 Name — RWB Reserved Transaction type for the error 0 The transaction for the error was a write transaction. 1 The transaction for the error was a read transaction. Reserved Description 4–10 — 11–15 SRCID Captures the source of the transaction when this information is provided on the internal interface to the LBC. 16–19 20–27 PB BNK Parity error on byte. There are four parity error status bits, one per byte lane. A bit is set for the byte that had a parity error (bit 16 represents byte 0, the most significant byte lane). Memory controller bank. There is one error status bit per memory controller bank (bit 20 represents bank 0). A bit is set for the local bus memory controller bank that had an error. Note that BNK is invalid if the error was not caused by parity checks. Extended address for the error. These bits capture the two msbs of the 34-bit address of the transaction resulting in an error. Reserved Error attribute capture is valid. Indicates that the captured error information is valid 0 Captured error attributes and address are not valid 1 Captured error attributes and address are valid 28–29 30 31 XA — V 13.3.1.14 Transfer Error Address Register (LTEAR) The transfer error address register (LTEAR) is shown in Figure 13-17. 0 31 R A W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x0_50C0 Figure 13-17. Transfer Error Address Register (LTEAR) Table 13-20 describes LTEAR[A]. Table 13-20. LTEAR Field Descriptions Bits 0–31 Name A Description Transaction address for the error. Holds the 32 lsbs of the 34-bit address of the transaction resulting in an error. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 13-30 Freescale Semiconductor Memory Map/Register Definition 13.3.1.15 Local Bus Configuration Register (LBCR) The local bus configuration register (LBCR) is shown in Figure 13-18. 0 1 7 8 9 10 11 12 13 14 15 16 23 24 31 R LDIS W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x0_50D0 BTCLC AHD LPBSE EPAR BMT Figure 13-18. Local Bus Configuration Register Table 13-21 describes LBCR fields. Table 13-21. LBCR Field Descriptions Bits 0 Name LDIS Description Local bus disable 0 Local bus is enabled 1 Local bus is disabled. No internal transactions will be acknowledged. Reserved 1–7 8–9 — BCTLC Defines the use of LBCTL 00 LBCTL is used as W/R control for GPCM or UPM accesses (buffer control). 01 LBCTL is used as LOE for GPCM accesses only. 10 LBCTL is used as LWE for GPCM accesses only. 11 Reserved. AHD Address hold disable. Removes part of the hold time for LAD with respect to LALE in order to lengthen the LALE pulse 0 During address phases on the local bus, the LALE signal negates one platform clock period prior to the address being invalidated. At 333 MHz, this provides 3 ns of additional address hold time at the external address latch. 1 During address phases on the local bus, the LALE signal negates half of one platform clock period prior to the address being invalidated. This halves the address hold time, but extends the latch enable duration. This may be necessary for very high frequency designs. Reserved 10 11–13 14 — LPBSE Enables parity byte select on LGTA/LGPL4/LUPWAIT/LPBSE pin. 0 Parity byte select is disabled. LGTA/LGPL4/LUPWAIT/LPBSE pin is available for memory control as LGPL4 (output) or LGTA/LUPWAIT (input). 1 Parity byte select is enabled. LGTA/LGPL4/LUPWAIT/LPBSE pin is dedicated as the parity byte select output, and LGTA/LUPWAIT is disabled. EPAR Determines odd or even parity. Writing the memory with EPAR = 1 and reading the memory with EPAR = 0 generates parity errors for testing. 0 Odd parity 1 Even parity 15 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 13-31 Local Bus Controller Table 13-21. LBCR Field Descriptions (continued) Bits 16–23 Name BMT Description Bus monitor timing. Defines the bus monitor time-out period. Clearing BMT (reset value) selects the maximum count of 2048 bus clock cycles. For non-zero values of BMT, the number of LCLK clock cycles to count down before a time-out error is generated is given by: bus cycles = BMT x 8. Apart from BMT = 0x00, the minimum value of BMT is 5, corresponding with 40 bus cycles. Shorter time-outs may result in spurious errors during SDRAM operation. Reserved 24–31 — 13.3.1.16 Clock Ratio Register (LCRR) The clock ratio register sets the system (CCB) clock to LBC bus frequency ratio. It also provides configuration bits for extra delay cycles for address and control signals. 0 1 2 3 4 5 6 7 8 13 14 15 16 27 28 31 R DBYP W Reset Offset 1000_0000_0000_0000_0000_0000_0000_1000 0x0_50D4 BUFCMDC ECL EADC CLKDIV Figure 13-19. Clock Ratio Register (LCRR) Table 13-22 describes LCRR fields. Table 13-22. LCRR Field Descriptions Bits 0 Name DBYP Description DLL bypass. This bit should be set when using low bus clock frequencies if the DLL is unable to lock. When in DLL bypass mode, incoming data is captured in the middle of the bus clock cycle. 0 The DLL is enabled. 1 The DLL is bypassed. Reserved 1 2–3 — BUFCMDC Additional delay cycles for SDRAM control signals. Defines the number of cycles to be added for each SDRAM command when LSDMR[BUFCMD] = 1. 00 4 01 1 10 2 11 3 — ECL Reserved Extended CAS latency. Determines the extended CAS latency for SDRAM accesses when LSDMR[CL] = 00. 00 4 01 5 10 6 11 7 Reserved 4–5 6–7 8–13 — MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 13-32 Freescale Semiconductor Functional Description Table 13-22. LCRR Field Descriptions (continued) Bits 14–15 Name EADC Description External address delay cycles. Defines the number of cycles for the assertion of LALE. Note that LALE negates prior to the end of the final local bus clock, as controlled by LBCR[AHD]. 00 4 01 1 10 2 11 3 Reserved System (CCB) clock divider. Sets the frequency ratio between the system (CCB) clock and the memory bus clock. Only the values shown in the table below are allowed. 0000–0001Reserved 0010 2 0011 Reserved 0100 4 0101–0111Reserved 1000 8 1001–1111Reserved 16–27 28–31 — CLKDIV 13.4 Functional Description The LBC allows the implementation of memory systems with very specific timing requirements. • The SDRAM machine provides an interface to SDRAMs using bank interleaving and back-to-back page mode to achieve high performance through a multiplexed address/data bus. An internal DLL for bus clock generation ensures improved data set-up margins for board designs. The GPCM provides interfacing for simpler, lower-performance memories and memory-mapped devices. It has inherently lower performance because it does not support bursting. For this reason, GPCM-controlled banks are used primarily for boot-loading and access to low-performance memory-mapped peripherals. The UPM supports refresh timers, address multiplexing of the external bus and generation of programmable control signals for row address and column address strobes, to allow for a minimal glue logic interface to DRAMs, burstable SRAMs, and almost any other kind of peripheral. The UPM can be used to generate flexible, user-defined timing patterns for control signals that govern a memory device. These patterns define how the external control signals behave during a read, write, burst-read, or burst-write access. Refresh timers are also available to periodically initiate user-defined refresh patterns. • • MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 13-33 Local Bus Controller Internal Memory Access Request Select 34-bit System Address Address Comparator Bank Select SDRAM Machine UPM A/B/C GPCM MSEL Field Signals Timing Generator MUX 32-Bit Physical RAM Address (A) External Signals Figure 13-20. Basic Operation of Memory Controllers in the LBC Each memory bank (chip select) can be assigned to any one of these three type of machines through the machine select bits of the base register for that bank (BRn[MSEL]), as illustrated in Figure 13-20. If a bank match occurs, the corresponding machine (GPCM, SDRAM or UPM) then takes ownership of the external signals that control the access and maintains control until the transaction ends. 13.4.1 Basic Architecture The following sections describe the basic architecture of the LBC. 13.4.1.1 Address and Address Space Checking The defined base addresses are written to the BRn registers, while the corresponding address masks are written to the ORn registers. Each time a local bus access is requested, the internal transaction address is compared with each bank. Addresses are decoded by comparing the 19 msbs of the address, masked by ORn[XAM] and ORn[AM], with the base address for each bank (BRn[XBA] and BRn[BA]). If a match is found on a memory controller bank, the attributes defined in the BRn and ORn for that bank are used to control the memory access. If a match is found in more than one bank, the lowest-numbered bank handles the memory access (that is, bank 0 has priority over bank 1). MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 13-34 Freescale Semiconductor Functional Description 13.4.1.2 External Address Latch Enable Signal (LALE) The local bus uses a multiplexed address/data bus. Therefore the LBC must distinguish between address and data phases, which take place on the same bus (LAD[0:31] pins). The LALE signal, when asserted, signifies an address phase during which the LBC drives the memory address on the LAD[0:31] pins. An external address latch uses this signal to capture the address and provide it to the address pins of the memory or peripheral device. When LALE is negated, LAD[0:31] then serves as the (bi-directional) data bus for the access. Any address phase initiates the assertion of LALE, which has a programmable duration of between 1 and 4 bus clock cycles. To ensure adequate hold time on the external address latch, LALE negates earlier than the address changes on LAD[0:31] during address phases. By default, LALE negates earlier by an amount equal to the platform clock period (which, divided by LCRR[CLKDIV], yields the bus clock). For example, if LBC is operating at 333 MHz internally, then an additional 3 ns of address hold time is introduced. However, when LCRR[CLKDIV] = 2 and the LCLK frequency exceeds 100 MHz, the duration of the shortened LALE pulse may not meet the minimum latch enable pulse width specifications of some latches. In such cases, setting LBCR[AHD] = 1 increases the LALE pulse width by half of one platform clock cycle, and decreases the address hold time by the same amount. At 333 MHz and with LCRR[CLKDIV] = 2, the duration of LALE would then be 4.5 ns, with 1.5 ns of hold time. If both longer hold time and longer LALE pulse duration are needed, then the address phase can be extended using the ORn[EAD] and LCRR[EADC] fields, and the LBCR[AHD] bit can be left at 0. However, this will add latency to all address tenures. The frequency of LALE assertion varies across the three memory controllers. For GPCM, every assertion of LCSn is considered an independent access, and accordingly, LALE asserts prior to each such access. For example, GPCM driving an 8-bit port would assert LALE and LCSn 32 times in order to satisfy a 32-byte cache line transfer. The SDRAM controller asserts LALE only to initiate a burst transfer with a starting address, therefore no more than one assertion of LALE may be required for SDRAM to transfer a 32-byte cache line through a 32-bit port. In the case of UPM, the frequency of LALE assertion depends on how the UPM RAM is programmed. UPM single accesses typically assert LALE once, upon commencement, but it is possible to program UPM to assert LALE several times, and to change the values of LA[27:31] with and without LALE being involved. In general, when using the GPCM and SDRAM controllers it is not necessary to use LA[27:31] if a sufficiently wide latch is used to capture the entire address during LALE phases. UPM may require LA[27:31] if the LBC is generating its own burst address sequence. To illustrate how a large transaction is handled by the LBC, Figure 13-21 shows LBC signals for GPCM performing a 32-byte write starting at address 0x5420. Note that during each of the 32 assertions of LALE, LA[27:31] exactly mirror LAD[27:31], but during data phases, only LAD[0:7] and LDP[0] are driven with valid data and parity, respectively. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 13-35 Local Bus Controller LCLK LALE LCSn LWE LA[27:31] 00 00 54 20 0 0 P(B0) D(B0) 00 54 21 01 D(B1) 00 54 22 P(B1) 0 0 02 D(B2) 03 00 54 23 1C 1D 00 D(B29) 54 3D 0 0 P(B29) 1E 00 D(B30) 54 3E 0 0 P(B30) 1F 00 D(B31) 54 3F 0 0 P(B31) LAD[0:7] LAD[8:31] LDP[0] 0 0 P(B2) 0 0 LDP[1:3] Note: All address and signal values are shown in hexadecimal. D(Bk) = kth of 32 data bytes, P(Bk) = parity bit of k th data byte. Figure 13-21. Example of 8-Bit GPCM Writing 32 Bytes to Address 0x5420 13.4.1.3 Data Transfer Acknowledge (TA) The three memory controllers in the LBC generate an internal transfer acknowledge signal, TA, to allow data on LAD[0:31] to be either sampled (for reads) or changed (on writes). The data sampling/data change always occurs at the end of the bus cycle in which the LBC asserts TA internally. In LBC debug mode, TA is also visible externally on the MDVAL pin. GPCM and SDRAM controllers automatically generate TA according to the timing parameters programmed for them in option and mode registers; a UPM generates TA only when a UPM pattern has the UTA RAM word bit set. Figure 13-22 shows LALE, TA (internal), and LCSn. Note that TA and LALE are never asserted together, and that for the duration of LALE, LCSn (or any other control signal) remains negated or frozen. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 13-36 Freescale Semiconductor Functional Description LCLK LAD LALE Address Data LCSn TA Figure 13-22. Basic LBC Bus Cycle with LALE, TA, and LCSn 13.4.1.4 Data Buffer Control (LBCTL) The memory controller provides a data buffer control signal for the local bus (LBCTL). This signal is activated when a GPCM or UPM controlled bank is accessed. LBCTL can be disabled by setting ORn[BCTLD]. Access to an SDRAM machine controlled bank does not activate the LBCTL control. LBCTL can be further configured by LBCR[BCTLC] to act as an extra LWE or an extra LOE signal when in GPCM mode. If LBCTL is configured as a data buffer control (LBCR[BCTLC] = 00), the signal is asserted (high) on the rising edge of the bus clock on the first cycle of the memory controller operation, coincident with LALE. If the access is a write, LBCTL remains high for the whole duration. However, if the access is a read, LBCTL is negated (low) with the negation of LALE so that the memory device is able to drive the bus. If back-to-back read accesses are pending, LBCTL is asserted (high) one bus clock cycle before the next transaction starts (that is, one bus clock cycle before LALE) to allow a whole bus cycle for the bus to turn around before the next address is driven. If an external bus transceiver is used, LBCTL should be used to signify the write direction when high. Note that the default (reset and bus idle) value of LBCTL is also high. 13.4.1.5 Atomic Operation The LBC supports the following kinds of atomic bus operations (set by BRn[ATOM]): • Read-after-write atomic (RAWA). When a write access hits a memory bank in which ATOM = 01, the LBC reserves the selected memory bank for the exclusive use of the accessing master. While the bank is reserved, no other device can be granted access to this bank. The reservation is released when the master that created it accesses the same bank with a read transaction. Additional write transactions prior to the releasing read do not change reservation status, but are otherwise processed normally. If the master fails to release the MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 13-37 Local Bus Controller • reservation within 256 bus clock cycles, the reservation is released and an atomic error is reported (if enabled). This feature is intended for CAM operations. Write-after-read atomic (WARA). When a read access hit a memory bank in which ATOM = 10, the LBC reserves the bus for the exclusive use of the accessing master. During the reservation period, no other device can be granted access to the atomic bank. The reservation is released when the device that created it accesses the same bank with a write transaction. Additional read transactions prior to the releasing write are otherwise processed normally and do not change the reservation status. If the device fails to release the reservation within 256 bus clock cycles, the reservation is released and an atomic error is reported (if enabled). 13.4.1.6 Parity Generation and Checking (LDP) Parity can be configured for any bank by programming BRn[DECC]. Parity is generated and checked on a per-byte basis using LDP[0:3] for the bank if BRn[DECC] = 01 (normal parity) or BRn[DECC] = 10 for read-modify-write (RMW) parity. Byte lane parity on LDP[0:3] is generated regardless of the BRn[DECC] setting. Note that RMW parity can be used only for 32-bit port size banks. LBCR[EPAR] determines the global type of parity (odd or even). 13.4.1.7 Bus Monitor A bus monitor is provided to ensure that each bus cycle is terminated within a reasonable (user defined) period. When a transaction starts, the bus monitor starts counting down from the time-out value (LBCR[BMT]) until a data beat is acknowledged on the bus. It then reloads the time-out value and resumes the countdown until the data tenure completes and then idles if there is no pending transaction. Setting LTEDR[BMD] disables bus monitor error reporting through LTESR[BM]; however, the bus monitor is still active and can generate a UPM exception or terminate a GPCM access. It is very important to ensure that the value of LBCR[BMT] is not set too low; otherwise spurious bus time-outs may occur during normal operation—particularly for SDRAMs— resulting in incomplete data transfers. Accordingly, apart from the reset value of 0x00 (corresponding with the maximum time-out of 2048 bus cycles), LBCR[BMT] must not be set below 0x05 (or 40 bus cycles for time-out) under any circumstances. 13.4.2 General-Purpose Chip-Select Machine (GPCM) The GPCM allows a minimal glue logic and flexible interface to SRAM, EPROM, FEPROM, ROM devices, and external peripherals. The GPCM contains two basic configuration register groups—BRn and ORn. Figure 13-23 shows a simple connection between an 8-bit port size SRAM device and the LBC in GPCM mode. Byte-write enable signals (LWE) are available for each byte written to memory. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 13-38 Freescale Semiconductor Functional Description Also, the output enable signal (LOE) is provided to minimize external glue logic. On system reset, a global (boot) chip-select is available that provides a boot ROM chip-select (LCS0) prior to the system being fully configured. LCSn LOE LWE[0:3] LA[30:31] LA[27:29] LBC in GPCM Mode LALE Memory/Peripheral A[19:5] Latch LAD[0:7] Data[7:0] CE OE W[3:0] A[0:1] A[4:2] LAD[12:26] Figure 13-23. Local Bus to GPCM Device Interface Figure 13-24 shows LCS as defined by the setup time required between the address lines and CE. The user can configure ORn[ACS] to specify LCS to meet this requirement. LCLK LAD LALE A[19:5] ACS = 10 ACS = 11 Address Read Data Latched Address TA LCSn LOE Figure 13-24. GPCM Basic Read Timing (XACS = 0, ACS = 1x, TRLX = 0, CLKDIV = 4,8) 13.4.2.1 Timing Configuration If BRn[MSEL] selects the GPCM, the attributes for the memory cycle are taken from ORn. These attributes include the CSNT, ACS, XACS, SCY, TRLX, EHTR, and SETA fields. Table 13-23 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 13-39 Local Bus Controller shows signal behavior and system response for a write access with LCRR[CLKDIV] = 4 or LCRR[CLKDIV] = 8. Table 13-24 shows the signal behavior and system response for a read access with LCRR[CLKDIV] = 4 or LCRR[CLKDIV] = 8. Table 13-25 and Table 13-26 show the write and read signal behavior, respectively, when LCRR[CLKDIV] = 2. Table 13-23. GPCM Write Control Signal Timing for LCRR[CLKDIV] = 4 or 8 Option Register Attributes TRLX 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 Signal Behavior (Bus Clock Cycles) LWE Negated to Address to LCSn Negated to Total Cycles1 LCSn Asserted Address Change Address/Data Invalid 0 1/4 1/2 0 1 2 0 1/4 1/2 0 1 2 0 1+1/4 1+1/2 0 2 3 0 1+1/4 1+1/2 0 2 3 0 0 0 0 0 0 0 –1/4 –1/4 0 –1/4 –1/4 0 0 0 0 0 0 0 –1–1/4 –1–1/4 0 –1–1/4 –1–1/4 0 0 0 0 0 0 –1/4 –1/4 –1/4 –1/4 –1/4 –1/4 0 0 0 0 0 0 –1–1/4 –1–1/4 –1–1/4 –1–1/4 –1–1/4 –1–1/4 3+SCY 3+SCY 3+SCY 3+SCY 3+SCY 4+SCY 3+SCY 3+SCY 3+SCY 3+SCY 3+SCY 4+SCY 3+2*SCY 4+2*SCY 4+2*SCY 3+2*SCY 4+2*SCY 5+2*SCY 4+2*SCY 5+2*SCY 5+2*SCY 4+2*SCY 5+2*SCY 6+2*SCY XACS 0 0 0 1 1 1 0 0 0 1 1 1 0 0 0 1 1 1 0 0 0 1 1 1 ACS 00 10 11 00 10 11 00 10 11 00 10 11 00 10 11 00 10 11 00 10 11 00 10 11 CSNT 0 0 0 0 0 0 1 1 1 1 1 1 0 0 0 0 0 0 1 1 1 1 1 1 Total cycles when LALE is asserted for one cycle only (ORn[EAD] = 0; ORn[EAD] = 1 and LCRR[EADC] = 01). Asserting LALE for more than one cycle increases the total cycle count accordingly. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 13-40 Freescale Semiconductor Functional Description Table 13-24. GPCM Read Control Signal Timing for LCRR[CLKDIV] = 4 or 8 Option Register Attributes TRLX 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 Signal Behavior (bus clock cycles) Address to LCSn Asserted LCSn Negated to Address Change Total Cycles1 0 1/4 1/2 0 1 2 0 1/4 1/2 0 1 2 0 1+1/4 1+1/2 0 2 3 0 1+1/4 1+1/2 0 2 3 1 1 1 1 1 1 2 2 2 2 2 2 5 5 5 5 5 5 9 9 9 9 9 9 4+SCY 4+SCY 4+SCY 4+SCY 4+SCY 5+SCY 5+SCY 5+SCY 5+SCY 5+SCY 5+SCY 6+SCY 8+2*SCY 9+2*SCY 9+2*SCY 8+2*SCY 9+2*SCY 10+2*SCY 12+2*SCY 13+2*SCY 13+2*SCY 12+2*SCY 13+2*SCY 14+2*SCY EHTR 0 0 0 0 0 0 1 1 1 1 1 1 0 0 0 0 0 0 1 1 1 1 1 1 XACS 0 0 0 1 1 1 0 0 0 1 1 1 0 0 0 1 1 1 0 0 0 1 1 1 ACS 00 10 11 00 10 11 00 10 11 00 10 11 00 10 11 00 10 11 00 10 11 00 10 11 Total cycles when LALE is asserted for one cycle only (ORn[EAD] = 0; ORn[EAD] = 1 and LCRR[EADC] = 01). Asserting LALE for more than one cycle increases the total cycle count accordingly. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 13-41 Local Bus Controller Table 13-25. GPCM Write Control Signal Timing for LCRR[CLKDIV] = 2 Option Register Attributes TRLX XACS ACS CSNT 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 Signal Behavior (bus clock cycles) LWE Negated to Address to LCSn Negated to Total Cycles1 LCSn Asserted Address Change Address/Data Invalid 0 1/2 1/2 0 1 2 0 1/2 1/2 0 1 2 0 1+1/2 1+1/2 0 2 3 0 1+1/2 1+1/2 0 2 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 –1 –1 0 –1 –1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 –1 –1 –1 –1 –1 –1 3+SCY 3+SCY 3+SCY 3+SCY 3+SCY 4+SCY 3+SCY 3+SCY 3+SCY 3+SCY 3+SCY 4+SCY 3+2*SCY 4+2*SCY 4+2*SCY 3+2*SCY 4+2*SCY 5+2*SCY 4+2*SCY 5+2*SCY 5+2*SCY 4+2*SCY 5+2*SCY 6+2*SCY 0 0 0 1 1 1 0 0 0 1 1 1 0 0 0 1 1 1 0 0 0 1 1 1 00 10 11 00 10 11 00 10 11 00 10 11 00 10 11 00 10 11 00 10 11 00 10 11 0 0 0 0 0 0 1 1 1 1 1 1 0 0 0 0 0 0 1 1 1 1 1 1 Total cycles when LALE is asserted for one cycle only (ORn[EAD]=0; ORn[EAD]=1 and LCRR[EADC]=01). Asserting LALE for more than one cycle increases the total cycle count accordingly. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 13-42 Freescale Semiconductor Functional Description Table 13-26. GPCM Read Control Signal Timing for LCRR[CLKDIV] = 2 Option Register Attributes Signal Behavior (Bus Clock cycles) TRLX EHTR XACS ACS Address to LCSn Asserted LCSn Negated to Address Change Total Cycles1 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 1 1 1 1 1 1 0 0 0 0 0 0 1 1 1 1 1 1 0 0 0 1 1 1 0 0 0 1 1 1 0 0 0 1 1 1 0 0 0 1 1 1 00 10 11 00 10 11 00 10 11 00 10 11 00 10 11 00 10 11 00 10 11 00 10 11 0 1/2 1/2 0 1 2 0 1/2 1/2 0 1 2 0 1+1/2 1+1/2 0 2 3 0 1+1/2 1+1/2 0 2 3 1 1 1 1 1 1 2 2 2 2 2 2 5 5 5 5 5 5 9 9 9 9 9 9 4+SCY 4+SCY 4+SCY 4+SCY 4+SCY 5+SCY 5+SCY 5+SCY 5+SCY 5+SCY 5+SCY 6+SCY 8+2*SCY 9+2*SCY 9+2*SCY 8+2*SCY 9+2*SCY 10+2*SCY 12+2*SCY 13+2*SCY 13+2*SCY 12+2*SCY 13+2*SCY 14+2*SCY Total cycles when LALE is asserted for 1 cycle only (ORn[EAD]=0; ORn[EAD]=1 and LCRR[EADC]=01). Asserting LALE for more than 1 cycle increases the total cycle count accordingly. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 13-43 Local Bus Controller 13.4.2.2 Chip-Select Assertion Timing The banks selected to work with the GPCM, support an option to drive the LCSn signal with different timings (with respect to the external address/data bus). LCSn can be driven in any of the following ways: • Simultaneous with the latched memory address. (This refers to the externally latched address and not the address timing on LAD[0:31]. That is, chip select does not assert during LALE). One quarter of a clock cycle later (for LCRR[CLKDIV] = 4 or 8). One half of a clock cycle later (for LCRR[CLKDIV] = 2, 4, or 8). One clock cycle later (for LCRR[CLKDIV] = 4), when ORn[XACS] = 1. Two clock cycles later (for LCRR[CLKDIV] = 2, 4, or 8), when ORn[XACS] = 1. Three clock cycles later (for LCRR[CLKDIV] = 2, 4, or 8), when ORn[XACS] = 1 and ORn[TRLX] = 1. • • • • • The timing diagram in Figure 13-24 shows two chip-select assertion timings for the case LCRR[CLKDIV] = 4 or 8. If LCRR[CLKDIV] = 2, LCSn asserts identically for ORn[ACS] = 10 or 11. 13.4.2.2.1 Programmable Wait State Configuration The GPCM supports internal generation of transfer acknowledge. It allows between zero and 30 wait states to be added to an access by programming ORn[SCY] and ORn[TRLX]. Internal generation of transfer acknowledge is enabled if ORn[SETA] = 0. If LGTA is asserted externally two bus clock cycles or more before the wait state counter has expired (to allow for synchronization latency), the current memory cycle is terminated by LGTA; otherwise it is terminated by the expiration of the wait state counter. Regardless of the setting of ORn[SETA], wait states prolong the assertion duration of both LOE and LWEn in the same manner. When TRLX = 1, the number of wait states inserted by the memory controller is doubled from ORn[SCY] cycles to 2 × ORn[SCY] cycles, allowing a maximum of 30 wait states. 13.4.2.2.2 Chip-Select and Write Enable Negation Timing Figure 13-23 shows a basic connection between the local bus and a static memory device. In this case, LCSn is connected directly to CE of the memory device. The LWE[0:3] signals are connected to the respective WE[3:0] signals on the memory device where each LWE[0:3] signal corresponds to a different data byte. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 13-44 Freescale Semiconductor Functional Description LCLK LAD LALE A Latched Address Address Write Data TA LCSn CSNT = 1 LWEn LOE SCY = 1 Figure 13-25. GPCM Basic Write Timing (XACS = 0, ACS = 00, CSNT = 1, SCY = 1, TRLX = 0, CLKDIV = 4 or 8) As Figure 13-25 shows, the timing for LCSn is the same as for the latched address. The strobes for the transaction are supplied by LOE or LWEn, depending on the transaction direction—read or write (write case shown in Figure 13-25). ORn[CSNT] controls the timing for the appropriate strobe negation in write cycles. When this attribute is asserted, the strobe is negated one quarter of a clock before the normal case provided that LCRR[CLKDIV] = 4 or 8. For example, when ACS = 00 and CSNT = 1, LWEn is negated one quarter of a clock earlier, as shown in Figure 13-25. If LCRR[CLKDIV] = 2, LWEn is negated either coincident with LCSn or one cycle earlier. 13.4.2.2.3 Relaxed Timing ORx[TRLX] is provided for memory systems that require more relaxed timing between signals. Setting TRLX = 1 has the following effect on timing: • • • • • An additional bus cycle is added between the address and control signals (but only if ACS ≠ 00). The number of wait states specified by SCY is doubled, providing up to 30 wait states. The extended hold time on read accesses (EHTR) is extended further. LCSn signals are negated 1 cycle earlier during writes (but only if ACS ≠ 00). LWE[0:3] signals are negated 1 cycle earlier during writes. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 13-45 Local Bus Controller LCLK LAD LALE A Latched Address ACS = 10 ACS = 11 LCSn LBCTL LWEn SCY = 1, TRLX = 1 LOE Address Read Data TA Figure 13-26. GPCM Relaxed Timing Read (XACS = 0, ACS = 1x, SCY = 1, CSNT = 0, TRLX = 1, CLKDIV = 4 or 8) LCLK LAD LALE A Latched Address 1 ACS = 10 ACS = 11 LCSn LBCTL LWEn LOE Latched Address 2 Address 1 Write Data 1 Address 2 Write Data 2 TA Figure 13-27. GPCM Relaxed Timing Back-to-Back Writes (XACS = 0, ACS = 1x, SCY = 0, CSNT = 0, TRLX = 1, CLKDIV = 4 or 8) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 13-46 Freescale Semiconductor Functional Description Figure 13-26 and Figure 13-27 show relaxed timing read and write transactions. The effect of CLKDIV = 2 for these examples is only to delay the assertion of LCSn in the ACS = 10 case to the ACS = 11 case. The example in Figure 13-27 also shows address and data multiplexing on LAD[0:31] for a pair of writes issued consecutively. When TRLX and CSNT are set in a write access, the LWE[0:3] strobe signals are negated one clock earlier than in the normal case, as shown in Figure 12-28 and Figure 12-29. If ACS ≠ 00, LCSn is also negated one clock earlier. LCLK LAD LALE A Latched Address ACS = 10 CSNT = 1 Address Write Data TA LCSn LBCTL LWEn LOE Figure 13-28. GPCM Relaxed Timing Write (XACS = 0, ACS = 10, SCY = 0, CSNT = 1, TRLX = 1, CLKDIV = 4 or 8) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 13-47 Local Bus Controller LCLK LAD LALE A Latched Address Address Write Data TA LCSn LBCTL SCY = 1, TRLX = 1 LWEn LOE CSNT = 1 Figure 13-29. GPCM Relaxed Timing Write (XACS = 0, ACS = 00, SCY = 1, CSNT = 1, TRLX = 1, CLKDIV = 4 or 8) 13.4.2.2.4 Output Enable (LOE) Timing The timing of the LOE is affected only by TRLX. It always asserts and negates on the rising edge of the bus clock. LOE asserts either on the rising edge of the bus clock after LCSn is asserted or coinciding with LCSn (if XACS = 1 and ACS = 10 or 11). Accordingly, assertion of LOE can be delayed (along with the assertion of LCSn) by programming TRLX = 1. LOE negates on the rising clock edge coinciding with LCSn negation 13.4.2.2.5 Extended Hold Time on Read Accesses Slow memory devices that take a long time to disable their data bus drivers on read accesses should choose some combination of ORn[TRLX,EHTR]. Any access following a read access to the slower memory bank is delayed by the number of clock cycles specified in Table 13-6 in addition to any existing bus turnaround cycle. The final bus turnaround cycle is automatically inserted by the LBC for reads, regardless of the setting of ORn[EHTR]. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 13-48 Freescale Semiconductor Functional Description LCLK LAD LALE A Latched Address 1 Address 1 Read Data 1 Bus turnaround Address 2 Data 2 Latched Address 2 TA LCSn LCSy LBCTL LOE Figure 13-30. GPCM Read Followed by Read (TRLX = 0, EHTR = 0, Fastest Timing) LCLK LAD Rd. Address LALE A Latched Read Address Read Data Extended hold Bus turnaround Wr. Address Wr. Data Wr. Address TA LCSn LCSy LBCTL LOE Figure 13-31. GPCM Read Followed by Write (TRLX = 0, EHTR = 1, 1-Cycle Extended Hold Time on Reads) 13.4.2.3 External Access Termination (LGTA) External access termination is supported by the GPCM using the asynchronous LGTA input signal, which is synchronized and sampled internally by the local bus. If, during assertion of LCSn, MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 13-49 Local Bus Controller the sampled LGTA signal is asserted, it is converted to an internal generation of transfer acknowledge, which terminates the current GPCM access (regardless of the setting of ORn[SETA]). LGTA should be asserted for at least one bus cycle to be effective. Note that because LGTA is synchronized, bus termination occurs two cycles after LGTA assertion, so in case of read cycle, the device still must drive data as long as LOE is asserted. The user selects whether transfer acknowledge is generated internally or externally (LGTA) by programming ORn[SETA]. Asserting LGTA always terminates an access, even if ORn[SETA] = 0 (internal transfer acknowledge generation), but it is the only means by which an access can be terminated if ORn[SETA] = 1. The timing of LGTA is illustrated by the example in Figure 13-32. LCLK LAD LALE A TA LGTA LCSn LBCTL LOE Latched Address Address Read Data Figure 13-32. External Termination of GPCM Access 13.4.2.4 Boot Chip-Select Operation Boot chip-select operation allows address decoding for a boot ROM before system initialization. LCS0 is the boot chip-select output; its operation differs from other external chip-select outputs after a system reset. When the core begins accessing memory after system reset, LCS0 is asserted for every local bus access until BR0 or OR0 is reconfigured. The boot chip-select also provides a programmable port size, which is configured during reset. The boot chip-select does not provide write protection. LCS0 operates this way until the first write to OR0 and it can be used as any other chip-select register after the preferred address range is loaded into BR0. After the first write to OR0, the boot chip-select can be restarted only with a hardware reset. Table 13-27 describes the initial values of the boot bank in the memory controller. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 13-50 Freescale Semiconductor Functional Description Table 13-27. Boot Bank Field Values after Reset Register BR0 Field BA XBA PS DECC WP MSEL ATOM V AM XAM BCTLD CSNT ACS XACS SCY SETA TRLX EHTR EAD Setting 0000_0000_0000_0000_0 00 From pin during reset. 00 0 000 00 1 0000_0000_0000_0000_0 00 0 1 11 1 1111 0 1 1 1 OR0 13.4.3 SDRAM Machine The LBC provides an SDRAM interface (machine) for the local bus. The machine provides the control functions and signals for Intel PC133 and JEDEC-compliant SDRAM devices. Each bank can control an SDRAM device on the local bus. 13.4.3.1 Supported SDRAM Configurations The memory controller supports any SDRAM configuration with the restrictions that all SDRAM devices that reside on the bus should have the same port size and timing parameters (as defined in LSDMR). Figure 13-33 shows an example connection between the LBC and a 32-bit SDRAM device with 12 address lines. Note that address signals A[4:0] of the SDRAM connect directly to LA[27:31], address pin A10 connects to the LBCs dedicated LSDA10 signal, while the remaining address bits (except A10) are latched from LAD[20:26]. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 13-51 Local Bus Controller LSDWE LSDRAS LSDCAS LCS1 LSDDQM[0:3] LSDA10 LA[27:29] LALE Latch Memory Address WE RAS CAS CS DQM[3:0] A10 A[2:0] A[11,9:3] Local Bus Controller LAD[18,20:26] Memory Data LAD[0:31] LCLK LCKE SDRAM 32-Bit Port Size DQ[31:0] CLK CKE Figure 13-33. Connection to a 32-Bit SDRAM with 12 Address Lines 13.4.3.2 SDRAM Power-On Initialization Following a system reset, initialization software must set up the programmable parameters in the memory controller banks registers (ORn, BRn, LSDMR). After all memory parameters are configured, system software should execute the following initialization sequence for each SDRAM device. • • • Issue a PRECHARGE-ALL-BANKS command Issue eight AUTO-REFRESH commands Issue a MODE-SET command to initialize the mode register The initial commands are executed by setting LSDMR[OP] and accessing the SDRAM with any write that hits the relevant bank. Note that software should ensure that no memory operations begin until this process completes. 13.4.3.3 Intel PC133 and JEDEC-Standard SDRAM Interface Commands The SDRAM machine performs all accesses to SDRAM by using Intel PC133 and JEDEC-standard SDRAM interface commands. The SDRAM device samples the command and MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 13-52 Freescale Semiconductor Functional Description data inputs on the rising edge of the bus clock. Data at the output of the SDRAM device is sampled on the rising edge of the bus clock. The following SDRAM interface commands are provided by setting LSDMR[OP] to a non-zero value (LSDMR[OP] = 000 sets normal read/write operation): Table 13-28. SDRAM Interface Commands Command (LSDMR[OP]) ACTIVATE (110) MODE-SET (011) Description Latches the row address and initiates a memory read of that row. Row data is latched in SDRAM sense amplifiers and must be restored with a PRECHARGE command before another ACTIVATE is issued. Allows setting of SDRAM options—CAS latency and burst length. CAS latency depends on the SDRAM device used. Although some SDRAMs provide burst lengths of 1, 2, 4, 8, or a page, the local bus memory controller supports only 8-beat bursts for 8-bit and 32-bit port size, or 4-beat bursts for 16-bit port size. The LBC does not support burst lengths of 1, 2 and a page for SDRAMs. The mode register data (CAS latency and burst length) is programmed into the LSDMR register by initialization software after reset. After the LSDMR is set, the LBC transfers the information to the SDRAM device by issuing a MODE-SET command. PRECHARGE Restores data from the sense amplifiers to the appropriate row in the SDRAM device array. Also (100: single bank) initializes the sense amplifiers to prepare for activating another row in the SDRAM device. Note that (101: all-banks) the LBC uses LSDA10 to distinguish between PRECHARGE-ALL-BANKS (LSDA10 is high) and PRECHARGE-SINGLE-BANK (LSDA10 is low). The SDRAMs must be compatible with this format. READ (111) Latches the column address and transfers data from the selected sense amplifier on the SDRAM device, to the output buffer as determined by the column address. During each successive clock, additional data is driven without additional read commands. At the end of the burst, the page remains open. Burst length is the one set for this bank. Read data is discarded by the LBC. Latches the column address and transfers data from the data signals to the selected sense amplifier on the SDRAM device, as determined by the column address. During each successive clock, additional data is transferred to the sense amplifiers from the data signals without additional write commands. At the end of the burst, the page remains open. Burst length is the one set for this bank. LSDDQM[0:3] are inactive and write data is undefined. WRITE (111) AUTO-REFRESH Causes a row to be read in all memory banks (JEDEC SDRAM) as determined by the refresh row (001) address counter (similar to CBR). The refresh row address counter is internal to the SDRAM device. After being read, a row is automatically rewritten into the memory array. All banks must be in a precharged state before executing refresh. SELF-REFRESH Allows data to be retained in the SDRAM device, even when the rest of the LBC is in a power saving (010) mode with clocks turned off. When placed in this mode, the SDRAM device is capable of issuing its own refresh commands, without external clocking from the LBC and the LCKE pin from the LBC is negated. This command can be issued at any time. Normal operation can be resumed only by setting LSDMR[OP] = 000, and waiting a minimum of 200 bus cycles before issuing reads or writes to the LBC. 13.4.3.4 Page Hit Checking The SDRAM machine supports page-mode operation. Each time a page is activated on the SDRAM device, the SDRAM machine stores its address in a page register. The page information, which the user writes to the ORn register, is used along with the bank size to compare page bits of MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 13-53 Local Bus Controller the address to the page register each time a bus-cycle access is requested. If a match is found, together with a bank match, the bus cycle is defined as a page hit. An open page is automatically closed by the SDRAM machine if the bus becomes idle, unless ORn[PMSEL] = 1. 13.4.3.5 Page Management The LBC can manage at most four open pages (one page per SDRAM bank) for a single SDRAM device. After a page is opened, it remains open unless: • • • • The next access is to a page in a different SDRAM device, in which case all open pages on the current device are closed with a PRECHARGE-ALL-BANKS command. The next access is to a page in an SDRAM bank that has a different page open on it, in which case the old page is closed with a PRECHARGE-SINGLE-BANK command. The current SDRAM device requires refresh services, in which case all open pages on the current device are closed with a PRECHARGE-ALL-BANKS command. The bus becomes idle and ORn[PMSEL] = 0, in which case all open pages in the current device are closed with a PRECHARGE-ALL-BANKS command. 13.4.3.6 SDRAM Address Multiplexing The lower address bus bits are connected to the memory device’s address port with the memory controller multiplexing the row/column and the internal bank select lines. The position of the bank select lines are set according to LSDMR[BSMA]. Figure 13-34 shows how the SDRAM controller shifts the row address down to the lower output address signals during activate and shifts the bank select bits up to the address pins specified by LSDMR[BSMA], supporting page-based interleaving. The lsb of the logical row address (An in Figure 13-34) is aligned with the connected lsb of LAD (bits 29, 30, and 31 for port sizes of 32, 16, and 8 bits, respectively). 0 Logical Address: msbs ROW An BS COLUMN 31 lsbs Activate Address (RAS): BS ROW lsbs R/W Address (CAS): ROW BS — COLUMN lsbs To memory device pins, except A10 Figure 13-34. SDRAM Address Multiplexing MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 13-54 Freescale Semiconductor Functional Description Note that during normal operation (read/write), a full 32-bit address that includes row and column is generated on LAD[0:31]. However, address/data signal multiplexing implies that the address must be latched by an external latch that is controlled by LALE. All SDRAM device address signals need to be connected to the latched address bits and burst address bits (LA[27:31]) of the LBC, with the exception of A10, which has a dedicated connection on LSDA10. LSDA10 is driven with the appropriate row address bit for SDRAM commands that require A10 to be an address. 13.4.3.7 SDRAM Device-Specific Parameters The software is responsible for setting correct values for device-specific parameters that can be extracted from the device’s data sheet. The values are stored in the ORn and LSDMR registers. These parameters include the following: • • • • • • Precharge to activate interval (LSDMR[PRETOACT]) Activate to read/write interval (LSDMR[ACTTORW]) CAS latency, column address to first data out (LSDMR[CL] and LCRR[ECL]) Write recovery, last data in to precharge (LSDMR[WRC]) Refresh recovery interval (LSDMR[RFRC]) External buffers on the control lines present (LSDMR[BUFCMD] and LCRR[BUFCMDC]) In addition, the LBC hardware ensures a default activate to precharge interval of 10 bus cycles. The following sections describe SDRAM parameters programmed in LSDMR. 13.4.3.7.1 Precharge-to-Activate Interval The precharge-to-activate interval parameter, controlled by LSDMR[PRETOACT]. defines the earliest timing for an ACTIVATE or REFRESH command after a PRECHARGE command to the same SDRAM bank. LCLK LALE LCSn LSDRAS LSDCAS LSDWE LSDDQM[0:3] LAD[0:31] 1111 ZZZZZZZZ RAS ADD XXXXX PRETOACT = 2 CAS ADD 0000 D0 1111 ZZZZZZZZ PRECHARGE Command Bank A ACTIVATE Command Bank A Figure 13-35. PRETOACT = 2 (2 Clock Cycles) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 13-55 Local Bus Controller 13.4.3.7.2 Activate-to-Read/Write Interval This parameter, controlled by LSDMR[ACTTORW], defines the earliest timing for a READ/WRITE command after an ACTIVATE command to the same SDRAM bank. LCLK LALE LCSn LSDRAS LSDCAS LSDWE LSDDQM[0:3] LAD[0:31] ZZZZZZZZ 1111 RAS ADD XXXX 0000 CAS ADD D0 D1 D2 D3 1111 ZZZZZZZZ ACTTORW = 2 ACTIVATE Command WRITE Command Figure 13-36. ACTTORW = 2 (2 Clock Cycles) 13.4.3.7.3 Column Address to First Data Out—CAS Latency This parameter, controlled by LSDMR[CL] for latency of 1, 2, or 3 and by LCRR[ECL] for latency of more than 3, defines the timing for first read data after a column address is sampled by the SDRAM. LCLK LALE LCSn LSDRAS LSDCAS LSDWE LSDDQM[0:3] LAD[0:31] 1111 ZZZZZZZZ RAS ADD XXXX CAS_ADD ZZZZZZZZ 0000 D0 D1 D2 1111 D3 CL = 2 READ Command First Data Out Figure 13-37. CL = 2 (2 Clock Cycles) 13.4.3.7.4 Last Data In to Precharge—Write Recovery This parameter, controlled by LSDMR[WRC], defines the earliest timing for a PRECHARGE command after the last data was written to the SDRAM. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 13-56 Freescale Semiconductor Functional Description LCLK LALE LCSn LSDRAS LSDCAS LSDWE LSDDQM[0:3] LAD[0:31] 1111 ZZZZZZZZ 0000 CAS ADD D0 D1 D2 D3 RAS ADD WRC = 2 X 1111 CAS ADD D0 0000 D1 WRITE Command Last Data In PRECHARGE Command Figure 13-38. WRC = 2 (2 Clock Cycles) 13.4.3.7.5 Refresh Recovery Interval (RFRC) This parameter, controlled by LSDMR[RFRC], defines the earliest timing for an ACTIVATE or REFRESH command after a REFRESH command to the same SDRAM device. LCLK LALE LCSn LSDRAS LSDCAS LSDWE LSDDQM[0:3] LAD[0:31] 1111 ZZZZZZZZ PRETOACT = 3 RAS ADD RFRC = 4 (6 clocks) XXXX CAS ADD PRECHARGE ALL Command (if needed) AUTO REFRESH Command ACTIVATE Command Figure 13-39. RFRC = 4 (6 Clock Cycles) 13.4.3.7.6 External Address and Command Buffers (BUFCMD) If the additional delay of any buffers placed on the command strobes (LSDRAS, LSDCAS, LSDWE and LSDA10), is endangering the device setup time, LSDMR[BUFCMD] should be set. Setting this bit causes the memory controller to add LCRR[BUFCMDC] extra bus cycles to the assertion of SDRAM control signals (LSDRAS, LSDCAS, LSDWE and LSDA10) for each SDRAM command. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 13-57 Local Bus Controller LCLK LALE LCSn LSDRAS LSDCAS LSDWE LSDDQM LAD ZZZZZZ RAS ADD 1111 XXXXXXXX CAS ADD XXXX D0 D1 0000 D2 1111 D3 ZZZZZZZZ Command setup cycle Command setup cycle Figure 13-40. BUFCMD = 1, LCRR[BUFCMDC] = 2 13.4.3.8 SDRAM Interface Timing The following figures show SDRAM timing for various types of accesses. LCLK LALE LCSn LSDRAS LSDCAS LSDWE LSDDQM[0:3] LAD[0:31] TA 1111 ZZZZZZZZ ROW ADD COL ADD ZZZZZZZZ 0000 ZZZZZZZZ 1111 D0 ZZZZZZZZ Figure 13-41. SDRAM Single-Beat Read, Page Closed, CL = 3 LCLK LALE LCSn LSDRAS LSDCAS LSDWE LSDDQM[0:3] LAD[0:31] TA 1111 ZZZZZZZZ COL ADD 0000 ZZZZZZZZ D0 1111 Figure 13-42. SDRAM Single-Beat Read, Page Hit, CL = 3 LCLK LALE LCSn LSDRAS LSDCAS LSDWE LSDDQM[0:3] LAD[0:31] TA 1111 ZZZZZZZZ COL ADD ROW ADD ZZZZZZZZ 0000 ZZZZZZZZ D0 D1 1111 ZZZZZZZZ Figure 13-43. SDRAM Two-Beat Burst Read, Page Closed, CL = 3 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 13-58 Freescale Semiconductor Functional Description Deactivate ctivate A LCLK LALE LCSn LSDRAS LSDCAS LSDWE LSDDQM[0:3] LAD[0:31] TA 1111 ZZZZZZZZ ROW ADD ZZZZZZZZCOL ADD 0000 ZZZZZZZZ D0 D1 D2 1111 D3 ZZZZZZZZ Figure 13-44. SDRAM Four-Beat Burst Read, Page Miss, CL = 3 LCLK LALE LCSn LSDRAS LSDCAS LSDWE LSDDQM[0:3] LAD TA 1111 ZZZZZZZZ COL ADD 0000 D0 1111 ZZZZZZZZ Figure 13-45. SDRAM Single-Beat Write, Page Hit. LCLK LALE LCSn LSDRAS LSDCAS LSDWE LSDDQM[0:3] LAD[0:31] TA 1111 ZZZZZZZZ RAS ADD XXXXXXXX COL ADD D0 0000 D1 D2 1111 ZZZZZZZZ Figure 13-46. SDRAM Three-Beat Write, Page Closed LCLK LALE LCSn LSDRAS LSDCAS LSDWE LSDDQM[0:3] LAD[0:31] TA Z 1111 COL ADD 1 0000 Z D0 1111 D1 Z COL ADD 2 0000 Z 1111 x D0 D1 Figure 13-47. SDRAM Read-after-Read Pipelined, Page Hit, CL = 3 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 13-59 Local Bus Controller LCLK LALE LCSn LSDRAS LSDCAS LSDWE LSDDQM[0:3] LAD[0:31] TA 1111 Z COL ADD 1 D0 D1 0000 D2 D3 1111 COL ADD 2 D0 D1 0000 D2 D3 Z Figure 13-48. SDRAM Write-after-Write Pipelined, Page Hit LCLK LALE LCSn LSDRAS LSDCAS LSDWE LSDDQM[0:3] LAD[0:31] TA Z 1111 COL ADD 1 D0 D1 0000 D2 D3 1111 COL ADD 2 0000 Z D0 D1 D2 1111 D3 Z Figure 13-49. SDRAM Read-after-Write Pipelined, Page Hit 13.4.3.9 SDRAM Read/Write Transactions The SDRAM interface supports read and write transactions of between 1 and 8 data beats for transaction sizes ranging from 1 to 32 bytes. A full burst is performed for each transaction, with the burst length dependent on the port size. A maximum burst of 8 beats is used for an 8-bit or 32-bit port size, while a maximum burst of 4 beats is used for a 16-bit port size, as programmed in LSDMR[BL]. For reads that require less than the full burst length, extraneous data in the burst is ignored and suppressed by the assertion of LSDDQM[0:3]. For writes that require less than the full burst length, the non-targeted addresses are protected by driving corresponding LSDDQM bits high (inactive) on the irrelevant cycles of the burst. However, system performance is not compromised because, if a new transaction is pending, the SDRAM controller begins executing it immediately, effectively terminating the burst early. 13.4.3.10 SDRAM MODE-SET Command Timing The LBC transfers mode register data (CAS latency and burst length) stored in the LSDMR register to the SDRAM device by issuing the MODE-SET command, as shown in Figure 13-50. In this case, the latched address carries the mode bits for the command. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 13-60 Freescale Semiconductor Functional Description LCLK LALE LCSn LSDRAS LSDCAS LSDWE LSDDQM[0:3] LAD[0:31] 1111 ZZZZZZZZ MODE ZZZZZZZZ MODE-SET Command Figure 13-50. SDRAM MODE-SET Command 13.4.3.11 SDRAM Refresh The memory controller supplies AUTO-REFRESH commands to any connected SDRAM device according to the interval specified in LSRT (and prescaled by MRTPR[PTP]). This represents the time period required between refreshes. The values of LSRT and MRTPR depend on the specific SDRAM devices used and the system clock frequency of the LBC. This value should allow for a potential collision between memory accesses and refresh cycles. The period of the refresh interval must be greater than the access time to ensure that read and write operations complete successfully. There are two levels of refresh request priority—low and high. The low priority request is generated as soon as the refresh timer expires; this request is granted only if no other requests to the memory controller are pending. If the request is not granted (memory controller is busy) and the refresh timer expires two more times, the request becomes high priority and is served when the current memory controller operation finishes. 13.4.3.11.1 SDRAM Refresh Timing The SDRAM memory controller implements bank staggering for the auto refresh function. This reduces instantaneous current consumption for memory refresh operations. After a refresh request is granted, the memory controller begins issuing an AUTO-REFRESH command to each device associated with the refresh timer. After a refresh command is issued to an SDRAM device, the memory controller waits for the number of bus clock cycles programmed in the SDRAM machine’s mode register (LSDMR[RFCR]) before issuing any subsequent ACTIVATE command to the same device. To avoid violating SDRAM device timing constraints, the user should ensure that the refresh request interval, defined by LSRT and MRTPR, is greater than the refresh recovery interval, defined by LSDMR[RFCR]. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 13-61 Local Bus Controller LSDMR[RFCR] Auto Auto Auto Auto Ref Ref Ref Ref LCLK LALE LCS1 LCS2 LCS3 LCS4 LSDRAS LSDCAS LSDWE LSDDQM[0:3] LAD[0:31] TA Activate 1111 Z ROW ADD X COL ADD 0000 D0 1111 Z Figure 13-51. SDRAM Bank-Staggered Auto-Refresh Timing 13.4.4 User-Programmable Machines (UPMs) UPMs are flexible interfaces that connect to a wide range of memory devices. At the heart of each UPM is an internal RAM array that specifies the logical value driven on the external memory control signals1 (LCSn, LBS[0:3] and LGPL[0:5]) for a given clock cycle. Each word in the RAM array provides bits that allow a memory access to be controlled with a resolution of up to one quarter of the external bus clock period on the byte-select and chip-select lines. Figure 13-52 shows the basic operation of each UPM. Internal / External Memory Access Request UPM Refresh Timer Request Run Command (issued in software) Exception Request Array Index Generator Index RAM Array LUPWAIT Wait Request Logic Hold Increment Index (LAST = 0) Internal Signals Latch Signals Timing Generator LGPLn LBSn LCSn WAEN Bit Internal Controls Figure 13-52. User-Programmable Machine Functional Block Diagram 1.If the LGPL4/LGTA /LUPWAIT/LPBSE signal is used as both an input and an output, a weak pullup is required. Refer to the hardware specification for details regarding termination options. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 13-62 Freescale Semiconductor Functional Description The following events initiate a UPM cycle: • • • Any internal device requests an external memory access to an address space mapped to a chip-select serviced by the UPM A UPM refresh timer expires and requests a transaction, such as a DRAM refresh A bus monitor time-out error during a normal UPM cycle redirects the UPM to execute an exception sequence The RAM array contains 64 words of 32-bits each. The signal timing generator loads the RAM word from the RAM array to drive the general-purpose lines, byte-selects, and chip-selects. If the UPM reads a RAM word with WAEN set, the external LUPWAIT signal is sampled and synchronized by the memory controller and the current request is frozen. 13.4.4.1 UPM Requests A special pattern location in the RAM array is associated with each of the possible UPM requests. An internal device’s request for a memory access initiates one of the following patterns (MxMR[OP] = 00): • • • • Read single-beat pattern (RSS) Read burst cycle pattern (RBS) Write single-beat pattern (WSS) Write burst cycle pattern (WBS) A UPM refresh timer request pattern initiates a refresh timer pattern (RTS). An exception (caused by a bus monitor time-out error) occurring while another UPM pattern is running initiates an exception condition pattern (EXS). Figure 13-53 and Table 13-29 show the start addresses of these patterns in the UPM RAM, according to cycle type. RUN commands (MxMR[OP] = 11), however, can initiate patterns starting at any of the 64 UPM RAM words. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 13-63 Local Bus Controller Array Index Generator Read Single-Beat Request Read Burst Request RSS RBS Write Single-Beat Request Write Burst Request WSS WBS RAM Array 64 RAM Words Refresh Timer Request Exception Condition Request RTS EXS Figure 13-53. RAM Array Indexing Table 13-29. UPM Routines Start Addresses UPM Routine Read single-beat (RSS) Read burst (RBS) Write single-beat (WSS) Write burst (WBS) Refresh timer (RTS) Exception condition (EXS) Routine Start Address 0x00 0x08 0x18 0x20 0x30 0x3C 13.4.4.1.1 Memory Access Requests The user must ensure that the UPM is appropriately initialized before a request occurs. The UPM supports two types of memory reads and writes: • A single-beat transfer transfers one operand consisting of up to a single word (dependent on port size). A single-beat cycle starts with one transfer start and ends with one transfer acknowledge. A burst transfer transfers exactly 4 double words regardless of port size. For 32-bit accesses, the burst cycle starts with one transfer start but ends after eight transfer acknowledges, whereas an 8-bit device requires 32 transfer acknowledges. • The user must ensure that patterns for single-beat transfers contain one, and only one, transfer acknowledge (UTA bit in RAM word set high) and for a burst transfer, contain the exact number of transfer acknowledges required. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 13-64 Freescale Semiconductor Functional Description Any transfers that do not naturally fit single or burst transfers are synthesized as a series of single transfers. These accesses are treated by the UPM as back-to-back, single-beat transfers. Burst transfers can also be inhibited by setting ORn[BI]. Burst performance can be achieved by ensuring that UPM transactions are 32-byte aligned with a transaction size being some multiple of 32-bytes, which is a natural fit for cache-line transfers, for example. 13.4.4.1.2 UPM Refresh Timer Requests Each UPM contains a refresh timer that can be programmed to generate refresh service requests of a particular pattern in the RAM array. Figure 13-54 shows the clock division hardware associated with memory refresh timer request generation. The UPM refresh timer register (LURT) defines the period for the timers associated with all three UPMs. System Clock PTP Prescaling Divide by LURT UPM refresh timer request Figure 13-54. Memory Refresh Timer Request Block Diagram By default, all local bus refreshes are performed using the refresh pattern of UPMA. This means that if refresh is required, MAMR[RFEN] must be set. It also means that only one refresh routine should be programmed and be placed in UPMA, which serves as the refresh executor. Any banks assigned to a UPM are provided with the refresh pattern if the RFEN bit of the corresponding UPM is set. UPMA assigned banks, therefore, always receive refresh services when MAMR[RFEN] is set, while UPMB and UPMC assigned banks also receive (the same) refresh services if the corresponding MxMR[RFEN] bits are set. Note that the UPM refresh timer request should not be used in a system with SDRAM refresh enabled. The system designer must choose to use either SDRAM refresh or UPM refresh. Using both may result in missing refresh periods to memory. 13.4.4.1.3 Software Requests—RUN Command Software can start a request to the UPM by issuing a RUN command to the UPM. Some memory devices have their own signal handshaking protocol to put them into special modes, such as self-refresh mode. Other memory devices require special commands to be issued on their control signals, such as for SDRAM initialization. For these special cycles, the user creates a special RAM pattern that can be stored in any unused areas in the UPM RAM. Then a RUN command is used to run the cycle. The UPM runs the pattern beginning at the specified RAM location until it encounters a RAM word with its LAST bit set. The RUN command is issued by setting MxMR[OP] = 11 and accessing UPMn memory region with any write transaction that hits the corresponding UPM machine. MxMR[MAD] determines the starting address in the RAM array for the pattern. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 13-65 Local Bus Controller Note that transfer acknowledges (UTA bit in the RAM word) are ignored for software (RUN command) requests, and hence the LAD signals remain high-impedance unless the normal initial LALE occurs or the RUN pattern causes assertion of LALE to occur on changes to the RAM word AMX field. 13.4.4.1.4 Exception Requests When the LBC under UPM control initiates an access to a memory device and an exception occurs (bus monitor time-out), the UPM provides a mechanism by which memory control signals can meet the device’s timing requirements without losing data. The mechanism is the exception pattern that defines how the UPM negates its signals in a controlled manner. 13.4.4.2 Programming the UPMs The UPM is a micro sequencer that requires microinstructions or RAM words to generate signal timings for different memory cycles. Follow these steps to program UPMs: 1. 2. 3. 4. Set up BRn and ORn registers. Write patterns into the RAM array. Program MRTPR, LURT and MAMR[RFEN] if refresh is required. Program MxMR. Patterns are written to the RAM array by setting MxMR[OP] = 01 and accessing the UPM with any write transaction that hits the relevant chip select. The entire array is thus programmed by an alternating series of writes: to MDR (RAM word to be written) each time followed by a (dummy) write transaction to the relevant UPM assigned bank. Note that the UPM memory region must be cache-inhibited or write-through (the MMU page must have the I or W bit set) during the time that the UPM array is being written. If the memory is to be cacheable and/or copyback, the MMU must be set accordingly after the UPM array is initialized. RAM array contents may also be read for debug purposes, for example, by alternating dummy read transactions, each time followed by reads of MDR (when MxMR[OP] = 10). 13.4.4.3 UPM Signal Timing RAM word fields specify the value of the various external signals at a granularity of up to four values for each bus clock cycle. The signal timing generator causes external signals to behave according to timing specified in the current RAM word. For LCRR[CLKDIV] = 4 or 8, each bit in the RAM word relating to LCSn and LBS timing specifies the value of the corresponding external signal at each quarter phase of the bus clock. If LCRR[CLKDIV] = 2, the external signal can change value only on each half phase of the bus clock. If the RAM word in this case (LCRR[CLKDIV] = 2) specifies a quarter phase signal change, the signal timing generator interprets this as a half cycle change. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 13-66 Freescale Semiconductor Functional Description The division of UPM bus cycles into phases is shown in Figure 13-55 and Figure 13-56. If LCRR[CLKDIV] = 2, the bus cycle comprises only two active phases, T1 and T3, which correspond with the first and second halves of the bus clock cycle, respectively. However, if LCRR[CLKDIV] = 4 or 8, four phases, T1–T4, define four quarters of the bus clock cycle. Because T2 and T4 are inactive when LCRR[CLKDIV] = 2, UPM ignores signal timing programmed for assertion in either of these phases in the case LCRR[CLKDIV] = 2. LCLK T1 T2 T3 T4 Figure 13-55. UPM Clock Scheme for LCRR[CLKDIV] = 2 LCLK T1 T2 T3 T4 Figure 13-56. UPM Clock Scheme for LCRR[CLKDIV] = 4 or 8 13.4.4.4 RAM Array The RAM array for each UPM is 64 locations deep and 32 bits wide, as shown in Figure 13-57. The signals at the bottom of the figure are UPM outputs. The selected LCSn is for the bank that matches the current address. The selected LBS is for the byte lanes read or written by the access. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 13-67 Local Bus Controller 32 Bits RAM Array 64 deep Clock Phases T1, T2, T3, T4 External Signals Timing Generator BRn[PS], LA[30,31] Current Bank CS Line Selector Byte Select Logic LCS[0:7] LGPL0 LGPL1 LGPL2 LGPL3 LGPL4 LGPL5 LBS[0:3] Figure 13-57. RAM Array and Signal Generation 13.4.4.4.1 RAM Words The RAM word is a 32-bit microinstruction stored in one of 64 locations in the RAM array. It specifies timing for external signals controlled by the UPM. Figure 13-58 shows the RAM word fields. When LCRR[CLKDIV] = 4 or 8, the CSTn and BSTn bits determine the state of UPM signals LCSn and LBS[0:3] at each quarter phase of the bus clock. When LCRR[CLKDIV] = 2, CST2 and CST4 are ignored and the external has the values defined by CST1 and CST3 but extended to half the clock cycle in duration. The same interpretation occurs for the BSTn bits when LCRR[CLKDIV] = 2. 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 R CST1 CST2 CST3 CST4 BST1 BST2 BST3 BST4 W Reset 16 17 18 19 20 21 G0L G0H G1T1 G1T3 G2T1 G2T3 0000_0000_0000_0000 22 23 24 25 26 27 28 29 30 31 R G3T1 G3T3 W Reset Offset G4T1/ G4T3/ G5T1 G5T3 DLT3 WAEN REDO LOOP EXEN AMX NA UTA TODT LAST 0000_0000_0000_0000 Figure 13-58. RAM Word Field Descriptions MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 13-68 Freescale Semiconductor Functional Description Table 13-30 describes RAM word fields. Table 13-30. RAM Word Field Descriptions Bits 0 Name CST1 Description Chip select timing 1. Defines the state (0 or 1) of LCSn during bus clock quarter phase 1 if LCRR[CLKDIV] = 4 or 8. Defines the state (0 or 1) of LCSn during bus clock half phase 1 if LCRR[CLKDIV] = 2. Chip select timing 2. Defines the state (0 or 1) of LCSn during bus clock quarter phase 2 if LCRR[CLKDIV] = 4 or 8. Ignored when LCRR[CLKDIV] = 2. Chip select timing 3. Defines the state (0 or 1) of LCSn during bus clock quarter phase 3 if LCRR[CLKDIV] = 4 or 8. Defines the state (0 or 1) of LCSn during bus clock half phase 2 if LCRR[CLKDIV] = 2. Chip select timing 4. Defines the state (0 or 1) of LCSn during bus clock quarter phase 4.if LCRR[CLKDIV] = 4 or 8. Ignored when LCRR[CLKDIV] = 2. Byte select timing 1. Defines the state (0 or 1) of LBS during bus clock quarter phase 1 (LCRR[CLKDIV] = 4 or 8) or bus clock half phase 1 (LCRR[CLKDIV] = 2), in conjunction with BRn[PS] and the state of LA[30:31]. Byte select timing 2:. Defines the state (0 or 1) of LBS during bus clock quarter phase 2 (LCRR[CLKDIV] = 4 or 8), in conjunction with BRn[PS] and the state of LA[30:31]. Ignored when LCRR[CLKDIV] = 2. Byte select timing 3. Defines the state (0 or 1) of LBS during bus clock quarter phase 3 (LCRR[CLKDIV] = 4 or 8) or bus clock half phase 2 (LCRR[CLKDIV] = 2), in conjunction with BRn[PS] and the state of LA[30:31]. Byte select timing 4. Defines the state (0 or 1) of LBS during bus clock quarter phase 4 (LCRR[CLKDIV] = 4 or 8), in conjunction with BRn[PS] and the state of LA[30:31]. Ignored when LCRR[CLKDIV] = 2. General-purpose line 0 lower. Defines the state of LGPL0 during the bus clock quarter phases 1 and 2 (first half phase). 00 Value defined by MxMR[G0CL] 01 Reserved 10 0 11 1 General-purpose line 0 higher. Defines the state of LGPL0 during the bus clock quarter phases 3 and 4 (second half phase). 00 Value defined by MxMR[G0CL] 01 Reserved 10 0 11 1 General-purpose line 1 timing 1. Defines the state (0 or 1) of LGPL1 during bus clock quarter phases 1 and 2 (first half phase). General-purpose line 1 timing 3. Defines the state (0 or 1) of LGPL1 during bus clock quarter phases 3 and 4 (second half phase) General-purpose line 2 timing 1. Defines state (0 or 1) of LGPL2 during bus clock quarter phases 1 and 2 (first half phase). General-purpose line 2 timing 3. Defines the state (0 or 1) of LGPL2 during bus clock quarter phases 3 and 4 (second half phase). 1 2 CST2 CST3 3 4 CST4 BST1 5 BST2 6 BST3 7 BST4 8–9 G0L 10–11 G0H 12 13 14 15 G1T1 G1T3 G2T1 G2T3 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 13-69 Local Bus Controller Table 13-30. RAM Word Field Descriptions (continued) Bits 16 17 18 Name G3T1 G3T3 Description General-purpose line 3 timing 1. Defines the state (0 or 1) of LGPL3 during bus clock quarter phases 1 and 2 (first half phase). General-purpose line 3 timing 3. Defines the state (0 or 1) of LGPL3 during bus clock quarter phases 3 and 4 (second half phase). G4T1/DLT3 General-purpose line 4 timing 1/delay time 3. The function of this bit is determined by MxMR[GPL4]. If M xMR[GPL4] = 0 and LGPL4/LUPWAIT pin functions as an output (LGPL4), G4T1/DLT3 defines the state (0 or 1) of LGPL4 during bus clock quarter phases 1 and 2 (first half phase). If MxMR[GPL4] = 1 and LGPL4/LUPWAIT functions as an input (LUPWAIT), if a read burst or single read is executed, G4T1/DLT3 defines the sampling of the data bus as follows: 0 In the current word, the data bus should be sampled at the start of bus clock quarter phase 1 of the next bus clock cycle. 1 In the current word, the data bus should be sampled at the start of bus clock quarter phase 3 of the current bus clock cycle. G4T3/WAEN General-purpose line 4 timing 3/wait enable. Bit function is determined by MxMR[GPL4]. If M xMR[GPL4] = 0 and LGPL4/LUPWAIT pin functions as an output (LGPL4), G4T3/WAEN defines the state (0 or 1) of LGPL4 during bus clock quarter phases 3 and 4 (second half phase). If M xMR[GPL4] = 1 and LGPL4/LUPWAIT functions as an input (LUPWAIT), G4T3/WAEN is used to enable the wait mechanism: 0 LUPWAIT detection is disabled. 1 LUPWAIT is enabled. If LUPWAIT is detected as being asserted, a freeze in the external signals logical values occurs until LUPWAIT is detected as being negated. G5T1 G5T3 REDO General-purpose line 5 timing 1. Defines the state (0 or 1) of LGPL5 during bus clock quarter phases 1 and 2 (first half phase). General-purpose line 5 timing 3. Defines the state (0 or 1) of LGPL5 during bus clock quarter phases 3 and 4 (second half phase). Redo current RAM word. Defines the number of times to execute the current RAM word. 00 Once (normal operation) 01 Twice 10 Three times 11 Four times Loop start/end. The first RAM word in the RAM array where LOOP is 1 is recognized as the loop start word. The next RAM word where LOOP is 1 is the loop end word. RAM words between, and including the start and end words, are defined as part of the loop. The number of times the UPM executes this loop is defined in the corresponding loop fields of the MxMR. 0 The current RAM word is not the loop start word or loop end word. 1 The current RAM word is the start or end of a loop. 19 20 21 22–23 24 LOOP MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 13-70 Freescale Semiconductor Functional Description Table 13-30. RAM Word Field Descriptions (continued) Bits 25 Name EXEN Description Exception enable. Allows branching to an exception pattern at the exception start address (EXS). When an internal bus monitor time-out exception is recognized and EXEN in the RAM word is set, the UPM branches to the special exception start address (EXS) and begins operating as the pattern defined there specifies. The user should provide an exception pattern to negate signals controlled by the UPM in a controlled fashion. For DRAM control, a handler should negate RAS and CAS to prevent data corruption. If EXEN = 0, exceptions are ignored by UPM (but not by Local Bus) and execution continues. After the UPM branches to the exception start address, it continues reading until the LAST bit is set in the RAM word. 0 The UPM continues executing the remaining RAM words, ignoring any internal bus monitor time-out. 1 The current RAM word allows a branch to the exception pattern after the current cycle if an exception condition is detected. Address multiplexing. Determines the source of LAD[0:31] during a LALE phase. Any change in the AMX field initiates a new LALE (address) phase. 00 LAD[0:31] is the non-multiplexed address. For example, column address. 01 Reserved 10 LAD[0:31] is the address multiplexed according to MxMR[AM]. For example, row address. 11 LAD[0:31] is the contents of MAR. Used, for example, to initialize a mode. Note that Source ID debug mode is only supported for the AMX = 00 setting. Next burst address. Determines when the address is incremented during a burst access. 0 The address increment function is disabled. 1 The address is incremented in the next cycle. In conjunction with the BRn[PS], the increment value of the state of LA[27:31] is 1, 2 or 4 for port sizes of 8-bits, 16-bits and 32-bits, respectively. UPM transfer acknowledge. Indicates assertion of transfer acknowledge in the current cycle. 0 Transfer acknowledge is not asserted in the current cycle. 1 Transfer acknowledge is asserted in the current cycle. Turn-on disable timer. The disable timer associated with each UPM allows a minimum time to be guaranteed between two successive accesses to the same memory bank. This feature is critical when DRAM requires a RAS precharge time. TODT turns the timer on to prevent another UPM access to the same bank until the timer expires.The disable timer period is determined in M xMR[DSn]. The disable timer does not affect memory accesses to different banks. Note that TODT must be set together with LAST, otherwise it is ignored. 0 The disable timer is turned off. 1 The disable timer for the current bank is activated preventing a new access to the same bank (when controlled by the UPMs) until the disable timer expires. For example, precharge time. Last word. When LAST is read in a RAM word, the current UPM pattern terminates and control signal timing set in the RAM word is applied to the current (and last) cycle. However, if the disable timer is activated and the next access is to the same bank, execution of the next UPM pattern is held off and the control signal values specified in the last word are extended in duration for the number of clock cycles specified in MxMR[DSn]. 0 The UPM continues executing RAM words. 1 Indicates the last RAM word in the program. The service to the UPM request is done after this cycle concludes. 26–27 AMX 28 NA 29 UTA 30 TODT 31 LAST MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 13-71 Local Bus Controller 13.4.4.4.2 Chip-Select Signal Timing (CSTn) If BRn[MSEL] of the accessed bank selects a UPM on the currently requested cycle, the UPM manipulates the LCSn for that bank with timing as specified in the UPM RAM word CSTn fields. The selected UPM affects only the assertion and negation of the appropriate LCSn signal. The state of the selected LCSn signal of the corresponding bank depends on the value of each CSTn bit. Figure 13-59 shows how UPMs control LCSn signals. Bank Selected Switch UPMA/B/C BRn[MSEL] LCS0 LCS1 LCS2 LCS3 SDRAM MUX LCS4 LCS5 LCS6 GPCM LCS7 Figure 13-59. LCSn Signal Selection 13.4.4.4.3 Byte Select Signal Timing (BSTn) If BRn[MSEL] of the accessed memory bank selects a UPM on the currently requested cycle, the selected UPM affects the assertion and negation of the appropriate LBS[0:3] signal. The timing of all four byte-select signals is specified in the RAM word. However, LBS[0:3] are also controlled by the port size of the accessed bank, the number of bytes to transfer, and the address accessed. Figure 13-60 shows how UPMs control LBS[0:3]. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 13-72 Freescale Semiconductor Functional Description Bank Selected A[29:31] Byte count BRn[MSEL] BRn[PS] UPMA Byte-Select Logic LBS0 LBS1 LBS2 LBS3 MUX UPMB UPMC Figure 13-60. LBS Signal Selection The uppermost byte select (LBS0), when asserted, indicates that LAD[0:7] contains valid data during a cycle. Likewise, LBS1 indicates that LAD[8:15] contains valid data, LBS2 indicates that LAD[16:23] contains valid data, and LBS3 indicates that LAD[24:31] contains valid data. For a UPM refresh timer request, all LBS[0:3] signals are asserted/negated by the UPM according to the refresh pattern only. Following any internal bus monitor exception, LBS[0:3] signals are negated regardless of the exception handling provided by any UPM exception pattern to prevent spurious writes to external RAM. 13.4.4.4.4 General-Purpose Signals (GnTn, GOn) The general-purpose signals (LGPL[0:5]) each have two bits in the RAM word that define the logical value of the signal to be changed at the rising edge of the bus clock and/or at the falling edge of the bus clock. LGPL0 offers enhancements beyond the other LGPLn lines. GPL0 can be controlled by an address line specified in MxMR[G0CL]. To use this feature, G0H and G0L should be set in the RAM word. For example, for a SIMM with multiple banks, this address line can be used to switch between internal memory device banks. 13.4.4.4.5 Loop Control (LOOP) The LOOP bit in the RAM word specifies the beginning and end of a set of UPM RAM words that are to be repeated. The first time LOOP = 1, the memory controller recognizes it as a loop start word and loads the memory loop counter with the corresponding contents of the loop field shown in Table 13-31. The next RAM word for which LOOP = 1 is recognized as a loop end word. When it is reached, the loop counter is decremented by one. Continued loop execution depends on the loop counter. If the counter is not zero, the next RAM word executed is the loop start word. Otherwise, the next RAM word executed is the one after the MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 13-73 Local Bus Controller loop end word. Loops can be executed sequentially but cannot be nested. Also, special care must be taken if LAST and LOOP must not be set together. Table 13-31. MxMR Loop Field Use Request Serviced Read single-beat cycle Read burst cycle Write single-beat cycle Write burst cycle Refresh timer expired RUN command Loop Field RLF RLF WLF WLF TLF RLF 13.4.4.4.6 Repeat Execution of Current RAM Word (REDO) The REDO function is useful for wait-state insertion in a long UPM routine that would otherwise need too many RAM words. Setting the REDO bits of the RAM word to a nonzero value causes the UPM to re-execute the current RAM word up to three more times, as defined in the REDO field of the current RAM word. Special care must be taken in the following cases: • • • • • When UTA and REDO are set together, TA is asserted the number of times specified by the REDO function. When NA and REDO are set together, the address is incremented the number of times specified by the REDO function. When LOOP and REDO are set together, the loop mechanism works as usual and the line is repeated according to the REDO function. LAST and REDO must not be set together. REDO should not be used within the exception routine. 13.4.4.4.7 Address Multiplexing (AMX) The address lines can be controlled by the pattern the user provides in the UPM. The address multiplex bits can choose between driving the transaction address, driving it according to the multiplexing specified by the MxMR[AM] field, or driving the MAR contents on the address signals. In all cases, LA[27:31] of the LBC are driven by the five lsbs of the address selected by AMX, regardless of whether the NA bit of the RAM word is used to increment the current address. The effect of NA = 1 is visible only when AMX = 00 chooses the column address. Table 13-32 shows how MxMR[AM] settings affect address multiplexing when the RAM word AMX = 10. The 16 msbs of the LAD[0:31] bus during an address phase are driven with zero in the AMX = 10 case. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 13-74 Freescale Semiconductor Functional Description Table 13-32. UPM Address Multiplexing LAD[0:31] AM as Address A0–A15 A16 A17 A18 A19 A20 A21 A22 A23 A24 A25 A26 A27 A28 A29 A30 A31 Signals 000 Signal driven on 001 external pin 010 when address 011 multiplexing 100 is enabled— RAM word 101 AMX = 10 0 0 0 0 0 0 A8 A7 A6 A5 A4 A3 A9 A8 A7 A6 A5 A4 A10 A11 A12 A13 A14 A15 A16 A17 A18 A19 A20 A21 A22 A23 A9 A8 A7 A6 A5 A10 A11 A12 A13 A14 A15 A16 A17 A18 A19 A20 A21 A22 A9 A8 A7 A6 A10 A11 A12 A13 A14 A15 A16 A17 A18 A19 A20 A21 A9 A8 A7 A10 A11 A12 A13 A14 A15 A16 A17 A18 A19 A20 A9 A8 A10 A11 A12 A13 A14 A15 A16 A17 A18 A19 A9 A10 A11 A12 A13 A14 A15 A16 A17 A18 Note that any change to the AMX field from one RAM word to the next RAM word executed results in an address phase on the LAD[0:31] bus with the assertion of LALE for the number of cycles set for LALE in the ORn and LCRR registers. The LGPL[0:5] signals maintain the value specified in the RAM word during the LALE phase. 13.4.4.4.8 Data Valid and Data Sample Control (UTA) When a read access is handled by the UPM, and the UTA bit is 1 (data is to be sampled by the LBC), the value of the DLT3 bit in the same RAM word, in conjunction with MxMR[GPLn4DIS], determines when the data input is sampled by the LBC as follows: • If MxMR[GPLn4DIS] = 1 (G4T4/DLT3 functions as DLT3) and DLT3 = 1 in the RAM word, data is latched on the falling edge of the bus clock instead of the rising edge. The LBC samples the data on the next falling edge of the bus clock, which is during the middle of the current bus cycle. This feature should be used only in systems without external synchronous bus devices that require mid-cycle sampling. If GPLn4DIS = 0 (G4T4/DLT3 functions as G4T4), or if GPLn4DIS = 1 but DLT3 = 0, data is latched on the rising edge of the bus clock, which occurs at the end of the current bus clock cycle (normal operation). • MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 13-75 Local Bus Controller Figure 13-61 shows how data sampling is controlled by the UPM. UPM read AND GPL4nDIS = 1 AND DLT3 = 1 To internal data bus M U1 L T I P L E0 X O R LCLK LAD[0:31] Figure 13-61. UPM Read Access Data Sampling 13.4.4.4.9 LGPL[0:5] Signal Negation (LAST) When the LAST bit is read in a RAM word, the current UPM pattern is terminated at the end of the current cycle. On the next cycle (following LAST) all the UPM signals are negated unconditionally (driven to logic 1), unless there is a back-to-back UPM request pending. In this case, the signal values for the cycle following the one in which the LAST bit was set are taken from the first RAM word of the pending UPM routine. 13.4.4.4.10Wait Mechanism (WAEN) The WAEN bit in the RAM array word can be used to enable the UPM wait mechanism in selected UPM RAM words. If the UPM reads a RAM word with WAEN set, the external LUPWAIT signal is sampled and synchronized by the memory controller as if it were an asynchronous signal. The WAEN bit is ignored if LAST = 1 in the same RAM word. Synchronization of LUPWAIT starts at the rising edge of the bus clock and takes at least 1 bus cycle to complete. If LUPWAIT is asserted and WAEN = 1 in the current UPM word, the UPM is frozen until LUPWAIT is negated. The value of external signals driven by the UPM remains as indicated in the previous RAM word. When LUPWAIT is negated, the UPM continues normal functions. Note that during WAIT cycles, the UPM does not handle data. Figure 13-62 shows how the WAEN bit in the word read by the UPM and the LUPWAIT signal are used to hold the UPM in a particular state until LUPWAIT is negated. As the example shows, the LCSn and LGPL1 states and the WAEN value are frozen until LUPWAIT is recognized as negated. WAEN is typically set before the line that contains UTA = 1. Note that if WAEN and NA are both set in the same RAM word, NA causes the burst address to increment once as normal regardless of whether the UPM freezes. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 13-76 Freescale Semiconductor Functional Description LCLK T1 T2 T3 T4 LCSn LGPL1 TA WAEN LUPWAIT Word n Word n+1 Word n+2 Wait Wait Word n+3 c1 c2 c3 c4 c5 c6 c7 c8 c9 c10 c11 A B C c12 c13 c14 D Figure 13-62. Effect of LUPWAIT Signal 13.4.4.5 Synchronous Sampling of LUPWAIT for Early Transfer Acknowledge If LUPWAIT is to be considered an asynchronous signal, which can be asserted/negated at any time, no UPM RAM word must contain both WAEN = 1 and UTA = 1 simultaneously. However, programming WAEN = 1 and UTA = 1 in the same RAM word allows UPM to treat LUPWAIT as a synchronous signal, which must meet set-up and hold times in relation to the rising edge of the bus clock. In this case, as soon as UPM samples LUPWAIT negated on the rising edge of the bus clock, it immediately generates an internal transfer acknowledge, which allows a data transfer one bus clock cycle later. The generation of transfer acknowledge is early because LUPWAIT is not re-synchronized, and the acknowledge occurs regardless of whether UPM was already frozen in WAIT cycles or not. This feature allows the synchronous negation of LUPWAIT to affect a data transfer, even if UTA, WAEN, and LAST are set simultaneously. 13.4.4.6 Extended Hold Time on Read Accesses Slow memory devices that take a long time to turn off their data bus drivers on read accesses should choose some non-zero combination of ORn[TRLX] and ORn[EHTR]. The next accesses after a read access to the slow memory device is delayed by the number of clock cycles specified in the ORn register in addition to any existing bus turn around cycle. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 13-77 Local Bus Controller 13.4.4.7 Memory System Interface Example Using UPM Connecting the local bus UPM controller to a DRAM device requires a detailed examination of the timing diagrams representing the possible memory cycles that must be performed when accessing this device. This section describes timing diagrams for various UPM configurations, using fast-page mode DRAM as an example, with LCRR[CLKDIV] = 4 or 8. These illustrative examples may not represent the timing necessary for any specific device used with the LBC. Here, LGPL1 is programmed to drive R/W of the DRAM, although any LGPLn signal may be used for this purpose. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 13-78 Freescale Semiconductor Functional Description LCLK LAD LALE A Row Column Address Address Read Data TA LA LCSn (RAS) LBSn (CAS) LGPL1 (R/W) LBCTL cst1 cst2 cst3 cst4 bst1 bst2 bst3 bst4 g0l0 g0l1 g0h0 g0h1 g1t1 g1t3 g2t1 g2t3 g3t1 g3t3 g4t1 g4t3 g5t1 g5t3 redo[0] redo[1] loop exen amx0 amx1 na uta todt last 0 0 0 0 1 1 1 1 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 Bit 0 Bit 1 Bit 2 Bit 3 Bit 4 Bit 5 Bit 6 Bit 7 Bit 8 Bit 9 Bit 10 Bit 11 Bit 12 Bit 13 Bit 14 Bit 15 Bit 16 Bit 17 Bit 18 Bit 19 Bit 20 Bit 21 Bit 22 Bit 23 Bit 24 Bit 25 Bit 26 Bit 27 Bit 28 Bit 29 Bit 30 Bit 31 Row lsbs Column lsbs LALE pause (due to change in AMX) 1 1 1 1 1 1 0 0 1 0 0 0 0 0 RSS RSS+1 0 0 0 0 0 0 0 0 RSS+1 0 0 0 0 0 1 1 1 RSS+2 Figure 13-63. Single-Beat Read Access to FPM DRAM MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 13-79 Local Bus Controller LCLK LAD LALE A Row Column Address Write Data Address Write Data TA LA LCSn (RAS) LBSn (CAS) LGPL1 (R/W) LBCTL cst1 cst2 cst3 cst4 bst1 bst2 bst3 bst4 g0l0 g0l1 g0h0 g0h1 g1t1 g1t3 g2t1 g2t3 g3t1 g3t3 g4t1 g4t3 g5t1 g5t3 redo[0] redo[1] loop exen amx0 amx1 na uta todt last 0 0 0 0 1 1 1 1 0 0 0 0 1 1 0 0 0 0 0 1 0 0 0 1 Bit 0 Bit 1 Bit 2 Bit 3 Bit 4 Bit 5 Bit 6 Bit 7 Bit 8 Bit 9 Bit 10 Bit 11 Bit 12 Bit 13 Bit 14 Bit 15 Bit 16 Bit 17 Bit 18 Bit 19 Bit 20 Bit 21 Bit 22 Bit 23 Bit 24 Bit 25 Bit 26 Bit 27 Bit 28 Bit 29 Bit 30 Bit 31 Row lsbs Column lsbs LALE pause (due to change in AMX) 0 0 0 0 0 0 0 0 1 0 0 0 0 0 WSS WSS+1 0 0 0 0 0 0 0 0 WSS+1 0 0 0 0 0 1 1 1 WSS+2 Figure 13-64. Single-Beat Write Access to FPM DRAM MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 13-80 Freescale Semiconductor Functional Description LCLK LAD LALE A Row Column 1 Column 2 Address Address Data 1 Data 2 TA LA LCSn (RAS) LBS n (CAS) LGPL1 (R/W) LBCTL cst1 cst2 cst3 cst4 bst1 bst2 bst3 bst4 g0l0 g0l1 g0h0 g0h1 g1t1 g1t3 g2t1 g2t3 g3t1 g3t3 g4t1 g4t3 g5t1 g5t3 redo[0] redo[1] loop exen amx0 amx1 na uta todt last 0 0 0 0 1 1 1 1 0 0 0 0 1 1 1 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 Bit 0 Bit 1 Bit 2 Bit 3 Bit 4 Bit 5 Bit 6 Bit 7 Bit 8 Bit 9 Bit 10 Bit 11 Bit 12 Bit 13 Bit 14 Bit 15 Bit 16 Bit 17 Bit 18 Bit 19 Bit 20 Bit 21 Bit 22 Bit 23 Bit 24 Bit 25 Bit 26 Bit 27 Bit 28 Bit 29 Bit 30 Bit 31 Row lsbs Column 1 lsbs Column 2 lsbs LALE pause (due to change in AMX) 1 1 1 1 1 1 1 1 0 0 1 0 0 0 0 0 RBS 1 0 0 0 0 0 0 0 RBS+1 1 1 0 0 1 1 0 0 RBS+2 0 0 0 0 0 0 1 1 RBS+3 Figure 13-65. Burst Read Access to FPM DRAM Using LOOP (Two Beats Shown) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 13-81 Local Bus Controller LCLK LAD LALE A TA LA LCSn (RAS) LBSn (CAS) LBCTL cst1 cst2 cst3 cst4 bst1 bst2 bst3 bst4 g0l0 g0l1 g0h0 g0h1 g1t1 g1t3 g2t1 g2t3 g3t1 g3t3 g4t1 g4t3 g5t1 g5t3 redo[0] redo[1] loop exen amx0 amx1 na uta todt last 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 1 1 Bit 0 Bit 1 Bit 2 Bit 3 Bit 4 Bit 5 Bit 6 Bit 7 Bit 8 Bit 9 Bit 10 Bit 11 Bit 12 Bit 13 Bit 14 Bit 15 Bit 16 Bit 17 Bit 18 Bit 19 Bit 20 Bit 21 Bit 22 Bit 23 Bit 24 Bit 25 Bit 26 Bit 27 Bit 28 Bit 29 Bit 30 Bit 31 0 0 0 0 0 0 0 0 PTS 0 0 0 0 0 0 0 0 PTS+1 0 0 0 0 0 0 1 1 PTS+2 Figure 13-66. Refresh Cycle (CBR) to FPM DRAM MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 13-82 Freescale Semiconductor Functional Description LCLK LAD LALE TA LCSn (RAS) LBSn (CAS) LBCTL cst1 cst2 cst3 cst4 bst1 bst2 bst3 bst4 g0l0 g0l1 g0h0 g0h1 g1t1 g1t3 g2t1 g2t3 g3t1 g3t3 g4t1 g4t3 g5t1 g5t3 redo[0] redo[1] loop exen amx0 amx1 na uta todt last 1 1 1 1 1 1 1 1 Bit 0 Bit 1 Bit 2 Bit 3 Bit 4 Bit 5 Bit 6 Bit 7 Bit 8 Bit 9 Bit 10 Bit 11 Bit 12 Bit 13 Bit 14 Bit 15 Bit 16 Bit 17 Bit 18 Bit 19 Bit 20 Bit 21 Bit 22 Bit 23 Bit 24 Bit 25 Bit 26 Bit 27 Bit 28 Bit 29 Bit 30 Bit 31 0 0 0 0 0 0 1 1 EXS Figure 13-67. Exception Cycle MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 13-83 Local Bus Controller 13.5 Initialization/Application Information 13.5.1 Interfacing to Peripherals 13.5.1.1 Multiplexed Address/Data Bus and Unmultiplexed Address Signals To save pins on the local bus, address and data are multiplexed onto the same 32 bit bus. An external latch is needed to demultiplex and reconstruct the original address. No external intelligence is needed, because the LALE signal provides the correct timing to control a standard logic latch. The LAD pins can be directly connected to the data signals of the memory/peripheral. Transactions on the local bus start with an address phase, where the LBC drives the transaction address on the LAD signals and asserts the LALE signal. This can be used to latch the address and then the LBC can continue with the data phase. The LBC supports port sizes of 8,16, and 32 bit. For devices smaller than 32 bits, transactions must be broken down. For this reason, LA[30:31] are driven unmultiplexed. For 8-bit devices, LA[30:31] should be used and for 16-bit devices, LA[30] should be used. 32-bit devices use neither of these signals. In addition, the LBC supports burst transfers (not in the GPCM machine). LA[27:29] are the burst addresses within a natural 32-byte burst. To minimize the amount of address phases needed on the local bus and to optimize the throughput, those signals are driven separately and should be used whenever a device requires the 5 least significant addresses. Those should not be used from LAD[27:31]. All other addresses, A[0:26], must be reconstructed through the latch. Local Bus Interface LAD[0:31] D Q LALE LA[27:31] Muxed Address/Data Unmuxed Address LE Latch A[27:31] A[0:26] D[0:31] Figure 13-68. Multiplexed Address/Data Bus 13.5.1.2 Peripheral Hierarchy on the Local Bus To achieve high bus speed interfaces for synchronous SRAMs or SDRAMs, a hierarchy of the memories/peripherals connected to the local bus is suggested, as shown in Figure 13-69. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 13-84 Freescale Semiconductor Initialization/Application Information Local Bus Interface LA[27:31] MA SDRAM A DQ LAD[0:31] D Q LALE LE Latch A B Buffer Muxed Address/Data Unmuxed Address Buffered Data DQ DQ Slower memories and peripherals SSRAM A A LBCTL DIR Figure 13-69. Local Bus Peripheral Hierarchy The multiplexed address/data bus sees the capacitive loading of the data pins of the fast SDRAMs or synchronous SRAMs plus one load for an address latch plus one load for a buffer to the slow memories. The loadings of all other memories and peripherals are hidden behind the buffer and the latch. The system designer needs to investigate the loading scenario and ensure that I/O timings can be met with the loading determined by the connected components. 13.5.1.3 Peripheral Hierarchy on the Local Bus for Very High Bus Speeds To achieve the highest possible bus speeds on the local bus, it is recommended to reduce the number of devices connected directly to the local bus even further. For those cases probably only one bank of synchronous SRAMs or SDRAMs should be used and instead of using a separate latch and a separate bus transceiver, a bus demultiplexer combining those two functions into one device should be used. Figure 13-70 shows an example of such a hierarchy. This section is only a guideline and the board designer must simulate the electric characteristics of his scenario to determine the maximum operating frequency. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 13-85 Local Bus Controller Local Bus Interface LA[27:31] MA SDRAM A LAD[0:31] Latch A/D Q LALE LE A LBCTL DIR B DQ DQ Slower memories and peripherals Muxed Address/Data Unmuxed Address Buffered Data Figure 13-70. Local Bus Peripheral Hierarchy for Very High Bus Speeds 13.5.1.4 GPCM Timings In case a system contains a memory hierarchy with high speed synchronous memories (SDRAM, synchronous SRAM) and lower speed asynchronous memories (for example, FLASH EPROM and peripherals) the GPCM-controlled memories should be decoupled by buffers to reduce capacitive loading on the bus. Those buffers have to be taken into account for the timing calculations. Local Bus Interface LAD[0:31] Latch LALE A Buffer Device A input pin Slower memories and peripherals LBCTL Muxed Address/Data Unmuxed Address Buffered Address Figure 13-71. GPCM Address Timings To calculate address setup timing for a slower peripheral/memory device, several parameters have to be added: propagation delay for the address latch, propagation delay for the buffer and the address setup for the actual peripheral. Typical values for the two propagation delays are in the order of 3–6 ns, so for a 166 MHz bus frequency, LCS should arrive on the order of 3 bus clocks later. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 13-86 Freescale Semiconductor Initialization/Application Information For data timings, only the propagation delay of one buffer plus the actual data setup time has to be considered. Local Bus Interface Device D input pin Slower memories and peripherals LAD[0:31] Buffer LBCTL Muxed Address/Data Buffered Data Figure 13-72. GPCM Data timings 13.5.2 Bus Turnaround Because the local bus uses multiplexed address and data, special consideration must be given to avoid bus contention at bus turnaround. The following cases must be examined: • • • • • • • Address phase after previous read Read data phase after address phase Read-modify-write cycle for parity protected memory banks UPM cycles with additional address phases Continued burst after the first beat Write data phase after address phase Address phase after previous write The bus does not change direction for the following cases so they need no special attention: 13.5.2.1 Address Phase After Previous Read During a read cycle, the memory/peripheral drives the bus and the bus transceiver drives LAD. After the data has been sampled, the output drivers of the external device must be disabled. This can take some time; for slow devices the EHTR feature of the GPCM or the programmability of the UPM should be used to guarantee that those devices have stopped driving the bus when the LBC memory controller ends the bus cycle. In this case, after the previous cycle ends, LBCTL goes high and changes the direction of the bus transceiver. The LBC then inserts a bus turnaround cycle to avoid contention. The external device has now already placed its data signals in high impedance and no bus contention will occur. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 13-87 Local Bus Controller 13.5.2.2 Read Data Phase After Address Phase During the address phase, LAD actively drives the address and LBCTL is high, driving the bus transceivers in the same direction as during a write. After the end of the address phase, LBCTL goes low and changes the direction of the bus transceiver. The LBC places the LAD signals in high impedance after its tdis(LB). The LBCTL will have its new state after ten(LB) and, because this is an asynchronous input, the transceiver starts to drive those signals after its ten(transceiver) time. The system designer has to ensure, that [ten(LB) + ten(transceiver)] is larger than tdis (LB) to avoid bus contention. 13.5.2.3 Read-Modify-Write Cycle for Parity Protected Memory Banks Principally, a read-modify-write cycle is a read cycle immediately followed by a write cycle. Because the write cycle will have a new address phase in any case, this basically is the same case as an address phase after a previous read. 13.5.2.4 UPM Cycles with Additional Address Phases The flexibility of the UPM allows the user to insert additional address phases during read cycles by changing the AMX field, therefore turning around the bus during one pattern. The LBC automatically inserts a single bus turnaround cycle if the bus (LAD) was previously high impedance for any reason, such as a read, before LALE is driven and LAD is driven with the new address. The turnaround cycle is not inserted on a write, because the bus was already driven to begin with. However, bus contention could potentially still occur on the far side of a bus transceiver. It is the responsibility of the designer of the UPM pattern to guarantee that enough idle cycles are inserted in the UPM pattern to avoid this. 13.5.3 Interface to Different Port-Size Devices The LBC supports 8-, 16-, and 32-bit data port sizes. However, the bus requires that the portion of the data bus used for a transfer to or from a particular port size be fixed. A 32-bit port must reside on D[0:31], a 16-bit port must reside on D[0:15], and an 8-bit port must reside on D[0:7]. The local bus always tries to transfer the maximum amount of data on all bus cycles. Figure 13-73 shows the device connections on the data bus. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 13-88 Freescale Semiconductor Initialization/Application Information Interface Output Register 0 OP0 D[0:7] OP0 OP4 OP1 D[8:15] OP1 OP5 OP2 D[15:23] OP2 OP6 OP3 D[24:31] OP3 OP7 32-Bit Port Size 31 OP4 OP5 OP6 OP7 63 OP0 OP2 OP4 OP6 OP1 OP3 OP5 OP7 16-Bit Port Size OP0 8-Bit Port Size OP7 Figure 13-73. Interface to Different Port-Size Devices Table 13-33 lists the bytes required on the data bus for read cycles. Table 13-33. Data Bus Requirements For Read Cycle Port Size/Data Bus Assignments Transfer Size Address State 1 A[29:31] 0–7 Byte 000 001 010 011 100 101 110 111 OP02 — — — OP4 — — — 32-Bit 8–15 —3 OP1 — — — OP5 — — 16–23 — — OP2 — — — OP6 — 24–31 — — — OP3 — — — OP7 16-Bit 0–7 OP0 — OP2 — OP4 — OP6 — 8–15 — OP1 — OP3 — OP5 — OP7 8-Bit 0–7 OP0 OP1 OP2 OP3 OP4 OP5 OP6 OP7 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 13-89 Local Bus Controller Table 13-33. Data Bus Requirements For Read Cycle (continued) Port Size/Data Bus Assignments Transfer Size Address State 1 A[29:31] 0–7 Half Word 000 001 010 100 101 110 Word 000 100 1 2 32-Bit 8–15 OP1 OP1 — OP5 OP5 — OP1 OP5 16–23 — OP2 OP2 — OP6 OP6 OP2 OP6 24–31 — — OP3 — — OP7 OP3 OP7 16-Bit 0–7 OP0 — OP2 OP4 — OP6 OP0 OP4 8–15 OP1 OP1 OP3 OP5 OP5 OP7 OP1 OP5 8-Bit 0–7 OP0 OP1 OP2 OP4 OP5 OP6 OP0 OP4 OP0 — — OP4 — — OP0 OP4 Address state is the calculated address for port size. OPn: These lanes are read or written during that bus transaction. OP0 is the most-significant byte of a word operand and OP3 is the least-significant byte. 3 — Denotes a byte not driven during that write cycle. 13.5.4 Interfacing to SDRAM The following subsections provide application information on interfacing to SDRAM. 13.5.4.1 Basic SDRAM Capabilities of the Local Bus The LBC provides one SDRAM machine for the local bus. Although there is only one machine, multiple chip selects (LCSn) can be programmed to support multiple SDRAM devices. Note that no limitation exists on the number of chip selects that can be programmed for SDRAM. This means that LCS[1:7] can be programmed to support SDRAM, assuming LCS0 is reserved for the GPCM to connect to Flash memory. If multiple chip selects are configured to support SDRAM on the local bus, each SDRAM device should have the same port size and timing parameters. This means that all option registers (ORn) for the SDRAM chip selects should be programmed exactly the same. NOTE Although in principle it is possible to mix different port sizes and timing parameters, combinations are limited and this operation is not recommended. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 13-90 Freescale Semiconductor Initialization/Application Information All the chip selects share the same local bus SDRAM mode register (LSDMR) for initialization along with the local bus-assigned SDRAM refresh timer register (LSRT) and the memory refresh timer prescaler register (MPTPR) for refresh. For refresh, the memory controller supplies auto refresh to SDRAM according to the time interval specified in LSRT and MPTPR as follows: LSRTx ( MPTPR [ PTP ] ) Refresh Period = ----------------------------------------------------------------System Frequency This represents the time period required between refreshes. When the refresh timer expires, the memory controller issues a CBR to each chip select. Each CBR is separated by one clock. A refresh timing diagram for multiple chip selects is shown in Figure 13-51 in Section 13.4.3.11.1, “SDRAM Refresh Timing.” During a memory transaction dispatched to the local bus, the memory controller compares the memory address with the address information of each chip select (programmed with BRn and ORn). If the comparison matches a chip select that is controlled by SDRAM, the memory controller requests service to the local bus SDRAM machine, depending on the information in BRn. Although multiple chip selects may be programmed for SDRAM, only one chip select is active at any given time; thus, multiple chip selects can share the same SDRAM machine. 13.5.4.2 Maximum Amount of SDRAM Supported Table 13-34 summarizes information based on SDRAM data sheets supplied by Micron. Table 13-34. Micron SDRAM Devices SDRAM Device I/O Port Bank Row Column x4 4 12 10 64 Mbit x8 4 12 9 x16 4 12 8 x32 4 11 8 x4 4 12 11 128 Mbit x8 4 12 10 x16 4 12 9 x32 4 12 8 x4 4 13 11 256 Mbit x8 4 13 10 x16 4 13 9 x32 4 13 8 x4 4 13 12 512 Mbit x8 4 13 11 x16 4 13 10 x32 4 TBD TBD The data port size is programmable, but the following examples use all 32 bits of the local bus. The 32-bit port size requires four SDRAM devices (with 8-bit I/O ports) connected in parallel to a single chip select. If 128-Mbit devices are used, one chip select provides 128-Mbit/device x 4 devices = 64 Mbytes. If 4 chip selects are programmed for SDRAM use, the result is 64 Mbytes x 4 = 256 Mbyte. If 256-Mbit SDRAM devices are used, the total available memory is 512 Mbyte. Consequently 512-Mbit devices allow for 1 Gbyte. Although there is no technical difficulty in supporting multiple chip select configurations, in practice, the user may want to maximize the amount of SDRAM assigned to each chip select to minimize cost. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 13-91 Local Bus Controller 13.5.4.3 SDRAM Machine Limitations This section describes limitations of the local bus SDRAM machine. 13.5.4.3.1 Analysis of Maximum Row Number Due to Bank Select Multiplexing LSDMR[BSMA] is used to multiplex the bank select address. The BSMA field and corresponding multiplexed address are shown below: 000 LA12–LA13 001 LA13–LA14 … 111 LA19–LA20 Note that LA12 is the latched value of LAD12. The highest address pins that the bank selects can be multiplexed with are LA[12:13], which limits the pins for the row address to LA[14:31]. For a 32-bit port, the maximum width of the local bus, LA[30:31] are not connected, and the maximum row is LA[14:29]. The local bus SDRAM machine supports 15 rows, which is sufficient for all devices. 13.5.4.3.2 Bank Select Signals Page-based interleaving allows bank signals to be multiplexed to the higher-order address pins to leave room for future upgrades. For example, a user could multiplex the bank select signals to LA[14:15], leaving LA16 to connect to the address pin for a larger memory size. This allows the system designer to design one board that can be used with a current generation of SDRAM devices and upgraded to the next generation without requiring a new board layout. 13.5.4.3.3 128-Mbyte SDRAM Figure 13-74 shows the connection to an SDRAM of 128 Mbytes. Note that all circuit diagrams are principal connection diagrams and do not show any means of signal integrity. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 13-92 Freescale Semiconductor Initialization/Application Information Local Bus Interface LSDDQM[0:3] DQM0 LCSn LSDRAS LSDWE LSDCAS ••• DQM3 32M x 8 SDRAM CAS CS RAS WE BA[1:0] DQM CKE CLK ADDR[12:0] A10 DQ[0:7] 32M x 8 SDRAM CAS CS RAS WE BA[1:0] DQM CKE CLK ADDR[12:0] A10 DQ[0:7] DATA[24:31] LALE Latch x4 ••• LAD[0:31] LSDA10 LCLK LCKE DATA[0:7] Figure 13-74. 128-Mbyte SDRAM Diagram Table 13-35 shows details about LADn signal connections for the example in Figure 13-74. Table 13-35. LADn Signal Connections to 128-Mbyte SDRAM LAD (latch address) LAD29 LAD28 LAD27 LAD26 LAD25 LAD24 LAD23 LAD22 LAD21 LAD20 LAD19 (no connect) SDRAM Address Pin A0 A1 A2 A3 A4 A5 A6 A7 A8 A9 A10 is connected to LSDA10 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 13-93 Local Bus Controller Table 13-35. LADn Signal Connections to 128-Mbyte SDRAM (continued) LAD (latch address) LAD18 LAD17 LAD16 LAD15 SDRAM Address Pin A11 A12 BA0 (if LSDMR[BSMA] = 011) BA1 (if LSDMR[BSMA] = 011) Consider the following SDRAM organization: • • The 32-bit port size is organized as 8 x 8 x 32 Mbit. Each device has four internal banks, 13 row address lines, and 10 column address lines. The logical address is partitioned as shown in Table 13-36. Table 13-36. Logical Address Bus Partitioning A[0:4] msb of start address A[5:17] Row A[18:19] Bank select A[20:29] Column A[30:31] lsb The following parameters are extracted: • • COLS = 011, 10 column lines ROWS = 100, 13 row lines During the address phase, the SDRAM address port is set as shown in Table 13-37. Table 13-37. SDRAM Device Address Port during Address Phase LA[0:14] — LA[15:16] Internal bank select A[18:19] LA[17:29] Row A[5:17] LA[30:31] No-connect Because the internal bank selects are multiplexed over LA[15:16], LSDMR[BSMA] must be set to 011. Table 13-38 shows the address port configuration during a READ/WRITE command. Table 13-38. SDRAM Device Address Port during READ/WRITE Command LA[0:14] msb of start address LA[15:16] Internal bank select LA[17:18] Don’t care LA[19] AP LA[20:29] Column LA[30:31] No-connect Table 13-39 shows the register configuration for this example. PSRT and MPTPR are not shown but should be programmed according to the device’s specific refresh requirements. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 13-94 Freescale Semiconductor Initialization/Application Information Table 13-39. Register Settings for 128-Mbytes SDRAMs Register BRn Field BA XBA PS MS V ORn AM XAM COLS ROWS LSDMR RFEN OP BSMA RFRC PRETOACT ACTTOROW BL WRC BUFCMD CL Value Base address Ext. Base Address 11 = 32-bit port size 011 = SDRAM-local bus 1 11_1111_1000_0000_0000_0 11 011 100 1 000 011 From device data sheet From device data sheet From device data sheet 0 From device data sheet 0 From device data sheet 13.5.4.3.4 256-Mbyte SDRAM This example uses the same Micron SDRAM as in the previous example, but doubles the number of devices connected and therefore uses two chip selects. 13.5.4.3.5 512-Mbyte SDRAM This example uses the MT48LC64M4A2FB from Micron to implement 512 Mbytes. In this SDRAM organization: • • The 32-bit port size is 8 × 4 × 64 Mbit × 2 chip select lines. Each device has 4 internal banks, 13 row address lines, and 11 column address lines. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 13-95 Local Bus Controller The logical address is partitioned as shown in Table 13-40. Table 13-40. Logical Address Partitioning A[0:3] msb of start address A[4:16] Row A[17:18] Bank select A[19:29] Column A[30:31] lsb The following parameters can be extracted: • • COLS = 100, 11 column lines ROWS = 100, 13 row lines During the address phase, the SDRAM address port is set as in Table 13-41. Table 13-41. SDRAM Device Address Port During Address Phase LA[0:13] — LA[15:16] Internal bank select (A[17:18]) LA[17:29] Row (A[4:16]) LA[30:31] No-connect Because the internal bank selects are multiplexed over LA[15:16], LSDMR[BSMA] must be set to 011. Table 13-42 shows the address port settings during a READ/WRITE command. Table 13-42. SDRAM Device Address Port During READ/WRITE Command LA[0:14] msb of start address LA[15:16] Internal bank select LA[17] Don’t care LA[18] AP LA[19:29] Column LA[30:31] No-connect Table 13-43 shows the register configuration. PSRT and MPTPR are not shown, but they should be programmed according to the specific device’s refresh requirements. Table 13-43. Register Settings for 512-Mbyte SDRAMs Register BRn Field BA XBA PS MS V ORn AM XAM BPD COLS ROWS Value Base address ext. Base address 11 = 32-bit port size 011 = SDRAM-local bus 1 11_1110_0000_0000_0000_0 11 01 100 100 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 13-96 Freescale Semiconductor Initialization/Application Information Table 13-43. Register Settings for 512-Mbyte SDRAMs (continued) Register PSDMR Field RFEN OP BSMA RFRC PRETOACT ACTTOROW BL WRC BUFCMD CL Value 1 000 011 From device data sheet From device data sheet From device data sheet 0 From device data sheet 0 From device data sheet 13.5.4.3.6 Power-Down Mode SDRAMs offer a power-down mode during which the device is not refreshed, and therefore, data is not maintained. This mode is invoked by driving CKE low, while all internal banks are idle; note that they must be precharged first. Figure 13-75 shows the timing. Note that the figure does not show the precharge-all command that is issued by the LBC automatically prior to the self-refresh command. Clock low-to-high transitions CKE Supplier specific minimum time COMMAND (As defined by S, RE, CE, W, & ADDR) NOP NOP NOP New command accepted here Cannot violate minimum refresh specification All banks idle Figure 13-75. SDRAM Power-Down Timing CKE remains low, as long as the device is powered down. After CKE transitions to high, the SDRAM exits the power-down mode. 13.5.4.3.7 Self-Refresh In order to be able to stop activity on the local bus (for power save or debug), while the content of the SDRAM is maintained, the self-refresh mode is supported. This mode is invoked by issuing a MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 13-97 Local Bus Controller self-refresh command to the SDRAM. The LBC applies the same timing as for the auto refresh, but also pulls the SDRAM CKE (LCKE) signal low in the same cycle. This can only be done with all banks being idle; the SDRAM machine must precharge them ahead of this. As long as CKE stays low, the device refreshes itself and does not need to see any refreshes from the local bus. To exit self refresh, CKE simply has to be pulled high. Note that after returning from self-refresh mode the SDRAM needs a supplier-specific time before it can accept new commands and the auto-refresh mechanism has to be started again. Figure 13-76 shows this timing. The SDRAM controller always uses 200 local bus clocks, which should satisfy any SDRAM requirements. As in the case of the power-down mode, the figure does not show the precharge-all command that is issued by the LBC automatically prior to the self-refresh command. Refer to Section 13.4.3.3, “Intel PC133 and JEDEC-Standard SDRAM Interface Commands,” for SDRAM interface commands and information on the self-refresh command. t EXTERNAL CK Stable clock CKE Supplier-specific minimum time S RE CE W NOP An Self-refresh entry All banks idle Self-refresh exit New command can occur here Figure 13-76. SDRAM Self-Refresh Mode Timing MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 13-98 Freescale Semiconductor Initialization/Application Information 13.5.4.3.8 SDRAM Timing To allow for very high speeds on the memory bus, the capacitive loading on the local bus must be taken into consideration as shown in Table 13-44. Table 13-44. SDRAM Capacitance Pin CLK RAS, CAS, WE, CS, CKE, DQM Address DQ0–DQ31 Min 2.0 2.0 2.0 3.5 Max 4.0 5.0 5.0 6.5 Unit pf pf pf pf Note: Capacitance values compiled from worst case numbers from various data sheets from Samsung and Micron To implement a system using the hierarchy described earlier for two synchronous memory banks, one address latch and one buffer loading the multiplexed address/data bus sees a loading of four loads of about 6.5 pF maximum. 30 pF can be used as a nominal load. Table 13-45. SDRAM AC Characteristics Device Speed Unit Parameter 166 MHz Min CLK cycle time CAS latency = 3 CAS latency = 2 CLK to valid output delay CAS latency = 3 CAS latency = 2 Output data hold time CAS latency = 3 CAS latency = 2 Input setup time Input hold time CLK to output in Hi-Z CAS latency = 3 CAS latency = 2 6 — — — 2.5 — 1.5 1 — — 5 — — — — — 5 — Max 1000 133 MHz Min 7.5 7.5 — — 3 3 2 1 5.4 5.4 5.4 5.4 — — — — — — ns ns ns ns ns Max 1000 ns Note: AC characteristics compiled from worst-case numbers from various data sheets from Samsung and Micron Setup and Hold Timing Calculations: Address TOF (time of flight): board layout delay Data TOF (time of flight): board layout delay MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 13-99 Local Bus Controller Clock skew (time of flight): clock skew between the LBC and the clock at the memory device. The local bus DLL feedback mechanism must be used to control this skew to optimize the timing margins, as described in the rest of this subsection. Address setup margin = cycle time – local bus address CTQ – SDRAM address input setup time – address TOF + clock skew Address hold margin = local bus address output hold time + address TOF – SDRAM address input hold time – clock skew Data write to SDRAM setup margin = cycle time – local bus data CTQ – SDRAM data input setup time – data TOF + clock skew Data write to SDRAM hold margin = local bus data output hold time + data TOF – SDRAM data input hold time – clock skew Data read from SDRAM setup margin = cycle time – SDRAM data CTQ – local bus data input setup time – data TOF – clock skew Data read from SDRAM hold margin = SDRAM data output hold time + data TOF – local bus data input hold time + clock skew To improve the timing margins a DLL is used to generate external clocks, which minimize the skew between the local bus and the memory clock. Figure 13-77 shows relative timings for the local bus clock DLL. tP LBC output signals on pins tD = tDLL+tP RBC (reference bus clock) LSYNC_IN LSYNC_OUT tL LCLK Figure 13-77. Local Bus DLL Operation MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 13-100 Freescale Semiconductor Initialization/Application Information 13.5.4.4 Parity Support for SDRAM Contrary to older DRAM technologies, SDRAM devices typically are organized either x4, x8, x16 or x32. There are no mainstream devices that include parity support. To allow for error protection on the local bus an additional SDRAM for the 4 parity bits must be used. Since the local bus allows for SDRAM accesses with less than the full port size, read-modify-write cycles are supported for SDRAM write cycles. Figure 13-78 shows a connection diagram. Local Bus Interface LPBSE LSDDQM[0:3] DQM0 LCS n LSDRAS LSDWE LSDCAS ••• DQM3 32M × 8 SDRAM CAS CS RAS WE BA[1:0] DQM CKE CLK ×4 ADDR[12:0] A10 ••• DQ[0:7] 32M × 8 SDRAM CAS CS RAS WE BA[1:0] DQM CKE CLK ADDR[12:0] A10 DQ[0:7] 32M × 4 SDRAM CAS CS RAS WE BA[1:0] DQM CKE CLK ADDR[12:0] A10 DQ[0:3] LALE Latch LAD[0:31] LSDA10 LCLK LCKE LDP[0:3] DATA[0:7] DATA[24:31] DP[0:3] Figure 13-78. Parity Support for SDRAM 13.5.5 Interfacing to ZBT SRAM In many applications, SDRAM provides sufficient performance for the local bus. However, especially in networking applications, memory access patterns are often random and SDRAM is not optimized for that case. ZBT SRAMs have been designed to optimize the performance in networking applications. This section describes how to interface to ZBT SRAMs. Figure 13-79 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 13-101 Local Bus Controller shows the connections. The UPM is used to generate control signals. The same interfacing is used for pipelined and flow-through versions of ZBT SRAMs. However different UPM patterns must be generated for those cases. Because ZBT SRAMs will mostly be used by performance-critical applications, we assume here that, typically, the maximum width of the local bus of 32 bits will be used. ZBT SRAMs allow different configurations. For the local bus the burst order should be set to linear burst order by tying the mode pin to GND; CKE should also be tied to ground. ZBT SRAMs perform four-beat bursts. Because the LBC generates eight-beat transactions (for 32 bit ports) the UPM breaks down each burst into two consecutive four-beat bursts. The internal address generator of the LBC generates the new A27 for the second burst. Local Bus Interface LBS[0:3] BW[0:1] LCSn LGPL0 LGPL1 LGPL2 GPIO 1M × 18 ZBT SRAM OE CE ADV/LD WE ZZ BW[0:1] MODE CKE CLK SA[19:0] DQ[0:17] 1M × 18 ZBT SRAM OE CE ADV/LD WE ZZ BW[0:1] MODE CKE CLK SA[19:0] DQ[0:17] BW[2:3] LA[27:29] LALE Latch LAD[0:31] LDP[0:3] LCLK DATA[0:15] DP[0:1] DATA[16:31] DP[2:3] Figure 13-79. Interface to ZBT SRAM Because we use linear burst on the SRAM, the device will itself burst with the burst addresses of [0:1:2:3]. The local bus always generates linear bursts and expects [0:1:2:3:4:5:6:7]. Therefore, two consecutive linear bursts of the ZBT SRAM with A27 = 0 for the first burst and A27 = 1 for the second burst give the desired burst pattern. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 13-102 Freescale Semiconductor Initialization/Application Information The UPM also supports single beat accesses. Because the ZBT SRAM does not support this and always responds with a burst, the UPM pattern has to take care that data for the critical beat is provided (for write) or sampled (for read), and that the rest of the burst is ignored (by negating WE). The UPM controller basically has to wait for the end of the SRAM burst to avoid bus contention with further bus activities. ZBT SRAMs have a power down mode, which is invoked by the ZZ pin. Connecting a GPIO pin to ZZ allows use of that power down mode; however accesses to the SRAM while in power down mode do not create valid results. This should be taken care of by the system software. Another observation is that SRAMs are available with natural parity. In the example, we use a ×18 SRAM, which holds two data bytes and two parity bits. While for the support of parity on SDRAM banks the local bus has to use read-modify-write cycles and compromise performance, SRAM banks can be used with natural parity and do not compromise performance for parity support. 13.5.6 Interfacing to DSP Host Ports In many applications, an integrated communications processor aggregates traffic for DSPs and distributes that traffic to the DSPs. The local bus allows connection to a variety of different DSP host ports and this section gives some information on how to interface to some example DSPs. 13.5.6.1 Interfacing to MSC8101 HDI16 This section describes how to interface to the HDI16 peripheral interface of the MSC8101. After initial set-up of the interface, the host and HDI16 device can communicate either by a read and write transaction from the core or, if the setup on the DSP and the host are implemented appropriately, by DMA transfers of the host DMA controller, which can be triggered automatically by signals generated by the HDI16 peripheral. 13.5.6.1.1 HDI16 Peripherals The host interface (HDI16) is a 16-bit-wide, full-duplex, double-buffered parallel port that can directly connect to the data bus of a host processor. It supports a variety of buses and gluelessly connects with a number of industry-standard microcomputers, microprocessors, and DSPs. The HDI16 also supports the 8-bit host data bus, which makes it fully compatible with the DSP56300 HI08 (as viewed by the host side, not from the DSP side). The host bus can operate asynchronously to the SC140 core clock, and the HDI16 registers are divided into two banks. The host register bank is accessible to the external host, and the core register bank is accessible to the SC140 core. The MSC8101HDI16 host port peripheral has two sets of 16-bit-wide registers—one set is only visible internally to the DSP, while the other set is visible only to the external host processor. Figure 13-81 illustrates the relationship between the two sides. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 13-103 Local Bus Controller All of the HDI16 peripheral’s registers are mapped directly onto the MSC8101’s QBus, as defined by the MSC8101 16-bit Digital Signal Processor Reference Manual (order number MSC8101RM); the transmit and receive FIFOs are mapped onto the DMA data bus such that the DMA controller can access them directly without core intervention. The addressing for each of these registers is defined in Section 13.5.6.1.2, “Physical Interconnections.” The HDI16 host port itself is a 16-bit-wide parallel port with various strobe and multiplexing options. The most important HDI16 host port facet is that it is specified as an asynchronous interface and so reduces concerns over clock skew between the HDI16 host port and the host device’s buses. Furthermore, with all the host port registers being accessed with a single chip select and four address lines, as far as the local bus is concerned, the DSP host port is akin to an asynchronous memory mapped region. So, for the HDI16 port in single strobe mode, the host device asserts a chip select, a single data strobe and a read/write line to select HDI16 read or write bus operations. The read and write strobes are also used as the data latch control to complete the bus transactions, obviating the need for any handshake termination signal from the DSP. The UPM programmer is responsible for satisfying the AC timings of the HDI16 transactions. Through appropriate mode selection, the HDI16 peripheral’s feature set can be fully supported by the local bus’s UPM controlled signals. The UPM defined interface can be used with any of the local bus’s eight chip selects to give the 16-bit port size and strobe generation that matches that of the HDI16 host port. Figure 13-80 shows the internal register diagram of the HDI16. DMA Data Bus QBus Data Bus 16 16 16 16 64 64 64 64 HCR HSR HCVR HPCR HOTX HORX ISR ICR CVR RX0 RX1 RX2 RX3 TX0 TX1 TX2 TX3 16 16 16 16 16 Host Bus 16 16 16 16 16 16 Figure 13-80. MSC8101 HDI16 Peripheral Registers MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 13-104 Freescale Semiconductor Initialization/Application Information 13.5.6.1.2 Physical Interconnections The physical interconnections between the UPM controlled local bus and the HDI16 of the MSC8101 HDI16 peripheral are given in Table 13-46 and Figure 13-81. Table 13-46. Local Bus to MSC8101 HDI16 Connections MSC8101 Signal(s) HD[0:15] HA[0] HA[1] HA[2] HA[3] HCS1 HCS2 HRDRW HDS HRRQ/HACK HTRQ/HREQ Type I/O/Z I I I I I I I I O O Host data bus Host address line Host address line Host address line Host address line Host chip select 1 Host chip select 2 Host R/W signal Host data strobe signal Receive host request OP Transmit host request OP Description Connect with Local Bus Signal LAD[0:15] Either LA[27] or latched A25 Either LA[28] or latched A26 LA[29] LA[30] LCSn Tie this to VDD LGPL1 or inverted LBCTL signal LGPLy As required in application As required in application Local Bus Interface LAD[0:31] LALE LA[27:28] LA[29:30] LCSn LGPL1 or inverted LBCTL LGPLy As required in application LAD[0:15] MSC8101 HD[0:15] Latch HA[0:1] HA[2:3] HCS1 HRDRW HDS HRRQ/HACK HTRQ/HREQ HCS2 Figure 13-81. Interface to MSC8101 HDI16 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 13-105 Local Bus Controller The connections are specified as follows: • One chip select, LCSn (whatever is available in the system), is used to memory map accesses from the host local bus to the HDI16 MSC8101’s HDI16 peripheral, and is connected to the HDI16 chip select line (HCS). This interface uses two separate general-purpose strobe lines (LGPL1 and LGPLy): — LGPL1 is programmed to generate the HDI16 read/write signal (HRDRW), which is typically high for a read access and low for write access. In any case, the host port requires a HR/W signal. This can be generated by using LGPL1, and allows it to adopt the timing virtually without restrictions. Alternatively the designer can invert LBCTL to generate this signal. It is the responsibility of the UPM pattern designer to plan for the additional delay of that inverter in the UPM pattern to satisfy the AC timings at the DSP host port. — LGPLy is programmed to generate the HDI16 data strobe (HDS), which must be asserted every 16-bit read or write transaction. Data lines—The bus data lines (LADn) are directly connected to the HDI16 data lines (HDn). DMA request/Service request signals (HRRQ & HTRQ)—as appropriate in the application Address lines—The address lines between the LBC and the HDI16 MSC8101 can either be connected in a straightforward or a specific manner to enable burst transfers across the HDI16. The connections are defined as follows and are described in more detail in the following sub-section: — Either LA[27] or latched value of A25 -> HA0 — Either LA[28] or latched value of A26 -> HA1 — LA[29] -> HA2 — LA[30] -> HA3 Ground lines—In order to provide the best ground plane, it is highly advised that all grounds are common and connected together. • • • • • 13.5.6.1.3 Supporting Burst Transfers As mentioned previously, to facilitate burst transfers the host’s local bus address lines can be connected in a very specific way. First, the local bus A31 signal is not required, as the HDI16 registers are 16-bit word addressed. Secondly, local bus LA[27] and LA[28] signals can be eliminated, so that the host side transmit and receive registers wrap around the same four 16-bit word addresses. For example, host transmit register 0 on the HDI16 peripheral (address 0x04) can be obtained by the host by accessing any of the following memory mapped addresses: 0x20, 0x28, 0x30, or 0x38. This is critical for burst accesses as the source or destination addresses increment after each 16-bit MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 13-106 Freescale Semiconductor Initialization/Application Information access to the interface for all 16 transactions within that burst. By using the addressing as defined, if the first access is at 0x20, the last will be at 0x3e but, more importantly, the four host Tx/Rx registers will have been looped around four times. 13.5.6.1.4 Host 60x Bus: HDI16 Peripheral Interface Hardware Timings The host UPM–controlled local bus and the HDI16 MSC8101’s HDI16 host interface are both programmable. Careful programming of the host chip select registers and UPM can meet the HDI16 MSC8101 host port timings. On any bus access the critical timing for both read and write is typically around the data latch point. For the UPM based read access, the host has the flexibility to latch data on a rising or falling LCLK edge. The falling LCLK edge is used here to latch the HDI16 data into the host MSC8101 at its earliest convenience. After the data is latched, appropriate HDI16 port data hold time is ensured before the data strobe (DS) and chip select (CS1) are negated. On a UPM write cycle, the critical action is in enveloping the DS assertion with CS asserted to ensure proper write data hold time after latching by the HDI16 host port. Special attention needs to be given to both the host read and write access strobe (DS) negation times (HDS assert). The HDI16 MSC8101 specifies some restrictions for consecutive register access, which results in a hold off negation time for the read and write access strobes. Rather than restrict the firmware to avoid consecutive bus accesses to host port registers, the negation hold off times should be accommodated in the UPM hardware interface settings. Additional clocks must be built into the end of UPM based cycle giving appropriate time before the next bus cycle starts. The timings can be readily adapted to allow external decode logic to be added to support chip selects for a larger number of DSP HDI16 host ports. 13.5.6.2 Interfacing to MSC8102 DSI The MSC8102 direct-slave interface (DSI) gives an external host direct access to the MSC8102. It provides the following slave interfaces to an external host: • • Asynchronous SRAM-like interface giving the host single accesses (with no external clock). Synchronous SSRAM-like interface giving the host single or burst accesses of 256 bits (eight beats of 32 bits or four beats of 64 bits) with its external clock decoupled from the MSC8102 internal bus clock. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 13-107 Local Bus Controller The DSI supports 32- or 64-bit data bus. For connection to the local bus the DSI has to be configured in 32-bit mode. This is achieved through the DSP reset configuration. The DSI supports two addressing modes, which are determined during the MSC8102 boot sequence. Refer to details in the MSC8102 documentation. • • Full address bus mode with HA[11:29] used in both 32-bit data mode and 64-bit data mode Sliding window mode with HA[14:29] used in both 32-bit data mode and 64-bit data mode 13.5.6.2.1 DSI in Asynchronous SRAM-Like Mode The local bus supports the DSI single strobe as well as the DSI double strobes of operation. As an example the dual strobe configuration is shown below. Figure 13-82 shows the interface to the MSC8102 DSI for asynchronous mode. Local Bus Interface MSC8102 HD[0:31] LAD[0:31] Latch LALE LA[27:29] LGPL2 LBS[0:3] LUPWAIT INTn LCS n LCSy HCID[0:3] HA[27:29] HRDS/HRDE HWBS/HDBS/ HWBE/HDBE[0:3] HTA HINT HBCS HCS HA[11:26] Figure 13-82. Interface to MSC8102 DSI in Asynchronous Mode The asynchronous SRAM-like mode of the DSI is inherently slower than the synchronous mode and should be used, if only relatively small amounts of data are transferred between the communications controller and the MSC8102. To allow for maximum timing flexibility, the UPM machine of the LBC should be used. The UPM programmer is responsible for ensuring correct setup and hold timings for all signals. The UPM allows sufficient control to satisfy any requirements here. Figure 13-83 shows an asynchronous write access. The DSI samples the host chip ID signals (HCID[0:3]) on the first falling edge of the host write byte strobe signals (HWBS) on which the host chip select signal (HCS) is asserted. If the HCID[0:3] signals match the CHIPID value, the MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 13-108 Freescale Semiconductor Initialization/Application Information DSI is accessed. The DSI will signal with the assertion of the host transfer acknowledge signal (HTA), whether it is ready to sample the host data bus (HD), and the host can terminate the access by immediately negating HWBS. The WAEN feature of the UPM must be used to insert wait states while the DSI is busy. The UWPL bit in the MxMR must be cleared to interpret the correct polarity of HTA. The DSI samples the host address bus (HA) and the host data bus (HD) on the rising edge of HWBS. In addition the assertion of HWBS[0:3] are sampled at the end and are part of the access attributes. Because the UPM is used for this mode, the DCR[4]:HTAAD should be set to 1 and DCR[9–10]:HTADT should be defined to a value different than 00. This mode is to be used in implementations with a pull-up resistor on HTA. The host can start its next access (back-to-back accesses) without negating HCS between accesses. If the next access is not to the same MSC8102, then to prevent contention on the HTA signal, the host must wait until the previous DSI stops driving HTA before it accesses the next device. If the next access is to the same MSC8102, the host must not start consecutive accesses before HTA is actively driven to a value of 1 by the previous access. The easiest way to achieve this is to insert idle cycles at the end of the UPM pattern to guarantee that HTA is inactive. HCS HCID[0:3] HA[11:29] HDST[0:1] HD[0:63] HRDS HWBS[0:7] HTAAD=1 & HTADT=01,10,11 HTA Legend: Timing conventions: 1 Valid value that can be 1 or 0 0 1 0 Don’t care Three-state output signal that is not driven by the DSI HTAAD=0 & HTADT=00 Figure 13-83. Asynchronous Write to MSC8102 DSI Figure 13-84 shows an asynchronous read access. The DSI samples the host address bus (HA) and the HCID on the first falling edge of the host read strobe signal (HRDS) on which the HCS is asserted. If HCID[0:3] match the CHIPID value, the DSI is accessed. When the DCR[8]:RPE bit is set, read access to the memory space (not to the register space) initiates data prefetching from consecutive addresses in the internal memory space. The DSI signals (with the assertion of the host transfer acknowledge signal (HTA)) that data is valid, and the host can sample the host data bus MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 13-109 Local Bus Controller (HD) and terminate the access by negating HRDS. If the data for this access is already in the read buffer due to the prefetch mechanism, assertion time of HTA is improved. The WAEN feature of the UPM must be used to insert wait states while the DSI is busy. MxMR[UWPL] has to be cleared to interpret the correct polarity of the HTA signal. Because the UPM is used in the mode, the DCR[4]:HTAAD should be set to 1 and the drive time control field, DCR[9–10]:HTADT, should be defined to a value different than 00. This mode is specially designed to be used for implementations with a pull-up resistor on HTA. The host can start its next access (back-to-back accesses) without negating the HCS signal between accesses. If the next access is not to the same MSC8102, then to prevent contention on HTA, the host must wait until the previous DSI stops driving HTA before it accesses the next device. If the next access is to the same MSC8102, the host must not start consecutive access before HTA is actively driven to 1 by the previous access. The easiest way to achieve this is to insert idle cycles at the end of the UPM pattern to guarantee that HTA is inactive. HCS HCID[0:3] HA[11:29] HDST[0:1] HD[0:63] HRDS HWBS[0:7] HTAAD = 1 & HTADT = 01,10,11 HTA Legend: Timing conventions: 1 Valid value that can be 1 or 0 0 1 0 Don’t care Three-state output signal that is not driven by the DSI HTAAD = 0 & HTADT = 00 Figure 13-84. Asynchronous Read from MSC8102 DSI 13.5.6.2.2 DSI in Synchronous Mode The synchronous SSRAM-like mode of the DSI is inherently faster than the asynchronous mode and should be used if larger amounts of data are transferred between the communications controller and the MSC8102. This will optimize the bus utilization, especially if several MSC8102s are connected to one local bus. The UPM machine of the LBC must be used to implement this interface. Figure 13-85 shows the interface for synchronous mode. Because the DSI will assert and negate HTA in synchronous mode even within a burst transfer on a clock-by-clock basis and because the MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 13-110 Freescale Semiconductor Initialization/Application Information DSI expects the host to react within one clock cycle, some tricks can be implemented to support the synchronous mode. HTA drives LUPWAIT of the UPM. MxMR[UWPL] must be cleared to interpret the correct polarity of HTA. Because this signal influences the internal state machine of the local bus clock, the local bus cannot react to HTA changes correctly within one local bus clock. Refer to Section 13.4.4.4.10, “Wait Mechanism (WAEN),” for more detailed information. The solution to this lies in that the local bus operates at a higher frequency than the DSI interface of the DSP. The local bus clock can be divided by an integer divider (1:2, 1:3 or 1:4) to generate the DSI clock. This should not be a problem because the local bus is designed for much higher frequencies than the DSI. Because all timings are given in DSP DSI clock cycles, the UPM patterns must be adjusted appropriately and need to assert a signal for 2, 3 or 4 clocks (as many as the divider ratio) instead of one. Fortunately, the UPM has the REDO feature, which allows every UPM RAM entry to be executed 1×, 2×, 3× or 4×, which should be sufficient for any divider ratio that would be used in this case. Local Bus Interface LCLK Clock divider MSC8102 HCLKIN LUPWAIT LGPLy Sync logic HTA HD[0:31] LAD[0:31] Latch LALE LA[27:29] LGPL2 LBS[0:3] INTn LGPLn LCSn LCSy HCID[0:3] HA[27:29] HRDS/HRDE HWBS/HDBS/ HWBE/HDBE[0:3] HINT HBRST HBCS HCS HA[11:26] Figure 13-85. Interface to MSC8102 DSI in Synchronous Mode This solution allows the local bus to react within multiple local bus clocks to the HTA signal and be still within one DSI clock. Typically, LUPWAIT is synchronized internally, and only 2 clocks after LUPWAIT changes, new data can be sampled or presented. For example, if the local bus MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 13-111 Local Bus Controller clock ratio is 3× the DSI clock, data can be sampled in the third local bus sub-clock, which is the last third of the DSI clock. If the local bus clock ratio is only 2× the DSI clock, there is a special mode, where the LUPWAIT is NOT synchronized. Refer to Section 13.4.4.5, “Synchronous Sampling of LUPWAIT for Early Transfer Acknowledge,” for more detailed information. In this mode, data is sampled in the 2nd subclock, which is the second half of the DSI clock. AC timing of LUPWAIT must be met in this mode; otherwise indeterministic behavior may occur. The remaining issue is the synchronization of the UPM cycles to the beginning of the DSI clock cycle. Because the UPM executes n cycles for every cycle of the DSI, a mechanism must be used to ensure that the UPM changes transitions in a way that is synchronized to the DSI clock. The solution is to use a special synchronize cycle at the beginning of the pattern. A GPL signal is used to control a multiplexer and to activate external synchronization logic, which uses the DSI clock to stall the UPM by asserting LUPWAIT until the beginning of the next DSI cycle. After that, this GPL signal must be negated and the multiplexer connects LUPWAIT to HTA instead, for the rest of the bus cycle. Note that the GPL signals should be used in the inverted state of their inactive state (GPL[0:4] are 1 when inactive, GPL5 is 0 when inactive) to start the synchronization process. Figure 13-86 shows an example for a synchronization mechanism for a clock divider of 3. Note that the length of the synchronization cycle depends on the relative start of the synchronization process and varies with every access. It can vary in length from one to n (clock ratio) local bus clocks. The second column (compensation cycle) is intended to compensate for the reaction time of LUPWAIT to get in lockstep with the DSI clock. For example, if the clock divider ratio is 1:3 and the LUPWAIT reaction time is two local bus clocks, because LUPWAIT is synchronized, then one local bus clock should be inserted. LCLK DSI_CLK LGPLy LUPWAIT Figure 13-86. UPM Synchronization Cycle Table 13-47. UPM Synchronization Cycles Sync CYCLE cst1–cst4 bst1–bst4 g0xx g1tx Compensation CYCLE DSI CYCLE 1 Bits 0–3 4–7 8–11 12–13 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 13-112 Freescale Semiconductor Initialization/Application Information Table 13-47. UPM Synchronization Cycles (continued) Sync CYCLE g2tx g3tx g4t1 g4t3 g5tx redo[0] redo[1] loop exen amx0 amx1 na uta todt last 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 1 0 Compensation CYCLE DSI CYCLE 1 Bits 14–15 16–17 18 19 20–21 22 23 24 25 26 27 28 29 30 31 This section describes synchronous single write and read, and synchronous burst write and read operations. The local bus supports the DSI single strobe as well as the DSI double strobes of operation. The dual strobe configuration is shown as an example. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 13-113 Local Bus Controller Synchronous Single Write Figure 13-87 shows a synchronous single write access. HCLKIN HCS HCID[0:3] HA[11:29] HDST[0:1] HD[0:63] HWBE [0:7] HRDE HBRST HTA Legend: Timing conventions: 1 Valid value that can be 1 or 0 0 1 0 Don’t care Three-state output signal that is not driven by the DSI Figure 13-87. Synchronous Single Write to MSC8102 DSI The DSI samples HA, HDST, HCID, HD, HWBE, HRDE, and HBRST on the first HCLKIN rising edge on which HCS is asserted. If HCID[0:3] match the CHIPID value, the DSI is accessed. At least one HWBE signal is asserted, and HRDE and HBRST are negated. Assertion of HTA indicates that the DSI is ready to complete the current access and the host must terminate this access. Because HTA is connected to the LUPWAIT signal of the UPM, all local bus signals are frozen until HTA goes to 0 and then the UPM continues in its pattern. Typically, HTA is asserted immediately. If the write buffer is full, HTA assertion is delayed. HTA is asserted for one HCLKIN cycle, driven to logic 1 in the next cycle, and stops being driven on the next rising edge of HCLKIN. The host can start its next access to the same MSC8102 immediately on the next HCLKIN rising edge without negating HCS between accesses. If the next access is not to the same MSC8102, then, to prevent contention on HTA, the host must wait to access the next device until the previous DSI stops driving HTA. The easiest way to achieve this is to insert idle cycles at the end of the UPM pattern to guarantee that HTA is inactive. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 13-114 Freescale Semiconductor Initialization/Application Information Synchronous Single Read Figure 13-88 shows a synchronous single read access. HCLKIN HCS HCID[0:3] HA[11:29] HDST[0:1] HD[0:63] HWBE[0:7] HRDE HBRST HTA Legend: Timing conventions: 1 Valid value that can be 1 or 0 0 1 Don’t care 0 Three-state output signal that is not driven by the DSI Figure 13-88. Synchronous Single Read from MSC8102 DSI The DSI samples HA, HDST, HCID, HWBE, HRDE, and HBRST on the first HCLKIN rising edge on which HCS is asserted. If the HCID[0:3] signals match the CHIPID value, the DSI is accessed. HRDE is asserted, and HWBE and HBRST are negated. If DCR[8]:RPE is set (see MSC8102 documentation), read access to the memory space (not to the register space) initiates prefetching data from consecutive addresses in the internal memory space. Assertion of HTA indicates that data is valid and the host must sample the HD and terminate the access. Because HTA is connected to the UPM LUPWAIT signal, all local bus signals are frozen until HTA goes to 0; then the UPM continues in its pattern. HTA is asserted earlier when the data for this access is already prefetched to the read buffer. It asserted for one HCLKIN cycle and driven to logic 1 in the next cycle. It stops being driven on the next rising edge of HCLKIN. The host can start its next access to the same MSC8102 immediately in the next HCLKIN rising edge without negating HCS between accesses. If the next access is not to the same MSC8102, then, to prevent contention on HTA, the host must wait to access the next device until the previous DSI stops driving HTA. The easiest way to achieve this is to insert idle cycles at the end of the UPM pattern to guarantee that HTA is inactive. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 13-115 Local Bus Controller Synchronous Burst Write Figure 13-89 shows a synchronous burst write access. HCLKIN HCS HCID[0:3] HA[11:29] HDST[0:1] HD[0:63] HWBE [0:7] HRDE HBRST HTA Legend: Timing conventions: 1 Valid value that can be 1 or 0 0 1 Don’t care 0 n = 3 in 64-bit data bus interface n = 7 in 32-bit data bus interface D(A) D(A+1) D(A+2) D(A+2) D(A+n) A Three-state output signal that is not driven by the DSI Figure 13-89. Synchronous Burst Write to MSC8102 DSI The DSI samples HA, HDST, HCID, HD, HWBE, HRDE, and HBRST on the first HCLKIN rising edge on which HCS is asserted. If HCID[0:3] match the CHIPID value, the DSI is accessed. HWBE are asserted, HBRST is asserted, and HRDE is negated. Assertion of HTA indicates that the DSI is ready to complete the current beat of the access and the host must proceed to the next beat of this access. When the host reaches the last beat of the access, it must terminate the burst access. Typically HTA is asserted immediately for each beat of the access. If the write buffer is full, HTA assertion is delayed. Because HTA is connected to the LUPWAIT signal of the UPM, all local bus signals are frozen until HTA goes to 0 and then the UPM continues in its pattern. After the last beat of the access, HTA is driven to logic 1 and stops being driven on the next rising edge of HCLKIN. The host can start its next access to the same MSC8102 immediately in the next HCLKIN rising edge without negating HCS between accesses. If the next access is not to the same MSC8102, then, to prevent contention on HTA, the host must wait to access the next device until the previous DSI stops driving HTA. The easiest way to achieve this is to insert idle cycles at the end of the UPM pattern to guarantee that HTA is inactive. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 13-116 Freescale Semiconductor Initialization/Application Information Synchronous Burst Read Figure 13-90 shows a synchronous burst read access. The DSI samples HA, HDST, HCID, HWBE, HRDE, and HBRST on the first HCLKIN rising edge on which HCS is asserted. If HCID[0:3] match the CHIPID value, the DSI is accessed. HCLKIN HCS HCID[0:3] HA[11:29] HDST[0:1] HD[0:63] HRDE HWBE[0:7] HBRST HTA Legend: Timing conventions: 1 Valid value that can be 1 or 0 0 1 Don’t care 0 n = 3 in 64-bit data bus interface n = 7 in 32-bit data bus interface D(A) D(A+1) D(A+2) D(A+2) D(A+n) A Three-state output signal that is not driven by the DSI Figure 13-90. Synchronous Burst Read from MSC8102 DSI HRDE and HBRST are asserted and HWBE are negated. When the DCR[8]:RPE bit (see the MSC8102 documentation) is set, a burst read access initiates data prefetching from consecutive addresses in the internal memory space. Assertion of HTA indicates that data is valid for the current beat of the access and the host must proceed to the next beat of this access. Because HTA is connected to the LUPWAIT signal of the UPM, all local bus signals are frozen until HTA goes to 0 and then the UPM continues in its pattern. When the host reaches the last beat of the access, it must terminate the burst access. The HTA is asserted earlier when the data for this access is already prefetched to the read buffer. Typically, after the first beat of the burst access, HTA remains asserted until the end of the access. After the last beat of the access, HTA is driven to 1 and stops being driven in the next rising edge of HCLKIN. The host can start its next access to the same MSC8102 immediately in the next HCLKIN rising edge without negating HCS between accesses. If the next access is not to the same MSC8102, to prevent contention on HTA, the host must wait to access the next device until the previous DSI stops driving the HTA signal. The easiest way to achieve this is insert idle cycles at the end of the UPM pattern to guarantee that HTA is inactive. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 13-117 Local Bus Controller 13.5.6.2.3 Broadcast Accesses Using HBCS, a host can share one chip-select signal between multiple MSC8102 devices for broadcasting write accesses. In broadcast mode, the DSI does not drive its HTA signal to prevent contention between multiple devices driving different values to the same signal. Also, the DSI does not decode HCID[0:3]. Note that broadcasting is allowed only for write accesses. The DSI sets the DSI error register (DER) OVF bit if there is an overflow during broadcast accesses. This bit can be cleared by writing a value of 1 to it. NOTE To avoid overflow when accessing DSI registers during broadcast accesses, wait at least 10 host clock cycles in synchronous mode or 8 internal clock cycles in asynchronous mode between each DSI register access. To avoid data corruption, if DER[0]:OVF is set, no broadcast access is written until the bit is reset. Therefore, after the last broadcast access, and before any regular write access, DER[0]:OVF must first be read and reset if it is set. NOTE In asynchronous mode, write data from a previous access (even a normal write access) may be lost due to overflow during broadcast accesses. To prevent such a loss, ensure that previous access data has propagated to the FIFO or DSI registers, depending on the type of previous access. This can be achieved by performing a read access prior to the first broadcast access. In broadcast accesses, the host must comply with the following rules: • In asynchronous mode, HWBS[0:3]/HDBS[0:3] assertion time should be at least the minimum, which is defined in the AC characteristics section of the MSC8102 Technical Data sheet. In synchronous mode single access, the host must wait 1 cycle before terminating the access. Access signals must be in the same valid state during two positive edges of the host clock cycles. Access duration is two clock cycles (the DSI may translate accesses lasting longer than two clock cycles as two or more back-to-back accesses). In synchronous mode burst accesses, broadcast accesses are not allowed. • • 13.5.6.3 Interfacing to EHPI from Texas Instruments TMS320Cxxxx DSPs The enhanced host port interface (EHPI) on DSPs from Texas Instruments provides a 16-bit-wide parallel port through which a host processor can directly access the memory of the DSP. The host MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 13-118 Freescale Semiconductor Initialization/Application Information and the DSP can exchange information through memory internal or external to the DSP and within the address reach of the EHPI. The EHPI uses 23-bit addresses, where each address is assigned to a 16-bit word in memory. The EHPI has two modes, one for multiplexed address/data and one with separate buses.To allow the connection of multiple DSPs and other peripherals on the local bus, the use of the EHPI non-multiplexed mode is recommended. The EHPI uses the signals in Figure 13-48. Table 13-48. EHPI Signals Signals HD[15:0] HA[19:4] HA[3:0] HBE[1:0] HCS HR/W HDS1, HDS2 I I I I Type Description Connect With LAD[0:15] Latched A[11:26] LA[27:30] LBS[0:1] LCSn LGPLy or inverted LBCTL I/O/Z Host data bus: Non-muxed mode: data only I Host address bus: Non-muxed mode: addresses Host address bus: lsbs Host byte-enable signals 00 Word, 0 MSB, 10 LSB, 11 Reserved Chip select signal R/W signal LGPLn Data strobe signals must be at least 2 DSP clocks wide • Host has separate active-low read and write strobe pins: connect one to HDS1, one to HDS2 • Host has one active-low strobe pin: connect to HDS1 or HDS2, the other to 1 • Host has one active high-strobe pin: connect to HDS1 or HDS2, the other to 0 EHPI ready signal EHPI control signals. Non-muxed mode: • HCNTL=1: access DSP data memory • HCNTL=0: access EHPI control register Address strobe signal EHPI mode signal High: non-muxed mode Low: muxed mode Reset mode signal DSP to host interrupt signal INTn High LUPWAIT Application specific: either GPIO pin or latched address lines HRDY HCNTL0 O I HAS HMODE I I RST_MODE HINT I O To achieve the timings required by the DSP host port and optimize the bus usage, the use of one of the UPMs is recommended. Note that the DSP address signals reflect 16-bit addresses, whereas local bus address signals reflect byte addresses. This essentially shifts address signals by one bit. The EHPI host port’s two strobe pins, HDS1 and HDS2, allow different options to control transfers. The UPM supports any of those options; however, the easiest is to use the single active-low strobe mode and connect one UPM LGPL signal (whatever is available) to HDS. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 13-119 Local Bus Controller In any case, the host port requires a HR/W signal. This can be generated by using another LGPL signal and allows to adopt the timing virtually without restrictions. Alternatively the designer can invert LBCTL to generate this signal. It is the responsibility of the UPM pattern designer to plan for the additional delay of that inverter in the UPM pattern to satisfy AC timings at the DSP host port. MPC8540 Local Bus Interface LAD[0:31] Latch LALE LA[27:30] LGPLy or inverted LBCTL LCSn LBS[1:0] LUPWAIT INT n LGPLn HCNTL0 HA[3:0] HR/W HCS HBE[1:0] HRDY HINT HDS1/HDS2 RST_MODE HMODE LAD[16:31] EHPI HD[15:0] HA[19:4] Figure 13-91. Interface to Texas Instruments EHPI in Non-Multiplexed Mode The DSP does not necessarily have deterministic access times. HRDY indicates whether the EHPI is ready for an access. If the signal is low, wait states must be inserted in the cycle. The LUPWAIT function of the UPM provides this mechanism. MxMR[UWPL] must be set to connect to this active-low signal. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 13-120 Freescale Semiconductor Initialization/Application Information Figure 13-92 shows read timing required by the EHPI in non-multiplexed mode. HCS HDS HR/W HA[18:0] Valid Address HD[15:0] HRDY† Data † While HCS is not selected, HRDY is in Hi-Z state. Figure 13-92. EHPI Non-Multiplexed Read Timings Figure 13-93 shows write timing required by the EHPI in non-multiplexed mode. HCS HDS HR/W HA[18:0] Valid Address Valid Address HD[15:0] HRDY† Data Valid Data Valid † While HCS is not selected, HRDY is in Hi-Z state. Figure 13-93. EHPI Non-Multiplexed Write Timings 13.5.6.3.1 Expansion to Multiple DSPs The connection shown above can be adapted easily to interface to multiple DSPs instead of only one. Each DSP host port needs to receive its own chip select and interrupt signals, and HRDY must be connected differently, depending on which DSP is used. All other signals can be connected to all host ports in parallel. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 13-121 Local Bus Controller HRDY signals can be bussed for certain DSPs (such as, TMS320VC5510); for others (such as, TMS320VC5509), an external multiplexer must be used to decide which HRDY signal is routed to the local bus. This multiplexer can be controlled by the respective chip selects. For a larger number of DSPs, this scheme must be extended with external logic, which uses additional address signals to generate multiple DSP HCSs for one local bus chip select and to multiplex the HINT and HRDY signals. The flexibility of the UPM allows for additional delay for that external logic. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 13-122 Freescale Semiconductor Chapter 14 Three-Speed Ethernet Controllers This chapter describes the two three-speed Ethernet controllers of the MPC8540. The two controllers are referenced as TSEC1 and TSEC2. Note that this chapter does not include the description of the separate 10/100 Ethernet controller (FEC) on the MPC8540. 14.1 Introduction The Ethernet IEEE 802.3 protocol is widely used on LANs that are based on the carrier-sense multiple access/collision detect (CSMA/CD) approach. Because Ethernet and IEEE 802.3 protocols are similar and can coexist on the same LAN, both are referred to as Ethernet in this manual, unless otherwise noted. 10/100 Ethernet provides increased Ethernet speed from 10 to 100 megabits per second (Mbps) and provides a simple, cost-effective option for backbone and server connectivity. This three-speed Ethernet controller (TSEC) also implements a gigabit Ethernet protocol, which builds on top of the Ethernet protocol, but increases speed tenfold over 10/100 Ethernet to 1000 Mbps, or one gigabit per second (Gbps). Gigabit Ethernet looks identical to Ethernet from the data link layer upward but it uses the ANSI X3T11 FiberChannel FC-0 (interface and media) and FC-1 (encode/decode) for the PHY Layer. In this manner, the standard takes advantage of the existing high-speed physical interface technology of FiberChannel while maintaining the IEEE 802.3 Ethernet frame format, backward compatibility for installed media, and the use of full- or half-duplex CSMA/CD. The Ethernet protocol implements the bottom two layers of the open systems interconnection (OSI) 7-layer model, that is, the data link and physical layers. Figure 14-1 shows the typical Ethernet protocol stack and the relationship to the OSI model. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-1 Three-Speed Ethernet Controllers OSI Reference Model Layers LAN CSMA/CD Layers Higher layers Logical link control (LLC) Data Link Media access control (MAC) Reconciliation MII Physical PCS PMA PMD MDI Physical medium 10/100 Mbps 10/100 Ethernet PHY Reconciliation GMII PCS PMA PMD MDI Physical medium 1-Gbps Ethernet Figure 14-1. Ethernet Protocol in Relation to the OSI Protocol Stack Gigabit Ethernet provides the following sublayers: • Media access control (MAC) sublayer—The MAC sublayer provides a logical connection between the MAC and its peer station. Its primary responsibility is to initialize, control, and manage the connection with the peer station. Reconciliation sublayer—The reconciliation sublayer acts as a command translator. It maps the terminology and commands used in the MAC layer into electrical formats appropriate for the physical layer entities. MII (media-independent interface) sublayer—The MII sublayer provides a standard interface between the MAC layer and the physical layer for 10/100 Mbps operations. It isolates the MAC layer and the physical layer, enabling the MAC layer to be used with various implementations of the physical layer. GMII (gigabit media-independent interface) sublayer—The GMII sublayer provides a standard interface between the MAC layer and the physical layer for 1-Gbps operation. It isolates the MAC layer and the physical layer, enabling the MAC layer to be used with various implementations of the physical layer. PCS (physical coding sublayer)—The PCS sublayer is responsible for encoding and decoding data stream to and from the MAC sublayer. Medium (1000BASEX) 8B/10B coding is used for fiber. Medium (1000BASET) 8B1Q coding is used for unshielded twisted pair (UTP). PMA (physical medium attachment) sublayer—The PMA sublayer is responsible for serializing code groups into a bit stream suitable for serial bit-oriented physical devices • • • • • MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-2 Freescale Semiconductor Introduction • • (SerDes) and vice versa. Synchronization is also performed for proper data decoding in this sublayer. The PMA sits between the PCS and the PMD sublayers. For fiber medium (1000BASEX) the interface on the PMD side of the PMA is a 1-bit 1250-MHz signal, while on the PMA’s PCS side the interface is a 10-bit interface (TBI) at 125 MHz. The TBI is an alternative to the GMII interface. If the TBI is used, the gigabit Ethernet controller must be capable of performing the PCS function. For UTP medium, the PMD interface side of the PMA consists of four pair of 62.5-MHz PAM5 encoded signals, while the PCS side provides the 1250 Mbps input to a 8B1Q4 PCS. PMD (physical medium dependent) sublayer—The PMD sublayer is responsible for signal transmission. The typical PMD functionality includes amplifier, modulation, and wave shaping. Different PMD devices may support different media. MDI (medium-dependent interface) sublayer—MDI is a connector. It defines different connector types for different physical media and PMD devices. Figure 14-2 describes the different physical interface standards. MAC full-duplex and/or half-duplex TBI 802.3z Gigabit Media-Independent Interface (GMII) (optional) 802.3ab 1000BASEX PCS 8 bytes/10 bytes auto-negotiation 1000BaseT PCS 1000BaseLX fiber optics Single-mode or multimode fiber (3 Km) 1000BaseSX fiber optics Multimode fiber (500 m) 1000BaseCX copper Shielded copper cable (25 m) 1000BaseT PMA Unshielded twisted pair (100 m) Figure 14-2. IEEE 802.3z and 802.3ab Physical Standards Ethernet/IEEE 802.3 frames are based on the frame structure shown in Figure 14-3. The term ‘packet’ is sometimes used to refer to the frame plus the preamble and start frame delimiter (SFD). Frame length is 64–1518 bytes Preamble 7 bytes Start frame delimiter 1 byte Destination address 6 bytes Source address 6 bytes Type/ Length 2 bytes Data 46–1500 bytes Frame check sequence 4 bytes Figure 14-3. Ethernet/IEEE 802.3 Frame Structure MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-3 Three-Speed Ethernet Controllers The elements of an Ethernet frame are as follows: • • • Preamble— The 7-byte preamble of alternating ones and zeros used for receiver timing synchronization (each byte containing the value 0x55). Start frame delimiter (SFD)—A sequence of 0xD5 (10101011 because the bit ordering is lsb first) indicates the beginning of the frame. 48-bit destination address (DA)—The first bit identifies the address as an individual address (0) or a group address (1). The second bit is used to indicate whether the address is locally-defined (1) or globally-defined (0). 48-bit source address (SA)— (Original versions of the IEEE 802.3 specification allowed 16-bit addressing, which has never been used widely.) Ethernet type field/IEEE 802.3 length field—The type field signifies the protocol (for example, TCP/IP) used in the rest of the frame. The length field specifies the length of the data portion of the frame. For both Ethernet and IEEE 802.3 frames to exist on the same LAN, the length field must be unique from any type fields used in Ethernet. This limitation requires that a type field be identified by a decimal number equal to or greater than 1536 (0x0600) but less than 65535 (0xFFFF). If the number, however, is between 0 and 1,500 (0x0000 through 0x05DC) then this field indicates the length of the MAC client data. The range from 1,501 to 1,535 (0x5DD through 0x5FF) was intentionally left undefined. Data and padding—Padding is optional. It is only needed if the data is smaller than 46 octets (one octet = one byte) to ensure the minimum frame size of 64 octets as specified in the IEEE 802.3 standard. In 802.3x the first two octets of the data field are used as opcode (OP) (pause = 0x0001) and the second two octets are used to transmit a pause time (PT) parameter (pausetime = 0x0000 for on and 0xFFFF for off). In addition, a third two-octet field can be used for an extended pause control parameter (PTE). Because the use of these fields varies with the protocol used, the ability to examine them and report their content can significantly accelerate Ethernet frame processing. Frame-check sequence (FCS)—Specifies the standard 32-bit cyclic redundancy check (CRC) obtained using the standard CCITT-CRC polynomial on all fields except the preamble, SFD and CRC. • • • • Figure 14-4 provides additional details of the Ethernet/IEEE 802.3 frame structure. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-4 Freescale Semiconductor Introduction Frame length is 64–1518 bytes Preamble 7 bytes Start frame delimiter 1 byte Destination address 6 bytes Source address 6 bytes Type/ Length 2 bytes Data 46–1500 bytes Frame check sequence 4 bytes Vendor 3 bytes Bit 0 yzxx_xxxx I/G U/L xxxx_xxxx UserValue DSAP SSAP Control Information LLC frame Destination address xxxx_xxxx xxxx_xxxx xxxx_xxxx Bit 47 xxxx_xxxx Example: 02608C: 876543 (02608C = 3COM, 876543 = UserValue assigned by 3COM) Figure 14-4. Ethernet/IEEE 802.3 Frame Structure With More Details Relative to Figure 14-4, the IEEE 802.3 section 3.11 (MAC frame format) defines the frame format such that the octets of a frame are transmitted from left to right (preamble first, the FCS last), and the bits of each octet are transmitted least-significant bit (lsb) first. The destination address example shown in Figure 14-4 (02608C:876543) which would normally be written as 0000 0010 0110 0000 1000 1100 1000 0111 0110 0101 0100 0011 is transmitted bit by bit as 0100 0000 0000 0110 0011 0001 1110 0001 1010 0110 1100 0010 which is an individual address (because the lsb is cleared) but locally-defined (because the second least-significant bit is set). When first originated, a type field was used for protocol identification. The IEEE 802.3 specification eliminated the type field, replacing it with the length field. The length field is used to identify the length, in bytes, of the data field. The protocol type in 802.3 frames are held within the data portion of the packet. The logical link control (LLC) is responsible for providing services to the network layer regardless of media type, such as FDDI, Ethernet, token ring, and others. The LLC layer makes use of LLC protocol data units (PDUs) in order to communicate between the media access control (MAC) layer and the upper layers of the protocol stack. Three variables determine access into the upper layers via the LLC-PDU. The variables include the destination service access point (DSAP), the source service access point (SSAP), and a control variable. The DSAP address specifies a unique identifier within the station providing protocol information for the upper layer. The SSAP provides the same information for the source address. The LLC defines service access for protocols that conform to the open system interconnection (OSI) model for network protocols. However, many protocols do not obey the rules for those MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-5 Three-Speed Ethernet Controllers layers and additional information must be added to the LLC in order to provide information regarding those protocols. Protocols that fall into this category include IP and IPX. The method used to provide this additional protocol information is called a subnetwork access protocol (SNAP) frame. A SNAP encapsulation is indicated by the DSAP and SSAP addresses being set to 0xAA and the LLC control field being set to 0x03. If that address is seen, a SNAP header follows. The SNAP header is five bytes long. The first three bytes consist of the organization code (SNAP OUI), which is assigned by the IEEE. The last two bytes become the type value set from the original Ethernet specifications if SNAP OUI = 0, or they become a SNAP protocol identifier if SNAP OUI is non zero. The three-speed Ethernet controller (TSEC) allows the flexibility to accelerate the identification and retrieval of all the standard and non-standard protocols mentioned above. Figure 14-5 shows the block diagram of the TSEC. Tx/Rx descriptors Tx/Rx data Tx/Rx status DMA Block interface Rx FIFO 2 Kbytes FIFO control Tx FIFO 2 Kbytes Address/Data filtering Pack words Unpack words Tx system CLK data path MAC Layer GMII MIB Counters Reduced Signal Interface TBI Reduced Signal Interface TBI Clocks To PHY GMII or To PHY RGMI or To PHY RTBI or To PHY TBI MII Mgmt Figure 14-5. TSEC Block Diagram MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-6 Freescale Semiconductor Features 14.1.1 Three-Speed Ethernet Controller Overview The TSEC is designed to support 10, 100, and 1000 Mbps Ethernet/802.3 networks and contains the following components: • • • • • Ethernet media access controller (MAC) First-in first-out (FIFO) controller Direct memory access (DMA) controller Ten-bit interface (TBI) Register-based statistical module that supports management information base (MIB) remote monitoring (RMON) The most-significant byte of data in a receive or transmit data buffer corresponds to the most-significant byte of a frame, respectively. The complete TSEC is designed for single MAC applications. The TSEC supports several standard MAC-PHY interfaces to connect to an external Ethernet transceiver: • • • • MII interface running at 10/100 Mbps GMII interface running at 1 Gbps TBI interface that can be connected to a SerDes device for fibre channel applications. Reduced signal count versions of the GMII (RGMII) and ten-bit (RTBI) interfaces While most of this document refers to the non-reduced signal count interfaces, it must be understood that these references also apply to the reduced signal count interfaces. The TSEC software programming model is similar to the MPC8260 (PowerQUICC II) device. Hence, it enables Freescale customers to leverage already implemented Ethernet drivers, reducing the software development cycle. 14.2 Features The MPC8540 TSEC includes these distinctive features: • • IEEE 802.3, 802.3u, 802.3x, 802.3z, 802.3ac, 802.3ab compliant Support for different Ethernet physical interfaces: — 10/100/1Gb IEEE 802.3 GMII — 10/100 Mbps IEEE 802.3 MII — 10-Mbps IEEE 802.3 MII — 1-Gbps IEEE 802.3z TBI — 10/100 Mbps RMII — 1-Gbps full-duplex RGMII — 1-Gbps RTBI MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-7 Three-Speed Ethernet Controllers • • • • • • • • • • Full- and half-duplex support (1 Gbps supports only full-duplex) IEEE 802.3 full-duplex flow control (automatic PAUSE frame generation or software programmed PAUSE frame generation and recognition) Support for out-of-sequence transmit queue (for initiating flow control) Programmable maximum frame length supports jumbo frames (up to 9.6 Kbytes) and IEEE 802.1 virtual local area network (VLAN) frames Retransmission from transmit FIFO following a collision Support for CRC generation and verification of inbound/outbound packets Address recognition — Each exact match can be programmed to be accepted or rejected — Broadcast address (accept/reject) — Exact match 48-bit individual (unicast) address — Hash (256-bit hash) check of individual (unicast) addresses — Hash (256-bit hash) check of group (multicast) addresses — Promiscuous mode Extraction data and its associated buffer descriptors can be directed to the processor’s L2 cache to reduce access latency Interrupt coalescing subject to a threshold frame counter and/or a threshold timer (Independent functionality for received and transmitted frames) RMON statistics support 14.3 Modes of Operation The primary TSEC operational modes are the following: • Full- and half-duplex operation This is determined by MACCFG2 register’s Full Duplex bit. Full-duplex mode is intended for use on point to point links between switches or end node to switch. Half-duplex mode is used in connections between an end node and a repeater or between repeaters. If configured in half-duplex mode (only 10- and 100-Mbps operation; MACCFG2 register’s Full Duplex bit is cleared), the MAC complies with the IEEE CSMA/CD access method. If configured in full-duplex mode (10/100/1Gb operation; MACCFG2 register’s Full Duplex bit is set), the MAC supports flow control. If flow control is enabled, it allows the MAC to receive or send PAUSE frames. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-8 Freescale Semiconductor External Signal Descriptions • • • • • 10- and 100-Mbps MII interface operation The MAC-PHY interface operates in MII mode by configuring bits 22–23 of MACCFG2 (MACCFG2[I/F Mode] = 01). The MII is the media-independent interface defined by the 802.3 standard for 10/100 Mbps operation. The speed of operation is determined by the TSECn_TX_CLK and TSECn_RX_CLK signals, which are driven by the transceiver. The transceiver either auto-negotiates the speed or it may be controlled by software via the serial management interface (EC_MDC/EC_MDIO signals) to the transceiver. 1-Gbps GMII and TBI interface operation The MAC-PHY interface operates in GMII mode by configuring bits 22–23 of MACCFG2 (MACCFG2[I/F Mode] = 10). The GMII is the gigabit media-independent interface defined by the 802.3 standard for 1-Gbps operation. Independently, the MAC-PHY interface can also operate in TBI mode. Note that either the TBI or GMII interface is chosen; not both at the same time. TBI is the ten-bit interface which contains PCS functions (ten-bit encoding/decoding) as defined by the 802.3 standard. In reduced signal count mode (RGMII or RTBI), the MAC remains configured in GMII or TBI but the TSEC multiplexes and decodes the input signals and provides the MAC with the expected interface. TSEC provides the TSECn_GTX_CLK to the PHY in either GMII or TBI mode of operation. Address recognition options The options supported are promiscuous, broadcast, exact individual address hash or exact match, and multicast hash match. For detailed descriptions refer to Section 14.5.3.8.1, “Individual Address Registers 0–7 (IADDRn),” Section 14.5.3.8.2, “Group Address Registers 0–7 (GADDRn),” Section 14.5.3.4.1, “Receive Control Register (RCTRL),” and Section 14.6.2.6, “Frame Recognition.” RMON support See Section 14.6.2.5, “RMON Support.” Internal loop back Internal loop back mode is selected through the Loop Back bit in the MACCFG1 register. See Section 14.6.2.10, “Internal and External Loop Back,” for details. 14.4 External Signal Descriptions This section defines the TSEC signals. The buses are described using the bus convention used in IEEE 802.3 because the PHY follows this same convention (that is, TxD[7:0] means 0 is the lsb). Notice that except for external physical interfaces the buses and registers follow a big-endian format. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-9 Three-Speed Ethernet Controllers The TSEC network interface supports multiple options: • The MII option requires 18 I/O signals (including the EC_MDIO and EC_MDC MII management interface) and supports both a data and a management interface to the PHY (transceiver) device. The MII option supports both 10- and 100-Mbps Ethernet rates. The GMII option is a superset of the MII signals and supports a 1-Gbps Ethernet rate. The TBI interface shares signals with the GMII interface signals. Finally, RGMII and RTBI options are reduced-signal implementations of the GMII and TBI interfaces. • • • 14.4.1 Detailed Signal Descriptions Table 14-1 contains detailed descriptions of the TSEC interface signals. For RGMII mode details please refer to the Hewlett-Packard reduced gigabit media-independent interface (RGMII) specification version 1.2a, dated 9/22/2000. All other modes follow the IEEE 802.3 standard, 2000 Edition. Input signals not used are internally disabled. Output signals not used are driven low. Table 14-1. TSEC Signals—Detailed Signal Descriptions Signal TSECn_COL I/O I Description Collision input. The behavior of this signal is not specified while in full-duplex mode. State Asserted/Negated—In MII mode, this signal is asserted upon detection of a collision, Meaning and must remain asserted while the collision persists. This signal is not used in the TSEC GMII, TBI, RTBI and RGMII modes. Timing Asserted/Negated—This signal is not required to transition synchronously with TSECn_TX_CLK or TSECn_RX_CLK. TSECn_CRS I Carrier sense input. In TBI and RTBI modes this signal can be used as SDET (signal detect), an optional signal that some PHYs generate in TBI or RTBI modes. This signal is not used in the TSEC GMII or RGMII modes. State Asserted/Negated—In MII mode, TSECn_TX_CLK is asserted while the transmit or Meaning receive medium is not idle. In the event of a collision, TSEC n_CRS must remain asserted for the duration of the collision. Timing Asserted/Negated—This signal is not required to transition synchronously with TSECn_TX_CLK or TSECn_RX_CLK. TSECn_GTX_CLK O Gigabit transmit clock. This signal is an output from the TSEC into the PHY. In GMII, TBI, or RTBI mode TSEC n_GTX_CLK is a 125-MHz clock that provides a timing reference for TX_EN, TXD, and TX_ER. In RGMII mode TSECn_GTX_CLK becomes the transmit clock and provides timing reference during 1000Base-T (125-MHz), 100Base-T (25-MHz) and 10Base-T (2.5-MHz) transmissions. This signal is not used in MII mode. Gigabit transmit 125-MHz source. This signal must be generated externally with a crystal or oscillator, or is sometimes provided by the PHY. In GMII, RGMII, RTBI or TBI mode, EC_GTX_CLK125 is a 125-MHz input into the TSEC and is used to generate all 125-MHz related signals and clocks. This input is not used in MII mode. EC_GTX_CLK125 I MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-10 Freescale Semiconductor External Signal Descriptions Table 14-1. TSEC Signals—Detailed Signal Descriptions (continued) Signal EC_MDC I/O O Description Management data clock. In GMII, MII, RGMII, RTBI or TBI mode this signal is a clock (typically 2.5 MHz) supplied by the MAC (IEEE–set minimum period of 400 ns or a frequency of 2.5 MHz, but the device may be configured up to 12.5 MHz if supported by the PHY at that speed). This clock is generated by dividing the core complex bus (CCB) clock by eight. The ratio can be modified further by writing to MIIMCFG[29:31]. Note that this signal is used during reset to configure the TSEC interface to ‘reduced-signal’ or ‘non-reduced-signal’ mode. See Section 4.4.3, “Power-On Reset Configuration,” for more information. EC_MDIO I/O Management data input/output, BIDI State Asserted/Negated—In GMII, MII, RGMII, RTBI or TBI mode EC_MDIO is a Meaning bidirectional signal to input PHY-supplied status during management read cycles and output control during MII management write cycles. Timing Asserted/Negated—This signal is required to be synchronous with the EC_MDC signal. TSECn_RX_CLK I Receive clock. In GMII, MII or RGMII mode, the receive clock TSEC n_RX_CLK is a continuous clock (2.5, 25, or 125 MHz) that provides a timing reference for TSECn_RX_DV, TSEC n_RXD, and TSEC n_RX_ER. In TBI mode, TSECn_RX_CLK is the input for a 62.5-MHz PMA receive clock, 0 split phase with PMA_RX_CLK1 and is supplied by the SerDes. In RTBI mode it is a 125-MHz receive clock. Receive data valid. In GMII or MII mode, if TSECn_RX_DV is asserted, the PHY is indicating that valid data is present on the GMII and MII interfaces. In RGMII mode, TSECn_RX_DV becomes RX_CTL. The RX_DV and RX_ERR are received on this signal on the rising and falling edges of TSECn_RX_CLK. In TBI mode, TSEC n_RX_DV represents receive code group (RCG) bit 8. Together, with RCG[9] and RCG[7:0], they represents the 10-bit encoded symbol of GMII receive signals. In RTBI mode, TSECn_RX_DV represents receive code group (RCG) bit 4 and 9. On the positive edge of the TSECn_RX_CLK, RCG[4] and RCG[3:0] represents the first half of the 10-bit encoded symbol. On the negative edge of the TSECn_RX_CLK, RCG[9] and RCG[8:4] represents the second half of the 10-bit encoded symbol. Receive data in. In GMII mode, TSECn_RXD[7:4] with TSECn_RXD[3:0], represent one complete octet of data to be transferred from the PHY to the MAC when TSEC n_RX_DV is asserted. In TBI mode, TSEC n_RXD[7:4] represents RCG[7:4]. Together, with RCG[9:8] and RCG[3:0], they represent the 10-bit encoded symbol of GMII receive signals. In GMII mode, TSECn_RXD[3:0] represents a nibble of data to be transferred from the PHY to the MAC when TSECn_RX_DV is asserted. A completely-formed SFD must be passed across the MII. While TSECn_RX_DV is not asserted, TSECn_RXD has no meaning. In RGMII or RTBI mode, TSEC n_RXD[3:0] are received on the rising edge of TSECn_RX_CLK and TSEC n_RXD[7:4] are received on the falling edge of TSECn_RX_CLK. In TBI mode, TSECn_RXD[3:0] represents RCG[3:0]. Together, with RCG[9:4], they represent the 10-bit encoded symbol of GMII receive signals. TSECn_RX_DV I TSECn_RXD[7:0] I MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-11 Three-Speed Ethernet Controllers Table 14-1. TSEC Signals—Detailed Signal Descriptions (continued) Signal TSECn_RX_ER I/O I Receive error State Asserted/Negated—In GMII or MII mode, if TSECn_RX_ER and TSECn_RX_DV Meaning are asserted, the PHY has detected an error in the current frame. In TBI mode, TSECn_RX_ER represents RCG[9]. Together, with RCG[8:0], they represent the 10-bit encoded symbol of GMII receive signals. This signal is not used in the TSEC RTBI or RGMII modes. TSECn_TX_CLK I Transmit clock in. In MII mode, TSEC n_TX_CLK is a continuous clock (2.5 or 25 MHz) that provides a timing reference for the TSECn_TX_EN, TSECn_TXD, and TSECn_TX_ER signals. In GMII mode, TSEC n_TX_CLK provides the 2.5 or 25-MHz timing reference during 10Base-T and 100Base-T and comes from the PHY. In 1000Base-T this clock is not used and TSEC n_GTX_CLK (125 MHz) becomes the timing reference. The TSECn_GTX_CLK is generated in the TSEC and provided to the PHY and the MAC. The TSECn_TX_CLK is generated in the PHY and provided to the MAC. In TBI mode, this signal is PMA receive clock 1 62.5 MHz, split phase with PMA_RX_CLK0, and is supplied by the SerDes. This signal is not used in the TSEC RTBI or RGMII modes. Transmit data out. In GMII mode, TSECn_TXD[7:4], together with TSECn_TXD[3:0], represent one complete octet of data to be sent from the MAC to the PHY when TSEC_TX_DV is asserted and has no meaning while TSECn_TX_EN is negated. In TBI mode, TSEC n_TXD[7:4] represents transmit code group (TCG) bits 7:4. Together, with TCG[9:8] and TCG[3:0], they represent the 10-bit encoded symbol. In GMII or MII mode, TSEC n_TXD[3:0] represent a nibble of data to be sent from the MAC to the PHY when TSECn_TX_EN is asserted and have no meaning while TSECn_TX_EN is negated. In RGMII or RTBI mode, TSECn_TXD[3:0] are transmitted on the rising edge of TSEC n_TX_CLK, and TSECn_TXD[7:4] are transmitted on the falling edge of TSEC n_TX_CLK. In TBI mode, TSEC n_TXD[3:0] represents TCG[3:0]. Together, with TCG[9:4], they represent the 10-bit encoded symbol. Note that some of these signals are also used during reset to configure the TSEC interface in GMII or TBI mode. Additionally, some of these signals are used for other reset configuration settings for the MPC8540. See Section 4.4.3, “Power-On Reset Configuration,” for more information. Transmit data valid. In GMII or MII mode, if TSECn_TX_EN is asserted, the MAC is indicating that valid data is present on the GMII’s or the MII’s TSECn_TXD signals. In RGMII mode, TSECn_TX_EN becomes TX_CTL. TX_EN and TX_ERR are asserted on this signal on rising and falling edges of the TSECn_TX_CLK, respectively. In TBI mode, TSEC n_TX_EN represents TCG[8]. Together, with TCG[9] and TCG[7:0], they represent the 10-bit encoded symbol. In RTBI mode, TSECn_TX_EN represents TCG[4]. Together with TCG[9], TCG[3:0] and TCG[8:5], they represent the 10-bit encoded symbol. Transmit error. In GMII or MII mode, assertion of TSEC n_TX_ER for one or more clock cycles while TSECn_TX_EN is asserted causes the PHY to transmit one or more illegal symbols. Asserting TSECn_TX_ER has no effect while operating at 10 Mbps or while TSECn_TX_EN is negated. This signal transitions synchronously with respect to TSECn_TX_CLK. In TBI mode, TSECn_TX_ER represents TCG[9]. Together, with TCG[8:0], they represents the 10-bit encoded symbol. This signal is not used in the TSEC RTBI or RGMII modes. Description TSECn_TXD[7:0] O TSECn_TX_EN O TSECn_TX_ER O MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-12 Freescale Semiconductor Memory Map/Register Definition 14.5 Memory Map/Register Definition The TSEC uses a software model similar to that employed by the Fast Ethernet function supported on the Freescale MPC8260 CPM FCC and in the FEC of the MPC860T. The TSEC device is programmed by a combination of control/status registers (CSR) and buffer descriptors. The CSRs are used for mode control, interrupts, and to access status information. The descriptors are used to pass data buffers and related buffer status or frame information between the hardware and software. All accesses to and from the registers must be made with 32-bit accesses. There is no support for accesses of sizes other than 32 bits. This section of the document defines the memory map and describes the registers in detail. The buffer descriptor is described in Section 14.6.3, “Buffer Descriptors.” The ten-bit interface (TBI) module MII registers are also described in this section. The TBI registers are defined like PHY registers and, as such, are accessed via the MII management interface in the same way as the PHYs are accessed. For detailed descriptions of the TBI registers (the MII register set for the ten-bit interface) please refer to Section 14.5.4, “Ten-Bit Interface (TBI).” 14.5.1 Top-Level Module Memory Map The TSEC implementation requires 4 Kbytes of memory-mapped space, of which more than 1 Kbyte is reserved for future expansion. The space is divided into the following sections: • • • • • • General control/status registers Transmit-specific control/status registers Receive-specific control/status registers MAC registers Event/statistic counters held in the MIB block Hash function registers Table 14-2 defines the top-level memory map. Table 14-2. Module Memory Map Summary Address 000–0FF 100–2FF 300–4FF 500–5FF Function TSEC general control/status registers TSEC transmit control/status registers TSEC receive control/status registers TSEC MAC registers MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-13 Three-Speed Ethernet Controllers Table 14-2. Module Memory Map Summary (continued) 600–7FF 800–8FF 900–AFF B00–BFF C00–FFF TSEC RMON MIB registers TSEC HASH function registers Reserved TSEC attribute registers Future expansion space 14.5.2 Detailed Memory Map—Control/Status Registers Table 14-3 lists the address, name, and a cross-reference to the complete description of each register. The offsets to the memory map table are defined for both TSECs. That is, TSEC1 starts at 0x2_4000 address offset and TSEC2 starts at 0x2_5000 address offset. The registers for TSEC1 are listed in Table 14-3, but the registers for TSEC2 are not. Note in Table 14-3 that the registers are the same for TSEC2 except that the offset changes from 0x2_4nnn to 0x2_5nnn. Table 14-3. Module Memory Map Offset Name Access1 Reset Section/Page TSEC1 General Control and Status Registers 0x2_4000– 0x2_400C 0x2_4010 0x2_4014 0x2_4018 0x2_401C 0x2_4020 0x2_4024 0x2_4028 0x2_402C 0x2_4030 0x2_4034– 0x2_4048 Reserved IEVENT—Interrupt event register IMASK—Interrupt mask register EDIS—Error disabled register Reserved ECNTRL—Ethernet control register MINFLR—Minimum frame length register PTV—Pause time value register DMACTRL—DMA control register TBIPA—TBI PHY address register Reserved R R/W R/W R/W R R/W R/W R/W R/W R/W R 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0040 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 — 14.5.3.1.1/14-20 14.5.3.1.2/14-23 14.5.3.1.3/14-24 — 14.5.3.1.4/14-25 14.5.3.1.5/14-26 14.5.3.1.6/14-27 14.5.3.1.7/14-28 14.5.3.1.8/14-29 — TSEC1 FIFO Control and Status Registers 0x2_404C 0x2_4050– 0x2_4088 0x2_408C FIFO_PAUSE_CTRL—FIFO pause control register Reserved FIFO_TX_THR—FIFO transmit threshold register R/W R R/W 0x0000_0000 0x0000_0000 0x0000_0100 14.5.3.2.1/14-30 — 14.5.3.2.2/14-31 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-14 Freescale Semiconductor Memory Map/Register Definition Table 14-3. Module Memory Map (continued) Offset 0x2_4090– 0x2_4094 0x2_4098 0x2_409C 0x2_40A0– 0x2_40FC Reserved FIFO_TX_STARVE—FIFO transmit starve register FIFO_TX_STARVE_SHUTOFF—FIFO transmit starve shutoff register Reserved Name Access1 R R/W R/W R Reset 0x0000_0000 0x0000_0080 0x0000_0100 0x0000_0000 Section/Page — 14.5.3.2.3/14-32 14.5.3.2.4/14-32 — TSEC1 Transmit Control and Status Registers 0x2_4100 0x2_4104 0x2_4108 0x2_410C 0x2_4110 0x2_4114– 0x2_4120 0x2_4124 0x2_4128– 0x2_4180 0x2_4184 0x2_4188– 0x2_4200 0x2_4204 0x2_4208– 0x2_42AC 0x2_42B0 0x2_42B4 0x2_42B8– 0x2_42FC TCTRL—Transmit control register TSTAT—Transmit status register Reserved TBDLEN—TxBD data length register TXIC—Transmit interrupt coalescing configuration register Reserved CTBPTR—Current TxBD pointer register Reserved TBPTR—TxBD pointer register Reserved TBASE—TxBD base address register Reserved OSTBD—Out-of-sequence TxBD register OSTBDP—Out-of-sequence Tx data buffer pointer register Reserved R/W R/W R R R/W R R R R/W R R/W R R/W R/W R 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0800_0000 0x0000_0000 0x0000_0000 14.5.3.3.1/14-33 14.5.3.3.2/14-34 — 14.5.3.3.3/14-35 14.5.3.3.4/14-35 — 14.5.3.3.5/14-36 — 14.5.3.3.6/14-37 — 14.5.3.3.7/14-37 — 14.5.3.3.8/14-38 14.5.3.3.9/14-40 — TSEC1 Receive Control and Status Registers 0x2_4300 0x2_4304 0x2_4308 0x2_430C RCTRL—Receive control register RSTAT—Receive status register Reserved RBDLEN—RxBD data length register R/W R/W R R 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 14.5.3.4.1/14-41 14.5.3.4.2/14-41 — 14.5.3.4.3/14-42 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-15 Three-Speed Ethernet Controllers Table 14-3. Module Memory Map (continued) Offset 0x2_4310 0x2_4314– 0x2_4320 0x2_4324 0x2_4328– 0x2_433C 0x2_4340 0x2_4344– 0x2_4380 0x2_4384 0x2_4388– 0x2_4400 0x2_4404 0x2_4408– 0x2_44FC Name RXIC—Receive interrupt coalescing configuration register Reserved CRBPTR—Current RxBD pointer register Reserved MRBLR—Maximum receive buffer length register Reserved RBPTR—RxBD pointer register Reserved RBASE—RxBD base address register Reserved Access1 R/W R R R R/W R R/W R R/W R Reset 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 Section/Page 14.5.3.4.4/14-43 — 14.5.3.4.5/14-44 — 14.5.3.4.6/14-44 — 14.5.3.4.7/14-45 — 14.5.3.4.8/14-45 — TSEC1 MAC Registers 0x2_4500 0x2_4504 0x2_4508 0x2_450C 0x2_4510 0x2_4514– 0x2_451C 0x2_4520 0x2_4524 0x2_4528 0x2_452C 0x2_4530 0x2_4534 0x2_4538 0x2_453C 0x2_4540 MACCFG1—MAC configuration register 1 MACCFG2—MAC configuration register 2 IPGIFG—Inter-packet gap/inter-frame gap register HAFDUP—Half-duplex register MAXFRM—Maximum frame length register Reserved MIIMCFG—MII management configuration register MIIMCOM—MII Management command register MIIMADD—MII Management address register MIIMCON—MII Management control register MIIMSTAT—MII Management status register MIIMIND—MII Management indicator register Reserved IFSTAT—Interface status register MACSTNADDR1—Station address register, part 1 R/W R/W R/W R/W R/W R R/W R/W R/W W R R R R/W R/W 0x0000_0000 0x0000_7000 0x4060_5060 0x00A1_F037 0x0000_0600 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 14.5.3.6.1/14-49 14.5.3.6.2/14-51 14.5.3.6.3/14-52 14.5.3.6.4/14-53 14.5.3.6.5/14-54 — 14.5.3.6.6/14-54 14.5.3.6.7/14-55 14.5.3.6.8/14-56 14.5.3.6.9/14-57 14.5.3.6.10/14-57 14.5.3.6.11/14-58 — 14.5.3.6.12/14-58 14.5.3.6.13/14-59 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-16 Freescale Semiconductor Memory Map/Register Definition Table 14-3. Module Memory Map (continued) Offset 0x2_4544 0x2_4548– 0x2_467C Name MACSTNADDR2—Station address register, part 2 Reserved Access1 R/W R Reset 0x0000_0000 0x0000_0000 Section/Page 14.5.3.6.14/14-60 — TSEC1 RMON MIB Registers TSEC1 Transmit and Receive Counters 0x2_4680 0x2_4684 0x2_4688 0x2_468C 0x2_4690 0x2_4694 0x2_4698 TR64—Transmit and receive 64-byte frame counter register TR127—Transmit and receive 65- to 127-byte frame counter register TR255—Transmit and receive 128- to 255-byte frame counter register TR511—Transmit and receive 256- to 511-byte frame counter register TR1K—Transmit and receive 512- to 1023-byte frame counter register TRMAX—Transmit and receive 1024- to 1518-byte frame counter register TRMGV—Transmit and receive 1519- to 1522-byte good VLAN frame count register TSEC1 Receive Counters 0x2_469C 0x2_46A0 0x2_46A4 0x2_46A8 0x2_46AC 0x2_46B0 0x2_46B4 0x2_46B8 0x2_46BC 0x2_46C0 0x2_46C4 0x2_46C8 0x2_46CC 0x2_46D0 0x2_46D4 RBYT—Receive byte counter register RPKT—Receive packet counter register RFCS—Receive FCS error counter register RMCA—Receive multicast packet counter register RBCA—Receive broadcast packet counter register RXCF—Receive control frame packet counter register RXPF—Receive PAUSE frame packet counter register RXUO—Receive unknown OP code counter register RALN—Receive alignment error counter register RFLR—Receive frame length error counter register RCDE—Receive code error counter register RCSE—Receive carrier sense error counter register RUND—Receive undersize packet counter register ROVR—Receive oversize packet counter register RFRG—Receive fragments counter register R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 14.5.3.7.8/14-64 14.5.3.7.9/14-65 14.5.3.7.10/14-65 14.5.3.7.11/14-66 14.5.3.7.12/14-66 14.5.3.7.13/14-67 14.5.3.7.14/14-67 14.5.3.7.15/14-68 14.5.3.7.16/14-68 14.5.3.7.17/14-69 14.5.3.7.18/14-69 14.5.3.7.19/14-70 14.5.3.7.20/14-70 14.5.3.7.21/14-71 14.5.3.7.22/14-71 R/W R/W R/W R/W R/W R/W R/W 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 14.5.3.7.1/14-61 14.5.3.7.2/14-61 14.5.3.7.3/14-62 14.5.3.7.4/14-62 14.5.3.7.5/14-63 14.5.3.7.6/14-63 14.5.3.7.7/14-64 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-17 Three-Speed Ethernet Controllers Table 14-3. Module Memory Map (continued) Offset 0x2_46D8 0x2_46DC Name RJBR—Receive jabber counter register RDRP—Receive drop register TSEC1 Transmit Counters 0x2_46E0 0x2_46E4 0x2_46E8 0x2_46EC 0x2_46F0 0x2_46F4 0x2_46F8 0x2_46FC 0x2_4700 0x2_4704 0x2_4708 0x2_470C 0x2_4710 0x2_4714 0x2_4718 0x2_471C 0x2_4720 0x2_4724 0x2_4728 0x2_472C TBYT—Transmit byte counter register TPKT—Transmit packet counter register TMCA—Transmit multicast packet counter register TBCA—Transmit broadcast packet counter register TXPF—Transmit PAUSE control frame counter register TDFR—Transmit deferral packet counter register TEDF—Transmit excessive deferral packet counter register TSCL—Transmit single collision packet counter register TMCL—Transmit multiple collision packet counter register TLCL—Transmit late collision packet counter register TXCL—Transmit excessive collision packet counter register TNCL—Transmit total collision counter register Reserved TDRP—Transmit drop frame counter register TJBR—Transmit jabber frame counter register TFCS—Transmit FCS error counter register TXCF—Transmit control frame counter register TOVR—Transmit oversize frame counter register TUND—Transmit undersize frame counter register TFRG—Transmit fragments frame counter register TSEC1 General Registers 0x2_4730 0x2_4734 0x2_4738 0x2_473C 0x2_4740– 0x2_47FC CAR1—Carry register one CAR2—Carry register two CAM1—Carry register one mask register CAM2—Carry register two mask register Reserved R/W R/W R/W R/W R 0x0000_0000 0x0000_0000 14.5.3.7.44/14-82 14.5.3.7.45/14-84 R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R R/W R/W R/W R/W R/W R/W R/W 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 14.5.3.7.25/14-73 14.5.3.7.26/14-73 14.5.3.7.27/14-74 14.5.3.7.28/14-74 14.5.3.7.29/14-75 14.5.3.7.30/14-75 14.5.3.7.31/14-76 14.5.3.7.32/14-76 14.5.3.7.33/14-77 14.5.3.7.34/14-77 14.5.3.7.35/14-78 14.5.3.7.36/14-78 — 14.5.3.7.37/14-79 14.5.3.7.38/14-79 14.5.3.7.39/14-80 14.5.3.7.40/14-80 14.5.3.7.41/14-81 14.5.3.7.42/14-81 14.5.3.7.43/14-82 Access1 R/W R/W Reset 0x0000_0000 0x0000_0000 Section/Page 14.5.3.7.23/14-72 14.5.3.7.24/14-72 0xFE01_FFFF 14.5.3.7.46/14-85 0x000F_FFFF 14.5.3.7.47/14-86 0x0000_0000 — MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-18 Freescale Semiconductor Memory Map/Register Definition Table 14-3. Module Memory Map (continued) Offset Name Access1 Reset Section/Page TSEC1 Hash Function Registers 0x2_4800 0x2_4804 0x2_4808 0x2_480C 0x2_4810 0x2_4814 0x2_4818 0x2_481C 0x2_4820– 0x2_487C 0x2_4880 0x2_4884 0x2_4888 0x2_488C 0x2_4890 0x2_4894 0x2_4898 0x2_489C 0x2_48A0– 0x2_4AFC IADDR0—Individual address register 0 IADDR1—Individual address register 1 IADDR2—Individual address register 2 IADDR3—Individual address register 3 IADDR4—Individual address register 4 IADDR5—Individual address register 5 IADDR6—Individual address register 6 IADDR7—Individual address register 7 Reserved GADDR0—Group address register 0 GADDR1—Group address register 1 GADDR2—Group address register 2 GADDR3—Group address register 3 GADDR4—Group address register 4 GADDR5—Group address register 5 GADDR6—Group address register 6 GADDR7—Group address register 7 Reserved R/W R/W R/W R/W R/W R/W R/W R/W R R/W R/W R/W R/W R/W R/W R/W R/W R 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 — — 14.5.3.8.2/14-88 14.5.3.8.1/14-88 TSEC1 Attribute Registers 0x2_4B00– 0x2_4BF4 0x2_4BF8 0x2_4BFC Reserved ATTR—Attribute register ATTRELI—Attribute EL & EI register TSEC1 Future Expansion Space 0x2_4C00– 0x2_4FFC Reserved R 0x0000_0000 — R R R 0x0000_0000 0x0000_0000 0x0000_0000 — 14.5.3.9.1/14-89 14.5.3.9.2/14-90 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-19 Three-Speed Ethernet Controllers Table 14-3. Module Memory Map (continued) Offset Name TSEC2 Registers 0x2_5000– 0x2_5FFC 1 2 Access1 Reset Section/Page TSEC2 registers2 R = means read-only, W = write only, R/W = read and write, LH = latches high, SC = self-clearing. TSEC2 has the same memory-mapped registers that are described for TSEC1 from 0x 2_4000 to 0x2_4FFF except the offsets are from 0x 2_5000 to 0x2_5FFF. 14.5.3 Memory-Mapped Register Descriptions This section provides a detailed description of all the TSEC registers. Because all of the TSEC registers are 32 bits wide, only 32-bit register accesses are supported. 14.5.3.1 TSEC General Control and Status Registers This section describes general control and status registers used for both transmitting and receiving Ethernet frames. All of the registers are 32 bits wide. 14.5.3.1.1 Interrupt Event Register (IEVENT) If an event occurs that sets a bit in the interrupt event (IEVENT) register, shown in Figure 14-6, an interrupt is generated if the corresponding bit in the interrupt enable register (IMASK) is also set. Clearing the IEVENT bit clears the interrupt signal. The bit in the IEVENT register is cleared if a 1 is written to that bit position. A write of 0 has no effect. These interrupts can be divided into operational interrupts, transceiver/network error interrupts, and internal error interrupts. Interrupts that may occur in normal operation are: • • • GTSC, GRSC, TXF, TXB, TXC, RXF, RXB, RXC, and MSRO BABR, BABT, LC, and CRL/XDA EBERR, XFUN, and BSY Interrupts resulting from errors/problems detected in the network or transceiver are: Interrupts resulting from internal errors are: Some of the error interrupts are independently counted in the management information base (MIB) block counters. Software may choose to mask off these interrupts because these errors are visible to network management through the MIB counters. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-20 Freescale Semiconductor Memory Map/Register Definition 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 R BABR RXC W Reset 16 17 0 BSY EBERR MSRO GTSC BABT TXC TXE TXB TXF 0 LC CRL/ XFUN XDA 0000_0000_0000_0000 22 23 24 25 31 R RXB W Reset Offset 0 0 0 0 0 0 GRSC RXF 0 0 0 0 0 0 0 0000_0000_0000_0000 TSEC1:0x2_4010; TSEC2:0x2_5010 Figure 14-6. IEVENT Register Definition Table 14-4 describes the fields of the IEVENT register. Table 14-4. IEVENT Field Descriptions Bits 0 Name BABR Description Babbling receive error. This bit indicates that a frame was received with length in excess of the MAC’s maximum frame length register while MACCFG2[Huge Frame] is set. 0 Excessive frame not received 1 Excessive frame received Receive control interrupt. A control frame was received. If MACCFG1[Rx_Flow] is set, a pause operation is performed lasting for the duration specified in the received pause control frame and beginning when the frame was received. 0 Control frame not received 1 Control frame received Busy condition interrupt. Indicates that a frame was received and discarded due to a lack of buffers. When IEVENT[BSY] is set RSTAT[QHLT] is also set. In order to begin receiving packets again, the user must clear RSTAT[QHLT]. This bit and RSTAT[QHLT] are set whenever the TSEC reads an RxBD with its Empty field cleared. 0 No frame received and discarded 1 Frame received and discarded Ethernet bus error. This bit indicates that a system bus error for a memory read occurred while a DMA transaction was underway. If the EBERR is set while transmission is in progress, the DMA stops sending data to the Tx FIFO which eventually causes an underrun error (XFUN) and TSTAT[THLT] is set. If the EBERR is set while receiving a frame, the DMA discards the frame and RSTAT[QHLT] is set. 0 No system bus error occurred 1 System bus error occurred Reserved MSTAT Register Overflow. This interrupt is asserted if the count for one of the MSTAT registers has exceeded the size of the register. 0 MSTAT count not exceeding register size 1 MSTAT count exceeds register size 1 RXC 2 BSY 3 EBERR 4 5 — MSRO MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-21 Three-Speed Ethernet Controllers Table 14-4. IEVENT Field Descriptions (continued) Bits 6 Name GTSC Description Graceful transmit stop complete. This interrupt is asserted following the completion of a graceful stop, which was intitiated by setting either DMACTRL[GTS] or TCTRL[TFC_PAUSE]. Graceful stop means that the transmitter is put into a pause state after completion of the frame currently being transmitted. 0 Graceful transmit stop not complete or not initiated 1 Graceful transmit stop completed Babbling transmit error. This bit indicates that the transmitted frame length has exceeded the value in the MAC’s Maximum Frame Length register and MACCFG2[Huge Frame] is cleared. Frame truncation occurs when this condition occurs. TxBD[TXTRUNC] is set in the last TxBD (TxBD[L] is set) of the frame. 0 Transmitted frame length not exceeding maximum frame length 1 Transmitted frame length exceeding maximum frame length Transmit control interrupt. This bit indicates that a control frame was transmitted. 0 Control frame not transmitted 1 Control frame transmitted Transmit error. This bit indicates that an error occurred on the transmitted channel that has caused TSTAT[THLT] to be set by TSEC. This bit is set whenever any transmit error occurs that causes the transmitter to halt (EBERR, LC, CRL/XDA, XFUN). It is not set if DMACTRL[WOP] is set and TSEC runs out of TxBDs to process. In order to begin transmitting packets again, the user must clear TSTAT[THLT]. 0 No transmit channel error occurred 1 Transmit channel error occurred Transmit buffer. This bit indicates that a transmit buffer descriptor was updated whose I (interrupt) bit was set in its status word and was not the last buffer descriptor of the frame. 0 No transmit buffer descriptor updated 1 Transmit buffer descriptor updated Transmit frame interrupt. This bit indicates that a frame was transmitted and that the last corresponding transmit buffer descriptor (TxBD) was updated. This only occurs if the I (interrupt) bit in the status word of the buffer descriptor is set. 0 No frame transmitted/TxBD not updated 1 Frame transmitted/TxBD updated Reserved Late collision. This bit indicates that a collision occurred beyond the collision window (slot time) in half-duplex mode. The frame is truncated with a bad CRC and the remainder of the frame is discarded. 0 No late collision occurred 1 Late collision occurred Collision retry limit/excessive defer abort. This bit indicates either one of two conditions occurred while attempting to transmit a frame: 1) the number of successive transmission collisions has exceeded the MAC’s HAFDUP[Retransmission Maximum] count or 2) an excessive defer abort condition has occurred. An excessive defer abort condition occurs when the TSEC waits more than 3036 bytes while attempting to send a frame and HAFDUP[EXCESS_DEFER] is 0. The TxBD[DEF] or OSTBD[DEF] is also set. In either case the frame is discarded without being transmitted and the TSEC is halted (TSTAT[THLT] is set). The CRL or XDA condition can only occur while in half-duplex mode. 0 Successive transmission collisions do not exceed maximum and no excessive defer abort condition has occurred. 1 Successive transmission collisions exceed maximum or an excessive defer abort condition has occurred. 7 BABT 8 TXC 9 TXE 10 TXB 11 TXF 12 13 — LC 14 CRL/ XDA MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-22 Freescale Semiconductor Memory Map/Register Definition Table 14-4. IEVENT Field Descriptions (continued) Bits 15 Name XFUN Description Transmit FIFO underrun. This bit indicates that the transmit FIFO became empty before the complete frame was transmitted. 0 Transmit FIFO not underrun 1 Transmit FIFO underrun Receive buffer. These bits indicate that a receive buffer descriptor was updated which had the I (Interrupt) bit set in its status word and was not the last buffer descriptor of the frame. 0 Receive buffer descriptor not updated 1 Receiver buffer descriptor updated Reserved Graceful receive stop complete. This interrupt is asserted if a graceful receive stop is completed. It allows the user to know if the system has completed the stop and it is safe to write to receive registers (status, control or configuration registers) that are used by the system during normal operation. 0 Graceful stop not completed 1 Graceful stop completed Receive frame interrupt. The last receive buffer descriptor (RxBD) of a frame was updated. This occurs only if the I (interrupt) bit in the buffer descriptor status word is set. 0 Frame not received 1 Frame received Reserved 16 RXB 17–22 23 — GRSC 24 RXF 25–31 — 14.5.3.1.2 Interrupt Mask Register (IMASK) The interrupt mask register provides control over which possible interrupt events are allowed to generate an actual interrupt. All implemented bits in this CSR are R/W. This register is cleared upon a hardware reset. If the corresponding bits in both the IEVENT and IMASK registers are set, an interrupt is generated. The interrupt signal can be cleared by clearing the corresponding IEVENT bit. Figure 14-7 shows the IMASK register. 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 R BREN RXCEN BSYEN EBERREN 0 MSROEN GTSCEN BTEN TXCEN TXEEN TXBEN TXFEN 0 LCEN W Reset 16 17 CRL/ XFUNEN XDAEN 0000_0000_0000_0000 22 23 24 25 31 R RXBEN 0 0 0 0 0 0 GRSCEN RXFEN 0 0 0 0 0 0 0 W Reset Offset 0000_0000_0000_0000 TSEC1:0x2_4014; TSEC2:0x2_5014 Figure 14-7. IMASK Register Definition MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-23 Three-Speed Ethernet Controllers Table 14-5 describes the fields of the IMASK register. Table 14-5. IMASK Field Descriptions Bits 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17–22 23 24 25–31 Name BREN RXCEN BSYEN EBERREN — MSROEN GTSCEN BTEN TXCEN TXEEN TXBEN TXFEN — LCEN CRL/XDAEN XFUNEN RXBEN — GRSCEN RXFEN — Babbling receiver interrupt enable Receive control interrupt enable Busy interrupt enable Ethernet controller bus error enable Reserved MSTAT register overflow interrupt enable Graceful transmit stop complete interrupt enable Babbling transmitter interrupt enable Transmit control interrupt enable Transmit error interrupt enable Transmit buffer interrupt enable Transmit frame interrupt enable Reserved Late collision enable Collision retry limit/excessive defer enable Transmit FIFO underrun enable Receive buffer interrupt enable Reserved Graceful receive stop complete interrupt enable Receive frame interrupt enable Reserved Description 14.5.3.1.3 Error Disabled Register (EDIS) The error disabled register, shown in Figure 14-8, controls error reporting by the TSEC. The IEVENT bit corresponding to an error will not be set if the error is disabled in EDIS. 01 2 3 4 12 13 14 31 R00 BSYDIS EBERRDIS W Reset Offset 000000 0 0 0 LCDIS 0 0 0 00000000000000 0 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_4018; TSEC2:0x2_5018 Figure 14-8. Error Disabled Register (EDIS) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-24 Freescale Semiconductor Memory Map/Register Definition Table 14-6 describes the fields of the EDIS register. Table 14-6. EDIS Field Descriptions Bits 0–1 2 Name — BSYDIS Reserved Busy disable 0 Allow TSEC to report IEVENT[BSY] status and halt buffer descriptor queue if BSY condition occurs. 1 Do not set IEVENT[BSY] and do not halt buffer descriptor queue if BSY condition occurs. Ethernet controller bus error disable 0 Allow TSEC to report IEVENT[EBERR] status and halt buffer descriptor queue if EBERR condition occurs. 1 Do not set IEVENT[EBERR] and do not halt buffer descriptor queue if EBERR condition occurs. Reserved Late collision disable 0 Allow TSEC to report IEVENT[LC] status, set the buffer descriptor LC field, and halt buffer descriptor queue if LC condition occurs. 1 Do not set IEVENT[LC] nor the buffer Descriptor LC field, and do not halt buffer descriptor queue if LC condition occurs. Reserved Description 3 EBERRDIS 4–12 13 — LCDIS 14–31 — 14.5.3.1.4 Ethernet Control Register (ECNTRL) ECNTRL, shown in Figure 14-9, is used to reset, configure, and initialize the TSEC. 0 16 17 18 19 20 25 26 27 28 29 31 R0000000000000000 0 CLRCNT AUTOZ STEN W Reset Offset 0 0 0 0 0 0 TBIM RPM R100M 000 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_4020; TSEC2:0x2_5020 Figure 14-9. ECNTRL Register Definition Table 14-7 describes the fields of the ECNTRL register. Table 14-7. ECNTRL Field Descriptions Bits 0–16 17 Name — Reserved Description CLRCNT Clear all statistics counters 0 Allow MSTAT counters to continue to increment. 1 Reset all MSTAT counters. This bit is self-resetting. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-25 Three-Speed Ethernet Controllers Table 14-7. ECNTRL Field Descriptions (continued) Bits 18 Name AUTOZ Description Automatically zero addressed statistical counter values, input to MSTAT module. 0 The user must write the addressed counter zero after a host read. 1 The addressed counter value is zeroed on a host read. This is a steady state signal and must be set prior to enabling the Ethernet controller and must not be changed without proper care. Statistics enabled, input to MSTAT module. 0 Statistics not enabled 1 Enables internal counters to update This is a steady state signal and must be set prior to enabling the Ethernet controller and must not be changed without proper care. Reserved Ten-bit interface mode. If this bit is set, ten-bit interface mode is enabled. This bit can be signal configured at reset to set or clear using the TSECn_GTX_CLK signal. 0 GMII or MII mode interface 1 TBI mode interface Reduced signal mode. If this bit is set, a reduced signal interface is expected. ECNTRL[TBI Mode] selects GMII or TBI. If RPM is selected, the MAC must be in GMII or TBI mode (RMII is not supported). This register can be signal configured at reset to 0 or 1 with the EC_MDC signal. 0 GMII or TBI in non-reduced signal mode configuration 1 RGMII or RTBI reduced signal mode RGMII 100 mode. This bit is ignored unless RPM is set and MACCFG2[I/F Mode] is assigned to 10/100 (01). If this bit is set, the TSEC interface is in 100 Mbps speed. 0 RGMII is in 10 Mbps mode 1 RGMII is in 100 Mbps mode Reserved 19 STEN 20–25 26 — TBIM 27 RPM 28 R100M 29–31 — 14.5.3.1.5 Minimum Frame Length Register (MINFLR) MINFLR is written by the user. It tells TSEC the smallest packet to accept and place in a receive buffer pointed to by the RxBD. Figure 14-10 shows the MINFLR register. 0 24 25 31 R0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 MINFLR W Reset Offset 0000_0000_0000_0000_0000_0000_0100_0000 TSEC1:0x2_4024; TSEC2:0x2_502 Figure 14-10. MINFLR Register Definition MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-26 Freescale Semiconductor Memory Map/Register Definition Table 14-8 describes the fields of the MINFLR register. Table 14-8. MINFLR Field Descriptions Bits 0–24 25–31 Name — Reserved Description MINFLR Minimum receive frame length (typically 64 decimal). If the Ethernet receives an incoming frame shorter than MINFLR, it discards that frame unless RCTRL[RSF] (receive short frames) is set, in which case RxBD[SH] (frame too short) is set in the last RxBD. The largest allowable value for MINFLR is 64. Unlike the MPC8260, in which PADs are added to make the transmit frame equal to MINFLR bytes, if padding is requested, TSEC always PADS transmit frames to 64 bytes ignoring MINFLR. MINFLR is only used to determine the minimum size of acceptable receive frames. 14.5.3.1.6 Pause Time Value Register (PTV) PTV, shown in Figure 14-11, is written by the user to store the pause duration used when the TSEC initiates a pause frame via TCTRL[TFC_PAUSE]. The low-order 16 bits (PT) represent the pause time and the high-order 16 bits (PTE) represent the extended pause control parameter. The pause time is measured in units of pause_quanta, equal to 512 bit times. The pause time can range from 0 to 65,535 pause_quanta, or 0 to 33,553,920 bit times. See Section 14.6.2.7, “Flow Control for additional details. 0 15 16 31 R PTE W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_4028; TSEC2:0x2_5028 PT Figure 14-11. PTV Register Definition Table 14-9 describes the fields of the PTV register. Table 14-9. PTV Field Descriptions Bits 0–15 Name PTE Description Extended pause control. This field allows software to add a 16-bit additional control parameter into the pause frame to be sent when TCTRL[TFC_PAUSE] is set. Note that current IEEE 802.3 pause frame format requires this parameter to be set to 0. Pause time value. Represents the 16-bit pause quanta (that is, 512 bit times). This pause value is used as part of the pause frame to be sent when TCTRL[TFC_PAUSE] is set. See Section 14.6.2.7, “Flow Control,” for more information. 16–31 PT MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-27 Three-Speed Ethernet Controllers 14.5.3.1.7 DMA Control Register (DMACTRL) DMACTRL, shown in Figure 14-12, is used to configure the DMA block. register. 0 23 24 25 26 27 28 29 30 31 R 0 000000000000000000000 0 0 TDSEN TBDSEN W Reset Offset 0 GRS GTS 0 WWR WOP 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_402C; TSEC2:0x2_502C Figure 14-12. DMACTRL Register Definition Table 14-10 describes the fields of the DMACTRL register. Table 14-10. DMACTRL Field Descriptions Bits 0–23 24 Name — TDSEN Reserved Tx Data snoop enable 0 Disables snooping of all transmit frames from memory. 1 Enables snooping of all transmit frames from memory. Description 25 TBDSEN TxBD snoop enable 0 Disables snooping of all transmit BD memory accesses. 1 Enables snooping of all transmit BD memory accesses. — GRS Reserved Graceful receive stop. If this bit is set, the Ethernet controller stops receiving frames following completion of the frame currently being received, that is, after a valid end of frame was received. The buffer of the receive frame associated with the EOF is closed and the IEVENT[GRSC] is set. Because the MAC’s receive enable bit (MACCFG[Rx_EN]) may still be set, the MAC may continue to receive but the TSEC ignores the receive data until GRS is cleared. If this bit is cleared, the TSEC scans the input data stream for the start of a new frame (preamble sequence and start of frame delimiter) and the first valid frame received uses the next RxBD. If GRS is set, the user must monitor the graceful receive stop complete (GRSC) bit in the IEVENT register to insure that the graceful receive stop was completed. The user can then clear IEVENT[GRSC] and can write to receive registers that are accessible to both user and the TSEC hardware without fear of conflict. 0 TSEC scans input data stream for valid frame. 1 TSEC stops receiving frames following completion of current frame. Graceful transmit stop. If this bit is set, the Ethernet controller stops transmission after the completion of any frame that is currently being transmitted and the GTSC interrupt in the IEVENT register is asserted. If frame transmission is not currently underway, the GTSC interrupt is asserted immediately. Once transmission has completed, a “restart” can be accomplished by clearing GTS. 0 Controller continues. 1 Controller stops transmission after completion of current frame. Reserved 26 27 28 GTS 29 — MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-28 Freescale Semiconductor Memory Map/Register Definition Table 14-10. DMACTRL Field Descriptions (continued) Bits 30 Name WWR Description Write with response. This bit gives the user the assurance that a BD was updated in memory before it receives an interrupt concerning a transmit or receive frame. 0 Do not wait for acknowledgement from system for BD writes before setting IEVENT bits. 1 Before setting IEVENT bits TXB, TXF, TXE, IE, XFUN, LC, CRL/XDA, RXB, RXF, the TSEC waits for acknowledgement from system that the transmit or receive BD being updated was stored in memory. Wait or poll. This bit, which is applicable only to the transmitter, provides the user the option for the TSEC to periodically poll TxBD or to wait for software to tell TSEC to fetch a buffer descriptor. While operating in the “Wait” mode, the TSEC allows two additional reads of a descriptor which is not ready before entering a halt state. No interrupt is driven. (IEVENT[TXE] is clear.) To resume transmission, software must clear TSTAT[THLT]. Note that if this bit is set, the user must ensure that all TxBDs involved with sending a frame have their ready bits set before any transmission begins. Otherwise, TSEC behaves in a boundedly undefined fashion. A buffer descriptor is considered not ready when its TxBD[Ready] field is clear or its TxBD[Data Length] field is zero. It is considered ready when both TxBD[Ready] is set and TxBD[Data Length] is non-zero. In Wait mode (DMACTRL[WOP] is set) when TSEC is processing a frame and an intermediate TxBD’s ready bit is cleared, TSEC does not halt, but instead continuously polls the same TxBD until the TxBD becomes ready or an Ethernet interface error or a memory error is encountered. CxBD becomes ready, TSEC reports the error in both the IEVENT register as well as the TxBD for Ethernet interface errors, or the IEVENT[EBERR] for a memory error and sets the TSTAT[THLT] bit. Note that software must eventually set all of its TxBDs for a frame, because TSEC continuously reads an intermediate TxBD until it becomes ready if insufficient data has been read to surpass the FIFO Transmit Threshold (FIFO_TX_THR) register value. 0 Poll TxBD every 512 clocks. 1 Do not poll, but wait for a write to TSTAT[THLT]. 31 WOP 14.5.3.1.8 TBI Physical Address Register (TBIPA) This TBIPA, shown in Figure 14-13, is writable by the user to assign a physical address to the TBI for MII management configuration. The TBI registers are accessed at the offset of TBIPA. For detailed descriptions of the TBI registers (the MII register set for the ten-bit interface) please refer to Section 14.5.4, “Ten-Bit Interface (TBI).” 0 26 27 31 R0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 TBIPA W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_4030; TSEC2:0x2_5030 Figure 14-13. TBIPA Register Definition MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-29 Three-Speed Ethernet Controllers Table 14-11 describes the fields of the TBIPA register. Table 14-11. TBIPA Field Descriptions Bits 0–26 27–31 Name — Reserved Description TBIPA TBI PHY address. This field is used to program the PHY address of the ten-bit interface’s MII management bus. To access the TBI register the user must write the TBIPA value to the MIIMADD [PHY Address] register located in the MAC register section. Refer to Section 14.5.3.6.8, “MII Management Address Register (MIIMADD).” 14.5.3.2 TSEC FIFO Control and Status Registers The following registers allow the user to change some of the default settings in the FIFO that can be used to optimize operation for performance or for safety. They must be set carefully in order to avoid an underrun condition. Underrun is an error condition in which data is not retrieved from external memory quickly enough, leaving the TX FIFO empty before the complete frame is transmitted. Because different combinations of events, several of which are determined by the user, can lead to underrun, the TSEC provides several FIFO registers that allow the user to select the proper setting to be able to tune the system and obtain the maximum performance with minimal chance of underrun. The principal causes for underrun in the TSEC are: • • • Misaligned data buffer addresses Small data buffer sizes Combinations of the above It is recommended that the minimum size data buffers be 64 bytes and that data buffers be 64-byte aligned. The user can deviate from these recommended values to try to increase performance or to use less memory, but unless the default values of some of the FIFO registers are adjusted, the probability of an underrun may also increase. The FIFO_TX_THR (default is 256 entries or 1 Kbytes) indicates the amount of data required to be in the FIFO before starting the transmission of a frame. The FIFO_TX_STARVE (default is 128 entries or 512 bytes) is used to indicate that the amount of data in the FIFO is so low that the risk of underrun is extremely high. The FIFO_TX_STARVE_SHUTOFF (default is 256 entries or 1 Kbyte) contains the watermark level to be used for exiting the starve state. These registers are intended to allow the user to make the proper trade-off. If triggered, the starve mode, for instance, automatically raises the priority of TSEC fetches from memory. 14.5.3.2.1 FIFO Pause Control Register (FIFO_PAUSE_CTRL) FIFO_PAUSE_CTRL, shown in Figure 14-14, is writable by the user to configure the properties of the TSEC FIFO. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-30 Freescale Semiconductor Memory Map/Register Definition 0 15 R W Reset 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0000_0000_0000_0000 16 29 30 31 R W Reset Offset 0 0 0 0 0 0 0 0 0 0 0 0 0 0 TFC_PAUSE_EN 0 0000_0000_0000_0000 TSEC1:0x2_404C; TSEC2:0x2_504C Figure 14-14. FIFO_PAUSE_CTRL Register Definition Table 14-12 describes the fields of the FIFO_PAUSE_CTRL register. Table 14-12. FIFO_PAUSE_CTRL Field Descriptions Bits 0–29 30 Name — Reserved Description TFC_PAUSE_EN TFC_PAUSE enable. This bit enables the ability to transmit a pause control frame by setting the TCTRL[TFC_PAUSE] bit. This bit is cleared at reset. 0 Pause control frame transmission disabled. 1 Pause control frame transmission enabled. — Reserved 31 14.5.3.2.2 FIFO Transmit Threshold Register (FIFO_TX_THR) The main purpose of the threshold register is to trigger the unloading of FIFO data to the PHY. It represents the numerical SRAM entry (0-511 for 2-Kbyte FIFO) to trigger the threshold function. If the number of valid entries in the FIFO is equal to or greater than the threshold register, transmission can begin. This register is read/write by software and is initialized to 0000_0000_0000_0000_0000_0001_0000_0000 at system reset. Figure 14-15 shows the FIFO_TX_THR register. 0 22 23 31 R0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 FIFO_TX_THR W Reset Offset 0000_0000_0000_0000_0000_0001_0000_0000 TSEC1:0x2_408C; TSEC2:0x2_508C Figure 14-15. FIFO_TX_THR Register Definition MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-31 Three-Speed Ethernet Controllers Table 14-13 describes the fields of the FIFO_TX_THR register. Table 14-13. FIFO_TX_THR Field Descriptions Bits 0–22 23–31 Name — Reserved Description FIFO_TX_THR FIFO transmit threshold. These bits mark the number of entries in the transmit FIFO that, if reached, trigger the unloading of frame data into the MAC. 14.5.3.2.3 FIFO Transmit Starve Register (FIFO_TX_STARVE) The purpose of the starve register, shown in Figure 14-16, is to inform the system of extremely imminent underrun conditions. It represents the numerical SRAM entry (0-511 for 2-Kbyte FIFO) to trigger the starve function. If the number of valid entries in the FIFO is less than or equal to the starve register, a starve alert is triggered. 0 22 23 31 R0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 FIFO_TX_STARVE W Reset Offset 0000_0000_0000_0000_0000_0000_1000_0000 TSEC1:0x2_4098; TSEC2:0x2_5098 Figure 14-16. FIFO_TX_STARVE Register Definition Table 14-14 describes the fields of the FIFO_TX_STARVE register. Table 14-14. FIFO_TX_STARVE Field Descriptions Bits 0–22 Name — Reserved Description 23–31 FIFO_TX_STARVE FIFO transmit starve. These bits indicate the value to trigger the transmit starve function. It triggers once the number of valid entries in the FIFO is less than or equal to the FIFO Tx starve. The starve state turns off if the number of valid entries in the FIFO becomes greater than or equal to the FIFO Tx starve shutoff register. 14.5.3.2.4 FIFO Transmit Starve Shutoff Register (FIFO_TX_STARVE_SHUTOFF) The starve shutoff register, shown in Figure 14-17, contains the watermark level to be used for exiting the starve state. If the starve state is in effect and the number of valid entries in the FIFO becomes greater than or equal to the value in the FIFO transmit starve shutoff register, the starve condition ends. This register is read/write by software. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-32 Freescale Semiconductor Memory Map/Register Definition 0 22 23 31 R0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 FIFO_TX_STARVE_SHUTOFF W Reset Offset 0000_0000_0000_0000_0000_0001_0000_0000 TSEC1:0x2_409C; TSEC2:0x2_509C Figure 14-17. FIFO_TX_STARVE_SHUTOFF Register Definition Table 14-15 describes the fields of the FIFO transmit starve shutoff register. Table 14-15. FIFO_TX_STARVE_SHUTOFF Field Descriptions Bits 0–22 23–31 Name — Reserved Description FIFO_TX_STARVE_ FIFO transmit starve shutoff. Indicates the value beyond which to exit the starve state. SHUTOFF The starve state turns off if the number of valid entries in the FIFO becomes greater than or equal to the FIFO Tx starve shutoff register. 14.5.3.3 TSEC Transmit Control and Status Registers This section describes the control and status registers that are used specifically for transmitting Ethernet frames. All of the registers are 32 bits wide. 14.5.3.3.1 Transmit Control Register (TCTRL) TCTRL, shown in Figure 14-18, is writable by the user to configure the transmit block. 0 19 20 21 26 27 28 29 31 R000000000000000 0 0 00 0 THDF W Reset Offset 0 0 0 0 0 0 RFC_PAUSE TFC_PAUSE 000 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_4100; TSEC2:0x2_5100 Figure 14-18. TCTRL Register Definition Table 14-16 describes the fields of the TCTRL register. Table 14-16. TCTRL Field Descriptions Bits 0–19 20 Name — THDF Reserved Transmit half-duplex flow control. Written by user. This bit is not self-resetting. 0 Disable back pressure. 1 Back pressure is applied to media. Description MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-33 Three-Speed Ethernet Controllers Table 14-16. TCTRL Field Descriptions (continued) Bits 21–26 27 Name — Reserved Description RFC_PAUSE Receive flow control pause frame. Written by TSEC. This read-only status bit is set if a flow control pause frame was received and the transmitter is paused for the duration defined in the received pause frame. This bit automatically clears after the pause duration is complete. 0 Pause duration complete. 1 Pause duration in progress. TFC_PAUSE Transmit flow control pause frame. Use this bit to transmit a PAUSE frame. To transmit a flow control pause frame, first set FIFO_PAUSE_CTRL[TFC_PAUSE_EN]. Next, set MACCFG1[GTS]. If TFC_PAUSE is then set, the MAC stops transmission of data frames after the current transmission completes. The GTSC interrupt in the IEVENT register is asserted. With transmission of data frames stopped, the MAC transmits a MAC control PAUSE frame with the duration value obtained from the PTV register. The TXC interrupt occurs after sending the control pause frame. Next, the MAC clears TFC_PAUSE and resumes transmitting data frames. Note that if the transmitter is paused due to user assertion of GTS or reception of a PAUSE frame, the MAC may still transmit a MAC control PAUSE frame. 0 No outstanding pause frame transmission request. 1 Pause frame transmission requested. — Reserved 28 29–31 14.5.3.3.2 Transmit Status Register (TSTAT) TSTAT, shown in Figure 14-19, is written by TSEC to convey DMA status information. 0 1 31 R THLT W Reset Offset 0000000000000000000000000000000 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_4104; TSEC2:0x2_5104 Figure 14-19. TSTAT Register Definition Table 14-17 describes the fields of the TSTAT register. Table 14-17. TSTAT Field Descriptions Bits 0 Name THLT Description Transmit halt 0 No hardware-initiated transmission halt. 1 TSEC transmission function halted. This bit is written by TSEC to inform the user that it is no longer processing transmit frames and that the transmit DMA function is disabled by hardware. To re-start the transmission function, the user must clear this bit by writing a one. Reserved 1–31 — MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-34 Freescale Semiconductor Memory Map/Register Definition 14.5.3.3.3 TxBD Data Length Register (TBDLEN) TBDLEN is a DMA register that contains the number of bytes remaining in the current transmit buffer. Figure 14-20 shows the TBDLEN register. 0 15 16 31 R0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 W Reset Offset TBDLEN 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_410C; TSEC2:0x2_510C Figure 14-20. TBDLEN Register Definition Table 14-18 describes the fields of the TBDLEN register. Table 14-18. TBDLEN Field Descriptions Bits 0–15 16–31 Name — TBDLEN Reserved Internally written by the DMA module. The transmit channel remains active until TBDLEN is 0. Description 14.5.3.3.4 Transmit Interrupt Coalescing Configuration Register (TXIC) TXIC, shown in Figure 14-21, enables and configures the operational parameters for interrupt coalescing associated with transmitted frames. Refer to Section 14.6.2.8.1, “Interrupt Coalescing,” for a functional description of interrupt coalescing. 0 1 2 3 10 11 15 16 31 R ICEN W Reset Offset 00 ICFCT 00000 ICTT 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_4110; TSEC2:0x2_5110 Figure 14-21. TXIC Register Definition MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-35 Three-Speed Ethernet Controllers Table 14-18 describes the fields of the TXIC register. Table 14-19. TXIC Field Descriptions Bits 0 Name ICEN Description Interrupt coalescing enable 0 Interrupt coalescing is disabled. Interrupts are raised as they are received if the TSEC transmit frame interrupt is enabled (IMASK[TXFEN] is set). 1 Interrupt coalescing is enabled. If the TSEC transmit frame interrupt is enabled (IMASK[TXFEN] is set), an interrupt is raised when the threshold number of frames is reached (defined by TXIC[ICFCT]) or when the threshold timer expires (defined by TXIC[ICTT]). Reserved Interrupt coalescing frame count threshold. While interrupt coalescing is enabled (TXIC[ICEN] is set), this value determines how many frames are transmitted before raising an interrupt1. Valid values for this field are from 1 to 255. A value of 0 is illegal. If set to 0, an interrupt can only be cleared by first clearing TXIC[ICEN] and then clearing the IEVENT[TXF] bit. Note that a value of 1 functionally defeats the advantages of interrupt coalescing since the frame threshold is reached with each frame transmitted. Reserved Interrupt coalescing timer threshold. While interrupt coalescing is enabled (TXIC[ICEN] is set), this value determines the maximum amount of time after transmitting a frame before raising an interrupt 1, subject also to IMASK[TXFEN]. If frames have been transmitted but the frame count threshold has not been met, an interrupt is raised when the threshold timer expires. The threshold timer is reset once an interrupt has been asserted. It begins counting once the interrupt is cleared and IEVENT[TXF] is set. The threshold value is represented in units equal to 64 TSEC interface clocks. Valid values for this field are from 1 to 65535. A value of 0 is illegal. If set to 0, an interrupt can only be cleared by first clearing TXIC[ICEN] and then clearing the IEVENT[TXF] bit. 1–2 3–10 — ICFCT 11–15 16–31 — ICTT 1 Interrupts resulting from the Interrupt bit (I) of the buffer descriptor in question and enabled subject to the IMASK register (IMASK[TXFEN] is set). 14.5.3.3.5 Current Transmit Buffer Descriptor Pointer Register (CTBPTR) CTBPTR contains the address of the transmit buffer descriptor either currently being processed, or processed most recently. Figure 14-22 shows the CTBPTR register. 0 28 29 31 R W Reset Offset CTBPTR 000 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_4124; TSEC2:0x2_5124 Figure 14-22. CTBPTR Register Definition MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-36 Freescale Semiconductor Memory Map/Register Definition Table 14-20 describes the fields of the CTBPTR register. Table 14-20. CTBPTR Field Descriptions Bits 0–28 29–31 Name Description CTBPTR The CTBPTR register is internally written by the DMA module. The value of this field increments by one (causing the register value to increment by eight) each time a descriptor is read from memory. — Reserved 14.5.3.3.6 Transmit Buffer Descriptor Pointer Register (TBPTR) TBPTR, shown in Figure 14-23, contains the low-order 32 bits of the next transmit buffer descriptor address. TBPTR takes on the TBASE value when TBASE is written by software. Although not necessary in most applications, the user can modify this register when the transmitter has been gracefully stopped or halted, as indicated by IEVENT[GTSC] or TSTAT[THLT]. 0 28 29 31 R TBPTR W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x2_4184; TSEC2:0x2_5184 000 Figure 14-23. TBPTR Register Definition Table 14-21 describes the fields of the TBPTR register. Table 14-21. TBPTR Field Descriptions Bits 0–28 29–31 Name TBPTR — Description The TBPTR register is internally written by the DMA module. The value of this field increments by one (causing the register value to increment by eight) each time a descriptor is read from memory. Reserved 14.5.3.3.7 Transmit Descriptor Base Address Register (TBASE) The TBASE register, shown in Figure 14-24, is written by the user with the TxBD base address. The value must be divisible by eight for the 8-byte data buffer descriptors. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-37 Three-Speed Ethernet Controllers 0 28 29 31 R TBASE W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x2_4204; TSEC2:0x2_5204 000 Figure 14-24. TBASE Register Definition Table 14-22 describes the fields of the TBASE register. Table 14-22. TBASE Field Descriptions Bits 0–28 Name Description TBASE Transmit base. TBASE defines the starting location in the memory map for the TSEC TxBDs. This field must be 8-byte aligned. Together with setting the W (wrap) bit in the last BD, the user can select how many BDs to allocate for the transmit packets. The user must initialize TBASE before enabling the TSEC transmit function. — Reserved 29–31 14.5.3.3.8 Out-of-Sequence TxBD Register (OSTBD) The out-of-sequence TxBD register, OSTBD, shown in Figure 14-25, includes the status/control and data length in the same format as a regular TxBD. It is useful for sending flow control frames. OSTBD[R] is always checked between frames. If it is not ready, a regular frame is sent. If a flow control frame is sent and OSTBD[I] is set, a TXC event is generated after frame transmission. This area must be cleared while not in use. Once the TSEC is in paused mode the out-of-sequence buffer descriptor cannot be used to send another flow control frame because the MAC regards it as a regular TxBD. 0 1 2 34 5 6 7 8 9 10 13 14 15 16 31 R R PAD/CRC W I W Reset Offset L TC DEF TO1 HFE/LC RL RC UN 0 OSTBDLEN 0000_1000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_42B0; TSEC2:0x2_52B0 Figure 14-25. OSTBD Register Definition MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-38 Freescale Semiconductor Memory Map/Register Definition Table 14-23 describes the fields of the OSTBD register. Table 14-23. OSTBD Field Descriptions Bits 0 Name R Description Ready. Written by TSEC and user. 0 The data buffer associated with this BD is not ready for transmission. The user is free to manipulate this BD or its associated data buffer. The TSEC clears this bit after the buffer is transmitted or after an error condition is encountered. 1 The data buffer, which was prepared for transmission by the user, was not transmitted or is currently being transmitted. No fields of this BD may be written by the user once this bit is set. Padding and CRC attachment for short frames. (Valid only when MACCFG2[PAD/CRC] is cleared, and MACCFG2[CRC EN] bit is cleared.) If MACCFG2[PAD/CRC] is set, pads are added to all short frames; however, this bit is ignored. 0 Do not add PADs to short frames unless OSTBD[TC] is set. 1 Add PADs to short frames. PAD bytes are inserted until the length of the transmitted frame equals 64 bytes. Unlike the MPC8260 which PADs up to MINFLR value, TSEC always PADs up to the IEEE minimum frame length of 64 bytes. Wrap. Written by user. This bit is ignored by TSEC. Interrupt. Written by user. 0 No interrupt is generated after this buffer is serviced. 1 IEVENT[TXF] is set after this buffer is serviced. This bit can cause an interrupt if IMASK[TXFEN] is enabled. Last in Frame. The OSTBD is always the last in the frame, so L is always set. (Hardwired to a value of 1.) Tx CRC. Written by user. (Valid only while it is set in the first BD and OSTBD[PAD/CRC] is cleared, MACCFG2[PAD/CRC] is cleared, and MACCFG2[CRC EN] is cleared.) If MACCFG2[PAD/CRC] is set or MACCFG2[CRC EN] is set, a CRC is added to all frames and this bit is ignored. 0 End transmission immediately after the last data byte, unless OSTBD[PAD/CRC] is set. 1 Transmit the CRC sequence after the last data byte. Defer indication. Written by TSEC. Hardware updates this bit after transmitting a frame if used as a ‘Defer indicator.’ Software/user updates this bit while building a transmit buffer descriptor if used as a ‘Hardware Event indicator.’ 0 This frame was not deferred. 1 This frame did not have a collision before it was sent but it was sent late because of deferring. Transmit software ownership. This read/write bit may be utilized by software, as necessary. Its state does not affect the hardware nor is it affected by the hardware. Huge frame enable (written by user)/Late collision (written by TSEC) Huge frame enable. Written by user. Valid only while it is set in first BD and the MACCFG2[Huge Frame] is cleared. If MACCFG2[Huge Frame] is set, this bit is ignored. 0 Truncate transmit frame if its length is greater than the MAC’s Maximum Frame Length register. 1 Do not truncate the transmit frame. Late collision. Written by TSEC. 0 No late collision. 1 A collision occurred after 64 bytes are sent. The TSEC terminates the transmission and updates LC. Retransmission limit. Written by TSEC. 0 Transmission before maximum retry limit is hit. 1 The transmitter failed (max. retry limit + 1) attempts to successfully send a message due to repeated collisions. The TSEC terminates the transmission and updates RL. 1 PAD/CRC 2 3 W I 4 5 L TC 6 DEF 7 8 TO1 HFE/LC 9 RL MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-39 Three-Speed Ethernet Controllers Table 14-23. OSTBD Field Descriptions (continued) Bits 10–13 Name RC Description Retry count. Written by TSEC. 0 The frame is sent correctly the first time. 1 More than zero attempts were needed to send the transmit frame. If this field is 15, 15 or more retries were needed. The Ethernet controller updates RC after sending the buffer. Underrun. Written by TSEC. 0 No underrun encountered (data was retrieved from external memory in time to send a complete frame). 1 The Ethernet controller encountered a transmitter underrun condition while sending the associated buffer. The TSEC terminates the transmission and updates UN. Reserved Out-of-sequence TxBD data length. Written by user. Data length is the number of octets the TSEC transmits from this BD’s data buffer. It is never modified by the TSEC. This field must be greater than zero in order for a transfer to take place. 14 UN 15 16–31 — OSTBDLEN 14.5.3.3.9 Out-of-Sequence Tx Data Buffer Pointer Register (OSTBDP) The out-of-sequence Tx data buffer pointer register (OSTBDP), shown in Figure 14-26, contains the data buffer pointer fields in the same format as a regular TxBD. With OSTBD, it provides the complete 8-byte descriptor. This area must be cleared while not in use. 0 31 R OSTBDP W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_42B4; TSEC2:0x2_52B4 Figure 14-26. OSTBDP Register Definition Table 14-24 describes OSTBDP. Table 14-24. OSTBDP Field Descriptions Bits 0–31 Name Description OSTDBP Out-of-sequence Tx Data Buffer Pointer. Written by user. The transmit data buffer pointer contains the address of the associated data buffer. There are no alignment requirements for this address. 14.5.3.4 TSEC Receive Control and Status Registers This section describes the control and status registers that are used specifically for receiving Ethernet frames. All of the registers are 32 bits wide. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-40 Freescale Semiconductor Memory Map/Register Definition 14.5.3.4.1 Receive Control Register (RCTRL) RCTRL, shown in Figure 14-27, controls the operational mode of the receive block. It must be written only after a system reset (at initialization) or if DMACTRL[GRS] is cleared. 0 26 27 28 29 30 31 R000000000000000 0 W Reset Offset 00000000000 BC_REJ PROM RSF 00 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_4300; TSEC2:0x2_5300 Figure 14-27. RCTRL Register Definition Table 14-25 describes the fields of the RCTRL register. Table 14-25. RCTRL Field Descriptions Bits 0–26 27 Name — BC_REJ Reserved Broadcast frame reject. If this bit is set, frames with DA (destination address) = FFFF_FFFF_FFFF are rejected unless RCTRL[PROM] is set. If both BC_REJ and RCTRL[PROM] are set, then frames with broadcast DA are accepted and the M (MISS) bit is set in the receive BD. 0 Broadcast frame reject is disabled. 1 Broadcast frame reject is enabled. Promiscuous mode. When set, all frames except pause frames are accepted. 0 Promiscuous mode is disabled. 1 Promiscuous mode is enabled. Receive short frame mode. When set, enables the reception of frames shorter than MINFLR bytes. Note that in order for short frames to be received when RSF is set, a DA hit must occur. When RSF is cleared, all frames shorter than MINFLR are automatically rejected. 0 Receive short fram mode is disabled. 1 Receive short frame mode is enabled. Reserved Description 28 PROM 29 RSF 30–31 — 14.5.3.4.2 Receive Status Register (RSTAT) TSEC writes to RSTAT, shown in Figure 14-28, under the following conditions: • • The receiver runs out of descriptors. The receiver was halted because an error condition was encountered while receiving a frame. Software must clear the QHLT bit to take the TSEC receiver function out of a halt state. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-41 Three-Speed Ethernet Controllers 0 7 8 9 31 R00000000 QHLT W Reset Offset 0000000 000 00 000 000 00 000 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_4304; TSEC2:0x2_5304 Figure 14-28. RSTAT Register Definition Table 14-26 describes the fields of the RSTAT register. Table 14-26. RSTAT Field Descriptions Bits 0–7 8 Name — Reserved Description QHLT RxBD queue is halted. When IEVENT[BSY] or IEVENT[EBERR] is set during reception of a packet, RSTAT[QHLT] is also set. In order to begin receiving packets again, the user must clear RSTAT[QHLT]. This bit is set whenever the TSEC reads an RxBD with its Empty field cleared or encounters a system bus error while processing an RX packet. It is a hardware-initiated stop indication (DMA_CTRL[GRS] being set by the user does not cause this bit to be set.). The current frame and all other frames directed to the halted queue are discarded. A write with a value of 1 re-enables the queue for receiving. 0 RxBD queue is enabled for Ethernet reception. (That is, it is not halted.) 1 All Ethernet controller receive activity to RxBD queue is halted. — Reserved 9–31 14.5.3.4.3 RxBD Data Length Register (RBDLEN) RBDLEN is a DMA register that contains the number of bytes remaining in the current receive buffer. Figure 14-29 shows the RBDLEN register. 0 15 16 31 R0 0 0 0 0 0 0 0 0 0 0 0 0 0 W Reset Offset 00 RBDLEN 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_430C; TSEC2:0x2_530C Figure 14-29. RBDLEN Register Definition Table 14-27 describes the fields of the RBDLEN register. Table 14-27. RBDLEN Field Descriptions Bits 0–15 Name — Reserved Description 16–31 RBDLEN The RBDLEN is internally written by the DMA module. If RBDLEN is cleared, all activity in the receive channel stops. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-42 Freescale Semiconductor Memory Map/Register Definition 14.5.3.4.4 Receive Interrupt Coalescing Configuration Register (RXIC) The RXIC register enables and configures the operational parameters for interrupt coalescing associated with received frames. Refer to Section 14.6.2.8.1, “Interrupt Coalescing,” for a functional description of interrupt coalescing as additional details regarding the use of this register. Figure 14-30 shows the RXIC register. 0 1 2 3 10 11 15 16 31 R ICEN W Reset Offset 0 0 ICFCT 0 0 0 0 0 ICTT 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_4310; TSEC2:0x2_5310 Figure 14-30. RXIC Register Definition Table 14-18 describes the fields of the RXIC register. Table 14-28. RXIC Field Descriptions Bits 0 Name ICEN Description Interrupt coalescing enable 0 Interrupt coalescing is disabled. Interrupts are raised as they are received if the TSEC receive frame interrupt is enabled (IMASK[RXFEN] is set). 1 Interrupt coalescing is enabled. If the TSEC receive frame interrupt is enabled (IMASK[RXFEN] is set), an interrupt is raised when the threshold number of frames is reached (defined by RXIC[ICFCT]) or when the threshold timer expires (defined by RXIC[ICTT]). Reserved 1–2 3–10 — ICFCT Interrupt coalescing frame count threshold. While interrupt coalescing is enabled (RXIC[ICEN] is set), this value determines how many frames are received before raising an interrupt1. Valid values for this field are from 1 to 255. A value of 0 is illegal. If set to 0, an interrupt can only be cleared by first clearing RXIC[ICEN] and then clearing the IEVENT[RXF] bit. Note that a value of 1 functionally defeats the advantages of interrupt coalescing since the frame threshold is reached with each frame received. — ICTT Reserved Interrupt coalescing timer threshold. While interrupt coalescing is enabled (RXIC[ICEN] is set), this value determines the maximum amount of time after receiving a frame before raising an interrupt 1, subject also to IMASK[RXFEN]. If frames have been received but the frame count threshold has not been met, an interrupt is raised when the threshold timer expires. The threshold timer is reset once an interrupt has been asserted. It begins counting once the interrupt is cleared and IEVENT[RXF] is set. The threshold value is represented in units equal to 64 TSEC interface clocks. Valid values for this field are from 1 to 65535. A value of 0 is illegal. If set to 0, an interrupt can only be cleared by first clearing RXIC[ICEN] and then clearing the IEVENT[RXF] bit. 11–15 16–31 1 Interrupts resulting from the Interrupt bit (I) of the buffer descriptor in question and enabled subject to the IMASK register (IMASK[RXFEN] is set). MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-43 Three-Speed Ethernet Controllers 14.5.3.4.5 Current Receive Buffer Descriptor Pointer Register (CRBPTR) CRBPTR contains the address of the receive buffer descriptor either currently being processed, or processed most recently. Figure 14-31 shows the CRBPTR register. 0 28 29 31 R W Reset Offset CRBPTR 000 0000_0000_0000_0000_0000_0000_0000_0000 0x2_4324; TSEC2:0x2_5324 Figure 14-31. CRBPTR Register Definition Table 14-29 describes the fields of the CRBPTR register. \ Table 14-29. CRBPTR Field Descriptions Bits 0–28 29–31 Name CRBPTR — Description The CRBPTR register is internally written by the DMA module. The value of this field increments by one (causing the register value to increment by eight) each time a descriptor is read from memory. Reserved 14.5.3.4.6 Maximum Receive Buffer Length Register (MRBLR) The MRBL register is written by the user. It informs the TSEC how much space is in the receive buffer pointed to by the RxBD. Figure 14-32 shows the MRBLR. 0 15 16 25 26 31 R0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 MRBL W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_4340; TSEC2:0x2_5340 000000 Figure 14-32. MRBL Register Definition MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-44 Freescale Semiconductor Memory Map/Register Definition Table 14-30 describes the fields of the MRBL register. \\ Table 14-30. MRBLR Field Descriptions Bits 0–15 16–25 Name — Description To ensure that MRBL is a multiple of 64, these bits are reserved and must be cleared. MRBL Maximum receive buffer length. MRBL is the number of bytes that the TSEC receiver writes to the receive buffer. The MRBL register is written by the user with a multiple of 64 for all modes. TSEC can write fewer bytes to the buffer than the value set in MRBL if a condition such as an error or end-of-frame occurs, but it never exceeds the MRBL value; therefore, user-supplied buffers must be at least as large as the MRBL. — To ensure that MRBL is a multiple of 64, these bits are reserved and must be cleared. 26–31 14.5.3.4.7 Receive Buffer Descriptor Pointer Register (RBPTR) RBPTR contains the receive buffer descriptor address. Figure 14-33 shows the RBPTR register. This register takes on the value of RBASE when the RBASE register is written by software. Although not necessary in most applications, the user can modify this register when the transmitter has been gracefully stopped or halted, as indicated by TSTAT[THLT]. 0 28 29 31 R RBPTR W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_4384; TSEC2:0x2_5384 0 00 Figure 14-33. RBPTR Register Definition Table 14-31 describes the fields of the RBPTR register. Table 14-31. RBPTR Field Descriptions Bits 0–28 29–31 Name Description RBPTR The RBPTR register is internally written by the DMA module. The value of this field increments by one (causing the register value to increment by eight) each time a descriptor is read from memory. — Reserved 14.5.3.4.8 Receive Descriptor Base Address Register (RBASE) The RBASE register is written by the user with the RxBD base address and must be divisible by eight for the 8-byte buffer descriptors. Figure 14-34 shows the RBASE register. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-45 Three-Speed Ethernet Controllers 0 28 29 31 R RBASE W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_4404; TSEC2:0x2_5404 00 0 Figure 14-34. RBASE Register Definition Table 14-32 describes the fields of the RBASE register. \ Table 14-32. RBASE Field Descriptions Bits 0–28 29–31 Name Description RBASE Receive base. RBASE defines the starting location in the memory map for the TSEC RxBD. — Reserved. These bits must be set to zero, to cause the value of RBASE to be a multiple of eight. 14.5.3.5 MAC Functionality This section describes the MAC registers and provides a brief overview of the functionality that can be exercised through the use of these registers, particularly those that provide functionality not explicitly required by the IEEE 802.3 standard. All of the MAC registers are 32 bits wide. 14.5.3.5.1 Configuring the MAC MAC configuration registers 1 and 2 provide a way to configure the MAC in multiple ways: • • Adjusting the preamble length—The length of the preamble can be adjusted from the nominal seven bytes to some other (non-zero) value. Varying pad/CRC combinations—Three pad/CRC combinations are provided to handle a variety of system requirements. The most simple are frames that already have a valid FCS field. In this case, the CRC is checked and reported via the transmit statistics vector (TSV[51:0]). The other two options include appending a valid CRC, or padding and then appending a valid CRC, resulting in a minimum frame of 64 octets. In addition to the programmable register set, the pad/CRC behavior can be dynamically adjusted on a per-packet basis. 14.5.3.5.2 Controlling CSMA/CD The half-duplex register (HAFDUP) allows control over the carrier-sense multiple access/collision detection (CSMA/CD) logic of the TSEC. Half-duplex is only supported for 10 Mbps and 100 Mbps operation. Following the completion of the packet transmission the part begins timing the inter packet gap (IPG) as programmed in the back-to-back IPG configuration register. The system is now free to begin another frame transfer. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-46 Freescale Semiconductor Memory Map/Register Definition In full-duplex mode both the carrier sense (CRS) and collision (COL) indications from the PHY are ignored, but in half-duplex mode the TSEC defers to CRS and, following a carrier event, times the IPG using the non-back-to-back IPG configuration values that include support for the optional two-thirds/one-third CRS deferral process. This optional IPG mechanism enhances system robustness and ensures fair access to the medium. During the first two-thirds of the IPG, the IPG timer is cleared if CRS is sensed. During the final one-third of the IPG, CRS is ignored and the transmission begins once IPG is timed. The two-thirds/one-third ratio is the recommended value. 14.5.3.5.3 Handling Packet Collisions While transmitting a packet in half-duplex mode, the TSEC is sensitive to COL. If a collision occurs, it aborts the packet and outputs the 32-bit jam sequence. The jam sequence is comprised of several bits of the CRC, inverted to guarantee an invalid CRC upon reception. A signal is sent to the system indicating that a collision occurred and that the start of the frame is needed for retransmission. The TSEC then backs off of the medium for a time determined by the truncated binary exponential back-off (BEB) algorithm. Following this back-off time, the packet is retried. The back-off time can be skipped if configured via the half-duplex register. However, this is non-standard behavior and its use must be carefully applied. Should any one packet experience excessive collisions, the packet is aborted. The system must flush the frame and move to the next one in line. If the system requests to send a packet while the TSEC is deferring to a carrier, the TSEC simply waits until the end of the carrier event and the timing of IPG before it honors the request. If packet transmission attempts experience collisions, the TSEC outputs the jam sequence and waits some amount of time before retrying the packet. This amount of time is determined by a controlled randomization process called truncated binary exponential back-off. The amount of time is an integer number of slot times. The number of slot times to delay before the nth retransmission attempt is chosen as a uniformly-distributed random integer r in the range: 0 ≤ r ≤ 2k, where k = min(n,10). So, after the first collision, TSEC backs-off either 0 or 1 slot times. After the fifth collision, TSEC backs-off between 0 and 32 slot times. After the tenth collision, the maximum number of slot times to back-off is 1024. This can be adjusted through the half-duplex register. An alternate truncation point, such as 7 for instance, can be programmed. On average, the MAC is more aggressive after seven collisions than other stations on the network. 14.5.3.5.4 Controlling Packet Flow Packet flow can be dealt with in a number of ways within TSEC. A default retransmit attempt limit of 15 can be reduced using the half-duplex register. The slot time or collision window can be used to gate the retry window and possibly reduce the amount of transmit buffering within the system. The slot time for 10/100 Mbps is 512 bit times. Because the slot time begins at the beginning of MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-47 Three-Speed Ethernet Controllers the packet, the end occurs around the 56th byte of the frame data. Slot time in 1-Gbps mode is not supported. Full-duplex flow control is provided for in IEEE 802.3x. Currently the standard does not address flow control in half-duplex environments. Common in the industry, however, is the concept of back pressure. The TSEC implements the optional back pressure mechanism using the raise carrier method. If the system receive logic wishes to stop the reception of packets in a network-friendly way, transmit half-duplex flow control (THDF) is set (TCTRL[THDF]). If the medium is idle, the TSEC raises carrier by transmitting preamble. Other stations on the half-duplex network then defer to the carrier. In the event the preamble transmission happens to cause a collision, TSEC ensures the minimum 96-bit presence on the wire, then drops preamble and waits a back-off time depending on the value of the configuration bit, back pressure no back-off (half-duplex) [BPNB]. These transmitting-preamble-for-back pressure collisions are not counted. If BPNB is set, the TSEC waits an inter-packet gap before resuming the transmission of preamble following the collision and does not defer. If cleared, the TSEC adheres to the truncated BEB algorithm that allows the possibility of packets being received. This also can be detrimental in that packets can now experience excessive collisions, causing them to be dropped in the stations from which they originate. To reduce the likelihood of lost packets and packets leaking through the back pressure mechanism, BPNB must be set. The TSEC drops carrier (cease transmitting preamble) periodically to avoid excessive defer conditions in other stations on the shared network. If, while applying back pressure, the TSEC is requested to send a packet, it stops sending preamble, and waits one IPG before sending the packet. BPNB applies for any collision that occurs during the sending of this packet. TSEC does not defer while attempting to send packets while in back pressure. Again, back pressure is non-standard, yet it can be effective in reducing the flow of receive packets. 14.5.3.5.5 Controlling PHY Links Control and status to and from the PHY is provided via the two-wire MII management interface described in IEEE 802.3u. The MII management registers (MII management configuration, command, address, control, status, and indicator registers) are used to exercise this interface between a host processor and one or more PHY devices (including the TBI). External PHYs may only be configured using the MII management interface of TSEC1 since the interface signals (EC_MDC and EC_MDIO) are only driven by TSEC1. The TSEC MII’s registers provide the ability to perform continuous read cycles, called a scan cycle. If requested (by setting MIIMCOM[Scan Cycle]), the part performs repetitive read cycles of the PHY status register, for example. In this way, link characteristics may be monitored more efficiently. The different fields in the MII management indicator register (scan, not valid and busy) are used to indicate availability of each read of the scan cycle to the host from MIIMSTAT[PHY Status]. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-48 Freescale Semiconductor Memory Map/Register Definition Yet another parameter that can be modified through the MII registers is the length of the MII management interface preamble. After establishing that a PHY supports preamble suppression, the host may so configure the TSEC. While enabled, the length of MII management frames are reduced from 64 clocks to 32 clocks. This effectively doubles the efficiency of the interface. 14.5.3.6 MAC Registers The following are the MAC registers. 14.5.3.6.1 MAC Configuration Register 1 (MACCFG1) The MACCFG1 register is written by the user. Figure 14-35 shows the MACCFG1 register. 0 1 11 12 13 14 15 R Soft_Reset W Reset 16 000000 0 00 0 0 Reset Rx MC Reset Tx MC Reset Rx Fun Reset Tx Fun 0000_0000_0000_0000 22 23 24 25 26 27 28 29 30 31 R W Reset Offset 0 000000 Loop Back 00 Rx_Flow Tx_Flow Sync’d Rx EN Rx_EN Sync’d Tx EN Tx_EN 0000_0000_0000_0000 TSEC1:0x2_4500; TSEC2:0x2_5500 Figure 14-35. MACCFG1 Register Definition Table 14-33 describes the fields of the MACCFG1 register. \ Table 14-33. MACCFG1 Field Descriptions Bits 0 Name Soft_Reset Description Soft reset. This bit is cleared by default. See Section 14.6.2.2, “Soft Reset and Reconfiguring Procedure,” for more information on setting this bit. 0 Normal operation 1 Place all modules within the MAC in reset. Reserved 1–11 12 — Reset Rx MC Reset receive MAC control block. This bit is cleared by default. 0 Normal operation 1 Place the receive MAC control block in reset. This block detects control frames and contains the pause timers. Reset Tx MC Reset transmit MAC control block. This bit is cleared by default. 0 Normal operation 1 Place the PETMC transmit MAC control block in reset. This block multiplexes data and control frame transfers. It also responds to XOFF PAUSE control frames. 13 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-49 Three-Speed Ethernet Controllers Table 14-33. MACCFG1 Field Descriptions (continued) Bits 14 Name Description Reset Rx Fun Reset receive function block. This bit is cleared by default. 0 Normal operation 1 Place the receive function block in reset. This block performs the receive frame protocol. Reset Tx Fun Reset transmit function block. This bit is cleared by default. 0 Normal operation 1 Place the transmit function block in reset. This block performs the frame transmission protocol. — Loop Back Reserved Loop back. This bit is cleared by default. 0 Normal operation 1 Loop back the MAC transmit outputs to the MAC receive inputs. Reserved Receive flow. This bit is cleared by default. 0 The receive MAC control ignores PAUSE flow control frames. 1 The receive MAC control detects and acts on PAUSE flow control frames. Transmit flow. This bit is cleared by default. 0 The transmit MAC control may not send PAUSE flow control frames if requested by the system. 1 The transmit MAC control may send PAUSE flow control frames if requested by the system. 15 16–22 23 24–25 26 — Rx_Flow 27 Tx_Flow 28 Sync’d Rx EN Receive enable synchronized to the receive stream (Read-only) 0 Frame reception is not enabled. 1 Frame reception is enabled. Rx_EN Receive enable. This bit is cleared by default. If set, prior to clearing this bit, set DMACTRL[GRS] then confirm subsequent occurrence of the graceful receive stop interrupt (IEVENT[GRSC] is set). 0 The MAC may not receive frames from the PHY. 1 The MAC may receive frames from the PHY. 29 30 Sync’d Tx EN Transmit enable synchronized to the transmit stream (Read-only) 0 Frame transmission is not enabled. 1 Frame transmission is enabled. Tx_EN Transmit enable. This bit is cleared by default. If set, prior to clearing this bit, set DMACTRL[GTS] then confirm subsequent occurrence of the graceful transmit stop interrupt (IEVENT[GTSC] is set). 0 The MAC may not transmit frames from the system. 1 The MAC may transmit frames from the system. 31 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-50 Freescale Semiconductor Memory Map/Register Definition 14.5.3.6.2 MAC Configuration Register 2 (MACCFG2) The MACCFG2 register is written by the user. Figure 14-36 shows the MACCFG2 register. 0 15 16 19 20 21 22 23 24 25 26 27 28 29 30 31 R 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Preamble 0 0 I/F 0 0 Huge Length 0 PAD/ CRC Full CRC EN Duplex Frame check Mode Length W Reset Offset 0000_0000_0000_0000_0111_0000_0000_0000 TSEC1:0x2_4504; TSEC2:0x2_5504 Figure 14-36. MACCFG2 Register Definition Table 14-34 describes the fields of the MACCFG2 register. Table 14-34. MACCFG2 Field Descriptions Bits 0–15 16–19 20–21 22–23 Name — Preamble Length — I/F Mode Reserved Determines the length in bytes of the preamble field of the packet. Its default is 0x7. A preamble length of 0 is not supported. Reserved Determines the type of interface to which the MAC is connected. Its default is 00. 00 Reserved 01 Nibble mode (MII) 10 Byte mode (GMII/TBI) 11 Reserved Reserved Huge frame enable. Cleared by default. 0 Limit the length of frames to the MAXIMUM FRAME LENGTH value. 1 Frames longer than the MAXIMUM FRAME LENGTH may be transmitted and received. Description 24–25 26 — Huge Frame 27 Length check Length check. This bit is cleared by default. 0 No length field checking is performed. 1 The MAC checks the frame’s length field to ensure it matches the actual data field length. — PAD/CRC Reserved Pad and append CRC. This bit is cleared by default. 0 Frames presented to the MAC have a valid length and contain a CRC. 1 The MAC pads all transmitted short frames and appends a CRC to every frame regardless of padding requirement. CRC enable. If the configuration bit PAD/CRC or the per-packet PAD/CRC bit is set, CRC EN is ignored. This bit is cleared by default. 0 Frames presented to the MAC have a valid length and contain a valid CRC. 1 The MAC appends a CRC on all frames. Clear this bit if frames presented to the MAC have a valid length and contain a valid CRC. Full duplex configure. This bit is cleared by default. 0 The MAC to operate in half-duplex mode only. 1 The MAC operates in full-duplex mode. 28 29 30 CRC EN 31 Full Duplex MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-51 Three-Speed Ethernet Controllers 14.5.3.6.3 Inter-Packet Gap/Inter-Frame Gap Register (IPGIFG) The IPGIFG register is written by the user. Figure 14-37 shows the IPGIFG register. 0 1 7 8 9 15 16 23 24 25 31 R0 W Reset Offset Non-Back-to-Back Inter-Packet-Gap, Part 1 0 Non-Back-to-Back Inter-Packet-Gap, Part 2 Minimum IFG Enforcement 0 Back-to-Back Inter-Packet-Gap 0100_0000_0110_0000_0101_0000_0110_0000 TSEC1:0x2_4508; TSEC2:0x2_5508 Figure 14-37. IPGIFG Register Definition Table 14-35 describes the fields of the IPGIFG register. Table 14-35. IPGIFG Field Descriptions Bits 0 1–7 Name — Reserved Description Programmable field representing the optional carrier Sense window referenced in Non-Back-to-Back Inter-Packet-Gap, Part IEEE 802.3/4.2.3.2.1 ‘carrier deference’. If carrier is detected during the timing of IPGR1, the MAC defers to carrier. If, however, carrier becomes active after IPGR1, the 1 MAC continues timing IPGR2 and transmits, knowingly causing a collision; thus, ensuring fair access to medium. Its range of values is 0x00 to IPGR2. Its default is 0x40 (64d) which follows the two-thirds/one-third guideline. — Reserved 8 9–15 Programmable field representing the non-back-to-back inter-packet-gap in bits. Its Non-Back-to-Back Inter-Packet-Gap, Part default is 0x60 (96d), which represents the minimum IPG of 96 bits. 2 Minimum IFG Enforcement Programmable field representing the minimum number of bits of IFG to enforce between frames. A frame is dropped whose IFG is less than that programmed. The default setting of 0x50 (80d) represents half of the nominal minimum IFG which is 160 bits. Reserved Programmable field representing the IPG between back-to-back packets. This is the IPG parameter used exclusively in full-duplex mode and in half-duplex mode if two transmit packets are sent back-to-back. Set this field to the number of bits of IPG desired. The default setting of 0x60 (96d) represents the minimum IPG of 96 bits. 16–23 24 25–31 — Back-to-Back Inter-Packet-Gap MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-52 Freescale Semiconductor Memory Map/Register Definition 14.5.3.6.4 Half-Duplex Register (HAFDUP) The HAFDUP register is written by the user. Figure 14-38 shows the HAFDUP register. 0 7 8 11 12 13 14 15 R W Reset 0 0 0 0 0 0 0 0 Alternate BEB Truncation Alt BEB BP No No Excess BackOff BackOff Defer 0000_0000_1010_0001 16 19 20 25 26 31 R Retransmission Maximum W Reset Offset 0 0 0 0 0 0 Collision Window 1111_0000_0011_0111 TSEC1:0x2_450C; TSEC2:0x2_550C Figure 14-38. Half-Duplex Register Definition Table 14-36 describes the fields of the HAFDUP register. Table 14-36. HAFDUP Field Descriptions Bits 0–7 8–11 12 Name — Alternate BEB Truncation Alt BEB Reserved This field is used while alternate binary exponential back-off enable is set. The value programmed is substituted for the Ethernet standard value of ten. Its default is 0xA. Alternate binary exponential back-off. This bit is cleared by default. 0 The Tx MAC follows the standard binary exponential back-off rule. 1 The Tx MAC uses the alternate binary exponential back-off truncation setting instead of the 802.3 standard tenth collision. The standard specifies that any collision after the tenth uses one less than 210 (210 – 1) as the maximum back-off time. Back pressure no back-off. This bit is cleared by default. 0 The Tx MAC follows the binary exponential back-off rule. 1 The Tx MAC immediately re-transmits, following a collision, during back pressure operation. No back-off. This bit is cleared by default. 0 The Tx MAC follows the binary exponential back-off rule. 1 The Tx MAC immediately re-transmits following a collision. Excessively deferred. This bit is set by default. 0 The Tx MAC aborts the transmission of a packet that is excessively deferred. 1 The Tx MAC allows the transmission of a packet that is excessively deferred. This is a programmable field specifying the number of retransmission attempts following a collision before aborting the packet due to excessive collisions. The standard specifies the attempt limit to be 0xF (15d). Its default value is 0xF. Description 13 BP No BackOff 14 No BackOff 15 Excess Defer 16–19 Retransmission Maximum MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-53 Three-Speed Ethernet Controllers Table 14-36. HAFDUP Field Descriptions (continued) Bits 20–25 Name — Reserved Description 26–31 Collision Window This is a programmable field representing the slot time or collision window during which collisions occur in properly configured networks. Because the collision window starts at the beginning of transmission, the preamble and SFD are included. Its default of 0x37 (55d) corresponds to the count of frame bytes at the end of the window. 14.5.3.6.5 Maximum Frame Length Register (MAXFRM) The MAXFRM register is written by the user. Figure 14-39 shows the MAXFRM register. 0 15 16 31 R0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Maximum Frame W Reset Offset 0000_0000_0000_0000_0000_0110_0000_0000 TSEC1:0x2_4510; TSEC2:0x2_5510 Figure 14-39. Maximum Frame Length Register Definition Table 14-37 describes the fields of the MAXFRM register. Table 14-37. MAXFRM Field Descriptions Bits 0–15 16–31 Name — Maximum Frame Reserved By default this field is set to 0x0600 (1536d). It sets the maximum frame size in both the transmit and receive directions. (Refer to MACCFG2[Huge Frame].) Description 14.5.3.6.6 MII Management Configuration Register (MIIMCFG) The MIIMCFG register is written by the user. Figure 14-40 shows the MIIMCFG register. 0 1 26 27 28 29 31 Reset W Mgmt Reset Offset R 00000000000000000000000000 No Pre 0 Mgmt Clock Select 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_4520; TSEC2:0x2_5520 Figure 14-40. MII Management Configuration Register Definition MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-54 Freescale Semiconductor Memory Map/Register Definition Table 14-38 describes the fields of the MIIMCFG register. Table 14-38. MIIMCFG Field Descriptions Bits 0 Name Description Reset Mgmt Reset management. This bit is cleared by default. 0 Allow the MII Mgmt to perform management read/write cycles if requested. 1 Reset the MII Mgmt. — No Pre Reserved Preamble suppress. This bit is cleared by default. 0 The MII Mgmt performs management read/write cycles with 32 clocks of preamble. 1 The MII Mgmt suppresses preamble generation and reduces the management cycle from 64 clocks to 32 clocks. This is in accordance with IEEE 802.3/22.2.4.4.2. Reserved 1–26 27 28 — 29–31 Mgmt Clock This field determines the clock frequency of the Mgmt Clock (EC_MDC). Its default value is 000. Select The source clock frequency is equal to the core complex bus clock divided first by eight and then futher divided by the following value: 000 Source clock divided by 4 001 Source clock divided by 4 010 Source clock divided by 6 011 Source clock divided by 8 100 Source clock divided by 10 101 Source clock divided by 14 110 Source clock divided by 20 111 Source clock divided by 28 14.5.3.6.7 MII Management Command Register (MIIMCOM) The MIIMCOM register is written by the user. Figure 14-41 shows the MIIMCOM register. 0 29 30 31 R0 W Reset Offset 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Scan Read Cycle Cycle 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_4524; TSEC2:0x2_5524 Figure 14-41. MIIMCOM Register Definition MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-55 Three-Speed Ethernet Controllers Table 14-39 describes the fields of the MIIMCOM register. Table 14-39. MIIMCOM Field Descriptions Bits 0–29 30 Name — Reserved Description Scan Scan cycle. This bit is cleared by default. Cycle 0 Normal operation. 1 The MII management continuously performs read cycles. This is useful for polling a PHY register, for example, monitoring link fails. Data is returned in register MIIMSTAT[PHY Status]. Read Read cycle. This bit is cleared by default but is not self-clearing once set. Cycle 0 Normal operation. 1 The MII management performs a single read cycle upon the transition of this bit from 0 to 1 using the PHY address (at MIIMADD[PHY Address]) and the register address (at MIIMADD[Register Address]). The 0 to 1 transition of this bit also causes the MIIMIND[BUSY] bit to be set. The read is complete when the MIIMIND[BUSY] bit clears. Data is returned in register MIIMSTAT[PHY Status]. 31 14.5.3.6.8 MII Management Address Register (MIIMADD) The MIIMADD register is written by the user. Figure 14-42 shows the MIIMADD register. 0 18 19 23 24 26 27 31 R0 0 00 0 0 0 0 00 0 0 0 0 0 00 0 0 PHY Address W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_4528; TSEC2:0x2_5528 000 Register Address Figure 14-42. MIIMADD Register Definition Table 14-40 describes the fields of the MIIMADD register. Table 14-40. MIIMADD Field Descriptions Bits 0–18 19–23 24–26 27–31 Name — PHY Address — Reserved This field represents the 5-bit PHY address field of management cycles. Up to 31 PHYs can be addressed (0 is reserved). Reserved Description Register Address This field represents the 5-bit register address field of management cycles. Up to 32 registers can be accessed. Its default value is 0x00. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-56 Freescale Semiconductor Memory Map/Register Definition 14.5.3.6.9 MII Management Control Register (MIIMCON) The MIIMCON register is written by the user. Figure 14-43 shows the MIIMCON register. 0 15 16 31 R00 00 000 00 00 00 00 0 PHY Control W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_452C; TSEC2:0x2_552C Figure 14-43. MII Management Control Register Definition Table 14-41 describes the fields of the MIIMCON register. Table 14-41. MIIMCON Field Descriptions Bits 0–15 16–31 Name — PHY Control Reserved If written, an MII Mgmt write cycle is performed using this 16-bit data, the pre-configured PHY address (at MIIMADD[PHY Address]) and the register address (at MIIMADD[Register Address]). Its default value is 0x0000. Description 14.5.3.6.10MII Management Status Register (MIIMSTAT) The MIIMSTAT, shown in Figure 14-44, is read-only by the user. 0 15 16 31 R00 0 00 0 00 0 00 0 00 0 0 W Reset Offset PHY Status 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_4530; TSEC2:0x2_5530 Figure 14-44. MIIMSTAT Register Definition Table 14-42 describes the fields of the MIIMSTAT register. Table 14-42. MIIMSTAT Field Descriptions Bits 0–15 16–31 Name — Reserved Description PHY Status Following an MII Mgmt read cycle, the 16-bit data can be read from this location. Its default value is 0x0000. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-57 Three-Speed Ethernet Controllers 14.5.3.6.11 MII Management Indicator Register (MIIMIND) The MIIMIND register is read-only by the user. Figure 14-45 shows the MIIMIND register. 0 28 29 30 31 R 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Not Valid Scan Busy W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_4534; TSEC2:0x2_5534 Figure 14-45. MII Management Indicator Register Definition Table 14-43 describes the fields of the MIIMIND register. Table 14-43. MIIMIND Field Descriptions Bits 0–28 29 Name — Not Valid Reserved Not valid 0 MII Mgmt read cycle has completed and the read data is valid. 1 MII Mgmt read cycle has not completed and the read data is not yet valid. Scan in progress 0 A scan operation (continuous MII Mgmt read cycles) is not in progress. 1 A scan operation (continuous MII Mgmt read cycles) is in progress. Busy 0 MII Mgmt block is not currently performing an MII Mgmt read or write cycle. 1 MII Mgmt block is currently performing an MII Mgmt read or write cycle. Description 30 Scan 31 Busy 14.5.3.6.12Interface Status Register (IFSTAT) Figure 14-46 shows the IFSTAT register. 0 21 22 23 27 28 29 31 R W Reset Offset 000000000000000000000 0 Excess Defer 0 000 0 Link Fail 00 0 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_453C; TSEC2:0x2_553C Figure 14-46. Interface Status Register Definition MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-58 Freescale Semiconductor Memory Map/Register Definition Table 14-44 describes the fields of the IFSTAT register. Table 14-44. IFSTAT Field Descriptions Bits 0–21 22 Name — Reserved Description Excess Defer Excessive transmission defer. This bit latches high and is cleared when read. This bit is cleared by default. 0 Normal operation 1 The MAC excessively defers a transmission. — Link Fail Reserved Link Fail. This bit indicates the status of signal detection. 0 The100X module has detected a “signal detect” for longer than 330 mS. 1 The100X module has detected a “signal detect” for less than 330 mS or not at all. Reserved 23–27 28 29–31 — 14.5.3.6.13Station Address Register Part 1 (MACSTNADDR1) The MACSTNADDR1 register is written by the user. Figure 14-47 shows the MACSTNADDR1 register. The value of the station address written by the user into MACSTNADDR1 and MACSTNADDR2 is byte-reversed from how it would appear in the DA field of a frame in memory. For example, for a station address of 0x12345678ABCD, perform a write to MACSTNADDR1 of 0xCDAB7856, and to MACSTNADDR2 of 0x34120000. When the user reads MACSTNADDR1, 0xCDAB7856 is returned. A read of MACSTNADDR2 returns a value of 0x34120000. Note, the I/G and U/L bits of the frame’s DA field is located at the LSBs of the 1st octet stored in MACSTNADDR2, where the I/G bit is bit 15, and the U/L bit is bit 14. 0 7 8 15 16 23 24 31 R Station Address, 6th Octet Station Address, 5th Octet Station Address, 4th Octet Station Address, 3rd Octet W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_4540; TSEC2:0x2_5540 Figure 14-47. Station Address Part 1 Register Definition MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-59 Three-Speed Ethernet Controllers Table 14-45 describes the fields of the MACSTNADDR1 register. Table 14-45. MACSTNADDR1 Field Descriptions Bits 0–7 Name Description Station Address, 6th Octet This field holds the sixth octet of the station address. The sixth octet (station address bits 40–47) defaults to a value of 0x00. 8–15 Station Address, 5th Octet This field holds the fifth octet of the station address. The fifth octet (station address bits 32–39) defaults to a value of 0x00. 16–23 Station Address, 4th Octet This field holds the fourth octet of the station address. The fourth octet (station address bits 24–31) defaults to a value of 0x00. 24–31 Station Address, 3rd Octet This field holds the third octet of the station address. The third octet (station address bits 16–23) defaults to a value of 0x00. 14.5.3.6.14Station Address Register Part 2 (MACSTNADDR2) The MACSTNADDR2 register is written by the user. Figure 14-48 shows the MACSTNADDR2 register. Note, the I/G and U/L bits of the frame’s DA field is located at the LSBs of the 1st octet stored in MACSTNADDR2, where the I/G bit is bit 15, and the U/L bit is bit 14. 0 7 8 15 16 31 R W Reset Offset Station Address, 2nd Octet 0000000000000000 Station Address, 1st Octet 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_4544; TSEC2:0x2_5544 Figure 14-48. Station Address Part 2 Register Definition Table 14-46 describes the fields of the MACSTNADDR2 register. Table 14-46. MACSTNADDR2 Field Descriptions Bits 0–7 8–15 16–31 Name Station Address, 2nd Octet Description This field holds the second octet of the station address. The second octet (station address bits 8–15) defaults to a value of 0x00. Station Address, 1st Octet This field holds the first octet of the station address. The first octet (station address bits 0–7) defaults to a value of 0x00. — Reserved 14.5.3.7 MIB Registers This section describes the MIB registers. The TSEC MSTAT module has 37 separate statistics counters, which simply count or accumulate statistical events that occur as packets are transmitted and received. These counters support MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-60 Freescale Semiconductor Memory Map/Register Definition RMON MIB group 1, RMON MIB group 2, RMON MIB group 3, RMON MIB group 9, RMON MIB 2, and the 802.3 Ethernet MIB. The detection of one or more of these statistical events triggers the MSTAT module to update its statistics counters. These counters are stored in internal data registers. The user may access the internal data registers at any time. An interrupt can be generated upon any one counter’s rollover condition via a carry interrupt output from the MSTAT. Each counter’s rollover condition can be discreetly masked from causing an interrupt by internal masking registers. In addition, each individual counter value may be reset on read access, or all counters may be simultaneously reset by assertion of an external module input signal. Figure 14-49 shows the TR64 register. 14.5.3.7.1 Transmit and Receive 64-Byte Frame Counter Register (TR64) 0 9 10 31 R0000000000 TR64 W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_4680; TSEC2:0x2_5680 Figure 14-49. Transmit and Receive 64-Byte Frame Register Definition Table 14-47 describes the fields of the TR64 register. Table 14-47. TR64 Field Descriptions Bits Name 0–9 — Reserved Description 10–31 TR64 Transmit and receive 64-byte frame counter—Increment for each good or bad frame transmitted and received which is 64 bytes in length, inclusive (excluding preamble and SFD but including FCS bytes) 14.5.3.7.2 Transmit and Receive 65- to 127-Byte Frame Counter Register (TR127) Figure 14-50 shows the TR127 register. 0 9 10 31 R0 000000000 TR127 W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_4684; TSEC2:0x2_5684 Figure 14-50. Transmit and Receive 65- to 127-Byte Frame Register Definition MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-61 Three-Speed Ethernet Controllers Table 14-48 describes the fields of the TR127 register. Table 14-48. TR127 Field Descriptions Bits 0–9 Name — Reserved Description 10–31 TR127 Transmit and receive 65- to 127-byte frame counter—Increment for each good or bad frame transmitted and received which is 65 to 127 bytes in length, inclusive (excluding preamble and SFD but including FCS bytes) 14.5.3.7.3 Transmit and Receive 128- to 255-Byte Frame Counter Register (TR255) Figure 14-51 shows the TR255 register. 0 9 10 31 R0000000000 TR255 W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_4688; TSEC2:0x2_5688 Figure 14-51. Transmit and Receive 128- to 255-Byte Frame Register Definition Table 14-49 describes the fields of the TR255 register. Table 14-49. TR255 Field Descriptions Bits 0–9 Name — Reserved Description 10–31 TR255 Transmit and receive 128- to 255-byte frame counter—Increments for each good or bad frame transmitted and received which is 128 to 255 bytes in length, inclusive (excluding preamble and SFD but including FCS bytes) 14.5.3.7.4 Transmit and Receive 256- to 511-Byte Frame Counter Register (TR511) Figure 14-52 shows the TR511 register. 0 9 10 31 R0 000000000 TR511 W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_468C; TSEC2:0x2_568C Figure 14-52. Transmit and Receive 256- to 511-Byte Frame Register Definition MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-62 Freescale Semiconductor Memory Map/Register Definition Table 14-50 describes the fields of the TR511 register. Table 14-50. TR511 Field Descriptions Bits 0–9 Name — Reserved Description 10–31 TR511 Increments for each good or bad frame transmitted and received which is 256 to 511 bytes in length, inclusive (excluding preamble and SFD but including FCS bytes) 14.5.3.7.5 Transmit and Receive 512- to 1023-Byte Frame Counter Register (TR1K) Figure 14-53 shows the TR1K register. 0 9 10 31 R0 000000000 TR1K W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_4690; TSEC2:0x2_5690 Figure 14-53. Transmit and Receive 512- to 1023-Byte Frame Register Definition Table 14-51 describes the fields of the TR1K register. Table 14-51. TR1K Field Descriptions Bits 0–9 Name — Reserved Description 10–31 TR1K Increments for each good or bad frame transmitted and received which is 512 to 1023 bytes in length, inclusive (excluding preamble and SFD but including FCS bytes) 14.5.3.7.6 Transmit and Receive 1024- to 1518-Byte Frame Counter Register (TRMAX) Figure 14-54 shows the TRMAX register. 0 9 10 31 R0 000000000 TRMAX W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_4694; TSEC2:0x2_5694 Figure 14-54. Transmit and Receive 1024- to 1518-Byte Frame Register Definition MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-63 Three-Speed Ethernet Controllers Table 14-52 describes the fields of the TRMAX register. Table 14-52. TRMAX Field Descriptions Bits 0–9 Name — Reserved Description 10–31 TRMAX Increments for each good or bad frame transmitted and received which is 1024 to 1518 bytes in length, inclusive (excluding preamble and SFD but including FCS bytes) 14.5.3.7.7 Transmit and Receive 1519- to 1522-Byte VLAN Frame Counter Register (TRMGV) Figure 14-55 shows the TRMGV register. 0 9 10 31 R0000000000 TRMGV W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_4698; TSEC2:0x2_5698 Figure 14-55. Transmit and Receive 1519- to 1522-Byte VLAN Frame Register Definition Table 14-53 describes the fields of the TRMGV register. Table 14-53. TRMGV Field Descriptions Bits 0–9 Name — Reserved Description 10–31 TRMGV Increments for each good or bad frame transmitted and received which is 1519 to 1522 bytes in length, inclusive (excluding preamble and SFD but including FCS bytes) 14.5.3.7.8 Receive Byte Counter Register (RBYT) Figure 14-56 shows the RBYT register. 0 1 31 R0 RBYT W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_469C; TSEC2:0x2_569C Figure 14-56. Receive Byte Counter Register Definition MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-64 Freescale Semiconductor Memory Map/Register Definition Table 14-54 describes the fields of the RBYT register. Table 14-54. RBYT Field Descriptions Bits Name 0 — Reserved Description 1–31 RBYT Receive byte counter. Increments by the byte count of frames received, including those in bad packets, excluding preamble and SFD but including FCS bytes. 14.5.3.7.9 Receive Packet Counter Register (RPKT) Figure 14-57 shows the RPKT register. 0 9 10 31 R0 000000000 RPKT W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_46A0; TSEC2:0x2_56A0 Figure 14-57. Receive Packet Counter Register Definition Table 14-55 describes the fields of the RPKT register. Table 14-55. RPKT Field Descriptions Bits 0–9 Name — Reserved Description 10–31 RPKT Receive packet counter. Increments for each frame received packet (including bad packets, all unicast, broadcast, and multicast packets). 14.5.3.7.10Receive FCS Error Counter Register (RFCS) Figure 14-58 shows the RFCS register. 0 15 16 31 R0 W Reset Offset 0 000000000 0 0000 RFCS 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_46A4; TSEC2:0x2_56A4 Figure 14-58. Receive FCS Error Counter Register Definition MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-65 Three-Speed Ethernet Controllers Table 14-56 describes the fields of the RFCS register. Table 14-56. RFCS Field Descriptions Bits Name 0–15 — Reserved Description 16–31 RFCS Receive FCS error counter. Increments for each frame received that has an integral 64 to 1518 length and contains a frame check sequence error. 14.5.3.7.11 Receive Multicast Packet Counter Register (RMCA) Figure 14-59 shows the RMCA register. 0 9 10 31 R0 W Reset Offset 00 000 00 00 RMCA 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_46A8; TSEC2:0x2_56A8 Figure 14-59. Receive Multicast Packet Counter Register Definition Table 14-57 describes the fields of the RMCA register. Table 14-57. RMCA Field Descriptions Bits 0–9 Name — Reserved Description 10–31 RMCA Receive multicast packet counter. Increments for each multicast good frame of lengths 64 to 1518 (non VLAN) or 1522 (VLAN), excluding broadcast frames. This count does not include range/length errors. 14.5.3.7.12Receive Broadcast Packet Counter Register (RBCA) Figure 14-59 shows the RBCA register. 0 9 10 31 R0 0 0 W Reset Offset 000 00 00 RBCA 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_46AC; TSEC2:0x2_56AC Figure 14-60. Receive Broadcast Packet Counter Register Definition MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-66 Freescale Semiconductor Memory Map/Register Definition Table 14-58 describes the fields of the RBCA register. Table 14-58. RBCA Field Descriptions Bits 0–9 Name — Reserved Description 10–31 RBCA Receive broadcast packet counter. Increments for each broadcast good frame of lengths 64 to 1518 (non VLAN) or 1522 (VLAN), excluding multicast frames. Does not include range/length errors. 14.5.3.7.13Receive Control Frame Packet Counter Register (RXCF) Figure 14-61 shows the RXCF register. 0 15 16 31 R0 0000000000 0 0 0 0 W Reset Offset 0 RXCF 0000_0000_0000_0000)0000_0000_0000_0000 TSEC1:0x2_46B0; TSEC2:0x2_56B0 Figure 14-61. Receive Control Frame Packet Counter Register Definition Table 14-59 describes the fields of the RXCF register. Table 14-59. RXCF Field Descriptions Bits Name 0–15 — Reserved Description 16–31 RXCF Receive control frame packet counter. Increments for each MAC control frame received (both pause control frames and control frames unsupported by IEEE). 14.5.3.7.14Receive Pause Frame Packet Counter Register (RXPF) Figure 14-62 shows the RXPF register. 0 15 16 31 R00000000000 0 0 0 0 W Reset Offset 0 RXPF 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_46B4; TSEC2:0x2_56B4 Figure 14-62. Receive Pause Frame Packet Counter Register Definition MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-67 Three-Speed Ethernet Controllers Table 14-60 describes the fields of the RXPF register. Table 14-60. RXPF Field Descriptions Bits 0–15 Name — Reserved Description 16–31 RXPF Receive PAUSE frame packet counter. Increments each time a valid PAUSE MAC control frame is received. 14.5.3.7.15Receive Unknown Opcode Packet Counter Register (RXUO) Figure 14-63 shows the RXUO register. 0 15 16 31 R0 0000000000 0 0 0 0 W Reset Offset 0 RXUO 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_46B8; TSEC2:0x2_56B8 Figure 14-63. Receive Unknown Opcode Packet Counter Register Definition Table 14-61 describes the fields of the RXUO register. Table 14-61. RXUO Field Descriptions Bits 0–15 Name — Reserved Description 16–31 RXUO Receive unknown opcode counter. Increments each time a MAC control frame is received which contains an opcode other than a PAUSE. 14.5.3.7.16Receive Alignment Error Counter Register (RALN) Figure 14-64 shows the RALN register. 0 15 16 31 R0 0000000000 0 0 0 0 W Reset Offset 0 RALN 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_46BC; TSEC2:0x2_56BC Figure 14-64. Receive Alignment Error Counter Register Definition MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-68 Freescale Semiconductor Memory Map/Register Definition Table 14-62 describes the fields of the RALN register. Table 14-62. RALN Field Descriptions Bits Name 0–15 — Reserved Description 16–31 RALN Receive alignment error counter. Increments for each received frame from 64 to 1518 (non VLAN) or 1522 (VLAN) which contains an invalid FCS and is not an integral number of bytes. 14.5.3.7.17Receive Frame Length Error Counter Register (RFLR) Figure 14-65 shows the RFLR register. 0 15 16 31 R0 000000000 0 0 0 0 0 W Reset Offset 0 RFLR 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_46C0; TSEC2:0x2_56C0 Figure 14-65. Receive Frame Length Error Counter Register Definition Table 14-63 describes the fields of the RFLR register. Table 14-63. RFLR Field Descriptions Bits 0–15 Name — Reserved Description 16–31 RFLR Receive frame length error counter. Increments for each frame received in which the 802.3 length field did not match the number of data bytes actually received (46 –1500 bytes). The counter does not increment if the length field is not a valid 802.3 length, such as an ethertype value. 14.5.3.7.18Receive Code Error Counter Register (RCDE) Figure 14-66 shows the RCDE register. 0 15 16 31 R000000000000 000 0 RCDE W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_46C4; TSEC2:0x2_56C4 Figure 14-66. Receive Code Error Counter Register Definition MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-69 Three-Speed Ethernet Controllers Table 14-64 describes the fields of the RCDE register. Table 14-64. RCDE Field Descriptions Bits 0–15 Name — Reserved Description 16–31 RCDE Receive code error counter. Increments each time a valid carrier is present and at least one invalid data symbol is detected. 14.5.3.7.19Receive Carrier Sense Error Counter Register (RCSE) Figure 14-67 shows the RCSE register. 0 15 16 31 R00000000000 0 0 00 0 RCSE W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_46C8; TSEC2:0x2_56C8 Figure 14-67. Receive Carrier Sense Error Counter Register Definition Table 14-65 describes the fields of the RCSE register. Table 14-65. RCSE Field Descriptions Bits 0–15 Name — Reserved Description 16–31 RCSE Receive false carrier counter. Increments each time a false carrier is detected during idle, as defined by a 1 on TSECn_RX_ER and an 0xE on TSEC n_RXD. The event is reported along with the statistics generated on the next received frame. Only one false carrier condition can be detected and logged between frames. 14.5.3.7.20Receive Undersize Packet Counter Register (RUND) Figure 14-68 shows the RUND register. 0 15 16 31 R0 0000000000 0 0 0 0 W Reset Offset 0 RUND 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_46CC; TSEC2:0x2_56CC Figure 14-68. Receive Undersize Packet Counter Register Definition MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-70 Freescale Semiconductor Memory Map/Register Definition Table 14-66 describes the fields of the RUND register. Table 14-66. RUND Field Descriptions Bits 0–15 Name — Reserved Description 16–31 RUND Receive undersize packet counter. Increments each time a frame is received which is less than 64 bytes in length and contains a valid FCS and were otherwise well formed. This count does not include range length errors. 14.5.3.7.21Receive Oversize Packet Counter Register (ROVR) Figure 14-69 shows the ROVR register. 0 15 16 31 R0 000000000 0 0 0 0 0 W Reset Offset 0 ROVR 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_46D0; TSEC2:0x2_56D0 Figure 14-69. Receive Oversize Packet Counter Register Definition Table 14-67 describes the fields of the ROVR register. Table 14-67. ROVR Field Descriptions Bits 0–15 Name — Reserved Description 16–31 ROVR Receive oversize packet counter. Increments each time a frame is received which exceeded 1518 (non VLAN) or 1522 (VLAN) and contains a valid FCS and was otherwise well formed. This count does not include range length errors. 14.5.3.7.22Receive Fragments Counter Register (RFRG) Figure 14-70 shows the RFRG register. 0 15 16 31 R0 0000000000 0 0 0 0 W Reset Offset 0 RFRG 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_46D4; TSEC2:0x2_56D4 Figure 14-70. Receive Fragments Counter Register Definition MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-71 Three-Speed Ethernet Controllers Table 14-68 describes the fields of the RFRG register. Table 14-68. RFRG Field Descriptions Bits 0–15 Name — Reserved Description 16–31 RFRG Receive fragments counter. Increments for each frame received which is less than 64 bytes in length and contains an invalid FCS. This includes integral and non-integral lengths. 14.5.3.7.23Receive Jabber Counter Register (RJBR) Figure 14-71 shows the RJBR register. 0 15 16 31 R0 0000000000 0 0 0 0 W Reset Offset 0 RJBR 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_46D8; TSEC2:0x2_56D8 Figure 14-71. Receive Jabber Counter Register Definition Table 14-69 describes the fields of the RJBR register. Table 14-69. RJBR Field Descriptions Bits 0–15 Name — Reserved Description 16–31 RJBR Receive jabber counter. Increments for frames received which exceed 1518 (non VLAN) or 1522 (VLAN) bytes and contain an invalid FCS. This includes alignment errors. 14.5.3.7.24Receive Dropped Packet Counter Register (RDRP) Figure 14-72 shows the RDRP register. 0 15 16 31 R0 0000000000 0 0 0 0 W Reset Offset 0 RDRP 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_46DC; TSEC2:0x2_56DC Figure 14-72. Receive Dropped Packet Counter Register Definition MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-72 Freescale Semiconductor Memory Map/Register Definition Table 14-70 describes the fields of the RDRP register. Table 14-70. RDRP Field Descriptions Bits 0–15 Name — Reserved Description 16–31 RDRP Receive dropped packets counter. Increments for frames received which are streamed to system but are later dropped due to lack of system resources. 14.5.3.7.25Transmit Byte Counter Register (TBYT) Figure 14-73 shows the TBYT register. 0 1 31 R0 TBYT W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_46E0; TSEC2:0x2_56E0 Figure 14-73. Transmit Byte Counter Register Definition Table 14-71 describes the fields of the TBYT register. Table 14-71. TBYT Field Descriptions Bits Name 0 — Reserved Description 1–31 TBYT Transmit byte counter. Increments by the number of bytes that were put on the wire including fragments of frames that were involved with collisions. This count does not include preamble/SFD or jam bytes. This counter does not count if the frame is truncated. 14.5.3.7.26Transmit Packet Counter Register (TPKT) Figure 14-74 shows the TPKT register. 0 9 10 31 R0 000000000 TPKT W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_46E4; TSEC2:0x2_56E4 Figure 14-74. Transmit Packet Counter Register Definition MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-73 Three-Speed Ethernet Controllers Table 14-72 describes the fields of the TPKT register. Table 14-72. TPKT Field Descriptions Bits 0–9 Name — Reserved Description 10–31 TPKT Transmit packet counter. Increments for each transmitted packet (including bad packets, excessive deferred packets, excessive collision packets, late collision packets, all unicast, broadcast, and multicast packets). 14.5.3.7.27Transmit Multicast Packet Counter Register (TMCA) Figure 14-75 shows the TMCA register. 0 9 10 31 R0 000000000 TMCA W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_46E8; TSEC2:0x2_56E8 Figure 14-75. Transmit Multicast Packet Counter Register Definition Table 14-73 describes the fields of the TMCA register. Table 14-73. TMCA Field Descriptions Bits 0–9 Name — Reserved Description 10–31 TMCA Transmit multicast packet counter. Increments for each multicast valid frame transmitted (excluding broadcast frames). 14.5.3.7.28Transmit Broadcast Packet Counter Register (TBCA) Figure 14-76 shows the TBCA register. 0 9 10 31 R0 0 0 0 0 0 0 0 0 0 TBCA W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_46EC; TSEC2:0x2_56EC Figure 14-76. Transmit Broadcast Packet Counter Register Definition MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-74 Freescale Semiconductor Memory Map/Register Definition Table 14-74 describes the fields of the TBCA register. Table 14-74. TBCA Field Descriptions Bits Name 0–9 — Reserved Description 10–31 TBCA Transmit broadcast packet counter. Increments for each broadcast frame transmitted (excluding multicast frames). 14.5.3.7.29Transmit Pause Control Frame Counter Register (TXPF) Figure 14-77 shows the TXPF register. 0 15 16 31 R0 0000000000 0 0 0 0 W Reset Offset 0 TXPF 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_46F0; TSEC2:0x2_56F0 Figure 14-77. Transmit Pause Control Frame Counter Register Definition Table 14-75 describes the fields of the TXPF register. Table 14-75. TXPF Field Descriptions Bits Name 0–15 — Reserved Description 16–31 TXPF Transmit PAUSE frame packet counter. Increments each time a valid PAUSE MAC control frame is transmitted. 14.5.3.7.30Transmit Deferral Packet Counter Register (TDFR) Figure 14-78 shows the TDFR register. 0 19 20 31 R0 0000000000 0 0 0 0 W Reset Offset 0 0 0 0 0 TDFR 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_46F4; TSEC2:0x2_56F4 Figure 14-78. Transmit Deferral Packet Counter Register Definition MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-75 Three-Speed Ethernet Controllers Table 14-76 describes the fields of the TDFR register. Table 14-76. TDFR Field Descriptions Bits 0–19 Name — Reserved Description 20–31 TDFR Transmit deferral packet counter. Increments for each frame, which was deferred on its first transmission attempt. This count does not include frames involved in collisions. 14.5.3.7.31Transmit Excessive Deferral Packet Counter Register (TEDF) Figure 14-79 shows the TEDF register. 0 19 20 31 R0 0000000000 0 0 0 0 W Reset Offset 0 0 0 0 0 TEDF 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_46F8; TSEC2:0x2_56F8 Figure 14-79. Transmit Excessive Deferral Packet Counter Register Definition Table 14-77 describes the fields of the TEDF register. Table 14-77. TEDF Field Descriptions Bits Name 0–19 — Reserved Description 20–31 TEDF Transmit excessive deferral packet counter. Increments for frames aborted which were deferred for an excessive period of time (3036 byte times). 14.5.3.7.32Transmit Single Collision Packet Counter Register (TSCL) Figure 14-80 shows the TSCL register. 0 19 20 31 R0 0000000000 0 0 0 0 W Reset Offset 0 0 0 0 0 TSCL 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_46FC; TSEC2:0x2_56FC Figure 14-80. Transmit Single Collision Packet Counter Register Definition MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-76 Freescale Semiconductor Memory Map/Register Definition Table 14-78 describes the fields of the TSCL register. Table 14-78. TSCL Field Descriptions Bits Name 0–19 — Reserved Description 20–31 TSCL Transmit single collision packet counter. Increments for each frame transmitted which experienced exactly one collision during transmission. 14.5.3.7.33Transmit Multiple Collision Packet Counter Register (TMCL) Figure 14-81 shows the TMCL register. 0 19 20 31 R0 0000000000 0 0 0 0 W Reset Offset 0 0 0 0 0 TMCL 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_4700; TSEC2:0x2_5700 Figure 14-81. Transmit Multiple Collision Packet Counter Register Definition Table 14-79 describes the fields of the TMCL register. Table 14-79. TMCL Field Descriptions Bits Name 0–19 — Reserved Description 20–31 TMCL Transmit multiple collision packet counter. Increments for each frame transmitted which experienced 2-15 collisions (including any late collisions) during transmission as defined using the Half_Duplex[RETRANSMISSION MAXIMUM] field. 14.5.3.7.34Transmit Late Collision Packet Counter Register (TLCL) Figure 14-82 shows the TLCL register. 0 19 20 31 R0 0000000000 0 0 0 0 W Reset Offset 0 0 0 0 0 TLCL 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_4704; TSEC2:0x2_5704 Figure 14-82. Transmit Late Collision Packet Counter Register Definition MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-77 Three-Speed Ethernet Controllers Table 14-80 describes the fields of the TLCL register. Table 14-80. TLCL Field Descriptions Bits Name 0–19 — Reserved Description 20–31 TLCL Transmit late collision packet counter. Increments for each frame transmitted which experienced a late collision during a transmission attempt. 14.5.3.7.35Transmit Excessive Collision Packet Counter Register (TXCL) Figure 14-83 shows the TXCL register. 0 19 20 31 R0 0000000000 0 0 0 0 W Reset Offset 0 0 0 0 0 TXCL 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_4708; TSEC2:0x2_5708 Figure 14-83. Transmit Excessive Collision Packet Counter Register Definition Table 14-81. TXCL Field Descriptions Bits Name 0–19 — Reserved Description 20–31 TXCL Transmit excessive collision packet counter. Increments for each frame that experienced 16 collisions during transmission and was aborted. 14.5.3.7.36Transmit Total Collision Counter Register (TNCL) Figure 14-84 shows the TNCL register. 0 19 20 31 R0 0000000000 0 0 0 0 W Reset Offset 0 0 0 0 0 TNCL 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_470C; TSEC2:0x2_570C Figure 14-84. Transmit Total Collision Counter Register Definition MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-78 Freescale Semiconductor Memory Map/Register Definition Table 14-82 describes the fields of the TNCL register. Table 14-82. TNCL Field Descriptions Bits Name 0–19 — Reserved Description 20–31 TNCL Transmit total collision counter. Increments by the number of collisions experienced during the transmission of a frame as defined as the simultaneous presence of signals on the DO and RD circuits (that is, transmitting and receiving at the same time). Note: This count does not include collisions that result in an excessive collision condition. 14.5.3.7.37Transmit Drop Frame Counter Register (TDRP) Figure 14-85 shows the TDRP register. 0 15 16 31 R000000000000 00 0 0 TDRP W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_4714; TSEC2:0x2_5714 Figure 14-85. Transmit Drop Frame Counter Register Definition Table 14-83 describes the fields of the TDRP register. Table 14-83. TDRP Field Descriptions Bits 0–15 16–31 Name — TDRP Reserved Transmit drop frame counter. Increments each time a memory error or an underrun has occurred. Description 14.5.3.7.38Transmit Jabber Frame Counter Register (TJBR) Figure 14-86 shows the TJBR register. 0 19 20 31 R0 0 00000000 0 0 0 0 0 W Reset Offset 0 0 0 0 0 TJBR 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_4718; TSEC2:0x2_5718 Figure 14-86. Transmit Jabber Frame Counter Register Definition MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-79 Three-Speed Ethernet Controllers Table 14-84 describes the fields of the TJBR register. Table 14-84. TJBR Field Descriptions Bits Name 0–19 — Reserved Description 20–31 TJBR Transmit jabber frame counter. Increments for each oversized transmitted frame with an incorrect FCS value. 14.5.3.7.39Transmit FCS Error Counter Register (TFCS) Figure 14-87 shows the TFCS register. 0 19 20 31 R0 0000000000 0 0 0 0 W Reset Offset 0 0 0 0 0 TFCS 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_471C; TSEC2:0x2_571C Figure 14-87. Transmit FCS Error Counter Register Definition Table 14-85 describes the fields of the TFCS register. Table 14-85. TFCS Field Descriptions Bits 0–19 20–31 Name — TFCS Reserved Transmit FCS error counter. Increments for every valid sized packet with an incorrect FCS value. Description 14.5.3.7.40Transmit Control Frame Counter Register (TXCF) Figure 14-88 shows the TXCF register. 0 19 20 31 R00000000000 0 0 0 0 W Reset Offset 0 0 0 0 0 TXCF 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_4720; TSEC2:0x2_5720 Figure 14-88. Transmit Control Frame Counter Register Definition MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-80 Freescale Semiconductor Memory Map/Register Definition Table 14-86 describes the fields of the TXCF register. Table 14-86. TXCF Field Descriptions Bits 0–19 Name — Reserved Description 20–31 TXCF Transmit control frame counter. Increments for every valid size frame with a type field signifying a control frame. 14.5.3.7.41Transmit Oversize Frame Counter Register (TOVR) Figure 14-89 shows the TOVR register. 0 19 20 31 R0 0000000000 0 0 0 0 W Reset Offset 0 0 0 0 0 TOVR 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_4724; TSEC2:0x2_5724 Figure 14-89. Transmit Oversized Frame Counter Register Definition Table 14-87 describes the fields of the TOVR register. Table 14-87. TOVR Field Descriptions Bits Name 0–19 — Reserved Description 20–31 TOVR Transmit oversize frame counter. Increments for each oversized transmitted frame with a correct FCS value. 14.5.3.7.42Transmit Undersize Frame Counter Register (TUND) Figure 14-90 shows the TUND register. 0 19 20 31 R0 0000000000 0 0 00 0 W Reset Offset 0 0 0 0 TDFR 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_4728; TSEC2:0x2_5728 Figure 14-90. Transmit Undersize Frame Counter Register Definition MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-81 Three-Speed Ethernet Controllers Table 14-88 describes the fields of the TUND register. Table 14-88. TUND Field Descriptions Bits Name 0–19 — Reserved Description 20–31 TDFR Transmit undersize frame counter. Increments for every frame less then 64 bytes, with a correct FCS value. 14.5.3.7.43Transmit Fragment Counter Register (TFRG) Figure 14-91 shows the TFRG register. 0 19 20 31 R00000000000 0 0 0 0 W Reset Offset 0 0 0 0 0 TFRG 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_472C; TSEC2:0x2_572C Figure 14-91. Transmit Fragment Counter Register Definition Table 14-89 describes the fields of the TFRG register. Table 14-89. TFRG Field Descriptions Bits 0–19 Name — Reserved Description 20–31 TFRG Transmit fragment counter. Increments for every frame less then 64 bytes, with an incorrect FCS value. 14.5.3.7.44Carry Register 1 (CAR1) Carry register bits are cleared when written with a one. Figure 14-92 shows the CAR1 register. 0 1 2 3 4 5 6 7 14 15 R W Reset C1 64 C1 127 C1 255 C1 511 C1 1K C1 C1 MAX MGV 0 0 0 0 0 0 0 0 C1 RBY 0000_0000_0000_0000 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 C1 C1 C1 C1 C1 RPK RFC RMC RBC RXC W Reset Offset R C1 RXP C1 C1 RXU RAL C1 RFL C1 C1 C1 C1 C1 C1 C1 RCD RCS RUN ROV RFR RJB RDR 0000_0000_0000_0000 TSEC1:0x2_4730; TSEC2:0x2_5730 Figure 14-92. Carry Register 1 (CAR1) Register Definition MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-82 Freescale Semiconductor Memory Map/Register Definition Table 14-90 describes the fields of the CAR1 register. Table 14-90. CAR1 Field Descriptions Bits 0 1 2 3 4 5 6 7–14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 Name C164 C1127 C1255 C1511 C11K C1MAX C1MGV — C1RBY C1RPK C1RFC C1RMC C1RBC C1RXC C1RXP C1RXU C1RAL C1RFL C1RCD C1RCS C1RUN C1ROV C1RFR C1RJB C1RDR Description Carry register 1 TR64 counter carry bit Carry register 1 TR127 counter carry bit Carry register 1 TR255 counter carry bit Carry register 1 TR511 counter carry bit Carry register 1 TR1K counter carry bit Carry register 1 TRMAX counter carry bit Carry register 1 TRMGV counter carry bit Reserved Carry register 1 RBYT counter carry bit Carry register 1 RPKT counter carry bit Carry register 1 RFCS counter carry bit Carry register 1 RMCA counter carry bit Carry register 1 RBCA counter carry bit Carry register 1 RXCF counter carry bit Carry register 1 RXPF counter carry bit Carry register 1 RXUO counter carry bit Carry register 1 RALN counter carry bit Carry register 1 RFLR counter carry bit Carry register 1 RCDE counter carry bit Carry register 1 RCSE counter carry bit Carry register 1 RUND counter carry bit Carry register 1 ROVR counter carry bit Carry register 1 RFRG counter carry bit Carry register 1 RJBR counter carry bit Carry register 1 RDRP counter carry bit MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-83 Three-Speed Ethernet Controllers 14.5.3.7.45Carry Register 2 (CAR2) Carry register bits are cleared when written with a one. Figure 14-93 shows the CAR2 register. 0 11 12 13 14 15 R W Reset 0 0 0 0 0 0 0 0 0 0 0 0 C2 C2 C2 C2 TJB TFC TCF TOV 0000_0000_0000_0000 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 C2 C2 TUN TFG W Reset Offset R C2 TBY C2 C2 C2 C2 C2 C2 TPK TMC TBC TPF TDF TED C2 C2 C2 TSC TMA TLC C2 C2 TXC TNC 0 C2 TDP 0000_0000_0000_0000 TSEC1:0x2_4734; TSEC2:0x2_5734 Figure 14-93. Carry Register 2 (CAR2) Register Definition Table 14-91 describes the fields of the CAR2 register. Table 14-91. CAR2 Field Descriptions Bits 0–11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 Name — C2TJB C2TFC C2TCF C2TOV C2TUN C2TFG C2TBY C2TPK C2TMC C2TBC C2TPF C2TDF C2TED C2TSC C2TMA C2TLC C2TXC Reserved Carry register 2 TJBR counter carry bit Carry register 2 TFCS counter carry bit Carry register 2 TXCF counter carry bit Carry register 2 TOVR counter carry bit Carry register 2 TUND counter carry bit Carry register 2 TFRG counter carry bit Carry register 2 TBYT counter carry bit Carry register 2 TPKT counter carry bit Carry register 2 TMCA counter carry bit Carry register 2 TBCA counter carry bit Carry register 2 TXPF counter carry bit Carry register 2 TDFR counter carry bit Carry register 2 TEDF counter carry bit Carry register 2 TSCL counter carry bit Carry register 2 TMCL counter carry bit Carry register 2 TLCL counter carry bit Carry register 2 TXCL counter carry bit Description MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-84 Freescale Semiconductor Memory Map/Register Definition Table 14-91. CAR2 Field Descriptions (continued) Bits 29 30 31 Name C2TNC — C2TDP Description Carry register 2 TNCL counter carry bit Reserved Carry register 2 TDRP counter carry bit 14.5.3.7.46Carry Mask Register 1 (CAM1) As long as one of the below mask bits is cleared, the corresponding interrupt bit is allowed to cause interrupt indications on output CARRY. These bits all default to a set state. Figure 14-94 shows the CAM1 register. 0 1 2 3 4 5 6 7 14 15 R W Reset M1 64 M1 127 M1 255 M1 511 M1 1K M1 M1 MAX MGV 0 0 0 0 0 0 0 0 M1 RBY 1111_1110_0000_0001 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 M1 M1 M1 M1 RPK RFC RMC RBC W Reset Offset R M1 RXC M1 RXP M1 M1 M1 RXU RAL RFL M1 M1 M1 M1 M1 M1 M1 RCD RCS RUN ROV RFR RJB RDR 1111_1111_1111_1111 TSEC1:0x2_4738; TSEC2:0x2_5738 Figure 14-94. Carry Mask Register 1 (CAM1) Register Definition Table 14-92 describes the fields of the CAM1 register. Table 14-92. CAM1 Field Descriptions Bits 0 1 2 3 4 5 6 7–14 15 16 Name M164 M1127 M1255 M1511 M11K M1MAX M1MGV — M1RBY M1RPK Description Mask register 1 TR64 counter carry bit mask Mask register 1 TR127 counter carry bit mask Mask register 1 TR255 counter carry bit mask Mask register 1 TR511 counter carry bit mask Mask register 1 TR1K counter carry bit mask Mask register 1 TRMAX counter carry bit mask Mask register 1 TRMGV counter carry bit mask Reserved Mask register 1 RBYT counter carry bit mask Mask register 1 RPKT counter carry bit mask MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-85 Three-Speed Ethernet Controllers Table 14-92. CAM1 Field Descriptions (continued) Bits 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 Name M1RFC M1RMC M1RBC M1RXC M1RXP M1RXU M1RAL M1RFL M1RCD M1RCS M1RUN M1ROV M1RFR M1RJB M1RDR Description Mask register 1 RFCS counter carry bit mask Mask register 1 RMCA counter carry bit mask Mask register 1 RBCA counter carry bit mask Mask register 1 RXCF counter carry bit mask Mask register 1 RXPF counter carry bit mask Mask register 1 RXUO counter carry bit mask Mask register 1 RALN counter carry bit mask Mask register 1 RFLR counter carry bit mask Mask register 1 RCDE counter carry bit mask Mask register 1 RCSE counter carry bit mask Mask register 1 RUND counter carry bit mask Mask register 1 ROVR counter carry bit mask Mask register 1 RFRG counter carry bit mask Mask register 1 RJBR counter carry bit mask Mask register 1 RDRP counter carry bit mask 14.5.3.7.47Carry Mask Register 2 (CAM2) As long as one of the below mask bits is cleared, the corresponding interrupt bit is allowed to cause interrupt indications on output CARRY. These bits default to a set state. Figure 14-95 shows the CAM2 register. 0 11 12 13 14 15 R W Reset 0 0 0 0 0 0 0 0 0 0 0 0 M2 M2 M2 M2 TJB TFC TCF TOV 0000_0000_0000_1111 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 M2 M2 M2 TUN TFG TBY W Reset Offset R M2 M2 TPK TMC M2 TBC M2 M2 M2 TPF TDF TED M2 TSC M2 M2 TMA TLC M2 M2 TXC TNC 0 M2 TDP 1111_1111_1111_1111 TSEC1:0x2_473C; TSEC2:0x2_573C Figure 14-95. Carry Mask Register 2 (CAM2) Register Definition MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-86 Freescale Semiconductor Memory Map/Register Definition Table 14-93 describes the fields of the CAM2 register. Table 14-93. CAM2 Field Descriptions Bits 0–11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 Name — M2TJB M2TFC M2TCF M2TOV M2TUN M2TFG M2TBY M2TPK M2TMC M2TBC M2TPF M2TDF M2TED M2TSC M2TMA M2TLC M2TXC M2TNC — M2TDP Reserved Mask register 2 TJBR counter carry bit mask Mask register 2 TFCS counter carry bit mask Mask register 2 TXCF counter carry bit mask Mask register 2 TOVR counter carry bit mask Mask register 2 TUND counter carry bit mask Mask register 2 TFRG counter carry bit mask Mask register 2 TBYT counter carry bit mask Mask register 2 TPKT counter carry bit mask Mask register 2 TMCA counter carry bit mask Mask register 2 TBCA counter carry bit mask Mask register 2 TXPF counter carry bit mask Mask register 2 TDFR counter carry bit mask Mask register 2 TEDF counter carry bit mask Mask register 2 TSCL counter carry bit mask Mask register 2 TMCL counter carry bit mask Mask register 2 TLCL counter carry bit mask Mask register 2 TXCL counter carry bit mask Mask register 2 TNCL counter carry bit mask Reserved Mask register 2 TDRP counter carry bit mask Description 14.5.3.8 Hash Function Registers This section provides detailed descriptions of the registers used for hash functions. All of the registers are 32 bits wide. If the DA field of a receive frame is processed through a 32-bit CRC generator, the 8 bits of the CRC remainder is mapped to a hash table entry. The user can enable a hash entry by setting the appropriate bit. A hash entry usually represents a set of addresses. A hash table hit occurs if the DA CRC result points to an enabled hash entry. The user must further filter the address. See Section 14.6.2.6.2, “Hash Table Algorithm,” for more information on the hash algorithm. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-87 Three-Speed Ethernet Controllers 14.5.3.8.1 Individual Address Registers 0–7 (IADDRn) The IADDRn registers, shown in Figure 14-96, are written by the user. These registers represent 256 entries of the individual (unicast) address hash table used in the address recognition process. When the DA field of a receive frame is processed through a 32-bit CRC generator, the 8 high-order bits (0–7) of the CRC remainder are mapped to one of the 256 entries. The user can enable a hash entry by setting the appropriate bit. A hash table hit occurs if the DA CRC result points to an enabled hash entry. 0 31 R IADDRn W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_4800, 0x2_4804, 0x2_4808, 0x2_480C, 0x2_4810, 0x2_4814, 0x2_4818, 0x2_481C TSEC2:0x2_5800, 0x2_5804, 0x2_5808, 0x2_580C, 0x2_5810, 0x2_5814, 0x2_5818, 0x2_581C Figure 14-96. IADDRn Register Definition Table 14-94 describes the field of the IADDRn registers. Table 14-94. IADDRn Field Description Bits Name Description 0–31 IADDRn Represents the 32-bit value associated with the corresponding register. IADDR0 contains the high-order 32 bits of the 256-entry hash table and IADDR7 represents the low-order 32 bits. For instance, the MSB of IADDR0 correlates to entry 0 and the LSB of IADDR7 correlates to entry 255. 14.5.3.8.2 Group Address Registers 0–7 (GADDRn) The GADDRn registers are written by the user. Together these registers represent 256 entries of the group (multicast) address hash table used in the address recognition process. While the DA field of a receive frame is processed through a 32-bit CRC generator, the 8 bits of the CRC remainder is mapped to one of the 256 entries. The user can enable a hash entry by setting the appropriate bit. A hash table hit occurs if the DA CRC result points to an enabled hash entry. Figure 14-97 shows the GADDRn registers. 0 31 R GADDR n W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_4880, 0x2_4884, 0x2_4888, 0x2_488C, 0x2_4890, 0x2_4894, 0x2_4898, 0x2_489C TSEC2:0x2_5880, 0x2_5884, 0x2_5888, 0x2_588C, 0x2_5890, 0x2_5894, 0x2_5898, 0x2_589C Figure 14-97. GADDRn Register Definition MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-88 Freescale Semiconductor Memory Map/Register Definition Table 14-95 describes GADDRn. Table 14-95. GADDRn Field Description Bits 0–31 Name Description GADDR n Represents the 32-bit value associated with the corresponding register. GADDR0 contains the high-order 32 bits of the 256-entry group hash table and GADDR7 represents the low-order 32 bits. For instance, the MSB of GADDR0 correlates to entry 0 and the LSB of GADDR7 correlates to entry 255. 14.5.3.9 Attribute Registers This section describes the two TSEC frame attribute registers. 14.5.3.9.1 Attribute Register (ATTR) The attribute register, shown in Figure 14-98, defines attributes and transaction types used to access buffer descriptors, to write receive data, and to read transmit data. Snoop enable attributes may be set for reading buffer descriptors and for reading transmit data. Buffer descriptors may be written with attributes that cause allocation into the L2 cache with or without line locking. Similarly, sections of a receive frame header may have attributes attached that cause allocation and locking in the L2 cache. This process of specifying a region of each frame to stash into the L2 cache is referred to as extraction, which is specified in conjunction with ATTRELI. ATTR[ELCWT] only has meaning if ATTRELI[EL] is non-zero. 0 16 17 18 19 20 21 22 23 24 25 26 31 R000000000000000 0 0 ELCWT W Reset Offset 0 BDLWT 00 RDSEN RBDSEN 000000 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_4BF8; TSEC2:0x2_5BF8 Figure 14-98. ATTR Register Definition Table 14-96 describes the fields of the ATTR register. Table 14-96. ATTR Field Descriptions Bits 0–16 Name — Reserved Description 17–18 ELCWT Extracted L2 cache write type. Specifies the write transaction type to perform for the extracted data. This occurs only if ATTRELI[EL] is non-zero. For maximum performance, it is recommended that if ELCWT is set to allocate, BDLWT must also be set to allocate. Writes to cache are always performed with snoop. This setting overrides the RDSEN bit setting. 00 No allocation performed 01 Reserved, no extraction occurs 10 Allocate L2 cache line 11 Allocate and lock L2 cache line MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-89 Three-Speed Ethernet Controllers Table 14-96. ATTR Field Descriptions (continued) Bits 19 Name — Reserved Description 20–21 BDLWT Buffer descriptor L2 cache write type. Specifies the write transaction type to perform for the buffer descriptor for a receive frame. Writes to cache are always performed with snoop. 00 No allocation performed. This setting overrides the RBDSEN bit setting. 01 Reserved 10 Allocate L2 cache line 11 Allocate and lock L2 cache line 22–23 24 — Reserved RDSEN Rx data snoop enable. This bit is superseded by the ELCWT settings. 0 Disables snooping of all receive frames data to memory unless ELCWT specifies L2 allocation 1 Enables snooping of all receive frames data to memory RBDSEN RxBD snoop enable. This bit is superseded by the BDLWT settings. 0 Disables snooping of all receive BD memory accesses unless BDLWT specifies L2 allocation 1 Enables snooping of all receive BD memory accesses — Reserved 25 26–31 14.5.3.9.2 Attribute Extract Length and Extract Index Register (ATTRELI) The ATTRELI registers are written by the user to specify the extract index and extract length. Figure 14-99 shows the ATTRELI register. 0 1 2 15 16 17 18 31 R W Reset Offset 00 EL 00 EI 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_4BFC; TSEC2:0x2_5BFC Figure 14-99. ATTRELI Register Definition Table 14-97 describes the fields of the ATTRELI register. Table 14-97. ATTRELI Field Descriptions Bits Name 0–1 2–15 16–17 18–31 — EL — EI Reserved Extracted length. Specifies the number of bytes to extract from the receive frame. The DMA controller uses this field to perform extraction. If set to zero, no extraction is performed. Reserved Extracted index. Points to the first byte within the receive frame from which to begin extracting data. Description MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-90 Freescale Semiconductor Memory Map/Register Definition 14.5.4 Ten-Bit Interface (TBI) This section describes the ten-bit interface (TBI) and the TBI MII set of registers. 14.5.4.1 TBI MII Set Register Descriptions This section describes the TBI MII registers. All TBI registers are 16 bits wide and are accessed at the offset of the TBI physical address. The TBI physical address of the TSEC is stored in TBIPA. Writing to the TBI registers is similar to writing to an external PHY, by using the MII management interface. By using TBIPA in place of the PHY address, in the MIIMADD[PHY Address] field, and setting MIIMADD[Register Address] to the address offset that corresponds to the desired register (see Table 14-98), the user can read (set MIIMCOM[Read Cycle]) or write (writing to MIIMCON[PHY control]) to the TBI block. Refer to the TBI physical address register in Section 14.5.3.1, “TSEC General Control and Status Registers,” and the TBI MII register set in Table 14-98. Note that jitter diagnostics and TBI control are not IEEE 802.3 required registers and are only used for test and control of the TSEC TBI block. The TBI’s TBI control register (TBI) is for configuring the TSEC ten-bit interface block. However, because this TBI block has an MII management interface (just like any other PHY), it has an IEEE 802.3 register called the control register (CR). Table 14-98. TBI MII Register Set Offset Name TEN-BIT INTERFACE (TBI) REGISTERS 0x00 0x01 0x02–0x03 0x04 0x05 0x06 0x07 0x08 0x0F 0x10 0x11 1 Access Size Section/Page 14.5.4/14-91 Control (CR) Status (SR) Reserved AN advertisement (ANA) AN link partner base page ability (ANLPBPA) AN expansion (ANEX) AN next page transmit (ANNPT) AN link partner ability next page (ANLPANP) Extended status (EXST) Jitter diagnostics (JD) TBI control (TBICON) R/W1 R, LH, LL R RW, R R R, LH R/W, R R R R/W R/W 16 bits 16 bits 2 bytes 16 bits 16 bits 16 bits 16 bits 16 bits 16 bits 16 bits 16 bits 14.5.4.2/14-92 14.5.4.3/14-93 — 14.5.4.3/14-93 14.5.4.5/14-96 14.5.4.6/14-98 14.5.4.7/14-98 14.5.4.8/14-99 14.5.4.9/14-100 14.5.4.10/14-101 14.5.4.11/14-102 R = means read only, WO = write only, R/W = read and write, LH = latches high, LL = latches low, SC = self-clearing. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-91 Three-Speed Ethernet Controllers 14.5.4.2 Control Register (CR) Figure 14-100 shows the CR register. 0 1 2 3 4 5 6 7 8 9 10 15 R PHY Reset W Reset Offset 0 Speed_0 AN Enable 00 Reset AN Full Duplex 0 Speed_1 000000 0001_0001_0100_0000 TBIPA offset + 0x00 Figure 14-100. Control Register Definition Table 14-99 describes the fields of the CR register. Table 14-99. CR Field Descriptions Bits 0 Name PHY Reset Description PHY reset. This bit is cleared by default. This bit is self-clearing. 0 Normal operation. 1 The control and status registers are returned to their default value. This in turn may change the internal state of the TBI and its link partner. Reserved Speed selection. This bit defaults to a cleared state and must always be cleared, which corresponds to 1-Gbps speed. Auto-negotiation enable. This bit is set by default. 0 The values programmed in bits 2, 7 and 9 determine the operating condition of the link. 1 Auto-negotiation process enabled Reserved Reset auto-negotiation. This bit is cleared by default and is self-clearing. 0 Normal operation 1 The auto-negotiation process restarts. This action is only available if auto-negotiation is enabled. Duplex mode. This bit is set by default. 0 Half-duplex operation 1 Full-duplex operation Reserved, must be cleared. 1 2 3 — Speed_0 AN Enable 4–5 6 — Reset AN 7 Full Duplex 8 — MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-92 Freescale Semiconductor Memory Map/Register Definition Table 14-99. CR Field Descriptions (continued) Bits 9 Name Speed_1 Description Speed selection. Defaults to a set state and must always be set, which corresponds to 1 Gbps Maximum Operating Speed Reserved Reserved 1 Gbps Reserved 10–15 — Reserved Bit 2 0 1 0 1 Bit 9 0 0 1 1 14.5.4.3 Status Register (SR) Figure 14-101 shows the SR register. 0 6 7 8 9 10 11 12 13 14 15 R W Reset Offset 0 0 0 0 0 0 0 Extend Status 0 No Pre AN Remote AN Link Done Fault Ability Status 0 Extend Ability 0000_0001_0100_1001 TBIPA offset + 0x01 Figure 14-101. Status Register Definition Table 14-100 describes the fields of the SR register. Table 14-100. SR Descriptions Bits 0–6 7 8 9 Name — Extend Status — No Pre AN Done Reserved, must be cleared. Indicates that PHY status information is also contained in the Extended Status Register. Returns 1 on read. This bit is read-only. Reserved, must be cleared. MF preamble suppression enable. This bit indicates whether or not the PHY is capable of handling MII management frames without the 32-bit preamble field. Returns 1, indicating support for suppressed preamble MII management frames. This bit is read-only. Auto-negotiation complete. This bit is read-only and is cleared by default. 0 Either the auto-negotiation process is underway or the auto-negotiation function is disabled. 1 The auto-negotiation process has completed. Description 10 11 Remote Remote fault. This bit is read-only and is cleared by default. Each read of SR clears this bit. Fault 0 Normal operation 1 A remote fault condition was detected. Latches high for software to detect the condition. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-93 Three-Speed Ethernet Controllers Table 14-100. SR Descriptions (continued) Bits 12 Name AN Ability Link Status — Extend Ability Description Auto-negotiation ability. While read as set, this bit indicates that the PHY has the ability to perform auto-negotiation. While read as cleared, this bit indicates the PHY lacks the ability to perform auto-negotiation. Returns 1 on read. This bit is read-only. Link status. This bit is read-only and is cleared by default. 0 A valid link is not established. Latches low allowing for software polling to detect a failure condition. 1 A valid link is established. Reserved, must be cleared. Extended capability. This bit indicates that the PHY contains the extended set of registers (those beyond control and status). Returns 1 on read. This bit is read-only. 13 14 15 14.5.4.4 AN Advertisement Register (ANA) Figure 14-102 shows the ANA register. 0 1 2 3 4 6 7 8 9 10 11 15 R Next Ack W Page Reset Offset Remote Fault 0 0 0 Pause Half Full Duplex Duplex 0 0 0 0 0 0000_0000_0000_0000 TBIPA offset + 0x04 Figure 14-102. AN Advertisement Register Definition Table 14-101 describes the fields of the ANA register. Table 14-101. ANA Field Descriptions Bits 0 Name Next Page Description Next page configuration. The local device sets this bit to either request next page transmission or advertise next page exchange capability. 0 The local device wishes not to engage in next page exchange. 1 The local device has no next pages but wishes to allow reception of next pages. If the local device has no next pages and the link partner wishes to send next pages, the local device shall send null message codes and have the message page set to 0b000_0000_0001, as defined in annex 28C of the IEEE 802.3 specification. Acknowledge. Write 0, ignore on read.This bit is read-only. (Reserved) 1 Ack MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-94 Freescale Semiconductor Memory Map/Register Definition Table 14-101. ANA Field Descriptions (continued) Bits 2–3 Name Remote Fault Description The local device’s remote fault condition is encoded in bits 2 and 3 of the base page. Values are shown in the following table. The default value is 00. Indicate a fault by setting a non-zero remote fault encoding and re-negotiating. RF1 bit[3] 0 0 1 1 4–6 7–8 — Pause Reserved, must be cleared. The local device’s PAUSE capability is encoded in bits 7 and 8, and the decodes are shown in the following table. For priority resolution information consult Table 14-102. PAUSE bit[8] 0 0 1 1 ASM_DIR bit[7] 0 1 0 1 No PAUSE Asymmetric PAUSE toward link partner Symmetric PAUSE Both symmetric PAUSE and Asymmetric PAUSE toward local device Capability RF2 bit[2] 0 1 0 1 Description No error, link OK Offline Link_Failure Auto-Negotiation_Error 9 Half Duplex Full Duplex — Half-duplex capability 0 Designates local device as not capable of half-duplex operation. 1 Designates local device as capable of half-duplex operation. Full-duplex capability 0 Designates the local device as not capable of full-duplex operation. 1 Designates the local device as capable of full-duplex operation. Reserved, must be cleared. 10 11–15 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-95 Three-Speed Ethernet Controllers Table 14-102 describes the resolution of pause priority. Table 14-102. PAUSE Priority Resolution Local Device PAUSE 0 ASM_DIR 0 Link Partner Local Resolution PAUSE x ASM_DIR x Disable PAUSE transmit Disable PAUSE receive Disable PAUSE transmit Disable PAUSE receive Disable PAUSE transmit Disable PAUSE receive Enable PAUSE transmit Disable PAUSE receive Disable PAUSE transmit Disable PAUSE receive Enable PAUSE transmit Enable PAUSE receive Disable PAUSE transmit Disable PAUSE receive Disable PAUSE transmit Enable PAUSE receive Enable PAUSE transmit Enable PAUSE receive Disable PAUSE transmit Disable PAUSE receive Disable PAUSE transmit Disable PAUSE receive Disable PAUSE transmit Disable PAUSE receive Disable PAUSE transmit Enable PAUSE receive Disable PAUSE transmit Disable PAUSE receive Enable PAUSE transmit Enable PAUSE receive Disable PAUSE transmit Disable PAUSE receive Enable PAUSE transmit Disable PAUSE receive Enable PAUSE transmit Enable PAUSE receive Link Partner Resolution 0 1 0 x 0 1 1 0 0 1 1 1 1 0 0 x 1 0 1 x 1 1 0 0 1 1 0 1 1 1 1 x 14.5.4.5 AN Link Partner Base Page Ability Register (ANLPBPA) Figure 14-103 shows the ANLPBPA register. 0 1 2 3 4 6 7 8 9 10 11 15 R Next Ack Page W Reset Offset Remote Fault 0 0 0 Pause Half Full Duplex Duplex 0 0 0 0 0 0000_0000_0000_0000 TBIPA offset + 0x05 Figure 14-103. AN Link Partner Base Page Ability Register Definition MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-96 Freescale Semiconductor Memory Map/Register Definition Table 14-103 describes the fields of the ANLPBPA register. Table 14-103. ANLPBPA Field Descriptions Bits 0 Name Next Page Description Next page. This bit is read-only. The link partner sets or clears this bit. 0 Link partner has no subsequent next pages or is not capable of receiving next pages 1 Link partner either requesting next page transmission or indicating the capability to receive next pages Acknowledge. Ignore on read. This bit is read-only The link partner’s remote fault condition is encoded in bits 2 and 3 of the base page. Values are shown in the remote fault encoding field table below. This bit is read-only. RF1 bit[3] 0 0 1 1 RF2 bit[2] 0 1 0 1 Description No error, link OK Offline Link_Failure Auto-Negotiation_Error 1 2–3 Ack Remote Fault 4–6 7–8 — Pause Reserved, must be cleared. Encoding of the link partner’s PAUSE capability is shown in the PAUSE encoding table below. For priority resolution information consult the IEEE 802.3 specification. This bit is read-only PAUSE bit[8] 0 0 1 1 ASM_DIR bit[7] 0 1 0 1 No PAUSE Asymmetric PAUSE toward link partner Symmetric PAUSE Both symmetric PAUSE and Asymmetric PAUSE toward local device Capability 9 Half Duplex Half-duplex capability. This bit is read-only. 0 Link partner is not capable of half-duplex mode 1 Link partner is capable of half-duplex mode Full-duplex capability. This bit is read-only. 0 Link partner is not capable of full-duplex mode 1 Link partner is capable of full-duplex mode Reserved, must be cleared. 10 Full Duplex 11–15 — MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-97 Three-Speed Ethernet Controllers 14.5.4.6 AN Expansion Register (ANEX) Figure 14-104 shows the ANEX register. 0 12 13 14 15 R W Reset Offset 0 0 0 0 0 0 0 0 0 0 0 0 0 NP Able Page Rx’d 0 0000_0000_0000_0110 TBIPA offset + 0x06 Figure 14-104. AN Expansion Register Definition Table 14-104 describes the fields of the ANEX register. Table 14-104. ANEX Field Descriptions Bits 0–12 13 14 Name — NP Able Page Rx’d Reserved, must be cleared. Next page able. This bit is read-only and returns 1 on read. While read as set, indicates local device supports next page function Page received. This bit is read-only. The bit clears on a read to the register. 0 Normal operation 1 A new page was received and stored in the applicable AN link partner ability or AN next page register. This bit latches high in order for software to detect while polling. Reserved, must be cleared. Description 15 — 14.5.4.7 AN Next Page Transmit Register (ANNPT) Figure 14-105 shows the ANNPT register. 0 1 2 3 4 5 15 R Next Page W Reset Offset Ack Msg Page Ack2 Toggle Message/Unformatted Code Field 0000_0000_0000_0000 TBIPA offset + 0x07 Figure 14-105. AN Next Page Transmit Register Definition MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-98 Freescale Semiconductor Memory Map/Register Definition Table 14-105 describes the fields of the ANNPT register. Table 14-105. ANNPT Field Descriptions Bits 0 Name Description Next Page Next page indication. [Reference MII bit 7.15 in IEEE 802.3, 2000 Edition Clause 28.2.4] 0 Last page 1 Additional next pages to follow Ack Msg Page Acknowledge. Write 0, ignore on read. [Reference MII bit 7.14] This bit is read-only. Message page. [Reference MII bit 7.13] 0 Unformatted page 1 Message page Acknowledge 2. Used by the next page function to indicate that the device has the ability to comply with the message. [Reference MII bit 7.12] 0 The local device cannot comply with message. 1 The local device complies with message. Toggle. Used to ensure synchronization with the link partner during next page exchange. This bit always takes the opposite value of the toggle bit of the previously-exchanged link code word. The initial value in the first next page transmitted is the inverse of bit 11 in the base link code word. [Reference MII bit 7.11] This bit is read-only. 0 Toggle bit of the previously-exchanged link code word was set 1 Toggle bit of the previously-exchanged link code word was clear 1 2 3 Ack2 4 Toggle 5–15 Message/ Message pages are formatted pages that carry a pre-defined message code, which is enumerated Unformatted in IEEE 802.3u/Annex 28C. Unformatted code fields take on an arbitrary value. [Reference MII field Code Field 7.10:0] 14.5.4.8 AN Link Partner Ability Next Page Register (ANLPANP) Figure 14-106 shows the ANLPANP register. 0 1 2 3 4 5 15 R Next Page W Reset Offset Ack Msg Page Ack2 Toggle Message/Unformatted Code Field 0000_0000_0000_0000 TBIPA offset + 0x08 Figure 14-106. AN Link Partner Ability Next Page Register Definition Table 14-106 describes the ANLPANP fields. Table 14-106. ANLPANP Field Descriptions Bits 0 Name Next Page Description Next page. The link partner sets and clears this bit. 0 Last page from link partner 1 Additional next pages to follow Acknowledge. Ignore on read. [Reference MII bit 8.14] This bit is read-only. 1 Ack MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-99 Three-Speed Ethernet Controllers Table 14-106. ANLPANP Field Descriptions (continued) Bits 2 Name Msg Page Message page 0 Unformatted page 1 Message page Acknowledge 2. Indicates the link partner’s ability to comply with the message 0 Link partner cannot comply with message 1 Link partner complies with message Toggle. Used to ensure synchronization with the link partner during next page exchange. This bit always takes the opposite value of the toggle bit of the previously-exchanged link code word. The initial value in the first next page transmitted is the inverse of bit 11 in the base link code word. This bit is read-only. 0 Toggle bit of the previously-exchanged link code word was set 1 Toggle bit of the previously-exchanged link code word was clear Message pages are formatted pages that carry a pre-defined message code, which is enumerated in IEEE 802.3u/Annex 28C. Unformatted code fields take on an arbitrary value. Description 3 Ack2 4 Toggle 5–15 Message/ Unformatted Code Field 14.5.4.9 Extended Status Register (EXST) Figure 14-107 shows the EXST register. 0 1 2 3 4 15 R 1000X 1000X 1000T 1000T Full Half Full Half W Reset Offset 0 0 0 0 0 0 0 0 0 0 0 0 1010_0000_0000_0000 TBIPA offset + 0x0F Figure 14-107. Extended Status Register Definition Table 14-107 describes the fields of the EXST register. Table 14-107. EXST Field Descriptions Bits 0 Name Description 1000X Full 1000X full-duplex capability. Returns 1 on read. This bit is read-only. 0 PHY cannot operate in 1000BASE-X full-duplex mode 1 PHY can operate in 1000BASE-X full-duplex mode 1000X Half 1000X half-duplex capability. Returns 0 on read. This bit is read-only. 0 PHY cannot operate in 1000BASE-X half-duplex mode 1 PHY can operate in 1000BASE-X half-duplex mode 1000T Full 1000T full-duplex capability. Returns 1 on read. This bit is read-only. 0 PHY cannot operate in 1000BASE-T full-duplex mode 1 PHY can operate in 1000BASE-T full-duplex mode 1 2 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-100 Freescale Semiconductor Memory Map/Register Definition Table 14-107. EXST Field Descriptions (continued) Bits 3 Name Description 1000T Half 1000T half-duplex capability. Returns 0 on read. This bit is read-only. 0 PHY cannot operate in 1000BASE-T half-duplex mode 1 PHY can operate in 1000BASE-T half-duplex mode — Reserve 4–15 14.5.4.10 Jitter Diagnostics Register (JD) Annex 36A in IEEE 802.3z describes several jitter test patterns. These can be configured to be sent by writing the jitter diagnostics register. See the register description for more information. It may be wise to auto-negotiate and advertise a remote fault, signaling offline, prior to beginning the test patterns. Figure 14-108 shows the JD register. 0 1 3 4 5 6 15 R Jitter W Enable Reset Offset Jitter Select 0 0 Custom Jitter Pattern 0000_0000_0000_0000 TBIPA offset + 0x10 Figure 14-108. Jitter Diagnostics Register Definition MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-101 Three-Speed Ethernet Controllers Table 14-108 describes the fields of the JD register. Table 14-108. JD Field Description Bits 0 Name Description Jitter Enable Jitter enable. This bit is cleared by default. 0 Normal transmit operation 1 Enable the TBI to transmit the jitter test patterns defined in annex 36A of the IEEE 802.3 specification. Jitter Select Selects the jitter pattern to be transmitted in diagnostics mode. Encoding of this field is shown in the following table. Default is 0 Jitter Pattern Select User defined uses custom jitter pattern, bits 6–15 High frequency (±D21.5) 1010101010101010101010101010101010101010… Mixed frequency (±K28.5) 1111101011000001010011111010110000010100… Low frequency 1111100000111110000011111000001111100000… Complex pattern (10’h17c,10’h0c9,10’h0e5,10’h2a3, 10’h17c,...) Square Wave (-K28.7) 0011111000001111100000111110000011111000… Reserved bit[1] 0 0 0 bit[2] 0 0 1 bit[3] 0 1 0 1–3 0 1 1 1 1 0 0 0 1 1 1 0 4–5 — Reserved 6–15 Custom Jitter Used in conjunction with jitter (pattern) select and jitter (diagnostic) enable; set this field to the Pattern desired custom pattern which is continuously transmitted. Its default is 0. 14.5.4.11 TBI Control Register (TBICON) Figure 14-109 shows the TBICON register. 0 1 2 3 4 6 7 8 9 10 11 12 13 14 15 R Soft_Reset W Reset Offset 0 Disable Disable Rx Dis Tx Dis 0 0 0 AN Sense 0 0 MII Clock Mode Select 0 0 0 0 0000_0000_0000_0000 TBIPA offset + 0x11 Figure 14-109. TBI Control Register Definition MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-102 Freescale Semiconductor Functional Description Table 14-109 describes the fields of the TBICON register. Table 14-109. TBICON Field Descriptions Bits 0 Name Soft_Reset Description Soft reset. This bit is cleared by default. 0 Normal operation 1 Resets the functional modules in theTBI Reserved. (Ignore on read) 1 2 — Disable Rx Dis Disable receive disparity. This bit is cleared by default. 0 Normal operation 1 Disables the running disparity calculation and checking in the receive direction Disable Tx Dis Disable transmit disparity. This bit is cleared by default. 0 Normal operation 1 Disables the running disparity calculation and checking in the transmit direction — AN Sense Reserved Auto-negotiation sense enable. This bit is cleared by default. 0 IEEE 802.3z Clause 37 behavior is desired, which results in the link not completing. 1 Allow the auto-negotiation function to sense either a Gigabit MAC in auto-negotiation bypass mode or an older Gigabit MAC without auto-negotiation capability. If sensed, auto-negotiation complete becomes true; however, the page received is low, indicating no page was exchanged. Management can then act accordingly. Reserved Clock select. This bit is cleared by default. 0 Allow the TBI to accept dual split-phase 62.5-MHz receive clocks. 1 Configure the TBI to accept a 125-MHz receive clock from the SerDes/PHY. The 125-MHz clock must be physically connected to ‘PMA receive clock 0’. This bit describes the configuration mode of the TBI. The user reads a 1 while the TBI is configured in GMII/MII mode (connected to a GMII/MII PHY) and a 0 while configured in TBI mode (connected to a 1000BASE-X SerDes). Its value is the inverse of ECNTRL[TBIM]. 0 TBI mode 1 GMII mode Reserved 3 4–6 7 8–9 10 — Clock Select 11 MII Mode 12–15 — 14.6 Functional Description This section describes many of the functions of the TSEC controller. 14.6.1 Connecting to Physical Interfaces This section describes how to connect the TSEC to various interfaces: MII, GMII, RGMII, RTBI and TBI. To avoid confusion, all of the buses follow the bus conventions used in the IEEE 802.3 specification, because each PHY follows the same convention. (For instance, in the bus TSECn_TXD[7:0], bit 7 is the MSB and bit 0 is the LSB). MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-103 Three-Speed Ethernet Controllers 14.6.1.1 Media-Independent Interface (MII) This section describes the media-independent interface (MII) intended to be used between the PHYs and the TSEC. Figure 14-110 shows the basic components of the MII including the signals required to establish TSEC module connection with a PHY. Transmit Error (TSECn_TX_ER) Transmit Data (TSECn_TXD[3:0]) Transmit Enable (TSECn_TX_EN) Transmit Clock (TSECn_TX_CLK) TSEC Collision Detect (TSECn_COL) Receive Data (TSECn_RXD[3:0]) Medium Receive Error (TSEC n_RX_ER) Receive Clock (TSECn_RX_CLK) Receive Data Valid (TSECn_RX_DV) Carrier Sense Output (TSEC n_CRS) Management Data Clock1 (EC_MDC) Management Data I/O1 (EC_MDIO) 1 10/100 Ethernet PHY The management signals (EC_MDC and EC_MDIO) are common to all of the Ethernet controllers’ connections in the system, assuming that each PHY has a different management address. Figure 14-110. TSEC-MII Connection An MII interface has 18 signals (including the EC_MDC and EC_MDIO signals), as defined by the IEEE 802.3u standard, for connecting to an Ethernet PHY. 14.6.1.2 Gigabit Media-Independent Interface (GMII) This section describes the gigabit media-independent interface (GMII) intended to be used between the PHYs and the TSEC. Figure 14-111 shows the basic components of the GMII including the signals required to establish the TSEC module connection with a PHY. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-104 Freescale Semiconductor Functional Description Gigabit Reference Clock (EC_GTX_CLK125) Transmit Error (TSECn_TX_ER) Transmit Nibble Data (TSECn_TXD[7:0]) Transmit Enable (TSECn_TX_EN) TSEC Gigabit Transmit Clock (TSECn_GTX_CLK) Transmit Clock (TSECn_TX_CLK) Collision Detect (TSECn_COL) Receive Nibble Data (TSECn_RXD[7:0]) Receive Error (TSEC n_RX_ER) Receive Clock (TSECn_RX_CLK) Receive Data Valid (TSEC n_RX_DV) Carrier Sense Output (TSEC n_CRS) Management Data Clock1 (EC_MDC) Management Data I/O1 (EC_MDIO) 1 Gigabit Ethernet PHY Medium The management signals (EC_MDC and EC_MDIO) are common to all of the Ethernet controllers’ connections in the system, assuming that each PHY has a different management address. Figure 14-111. TSEC-GMII Connection A GMII interface has 28 signals (TSECn_GTX_CLK + EC_GTX_CLK125 included), as defined by the IEEE 802.3u standard, for connecting to an Ethernet PHY. 14.6.1.3 Reduced Gigabit Media-Independent Interface (RGMII) This section describes the reduced gigabit media-independent interface (RGMII) intended to be used between the PHYs and the GMII MAC. The RGMII is an alternative to the IEEE802.3u MII, the IEEE802.3z GMII, and the TBI. The RGMII reduces the number of signals required to interconnect the MAC and the PHY from a maximum of 28 signals (GMII) to 15 signals (EC_GTX_CLK125 included) in a cost-effective and technology-independent manner. To accomplish this objective, the data paths and all associated control signals are multiplexed using both edges of the clock. For gigabit operation, the clocks operate at 125 MHz, and for 10/100 operation, the clocks operate at 2.5 MHz or 25 MHz, respectively. Figure 14-112 shows the basic components of the gigabit reduced media-independent interface and the signals required to establish the gigabit Ethernet controllers’ module connection with a PHY. The RGMII is implemented as defined by the RGMII specification Version 1.2a 9/22/00. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-105 Three-Speed Ethernet Controllers Gigabit Reference Clock (EC_GTX_CLK125) Gigabit Transmit Clock (TSECn_GTX_CLK) Transmit Control (TX_EN/f(TX_EN,TX_ER)) Transmit Data (TSEC n_TXD[3:0]/TSECn_TXD[7:4]) Receive Clock (TSECn_RX_CLK) TSEC Receive Control (RX_DV/f(RX_DV,RX_ER)) Receive Data (TSEC n_RXD[3:0]/TSECn_RXD[7:4]) Gigabit Ethernet PHY Medium Management Data Clock1 (EC_MDC) Management Data I/O1 (EC_MDIO) 1 The management signals (EC_MDC and EC_MDIO) are common to all of the gigabit Ethernet controllers module connections in the system, assuming that each PHY has a different management address. Figure 14-112. TSEC-RGMII Connection 14.6.1.4 Ten-Bit Interface (TBI) This section describes the ten-bit interface (TBI) intended to be used between the PHYs and the TSEC to implement a standard SerDes interface for optical-fiber devices in 1000BASE-SX/LX applications. Figure 14-113 shows the basic components of the TBI including the signals required to establish TSEC module connection with a PHY. RBC0 and RBC1 are differential 62.5-MHz receive clocks. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-106 Freescale Semiconductor Functional Description Gigabit Reference Clock (EC_GTX_CLK125) Transmit Data (TSEC n_TXD[9:0]) GigaBit Transmit Clock (TSECn_GTX_CLK) TSEC Receive Data (TSECn_RXD[9:0]) Medium TBI Receive Clock 0 (TSECn_RX_CLK0) TBI Receive Clock 1 (TSECn_RX_CLK1) SIGNAL DETECT (SDET) Gigabit Ethernet PHY Management Data Clock1 (EC_MDC) Management Data I/O1 (EC_MDIO) 1 The management signals (EC_MDC and EC_MDIO) are common to all of the Ethernet controllers’ connections in the system, assuming that each PHY has a different management address. Figure 14-113. TSEC-TBI Connection A TBI interface has 27 signals (EC_GTX_CLK125 included) for connecting to an Ethernet PHY, as defined by IEEE 802.3z GMII and TBI standards. 14.6.1.5 Reduced Ten-Bit Interface (RTBI) This section describes the reduced ten-bit interface (RTBI) intended to be used between the PHYs and the TSEC to implement a reduced signal count version of a SerDes interface for optical-fiber devices in 1000BASE-SX/LX applications. Figure 14-114 shows the basic components of the RTBI including the signals required to establish TSEC module connection with a PHY. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-107 Three-Speed Ethernet Controllers Gigabit Reference Clock (EC_GTX_CLK125) Transmit Data (TSECn_TXD[4:0]/TSECn_TXD[9:5]) GigaBit Transmit Clock (TSECn_GTX_CLK) TSEC Receive Data (TSEC n_RXD[4:0]/TSECn_RXD[9:5]) Receive Clock (TSECn_RX_CLK) Gigabit Ethernet PHY Medium Management Data Clock1 (EC_MDC) Management Data I/O1 (EC_MDIO) 1 The management signals (EC_MDC and EC_MDIO) are common to all of the Ethernet controllers’ connections in the system, assuming that each PHY has a different management address. Figure 14-114. TSEC-RTBI Connection A RTBI interface has 15 signals (EC_GTX_CLK125 included), as defined by the RGMII specification Version 1.2a 9/22/00, and is intended to be an alternative to the IEEE 802.3u MII, the IEEE 802.3z GMII and the TBI standard for connecting to an Ethernet PHY. Table 14-110 describes the signal multiplexing for GMII, MII and TBI interfaces. Table 14-110. GMII, MII, and TBI Signal Multiplexing TSEC Signals Frequency [MHz] 125 Voltage[V] 3.3/2.5 Signals No. of Signals 1 1 1 1 1 1 1 1 1 1 GMII Interface Frequency [MHz] 125 Voltage[V] 3.3 Signals MII Interface Frequency [MHz] 25 Voltage[V] 3.3 No. of Signal s TBI Interface Frequency [MHz] 62.5 Voltage[V] 3.3 Signals No. of Signal s 1 1 1 1 1 1 1 1 1 1 (TSECn_) GTX_CLK TX_CLK TxD[0] TxD[1] TxD[2] TxD[3] TxD[4] TxD[5] TxD[6] TxD[7] I/O (TSEC n_) GTX_CLK TX_CLK TxD[0] TxD[1] TxD[2] TxD[3] TxD[4] TxD[5] TxD[6] TxD[7] I/O No. of Signals Signal (TSECn_) s 1 1 1 1 1 1 1 1 1 1 TX_CLK TxD[0] TxD[1] TxD[2] TxD[3] I/O (TSECn_) GTX_CLK I/O O I O O O O O O O O O I O O O O O O O O O I O O O O O O O O I O O O O 1 1 1 1 1 RX_CLK1 TCG[0] TCG[1] TCG[2] TCG[3] TCG[4] TCG[5] TCG[6] TCG[7] MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-108 Freescale Semiconductor Functional Description Table 14-110. GMII, MII, and TBI Signal Multiplexing (continued) TSEC Signals Frequency [MHz] 125 Voltage[V] 3.3/2.5 TX_EN TX_ER RX_CLK RxD[0] RxD[1] RxD[2] RxD[3] RxD[4] RxD[5] RxD[6] RxD[7] RX_DV RX_ER COL CRS O O I I I I I I I I I I I I I 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 GMII Interface Frequency [MHz] 125 Voltage[V] 3.3 TX_EN TX_ER RX_CLK RxD[0] RxD[1] RxD[2] RxD[3] RxD[4] RxD[5] RxD[6] RxD[7] RX_DV RX_ER O O I I I I I I I I I I I 1 1 1 1 1 1 1 1 1 1 1 1 1 RX_DV RX_ER COL CRS I I I I 1 1 1 1 SDET I 1 MII Interface Frequency [MHz] 25 Voltage[V] 3.3 TX_EN TX_ER RX_CLK RxD[0] RxD[1] RxD[2] RxD[3] O O I I I I I 1 1 1 1 1 1 1 TBI Interface Frequency [MHz] 62.5 Voltage[V] 3.3 TCG[8] TCG[9] RX_CLK0 RCG[0] RCG[1] RCG[2] RCG[3] RCG[4] RCG[5] RCG[6] RCG[7] RCG[8] RCG[9] O O I I I I I I I I I I I 1 1 1 1 1 1 1 1 1 1 1 1 1 Sum 25 Sum 23 Sum 16 Sum 24 Table 14-111 describes the signal multiplexing for RGMII and RTBI interfaces. Table 14-111. RGMII and RTBI Signal Multiplexing TSEC Signals Frequency [MHz] 125 Voltage[V] 3.3/2.5 Signals No. of Signal s 1 1 1 1 1 1 TxD[0]/TxD[4] TxD[1]/TxD[5] TxD[2]/TxD[6] TxD[3]/TxD[7] O O O O 1 1 1 1 TCG[0]/TCG[5] TCG[1]/TCG[6] TCG[2]/TCG[7] TCG[3]/TCG[8] O O O O 1 1 1 1 RGMII Interface Frequency [MHz] 125 Voltage[V] 2.5 Signals No. of Signal s 1 RTBI Interface Frequency [MHz] 62.5 Voltage[V] 2.5 Signals No. of Signals 1 (TSECn_) GTX_CLK TX_CLK TxD[0] TxD[1] TxD[2] TxD[3] I/O (TSECn_) GTX_CLK I/O (TSECn_) GTX_CLK I/O O I O O O O O O MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-109 Three-Speed Ethernet Controllers Table 14-111. RGMII and RTBI Signal Multiplexing (continued) TSEC Signals Frequency [MHz] 125 Voltage[V] 3.3/2.5 Signals No. of Signal s 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 RX_CTL (RX_DV/RX_ERR) I 1 RCG[4]/RCG[9] I 1 RX_CLK RxD[0]/RxD[4] RxD[1]/RxD[5] RxD[2]/RxD[6] RxD[3]/RxD[7] I I I I I 1 1 1 1 1 RX_CLK RCG[0]/RCG[5] RCG[1]/RCG[6] RCG[2]/RCG[7] RCG[3]/RCG[8] I I I I I 1 1 1 1 1 TX_CTL (TX_EN/TX_ERR) O 1 TCG[4]/TCG[9] O 1 RGMII Interface Frequency [MHz] 125 Voltage[V] 2.5 Signals No. of Signal s RTBI Interface Frequency [MHz] 62.5 Voltage[V] 2.5 Signals No. of Signals (TSECn_) TxD[4] TxD[5] TxD[6] TxD[7] TX_EN TX_ER RX_CLK RxD[0] RxD[1] RxD[2] RxD[3] RxD[4] RxD[5] RxD[6] RxD[7] RX_DV RX_ER COL CRS I/O (TSECn_) I/O (TSECn_) I/O O O O O O O I I I I I I I I I I I I I Sum 25 Sum 12 Sum 12 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-110 Freescale Semiconductor Functional Description Table 14-112 describes the signals shared by all interfaces. Table 14-112. Shared Signals Signals EC_MDIO EC_MDC EC_GTX_CLK125 I/O I/O O I # of signals 1 1 1 Function Management interface I/O Management interface clock Reference clock Sum 3 14.6.2 Gigabit Ethernet Channel Operation This section describes the operation of the TSEC. First, the software initialization sequence is described. Next, the software (Ethernet driver) interface for transmitting and receiving frames is reviewed. Address recognition and hash table algorithm features are also discussed. The section concludes with interrupt handling, inter-packet gap time, and loop back descriptions. 14.6.2.1 Initialization Sequence This section describes which registers are reset due to a hard or software reset and what registers the user must initialize prior to enabling the TSEC. 14.6.2.1.1 Hardware-Controlled Initialization A hard reset occurs when the system powers up. All TSEC registers and control logic are reset to their default states after a hard reset has occurred. 14.6.2.1.2 User Initialization After the system has undergone a hard reset, software must initialize certain basic TSEC registers. Other registers can also be initialized during this time, but they are optional and must be determined based on the requirements of the system. The module memory map in Table 14-3 lists all the TSEC registers. Table 14-113 describes the minimum steps for register initialization. Table 14-113. Steps of Minimum Register Initialization Description 1. Set, then clear MACCFG1 [Soft_Reset] 2. Initialize MACCFG2 3. Initialize MAC station address MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-111 Three-Speed Ethernet Controllers Table 14-113. Steps of Minimum Register Initialization (continued) Description 4. Set up the PHY using the MII Mgmt Interface 5. Configure the TBI control to TBI or GMII 6. Clear IEVENT 7. Initialize IMASK 8. Initialize IADDRn 9. Initialize GADDRn 10. Initialize RCTRL 11. Initialize DMACTRL After the registers are initialized, the user must execute the following steps in the order described below to bring the TSEC into a functional state (out of reset): 1. For the transmission of Ethernet frames, TxBDs must first be built in memory, linked together as a ring, and pointed to by the TBASE register. A minimum of two buffer descriptors per ring is required. Setting the ring to a size of one causes the same frame to be transmitted twice. 2. Likewise, for the reception of Ethernet frames, the receive queue must be ready, with its RxBD pointed to by the RBASE register. Both transmit and receive can be gracefully stopped after transmission and reception begins. 3. Write to MACCFG1 and set the appropriate bits. These need to include Rx_EN and Tx_EN. To enable flow control, Rx_Flow and Tx_Flow must also be set. 4. Clearing DMACTRL[GTS] triggers the transmission of frame data if the transmitter had been previously stopped. DMACTRL[GRS] must be cleared if the receiver had been previously stopped. See Section 14.5.3.1.7, “DMA Control Register (DMACTRL),” and Section 14.6.3.1, “Transmit Data Buffer Descriptor (TxBD),” for more information. 14.6.2.2 Soft Reset and Reconfiguring Procedure Before issuing a soft reset to and/or reconfiguring the MAC with new parameters, user must properly shutdown the DMA and make sure it is in an idle state for the entire duration. User must gracefully stop the DMA by setting both GRS and GTS bits in the DMACTRL register, then wait for both GRSC and GTSC bits to be set in the IEVENT register before resetting the MAC or changing parameters. Both GRS and GTS bits must be cleared before re-enabling the MAC to resume the DMA. During the MAC configuration, if a new set of Tx buffer descriptors are used, the user must load the pointers into the TBASE register. Likewise if a new set of Rx buffer descriptors are used, the RBASE register must be written with the new pointer. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-112 Freescale Semiconductor Functional Description Following is a procedure to gracefully reset and reconfigure the MAC: 1. Set GTS bit in DMACTRL register. 2. Poll GTSC bit in IEVENT register until detected as set. 3. Clear both Rx_EN and Tx_EN bits in MACCFG1. 4. Wait for a period of 9.6 Kbytes worth of data on the interface (~8ms worst case). 5. Set GRS bit in DMACTRL register. 6. Poll GRSC bit in IEVENT register until detected as set. 7. Set Soft_Reset bit in MACCFG1 register. 8. Clear Soft_Reset bit in MACCFG1 register. 9. Load TBASE with new TxBD pointer. 10. Load RBASE with new RxBD pointer. 11. Set up other MAC registers (MACCFG2, MAXFRM, and so on). 12. Set WWR and WOP bits in DMACTRL register. 13. Clear THLT bit in TSTAT register and QHLT bit in RSTAT register by writing 1 to these bits. 14. Clear GRS/GTS bits in DMACTRL. (Do not change other bits.) 15. Enable Tx_EN/Rx_EN in MACCFG1 register. 14.6.2.3 Gigabit Ethernet Frame Transmission The Ethernet transmitter requires little core intervention. After the software driver initializes the system, the TSEC begins to poll the first transmit buffer descriptor (TxBD) in the TxBD ring every 512 transmit clocks. If TxBD[R] is set, TSEC begins moving transmit buffer from memory to its Tx FIFO. The transmitter takes data from the Tx FIFO and transmits data to the MAC. The MAC transmits the data through the GMII interface to the physical media. The transmitter, once initialized, runs until the end-of-frame (EOF) condition is detected unless a collision within the collision window occurs (half-duplex mode) or an abort condition is encountered. If the user has a frame ready to transmit, a transmit-on-demand function may be emulated while in polling mode by using the graceful-transmit-stop feature. First, clear the IMASK[GTSCEN] bit to mask the graceful-transmit-stop complete interrupt. Next set, then immediately clear the DMACTRL[GTS] bit. Clear the resulting IEVENT[GTSC] bit. Finally, the IMASK[GTSCEN] bit may be set once again. There is one internal buffer for out-of-sequence flow control frames. While the TSEC is between frames, this buffer is polled. The buffer must contain the whole frame. Once the TSEC is in paused mode, the out-of-sequence buffer descriptor cannot be used to send another flow control frame because the MAC regards it as a regular TxBD. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-113 Three-Speed Ethernet Controllers In half-duplex mode (MACCFG2[Full Duplex] is cleared) the MAC defers transmission if the line is busy (CRS asserted). Before transmitting, the MAC waits for carrier sense to become inactive, at which point it then determines if CRS remains negated for 60 clocks. If so, transmission begins after an additional 36 bit times (96 bit times after CRS originally became negated). If CRS continues to be asserted, the MAC follows a specified back-off procedure and tries to retransmit the frame until the retry limit is reached. Data stored in the Tx FIFO is re-transmitted in case of a collision. This improves bus usage and latency. The transmitter also monitors for an abort condition and terminates the current frame if an abort condition is encountered. In full-duplex mode the protocol is independent of network activity, and only the transmit inter-frame gap must be enforced. The transmit block also implements full-duplex flow control. If a flow control frame is received, the MAC does not service the transmitter’s request to send data until the pause duration is over. If the MAC is currently sending data when a pause frame is received, the MAC finishes sending the current frame, then suspends subsequent frames (except a pause frame) until the pause duration is over. The pause duration is defined by the received pause control frame and begins when the frame was first received. In addition, the transmitter supports transmission of flow control frames via TCTRL[TFC_PAUSE]. The transmit pause frame is generated internally based on the PAUSE register that defines the pause value to be sent. Note that it is possible to send a pause frame while the pause timer has not expired. The MAC automatically appends FCS (32-bit CRC) bytes to the frame if any of the following values are set: • • • • TxBD[PAD/CRC] is set in first TxBD TxBD[TC] is set in first TxBD MACCFG2[PAD/CRC] is set MACCFG2[CRC EN] is set Following the transmission of the FCS, the Ethernet controller writes the frame status bits into the BD and clears TxBD[R]. If the end of the current buffer is reached and TxBD[L] is cleared (a frame is comprised of multiple buffer descriptors), only TxBD[R] is cleared. For both half- and full-duplex modes, an interrupt can be issued depending on TxBD[I]. The Ethernet controller then proceeds to the next TxBD in the table. In this way, the core can be interrupted after each frame, after each buffer, or after a specific buffer is sent. If TxBD[PAD/CRC] is set, the Ethernet controller pads any frame shorter than 64 bytes. To pause transmission or rearrange the transmit queue, set DMACTRL[GTS]. This can help in transmitting expedited data ahead of previously linked buffers or for error situations. If GTS is set, the TSEC transmitter performs a graceful transmit stop. The Ethernet controller stops immediately if no transmission is in progress or continues transmission until the current frame either finishes MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-114 Freescale Semiconductor Functional Description or terminates with an error. The IEVENT[GTSC] interrupt occurs once the graceful transmit stop operation is completed. After GTS is cleared, the TSEC resumes transmission with the next frame. While the TSEC is in 10/100 Mbps mode it sends bytes least-significant nibble first and each nibble is sent lsb first. While it is in 1-Gbps mode it sends bytes lsb first. 14.6.2.4 Gigabit Ethernet Frame Reception The TSEC Ethernet receiver is designed to work with little core intervention and can perform address recognition, CRC checking, short frame checking, and maximum frame-length checking. The receiver can also force frame headers and buffer descriptors to be allocated into the L2 cache. See Section 14.6.4, “Data Extraction to the L2 Cache,” for additional information. After a hardware reset, the software driver clears the RSTAT register and sets MACCFG1[RX_EN]. The Ethernet receiver is enabled and immediately starts processing receive frames. If TSECn_RX_DV is asserted and TSECn_COL remains negated, the MAC strips a valid preamble/SFD (start of frame delimiter) header and begins processing the frame. If a valid header is not found, the frame is ignored. If the receiver detects the first bytes of a frame, the TSEC controller begins to perform the frame recognition function through destination address (DA) recognition (See Section 14.6.2.6, “Frame Recognition,” for additional information.). Based on this match the frame can be accepted or rejected. Once accepted, the TSEC processes the frame based on user-defined attributes. The receiver can also filter frames based on physical (individual), group (multicast), and broadcast addresses. Because Ethernet receive frame data is not written to memory until the internal frame recognition algorithm is complete, system bus usage is not wasted on frames unwanted by this station. If a frame is accepted, the Ethernet controller fetches the receive buffer descriptor (RxBD) from the queue. If RxBD is not being used by software (RxBD[E] is set), the TSEC starts transferring the incoming frame. RxBD[F] is set for the first RxBD used for any particular receive frame. If the buffer is filled, the TSEC controller clears RxBD[E] and, if RxBD[I] is set, generates an interrupt. If the incoming frame is larger than the buffer, the Ethernet controller fetches the next RxBD in the table. If it is empty, the controller continues receiving the rest of the frame. In half-duplex mode, if a collision is detected during the frame, no RxBDs are used; thus, no collision frames are presented to the user except late collisions, which indicate LAN problems. The RxBD length is determined by the MRBL field in the maximum receive buffer length register (MRBL). The smallest valid value is 64 bytes. During reception, the Ethernet controller checks for frames that are too short or too long. After the frame ends (CRS is negated), the receive CRC field is checked and written to the data buffer. The data length written to the last RxBD in the Ethernet frame is the length of the entire frame, which enables the software to recognize a frame-too-long condition. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-115 Three-Speed Ethernet Controllers Receive frames are not truncated if they exceed maximum frame bytes in the MAC’s maximum frame register if MACCFG2[Huge Frame] is set, yet the babbling receiver error interrupt occurs (IEVENT[BABR] is set) and RxBD[LG] is set. After the receive frame is complete, the Ethernet controller sets RxBD[L], updates the frame status bits in the RxBD, and clears RxBD[E]. If RxBD[I] is set, the Ethernet controller next generates an interrupt (that can be masked) indicating that a frame was received and is in memory. The Ethernet controller then waits for a new frame. To interrupt reception or rearrange the receive queue, DMACTRL[GRS] must be set. If this bit is set, the TSEC receiver performs a graceful receive stop. The Ethernet controller stops immediately if no frames are being received or continues receiving until the current frame either finishes or an error condition occurs. The IEVENT[GRSC] interrupt event is signalled after the graceful receive stop operation is completed. While in this mode the user can then clear IEVENT[GRSC] and can write to registers that are accessible to both the user and the TSEC hardware without fear of conflict. After DMACTRL[GRS] is cleared, the TSEC scans the input data stream for the start of a new frame (preamble sequence and start of frame delimiter), it resumes receiving, and the first valid frame received is placed in the next available RxBD. 14.6.2.5 RMON Support TSEC automatically gathers network statistics required for RMON without needing to receive all addresses. The RMON MIB group 1, RMON MIB group 2, RMON MIB group 3, RMON MIB group 9, RMON MIB 2, and the 802.3 Ethernet MIB are supported. For RMON statistics and their corresponding counters see the memory map. 14.6.2.6 Frame Recognition The Ethernet controller performs frame recognition using destination address (DA) recognition. A frame can be rejected or accepted based on the outcome. 14.6.2.6.1 Destination Address Recognition The Ethernet controller can also perform the frame filtering using the traditional destination address (DA) recognition methods. Figure 14-115 is a flowchart for address recognition on received frames that is used to explain the concept. In the actual implementation most of the decision points shown in the figure actually occur simultaneously. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-116 Freescale Semiconductor Functional Description Incoming Frame G I/G Address I Broadcast Address T F T Station Address Match? F Hash Search (Use Group Table) Hash Search (Use Individual Table) T Hash Match? F Broadcast Reject T F F T Promiscuous? Discard Frame Receive Frame Figure 14-115. Ethernet Address Recognition Flowchart The Ethernet controller compares the destination address field of the received frame with the physical address the user programs in the station address registers (MACSTNADDR1 and MACSTNADDR2). If the DA does not match the station address, then the controller performs address recognition on multiple individual addresses using the IADDRn hash table. The user must MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-117 Three-Speed Ethernet Controllers write zeros to the hash in order to avoid a hash match, and ones to station address in order to avoid individual address match, or the user can turn on promiscuous mode. (See Section 14.5.3.4.1, “Receive Control Register (RCTRL).”) In the group type of address recognition, the Ethernet controller determines whether the group address is a broadcast address. If it is a broadcast, and broadcast addresses are enabled, the frame is accepted. If the group address is not a broadcast address, the user can perform address recognition on multiple group addresses using the GADDRn hash table. In promiscuous mode, the Ethernet controller receives all of the incoming frames regardless of their address. 14.6.2.6.2 Hash Table Algorithm The hash table process used in the individual and group hash filtering operates as follows. The Ethernet controller maps any 48-bit destination address into one of 256 bins, represented by the 256 bits in GADDR0–7 or IADDR0–7. The eight high-order bits of a cyclic redundancy check (CRC) checksum are used to index into the hash table. The high-order three bits of this 8-bit field are used to select one of the eight registers in either the individual or group hash table. The low-order five bits select a bit within the 32-bit register. A value of 0 in the high-order three bits selects IADDR0/GADDR0. The same process is used if the Ethernet controller receives a frame. If the CRC checksum selects a bit that is set in the group/individual hash table, the frame is accepted. If 32 group addresses are stored in the hash table and random group addresses are received, the hash table prevents roughly 224/256 (87.5%) of the group address frames from reaching memory. Software must further filter those that reach memory to determine if they contain the correct addresses. Better performance is achieved by using the group and individual hash tables in combination. For instance, if 32 group and 32 physical addresses are stored in their respective hash tables, because 87.5% of all group addresses and 87.5% of all individual address are rejected, then 87.5% of all frames are prevented from reaching memory. The effectiveness of the hash table declines as the number of addresses increases. For instance, as the number of addresses stored in the 256-bin hash table increases, the vast majority of the hash table bits are set, preventing only a small fraction of frames from reaching memory. 14.6.2.6.3 CRC Computation Examples There are many algorithms for calculating the CRC value of a number. Refer to the RFC 3309 standard, which can be found at http://www.faqs.org/rfcs/rfc3309.html, to compute the CRC value for the purposes of TSEC. The RFC 3309 algorithm uses the following polynomial to calculate the CRC value: x32+x26+x23+x22+x16+x12+x11+x10+x8+x7+x5+x4+x2+x1+x0 or 0x04c11db7. Given a destination MAC address of DA=01000CCCCCCC, the algorithm results in a CRC remainder value of 0xA29F4BBC. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-118 Freescale Semiconductor Functional Description Bit-reversing the low-order byte of the CRC value (0xBC) yields BR_CRC = 0x3D = 0b00111101 The high-order 3-bits of the new BR_CRC value are used to select which 32-bit register (of the 8) to use. This example maps the DA to register 1. High-order 3 bits of BR_CRC: HO_CRC = 0b001 = 1 The low-order 5 bits are used to select which bit to set in the given register (with a value of 0 setting 0x8000_0000 and 31 setting 0x0000_0001). Therefore, the example DA maps to bit 29 of register 1. Low-order 5 bits of BR_CRC: LO_CRC = 0b11101 = 29 Therefore, GADDR1 is ORed with the value 0x0000_0004. Additional calculated examples follow: Example 1: • • • • • • • • • • • • Destination MAC address: DA = 01005E000128 CRC remainder value: CRC = 0x821D6CD3 Bit-reversed least-significant byte of CRC value: BR_CRC = 0xCB = 0b11001011 High-order 3 bits of BR_CRC: HO_CRC = 0b110 = 6 Low-order 5 bits of BR_CRC: LO_CRC = 0b01011 = 11 GADDR6 = 0x0010_0000 Destination MAC address: DA = 0004F0604F10 CRC remainder value: CRC = 0x1F5A66B5 Bit-reversed least-significant byte of CRC value: BR_CRC = 0xAD = 0b10101101 High-order 3 bits of BR_CRC: HO_CRC = 0b101 = 5 Low-order 5 bits of BR_CRC: LO_CRC = 0b01101 = 13 GADDR5 = 0x0004_0000 Example 2: 14.6.2.7 Flow Control Because collisions cannot occur in full-duplex mode, gigabit Ethernet can operate at the maximum rate. If the rate becomes too fast for a station’s receiver, the station’s transmitter can send flow-control frames to reduce the rate. Flow-control instructions are transferred by special frames of minimum frame size. The length/type fields of these frames have a special value. Table 14-114 lists the flow-control frame structure. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-119 Three-Speed Ethernet Controllers Table 14-114. Flow Control Frame Structure Size [Octets] 7 1 6 6 2 2 2 Description Preamble SFD Destination address Source address Length/type MAC opcode MAC parameter 88-08 00-01 Control frame type Pause command Pause time as defined by the PTV[PT] field. The pause period is measured in pause_quanta, a speed-independent constant of 512 bit-times (unlike slot time). The most-significant octet is sent first. Extended pause control parameter as defined by the PTV[PTE] field. The most-significant octet is sent first. — Frame check sequence (CRC) Start frame delimiter 01-80C2-00-00-01 Multicast address reserved for use in MAC frames Value Comment 2 40 4 Extended MAC parameter Reserved FCS If flow-control mode is enabled (MACCFG1[Rx_Flow] is set) and the receiver identifies a pause-flow control frame, transmission stops for the time specified in the control frame. During this pause, only a control frame can be sent (TCTRL[TFC_PAUSE] is set). Normal transmission resumes after the pause timer stops counting. If another pause-control frame is received during the pause, the period changes to the new value received. 14.6.2.8 Interrupt Handling The following describes what usually occurs within a TSEC interrupt handler: • • If an interrupt occurs, read IEVENT to determine interrupt sources. IEVENT bits to be handled in this interrupt handler are normally cleared at this time. Process the TxBDs to reuse them if the IEVENT[TXB or TXF] were set. If the transmit speed is fast or the interrupt delay is long, more than one transmit buffer may have been sent by the TSEC; thus, it is important to check more than just one TxBD during the interrupt handler. One common practice is to process all TxBDs in the interrupt handler until one is found with R set. See Table 14-115. Obtain data from the RxBD if IEVENT[RXC,RXB or RXF] is set. If the receive speed is fast or the interrupt delay is long, the TSEC may have received more than one RxBD; thus, it is important to check more than just one RxBD during interrupt handling. Typically, all RxBDs in the interrupt handler are processed until one is found with E set. Because the TSEC pre-fetches BDs, the BD table must be big enough so that there is always another empty BD to pre-fetch. See Table 14-116. • MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-120 Freescale Semiconductor Functional Description • • Clear any set halt bits in TSTAT and RSTAT registers, or DMACTRL[GTS] and DMACTRL[GRS]. Continue normal execution. Table 14-115. Non-Error Transmit Interrupts Interrupt GTSC Description Graceful transmit stop complete: transmitter is put into a pause state after completion of the frame currently being transmitted. Transmit control: Instead of the next transmit frame, a control frame was sent. Action taken by TSEC None TXC TXB None Transmit buffer: A transmit buffer descriptor, that is not the Programmable “write with response” last one in the frame, was updated. TxBD to memory before setting IEVENT[TXB]. Transmit frame: A frame was transmitted and the last transmit buffer descriptor (TxBD) of that frame was updated. Programmable “write with response” to memory on the last TxBD before setting IEVENT[TXF]. TXF Table 14-116. Non-Error Receive Interrupts Interrupt GRSC Description Graceful receive stop complete: Receiver is put into a pause state after completion of the frame currently being received. Receive control: A control frame was received. As soon as the transmitter finishes sending the current frame, a pause operation is performed lasting for the duration specified in the received pause control frame and beginning when the frame was first received. Receive buffer: A receive buffer descriptor, that is not the last one of the frame, was updated. Action taken by TSEC None RXC None RXB Programmable “write with response” RxBD to memory before setting IEVENT[RXB]. RXF Receive frame: A frame was received and the last receive Programmable “write with response” to buffer descriptor (RxBD) of that frame was updated. memory on the last RxBD before setting IEVENT[RXF]. 14.6.2.8.1 Interrupt Coalescing Interrupt coalescing offers the user the ability to contour the behavior of the TSEC with regard to frame interrupts. Separate but identical mechanisms exist for handling both transmitted frames and received frames. While interrupt coalescing is enabled a transmit or receive frame interrupt (resulting from the interrupt bit (I) of the buffer descriptor in question and appropriately-enabled by the IMASK register) is raised either when a counter threshold-defined number of frames is received/transmitted or the timer threshold-defined period of time has lapsed, whichever occurs first. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-121 Three-Speed Ethernet Controllers 14.6.2.8.2 Interrupt Coalescing By Frame Count Threshold To avoid interrupt bandwidth congestion due to frequent, consecutive interrupts, the user may enable and configure interrupt coalescing to deliberately group frame interrupts, reducing the total number of interrupts raised. The number of frames received or transmitted prior to an interrupt being raised is determined by the frame threshold field (ICFCT) in the appropriate interrupt coalescing configuration register (RXIC or TXIC). The frame threshold field may be assigned a value between 1 and 255. A value of 0 results in boundedly undefined behavior. Note that a value of 1 functionally defeats the advantages of interrupt coalescing since the frame threshold is reached with each frame received or transmitted. Once the number of frames transmitted or received reaches the threshold limit, an interrupt is raised (if the interrupt bit of the frame buffer descriptor(s) is set and if appropriately-enabled in the IMASK register), the threshold counter is reset, and then continues counting frames while the interrupt is active. The threshold counter is also reset if an interrupt is raised subject to the corresponding threshold timer. 14.6.2.8.3 Interrupt Coalescing By Timer Threshold To avoid stale frame interrupts, the user may also assign a timer threshold, beyond which any frame interrupts not yet raised are forced (if the interrupt bit of the frame buffer descriptor(s) is set and if enabled in the IMASK register). The timer threshold fields of the receive and transmit interrupt coalescing configuration registers (RXIC[ICTT] and TXIC[ICTT]) are defined in units equivalent to 64 TSEC interface clocks. That is, one timer threshold unit is 26.5 µs, 2.56 µs, or 512 ns, corresponding to interface modes 10 Mbps, 100 Mbps, or 1 Gbps, respectively. After transmitting a frame (with TXBD[I] set, causing IEVENT[TXF] to be set), the transmit interrupt coalescing threshold timer begins counting. When the timer threshold has been reached an interrupt is raised if the interrupt bit of the buffer descriptor is set and IMASK[TXFEN] is set. After receiving a frame (with RXBD[I] set, causing IEVENT[RXF] to be set), the receive interrupt coalescing threshold timer begins counting. When the timer threshold has been reached an interrupt is raised if the interrupt bit of the buffer descriptor is set and IMASK[RXFEN] is set. The interrupt coalescing timer thresholds (transmit and receive, operating independently) may be values ranging from 1 to 65535. A value of 0 results in boundedly undefined behavior. Table 14-117 specifies the range of possible timing thresholds subject to the interface frequency and the value of RXIC[ICTT] or TXIC[ICTT]. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-122 Freescale Semiconductor Functional Description Table 14-117. Interrupt Coalescing Timing Threshold Ranges TSEC Interface Format and Frequency 10Base-T at 2.5 MHz 100Base-T at 25 MHz 1000Base-T at 125 MHz Interrupt Coalescing Threshold Time Minimum (ICTT=1) 25.6 µs 2.56 µs 512 ns Maximum (ICTT=65535) 1.678 s 167.8 ms 33.55 ms The transmit or receive timer threshold counter is reset when a corresponding interrupt is raised and begins counting again upon deassertion of that interrupt and once IEVENT[TXF or RXF] is set. 14.6.2.9 Inter-Packet Gap Time If a station must transmit, it waits until the LAN becomes silent for a specified period (inter-packet gap). After a station begins sending, it continually checks for collisions on the LAN. If a collision is detected, the station forces a jam signal (all ones) on its frame and stops transmitting. Collisions usually occur close to the beginning of a packet. The station then waits a random time period (back-off) before attempting to send again. After the back-off completes, the station waits for silence on the LAN and then begins retransmission on the LAN. This process is called a retry. If the packet is not successfully sent within a specified number of retries, an error is indicated. The minimum inter-packet gap time for back-to-back transmission is 96 serial clocks. The receiver receives back-to-back packets with this minimum spacing. In addition, after waiting a required number of clocks (based on the back-off algorithm), the transmitter waits for carrier sense to be negated before retransmitting the packet. Retransmission begins 36 serial clocks after carrier sense is negated for at least 60 serial clocks. 14.6.2.10 Internal and External Loop Back Setting MACCFG1[Loop Back] causes the MAC transmit outputs to be looped back to the MAC receive inputs. Clearing this bit results in normal operation. This bit is cleared by default. 14.6.2.11 Error-Handling Procedure The Ethernet controller reports frame reception and transmission error conditions using the channel BDs, the error counters, and the IEVENT register. Programming note: When the TSEC encounters a halt condition (TSTAT[THLT] is set), it stops processing the frame at the current TxBD. The TSEC relies on the user to manage the buffer descriptor pointer, TBPTR, or the buffer descriptor queue before resuming transmissions. Once the TSEC resumes, it fetches the TxBD pointed to by TBPTR. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-123 Three-Speed Ethernet Controllers Transmission errors are described in Table 14-118. Table 14-118. Transmission Errors Error Transmitter underrun Response The controller sends 32 bits that ensure a CRC error, terminates buffer transmission, sets TxBD[UN], closes the buffer, sets IEVENT[XFUN] and IEVENT[TXE]. The controller resumes transmission after TSTAT[THLT] is cleared (and DMACTRL[GTS] is cleared). Retransmission The controller terminates buffer transmission, sets TxBD[RL], closes the buffer, sets IEVENT[CRL/XDA] and IEVENT[TXE]. Transmission resumes after TSTAT[THLT] is cleared (and attempts limit DMACTRL[GTS] is cleared). expired Excessive defer abort Late collision The controller terminates buffer transmission, sets TxBD[DEF], closes the buffer, sets IEVENT[CRC/XDA], and IEVENT[TXE]. Transmission resumes after TSTAT[THLT] is cleared. The controller terminates buffer transmission, sets TxBD[LC], closes the buffer, sets IEVENT[LC] and IEVENT[TXE]. The controller resumes transmission after TSTAT[THLT] is cleared (and DMACTRL[GTS] is cleared). A system bus error occurred during a DMA transaction. The controller sets IEVENT[EBERR], DMA stops sending data to the FIFO which causes an underrun error but IEVENT[XFUN] is not set. The TSTAT[THLT] is set. Transmits are continued once TSTAT[THLT] is cleared. A frame is transmitted which exceeds the MAC’s Maximum Frame Length and MACCFG2[Huge Frame] is a 0. The controller sets IEVENT[BABT] and continues without interruption. TxBD[TXTRUNC] is set in the last TxBD (TxBD[L] is set) of the frame. Memory Read Error Babbling Transmit Error Reception errors are described in Table 14-119. Table 14-119. Reception Errors Error Overrun error Description The Ethernet controller maintains an internal FIFO buffer for receiving data. If a receiver FIFO buffer overrun occurs, the controller sets RxBD[OV], sets RxBD[L], closes the buffer, and sets IEVENT[RXF], The receiver then enters hunt mode (seeking start of a new frame). A frame is received and discarded due to a lack of buffers. The controller sets IEVENT[BSY]. In addition, the RSTAT[QHLT] bit is set. The halted queue resumes reception once the user clears the RSTAT[QHLT] bit. Busy error Non-octet error The Ethernet controller handles a nibble of dribbling bits if the receive frame terminates as non-octet (dribbling bits) aligned and it checks the CRC of the frame on the last octet boundary. If there is a CRC error, the frame non-octet aligned (RxBD[NO]) error is reported, IEVENT[RXF] is set, and the alignment error counter increments. The TSEC relies on the statistics collector block to increment the receive alignment error counter (RALN). If there is no CRC error, no error is reported. CRC error If a CRC error occurs, the controller sets RxBD[CR], closes the buffer, and sets IEVENT[RXF]. This TSEC relies on the statistics collector block to record the event. After receiving a frame with a CRC error, the receiver then enters hunt mode. A system bus error occurred during a DMA transaction. The controller sets IEVENT[EBERR] and discards the frame. In addition the RSTAT[QHLT] bit is set. The halted queue resumes reception once the RSTAT[QHLT] bit is cleared. A frame is received that exceeds the MAC’s maximum frame length. The controller sets IEVENT[BABR] and continues Memory Read Error Babbling Receive Error MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-124 Freescale Semiconductor Functional Description 14.6.3 Buffer Descriptors The TSEC buffer descriptor (BD) is modeled after the MPC8260 Fast Ethernet controller BD for ease of reuse. Drawing from the MPC8260 FEC BD programming model, the TSEC descriptor base registers point to the beginning of BD rings. The 8-byte data BD format is similar to the MPC8260 BD model. Data buffers are used in the transmission and reception of Ethernet frames (see Figure 14-116). Data BDs encapsulate all information necessary for the TSEC to transmit or receive an Ethernet frame. Within each data BD there is a status field, a data length field, and a data pointer. The BD completely describes an Ethernet packet by centralizing status information for the data packet in the status field of the BD and by containing a data BD pointer to the location of the data buffer. Software is responsible for setting up the BDs in memory. Because of pre-fetching, a minimum of two buffer descriptors per ring are required. This applies to both the transmit and the receive descriptor rings. Software also must have the data pointer pointing to memory. Within the status field, there exists an ownership bit which defines the current state of the buffer (pointed to by the data pointer). Other bits in the status field in the buffer descriptor are used to communicate status/control information between the TSEC and the software driver. The status field of the BD is 16-bit field, as is the length field. The data buffer pointer is a 32-bit field. Therefore, the BDs should be accessed with the following C structure: typedef unsigned short uint_16; /* choose 16-bit native type */ typedef unsigned int uint_32; /* choose 32-bit native type */ typedef struct bd_struct { uint_16 flags; uint_16 length; uint_32 bufptr; }; Because there is no next BD pointer in the transmit/receive BD (see Figure 14-117), all BDs must reside sequentially in memory. The TSEC increments the current BD location appropriately to the next BD location to be processed. There is a wrap bit in the last BD that informs the TSEC to loop back to the beginning of the BD chain. Software must initialize TBASE and RBASE that point to the beginning transmit and receive BDs for the TSEC. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-125 Three-Speed Ethernet Controllers Memory Map System Memory Tx Buffer Descriptors Status & Control TxBD Table Pointer (TBASE) RxBD Table Pointer (RBASE) TxBD Table Data Length Buffer Pointer Tx Buffer Rx Buffer Descriptors RxBD Table Status & Control Data Length Buffer Pointer Rx Buffer Figure 14-116. Example of TSEC Memory Structure for BD Beginning BD pointer 0 1 4 W =1 2 3 Figure 14-117. Buffer Descriptor Ring 14.6.3.1 Transmit Data Buffer Descriptor (TxBD) Data is presented to the TSEC for transmission by arranging it in memory buffers referenced by the TxBDs. In the TxBD the user initializes the R, PAD/CRC, W, L, and TC bits and the length (in bytes) in the first word, and the buffer pointer in the second word. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-126 Freescale Semiconductor Functional Description The TSEC clears the R bit in the first word of the BD after it finishes using the data buffer. The transfer status bits are then updated. Additional transmit frame status can be found in statistic counters in the MIB block. Figure 14-118 shows the TxBD. 0 1 2 3 4 5 6 7 8 9 10 13 14 15 Offset + 0 R Offset + 2 Offset + 4 PAD/CRC W IL TC DEF TO1 HFE/LC RL RC UN TXTRUNC Data Length TX Data Buffer Pointer Offset + 6 Figure 14-118. Transmit Buffer Descriptor The TxBD fields are detailed in Table 14-120. Table 14-120. Transmit Data Buffer Descriptor (TxBD) Field Descriptions Offset Offset + 0 Bits 0 Name R Description Ready. Written by TSEC and user 0 The data buffer associated with this BD is not ready for transmission. The user is free to manipulate this BD or its associated data buffer. The TSEC clears this bit after the buffer is transmitted or after an error condition is encountered. 1 The data buffer, which is prepared for transmission by the user, was not transmitted or is currently being transmitted. No fields of this BD may be written by the user once this bit is set. PAD/CRC. Padding and CRC attachment for frames. (Valid only while it is set in the first BD and MACCFG2[PAD/CRC] is cleared.) If MACCFG2[PAD/CRC] is set, this bit is ignored. 0 Do not add padding to short frames. No CRC is appended unless TxBD[TC] is set. 1 Add PAD/CRCs to frames. PAD bytes are inserted until the length of the transmitted frame equals 64 bytes. Unlike the MPC8260 which PADs up to MINFLR value, TSEC PADs always up to the IEEE minimum frame length of 64 bytes. CRC is always appended to frames. Wrap. Written by user 0 The next buffer descriptor is found in the consecutive location. 1 The next buffer descriptor is found at the location defined in TBASE. Interrupt. Written by user 0 No interrupt is generated after this buffer is serviced. 1 IEVENT[TXB] or IEVENT[TXF] are set after this buffer is serviced. These bits can cause an interrupt if they are enabled (That is, IEVENT[TXBEN] or IEVENT[TXFEN] are set). Last in frame. Written by user 0 The buffer is not the last in the transmit frame. 1 The buffer is the last in the transmit frame. Tx CRC. Written by user. (Valid only while it is set in first BD and TxBD[PAD/CRC] is cleared and MACCFG2[PAD/CRC] is cleared and MACCFG2[CRC EN] is cleared.) If MACCFG2[PAD/CRC] is set or MACCFG2[CRC EN] is set, this bit is ignored. 0 End transmission immediately after the last data byte with no hardware generated CRC appended, unless TxBD[PAD/CRC] is set. 1 Transmit the CRC sequence after the last data byte. Offset + 0 1 PAD/ CRC Offset + 0 2 W Offset + 0 3 I Offset + 0 4 L Offset + 0 5 TC MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-127 Three-Speed Ethernet Controllers Table 14-120. Transmit Data Buffer Descriptor (TxBD) Field Descriptions Offset Offset + 0 Bits 6 Name DEF Description Defer indication. Hardware updates this bit if an excessive defer condition occurs. 0 This frame was not deferred. 1 If HAFDUP[EXCESS_DEFER]=1, this frame did not have a collision before it was sent but it was sent late because of deferring. If HAFDUP[EXCESS_DEFER]=0, this frame was aborted and not sent. Transmit software ownership. This read/write bit may be utilized by software, as necessary. Its state does not affect the hardware nor is it affected by the hardware. Huge Frame Enable (written by user)/Late collision (written by TSEC) Valid only while it is set in first BD and the MACCFG2[Huge Frame] is cleared. If MACCFG2[Huge Frame] is set, this bit is ignored. 0 Truncate transmit frame if its length is greater than the MAC’s Maximum Frame Length register. 1 Do not truncate the transmit frame. Late collision. Written by TSEC 0 No late collision 1 A collision occurred after 64 bytes are sent. The TSEC terminates the transmission and updates LC. Offset + 0 9 RL Retransmission Limit. Written by TSEC 0 Transmission before maximum retry limit is hit 1 The transmitter failed (max. retry limit + 1) attempts to successfully send a message due to repeated collisions. The TSEC terminates the transmission and updates RL. Retry Count. Written by TSEC 0000 The frame is sent correctly the first time or if RL is set then the retry limit has been reached nnnn One or more attempts were needed to send the transmit frame. If this field is 15, then 15 or more retries were needed. The Ethernet controller updates RC after sending the buffer. Underrun. Written by TSEC 0 No underrun encountered (data was retrieved from external memory in time to send a complete frame) 1 The Ethernet controller encountered a transmitter underrun condition while sending the associated buffer. The TSEC terminates the transmission and updates UN. Offset + 0 Offset + 0 7 8 TO1 HFE/LC Offset + 0 10–13 RC Offset + 0 14 UN Offset + 0 Offset + 2 Offset + 4 15 0–15 0–31 TXTRUNC TX truncation. Set in the last TxBD (TxBD[L] is set) when IEVENT[BABT] occurs for the frame Data Length TX Data Buffer Pointer Data length is the number of octets the TSEC transmits from this BD’s data buffer. It is never modified by the TSEC. This field must be greater than zero. The transmit buffer pointer contains the address of the associated data buffer. There are no alignment requirements for this address. 14.6.3.2 Receive Buffer Descriptor (RxBD) In the RxBD the user initializes the E, I, and W bits in the first word and the pointer in second word. If the data buffer is used, the TSEC modifies the E, L, F, M, BC, MC, LG, NO, SH, CR, OV, and TR bits and writes the length of the used portion of the buffer in the first word. The M, BC, MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-128 Freescale Semiconductor Functional Description MC, LG, NO, SH, CR, OV, and TR bits in the first word of the buffer descriptor are only modified by the TSEC if the L bit is set. The first word of the RxBD contains control and status bits. Its format is detailed in Figure 14-119 below. 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Offset + 0 Offset + 2 Offset + 4 E RO1 W I L F 0 M BC MC LG NO SH CR OV TR Data Length RX Data Buffer Pointer Offset + 6 Figure 14-119. Receive Buffer Descriptor Table 14-121 describes the fields of the RxBD. Table 14-121. Receive Buffer Descriptor Field Descriptions Offset Offset + 0 Bits 0 Name E Description Empty. Written by TSEC (when cleared) and by user (when set). 0 The data buffer associated with this BD is filled with received data, or data reception is aborted due to an error condition. The status and length fields have been updated as required. 1 The data buffer associated with this BD is empty, or reception is currently in progress. Receive software ownership bit. Reserved for use by software. This read/write bit is not modified by hardware, nor does its value affect hardware. Wrap. Written by user 0 The next buffer descriptor is found in the consecutive location. 1 The next buffer descriptor is found at the location defined in RBASE. Interrupt. Written by user 0 No interrupt is generated after this buffer is serviced. 1 IEVENT[RXB] or IEVENT[RXF] are set after this buffer is serviced. This bit can cause an interrupt if enabled (IMASK[RXBEN] is set or IMASK[RXFEN] is set). If the user wants to be interrupted only if RXF occurs, then the user must disable RXB (IMASK[RXBEN] is cleared) and enable RXF (IMASK[RXFEN] is set). Last in frame. Written by TSEC 0 The buffer is not the last in a frame. 1 The buffer is the last in a frame. First in frame. Written by TSEC 0 The buffer is not the first in a frame. 1 The buffer is the first in a frame. Reserved Miss. Written by TSEC. (This bit is valid only if the L-bit is set and TSEC is in promiscuous mode.) This bit is set by the TSEC for frames that were accepted in promiscuous mode, but were flagged as a ‘miss’ by the internal address recognition; thus, while in promiscuous mode, the user can use the M-bit to quickly determine whether the frame was destined to this station. 0 The frame was received because of an address recognition hit. 1 The frame was received because of promiscuous mode. Offset + 0 Offset + 0 1 2 RO1 W Offset + 0 3 I Offset + 0 4 L Offset + 0 5 F Offset + 0 Offset + 0 6 7 — M MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-129 Three-Speed Ethernet Controllers Table 14-121. Receive Buffer Descriptor Field Descriptions (continued) Offset Offset + 0 Offset + 0 Offset + 0 Bits 8 9 10 Name BC MC LG Description Broadcast. Written by TSEC. (Only valid if L is set.) Is set if the DA is broadcast (FF-FF-FF-FF-FF-FF) Multicast. Written by TSEC. (Only valid if L is set.) Is set if the DA is multicast and not BC Rx frame length violation. Written by TSEC. (Only valid if L is set.) A frame length greater than maximum frame length was recognized while MACCFG2[Huge Frame] was set. Note, if MACCFG2[Huge Frame] is cleared, the frame is truncated to the value programmed in the Maximum Frame Length register. Rx non-octet aligned frame. Written by TSEC. (Only valid if L is set.) A frame that contained a number of bits not divisible by eight was received. Short frame. Written by TSEC. (only valid if L is set.) A frame length that was less than the minimum length defined for this channel (MINFLR) was recognized, provided RCTRL[RSF] is set. Rx CRC error. Written by TSEC. (Only valid if L is set.) This frame contains a CRC error and is an integral number of octets in length.This bit is also set if a receive code group error is detected. Overrun. Written by TSEC. (Only valid if L is set.) A receive FIFO overrun occurred during frame reception. If this bit is set, the other status bits, M, LG, NO, SH, CR, and CL lose their normal meaning and are zero. Truncation. Written by TSEC. (Only valid if L is set.) Is set if the receive frame is truncated. This can happen if a frame length greater than maximum frame length was received and the MACCFG2[Huge Frame] is cleared. If this bit is set, the frame must be discarded and the other error bits must be ignored as they may be incorrect. Data length. Written by TSEC. Data length is the number of octets written by the TSEC into this BD’s data buffer if L is cleared (the value is equal to MRBL), or the length of the frame including CRC, if L is set. Offset + 0 Offset + 0 11 12 NO SH Offset + 0 13 CR Offset + 0 14 OV Offset + 0 15 TR Offset + 2 0–15 Data Length Offset + 4 0–31 Rx Data Receive buffer pointer. Written by user. The receive buffer pointer, which always points to the Buffer first location of the associated data buffer, must be 64-byte aligned. The buffer must reside in Pointer memory external to the TSEC. 14.6.4 Data Extraction to the L2 Cache Some applications require the ability to identify selected portions of data within a frame’s data payload. This process is called extraction; although, the data is not truly removed. Rather than literarily extracting a section of the data and copying it into a new memory location, the data is placed in the L2 cache. This allows the processor to quickly access critical frame information as soon as the processor is ready without having to first fetch the data from main memory. This reduction in latency can result in a substantial improvement in packet processing throughput. Extraction functionality is controlled and configured with ATTR and ATTRELI. See Section 14.5.3.9.1, “Attribute Register (ATTR),” and Section 14.5.3.9.2, “Attribute Extract Length and Extract Index Register (ATTRELI),” for specific register information. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-130 Freescale Semiconductor Initialization/Application Information 14.7 Initialization/Application Information 14.7.1 Interface Mode Configuration This section describes how to configure the TSEC in different supported interface modes. These include MII, GMII, TBI, RGMII, RTBI. The pinout, the data registers that must be initialized, as well as speed selection options are described. The ECNTRL[TBIM] and ECNTRL[RPM] bits are written, assuming the part was not pin-configured at initialization to the correct mode. 14.7.1.1 MII Interface Mode Table 14-122 describes the signal configurations required for MII interface mode. Table 14-122. MII Interface Mode Signal Configuration TSEC Signals Frequency [MHz] 125 Voltage[V] 3.3/2.5 Signals GTX_CLK TX_CLK TxD[0] TxD[1] TxD[2] TxD[3] TxD[4] TxD[5] TxD[6] TxD[7] TX_EN TX_ER RX_CLK RxD[0] RxD[1] RxD[2] RxD[3] RxD[4] RxD[5] I/O O I O O O O O O O O O O I I I I I I I No. of Signals 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 MII Interface Frequency [MHz] 25 Voltage[V] 3.3 Signals Leave Unconnected TX_CLK TxD[0] TxD[1] TxD[2] TxD[3] Leave Unconnected Leave Unconnected Leave Unconnected Leave Unconnected TX_EN TX_ER RX_CLK RxD[0] RxD[1] RxD[2] RxD[3] not used not used I/O O I O O O O O O O O O O I I I I I I I 1 1 1 1 1 1 1 1 1 1 1 1 No. of Signals MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-131 Three-Speed Ethernet Controllers Table 14-122. MII Interface Mode Signal Configuration (continued) TSEC Signals Frequency [MHz] 125 Voltage[V] 3.3/2.5 Signals RxD[6] RxD[7] RX_DV RX_ER COL CRS Sum I/O I I I I I I No. of Signals 1 1 1 1 1 1 25 MII Interface Frequency [MHz] 25 Voltage[V] 3.3 Signals not used not used RX_DV RX_ER COL CRS Sum I/O I I I I I I 1 1 1 1 16 No. of Signals Table 14-123 describes the shared signals of the MII interface. Table 14-123. Shared MII Signals TSEC Signals EC_MDIO EC_MDC EC_GTX_CLK125 I/O I/O O I No. of signals 1 1 1 MII Signals EC_MDIO EC_MDC not used I/O I/O O I No. of signals 1 1 0 Function Management interface I/O Management interface Clock Reference Clock Sum 3 Sum 2 Table 14-124 describes the register initializations required to reconfigure the PHY to MII mode following initial auto-negotiation. Table 14-124. MII Mode Register Initialization Steps Have the TSECn_GTX_CLK configuration signal pulled low for G/MII mode. Set Soft_Reset, MACCFG1[1000_0000_0000_0000_0000_0000_0000_0000] Clear Soft_Reset, MACCFG1[0000_0000_0000_0000_0000_0000_0000_0000] Initialize MACCFG2, for MII, half duplex operation. Set I/F Mode bit, MACCFG2[0000_0000_0000_0000_0111_0001_0000_0100] (This example has Full Duplex = 0, Preamble count = 7, PAD/CRC append = 1) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-132 Freescale Semiconductor Initialization/Application Information Table 14-124. MII Mode Register Initialization Steps (continued) Initialize ECNTRL, ECNTRL[0000_0000_0000_0000_0001_0000_0000_0000] (This example has Statistics Enable = 1) Initialize MAC Station Address, MACSTNADDR2[0110_0000_0000_0010_0000_0000_0000_0000] Set station address to 02_60_8C_87_65_43, for example. (Note that PHY configuration register addresses are not necessarily consistent in all PHYs.) Initialize MAC Station Address, MACSTNADDR1[0100_0011_0110_0101_1000_0111_1000_1100] Set station address to 02_60_8C_87_65_43, for example. (Note that PHY configuration register addresses are not necessarily consistent in all PHYs.) Assign a Physical address to the TBI so as to not conflict with the external PHY Physical address, TBIPA[0000_0000_0000_0000_0000_0000_0000_0101] Set to 05, for example. (Note that PHY configuration register addresses are not necessarily consistent in all PHYs.) Reset the management interface. MIIMCFG[1000_0000_0000_0000_0000_0000_0000_0000] Setup the MII Mgmt clock speed, MIIMCFG[0000_0000_0000_0000_000_0000_0000_0101] set source clock divide by 14, for example, to insure that EC_MDC clock speed is approximately 2.5 MHz. Read MII Mgmt Indicator register and check for Busy = 0, MIIMIND ---> [0000_0000_0000_0000_0000_0000_0000_0000] This indicates that the TSEC MII Mgmt bus is idle. Set up the MII Mgmt for a write cycle to the external PHY Auxiliary Control and Status Register to configure the PHY through the Management interface (overrides configuration signals of the PHY). MIIMADD[0000_0000_0000_0000_0000_0000_0001_1100] Perform an MII Mgmt write cycle to the external PHY Writing to MII Mgmt Control with 16-bit data intended for the external PHY register, MIIMCON[0000_0000_0000_0000_0000_0000_0000_0100] Check to see if MII Mgmt write is complete Read MII Mgmt Indicator register and check for Busy = 0, MIIMIND ---> [0000_0000_0000_0000_0000_0000_0000_0000] This indicates that the write cycle was completed. Set up the MII Mgmt for a write cycle to the external PHY Extended PHY control register 1 to set up the interface mode selection. MIIMADD[0000_0000_0000_0000_0000_0000_0001_0111] Perform an MII Mgmt write cycle to the external PHY. Write to MII Mgmt Control with 16-bit data intended for the external PHY register, MIIMCON[0000_0000_0000_0000_0000_0000_0000_0000] Check to see if MII Mgmt write is complete. Read MII Mgmt Indicator register and check for Busy = 0, MIIMIND ---> [0000_0000_0000_0000_0000_0000_0000_0000] This indicates that the write cycle was completed. Set up the MII Mgmt for a write cycle to the external PHY Mode control register to set up the interface mode selection. MIIMADD[0000_0000_0000_0000_0000_0000_0000_0000] MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-133 Three-Speed Ethernet Controllers Table 14-124. MII Mode Register Initialization Steps (continued) Perform an MII Mgmt write cycle to the external PHY. Write to MII Mgmt Control with 16-bit data intended for the external PHY register, MIIMCON[0000_0000_0000_0000_00uu_00uu_0u00_0000] where u is user defined based on desired configuration. Check to see if MII Mgmt write is complete Read MII Mgmt Indicator register and check for Busy = 0, MIIMIND ---> [0000_0000_0000_0000_0000_0000_0000_0000] This indicates that the write cycle was completed. If auto-negotiation was enabled in the PHY, check to see if PHY has completed Auto-Negotiation. Set up the MII Mgmt for a read cycle to PHY MII Mgmt register (write the PHY address and Register address), MIIMADD[0000_0000_0000_0000_0000_0000_0000_0001] The PHY Status register is at address 0x1 and in this case the PHY Address is 0x00. Perform an MII Mgmt read cycle of Status Register. Clear MIIMCOM[Read Cycle]. Set MIIMCOM[Read Cycle]. (Uses the PHY address (0) and Register address (1) placed in MIIMADD register), When MIIMIND[BUSY]=0, read the MIIMSTAT register and check bit 10 (AN Done and Link is up) MIIMSTAT ---> [0000_0000_0000_0000_0000_000_0010_0100] Other information about the link is also returned.(Extend Status, No pre, Remote Fault, An Ability, Link status, extend Ability) Check auto-negotiation attributes in the PHY as necessary. Clear IEVENT register, IEVENT[0000_0000_0000_0000_0000_0000_0000_0000] Initialize IMASK, (Optional) IMASK[0000_0000_0000_0000_0000_0000_0000_0000] Initialize IADDRn, (Optional) IADDRn[0000_0000_0000_0000_0000_0000_0000_0000] Initialize GADDRn, (Optional) GADDRn[0000_0000_0000_0000_0000_0000_0000_0000] Initialize RCTRL, (Optional) RCTRL[0000_0000_0000_0000_0000_0000_0000_0000] Initialize DMACTRL, (Optional) DMACTRL[0000_0000_0000_0000_0000_0000_0000_0000] Initialize FIFO_PAUSE_CTRL, FIFO_PAUSE_CTRL[0000_0000_0000_0000_0000_0000_0000_0010] Initialize (Empty) Transmit Descriptor ring and fill buffers with Data. Initialize TBASE, TBASE[LLLL_LLLL_LLLL_LLLL_LLLL_LLLL_LLLL_L000] Initialize (Empty) Receive Descriptor ring and fill with empty buffers. Initialize RBASE, RBASE[LLLL_LLLL_LLLL_LLLL_LLLL_LLLL_LLLL_L000] Enable Rx and Tx, MACCFG1[0000_0000_0000_0000_0000_0000_0000_0101] MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-134 Freescale Semiconductor Initialization/Application Information 14.7.1.2 GMII Interface Mode Table 14-125 describes the signal configurations required for GMII interface mode. Table 14-125. GMII Interface Mode Signal Configuration TSEC SIGNALS Frequency [MHz] 125 Voltage[V] 3.3/2.5 Signals GTX_CLK TX_CLK TxD[0] TxD[1] TxD[2] TxD[3] TxD[4] TxD[5] TxD[6] TxD[7] TX_EN TX_ER RX_CLK RxD[0] RxD[1] RxD[2] RxD[3] RxD[4] RxD[5] RxD[6] RxD[7] RX_DV RX_ER COL CRS Sum I/O O I O O O O O O O O O O I I I I I I I I I I I I I No. of Signals 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 25 Sum 22 TxD[0] TxD[1] TxD[2] TxD[3] TxD[4] TxD[5] TxD[6] TxD[7] TX_EN TX_ER RX_CLK RxD[0] RxD[1] RxD[2] RxD[3] RxD[4] RxD[5] RxD[6] RxD[7] RX_DV RX_ER O O O O O O O O O O I I I I I I I I I I I 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 GMII INTERFACE Frequency [MHz] 125 Voltage[V] 3.3 Signals GTX_CLK I/O O No. of Signals 1 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-135 Three-Speed Ethernet Controllers Table 14-126 describes the shared signals of the GMII interface. Table 14-126. Shared GMII Signals TSEC Signals EC_MDIO EC_MDC EC_GTX_CLK125 Sum I/O I/O O I No. of Signals 1 1 1 3 GMII Signals EC_MDIO EC_MDC EC_GTX_CLK125 Sum I/O I/O O I No. of Signals 1 1 1 3 Function Management interface I/O Management interface Clock Reference Clock Table 14-127 describes the register initializations required to reconfigure the PHY to GMII mode following initial auto-negotiation. Table 14-127. GMII Mode Register Initialization Steps Have the TSECn_GTX_CLK configuration signal pulled low for G/MII mode. Set Soft_Reset, MACCFG1[1000_0000_0000_0000_0000_0000_0000_0000] Clear Soft_Reset, MACCFG1[0000_0000_0000_0000_0000_0000_0000_0000] Initialize MACCFG2, for GMII, Full duplex operation. Set I/F Mode bit. MACCFG2[0000_0000_0000_0000_0111_0010_0000_0101] (This example has Full Duplex = 1, Preamble count = 7, PAD/CRC append = 1) Initialize ECNTRL, ECNTRL[0000_0000_0000_0000_0001_0000_0000_0000] (This example has Statistics Enable = 1) Initialize MAC Station Address, MACSTNADDR2[0110_0000_0000_0010_0000_0000_0000_0000] Set station address to 02_60_8C_87_65_43, for example. (Note that PHY configuration register addresses are not necessarily consistent in all PHYs.) Initialize MAC Station Address, MACSTNADDR1[0100_0011_0110_0101_1000_0111_1000_1100] Set station address to 02_60_8C_87_65_43, for example. (Note that PHY configuration register addresses are not necessarily consistent in all PHYs.) Assign a Physical address to the TBI so as to not conflict with the external PHY Physical address, TBIPA[0000_0000_0000_0000_0000_0000_0000_0101] Set to 05, for example. (Note that PHY configuration register addresses are not necessarily consistent in all PHYs.) Reset the management interface, MIIMCFG[1000_0000_0000_0000_0000_0000_0000_0000] Setup the MII Mgmt clock speed, MIIMCFG[0000_0000_0000_0000_0000_0000_0000_0101] Set source clock divide by 14, for example, to insure that EC_MDC clock speed is approximately 2.5 MHz. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-136 Freescale Semiconductor Initialization/Application Information Table 14-127. GMII Mode Register Initialization Steps (continued) Read MII Mgmt Indicator register and check for Busy = 0, MIIMIND ---> [0000_0000_0000_0000_0000_0000_0000_0000] This indicates that the TSEC MII Mgmt bus is idle. Set up the MII Mgmt for a write cycle to the external PHY Auxiliary Control and Status Register to configure the PHY through the Management interface (overrides configuration signals of the PHY), MIIMADD[0000_0000_0000_0000_0000_0000_0001_1100] Perform an MII Mgmt write cycle to the external PHY. Write to MII Mgmt Control with 16-bit data intended for the external PHY register, MIIMCON[0000_0000_0000_0000_0000_0000_0000_0100] Check to see if MII Mgmt write is complete. Read MII Mgmt Indicator register and check for Busy = 0, MIIMIND ---> [0000_0000_0000_0000_0000_0000_0000_0000] This indicates that the write cycle was completed Set up the MII Mgmt for a write cycle to the external PHY Extended PHY control register 1 to set up the interface mode selection MIIMADD[0000_0000_0000_0000_0000_0000_0001_0111] Perform an MII Mgmt write cycle to the external PHY. Write to MII Mgmt Control with 16-bit data intended for the external PHY register, MIIMCON[0000_0000_0000_0000_0000_0000_0000_0000] Check to see if MII Mgmt write is complete. Read MII Mgmt Indicator register and check for Busy = 0, MIIMIND ---> [0000_0000_0000_0000_0000_0000_0000_0000] This indicates that the write cycle was completed. Set up the MII Mgmt for a write cycle to the external PHY Mode control register to set up the interface mode selection, MIIMADD[0000_0000_0000_0000_0000_0000_0000_0000] Perform an MII Mgmt write cycle to the external PHY. Write to MII Mgmt Control with 16-bit data intended for the external PHY register, MIIMCON[0000_0000_0000_0000_000u_00u1_0100_0000] where u is user defined based on desired configuration. Check to see if MII Mgmt write is complete. Read MII Mgmt Indicator register and check for Busy = 0, MIIMIND ---> [0000_0000_0000_0000_0000_0000_0000_0000] This indicates that the write cycle was completed. If auto-negotiation was enabled in the PHY, check to see if PHY has completed Auto-Negotiation. Set up the MII Mgmt for a read cycle to PHY MII Mgmt register (write the PHY address and Register address), MIIMADD[0000_0000_0000_0000_0000_0000_0000_0001] The PHY Status register is at address 0x1 and in this case the PHY Address is 0x00 Perform an MII Mgmt read cycle of Status Register. Clear MIIMCOM[Read Cycle]. Set MIIMCOM[Read Cycle]. (Uses the PHY address (0) and Register address (1) placed in MIIMADD register), When MIIMIND[BUSY]=0, Read the MIIMSTAT register and check bit 10 (AN Done and Link is up), MIIMSTAT ---> [0000_0000_0000_0000_0000_000_0010_0100] Other information about the link is also returned.(Extend Status, No pre, Remote Fault, An Ability, Link status, extend Ability) Check auto-negotiation attributes in the PHY as necessary. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-137 Three-Speed Ethernet Controllers Table 14-127. GMII Mode Register Initialization Steps (continued) Clear IEVENT register, IEVENT[0000_0000_0000_0000_0000_0000_0000_0000] Initialize IMASK, (Optional) IMASK[0000_0000_0000_0000_0000_0000_0000_0000] Initialize IADDRn, (Optional) IADDRn[0000_0000_0000_0000_0000_0000_0000_0000] Initialize GADDRn, (Optional) GADDRn[0000_0000_0000_0000_0000_0000_0000_0000] Initialize RCTRL, (Optional) RCTRL[0000_0000_0000_0000_0000_0000_0000_0000] Initialize DMACTRL, (Optional) DMACTRL[0000_0000_0000_0000_0000_0000_0000_0000] Initialize FIFO_PAUSE_CTRL, FIFO_PAUSE_CTRL[0000_0000_0000_0000_0000_0000_0000_0010] Initialize (Empty) Transmit Descriptor ring and fill buffers with Data. Initialize TBASE, TBASE[LLLL_LLLL_LLLL_LLLL_LLLL_LLLL_LLLL_L000] Initialize (Empty) Receive Descriptor ring and fill with empty buffers. Initialize RBASE, RBASE[LLLL_LLLL_LLLL_LLLL_LLLL_LLLL_LLLL_L000] Enable Rx and Tx, MACCFG1[0000_0000_0000_0000_0000_0000_0000_0101] 14.7.1.3 TBI Interface Mode Table 14-128 describes the signal configurations required for TBI interface mode. Table 14-128. TBI Interface Mode Signal Configuration TSEC Signals Frequency [MHz] 125 Voltage[V] 3.3/2.5 Signals GTX_CLK TX_CLK TxD[0] TxD[1] TxD[2] TxD[3] TxD[4] I/O O I O O O O O No. of Signals 1 1 1 1 1 1 1 TBI Interface Frequency [MHz] 62.5 Voltage[V] 3.3 Signals GTX_CLK RX_CLK1 TCG[0] TCG[1] TCG[2] TCG[3] TCG[4] I/O O I O O O O O No. of Signals 1 1 1 1 1 1 1 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-138 Freescale Semiconductor Initialization/Application Information Table 14-128. TBI Interface Mode Signal Configuration (continued) TSEC Signals Frequency [MHz] 125 Voltage[V] 3.3/2.5 Signals TxD[5] TxD[6] TxD[7] TX_EN TX_ER RX_CLK RxD[0] RxD[1] RxD[2] RxD[3] RxD[4] RxD[5] RxD[6] RxD[7] RX_DV RX_ER COL CRS Sum I/O O O O O O I I I I I I I I I I I I I No. of Signals 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 25 SDET Sum TBI Interface Frequency [MHz] 62.5 Voltage[V] 3.3 Signals TCG[5] TCG[6] TCG[7] TCG[8] TCG[9] RX_CLK0 RCG[0] RCG[1] RCG[2] RCG[3] RCG[4] RCG[5] RCG[6] RCG[7] RCG[8] RCG[9] I/O O O O O O I I I I I I I I I I I I I 1 24 No. of Signals 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 Table 14-129 describes the shared signals for the TBI interface. Table 14-129. Shared TBI Signals TSEC Signals EC_MDIO EC_MDC EC_GTX_CLK125 Sum I/O I/O O I No. of Signals 1 1 1 3 TBI Signals EC_MDIO EC_MDC EC_GTX_CLK12 5 Sum I/O I/O O I No. of Signals 1 1 1 3 Function Management interface I/O Management interface Clock Reference Clock MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-139 Three-Speed Ethernet Controllers Table 14-130 describes the register initializations required to reconfigure the PHY to TBI mode following initial auto-negotiation. Table 14-130. TBI Mode Register Initialization Steps Have the TSECn_GTX_CLK configuration signal pulled high and EC_MDC signal pulled high for TBI mode. (TSEC attempts to auto-negotiate after system reset.) Set Soft_Reset, MACCFG1[1000_0000_0000_0000_0000_0000_0000_0000] Clear Soft_Reset, MACCFG1[0000_0000_0000_0000_0000_0000_0000_0101] Initialize MACCFG2, MACCFG2[0000_0000_0000_0000_0111_0010_0000_0101] (I/F Mode = 2, Full Duplex = 1) Initialize ECNTRL, ECNTRL[0000_0000_0000_0000_0001_0000_0000_0000] (This example has Statistics Enable = 1) Initialize MAC Station Address MACSTNADDR2[0110_0000_0000_0010_0000_0000_0000_0000] to 02608C:876543, for example. (Note that PHY configuration register addresses are not necessarily consistent in all PHYs.) Initialize MAC Station Address MACSTNADDR1[0100_0011_0110_0101_1000_0111_1000_1100] to 02608C:876543, for example. (Note that PHY configuration register addresses are not necessarily consistent in all PHYs.) Assign a Physical address to the TBI, TBIPA[0000_0000_0000_0000_0000_0000_0001_0000] set to 16, for example. (Note that PHY configuration register addresses are not necessarily consistent in all PHYs.) Setup the MII Mgmt clock speed, MIIMCFG[0000_0000_0000_0000_0000_0000_0000_0111] Set source clock divide by 28, for example, to insure that EC_MDC clock speed is not greater than 2.5 MHz. Read MII Mgmt Indicator register and check for Busy = 0, MIIMIND ---> [0000_0000_0000_0000_0000_0000_0000_0000] This indicates that the TSEC MII Mgmt bus is idle. Set up the MII Mgmt for a read cycle to TBI Control register (write the TBI address and Register address), MIIMADD[0000_0000_0000_0000_0001_0000_0000_0000] The TBI Control register is at offset address 0x0 from TBIPA. Perform an MII Mgmt read cycle to verify state of TBI Control Register (Optional) Clear MIIMCOM[Read Cycle] Set MIIMCOM[Read Cycle] (Uses the TBI address and Register address placed in MIIMADD register), When MIIMIND[BUSY]=0, read the MIIMSTAT and look for AN Enable and other bit information. Set up the MII Mgmt for a write cycle to TBI’s AN Advertisement register (write the PHY address and Register address), MIIMADD[0000_0000_0000_0000_0001_0000_0000_0100] The AN Advertisement register is at offset address 0x04 from the TBI’s address. (in this case 0x10) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-140 Freescale Semiconductor Initialization/Application Information Table 14-130. TBI Mode Register Initialization Steps (continued) Perform an MII Mgmt write cycle to TBI. Writing to MII Mgmt Control with 16-bit data intended for TBI’s AN Advertisement register, MIIMCON[0000_0000_0000_0000_0000_0001_1010_0000] This advertises to the Link Partner that the TBI supports PAUSE and Full Duplex mode and does not support Half Duplex mode. Check to see if MII Mgmt write is complete. Read MII Mgmt Indicator register and check for Busy = 0, MIIMIND ---> [0000_0000_0000_0000_0000_0000_0000_0000] This indicates that the write cycle was completed. Set up the MII Mgmt for a write cycle to TBI’s Control register (write the PHY address and Register address), MIIMADD[0000_0000_0000_0000_0001_0000_0000_0000] the Control register is at offset address 0x00 from the TBI’s address. (in this case 0x10) Perform an MII Mgmt write cycle to TBI. Writing to MII Mgmt Control with 16-bit data intended for TBI’s Control register, MIIMCON[0000_0000_0000_0000_0001_0010_0000_0000] This enables the TBI to restart Auto-Negotiations using the configuration set in the AN Advertisement register. Check to see if MII Mgmt write is complete. Read MII Mgmt Indicator register and check for Busy = 0, MIIMIND ---> [0000_0000_0000_0000_0000_0000_0000_0000] This indicates that the write cycle was completed. Check to see if PHY has completed Auto-Negotiation. Set up the MII Mgmt for a read cycle to PHY MII Mgmt register (write the PHY address and Register address), MIIMADD[0000_0000_0000_0000_0001_0000_0000_0001] The Phy Status control register is at address 0x1 and in this case the PHY Address is 0x10. Perform an MII Mgmt read cycle of Status Register. Clear MIIMCOM[Read Cycle] Set MIIMCOM[Read Cycle] (Uses the PHY address (2) and Register address (2) placed in MIIMADD register), When MIIMIND[BUSY]=0, read the MIIMSTAT register and check bit 10 (AN Done) MIIMSTAT ---> [0000_0000_0000_0000_0000_0000_0010_0000] Other information about the link is also returned. (Extend Status, No pre, Remote Fault, An Ability, Link status, extend Ability) Perform an MII Mgmt read cycle of AN Expansion Register. Setup MIIMADD[0000_0000_0000_0000_0001_0000_0000_0110] Clear MIIMCOM[Read Cycle] Set MIIMCOM[Read Cycle] (Uses the PHY address (0x10) and Register address (6) placed in MIIMADD register), When MIIMIND[BUSY]=0, read the MII Mgmt AN Expansion register and check bits 13 and 14 (NP Able and Page Rx’d) MII Mgmt AN Expansion ---> [0000_0000_0000_0000_0000_0000_0000_0110] MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-141 Three-Speed Ethernet Controllers Table 14-130. TBI Mode Register Initialization Steps (continued) Perform an MII Mgmt read cycle of AN Link Partner Base Page Ability Register. (Optional) Setup MIIMADD[0000_0000_0000_0000_0001_0000_0000_0101] Clear MIIMCOM[Read Cycle] Set MIIMCOM[Read Cycle] (Uses the PHY address (0x10) and Register address (5) placed in MIIMADD register), When MIIMIND[BUSY]=0, read the MII Mgmt AN Link Partner Base Page Ability register and check bits 2 and 3 (Remote Fault) and bits 9 and 10. (Half and Full Duplex) MII Mgmt AN Link Partner Base Page Ability ---> [0000_0000_0000_0000_0000_000x_x110_0000] Clear IEVENT register, IEVENT[0000_0000_0000_0000_0000_0000_0000_0000] Initialize IMASK (Optional) IMASK[0000_0000_0000_0000_0000_0000_0000_0000] Initialize IADDRn (Optional) IADDRn[0000_0000_0000_0000_0000_0000_0000_0000] Initialize GADDR n (Optional) GADDRn[0000_0000_0000_0000_0000_0000_0000_0000] Initialize RCTRL (Optional) RCTRL[0000_0000_0000_0000_0000_0000_0000_0000] Initialize DMACTRL (Optional) DMACTRL[0000_0000_0000_0000_0000_0000_0000_0000] Initialize FIFO_PAUSE_CTRL, FIFO_PAUSE_CTRL[0000_0000_0000_0000_0000_0000_0000_0010] Initialize (Empty) Transmit Descriptor ring and fill buffers with Data Initialize TBASE, TBASE[LLLL_LLLL_LLLL_LLLL_LLLL_LLLL_LLLL_L000] Initialize (Empty) Receive Descriptor ring and fill with empty buffers Initialize RBASE, RBASE[LLLL_LLLL_LLLL_LLLL_LLLL_LLLL_LLLL_L000] Enable Rx and Tx, MACCFG1[0000_0000_0000_0000_0000_0000_0000_0101] MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-142 Freescale Semiconductor Initialization/Application Information 14.7.1.4 RGMII Interface Mode Table 14-131 shows the signals configurations required for RGMII interface mode. Table 14-131. RGMII Interface Mode Signal Configuration TSEC SIGNALS Frequency [MHz] 125 Voltage[V] 3.3/2.5 Signals GTX_CLK TX_CLK TxD[0] TxD[1] TxD[2] TxD[3] TxD[4] TxD[5] TxD[6] TxD[7] TX_EN TX_ER RX_CLK RxD[0] RxD[1] RxD[2] RxD[3] RxD[4] RxD[5] RxD[6] RxD[7] RX_DV I/O O I O O O O O O O O O O I I I I I I I I I I No. of Signals 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 RX_CTL (RX_DV/RX_ERR) I 1 RX_CLK RxD[0]/RxD[4] RxD[1]/RxD[5] RxD[2]/RxD[6] RxD[3]/RxD[7] I I I I I 1 1 1 1 1 TX_CTL (TX_EN/TX_ERR) O 1 TxD[0]/TxD[4] TxD[1]/TxD[5] TxD[2]/TxD[6] TxD[3]/TxD[7] O O O O 1 1 1 1 RGMII INRTERFACE Frequency [MHz] 125 Voltage[V] 2.5 Signals GTX_CLK I/O O No. of Signals 1 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-143 Three-Speed Ethernet Controllers Table 14-131. RGMII Interface Mode Signal Configuration (continued) TSEC SIGNALS Frequency [MHz] 125 Voltage[V] 3.3/2.5 Signals RX_ER COL CRS I/O I I I No. of Signals 1 1 1 RGMII INRTERFACE Frequency [MHz] 125 Voltage[V] 2.5 Signals I/O No. of Signals Sum 25 Sum 12 Table 14-132 describes the shared signals for the RGMII interface. Table 14-132. Shared RGMII Signals TSEC Signals EC_MDIO EC_MDC EC_GTX_CLK12 5 Sum I/O I/O O I No. of Signals 1 1 1 3 RGMII Signals EC_MDIO EC_MDC EC_GTX_CLK12 5 Sum I/O I/O O I No. of Signals 1 1 1 3 Function Management interface I/O Management interface Clock Reference Clock Table 14-133 describes the register initializations required to reconfigure the PHY to RGMII mode following initial auto-negotiation. Table 14-133. RGMII Mode Register Initialization Steps Have the TSEC n_GTX_CLK configuration signal pulled low and EC_MDC signal pulled low for RGMII mode. TBI Control register’s Auto-negotiation Enable and Reset bits are ignored. Set Soft_Reset, MACCFG1[1000_0000_0000_0000_0000_0000_0000_0000] Clear Soft_Reset, MACCFG1[0000_0000_0000_0000_0000_0000_0000_0101] Initialize MACCFG2, MACCFG2[0000_0000_0000_0000_0111_0010_0000_0101] (I/F Mode = 2, Full Duplex = 1) Initialize ECNTRL, ECNTRL[0000_0000_0000_0000_0001_0000_0000_0000] (This example has RGMII 10-Mbps mode, Statistics Enable = 1) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-144 Freescale Semiconductor Initialization/Application Information Table 14-133. RGMII Mode Register Initialization Steps (continued) Initialize MAC Station Address, MACSTNADDR2[0110_0000_0000_0010_0000_0000_0000_0000] to 02608C:876543, for example. (Note that PHY configuration register addresses are not necessarily consistent in all PHYs.) Initialize MAC Station Address, MACSTNADDR1[0100_0011_0110_0101_1000_0111_1000_1100] to 02608C:876543, for example. (Note that PHY configuration register addresses are not necessarily consistent in all PHYs.) Assign a Physical address to the TBI, TBIPA[0000_0000_0000_0000_0000_0000_0001_0000] set to 16, for example. (Note that PHY configuration register addresses are not necessarily consistent in all PHYs.) Setup the MII Mgmt clock speed, MIIMCFG[0000_0000_0000_0000_000_0000_0000_0111] Set source clock divide by 28, for example, to insure that EC_MDC clock speed is not greater than 2.5 MHz. Read MII Mgmt Indicator register and check for Busy = 0, MIIMIND ---> [0000_0000_0000_0000_0000_0000_0000_0000] This indicates that the TSEC MII Mgmt bus is idle. Set up the MII Mgmt for a write cycle to external the PHY AN Advertisement register (write the PHY address and Register address), MIIMADD[0000_0000_0000_0000_0001_0001_0000_0100] The AN Advertisement register is at offset address 0x04 from the external PHY address. (in this case 0x11) Perform an MII Mgmt write cycle to the external PHY. Write to MII Mgmt Control with 16-bit data intended for the external PHY AN Advertisement register, MIIMCON[0000_0000_0000_0000_u0uu_uuuu_uuuu_uuuu] Where u must be selected by the user for proper system configuration. Check to see if MII Mgmt write is complete. Read MII Mgmt Indicator register and check for Busy = 0, MIIMIND ---> [0000_0000_0000_0000_0000_0000_0000_0000] This indicates that the write cycle was completed. Set up the MII Mgmt for a write cycle to the external PHY Control register (write the PHY address and Register address), MIIMADD[0000_0000_0000_0000_0001_0001_0000_0000] The Control register is at offset address 0x00 from the external PHY address. (in this case 0x11) Perform an MII Mgmt write cycle to the external PHY. Write to MII Mgmt Control with 16-bit data intended for the external PHY Control register, MIIMCON[0000_0000_0000_0000_0001_0010_0000_0000] This enables the external PHY to restart Auto-Negotiations using the configuration set in the AN Advertisement register. Check to see if MII Mgmt write is complete. Read MII Mgmt Indicator register and check for Busy = 0, MIIMIND ---> [0000_0000_0000_0000_0000_0000_0000_0000] This indicates that the write cycle was completed. Check to see if PHY has completed Auto-Negotiation. Set up the MII Mgmt for a read cycle to the PHY MII Mgmt register (write the PHY address and Register address), MIIMADD[0000_0000_0000_0000_0000_0010_0000_0001] The PHY Status register is at address 0x1 and in this case the PHY Address is 0x2. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-145 Three-Speed Ethernet Controllers Table 14-133. RGMII Mode Register Initialization Steps (continued) Perform an MII Mgmt read cycle of Status Register. Clear MIIMCOM[Read Cycle] Set MIIMCOM[Read Cycle] (Uses the PHY address (2) and Register address (2) placed in MIIMADD register) When MIIMIND[BUSY]=0, read the MIIMSTAT register and check bit 10. (AN Done) MIIMSTAT ---> [0000_0000_0000_0000_0000_0000_0010_0000] Other information about the link is also returned. (Extend Status, No pre, Remote Fault, An Ability, Link status, extend Ability) Perform an MII Mgmt read cycle of AN Expansion Register. Setup MIIMADD[0000_0000_0000_0000_0001_0001_0000_0110] Clear MIIMCOM[Read Cycle] Set MIIMCOM[Read Cycle] (Uses the PHY address (0x11) and Register address (6) placed in MIIMADD register) When MIIMIND[BUSY]=0, read the MII Mgmt AN Expansion register and check bits 13 and 14. (NP Able and Page Rx’d) MII Mgmt AN Expansion ---> [0000_0000_0000_0000_0000_0000_0000_0110] Perform an MII Mgmt read cycle of AN Link Partner Base Page Ability Register. (Optional) Setup MIIMADD[0000_0000_0000_0000_0001_0001_0000_0101] Clear MIIMCOM[Read Cycle] Set MIIMCOM[Read Cycle] (Uses the PHY address (0x11) and Register address (5) placed in MIIMADD register) When MIIMIND[BUSY]=0, read the MII Mgmt AN Link Partner Base Page Ability register and check bits 9 and 10. (Half and Full Duplex) MII Mgmt AN Link Partner Base Page Ability ---> [0000_0000_0000_0000_0000_000x_x110_0000] Clear IEVENT register, IEVENT[0000_0000_0000_0000_0000_0000_0000_0000] Initialize IMASK (Optional) IMASK[0000_0000_0000_0000_0000_0000_0000_0000] Initialize IADDRn (Optional) IADDRn[0000_0000_0000_0000_0000_0000_0000_0000] Initialize GADDR n (Optional) GADDRn[0000_0000_0000_0000_0000_0000_0000_0000] Initialize RCTRL (Optional) RCTRL[0000_0000_0000_0000_0000_0000_0000_0000] Initialize DMACTRL (Optional) DMACTRL[0000_0000_0000_0000_0000_0000_0000_0000] Initialize FIFO_PAUSE_CTRL, FIFO_PAUSE_CTRL[0000_0000_0000_0000_0000_0000_0000_0010] Initialize (Empty) Transmit Descriptor ring and fill buffers with Data Initialize TBASE, TBASE[LLLL_LLLL_LLLL_LLLL_LLLL_LLLL_LLLL_L000] Initialize (Empty) Receive Descriptor ring and fill with empty buffers Initialize RBASE, RBASE[LLLL_LLLL_LLLL_LLLL_LLLL_LLLL_LLLL_L000] Enable Rx and Tx, MACCFG1[0000_0000_0000_0000_0000_0000_0000_0101] MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-146 Freescale Semiconductor Initialization/Application Information 14.7.1.5 RTBI Interface Mode Table 14-134 describes the signal configurations required for RTBI interface mode. Table 14-134. RTBI Interlace Mode Signal Configuration TSEC Signals Frequency [MHz] 125 Voltage[V] 3.3/2.5 Signals GTX_CLK TX_CLK TxD[0] TxD[1] TxD[2] TxD[3] TxD[4] TxD[5] TxD[6] TxD[7] TX_EN TX_ER RX_CLK RxD[0] RxD[1] RxD[2] RxD[3] RxD[4] RxD[5] RxD[6] RxD[7] RX_DV I/O O I O O O O O O O O O O I I I I I I I I I I No. of Signals 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 RCG[4]/RCG[9] I 1 RX_CLK RCG[0]/RCG[5] RCG[1]/RCG[6] RCG[2]/RCG[7] RCG[3]/RCG[8] I I I I I 1 1 1 1 1 TCG[4]/TCG[9] O 1 TCG[0]/TCG[5] TCG[1]/TCG[6] TCG[2]/TCG[7] TCG[3]/TCG[8] O O O O 1 1 1 1 RTBI Interface Frequency [MHz] 62.5 Voltage[V] 2.5 Signals GTX_CLK I/O O No. of Signals 1 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-147 Three-Speed Ethernet Controllers Table 14-134. RTBI Interlace Mode Signal Configuration (continued) TSEC Signals Frequency [MHz] 125 Voltage[V] 3.3/2.5 Signals RX_ER COL CRS Sum I/O I I I No. of Signals 1 1 1 25 Sum I 12 RTBI Interface Frequency [MHz] 62.5 Voltage[V] 2.5 Signals I/O No. of Signals Table 14-135 describes the shared signals for the RTBI interface. Table 14-135. Shared RTBI Signals TSEC Signals EC_MDIO EC_MDC EC_GTX_CLK12 5 Sum I/O I/O O I No. of Signals 1 1 1 3 RTBI Signals EC_MDIO EC_MDC EC_GTX_CLK12 5 Sum I/O I/O O I No. of Signals 1 1 1 3 Function Management interface I/O Management interface Clock Reference Clock Table 14-136 describes the register initializations required to reconfigure the PHY to RTBI mode following initial auto-negotiation. Table 14-136. RTBI Mode Register Initialization Steps Have the TSECn_GTX_CLK configuration signal pulled high and the EC_MDC signal pulled low for RTBI mode. (TSEC attempts to auto-negotiate after system reset.) Set Soft_Reset, MACCFG1[1000_0000_0000_0000_0000_0000_0000_0000] Clear Soft_Reset, MACCFG1[0000_0000_0000_0000_0000_0000_0000_0101] Initialize MACCFG2, MACCFG2[0000_0000_0000_0000_0111_0010_0000_0101] (I/F Mode = 2, Full Duplex = 1) Initialize ECNTRL, ECNTRL[0000_0000_0000_0000_0001_0000_0000_0000] (This example has Statistics Enable = 1) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-148 Freescale Semiconductor Initialization/Application Information Table 14-136. RTBI Mode Register Initialization Steps (continued) Initialize MAC Station Address, MACSTNADDR2[0110_0000_0000_0010_0000_0000_0000_0000] to 02608C:876543, for example. (Note that PHY configuration register addresses are not necessarily consistent in all PHYs.) Initialize MAC Station Address, MACSTNADDR1[0100_0011_0110_0101_1000_0111_1000_1100] to 02608C:876543, for example. (Note that PHY configuration register addresses are not necessarily consistent in all PHYs.) Assign a Physical address to the TBI, TBIPA[0000_0000_0000_0000_0000_0000_0001_0000] set to 16, for example. (Note that PHY configuration register addresses are not necessarily consistent in all PHYs.) Setup the MII Mgmt clock speed, MIIMCFG[0000_0000_0000_0000_0000_0000_0000_0111] Set source clock divide by 28, for example, to insure that EC_MDC clock speed is not greater than 2.5 MHz. Read MII Mgmt Indicator register and check for Busy = 0, MIIMIND ---> [0000_0000_0000_0000_0000_0000_0000_0000] This indicates that the TSEC MII Mgmt bus is idle. Set up the MII Mgmt for a read cycle to TBI Control register (write the TBI’s address and Register address), MIIMADD[0000_0000_0000_0000_0001_0000_0000_0000] The TBI Control register is at offset address 0x0 from TBIPA. Perform an MII Mgmt read cycle to verify state of TBI Control Register(0ptional) Clear MIIMCOM[Read Cycle] Set MIIMCOM[Read Cycle] (Uses the TBI address and Register address placed in MIIMADD register), When MIIMIND[BUSY]=0, read the MIIMSTAT and look for AN Enable and other bit information. Set up the MII Mgmt for a write cycle to TBI’s AN Advertisement register (write the PHY address and Register address), MIIMADD[0000_0000_0000_0000_0001_0000_0000_0100] The AN Advertisement register is at offset address 0x04 from the TBI’s address. (in this case 0x10) Perform an MII Mgmt write cycle to TBI. Write to MII Mgmt Control with 16-bit data intended for TBI’s AN Advertisement register, MIIMCON[0000_0000_0000_0000_0000_0001_1010_0000] This advertises to the Link Partner that the TBI supports PAUSE and Full Duplex mode and does not support Half Duplex mode. Check to see if MII Mgmt write is complete. Read MII Mgmt Indicator register and check for Busy = 0, MIIMIND ---> [0000_0000_0000_0000_0000_0000_0000_0000] This indicates that the write cycle was completed. Set up the MII Mgmt for a write cycle to TBI’s Control register (write the PHY address and Register address), MIIMADD[0000_0000_0000_0000_0001_0000_0000_0000] The Control register is at offset address 0x00 from the TBI’s address. (in this case 0x10) Perform an MII Mgmt write cycle to TBI. Writing to MII Mgmt Control with 16-bit data intended for TBI’s Control register, MIIMCON[0000_0000_0000_0000_0001_0010_0000_0000] This enables the TBI to restart Auto-Negotiations using the configuration set in the AN Advertisement register. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-149 Three-Speed Ethernet Controllers Table 14-136. RTBI Mode Register Initialization Steps (continued) Check to see if MII Mgmt write is complete. Read MII Mgmt Indicator register and check for Busy = 0, MIIMIND ---> [0000_0000_0000_0000_0000_0000_0000_0000] This indicates that the write cycle was completed. Check to see if PHY has completed Auto-Negotiation. Set up the MII Mgmt for a read cycle to the PHY MII Mgmt register (write the PHY address and Register address), MIIMADD[0000_0000_0000_0000_0001_0000_0000_0001] The Phy Status control register is at address 0x1 and in this case the PHY Address is 0x10. Perform an MII Mgmt read cycle of Status Register. Clear MIIMCOM[Read Cycle] Set MIIMCOM[Read Cycle] (Uses the PHY address (2) and Register address (2) placed in MIIMADD register), When MIIMIND[BUSY]=0, read the MIIMSTAT register and check bit 10 (AN Done) MIIMSTAT ---> [0000_0000_0000_0000_0000_0000_0010_0000] Other information about the link is also returned. (Extend Status, No pre, Remote Fault, An Ability, Link status, extend Ability) Perform an MII Mgmt read cycle of AN Expansion Register. Setup MIIMADD[0000_0000_0000_0000_0001_0000_0000_0110] Clear MIIMCOM[Read Cycle] Set MIIMCOM[Read Cycle] (Uses the PHY address (0x10) and Register address (6) placed in MIIMADD register), When MIIMIND[BUSY]=0, read the MII Mgmt AN Expansion register and check bits 13 and 14. (NP Able and Page Rx’d) MII Mgmt AN Expansion ---> [0000_0000_0000_0000_0000_0000_0000_0110] Perform an MII Mgmt read cycle of AN Link Partner Base Page Ability Register. (Optional) Setup MIIMADD[0000_0000_0000_0000_0001_0000_0000_0101] Clear MIIMCOM[Read Cycle] Set MIIMCOM[Read Cycle] (Uses the PHY address (0x10) and Register address (5) placed in MIIMADD register), When MIIMIND[BUSY]=0, read the MII Mgmt AN Link Partner Base Page Ability register and check bits 9 and 10. (Half and Full Duplex) MII Mgmt AN Link Partner Base Page Ability ---> [0000_0000_0000_0000_0000_000x_x110_0000] Clear IEVENT register, IEVENT[0000_0000_0000_0000_0000_0000_0000_0000] Initialize IMASK (Optional) IMASK[0000_0000_0000_0000_0000_0000_0000_0000] Initialize IADDRn (Optional) IADDRn[0000_0000_0000_0000_0000_0000_0000_0000] Initialize GADDR n (Optional) GADDRn[0000_0000_0000_0000_0000_0000_0000_0000] Initialize RCTRL (Optional) RCTRL[0000_0000_0000_0000_0000_0000_0000_0000] Initialize DMACTRL (Optional) DMACTRL[0000_0000_0000_0000_0000_0000_0000_0000] Initialize FIFO_PAUSE_CTRL, FIFO_PAUSE_CTRL[0000_0000_0000_0000_0000_0000_0000_0010] MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-150 Freescale Semiconductor Initialization/Application Information Table 14-136. RTBI Mode Register Initialization Steps (continued) Initialize (Empty) Transmit Descriptor ring and fill buffers with Data Initialize TBASE, TBASE[LLLL_LLLL_LLLL_LLLL_LLLL_LLLL_LLLL_L000] Initialize (Empty) Receive Descriptor ring and fill with empty buffers Initialize RBASE, RBASE[LLLL_LLLL_LLLL_LLLL_LLLL_LLLL_LLLL_L000] Enable Rx and Tx, MACCFG1[0000_0000_0000_0000_0000_0000_0000_0101] MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 14-151 Three-Speed Ethernet Controllers MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 14-152 Freescale Semiconductor Introduction Chapter 15 DMA Controller This chapter describes the DMA controller of this device. 15.1 Introduction The DMA controller transfers blocks of data between the RapidIO controller, PCI, the local bus controller (LBC) interface, and the local address space, independent of the e500 core or external hosts. 15.1.1 Block Diagram Figure 15-1 shows the block diagram of the DMA controller. CHN 0 CHN 1 CHN 2 CHN 3 Arbitration and Bandwidth Control Source Address Controls Destination Address Controls DMA Controller Address Tenure Controls Data Tenure Controls Address Interface Data Interface Figure 15-1. DMA Block Diagram 15.1.2 Overview The DMA controller has four high-speed DMA channels. Both the e500 core and external devices can initiate DMA transfers. All channels are capable of complex data movement and advanced MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 15-1 DMA Controller transaction chaining. Figure 15-1 is a high-level block diagram of the DMA controller. Operations such as descriptor fetches and block transfers are initiated by each channel. A channel is selected by the arbitration logic and information is passed to the source and destination control blocks for processing. The source and destination blocks generate read and write requests to the address tenure engine, which manages the DMA master port address interface. After a transaction is accepted by the master port, control is transferred to the data tenure engine that manages the read and write data transfers. A channel remains active in the shared resources for the duration of the data transfer unless the allotted bandwidth per channel is reached. 15.1.3 Features The DMA controller offers the following features: • • • • • • • • • • Four high-speed/high-bandwidth channels accessible by local and remote masters Basic DMA operation modes (direct, simple chaining) Extended DMA operation modes (advanced chaining and stride capability) Cascading descriptor chains Misaligned transfers Programmable bandwidth control between channels Up to 256 bytes for DMA sub-block transfers to maximize performance over the RapidIO interface Three priority levels supported for source and destination transactions Interrupt on error and completed segment, list, or link Externally-controlled transfer using DMA_DREQ, DMA_DACK, and DMA_DDONE 15.1.4 Modes of Operation The MPC8540 has two modes of operation: basic and extended. Basic mode is the DMA legacy mode. It does not support advanced features. Extended mode supports advanced features like striding and flexible descriptor structures. These two basic modes allow users to initiate and end DMA transfers in various ways. Table 15-1 summarizes the relationship between the modes and the following features: • • • Direct mode. No descriptors are involved. Software must initialize the required fields as described in Table 15-2 before starting a transfer. Chaining mode. Software must initialize descriptors in memory and the required fields as described in Table 15-2 before starting a transfer. Single-write start mode. The DMA process can be started by using a single-write command to either the descriptor address register in one of the chaining modes or the source/destination address registers in one of the direct modes. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 15-2 Freescale Semiconductor Introduction • • • External control capability: This allows an external agent to start, pause, and check the status of a DMA transfer which has already been initialized. Channel continue capability: The channel continue capability allows software the flexibility of having the DMA controller start with descriptors that have already been programmed while software continues to build more descriptors in memory. Channel abort capability: The software can abort a previously initiated transfer by setting the bit MRn[CA]. The DMA controller terminates all outstanding transfers initiated by the channel without generating any errors before entering an idle state. Table 15-1. Relationship of Modes and Features Mode B (Basic) Mode with One Additional Feature Mode with Two Additional Features BD (Basic direct) BDS (BD single-write start) BDE (BD external control) BC (Basic chaining) BCE (BC external control) BCS (BC single-write start) Ext (Extended) ExtD (Extended direct) ExtDS (ExtD single-write start) ExtDE (ExtD external control) ExtC (Extended chaining) ExtCE (ExtC external control) ExtCS (ExtC single-write start) Table 15-2 describes bit settings required for each DMA mode of operation Table 15-2. DMA Mode Field Descriptions Modes with Features MRn[XFE] MRn[CTM] MRn[SRW] MRn[EM] MRn[CDSM/SWSM] MRn[EMS_EN] Basic Direct Modes Basic direct Basic direct external control Basic direct single-write start 0 0 0 1 1 1 0 0 1 0 0 0 0 0 1 or 0 0 1 0 Basic Chaining Modes Basic chaining Basic chaining external control Basic chaining single-write start 0 0 0 0 0 0 Reserved Reserved Reserved 0 0 0 0 0 1 0 1 0 Extended Direct Modes Extended direct Extended direct external control Extended direct single-write start 1 1 1 1 1 1 0 0 1 0 0 0 0 0 1 or 0 0 1 0 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 15-3 DMA Controller Table 15-2. DMA Mode Field Descriptions (continued) Modes with Features MRn[XFE] MRn[CTM] MRn[SRW] MRn[EM] MRn[CDSM/SWSM] MRn[EMS_EN] Extended Chaining Modes Extended chaining Extended chaining external control Extended chaining single-write start 1 1 1 0 0 0 Reserved Reserved Reserved 0 0 0 0 0 1 0 1 0 Refer to Section 15.4, “Functional Description,” for details on these modes. Figure 15-2 shows the general DMA operational flow chart. Software sets 1st link Process Link Y Is last link? N Advance Link In Chain or Extended Mode? N Y Y Is last list? N Done - DMA halts Advance List Figure 15-2. DMA Operational Flow Chart MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 15-4 Freescale Semiconductor External Signal Descriptions 15.2 External Signal Descriptions This section describes the DMA signals. 15.2.1 Signal Overview Figure 15-3 summarizes the DMA controller signals. 1 1 1 1 1 1 1 1 1 1 1 1 DMA_DREQ0 DMA_DACK0 DMA_DDONE0 DMA_DREQ1 DMA_DACK1 DMA_DDONE1 DMA_DREQ2 DMA_DACK2 DMA_DDONE2 DMA_DREQ3 DMA_DACK3 DMA_DDONE3 DMA External Control Interface DMA Unit Figure 15-3. DMA Signal Summary Note that the three DMA signals for DMA channel 3 are multiplexed with the IRQ9–11 signals on the MPC8540 device. These functions are mutually exclusive and the active function is specified in the PMUXCR register of the global utilities block as described in Section 18.4.1.10, “Alternate Function Signal Multiplex Control Register (PMUXCR).” MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 15-5 DMA Controller 15.2.2 Detailed Signal Descriptions Table 15-3 describes the DMA signals. Table 15-3. DMA Signals—Detailed Signal Descriptions Signal DMA_DREQn DMA request I/O I Description DMA request. The DMA request signal indicates the start of a DMA transfer or a restart from a paused request. The falling edge of DMA_DREQn causes MRn[CS] to be set, thereby activating the corresponding DMA channel. State Asserted—Assertion of DMA_DREQn while DMA_DACKn is negated causes a new Meaning transfer to start OR resumes a paused transfer if the EMP_EN bit is set. Assertion while DMA_DACKn is asserted results in an illegal condition Negated—Negation while DMA_DACKn is asserted has no effect. Negation before the assertion of DMA_DACKn results in an illegal condition Timing Assertion—Can be asserted asynchronously Negation— Must remain asserted at least until the assertion of the corresponding DMA_DACKn DMA_DACKn O DMA acknowledge. Indicates that a DMA transfer is currently in progress State Asserted—Indicates that a DMA transfer is currently in progress. Asserted after the Meaning assertion of DMA_DREQ to indicate the start of a transfer Negated—Negated after finishing a complete transfer or after entering a paused state if MRn[EMP_EN] is set Timing Assertion—Asynchronous assertion; asserted for more than three system clocks Negation—Asynchronous negation; negated for more than three system clocks DMA_DDONEn O DMA done. Indicates that a DMA transfer is complete. State Asserted—Indicates transfer completion. SRn[CB] is clear. Note, however, that write Meaning data may still be queued at the target interface or in the process of transfer on an external interface. Negated—Indicates that the current transfer is in process Timing Assertion—Always asserts asynchronously after the negation of the final DMA_DACKn to indicate completion of a transfer. For a paused transfer, DMA_DDONE n is asserted asynchronously after the negation of the final DMA_DACKn. Negation—Negated asynchronously after the assertion of DMA_DREQn for the next transfer 15.3 Memory Map/Register Definition This section provides a detailed description of all accessible DMA memory and registers. The descriptions include individual bit level descriptions and reset states of each register. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 15-6 Freescale Semiconductor Memory Map/Register Definition 15.3.1 Module Memory Map Figure 15-4 shows the address map of DMA register space. 0x2_1000 General DMA Registers 0x2_1300 Reserved 0x2_1FFF Figure 15-4. DMA Register Space (Memory-Mapped) Table 15-4 lists the DMA registers. Table 15-4. DMA Register Summary Offset Register General Registers 0x2_1100 0x2_1104 0x2_1108 0x2_110C 0x2_1110 0x2_1114 0x2_1118 0x2_111C 0x2_1120 0x2_1124 0x2_1128 0x2_112C– 0x2_1130 MR0—DMA mode register 0 SR0—DMA status register 0 Reserved CLNDAR0—DMA current link descriptor address register 0 SATR0—DMA source attributes register 0 SAR0—DMA source address register 0 DATR0—DMA destination attributes register 0 DAR0—DMA destination address register 0 BCR0—DMA byte count register 0 Reserved NLNDAR0—DMA next link descriptor address register 0 Reserved R/W Special — R/W R/W R/W R/W R/W R/W — R/W — 0x0000_0000 0x0000_0000 — 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 — 0x0000_0000 — 15.3.2.1/15-10 15.3.2.2/15-13 — 15.3.2.3/15-14 15.3.2.4/15-16 15.3.2.5/15-18 15.3.2.6/15-19 15.3.2.7/15-21 15.3.2.8/15-23 — 15.3.2.9/15-23 — Access Reset Section/Page MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 15-7 DMA Controller Table 15-4. DMA Register Summary (continued) Offset 0x2_1134 0x2_1138 0x2_113C 0x2_1140 0x2_1144 0x2_1148– 0x2_117C 0x2_1180 0x2_1184 0x2_1188 0x2_118C 0x2_1190 0x2_1194 0x2_1198 0x2_119C 0x2_11A0 0x2_11A4 0x2_11A8 0x2_11AC– 0x2_11B0 0x2_11B4 0x2_11B8 0x2_11BC 0x2_11C0 0x2_11C4 0x2_11C8– 0x2_11FC 0x2_1200 0x2_1204 0x2_1208 0x2_120C 0x2_1210 0x2_1214 Register CLSDAR0—DMA current list alternate base descriptor address register 0 Reserved NLSDAR0—DMA next list descriptor address register 0 SSR0—DMA source stride register 0 DSR0—DMA destination stride register 0 Reserved MR1—DMA mode register 1 SR1—DMA status register 1 Reserved CLNDAR1—DMA current link descriptor address register 1 SATR1—DMA source attributes register 1 SAR1—DMA source address register 1 DATR1—DMA destination attributes register 1 DAR1—DMA destination address register 1 BCR1—DMA byte count register 1 Reserved NLNDAR1—DMA next link descriptor address register 1 Reserved CLSDAR1—DMA current list alternate base descriptor address register 1 Reserved NLSDAR1—DMA next list descriptor address register 1 SSR1—DMA source stride register 1 DSR1—DMA destination stride register 1 Reserved MR2—DMA mode register 2 SR2—DMA status register 2 Reserved CLNDAR2—DMA current link descriptor address register 2 SATR2—DMA source attributes register 2 SAR2—DMA source address register 2 Access R/W — R/W R/W R/W — R/W Special — R/W R/W R/W R/W R/W R/W — R/W — R/W — R/W R/W R/W — R/W Special — R/W R/W R/W Reset 0x0000_0000 — 0x0000_0000 0x0000_0000 0x0000_0000 — 0x0000_0000 0x0000_0000 — 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 — 0x0000_0000 — 0x0000_0000 — 0x0000_0000 0x0000_0000 0x0000_0000 — 0x0000_0000 0x0000_0000 — 0x0000_0000 0x0000_0000 0x0000_0000 Section/Page 15.3.2.10/15-24 — 15.3.2.11/15-25 15.3.2.12/15-25 15.3.2.13/15-26 — 15.3.2.1/15-10 15.3.2.2/15-13 — 15.3.2.3/15-14 15.3.2.4/15-16 15.3.2.5/15-18 15.3.2.6/15-19 15.3.2.7/15-21 15.3.2.8/15-23 — 15.3.2.9/15-23 — 15.3.2.10/15-24 — 15.3.2.11/15-25 15.3.2.12/15-25 15.3.2.13/15-26 — 15.3.2.1/15-10 15.3.2.2/15-13 — 15.3.2.3/15-14 15.3.2.4/15-16 15.3.2.5/15-18 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 15-8 Freescale Semiconductor Memory Map/Register Definition Table 15-4. DMA Register Summary (continued) Offset 0x2_1218 0x2_121C 0x2_1220 0x2_1224 0x2_1228 0x2_122C– 0x2_1230 0x2_1234 0x2_1238 0x2_123C 0x2_1240 0x2_1244 0x2_1248– 0x2_127C 0x2_1280 0x2_1284 0x2_1288 0x2_128C 0x2_1290 0x2_1294 0x2_1298 0x2_129C 0x2_12A0 0x2_12A4 0x2_12A8 0x2_12AC– 0x2_12B0 0x2_12B4 0x2_12B8 0x2_12BC 0x2_12C0 0x2_12C4 Register DATR2—DMA destination attributes register 2 DAR2—DMA destination address register 2 BCR2—DMA byte count register 2 Reserved NLNDAR2—DMA next link descriptor address register 2 Reserved CLSDAR2—DMA current list alternate base descriptor address register 2 Reserved NLSDAR2—DMA next list descriptor address register 2 SSR2—DMA source stride register 2 DSR2—DMA destination stride register 2 Reserved MR3—DMA mode register 3 SR3—DMA status register 3 Reserved CLNDAR3—DMA current link descriptor address register 3 SATR3—DMA source attributes register 3 SAR3—DMA source address register 3 DATR3—DMA destination attributes register 3 DAR3—DMA destination address register 3 BCR3—DMA byte count register 3 Reserved NLNDAR3—DMA next link descriptor address register 3 Reserved CLSDAR3—DMA current list alternate base descriptor address register 3 Reserved NLSDAR3—DMA next list descriptor address register 3 SSR3—DMA source stride register 3 DSR3—DMA destination stride register 3 Access R/W R/W R/W — R/W — R/W — R/W R/W R/W — R/W Special — R/W R/W R/W R/W R/W R/W — R/W — R/W — R/W R/W R/W Reset 0x0000_0000 0x0000_0000 0x0000_0000 — 0x0000_0000 — 0x0000_0000 — 0x0000_0000 0x0000_0000 0x0000_0000 — 0x0000_0000 0x0000_0000 — 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 — 0x0000_0000 — 0x0000_0000 — 0x0000_0000 0x0000_0000 0x0000_0000 Section/Page 15.3.2.6/15-19 15.3.2.7/15-21 15.3.2.8/15-23 — 15.3.2.9/15-23 — 15.3.2.10/15-24 — 15.3.2.11/15-25 15.3.2.12/15-25 15.3.2.13/15-26 — 15.3.2.1/15-10 15.3.2.2/15-13 — 15.3.2.3/15-14 15.3.2.4/15-16 15.3.2.5/15-18 15.3.2.6/15-19 15.3.2.7/15-21 15.3.2.8/15-23 — 15.3.2.9/15-23 — 15.3.2.10/15-24 — 15.3.2.11/15-25 15.3.2.12/15-25 15.3.2.13/15-26 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 15-9 DMA Controller Table 15-4. DMA Register Summary (continued) Offset 0x2_12C8– 0x2_12FC 0x2_1300 Reserved DGSR—DMA general status register Register Access — Read Reset — 0x0000_0000 Section/Page — 15.3.2.14/15-26 15.3.2 DMA Register Descriptions The following sections describe the DMA registers. The majority of these registers are channel-specific and can be identified by one of the four offsets that describe the register. 15.3.2.1 Mode Registers (MRn) The mode register allows software to start a DMA transfer and to control various DMA transfer characteristics. Figure 15-5 describes the MRn. 0 3 4 7 8 9 10 11 12 13 14 15 R0 W Reset 16 0 0 0 BWC 0 0 EMP_EN 0 0 EMS_EN DAHTS 0000_0000_0000_0000 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 R SAHTS DAHE SAHE W Reset Offset 0 SRW EOSIE EOLNIE EOLSIE EIE 0000_0000_0000_0000 0x2_1100, 0x2_1180, 0x2_1200, 0x2_1280 XFE CDSM/ CA SWSM CTM CC CS Figure 15-5. DMA Mode Registers (MRn) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 15-10 Freescale Semiconductor Memory Map/Register Definition Table 15-5 describes the MRn fields. Table 15-5. MRn Field Descriptions Bits 0–3 4–7 Name — BWC Reserved Bandwidth/pause control. Defined only if multiple channels are executing transfers concurrently or if MRn[EMP_EN] is set. Determines how many bytes a given channel is allowed to transfer before the DMA engine switches to the next channel. The DMA channel pauses during this switch if MRn[EMP_EN] is set for this channel. 0111 128 bytes 0000 1 byte 1000 256 bytes 0001 2 bytes 1001 512 bytes 0010 4 bytes 1010 1024 bytes 0011 8 bytes 1011–1110 Reserved 0100 16 bytes 1111 Disable bandwidth sharing to allow 0101 32 bytes uninterrupted transfers from each channel. 0110 64 bytes Reserved Description 8–9 10 — EMP_EN External master pause enable. Valid only if MRn[EMS_EN] is set. 0 Disable the channel from being paused by an external master. 1 Enable the channel to be paused by an external master. — Reserved 11–12 13 EMS_EN External master start enable. This bit does not apply to single-write start modes (direct or chaining). 0 Disable the channel from being started by an external DMA start pin. 1 Enable the channel to be started by an external DMA start pin, which sets MRn[CS]. DAHTS Destination address hold transfer size. Indicates the transfer size used for each transaction while MRn[DAHE] is set. The byte count register must be in multiples of the size and the destination address register must be aligned based on the size. The transfer size assigned to MRn[DAHTS] must be equal to or smaller than that assigned to MRn[BWC] to avoid undefined behavior. 00 1 byte 01 2 bytes 10 4 bytes 11 8 bytes Source address hold transfer size. Indicates the transfer size used for each transaction while MRn[SAHE] is set. The byte count register must be in multiples of the size and the source address register must be aligned based on the size. The transfer size assigned to MRn[SAHTS] must be equal to or smaller than that assigned to MRn[BWC] to avoid undefined behavior. 00 1 byte 01 2 bytes 10 4 bytes 11 8 bytes Destination address hold enable 0 Disable destination address hold. 1 Enable the DMA controller to hold the destination address of a transfer to the size specified by MRn[DAHTS]. Hardware only supports aligned transfers for this feature. Source address hold enable 0 Disable source address hold. 1 Enable the DMA controller to hold the source address of a transfer to the size specified by MRn[SAHTS]. Hardware only supports aligned transfers for this feature. 14–15 16–17 SAHTS 18 DAHE 19 SAHE MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 15-11 DMA Controller Table 15-5. MRn Field Descriptions (continued) Bits 20 21 Name — SRW Reserved Single register write (Direct mode only; reserved for chaining mode.) 0 Normal operation. 1 Enable a write to the source address register to simultaneously set MRn[CS], starting a DMA transfer, when MRn[CDSM/SWSM] is also set. Setting this bit and clearing CDSM/SWSM causes a write to the destination address register to simultaneously set MRn[CS], starting a DMA transfer. End-of-segments interrupt enable 0 Do not generate an interrupt at the completion of a data transfer. CLNDARn[EOSIE] overrides this bit on a link descriptor basis. 1 Generate an interrupt at the completion of a data transfer (That is, SRn[EOSI] is set). This bit overrides the CLNDARn[EOSIE]. End-of-links interrupt enable 0 Do not generate an interrupt at the completion of a list of DMA transfers. 1 Generate an interrupt at the completion of a list of DMA transfers (That is, NLNDARn[EOLND] is set). End-of-lists interrupt enable 0 Do not generate an interrupt at the completion of all DMA transfers. 1 Generate an interrupt at the completion of all DMA transfers (That is, NLNDARn[EOLND] and NLSDARn[EOLSD] are set). Error interrupt enable 0 Do not generate an interrupt if a programming or transfer error is detected. 1 Generate an interrupt if a programming or transfer error is detected. Extended features enable 0 Disable the new chaining features. 1 Enable the new chaining features. • In chaining mode: Current descriptor start mode/single-write start mode. • In basic mode (MRn[XFE] is cleared), setting this bit causes a write to the current link descriptor address register to simultaneously set MRn[CS], starting a DMA transfer. • In extended chaining mode (MRn[XFE] is set), setting this bit causes a write to the current list descriptor address register to simultaneously set MRn[CS], starting a DMA transfer. • In direct mode: Setting this bit and MRn[SRW] causes a write to the source address register to simultaneously set MRn[CS], starting a DMA transfer. Clearing this bit and setting MR n[SRW] causes a write to the destination address register to simultaneously set MRn[CS], starting a DMA transfer. This bit must be cleared when MRn[SRW] is cleared. Channel abort 0 No effect. 1 Cause the current transfer to be aborted and SRn[CB] to be cleared if the channel is busy. The channel remains in the idle state until a new transfer is programmed. Channel transfer mode 0 Configure the channel in chaining mode. 1 Configure the channel into direct mode. This means that software is responsible for placing all the required parameters into necessary registers to start the DMA process. Description 22 EOSIE 23 EOLNIE 24 EOLSIE 25 EIE 26 XFE 27 CDSM/ SWSM 28 CA 29 CTM MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 15-12 Freescale Semiconductor Memory Map/Register Definition Table 15-5. MRn Field Descriptions (continued) Bits 30 Name CC Description Channel continue. This bit applies only to chaining mode and is cleared by hardware after the first descriptor read when continuing a transfer. This bit is reserved for external master mode. 0 No effect. 1 The DMA transfer restarts the transferring process starting at the current descriptor address. Channel start. This bit is also automatically set by hardware during single-write start mode and external master start enable mode. 0 Halt the DMA process if channel is busy (SRn[CB] is set). No effect if the channel is not busy. 1 Start the DMA process if channel is not busy (CB is cleared). If the channel was halted (CS = 0 and CB = 1), the transfer continues from the point at which it was halted. 31 CS 15.3.2.2 Status Registers (SRn) The status registers, shown in Figure 15-6, report various DMA conditions during and after a DMA transfer. 0 23 24 25 26 27 28 29 30 31 R00000000000000000000000 0 TE W Reset Offset 0 CH PE EOLNI CB EOSI EOLSI 0000_0000_0000_0000_0000_0000_0000_0000 0x2_1104, 0x2_1184, 0x2_1204, 0x2_1284 Figure 15-6. Status Registers (SRn) Table 15-6 describes the bits of the SRn. Table 15-6. SRn Field Descriptions Bits 0–23 24 Name — TE Reserved Transfer error (Bit reset, write 1 to clear) 0 No error condition during the DMA transfer. 1 Error condition during the DMA transfer. See Section 15.4.3, “DMA Errors” for additional information. Reserved Channel halted. Cleared automatically by hardware if MRn[CS] is set again for resuming a halted transfer. 0 Channel is not halted. If software attempts to halt an idle channel (SRn[CB] is cleared), this bit will remain 0. 1 DMA transfer was successfully halted by software and it can be resumed. Programming error (Bit reset, write 1 to clear) 0 No programming error detected. 1 A programming error is detected that prevents the DMA transfer from occurring. Description 25 26 — CH 27 PE MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 15-13 DMA Controller Table 15-6. SRn Field Descriptions (continued) Bits 28 Name Description EOLNI End-of-links interrupt. After transferring the last block of data in the last link descriptor, if MRn[EOLSIE] is set, then this bit is set and an interrupt is generated. (Bit reset, write 1 to clear) CB Channel busy 0 DMA transfer is finished, an error occurred, or a channel abort occurred. 1 A DMA transfer is currently in progress. End-of-segment interrupt. In chaining mode, after finishing a data transfer, if MRn[EOSIE] is set or if CLNDARn[EOSIE] is set, this bit gets set and an interrupt is generated. In direct mode, if MRn[EOSIE] is set, this bit gets set and an interrupt is generated. (Bit reset, write 1 to clear) 29 30 EOSI 31 EOLSI End-of-list interrupt. After transferring the last block of data in the last list descriptor, if MRn[EOLSIE] is set, then this bit is set and an interrupt is generated. (Bit reset, write 1 to clear) 15.3.2.3 Current Link Descriptor Address Registers (CLNDARn) Current link descriptor address registers contain the address of the current link descriptor. In basic chaining mode, shown in Figure 15-7, software must initialize this register to point to the first link descriptor in memory. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 15-14 Freescale Semiconductor Memory Map/Register Definition Software initializes CLNDAR n with 1st link descriptor. 1st link descriptor is processed. Current link descriptor ← Next link descriptor. Y Is NLNDARn[EOLND] set? N Last link descriptor has been processed. Is Extended Chaining enabled? N Y Y Is NLSDARn[EOLSD] set? N Done — DMA halts. CLSDARn ← NLSDARn Figure 15-7. Basic Chaining Mode Flow Chart After the current descriptor is processed, the current link descriptor address register is loaded from the next link descriptor address register and NLNDARn[EOLND] in the next link descriptor address register is examined. If EOLND is zero, the DMA controller reads in the new current link descriptor for processing. If EOLND is set, the last descriptor of the list was just completed. If extended chaining mode is not enabled, all DMA transfers are complete and the DMA controller halts. If extended chaining mode is enabled, the DMA controller examines the state of NLSDARn[EOLSD] in the next list descriptor address register. If EOLSD is clear, the controller loads the contents of the next list descriptor address register into the current list descriptor address register and reads the new list descriptor from memory. If EOLSD is set, all DMA transfers are complete and the DMA controller halts. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 15-15 DMA Controller Figure 15-8 shows CLNDARn. 0 26 27 28 29 30 31 R CLNDA W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x2_110C, 0x2_118C, 0x2_120C, 0x2_128C 0 EOSIE 0 0 0 Figure 15-8. Current Link Descriptor Address Registers (CLNDARn) Table 15-7 describes the fields of the CLNDARn. Table 15-7. CLNDARn Field Descriptions Bits 0–26 27 28 Name Description CLNDA Current link descriptor address. Contains the current descriptor address of the buffer descriptor in memory. The descriptor must be aligned to a 32-byte boundary. — EOSIE Reserved End-of-segment interrupt enable 0 Do not generate an interrupt upon completion of the current DMA transfer for the current link descriptor. 1 Generate an interrupt upon completion of the current DMA transfer for the current link descriptor. Reserved 29–31 — 15.3.2.4 Source Attributes Registers (SATRn) The source attributes registers, shown in Figure 15-9, contain the transaction attributes to be used for the DMA operation. Stride mode is enabled by setting SATRn[SSME]. Source read transaction type is specified using the SREADTTYPE field. ATMU bypass mode is enabled by setting SATRn[SBPATMU]. ATMU bypass mode is only applicable for accesses to RapidIO. If SBPATMU is set, STRANSINT must be set to RapidIO and attributes that would otherwise come from the ATMU must be specified in this register. If SBPATMU is not set, attributes are derived from local access windows and outbound ATMU registers according to the transaction address. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 15-16 Freescale Semiconductor Memory Map/Register Definition 0 1 2 3 4 5 6 7 8 11 12 15 R W Reset 0 0 SBPATMU 0 STFLOWLVL SPCIORDER SSME STRANSINT SREADTTYPE 0000_0000_0000_0000 16 21 22 31 R W Reset Offset 0 0 0 0 0 0 ESAD 0000_0000_0000_0000 0x2_1110, 0x2_1190, 0x2_1210, 0x2_1290 Figure 15-9. Source Attributes Registers (SATRn) Table 15-8 describes the fields of the SATRn. Table 15-8. SATRn Field Descriptions Bits 0–1 2 Name — SBPATMU Reserved Bypass ATMU for this DMA operation 0 Route the operation through the ATMU outbound windows. SATRn[SREADTTYPE] should specify a local address space transaction type. 1 Bypass ATMU. Never generate an address match. Always use the attributes in the source attribute registers. Route the transaction to the interface specified in SATRn[STRANSINT]. Applicable only to RapidIO interface. Reserved RapidIO transaction flow level 00 Lowest priority transaction flow 01 Next highest priority transaction flow 10 Highest priority level transaction flow 11 Reserved Applicable only to RapidIO interface, while SATRn[SBPATMU] is set. Follow PCI transaction ordering rules (elevate write priority one level over reads). Applicable only while SATRn[SBPATMU] is set. Source stride mode enable 0 Stride mode disabled. 1 Stride mode enabled. Ignored in basic mode (MRn[XFE] is cleared). Striding on the source address can be accomplished by enabling SATRn[SSME] and setting the desired stride size and distance in the SSRn. DMA source transaction interface 0000–1011 Reserved 1100 RapidIO interface 1101–1111 Reserved Applicable only to RapidIO interface, while SATRn[SBPATMU] is set. Description 3 4–5 — STFLOWLVL 6 SPCIORDER 7 SSME 8–11 STRANSINT MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 15-17 DMA Controller Table 15-8. SATRn Field Descriptions (continued) Bits Name Description 12–15 SREADTTYPE DMA source transaction type. Reserved values will result in a programming error being detected and logged in SR[PE]. • Transaction type to run on RapidIO interface in ATMU bypass mode 0000–0001 Reserved 0010 I/O read home 0011 Reserved 0100 nread 0101–0110 Reserved 0111 maintenance read 1000–1111 Reserved • Transaction type to run on local address space - used even in non-ATMU bypass mode 0000–0001 Reserved 0011 Reserved 0100 Read, don’t snoop local processor 0101 Read, snoop local processor 0111 Read, unlock L2 cache line 1000–1111 Reserved 16–21 22–31 — ESAD Reserved Extended source address. ESAD[0–7] represent the target ID. ESAD[8–9] represent the two high-order bits of the local device offset over the RapidIO interface. Applicable only to RapidIO interface, while SATRn[SBPATMU] is set. 15.3.2.5 Source Address Registers (SARn) The source address registers, shown in Figure 15-10, contain the address from which the DMA controller reads data. In direct mode, if MRn[CDSM/SWSM] and MRn[SRW] are set, a write to this register simultaneously sets MRn[CS], starting a DMA transfer. Software must ensure that this is a valid address. 0 31 R SAD W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x2_1114, 0x2_1194, 0x2_1214, 0x2_1294 Figure 15-10. Source Address Registers (SARn) Table 15-9 describes the field of the SARn. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 15-18 Freescale Semiconductor Memory Map/Register Definition Table 15-9. SARn Field Descriptions Bits 0–31 Name SAD Description Source address. This register contains the source address of the DMA transfer. The contents are updated after every DMA write operation unless the final stride of a striding operation is less than the stride size, in which case it remains equal to the address from which the last stride began. 15.3.2.5.1 Source Address Registers for RapidIO Maintenance Reads (SARn) If RapidIO is the source of a transaction, the SARn registers are redefined as shown below. Several options exist for the packet type that can be specified. A maintenance read is one of many possible reads. Maintenance packets have an offset instead of an address. Figure 15-11 describes the SARn. 0 7 8 29 30 31 R HOP_COUNT W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x2_1114, 0x2_1194, 0x2_1214, 0x2_1294 CONFIG_OFFSET 0 0 Figure 15-11. Source Address Registers for RapidIO Maintenance Reads (SARn) Table 15-10 describes the fields of the SARn. Table 15-10. SARn Field Descriptions Bits 0–7 8–29 30–31 Name HOP_COUNT Description Maintenance packet hop_count as defined by the RapidIO Interconnect Specification 1.2 CONFIG_OFFSET Maintenance packet word offset as defined by the R apidIO Interconnect Specification 1.2 — Reserved 15.3.2.6 Destination Attributes Registers (DATRn) The destination attributes registers, shown in Figure 15-12, contain the transaction attributes for the DMA operation. Stride mode is enabled by setting DATRn[DSME]. Destination write transaction type is specified using the DWRITETTYPE field. ATMU bypass mode, which is applicable only for accesses to RapidIO, is enabled by setting DATRn[DBPATMU]. If DBPATMU is set, DTRANSINT must be set to RapidIO and attributes that would otherwise come from the ATMU must be specified in this register. If DBPATMU is zero, the target interface is derived from the local access ATMU mapping and the transaction is obtained from the value specified in DWRITETTYPE using the local address space category. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 15-19 DMA Controller 0 1 2 3 4 5 6 7 8 11 12 15 R W Reset 0 0 DBPATMU 0 DTFLOWLVL DPCI_ORDER DSME DTRANSINT DWRITETTYPE 0000_0000_0000_0000 16 21 22 31 R W Reset Offset 0 0 0 0 0 0 EDAD 0000_0000_0000_0000 0x2_1118, 0x2_1198, 0x2_1218, 0x2_1298 Figure 15-12. Destination Attributes Registers (DATRn) Table 15-11 describes the fields of the DATRn. Table 15-11. DATRn Field Descriptions Bits 0–1 2 Name — DBPATMU Reserved Bypass ATMU for this DMA operation 0 Route the operation through the ATMU outbound windows. DATRn[DWRITETTYPE] should specify a local address space transaction type. 1 Bypass ATMU. Never generate an address match. Always use the attributes in the destination attribute registers. Route the transaction to the interface specified in the DATRn[DTRANSINT] field. Applicable only to RapidIO interface. Reserved RapidIO transaction flow level 00 Lowest priority transaction flow 01 Next highest priority transaction flow 10 Highest priority transaction flow 11 Reserved Applicable only to RapidIO interface, while DATRn[DBPATMU] is set. Follow PCI transaction ordering rules on RapidIO (elevate write priority one level over reads). Applicable only while DATRn[DBPATMU] is set. Destination stride mode enable 0 Stride mode disabled. 1 Stride mode enabled. Ignored in basic mode (MR n[XFE] is cleared). Striding on the destination address can be accomplished by setting DSME and setting the desired stride size and distance in DSRn. DMA destination transaction interface 0000–1011 Reserved 1100 RapidIO interface 1101–1111 Reserved Applicable only while DATRn[DBPATMU] is set. Description 3 4–5 — DTFLOWLVL 6 DPCI_ORDER 7 DSME 8–11 DTRANSINT MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 15-20 Freescale Semiconductor Memory Map/Register Definition Table 15-11. DATRn Field Descriptions (continued) Bits 12–15 Name Description DWRITETTYPE DMA destination transaction type. Reserved values will result in a programming error being detected and logged in SR[PE]. • Transaction type to run on RapidIO interface in ATMU bypass mode 0000 Reserved 0001 Flush 0010 Reserved 0011 SWRITE for all but last, NWRITE_R for last transaction 0100 NWRITE for all but last, NWRITE_R for last transaction 0101 NWRITE_R 0110 Message of size 8, 16, 32, 64, 128 or 256 bytes. (Other message sizes produce boundedly undefined behavior.) 0111 MAINTENANCE write 1000–1111Reserved • Transaction type to run on local address space—Used even in non-ATMU bypass mode 0000–0011Reserved 0100 Write, don’t snoop local processor 0101 Write, snoop local processor 0110 Write, allocate L2 cache line 0111 Write, allocate and lock L2 cache line 1000–1111 Reserved — EDAD Reserved Extended destination address. Applicable only for RapidIO interface with DATRn[DBPATMU] set. EDAD[0–7] represents the RapidIO target ID. EDAD[8–9] is defined as follows, subject to the transaction type: Message: EDAD[8–9] represents the value for the mbox field in the message packet. Other: EDAD[8–9] represents the two high-order bits of the local device offset. 16–21 22–31 15.3.2.7 Destination Address Registers (DARn) The destination address registers, shown in Figure 15-13, contain the addresses to which the DMA controller writes data. In direct mode, if MRn[SRW] is set and MRn[CDSM/SWSM] is cleared, a write to this register simultaneously sets MRn[CS], starting a DMA transfer. Software must ensure that this is a valid address. 0 31 R DAD W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x2_111C, 0x2_119C, 0x2_121C, 0x2_129C Figure 15-13. Destination Address Registers (DARn) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 15-21 DMA Controller Table 15-12 describes the field of the DARn. Table 15-12. DARn Field Descriptions Bits 0–31 Name DAD Description Destination address. This register contains the destination address of the DMA transfer. The contents are updated after every DMA write operation unless the final stride of a striding operation is less than the stride size, in which case it remains equal to the address from which the last stride began. 15.3.2.7.1 Destination Address Registers for RapidIO Maintenance Writes (DARn) If RapidIO is the destination of a transaction, the DARn registers are redefined as shown below. Several options exist for the transaction type that can be specified. There are a number of noncoherent write and flush types for address-based write transactions, and message types for port-based write transactions. Maintenance packets have an offset instead of an address. Figure 15-14 shows the DARn. 0 7 8 29 30 31 R HOP_COUNT W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x2_111C, 0x2_119C, 0x2_121C, 0x2_129C CONFIG_OFFSET 0 0 Figure 15-14. Destination Address Registers for RapidIO Maintenance Writes (DARn) Table 15-13 describes the fields of the DARn. Table 15-13. DARn Field Descriptions Bits 0–7 8–29 30–31 Name HOP_COUNT Description Maintenance packet hop count as defined by the RapidIO Interconnect Specification 1.2 CONFIG_OFFSET Maintenance packet word offset as defined by the RapidIO Interconnect Specification 1.2 — Reserved MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 15-22 Freescale Semiconductor Memory Map/Register Definition 15.3.2.8 Byte Count Registers (BCRn) The byte count register, shown in Figure 15-15, contains the number of bytes to transfer. 0 5 6 31 R0000 00 BC W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x2_1120, 0x2_11A0, 0x2_1220, 0x2_12A0 Figure 15-15. Byte Count Registers (BCRn) Table 15-14 describes the fields of the BCRn. Table 15-14. BCRn Field Descriptions Bits Name 0–5 6–31 — BC Reserved Byte count. Contains the number of bytes to transfer. The value in this register is decremented after each DMA read operation. The maximum transfer size is 226 – 1 Bytes. Description 15.3.2.9 Next Link Descriptor Address Registers (NLNDARn) The next link descriptor address registers, shown in Figure 15-16, contain the address for the next link descriptor in memory. Contents transferred to the current descriptor address registers become effective for the current transfer in basic and extended chaining modes. 0 26 27 28 29 30 31 R NLNDA W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x2_1128, 0x2_11A8, 0x2_1228, 0x2_12A8 0 NDEOSIE 00 EOLND Figure 15-16. Next Link Descriptor Address Registers (NLNDARn) Table 15-15 describes the fields of the NLNDARn registers. Table 15-15. NLNDARn Field Descriptions Bits 0–26 27 Name NLNDA — Description Next link descriptor address. Contains the next link descriptor address in memory. The descriptor must be aligned to a 32-byte boundary. Reserved MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 15-23 DMA Controller Table 15-15. NLNDARn Field Descriptions (continued) Bits 28 Name NDEOSIE Description Next descriptor end-of-segment interrupt enable 0 Do not generate an interrupt if the current DMA transfer for the current descriptor is finished. 1 Generate an interrupt if the current DMA transfer for the current descriptor is finished. Reserved End-of-links descriptor. This bit is ignored in direct mode. 0 This descriptor is not the last link descriptor in memory for this list. 1 This descriptor is the last link descriptor in memory for this list. If this bit is set, the DMA controller advances to the next list descriptor in memory if NLSDARn[EOLSD] is also set in extended mode. 29–30 31 — EOLND 15.3.2.10 Current List Descriptor Address Registers (CLSDARn) The current list descriptor address registers, shown in Figure 15-17, contain the current address of the list descriptor in memory in extended chaining mode. In extended chaining mode, software must initialize CLSDARn to point to the first list descriptor in memory. After finishing the last link descriptor in the current list, the DMA controller loads the contents of the next list descriptor address register into the current list descriptor address register. If NLSDARn[EOLSD] in the next list descriptor address register is clear, the DMA controller reads the new current list descriptor from memory to process that list. If EOLSD in the next list descriptor address register is set and the last link in the current list is finished all DMA transfers are complete. n 0 26 27 31 R CLSDA W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x2_1134, 0x2_11B4, 0x2_1234, 0x2_12B4 0 00 0 0 Figure 15-17. Current List Descriptor Address Registers (CLSDARn) Table 15-16 describes the fields of the CLSDARn. Table 15-16. CLSDARn Field Descriptions Bits 0–26 27–31 Name CLSDA — Description Current list descriptor address. Contains the current list descriptor address of the buffer descriptor in memory in extended chaining mode. The descriptor must be aligned to a 32-byte boundary. Reserved MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 15-24 Freescale Semiconductor Memory Map/Register Definition 15.3.2.11 Next List Descriptor Address Registers (NLSDARn) The next list descriptor address register, shown in Figure 15-18, contains the address for the next list descriptor in memory. If the contents are transferred to the current list descriptor address register they become effective for the current transfer in extended chaining mode. 0 26 27 31 R NLSDA W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x2_113C, 0x2_11BC, 0x2_123C, 0x2_12BC 0 0 0 0 EOLSD Figure 15-18. Next List Descriptor Address Registers (NLSDARn) Table 15-17 describes the fields of the NLSDARn. Table 15-17. NLSDARn Field Descriptions Bits 0–26 27–30 31 Name Description NLSDA Next list descriptor address. Contains the next descriptor address of the buffer descriptor in memory. The descriptor must be aligned on a 32-byte boundary. — Reserved EOLSD End-of-lists descriptor. This bit is ignored in direct mode. 0 This list descriptor is not the last list descriptor in memory. 1 This list descriptor is the last list descriptor in memory. If this bit is set, then the DMA controller halts after the last link descriptor transaction is finished. 15.3.2.12 Source Stride Registers (SSRn) The source stride register, shown in Figure 15-19, contains the stride size and distance. Note that the source stride information is loaded when a new list descriptor is read from memory. Therefore, the source stride register is applicable for all link descriptors in the new list. Changing the source stride information for a link requires that a new list be generated. 0 7 8 19 20 31 R0 0 0 0 0 0 W Reset Offset 0 0 SSS 0000_0000_0000_0000_0000_0000_0000_0000 0x2_1140, 0x2_11C0, 0x2_1240, 0x2_12C0 SSD Figure 15-19. Source Stride Registers (SSRn) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 15-25 DMA Controller Table 15-18 describes the fields of the SSRn. Table 15-18. SSRn Field Descriptions Bits 0–7 8–19 Name — Reserved Description SSS Source stride size. Number of bytes to transfer before jumping to the next address as specified in the source stride distance field. 20–31 SSD Source stride distance. The source stride distance in bytes from start byte to start byte. 15.3.2.13 Destination Stride Registers (DSRn) The destination stride register contains the stride size, and distance. Note that the destination stride information is loaded when a new list descriptor is read from memory. Therefore, the destination stride register is applicable for all link descriptors in the new list. Changing the destination stride information for a link requires that a new list be generated. Figure 15-20 describes the DSRn. 0 7 8 19 20 31 R00 0 0 0 0 W Reset Offset 0 0 DSS 0000_0000_0000_0000_0000_0000_0000_0000 0x2_1144, 0x2_11C4, 0x2_1244, 0x2_12C4 DSD Figure 15-20. Destination Stride Registers (DSRn) Table 15-19 describes the fields of the DSRn. Table 15-19. DSRn Field Descriptions Bits Name 0–7 8–19 — Reserved Description DSS Destination stride size. Number of bytes to transfer before jumping to the next address as specified in the destination stride distance field. 20–31 DSD Destination stride distance. The destination stride distance in bytes from start byte to start byte. 15.3.2.14 DMA General Status Register (DGSR) The DMA general status register combines all of the status bits from each channel into one register, including the enhanced DMA channel. This register is read-only. Figure 15-21 describes the DGSR. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 15-26 Freescale Semiconductor Memory Map/Register Definition 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 R TE0 0 W Reset 16 17 CH0 PE0 EOLNI0 CB0 EOSI0 EOLSI0 TE1 0 CH1 PE1 EOLNI1 CB1 EOSI1 EOLSI1 0000_0000_0000_0000 18 19 20 21 22 23 24 25 26 27 28 29 30 31 R TE2 0 W Reset Offset CH2 PE2 EOLNI2 CB2 EOSI2 EOLSI2 TE3 0 CH3 PE3 EOLNI3 CB3 EOSI3 EOLSI3 0000_0000_0000_0000 0x2_1300 Figure 15-21. DMA General Status Register (DGSR) Table 15-20 describes the fields of the DGSR. Table 15-20. DGSR Field Descriptions Bits 0 Name TE0 Description Transfer error, Channel 0 0 Normal operation 1 An error condition occurred during the DMA transfer. Reserved Channel halted, Channel 0 Programming error, Channel 0 End-of-links interrupt, Channel 0 Channel busy, Channel 0 End-of-segment interrupt, Channel 0 End-of-lists/direct interrupt, Channel 0 Transfer error, Channel 1 0 Normal operation 1 An error condition occurred during the DMA transfer. Reserved Channel halted, Channel 1 Programming error, Channel 1 End-of-links interrupt, Channel 1 Channel busy, Channel 1 End-of-segment interrupt, Channel 1 End-of-lists/direct interrupt, Channel 1 1 2 3 4 5 6 7 8 — CH0 PE0 EOLNI0 CB0 EOSI0 EOLSI0 TE1 9 10 11 12 13 14 15 — CH1 PE1 EOLNI1 CB1 EOSI1 EOLSI1 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 15-27 DMA Controller Table 15-20. DGSR Field Descriptions (continued) Bits 16 Name TE2 Description Transfer error, Channel 2 0 Normal operation 1 An error condition occurred during the DMA transfer. Reserved Channel halted, Channel 2 Programming error, Channel 2 End-of-links interrupt, Channel 2 Channel busy, Channel 2 End-of-segment interrupt, Channel 2 End-of-lists/direct interrupt, Channel 2 Transfer error, Channel 3 0 Normal operation 1 An error condition occurred during the DMA transfer. Reserved Channel halted, Channel 3 Programming error, Channel 3 End-of-links interrupt, Channel 3 Channel busy, Channel 3 End-of-segment interrupt, Channel 3 End-of-lists/direct interrupt, Channel 3 17 18 19 20 21 22 23 24 — CH2 PE2 EOLNI2 CB2 EOSI2 EOLSI2 TE3 25 26 27 28 29 30 31 — CH3 PE3 EOLNI3 CB3 EOSI3 EOLSI3 15.4 Functional Description This section describes the function of the the DMA controller. 15.4.1 DMA Channel Operation All DMA channels support two different modes of operation; a basic mode (MRn[XFE] is cleared) and an extended mode (MRn[XFE] is set). In both modes, a channel can be activated by clearing and setting MRn[CS], or via the single-write start mode using MRn[CDSM/SWSM] and MRn[SRW], or via an external control mode using MRn[ECS_EN]. In basic mode, the channel can be programmed in basic direct mode or basic chaining mode. In extended mode, the channel can be programmed in extended direct mode or extended chaining mode. Extended mode provides more capabilities, such as extended descriptor chaining, striding capabilities, and a more flexible descriptor structure. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 15-28 Freescale Semiconductor Functional Description The DMA controller supports misaligned transfers for both the source and destination addresses. In order to maximize performance, the source and destination engines align the source and destination addresses to a 64-byte boundary. The DMA always reads/writes the maximum number of bytes for a given transfer as described by the capability inputs of the DMA controller except for globally coherent transactions that use the size of the cache coherence granule as described by the mode select input. Using 256 bytes over the RapidIO interface reduces packet overhead which translates to increased bandwidth utilization through the interface. The DMA controller supports bandwidth control, which prevents a channel from consuming all the data bandwidth in the controller. Each channel is allowed to consume the bandwith of the shared resources as specified by the bandwidth control value. After the channel uses its allotted bandwidth, the arbiter grants the next channel access to the shared resources. The arbitration is round robin between the channels. This feature is also used to implement the external control pause feature. If the external control start and pause are enabled in the MRn, the channel enters a paused state after transferring the data described in the bandwidth control. External control can restart the channel from a paused state. The DMA controller is designed to support RapidIO transaction types, including various priority level support. The DMA controller offers additional features from previous generations in which the reads can be mapped to globally coherent (IO_READ), non-coherent (NREAD), or maintenance reads. In addition, the writes can be mapped to coherent (flush with data) and non-coherent (NWRITE, NWRITE_R) writes, messages, and maintenance writes. The DMA programming model permits software to program each DMA engine independently to interrupt on completed segment, chain, or error. It also provides the capability for software to resume the DMA engine from a hardware halted condition by setting the channel continue bit, MRn[CC]. See Table 15-21 on page 15-36 for more complete descriptions of the channel states and state transitions. 15.4.1.1 Basic DMA Mode Transfer This mode is primarily included for backward compatibility with existing DMA controllers which use a simple programming model. This is the default mode out of reset. The different modes of operation under the basic mode are explained in the following sections. 15.4.1.1.1 Basic Direct Mode In basic direct mode, the DMA controller does not read descriptors from memory, but instead uses the current parameters programmed in the DMA registers to start the DMA transfer. Software is responsible for initializing SARn, SATRn, DARn, DATRn, and BCRn registers. The DMA transfer is started when MRn[CS] is set. Software is expected to program all the appropriate registers before setting MRn[CS] to a 1. The transfer is finished after all the bytes specified in the MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 15-29 DMA Controller byte count register have been transferred or if an error condition occurs. The sequence of events to start and complete a transfer in basic direct mode is as follows: 1. Poll the channel state (see Table 15-21), to confirm that the specific DMA channel is idle. 2. Initialize SARn, SATRn, DARn, DATRn and BCRn. 3. Set the mode register channel transfer mode bit, MRn[CTM], to indicate direct mode. Other control parameters may also be initialized in the mode register. 4. Clear then set the mode register channel start bit, MRn[CS], to start the DMA transfer. 5. SRn[CB] is set by the DMA controller to indicate the DMA transfer is in progress. 6. SRn[CB] is automatically cleared by the DMA controller after the transfer is finished, or if the transfer is aborted (MRn[CA] transitions from a 0 to 1), or if a transfer error occurs. 7. End of segment interrupt is generated if MRn[EOSIE] is set. 15.4.1.1.2 Basic Direct Single-Write Start Mode In basic direct single-write start mode, the DMA controller does not read descriptors from memory, but instead uses the current parameters programmed in the DMA registers to start the DMA transfer. Software is responsible for initializing the SATRn, DATRn, and BCRn registers. Setting MRn[SRW] configures the DMA controller to begin the DMA transfer either when SARn is written or when DARn is written, determined by the state of MRn[CDSM/SWSM]. Writing to SARn initiates the DMA transfer if MRn[CDSM/SWSM] is set. Writing to DARn initiates the DMA transfer if MRn[CDSM/SWSM] is cleared. The DMA controller automatically sets the channel start bit, MRn[CS]. Software is expected to program all the appropriate registers before writing the source or destination address registers. The transfer is finished after all the bytes specified in the byte count register have been transferred or if an error condition occurs. The sequence of events to start and complete a transfer in single-write start basic direct mode is as follows: 1. Poll the channel state (see Table 15-21), to confirm that the specific DMA channel is idle. 2. Initialize the source attributes (SATRn), DATRn, and BCRn registers. 3. Set the mode register channel transfer mode bit, MRn[CTM], and the single-write start direct mode bit, MRn[SRW]. Other control parameters may also be initialized in the mode register. Set MRn[CDSM/SWSM] for transfers started using SARn. Clear MRn[CDSM/SWSM] for transfers started using the DARn. 4. A write to the source or destination address register starts the DMA transfer and automatically sets MRn[CS]. 5. SRn[CB] is set by the DMA controller to indicate the DMA transfer is in progress. 6. SRn[CB] is automatically cleared by the DMA controller after the transfer is finished, or if the transfer is aborted (MRn[CA] transitions from a 0 to 1), or if a transfer error occurs. 7. End of segment interrupt is generated if MRn[EOSIE] is set. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 15-30 Freescale Semiconductor Functional Description 15.4.1.1.3 Basic Chaining Mode In basic chaining mode, software must first build link descriptor segments in memory. Then the current link descriptor address register must be initialized to point to the first descriptor in memory. The DMA controller loads descriptors from memory prior to a DMA transfer. The DMA controller begins the transfer according to the link descriptor information loaded for the segment. After the current segment is finished, the DMA controller reads the next link descriptor from memory and begins another DMA transfer. The transfer is finished if the current link descriptor is the last one in memory or if an error condition occurs. The sequence of events to start and complete a transfer in chaining mode is as follows: 1. Build link descriptor segments in memory. See Section 15.4.4, “DMA Descriptors,” for more information on descriptors. 2. Poll the channel state (see Table 15-21), to confirm that the specific DMA channel is idle. 3. Initialize CLNDARn to point to the first link descriptor in memory. 4. Clear the mode register channel transfer mode bit, MRn[CTM], as well as MRn[XFE], to indicate basic chaining mode. Other control parameters may also be initialized in the mode register. 5. Clear, then set the mode register channel start bit, MRn[CS], to start the DMA transfer. 6. SRn[CB] is set by the DMA controller to indicate the DMA transfer is in progress. 7. SRn[CB] is automatically cleared by the DMA controller after finishing the transfer of the last descriptor segment, or if the transfer is aborted (MRn[CA] transitions from a 0 to 1), or if an error occurs during any of the transfers. 15.4.1.1.4 Basic Chaining Single-Write Start Mode Basic chaining single-write start mode allows a chain to be started by writing the current link descriptor address register (CLNDARn). Setting MRn[CDSM/SWSM] in the mode register causes MRn[CS] to be automatically set when the current link descriptor address register is written. The sequence of events to start and complete a chain using single-write start mode is as follows: 1. Set the mode register current descriptor start mode bit, MRn[CDSM/SWSM], and the extended features enable bit MRn[XFE]. Also, clear the channel transfer mode bit, MRn[CTM]. This initialization indicates basic chaining and single-write start mode. Also other control parameters may be initialized in the mode register. 2. Build link descriptor segments in memory. See Section 15.4.4, “DMA Descriptors,” for more information on descriptors. 3. Poll the channel state (see Table 15-21), to confirm that the specific DMA channel is idle. 4. Initialize CLNDARn to point to the first descriptor segment in memory. This write automatically causes the DMA controller to begin the link descriptor fetch and set MRn[CS]. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 15-31 DMA Controller 5. SRn[CB] is set by the DMA controller to indicate the DMA transfer is in progress. 6. SRn[CB] is automatically cleared by the DMA controller after finishing the transfer of the last descriptor segment, or if the transfer is aborted (MRn[CA] transitions from a 0 to 1), or if an error occurs during any of the transfers. 15.4.1.2 Extended DMA Mode Transfer The extended DMA mode also operates in chaining and direct mode. It offers additional capability over the basic mode by supporting striding and a more flexible descriptor structure. This additional functionality also requires a new and more complex programming model. The extended DMA mode is activated by setting MRn[XFE]. 15.4.1.2.1 Extended Direct Mode Extended direct mode has the same functionality as basic direct mode with the addition of stride capabilities. The bit settings are the same as in direct mode with the exception of the MRn[XFE] being set. Striding on the source address can be accomplished by setting SATRn[SSME] and setting the desired stride size and distance in SSRn. Striding on the destination address can be accomplished by setting DATRn[DSME] and setting the desired stride size and distance in DSRn. 15.4.1.2.2 Extended Direct Single-Write Start Mode Extended direct single-write start mode has the same functionality as the basic direct single-write start mode with the addition of stride capabilities. The bit settings are also the same with the exception of MRn[XFE] being set. Striding on the source address can be accomplished by setting SATRn[SSME] and setting the desired stride size and distance in SSRn. Striding on the destination address can be accomplished by setting DATRn[DSME] and setting the desired stride size and distance in DSRn. 15.4.1.2.3 Extended Chaining Mode In extended chaining mode, the software must first build list and link descriptor segments in memory. Then CLSDARn must be initialized to point to the first list descriptor in memory. The DMA controller loads list descriptors and link descriptors from memory prior to a DMA transfer. The DMA controller begins the transfer according to the link descriptor information loaded. Once the current link descriptor is finished, the DMA controller reads the next link descriptor from memory and begins another DMA transfer. If the current link descriptor is the last in the list, the DMA controller reads the next list descriptor in memory. The transfer is finished if the current link descriptor is the last one in the last list in memory or if an error condition occurs. The sequence of events to start and complete a transfer in extended chaining mode is as follows: 1. Build link and list descriptor segments in memory. See Section 15.4.4, “DMA Descriptors,” for more information on descriptors. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 15-32 Freescale Semiconductor Functional Description 2. Poll the channel state (see Table 15-21), to confirm that the specific DMA channel is idle. 3. Initialize CLSDARn to point to the first list descriptor in memory. 4. Clear the mode register channel transfer mode bit, MRn[CTM], to indicate chaining mode. MRn[XFE] must be set to indicate extended DMA mode. Other control parameters may also be initialized in the mode register. 5. Clear, then set the mode register channel start bit, MRn[CS], to start the DMA transfer. 6. SRn[CB] is set by the DMA controller to indicate the DMA transfer is in progress. 7. SRn[CB] is automatically cleared by the DMA controller after finishing the transfer of the last descriptor segment, or if the transfer is aborted (MRn[CA] transitions from a 0 to 1), or if an error occurs during any of the transfers. 15.4.1.2.4 Extended Chaining Single-Write Start Mode In the extended mode, the single-write start feature allows a chain to be started by writing the current list descriptor pointer. Setting MRn[CDSM/SWSM] causes MRn[CS] to be set automatically when CLSDARn is written. The sequence of events to start and complete an extended chain using single-write start mode is as follows: 1. Set MRn[CDSM/SWSM], MRn[CTM], and MRn[XFE] to indicate extended chaining and single-write start mode. Also other control parameters may be initialized in the mode register. 2. Build list and link descriptor segments in local memory. 3. Poll the channel state (see Table 15-21), to confirm that the specific DMA channel is idle. 4. Initialize the current list descriptor address register to point to the first list descriptor segment in memory. This write automatically causes the DMA controller to begin the list descriptor fetch and set MRn[CS]. 5. SRn[CB] is set by the DMA controller to indicate the DMA transfer is in progress. 6. SRn[CB] is automatically cleared by the DMA controller after finishing the transfer of the last descriptor segment, or if the transfer is aborted (MRn[CA] transitions from a 0 to 1), or if an error occurs during any of the transfers. 15.4.1.3 External Control Mode Transfer An external control can be used to control all DMA channels by setting MRn[EMS_EN]. The external control can control the DMA channel in the following transfer modes: • • • • Basic direct Basic chaining Extended direct Extended chaining MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 15-33 DMA Controller The external control and the DMA controller use a well defined protocol to communicate. The external control can start or pause a DMA transfer. The DMA controller acknowledges a DMA transfer in progress and also indicates a transfer completion. The pause feature can be enabled by setting MRn[EMP_EN]. MRn[BWC] specifies how much data to allow a specific channel to transfer before entering a paused state by clearing MRn[CS]. The channel can be restarted from a paused state by the external master. In chaining modes, the channel does not pause for descriptor fetch transfer. It only pauses during the actual data transfer. The following signals are defined for the external control interface: • • • DMA_DREQ, Indicates a DMA transfer start or restart from a pause request. The falling edge of DMA_DREQ sets MRn[CS]. DMA_DACK, Indicates a DMA transfer currently in progress. SRn[CB] is set. DMA_DDONE, Indicates that the DMA engine has completed the transfer. SRn[CB] is clear. Note, however, that write data may still be queued at the target interface or in the process of transfer on an external interface. Detailed descriptions of the external control interface are in Table 15-4. The timing diagram of the external control interface is shown in Figure 15-22. CLOCK DREQ DACK DDONE EMP_EN Transfer Start Transfer Done Transfer Start Transfer Pause Transfer Restart Transfer In Progress Figure 15-22. External Control Interface Timing 15.4.1.4 Channel Continue Mode for Cascading Transfer Chains The channel continue mode (enabled when MRn[CC] is set) offers software the flexibility of having the DMA controller get started on descriptors that have already been programmed while software continues to build more descriptors in memory. Software can set the end-of-links descriptor (EOLND) in basic mode, or end-of-lists descriptor (EOLSD) in extended mode, to cause the channel to go into a halted state while software continues to build other descriptors in memory. Software can then set CC to force hardware to continue where it left off. Channel continue is only meaningful for chaining modes, not direct mode. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 15-34 Freescale Semiconductor Functional Description If CC is set by software while the channel is busy with a transfer, the DMA controller finishes all transfers until it reaches the EOLND in basic mode or EOLSD in extended mode. The DMA controller then refetches the last link descriptor in basic mode, or the last list descriptor in extended mode and clears the channel continue bit. If EOLND or EOLSD is still set for their respective modes, the DMA controller remains in the idle state. If EOLND or EOLSD is not set, the DMA controller continues the transfer by refetching the new descriptor. If CC is set by software while the channel is not busy with a transfer, the DMA controller refetches the last link descriptor in basic mode, or the last list descriptor in extended mode and clears the channel continue bit. If EOLND or EOLSD is still set for their respective modes, the DMA controller remains in the idle state. If the EOLND or EOLSD bits are not set, the DMA controller continues the transfer by refetching the new descriptor. 15.4.1.4.1 Basic Mode On a channel continue, the descriptor at the current link descriptor address register (CLNDARn) is refetched to get the next link descriptor address field as updated by software. The channel halts if NLNDARn[EOLND] is still set. If EOLND is zero, the next link descriptor address is copied into CLNDARn and the channel continues with another descriptor fetch of the current link descriptor address. As a result, two link descriptor fetches always exist after channel continue before starting the first transfer. 15.4.1.4.2 Extended Mode On a channel continue, the descriptor at the current list descriptor (CLSDARn) address register is refetched to get the next list descriptor address field as updated by software. The channel halts if NLSDARn[EOLSD] is still set. If not, the next list descriptor address is copied into the CLSDARn register and the channel continues with another descriptor fetch of the current list descriptor address. As a result, two list descriptor fetches always exist after channel continue before the first link descriptor fetch and the first transfer. 15.4.1.5 Channel Abort Software can abort a previously initiated transfer by setting MRn[CA]. Once the DMA channel controller detects a zero-to-one transition of MRn[CA], it finishes the current sub-block transfer and halts all further activity. The controller then waits for all previously initiated transfers from the specified channel to drain and clears SRn[CB]. Successful completion of a software initiated abort request can be recognized by MRn[CA] being set and SRn[CB] being cleared. Obviously, if the controller was already halted because of an error condition (SRn[TE] is set), or the channel has completed all transfers, then SRn[CB] being cleared may not signify that the controller entered a halt state due to the abort request. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 15-35 DMA Controller 15.4.1.6 Bandwidth Control MRn[BWC] specifies how much data to allow a specific channel to transfer before allowing the next channel to use the shared data transfer hardware. This promotes equitable bandwidth allocation between channels. However, if only one channel is busy, hardware overrides the specified bandwidth control size value. The DMA controller allows a channel to transfer up to 1 Kbyte at a time when no other channel is active. 15.4.1.7 Channel State Table 15-21 defines the state of a channel based on the values of the channel start (MRn[CS]), channel busy (SRn[CB]), transfer error (SRn[TE]), and channel continue (MRn[CC]) bits. Table 15-21. Channel State Table MRn[CS] SRn[CB] 0 0 0 0 0 0 — 1 1 1 1 1 1 0 0 0 0 1 1 1 0 0 0 0 1 1 SRn[TE] MRn[CC] 0 0 1 1 0 0 1 0 0 1 1 0 0 0 1 0 1 0 1 — 0 1 0 1 0 1 Channel State Idle state. This is the state of the bits out of reset. Channel continue unexpected. Channel remains idle. Error occurred after software halted the channel. Channel continue unexpected. Channel remains in error halt state. Software halted channel. The channel was busy and software cleared MRn[CS]. Channel remains in halt state. The channel has encountered an error condition and it is trying to halt. Ready to start transfer (byte count > 0), or transfer complete (byte count = 0) Continue transfer (only meaningful in chaining mode, not direct mode). In direct mode, the channel continue has no effect. Error occurred during transfer. Channel remains in error halt state. Transfer in progress. Continue after reaching the end of list/link, or the first descriptor fetch after channel continue. 15.4.1.8 Illustration of Stride Size and Stride Distance If operating in stride mode, the stride size defines the amount of data to transfer before jumping to the next quantity of data as specified by the stride distance. The stride distance is added to the current base address to point to the next quantity of data to be transferred. Figure 15-23 illustrates the stride size and distance parameters. As shown, each time the stride distance is added to the base address, the resulting address becomes the new base address. This sequence repeats until the amount of data transferred equals the transfer size. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 15-36 Freescale Semiconductor Functional Description Stride Distance Stride Size Base Address New Base Address New Base Address Figure 15-23. Stride Size and Stride Distance 15.4.2 DMA Transfer Interfaces The DMA can be used to achieve data transfers across the entire memory map. 15.4.3 DMA Errors On a transfer error (uncorrectable ECC errors on memory accesses, parity errors on local bus or PCI, address mapping errors, for example), the DMA halts by setting SRn[TE] and generates an interrupt if MRn[EIE] is set. On a programming error, the DMA sets SRn[PE] and generates an interrupt if MRn[EIE] is set. The DMA controller detects the following programming errors: • • • • • Transfer started with a byte count of zero. Stride transfer started with a stride size of zero. Transfer started with a priority of three. Illegal type—Defined by SATRn[SREADTTYPE] and DATRn[DWRITETTYPE]. Used for the transfer. Invalid interface—Defined by SATRn[STRANSINT] and DATRn[DTRANSINT]. Used for the transfer when in ATMU bypass mode. 15.4.4 DMA Descriptors The DMA engine recognizes list descriptors and link descriptors. List descriptors connect lists of link descriptors. Link descriptors describe the DMA activity that is to take place. DMA descriptors are built in either local or remote memory and are connected by the next descriptor fields. Only link descriptors contain information for the DMA controller to transfer data. Software must ensure that each descriptor is 32-byte aligned. The last link descriptor in the last list in memory sets the EOLND bit in the next link descriptor; the next list descriptor fields indicating that these are the last descriptors in memory. Software initializes the current list descriptor address register to point to the first list descriptor in memory. The DMA controller traverses through the descriptor lists until the last link descriptor is met. For each link descriptor in the chain, the DMA controller starts a new DMA transfer with the control parameters specified by that descriptor. Link and list descriptor fetches always snoop the local memory space. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 15-37 DMA Controller NOTE Software must ensure that each descriptor is aligned on a 32-byte boundary. The last link descriptor in the last list in memory sets NLNDARn[EOLND] in the next link descriptor and NLSDARn[EOLSD] in the next list descriptor fields indicating that these are the last descriptors in memory. Software initializes the current list descriptor address register to point to the first list descriptor in memory. The DMA controller traverses through the descriptor lists until the last link descriptor is met as shown in Figure 15-24. For each link descriptor in the chain, the DMA controller starts a new DMA transfer with the control parameters specified by that descriptor. Table 15-22 summarizes the DMA list descriptors. Table 15-22. DMA List Descriptor Summary Descriptor Field Next list descriptor address First link descriptor address Source stride Destination stride Description Points to the next list descriptor in memory. After the DMA controller reads the descriptor from memory, this field is loaded into the next list descriptor address registers. Points to the first link descriptor in memory for this list. After the DMA controller reads the descriptor from memory, this field is loaded into the current link descriptor address registers. Contains the stride information used for the data source if striding is enabled for a link in the list. Contains the stride information used for the data destination if striding is enabled for a link in the list. Table 15-23 summarizes the DMA link descriptors. Table 15-23. DMA List Descriptor Summary Descriptor Field Source attributes register Source address Destination attributes register Destination address Next link descriptor address Description Contains source transaction attributes. Contains the source address of the DMA transfer. After the DMA controller reads the descriptor from memory, this field is loaded into the Source address register. Contains destination transaction attributes. Contains the destination address of the DMA transfer. After the DMA controller reads the descriptor from memory, this field is loaded into the destination address register. Points to the next link descriptor in memory. After the DMA controller reads the link descriptor from memory, this field is loaded into the next link descriptor address registers. Contains the number of bytes to transfer. After the DMA controller reads the descriptor from memory, this field is loaded into the byte count register. Byte count MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 15-38 Freescale Semiconductor Functional Description Figure 15-24 describes the DMA transaction flow. Descriptors In Memory DMA Controller Flow (Extended Mode) MRn[CS] is set by software after programming descriptors in memory. Fetch the first list descriptor using the CLSDARn address. Write NLSDAR n, CLNDAR n. Fetch the first link descriptor. Write programming model registers. Begin DMA transfer using SAR n, SATR n, DARn, DATRn. After transfer check if NLNDARn[EOLND] is set. Move NLNDAR n to CLNDAR n. Fetch the current link descriptor. Write programming model registers. Begin DMA transfer using SAR n, SATR n, DARn, DATRn. After DMA transfer check if NLNDAR n[EOLND] is set. Check if NLSDAR n[EOLSD] is set. Move NLSDARn to CLSDARn. Fetch the current list descriptor. Write NLSDAR n, CLNDAR n. Move NLNDAR n to CLNDAR n. Fetch the first link descriptor. Write programming model registers. Begin DMA transfer using SAR n, SATR n, DARn, DATRn. After DMA transfer check if NLNDAR n[EOLND] is set. Move NLNDAR n to CLNDAR n. Fetch the current link descriptor. Write programming model registers. Begin DMA transfer using SAR n, SATR n, DARn, DATRn. After DMA transfer check if NLNDAR n[EOLND] is set. Check if NLSDAR n[EOLSD] is set. Move NLSDARn to CLSDARn. Fetch the current list descriptor. Write NLSDAR n, CLNDAR n. Fetch the first link descriptor. Write programming model registers. Begin DMA transfer using SAR n, SATR n, DARn, DATRn. After DMA transfer check if NLNDAR n[EOLND] is set. Move NLNDAR n to CLNDAR n. Fetch the current link descriptor. Write programming model registers. Begin DMA transfer using SAR n, SATR n, DARn, DATRn. After DMA transfer check if NLNDAR n[EOLND] is set. Check if NLSDAR n[EOLSD] is set. Clear SRn[CB]. DMA Transfer Complete Start Current list descriptor Address Register List Descriptor Link Descriptor Link Descriptor . . . Link Descriptor (EOLND) List Descriptor Link Descriptor Link Descriptor . . Link Descriptor (EOLND) List Descriptor Link Descriptor Link Descriptor . . Link Descriptor (EOLND/EOLSD) Finish Figure 15-24. DMA Transaction Flow with DMA Descriptors MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 15-39 DMA Controller Figure 15-25 describes the format of the list descriptors. Offset 0x00 0x04 0x08 0x0C 0x10 0x14 0x18 0x1C R eserved Next list descriptor address Reserved First link descriptor address Source stride Destination stride Reserved Reserved Figure 15-25. List Descriptor Format Figure 15-26 describes the format of the link descriptors. Offset 0x00 0x04 0x08 0x0C 0x10 0x14 0x18 0x1C Source attributes Source address Destination attributes Destination address Reserved Next link descriptor address Byte count Reserved Figure 15-26. Link Descriptor Format 15.4.5 Limitations and Restrictions This section addresses some of the limitations and restrictions of the DMA controller and is intended to help software maximize the DMA performance and avoid DMA programming errors. The limitations of the DMA controller are the following: • Due to the limited number of buffers that the DMA controller can use, stride sizes less than 64 bytes should be avoided. Maximum utilization is obtained from strides greater than or equal to 256-bytes. However, small stride sizes can be used for scatter-gather functions. Coherent reads or writes are broken up into cache line accesses in the DMA. • MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 15-40 Freescale Semiconductor DMA System Considerations The DMA controller restrictions are as follows: • • • Setting the source or destination priority level (STFLOWLVL or DTFLOWLVL) to a value of three (0b11) is considered a programming error. All interface capabilities from where descriptors are being fetched must support read sizes 32-bytes or greater. If MRn[SAHE] is set, the source interface transfer size capability must be greater than or equal to MRn[SAHTS]. The source address must be aligned to the size specified by SAHTS. If MRn[DAHE] is set, the destination interface transfer size capability must be greater than or equal to MRn[DAHTS]. The destination address must be aligned to the size specified by DAHTS. Destination striding is not supported if MRn[DAHE] is set and source striding is not supported if MRn[SAHE] is set. If the DMA is programmed to send SWRITEs over RapidIO, the programmer must ensure that the destination address is double-word aligned and that the byte count is a double-word multiple. If the DMA is programmed to send messages over RapidIO, the programmer must ensure that the message length (BCRn[BC]) is 8, 16, 32, 64, 128 or 256 bytes. This can be achieved by setting the byte count register (BCR) to a power of 2 value equal to 8 or greater. Striding does not work if the destination transaction type is MESSAGE (DATR[DWRITETTYPE] = 0x0110) because messages have no memory addresses. As well, destination address hold should be disabled (MRn[DAHE] is cleared) unless the destination address hold transfer size indicates an 8-byte message (MRn[DAHTS] = 0x11). Software is responsible for disabling striding and DAHE, in this case, and for ensuring that the bandwidth control is large enough to support the desired message size. Failure to adhere to these restrictions results in boundedly undefined behavior. When DMA is used to issue maintenance reads and writes in bypass mode, the sum of the starting offset field in DAR and the byte count must not cause the offset to rollover. Failure to adhere to this restriction results in boundedly undefined behavior. • • • • • • 15.5 DMA System Considerations Figure 15-27 shows the most important data paths within the MPC8540. Many of these paths are served by captive DMA controllers and virtually all can be served by the general purpose four-channel DMA controller. This section provides information about how to make most effective use of these DMA channels including the following topics: • • DMA transaction initiators (masters) DMA targets, that is, data sources and destinations MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 15-41 DMA Controller • • Transparency of the bus interfaces to DMA operations What is useful as opposed to what is possible. For example, it is possible to address any internal register via the internal control bus, which means configuration and control registers can be DMA source or destination targets. However, the typical use of DMA functionality is to reduce host processor loading by moving large amounts of data with minimal CPU involvement. Using a general-purpose DMA controller to load small amounts of configuration data only makes sense in special circumstances (perhaps during system boot, for example). ECM e500 Coherency Module e500 core GMII RGMII TBI RTBI Three speed Ethernet controller DMA L2 cache 10/100/1G MAC Three speed Ethernet controller DMA DDR SDRAM controller DDR SDRAM bus 10/100/1G MAC I/O voltages supported = 2.5V Bus speeds: 100MHz(200)/ 133MHz(266)/150MHz(300)/166MHz(333) Support 64-bit wide+ECC (error injection) PCI/PCI-X bus PCI/PCI-X controller OCeaN Fast Ethernet controller 10/100 MAC TBI RTBI DMA RapidIO RapidIO endpoint DUART Serial In/Out DMA Ext DMA pins (4 channel) controller Open PIC I/O voltages supported = 3.3V I2C Register & control Boot sequencer Local bus controller Local bus I/O voltages supported = 3.3V Bus speeds: 66MHz/83MHz/ 100MHz/133MHz/166MHz Peripheral sizes 8/16/32-bit Figure 15-27. DMA Data Paths Table 15-24 lists all of the DMA controllers (both captive and general-purpose) on the device and the most likely DMA targets on and off-chip. The bus controllers themselves cannot initiate DMA MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 15-42 Freescale Semiconductor DMA System Considerations transfers; rather, they operate transparently with respect to both on- and off-chip DMA controllers. The codes in the table cells have the following meaning: • • • Y—Supported NR—Possible, but not recommended. Inefficient use of system resources. NS—Possible, but not supported. Resulting system behavior is not defined. Table 15-24. DMA Paths On-Chip Targets DMA Controllers Ethernet2 4 Channel3 Off-chip PCI-X Controller RapidIO Controller 1 2 2 Off-Chip Targets DDR Y Y Y Y Local Bus Y Y Y Y PCI/ DUART RapidIO PCI-X FIFO Y Y — Y Y Y Y — Y Y Y Y Ethernet Buffers Y NS Y Y L2 Y Y Y Y Configuration Registers1 NS NR NS NS On-chip Includes I C data register. Captive resource. Not available to external masters. 3 Can serve external masters. 15.5.1 Unusual DMA Scenarios The following is a description of unusual DMA paths including explanations of why some functional blocks cannot serve as DMA targets. The following topics are addressed: • • • • Transaction initiators (masters) DMA targets, that is, data sources or destinations Transparency of the bus controllers to DMA transactions What is useful as opposed to what is possible. For example, any register can be addressed through an internal control bus, which means configuration and control registers can be DMA targets. 15.5.1.1 DMA to e500 Core The L1 cache cannot be a direct DMA target because it cannot be directly addressed by software. However, DMA access into the L1 cache occurs indirectly if a block of memory that is cached in the L1 is specified as the DMA target. This effect is deterministic if the target memory block was locked into the L1 with cache locking instructions. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 15-43 DMA Controller 15.5.1.2 DMA to Ethernet The Ethernet controllers cannot serve as DMA targets because they have no suitable internal memory for this purpose. The Ethernet controllers have dedicated DMA channels to move data between the external transmit and receive buffers and the internal packet buffer. This dedicated channel is the only DMA service to the internal packet buffers. However, Ethernet ports can serve as DMA targets by using a general-purpose DMA controller to access the transmit and receive buffers defined by the Ethernet buffer descriptors. Because Ethernet data buffers are located in RAM outside of the Ethernet controllers, general-purpose DMA engines can move data to or from these memory regions. Also, because Ethernet controllers automatically read buffer descriptors and send (or load) data buffers, a DMA transfer into (or out of) these buffers is effectively a transfer into (or out of) the Ethernet ports. 15.5.1.3 DMA to Configuration and Control Registers Because any internal register can be addressed with the four-channel DMA controller, configuration and control registers throughout the device are valid DMA targets. However, the primary purpose of DMA—to reduce processor load by moving large blocks of data— is not served by DMA transfers of configuration data. For example, while it is possible to DMA into the I2C controller or programmable interrupt controller (PIC), doing so is extremely inefficient and is seldom beneficial in normal operation. The overhead of creating DMA descriptors far exceeds any savings in CPU cycles. 15.5.1.4 DMA to I2C The I2C controller is not transparent to DMA transfers. Observe the caveats listed in Section 15.5.1.3, “DMA to Configuration and Control Registers,” when accessing any I2C register, including the data register (I2CDR). 15.5.1.5 DMA to DUART The DUART provides complete and sophisticated DMA support which is described in Chapter 12, “DUART,” specifically Section 12.4.5, “FIFO Mode.” MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 15-44 Freescale Semiconductor Chapter 16 PCI/PCI-X Bus Interface The MPC8540 PCI/PCI-X interface complies with the PCI Local Bus Specification, Rev. 2.2 and the PCI-X Addendum to the PCI Local Bus Specification, Rev. 1.0a. It is beyond the scope of this manual to document the intricacies of the PCI and PCI-X buses. This chapter describes the PCI/PCI-X controller (referenced as PCI/X throughout this chapter) of this device and provides a basic description of the PCI and PCI-X bus operations. The specific emphasis is directed at how the MPC8540 implements the PCI and PCI-X buses. Designers of systems incorporating PCI/PCI-X devices should refer to the respective specification for a thorough description of the PCI/PCI-X buses. NOTE Much of the available PCI literature refers to a 16-bit quantity as a WORD and a 32-bit quantity as a DWORD. Note that this is inconsistent with the terminology in the rest of this manual where the terms ‘word’ and ‘double word’ refer to a 32-bit and 64-bit quantity respectively. Where necessary to avoid confusion, the precise number of bits or bytes is specified. 16.1 Introduction The PCI/X controller acts as a bridge between the PCI/X interface and the OCeaN switch fabric. Figure 16-1 is a high-level block diagram of the PCI/X controller. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 16-1 PCI/PCI-X Bus Interface PCI/X Interface Regs PCI Arb Regs Arb PCI-X Interface Fabric Request Data Out Data In OCeaN Physical Interface Figure 16-1. PCI/X Controller Block Diagram 16.1.1 Overview The PCI/X controller connects the OCeaN to the PCI/X bus, to which I/O components are connected. The PCI bus uses a 32- or 64-bit multiplexed address/data bus, plus various control and error signals. The PCI/X interface supports address and data parity with error checking and reporting. The PCI/X interface of the MPC8540 functions both as a master (initiator) and a target device. Internally, the design is divided into the following: • • • Data path blocks Control logic blocks Memory The data path blocks contain the queues, tables for transaction tracking, and ordering. The control blocks contain control logic and state-machines for buffer control, bus protocol, tag generation, and transaction resizing. The memory blocks are used solely for inbound and outbound data storage. This allows the MPC8540 to handle separate PCI/X transactions simultaneously. For example, consider the case where a burst-write transaction from the MPC8540 to another PCI device MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 16-2 Freescale Semiconductor Introduction terminates with a disconnect before finishing the transaction. If another PCI device is granted the PCI bus and requests a burst-read from local memory, the MPC8540, as a target, can accept the burst-read transfer. When the MPC8540 is granted mastership of the PCI bus, the burst-write transaction continues. There are two blocks of memory in the design: • • The inbound buffers The outbound read buffers combined with the outbound write buffers There are many blocks of control logic in the block. On the PCI/X side there are machines for PCI/X controller-initiated address and data tenures for inbound and outbound data, respectively. On the OCeaN side there are machines for fabric arbitration, outbound data, and inbound data. As an initiator, the MPC8540 supports read and write operations to the PCI/X memory space, the PCI/X I/O space, and the 256-byte PCI/X configuration space. As an initiator, the MPC8540 also supports generating PCI/X special-cycle and interrupt-acknowledge transactions. As a target, the MPC8540 supports read and write operations to local memory, and, when configured in agent mode, read and write operations to the internal PCI/X configuration registers. The MPC8540 can function as either a PCI/X host bridge (host mode) or a peripheral device on the PCI/X bus (agent mode). See Section 16.1.3.1.1, “Host Mode,” for more information. In agent mode, all of the PCI/X configuration registers in the MPC8540 can be programmed from the PCI/X bus. See Section 16.4.2.11.3, “PCI Configuration in Agent and Agent Lock Modes,” for more information. The PCI/X interface provides bus arbitration for the MPC8540 and up to five other PCI/X bus masters. The arbitration algorithm is a programmable two-level, round-robin priority selector. The on-chip PCI/X arbiter can operate in both host and agent modes or it can be disabled to allow for an external PCI/X arbiter. The MPC8540 also provides an address translation mechanism to map inbound PCI/X to OCeaN accesses and outbound OCeaN to PCI/X accesses. 16.1.1.1 MPC8540 as a PCI/X Initiator Upon detecting an OCeaN-to-PCI/X transaction, the MPC8540 requests the use of the PCI/X bus. For OCeaN-to-PCI/X bus write operations, the MPC8540 requests mastership of the PCI/X bus when the source completes the write operation to the OCeaN. For OCeaN-to-PCI/X read operations, the MPC8540 requests mastership of the PCI/X bus when it decodes that the access is for PCI/X address space. Once granted, the MPC8540 drives the address (PCI_AD[63:0]) and the bus command (PCI_C/BE[7:0]) signals. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 16-3 PCI/PCI-X Bus Interface The master part of the interface can initiate master-abort cycles, recognizes target-abort, target-retry, and target-disconnect cycles, and supports various device selection timings. The master interface does not run fast back-to-back or exclusive accesses. 16.1.1.2 MPC8540 as a PCI/X Target Upon detection of a PCI/X address phase, the MPC8540 decodes the address and bus command to determine if the transaction is within the local memory access boundaries. If the transaction is destined for local memory, the target interface latches the address, decodes the PCI/X bus command, and forwards the transaction to the OCeaN control unit. On writes to local memory, data is forwarded along with the byte enables (if applicable) to the internal control unit. On reads, the data is driven on the bus and the byte enables (if applicable) determine which byte lanes contain meaningful data. The target interface of the MPC8540 can issue target-abort, target-retry, and target-disconnect cycles. The target interface supports fast back-to-back transactions. The target interface uses the fastest device selection timing. As a PCI/X target, the MPC8540 can issue split response and split completion transactions. The MPC8540 supports data streaming to and from local memory. This means that data can flow between the MPC8540’s PCI/X interface and local memory as long as the internal buffers are not filled. 16.1.2 Features The following is a list of PCI features that are supported: • • • • • • • • • • • • • PCI interface 2.2 compatible 66- and 33-MHz support 64- and 32-bit PCI interface support on primary PCI port Host and agent mode support 64-bit dual address cycle (DAC) support On-chip arbitration with support for 5 high-priority request and grant signal pairs Accesses to all PCI memory and I/O address spaces PCI-to-memory and memory-to-PCI streaming Memory prefetching of PCI read accesses Posting of processor-to-PCI and PCI-to-memory writes Selectable snoop for inbound accesses PCI configuration registers PCI 3.3-V compatible MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 16-4 Freescale Semiconductor Introduction The following is a list of PCI-X features supported: • • • • • • • • • • • PCI-X revision 1.0a compatible Support for up to 133-MHz point-to-point connection Support for 32- and 64-bit interface 64-bit dual address cycle (DAC) support On-chip arbitration with support for 5 high-priority request and grant signal pairs Support for accesses to all PCI-X memory and I/O address spaces Support for four split transactions Complete allowable disconnect boundary (ADB) support All PCI-X ordering rules enforced Implementation of relaxed-ordering PCI-X 3.3-V compatible 16.1.3 Modes of Operation A number of parameters that affect the PCI/X controller modes of operation are determined at power-on reset (POR) by reset configuration signals as described in Chapter 4, “Reset, Clocking, and Initialization.” Table 16-1 provides a summary of these modes. Table 16-1. POR Parameters for PCI/X Controller Parameter Host/agent configuration PCI width selection PCI I/O impedance PCI arbiter enable Description Selects between host and agent mode for the PCI/X and RapidIO interfaces. Selects between 32-bit or 64-bit data bus width. Selects the impedance of the PCI/X I/O drivers Enables the on-chip PCI/X bus arbiter Section/Pag e 4.4.3.4/4-14 4.4.3.12/4-19 4.4.3.13/4-19 4.4.3.14/4-19 4.4.3.15/4-20 4.4.3.16/4-20 4.4.3.19/4-21 PCI debug mode enable Selects between normal operation or debug mode for PCI_AD[62:58]. PCI-X configuration PCI/X output hold Selects between PCI or PCI-X operation Selects the number of buffer delays for PCI output signals depending on whether PCI or PCI-X operation is selected. This provides flexibility in meeting minimum output hold specifications relative to SYSCLK for both PCI and PCI-X systems 16.1.3.1 Host/Agent Modes The PCI/X controller can function as either a PCI/X host bridge (host mode) or a peripheral device on the PCI bus (agent mode). The PCI/X controller can also operate in agent configuration lock mode. Note that host/agent mode selection is determined at power-up as summarized in Section 16.5.1, “Power-On Reset Configuration Modes.” MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 16-5 PCI/PCI-X Bus Interface 16.1.3.1.1 Host Mode When the device powers up in host mode, all inbound configuration accesses are ignored (and thus master aborted). See Section 16.5.1.1, “Host Mode,” for more information. 16.1.3.1.2 Agent Mode When the device powers up in agent mode, it acknowledges inbound configuration accesses. See Section 16.5.1.2, “Agent Mode,” for more information. Note that in PCI agent mode, the PCI/X controller ignores all PCI/X memory accesses except those to the memory-mapped registers) until inbound address translation is enabled. 16.1.3.1.3 Agent Configuration Lock Mode When the device powers up in agent configuration lock mode, it retries inbound configuration accesses until the ACL bit in the PCI bus function register is cleared. See Section 16.5.1.3, “Agent Configuration Lock Mode,” for more information. 16.1.3.2 PCI/X Bus Width (64-/32-Bit Bus) This input configures the PCI/X interface to be in 32-bit or 64-bit extended mode of operation. The initial value is determined by the value on the PCI_REQ64 power-on reset configuration signal. See Chapter 4, “Reset, Clocking, and Initialization,” and the MPC8540 Integrated Processor Hardware Specifications, for more information. 16.1.3.3 PCI/X Arbiter (Internal/External Arbiter) This input configures the on-chip PCI/X arbiter. The initial value is determined by the value on the PCI_GNT2 power-on reset configuration signal. See Chapter 4, “Reset, Clocking, and Initialization,” and the MPC8540 Integrated Processor Hardware Specifications. 16.1.3.4 PCI/X Bus Mode This input configures PCI or PCI-X mode on the PCI/X port. Note that this input does not support the three states of the PCIXCAP input defined in the PCI-X specification. It is sampled as either a 1 or a 0 during reset only. The initial value is determined by the value on the PCI_GNT4 power-on reset configuration signal. See Chapter 4, “Reset, Clocking, and Initialization,” and the MPC8540 Integrated Processor Hardware Specifications. 16.1.3.5 PCI/X Signal Output Hold Timing To meet minimum output hold specifications relative to SYSCLK for both PCI and PCI-X systems, the MPC8540 has a programmable output hold delay for bus signals. The initial value of the output hold delay is determined by the values on the LWE[0:1] power-on reset configuration MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 16-6 Freescale Semiconductor External Signal Descriptions signals. See Chapter 4, “Reset, Clocking, and Initialization,” and the MPC8540 Integrated Processor Hardware Specifications, for more information on these values and signal timing. 16.1.3.6 PCI/X Impedance The MPC8540 has a programmable impedance for PCI/X bus signals. The initial value is determined by the value on the PCI_GNT1 power-on reset configuration signal. See Chapter 4, “Reset, Clocking, and Initialization,” and the MPC8540 Integrated Processor Hardware Specifications, for more information. 16.2 External Signal Descriptions Figure 16-2 shows the external PCI/X signals. PCI/X Required Signals PCI_AD[31:0] Address and Data (37 Signals) 32 PCI_C/BE[3:0] PCI_PAR 1 PCI_FRAME 1 PCI_TRDY 1 PCI_IRDY 4 Interface Command (6 Signals) 1 PCI_STOP 1 PCI_DEVSEL PCI_IDSEL 1 PCI_PERR 1 PCI_SERR PCI_REQ[0:4] 1 5 PCI_GNT[0:4] 5 PCI_AD[63:32] PCI_C/BE[7:4] 32 4 1 PCI_REQ64 PCI_ACK64 1 1 1 Error Reporting (2 Signals) Arbitration (10 Signals) PCI/X Optional Signals 64-Bit Extension (39 Signals) PCI_PAR64 Figure 16-2. PCI/X Interface External Signals MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 16-7 PCI/PCI-X Bus Interface Table 16-2 contains the detailed descriptions of the external PCI/X interface signals. Table 16-2. PCI Interface Signals—Detailed Signal Descriptions Signal PCI_ACK64 I/ O Description I/ 64-bit transaction acknowledge.Indicates that the current target supports 64-bit transfers during the O data phase of the current transaction. O As an output for the bidirectional 64-bit transaction acknowledge, this signal operates as follows: State Asserted—Indicates that this PCI controller, as the target of a PCI transaction, may use the full 64-bit data bus for the data phase of the transaction. Meanin g Negated—Indicates that this PCI controller, as the target of a PCI transaction, may use only 32 bits of the data bus in servicing a data transfer. Timing Assertion/Negation—As specified by PCI Local Bus Specification Rev 2.2 or by PCI-X Addendum Rev 1.0a I As an input for the bidirectional 64-bit transaction acknowledge, this signal operates as follows: State Asserted—Indicates that the target of a PCI transaction may use the full 64-bit data bus for the data phase of the transaction. Meanin g Negated—Indicates that the target of a PCI transaction may only use 32 bits of the data bus during the data phase of the transaction. Timing Assertion/Negation—As specified by PCI Local Bus Specification Rev 2.2 or by PCI-X Addendum Rev 1.0a PCI_AD[63:0] I/ PCI address/data bus O O As outputs for the bidirectional PCI address/data bus, these signals operate as described below. State Asserted/Negated—Represents the physical address during the address phase of a PCI transaction. During the data phase(s) of a PCI transaction, the PCI address/data Meanin bus contain the data being written. g The PCI_AD[7:0] signals define the LSB and PCI_AD[63:56] define the MSB. Timing Assertion/Negation—As specified by PCI Local Bus Specification Rev 2.2 or by PCI-X Addendum Rev 1.0a I As inputs for the bidirectional PCI address/data bus, these signals operate as described below. State Asserted/Negated—Represents the address to be decoded as a check for device select during the address phase of a PCI transaction or the data being received during the Meanin data phase(s) of a PCI transaction. g The PCI_AD[7:0] signals define the LSB and PCI_AD[63:56] define the MSB. Timing Assertion/Negation—As specified by PCI Local Bus Specification Rev 2.2 or by PCI-X Addendum Rev 1.0a MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 16-8 Freescale Semiconductor External Signal Descriptions Table 16-2. PCI Interface Signals—Detailed Signal Descriptions (continued) Signal I/ O Description PCI_C/BE[7:0 I/ Command/byte enable. Command encodings are described in Section 16.4.2.2, “PCI Bus ] O Commands,” and Section 16.4.3.2, “PCI-X Command Encodings.” O As outputs for the bidirectional command/byte enable, these signals operate as described below. State Asserted/Negated—During the address phase, PCI_C/BE[7:0] define the bus command. Meanin Byte enables determine which byte lanes carry meaningful data for PCI bus data g phases. The PCI_C/BE0 signal applies to the LSB. Timing Assertion/Negation—As specified by PCI Local Bus Specification Rev 2.2 or by PCI-X Addendum Rev 1.0a I As inputs for the bidirectional command/byte enable, these signals operate as described below. State Asserted/Negated—During the address phase, PCI_C/BE[7:0] indicate the command Meanin that another master is sending. During the PCI bus data phase, PCI_C/BE[7:0] g indicate which byte lanes are valid. Timing Assertion/Negation—As specified by PCI Local Bus Specification Rev 2.2 or by PCI-X Addendum Rev 1.0a PCI_DEVSEL I/ Device select O O As outputs for the bidirectional device select, these signals operate as described below. State Asserted—Indicates that this PCI controller has decoded the address and is the target of the current access. Meanin g Negated—Indicates that this PCI controller has decoded the address and is not the target of the current access. Timing Assertion/Negation—As specified by PCI Local Bus Specification Rev 2.2 or by PCI-X Addendum Rev 1.0a I As inputs for the bidirectional device select, these signals operate as described below. State Asserted—Indicates that some PCI agent (other than this PCI controller) has decoded its address as the target of the current access. Meanin g Negated—Indicates that no PCI agent has been selected. Timing Assertion/Negation—As specified by PCI Local Bus Specification Rev 2.2 or by PCI-X Addendum Rev 1.0a MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 16-9 PCI/PCI-X Bus Interface Table 16-2. PCI Interface Signals—Detailed Signal Descriptions (continued) Signal PCI_FRAME I/ O I/ Frame O O As outputs for the bidirectional frame, these signals operate as described below. State Asserted—Indicates that this PCI controller, acting as a PCI master, is initiating a bus transaction. While PCI_FRAME is asserted, data transfers may continue. Meanin g Negated—If PCI_IRDY is asserted, indicates that the PCI transaction is in the final data phase; if PCI_IRDY is negated, indicates that the PCI bus is idle. Timing Assertion/Negation—As specified by PCI Local Bus Specification Rev 2.2 or by PCI-X Addendum Rev 1.0a I As inputs for the bidirectional frame, these signals operate as described below. State Asserted—Indicates that another PCI master is initiating a bus transaction. Meanin Negated—Indicates that the transaction is in the final data phase or that the bus is idle. g Timing Assertion/Negation—As specified by PCI Local Bus Specification Rev 2.2 or by PCI-X Addendum Rev 1.0a PCI_GNT[4:0] O PCI bus grant. Output signals on this PCI controller when the arbiter is enabled. When the arbiter is disabled, PCI_GNT0 is an input. Note that PCI_GNT[n] is a point-to-point signal. Every master has its own bus grant signal. Also, note that these signals are also used as reset configuration signals in the MPC8540 as described in Section 4.4.3, “Power-On Reset Configuration." State Asserted—Indicates that this PCI controller granted control of the PCI bus to agent n. Meanin Negated—Indicates that this PCI controller did not grant control of the PCI bus to agent n. g Timing Assertion/Negation—As specified by PCI Local Bus Specification Rev 2.2 or by PCI-X Addendum Rev 1.0a PCI_IDSEL I Initialization device select. Used as a chip select during configuration read and write transactions. State Asserted—Indicates this PCI controller is being selected as a target of a configuration read or write transactions. Meanin g Negated—Indicates this PCI controller is not being selected as a target of configuration read or write transactions. Timing Assertion/Negation—As specified by PCI Local Bus Specification Rev 2.2 or by PCI-X Addendum Rev 1.0a Description MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 16-10 Freescale Semiconductor External Signal Descriptions Table 16-2. PCI Interface Signals—Detailed Signal Descriptions (continued) Signal PCI_IRDY I/ O I/ Initiator ready O O As outputs for the bidirectional initiator ready, these signals operate as described below. State Asserted—Indicates that this PCI controller, acting as a PCI master, can complete the current data phase of a PCI transaction. During a write, this PCI controller asserts Meanin PCI_IRDY to indicate that valid data is present on PCI_AD[63:0]. During a read, g this PCI controller asserts PCI_IRDY to indicate that it is prepared to accept data. Negated—Indicates that the PCI target needs to wait before this PCI controller, acting as a PCI master, can complete the current data phase. During a write, this PCI controller negates PCI_IRDY to insert a wait cycle when it cannot provide valid data to the target. During a read, this PCI controller negates PCI_IRDY to insert a wait cycle when it cannot accept data from the target. Timing Assertion/Negation—As specified by PCI Local Bus Specification Rev 2.2 or by PCI-X Addendum Rev 1.0a I As inputs for the bidirectional initiator ready, these signals operate as described below. State Asserted—Indicates another PCI master can complete the current data phase of a transaction. Meanin g Negated—If PCI_FRAME is asserted, indicates a wait cycle from another master. If PCI_FRAME is negated, indicates the PCI bus is idle. Timing Assertion/Negation—As specified by PCI Local Bus Specification Rev 2.2 or by PCI-X Addendum Rev 1.0a PCI_PAR I/ PCI parity. O O As outputs for the bidirectional PCI parity, these signals operate as described below. State Asserted—Indicates odd parity across PCI_AD[31:0] and PCI_C/BE[3:0] during address Meanin and data phases. g Negated—Indicates even parity across PCI_AD[31:0] and PCI_C/BE[3:0] during address and data phases. Timing Assertion/Negation—As specified by PCI Local Bus Specification Rev 2.2 or by PCI-X Addendum Rev 1.0a I As inputs for the bidirectional PCI parity, these signals operate as described below. State Asserted—Indicates odd parity driven by another PCI master or the PCI target during read data phases. Meanin g Negated—Indicates even parity driven by another PCI master or the PCI target during read data phases. Timing Assertion/Negation—As specified by PCI Local Bus Specification Rev 2.2 or by PCI-X Addendum Rev 1.0a Description MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 16-11 PCI/PCI-X Bus Interface Table 16-2. PCI Interface Signals—Detailed Signal Descriptions (continued) Signal PCI_PAR64 I/ O Description I/ Upper DWORD parity. The even parity bit that protects the upper 32 bits of data and upper 4 bits O of command/byte enable. O As outputs for the bidirectional Upper DWORD Parity, these signals operate as described below. State Asserted—Indicates odd parity across the PCI_AD[63:32] and PCI_C/BE[7:4] signals Meanin during address and data phases. g Negated—Indicates even parity across the PCI_AD[63:32] and PCI_C/BE[7:4] signals during address and data phases. Timing Assertion/Negation—As specified by PCI Local Bus Specification Rev 2.2 or by PCI-X Addendum Rev 1.0a I As inputs for the bidirectional Upper DWORD Parity, these signals operate as described below. State Asserted—Indicates odd parity driven by another PCI master or the PCI target during read data phases. Meanin g Negated—Indicates even parity driven by another PCI master or the PCI target during read data phases. Timing Assertion/Negation—As specified by PCI Local Bus Specification Rev 2.2 or by PCI-X Addendum Rev 1.0a PCI_PERR I/ PCI parity error O O As outputs for the bidirectional PCI parity error, these signals operate as described below. State Asserted—Indicates that this PCI controller, acting as a PCI agent, detected a data parity error. (Driven by the PCI initiator on reads; driven by the PCI target on writes.) Meanin g Negated—Indicates no error. Timing Assertion/Negation—As specified by PCI Local Bus Specification Rev 2.2 or by PCI-X Addendum Rev 1.0a I As inputs for the bidirectional PCI parity error, these signals operate as described below. State Asserted—Indicates that another PCI agent detected a data parity error while this PCI controller was sourcing data (this PCI controller was acting as the PCI initiator Meanin during a write, or was acting as the PCI target during a read). g Negated—Indicates no error. Timing Assertion/Negation—As specified by PCI Local Bus Specification Rev 2.2 or by PCI-X Addendum Rev 1.0a Note: If a parity error occurs on the last data beat of a PCI-X transaction with inbound data (outbound read with split completion data or inbound write), the MPC8540 asserts PERR longer than permitted by the PCI-X specification. This condition may cause a subsequent transaction to erroneously detect a parity error. The condition may be remedied by placing a stronger pull-up resistors on the PERR signal. For example: • 1 Kohm for 133 MHz point-to-point operation • 2 Kohm for 66 MHz operation • 4 Kohm for 33 MHz operation. These values are stronger than the specification allows (5 Kohm minimum for pull-up resistors). MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 16-12 Freescale Semiconductor External Signal Descriptions Table 16-2. PCI Interface Signals—Detailed Signal Descriptions (continued) Signal PCI_REQ [4:0] I/ O I Description PCI bus request. Input signals on this PCI controller when the arbiter is enabled. When the arbiter is disabled, PCI_REQ[0] is an output. Note that PCI_REQ[n] is a point-to-point signal. Every master has its own bus request signal. Following is the state meaning for the PCI_REQ[n] input. State Asserted—Indicates that agent n is requesting control of the PCI bus to perform a transaction. Meanin g Negated—Indicates that agent n does not require use of the PCI bus. Timing Assertion/Negation—As specified by PCI Local Bus Specification Rev 2.2 or by PCI-X Addendum Rev 1.0a PCI_REQ64 I/ 64-bit transaction request. Indicates that the current master desires to transfer data using 64-bit O transfers. Also used as a reset configuration signal in the MPC8540 as described in Section 4.4.3, “Power-On Reset Configuration.” O As an output for the bidirectional 64-bit transaction request, this signal operates as follows: State Asserted—Indicates that this PCI controller, as the master of a PCI transaction, desires to use all 64 bits. Meanin g Negated—Indicates that this PCI controller, as the master of a PCI transaction, uses only 32 bits of the data bus in servicing a data transfer. Timing Assertion/Negation—As specified by PCI Local Bus Specification Rev 2.2 or by PCI-X Addendum Rev 1.0a I As an input for the bidirectional 64-bit transaction request, this signal operates as described below. State Asserted—Indicates that the master of a PCI transaction is requesting to use the full 64-bit data bus for the data phase of the transaction. Meanin g Negated—Indicates that the master of a PCI transaction uses only 32 bits of the data bus during the data phase of the transaction. Timing Assertion/Negation—As specified by PCI Local Bus Specification Rev 2.2 or by PCI-X Addendum Rev 1.0a PCI_SERR I/ PCI system error O O As outputs for the bidirectional PCI system error, these signals operate as described below. State Asserted—Indicates that an address parity error, a target-abort (when this PCI controller is acting as the initiator), or some other system error (where the result is a Meanin catastrophic error) was detected. g Negated—Indicates no error. Timing Assertion/Negation—As specified by PCI Local Bus Specification Rev 2.2 or by PCI-X Addendum Rev 1.0a I As inputs for the bidirectional PCI system error, these signals operate as described below. State Asserted—Indicates that a target (other than this PCI controller) has detected a catastrophic error. Meanin g Negated—Indicates no error. Timing Assertion/Negation—As specified by PCI Local Bus Specification Rev 2.2 or by PCI-X Addendum Rev 1.0a MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 16-13 PCI/PCI-X Bus Interface Table 16-2. PCI Interface Signals—Detailed Signal Descriptions (continued) Signal PCI_STOP I/ O I/ Stop. O O As outputs for the bidirectional stop, these signals operate as described below. State Asserted—Indicates that this PCI controller, acting as a PCI target, is requesting that the initiator stop the current transaction. Meanin g Negated—Indicates that the current transaction can continue. Timing Assertion/Negation—As specified by PCI Local Bus Specification Rev 2.2 or by PCI-X Addendum Rev 1.0a I As inputs for the bidirectional stop, these signals operate as described below. State Asserted—Indicates that a target is requesting that the PCI initiator stop the current transaction. Meanin g Negated—Indicates that the current transaction can continue. Timing Assertion/Negation—As specified by PCI Local Bus Specification Rev 2.2 or by PCI-X Addendum Rev 1.0a PCI_TRDY I/ Target ready. O O As outputs for the bidirectional target ready, these signals operate as described below. State Asserted—Indicates that this PCI controller, acting as a PCI target, can complete the current data phase of a PCI transaction. During a read, this PCI controller asserts Meanin PCI_TRDY to indicate that valid data is present on PCI_AD[31:0]. During a write, g this PCI controller asserts PCI_TRDY to indicate that it is prepared to accept data. Negated—Indicates that the PCI initiator needs to wait before this PCI controller, acting as a PCI target, can complete the current data phase. During a read, this PCI controller negates PCI_TRDY to insert a wait cycle when it cannot provide valid data to the initiator. During a write, this PCI controller negates PCI_TRDY to insert a wait cycle when it cannot accept data from the initiator. Timing Assertion/Negation—As specified by PCI Local Bus Specification Rev 2.2 or by PCI-X Addendum Rev 1.0a I As inputs for the bidirectional target ready, these signals operate as described below. State Asserted—Another PCI target is able to complete the current data phase of a transaction. Meanin g Negated—Indicates a wait cycle from another target. Timing Assertion/Negation—As specified by PCI Local Bus Specification Rev 2.2 or by PCI-X Addendum Rev 1.0a Description MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 16-14 Freescale Semiconductor Memory Map/Register Definitions 16.3 Memory Map/Register Definitions The PCI/X interface unit of the MPC8540 supports the following register types: • Memory-mapped registers—these registers control PCI address translation, PCI error management, and PCI configuration register access on the MPC8540. These registers are described in Section 16.3.1, “PCI/X Memory Mapped Registers,” and its subsections. PCI/X configuration registers contained within the PCI/X configuration header—these registers are specified by the PCI bus specification for every PCI device. These registers are described in Section 16.3.2, “PCI/X Configuration Header,” and its subsections. • 16.3.1 PCI/X Memory Mapped Registers The PCI/X memory mapped registers are accessed by reading and writing to an address comprised of the base address (specified in the CCSRBAR on the local side or the PCSRBAR on the PCI/X side) plus the offset of the specific register to be accessed. Note that all memory-mapped registers (except the PCI/X configuration data register, PCI CFG_DATA) must only be accessed as 32-bit quantities. Table 16-3 lists the memory-mapped registers. Table 16-3. PCI/X Memory-Mapped Register Map Offset Register Access Reset Section/Page PCI/X Configuration Access Registers 0x0_8000 0x0_8004 0x0_8008 0x0_800C– 0x0_8BFC CFG_ADDR—PCI/X configuration address CFG_DATA—PCI/X configuration data INT_ACK—PCI/X interrupt acknowledge Reserved R/W R/W R — PCI/X ATMU Registers—Outbound and Inbound 0x0_8C00–0x0_8C3C–Outbound Window 0 (default) 0x0_8C00 0x0_8C04 0x0_8C08 0x0_8C0C 0x0_8C10 0x0_8C14– 0x0_8C1C POTAR0—PCI/X outbound window 0 (default) translation address register POTEAR0—PCI/X outbound window 0 (default) translation extended address register Reserved Reserved POWAR0—PCI/X outbound window 0 (default) attributes register Reserved R/W R/W — — R/W — 0x0000_0000 0x0000_0000 — — 0x8004_401F — 16.3.1.2.4/16-22 16.3.1.2.1/16-21 16.3.1.2.2/16-21 0x0000_0000 0x0000_0000 0x0000_0000 — 16.3.1.1.1/16-18 16.3.1.1.1/16-18 16.3.1.1.3/16-20 — MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 16-15 PCI/PCI-X Bus Interface Table 16-3. PCI/X Memory-Mapped Register Map (continued) Offset Register Access Reset Section/Page 0x0_8C20–0x0_8C3C—Outbound Window 1 0x0_8C20 0x0_8C24 0x0_8C28 0x0_8C2C 0x0_8C30 0x0_8C34– 0x0_8C3C POTAR1—PCI/X outbound window 1 translation address register POTEAR1—PCI/X outbound window 1 translation extended address register POWBAR1—PCI/X outbound window 1 base address register Reserved POWAR1—PCI/X outbound window 1 attributes register Reserved R/W R/W R/W — R/W — 0x0_8C40–0x0_8C5C—Outbound Window 2 0x0_8C40 0x0_8C44 0x0_8C48 0x0_8C4C 0x0_8C50 0x0_8C54– 0x0_8C5C POTAR2—PCI/X outbound window 2 translation address register POTEAR2—PCI/X outbound window 2 translation extended address register POWBAR2—PCI/X outbound window 2 base address register Reserved POWAR2—PCI/X outbound window 2 attributes register Reserved R/W R/W R/W — R/W — 0x0_8C60–0x0_8C7C—Outbound Window 3 0x0_8C60 0x0_8C64 0x0_8C68 0x0_8C6C 0x0_8C70 0x0_8C74– 0x0_8C7C POTAR3—PCI/X outbound window 3 translation address register POTEAR3—PCI/X outbound window 3 translation extended address register POWBAR3—PCI/X outbound window 3 base address register Reserved POWAR3—PCI/X outbound window 3 attributes register Reserved R/W R/W R/W — R/W — 0x0_8C80–0x0_8C9C—Outbound Window 4 0x0_8C80 0x0_8C84 POTAR4—PCI/X outbound window 4 translation address register POTEAR4—PCI/X outbound window 4 translation extended address register R/W R/W 0x0000_0000 0x0000_0000 16.3.1.2.1/16-21 16.3.1.2.2/16-21 0x0000_0000 0x0000_0000 0x0000_0000 — 0x0000_0000 — 16.3.1.2.4/16-22 16.3.1.2.1/16-21 16.3.1.2.2/16-21 16.3.1.2.3/16-22 0x0000_0000 0x0000_0000 0x0000_0000 — 0x0000_0000 — 16.3.1.2.4/16-22 16.3.1.2.1/16-21 16.3.1.2.2/16-21 16.3.1.2.3/16-22 0x0000_0000 0x0000_0000 0x0000_0000 — 0x0000_0000 — 16.3.1.2.4/16-22 16.3.1.2.1/16-21 16.3.1.2.2/16-21 16.3.1.2.3/16-22 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 16-16 Freescale Semiconductor Memory Map/Register Definitions Table 16-3. PCI/X Memory-Mapped Register Map (continued) Offset 0x0_8C88 0x0_8C8C 0x0_8C90 0x0_8C94– 0x0_8D9C Register POWBAR4—PCI/X outbound window 4 base address register Reserved POWAR4—PCI/X outbound window 4 attributes register Reserved Access R/W — R/W — 0x0_8DA0–0x0_8DBC–Inbound Window 3 0x0_8DA0 0x0_8DA4 0x0_8DA8 0x0_8DAC 0x0_8DB0 PITAR3—PCI/X inbound window 3 translation address register Reserved PIWBAR3—PCI/X inbound window 3 base address register PIWBEAR3—PCI/X inbound window 3 base extended address register PIWAR3—PCI/X inbound window 3 attributes register R/W — R/W R/W R/W — 0x0_8DC0–0x0_8DDC–Inbound Window 2 0x0_8DC0 0x0_8DC4 0x0_8DC8 0x0_8DCC 0x0_8DD0 PITAR2—PCI/X inbound window 2 translation address register Reserved PIWBAR2—PCI/X inbound window 2 base address register PIWBEAR2—PCI/X inbound window 2 base extended address register PIWAR2—PCI/X inbound window 2 attributes register R/W — R/W R/W R/W — 0x0_8DE0–0x0_8DFC–Inbound Window 1 0x0_8DE0 0x0_8DE4 0x0_8DE8 0x0_8DEC 0x0_8DF0 PITAR1—PCI/X inbound window 1 translation address register Reserved PIWBAR1—PCI/X inbound window 1 base address register Reserved PIWAR1—PCI/X inbound window 1 attributes register R/W — R/W — R/W — 0x0000_0000 — 0x0000_0000 — 0x0000_0000 — 16.3.1.3.4/16-26 16.3.1.3.2/16-25 16.3.1.3.1/16-25 0x0000_0000 — 0x0000_0000 0x0000_0000 0x0000_0000 — 16.3.1.3.2/16-25 16.3.1.3.3/16-26 16.3.1.3.4/16-26 16.3.1.3.1/16-25 0x0000_0000 — 0x0000_0000 0x0000_0000 0x0000_0000 — 16.3.1.3.2/16-25 16.3.1.3.3/16-26 16.3.1.3.4/16-26 16.3.1.3.1/16-25 Reset 0x0000_0000 — 0x0000_0000 — 16.3.1.2.4/16-22 Section/Page 16.3.1.2.3/16-22 0x0_8DB4– Reserved 0x0_8DBC 0x0_8DD4– Reserved 0x0_8DDC 0x0_8DF4– Reserved 0x0_8DFC PCI/X Error Management Registers MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 16-17 PCI/PCI-X Bus Interface Table 16-3. PCI/X Memory-Mapped Register Map (continued) Offset 0x0_8E00 0x0_8E04 0x0_8E08 0x0_8E0C 0x0_8E10 0x0_8E14 0x0_8E18 0x0_8E1C 0x0_8E20 0x0_8E24 0x0_8E28– 0x0_8EFC 0x0_8F00– 0x0_8FFC Register ERR_DR—PCI/X error detect register ERR_CAP_DR—PCI/X error capture disabled register ERR_EN—PCI/X error enable register ERR_ATTRIB—PCI/X error attributes capture register ERR_ADDR—PCI/X error address capture register ERR_EXT_ADDR—PCI/X error extended address capture register ERR_DL—PCI/X error data low capture register ERR_DH—PCI/X error data high capture register GAS_TIMR—PCI/X gasket timer register PCIX_TIMR—PCI-X split completion timer register Reserved Reserved for debug Access Special R/W R/W R/W R/W R/W R/W R/W R/W R/W — — Reset 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 — — Section/Page 16.3.1.4.1/16-29 16.3.1.4.2/16-30 16.3.1.4.3/16-31 16.3.1.4.4/16-32 16.3.1.4.5/16-33 16.3.1.4.6/16-34 16.3.1.4.7/16-34 16.3.1.4.8/16-34 16.3.1.4.9/16-35 16.3.1.4.10/16-3 6 16.3.1.1 PCI/X Configuration Access Registers The PCI/X configuration header, shown in Figure 16-25 and Figure 16-26, is accessed through an indirect method using a pair of 32-bit memory-mapped access registers, CFG_ADDR at offset 0x0_8000 and CFG_DATA at offset 0x0_8004. 16.3.1.1.1 PCI/X Configuration Address Register (CFG_ADDR) The CFG_ADDR register is shown in Figure 16-3. 0 7 8 15 16 20 21 23 24 29 30 31 R EN W Reset Offset 0 0 0 0000 Bus Number Device Number 00 Function Register Number Number 0000_0000_0000_0000_0000_0000_0000_0000 0x0_8000 Figure 16-3. PCI/X CFG_ADDR Register MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 16-18 Freescale Semiconductor Memory Map/Register Definitions Table 16-4 describes the bit settings for the CFG_ADDR register. Table 16-4. PCI/X CFG_ADDR Field Descriptions Bits 0 1–7 8–15 Enable — Bus Number Name Description Allow a PCI configuration access when PCI CFG_DATA is accessed Reserved PCI bus number to access Device number to access on specified bus Function to access within specified device 32-bit register to access within specified device Reserved, hardwired to logic 00 16–20 Device Number 21–23 Function Number 24–29 Register Number 30–31 — Bus number 0xb00 and device number 0b0_0000 are used to configure the internal PCI/X configuration header of the MPC8540 itself. See Section 16.4.2.11.2, “Accessing the PCI Configuration Space in Host Mode,” and Section 16.4.2.11.3, “PCI Configuration in Agent and Agent Lock Modes,” for usage of PCI/X CFG_ADDR. 16.3.1.1.2 PCI/X Configuration Data Register (CFG_DATA) The CFG_DATA register is shown in Figure 16-3. 0 31 R Data W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x0_8004 Figure 16-4. PCI/X CFG_DATA Register Table 16-5 describes the bit settings for the CFG_DATA register Table 16-5. PCI/X CFG_DATA Field Descriptions Bits 0–31 Name Data Description A read or write to this register starts a PCI configuration cycle if the PCI CFG_ADDR enable bit is set. If the enable bit is not set, a PCI I/O transaction is generated. The CFG_DATA register is a 4-byte window into the little-endian PCI/X configuration header data structure; therefore byte addressing within the CFG_DATA register uses little-endian convention. Note that CFG_DATA may contain 1, 2, 3, or 4 bytes depending on the size of the register being accessed. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 16-19 PCI/PCI-X Bus Interface See Section 16.4.2.11.2, “Accessing the PCI Configuration Space in Host Mode,” and Section 16.4.2.11.3, “PCI Configuration in Agent and Agent Lock Modes,” for use of CFG_DATA. 16.3.1.1.3 PCI/X Interrupt Acknowledge Register (INT_ACK) An external PCI/X interrupt acknowledge transaction is generated by reading the INT_ACK register at offset 0x0_8008. PCI INT_ACK is read-only and a write to that address results in nothing. The INT_ACK register is shown in Figure 16-5. 0 31 R W Reset Offset Data 0000_0000_0000_0000_0000_0000_0000_0000 0x0_8008 Figure 16-5. PCI/X INT_ACK Register Table 16-6 describes the bit settings for the INT_ACK register. Table 16-6. PCI/X INT_ACK Field Descriptions Bits 0–31 Name Data Description A read to this register generates a PCI interrupt acknowledge cycle. 16.3.1.2 PCI/X ATMU Outbound Registers The outbound address translation and mapping unit controls the mapping of transactions from the internal 32-bit address space of the MPC8540 to the external PCI address space. The outbound ATMU consists of four translation windows plus a default translation for transactions that do not hit in one of the four windows. Each window contains a base address that points to the beginning of the window in the local address map, a translation address that specifies the high-order bits of the transaction in the external PCI address space, and a set of attributes including window size and external transaction type. Each window must be aligned based on the granularity specified by the window size. If two outbound ATMU windows overlap in the local address space, the mapping of the lower numbered window has precedence over the higher numbered window. Note that outbound translation windows must not overlap the configuration access registers. Window 0 is the default window and is the only window enabled upon reset. The default outbound register set is used when a transaction misses in all other outbound windows. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 16-20 Freescale Semiconductor Memory Map/Register Definitions 16.3.1.2.1 PCI/X Outbound Translation Address Registers (POTARn) The PCI/X outbound translation address registers (POTARn) select the starting addresses in the PCI address space for hits in the PCI/X outbound windows. The translated address is created by concatenating the transaction offset to this translation address. The format of the POTARn is shown in Figure 16-6. 0 11 12 31 R TEA W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x0_8C00, 0x0_8C20, 0x0_8C40, 0x0_8C60, 0x0_8C80 TA Figure 16-6. PCI/X Outbound Translation Address Registers Table 16-7 describes the fields of the POTARn registers. Table 16-7. POTARn Field Descriptions Bits 0–11 12–3 1 Nam e TEA TA Description Translation extended address. Represents bits 43–32 of a 64-bit PCI address (bit 0 is lsb). Translation address. Represents bits 31–12 of the PCI address. Based on the size of the window specified in POWAR n[OWS], the low-order bits of this field may be ignored. 16.3.1.2.2 PCI/X Outbound Translation Extended Address Registers (POTEARn) The PCI/X outbound translation extended address registers (POTEARn), shown in Figure 16-7, contain the most significant bits of a 64-bit translation address. 0 11 12 31 R0 0 0 0 0 0 0 0 0 0 0 0 TEA W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x0_8C04, 0x0_8C24, 0x0_8C44, 0x0_8C64, 0x0_8C84 Figure 16-7. PCI/X Outbound Translation Extended Address Registers MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 16-21 PCI/PCI-X Bus Interface Table 16-8 describes the fields of the POTEARn. Table 16-8. POTEARn Field Descriptions Bits 0–11 12–31 Name — TEA Reserved Translation extended address. Represents bits [63–44] of a 64-bit PCI address (used in conjunction with POTARn) Description 16.3.1.2.3 PCI/X Outbound Window Base Address Registers (POWBARn) The PCI/X outbound window base address registers (POWBARn), shown in Figure 16-8, point to the beginning of each translation window in the local 32-bit address space. Addresses for outbound transactions are compared to the appropriate bits in these registers, according to the sizes of the windows. If a transaction does not fall within one of these windows, the default translation and mapping is used. The default window is always enabled and used when the other windows miss. Note that POWBAR0 (for outbound ATMU window 0) is not used, because window 0 is the default window used when no other windows match. POWBAR0 may be read from and written to, but the value is ignored. 0 11 12 31 R — W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x0_8C08, 0x0_8C28, 0x0_8C48, 0x0_8C68, 0x0_8C88 WBA Figure 16-8. PCI/X Outbound Window Base Address Registers Table 16-9 describes the field of the POWBARn. Table 16-9. POWBARn Field Descriptions Bits 0–11 12–3 1 Name — Reserved, should be cleared. Description WBA Window base address. Source address that is the starting point for the outbound translation window. The window must be aligned based on the size selected in the window size bits. 16.3.1.2.4 PCI/X Outbound Window Attributes Registers (POWARn) The PCI/X outbound window attributes registers (POWARn) define the window sizes to translate and other attributes for the translations. The minimum window size is 4 Kbytes. The maximum window size is 4 Gbytes. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 16-22 Freescale Semiconductor Memory Map/Register Definitions The default window attribute register, POWAR0, is shown in Figure 16-9. Note that the fields for all of the POWARn registers are the same, only the reset values are different. 0 1 11 12 15 16 19 20 25 26 31 R W Reset Offset 1 EN1 00000000000 RTT WTT 000000 OWS 1000_0000_0000_0100_0100_0000_0001_1111 2 0x0_8C10 For POWAR0, translation is always enabled. The enable field (EN) may be read and written, but the value is ignored. 2 The default window is enabled, configured for memory read and memory write, and set to an OWS size of 4 Gbytes. Figure 16-9. PCI/X Outbound Window Attributes Register 0 (Default) POWAR1–POWAR4 are shown in Figure 16-10. 0 1 11 12 15 16 19 20 25 26 31 R EN W Reset Offset 00000000000 RTT WTT 000000 OWS 0000_0000_0000_0000_0000_0000_0000_0000 0x0_8C30, 0x0_8C50, 0x0_8C70, 0x0_8C90 Figure 16-10. PCI/X Outbound Window Attributes Registers 1–4 Table 16-10 describes the fields for the POWARn registers. Table 16-10. POWARn Field Descriptions Bits 0 1–11 12–15 Name EN — RTT Description Enable. Enables this address translation. Note that for POWAR0, translation is always enabled. The enable field (EN) may be read and written, but the value is ignored. Reserved Read transaction type to run on PCI 0000–0011Reserved 0100 Memory Read 0101–0111Reserved 1000 I/O Read 1001–1111Reserved Writetransaction type to run on PCI 0000–0011Reserved 0100 Memory Write 0101–0111Reserved 1000 I/O Write 1001–1111Reserved 16–19 WTT MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 16-23 PCI/PCI-X Bus Interface Table 16-10. POWARn Field Descriptions (continued) Bits 20–25 26–31 Name — OWS Reserved Outbound window size. Outbound translation window size N which is the encoded 2(N+1) bytes window size. The smallest window size is 4 Kbytes. 000000–0010114 Kbyte window size 0011008-Kbyte window size ... 0111114-Gbyte window size 100000–111111Reserved The default POWAR register (0x0_8C10) has an OWS value of 011111. Also note that for POWAR0, setting OWS to less than 4 Gbytes causes addresses that miss in the other outbound windows to be aliased to the smaller address range defined by POWAR0[OWS] and POTAR0. Description 16.3.1.3 PCI/X ATMU Inbound Registers The inbound address translation and mapping unit controls the mapping of transactions from the external PCI address space to the local address space of the MPC8540. The inbound ATMU is comprised of four windows—a configuration window and three general translation windows. The configuration window has higher priority than all other inbound ATMU windows and takes precedence over them if there is an overlap. Each window contains the following: • A base address, which points to the beginning of the window in the external PCI address map. The base address of each window is also accessible by PCI configuration transactions as base address registers within the PCI configuration header, as shown in Figure 16-25. The registers may be read or updated equivalently through the ATMU memory map or through PCI configuration transactions to the PCI configuration header. A translation address, which specifies the upper order bits of the transaction in the local address space. A set of attributes including window size and internal transaction attributes. • • Each window’s base address and translation address must be aligned to the size of the window. If two general inbound ATMU windows overlap in the external PCI address space, the mappings of the lower numbered window are applied; however, it is illegal for an inbound window to overlap the PCSRBAR window. In addition, if inbound ATMU windows are overlapped, the ATMU windows must not map to the same address with different sets of attributes. Note that PCSRBAR in the PCI configuration header acts as a fourth inbound window that translates a 1-Mbyte region of PCI space to the local configuration space pointed to by CCSRBAR. PCSRBAR can be accessed by PCI configuration cycles or by accessing the PCI configuration header through the PCI CFG_ADDR and PCI CFG_DATA registers. See Section 16.3.1.1.1, “PCI/X Configuration Address Register (CFG_ADDR),” Section 16.3.1.1.2, “PCI/X Configuration Data Register (CFG_DATA),” and Section 16.3.2.11, “PCI Base Address MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 16-24 Freescale Semiconductor Memory Map/Register Definitions Registers.” All accesses to PCSRBAR have an automatic internal byte lane redirection from the little-endian PCI bus to the big-endian CCSRBAR configuration space. 16.3.1.3.1 PCI/X Inbound Translation Address Registers (PITARn) The PCI/X inbound translation address registers (PITARn) points to the beginning of the local address space for the inbound window. The translated address is created by concatenating the transaction offset to this translation address. The format of the PITARn is shown in Figure 16-11. 0 11 12 31 R — W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x0_8DA0, 0x0_8DC0, 0x0_8DE0 TA Figure 16-11. PCI/X Inbound Translation Address Registers Table 16-11 describes the fields of the PITARn registers. Table 16-11. PITARn Field Descriptions Bits 0–11 12–3 1 Nam e — TA Reserved, should be cleared. Translation address. Indicates the starting point of the inbound translated address. The translation address must be aligned based on the size field. TA corresponds to the high-order 20 bits of a 32-bit local address. Description 16.3.1.3.2 PCI/X Inbound Window Base Address Registers (PIWBARn) PCI/X inbound window base address registers (PIWBARn), shown in Figure 16-12, select the PCI/X base address for windows translated to the MPC8540 local address space. Inbound transaction addresses are compared to these windows. If a PCI/X transaction does not fall in one of these spaces, the PCI/X interface does not assert DEVSEL. 0 11 12 31 R BEA W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x0_8DA8, 0x0_8DC8, 0x0_8DE8 BA Figure 16-12. PCI/X Inbound Window Base Address Registers MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 16-25 PCI/PCI-X Bus Interface Table 16-12 describes the fields of the PIWBARn registers. Table 16-12. PIWBARn Field Descriptions Bits 0–11 12–31 Name BEA BA Description Base extended address. Corresponds to bits 43–32 of a 64-bit PCI base address. Base address. Corresponds to bits 31–12 of a PCI base address. 16.3.1.3.3 PCI/X Inbound Window Base Extended Address Registers (PIWBEARn) The PCI/X inbound window base extended address registers (PIWBEARn), shown in Figure 16-13, contain the msbs of a 64-bit base address. Note that inbound window 1 supports only a 32-bit base address and does not define an inbound window base extended address register. 0 11 12 31 R0 0 0 0 0 0 0 0 0 0 0 0 BEA W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x0_8DAC, 0x0_8DCC Figure 16-13. PCI/X Inbound Window Base Extended Address Registers Table 16-13 describes the fields of the PIWBEARn registers. Table 16-13. PIWBEARn Field Descriptions Bits 0–11 12–31 Name — BEA Reserved Base extended address. Corresponds to bits 63–44 of a 64-bit PCI base address. Description 16.3.1.3.4 PCI/X Inbound Window Attributes Registers (PIWARn) The PCI/X inbound window attributes registers (PIWARn) define the window sizes to translate and other attributes for the translations. 16 Gbytes is the largest window size allowed. The format of the PIWBARn is shown in Figure 16-14. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 16-26 Freescale Semiconductor Memory Map/Register Definitions 0 1 2 3 7 8 11 12 15 16 19 20 25 26 31 R EN W Reset Offset 0 PF 00000 TGI RTT WTT 000000 IWS 0000_0000_0000_0000_0000_0000_0000_0000 0x0_8DB0, 0x0_8DD0, 0x0_8DF0 Figure 16-14. PCI/X Inbound Window Attributes Registers Table 16-14 describes the fields of the PIWARn registers. Table 16-14. PIWARn Field Descriptions Bits 0 1 2 Name EN — PF Enable. Enables this address translation Reserved Prefetchable. Indicates that the address space is prefetchable so that prefetching and streaming are attempted. 0 Not prefetchable 1 Prefetchable Reserved Target Interface. 0000–1011 Reserved 1100 RapidIO 1101–1110 Reserved 1111 Local Memory (DDR SDRAM, Local Bus, SRAM) Note: If this field is set to an I/O port rather than local memory space, attributes for the external I/O transaction are assigned in an outbound ATMU of that I/O controller. Read transaction type. Transaction type to run if access is a read. The field description differs subject to the transaction being targeted to I/O interface or to local memory. Following are the transaction type settings for reads to the RapidIO interface: 0000–0011 Reserved 0100 Read 0101–1111 Reserved Following are the transaction type settings for reads to local memory: 0000–0011 Reserved 0100 Read, don’t snoop local processor 0101 Read, snoop local processor 0110 Reserved 0111 Read, unlock L2 cache line 1000–1111 Reserved Description 3–7 8–11 — TGI 12–15 RTT MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 16-27 PCI/PCI-X Bus Interface Table 16-14. PIWARn Field Descriptions (continued) Bits 16–19 Name WTT Description Write transaction type. Transaction type to run if access is a write. The field description differs subject to the transaction being targeted to an I/O interface or to local memory. Following are the transaction type settings for writes to the RapidIO interface: 0000–0011 Reserved 0100 Write 0101–1111 Reserved Following are the transaction type settings for writes to local memory: 0000–0011 Reserved 0100 Write, don’t snoop local processor 0101 Write, snoop local processor 0110 Write, allocate L2 cache line 0111 Write, allocate and lock L2 cache line 1000–1111 Reserved Reserved Inbound window size. Inbound translation window size N which is the encoded 2^(N+1) bytes window size. The smallest window is 4 Kbytes. 000000–001010 Reserved 001011 4-Kbyte window size 001100 8-Kbyte window size ... 011111 4-Gbyte window size 100000–111111 Reserved For configuration and run-time registers, the window size is fixed at 010011 1-Mbyte window size For register set 0, the window size is limited to 4 Gbytes or smaller. 20–25 26–31 — IWS 16.3.1.4 PCI/X Error Management Registers When a PCI/X error is detected, the appropriate error bit is set in the PCI/X error detect register. Subsequent errors set the appropriate error bits in the error detection registers, but relevant information (attributes, address, and data) is captured only for the first error. The PCI/X error detect register is a write-1-to-clear type register. That is, reading from this register occurs normally; however, write operations are different in that the bits can be cleared but not set. A bit is cleared whenever the register is written, and the data in the corresponding bit location is a 1. For example, to clear bit 25 and not affect any other bits in the register, the value 0x0000_0040 is written to the register. The error bit is set regardless of the state of the corresponding error enable bit in the PCI/X error enable register. The error enable bits are used to send or block the error reporting to the interrupt mechanism. The interrupt can be cleared by writing 0xFFFF_FFFF to the PCI/X error detect register. Note that some errors are reported in two bits—one in the PCI/X error detect register (ERR_DR) and another in the PCI bus status register in the PCI/X configuration header. These bits must be cleared separately; that is, clearing one does not clear the other. For example, clearing the MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 16-28 Freescale Semiconductor Memory Map/Register Definitions ERR_DR[Mstr abort error] does not clear the received master abort bit in the PCI bus status register. In these cases, both bits must be cleared before further error reporting can occur. Refer to Table 16-57 for PCI mode error actions and Table 16-65 for PCI-X mode error actions. Likewise, some errors are enabled by programming two bits—one in the PCI/X error enable register and another in the PCI bus command register in the PCI/X configuration header. A master-abort condition during a configuration cycle is not necessarily an error. In this case, if relevant, the master-abort error enable can be disabled to prevent the reporting of master-aborts during outbound configuration cycles. Master-aborts during configuration reads return 0xFFFF_FFFF. If a data parity error occurs during an inbound configuration write access, the error is reported and captured. However, the erroneous data is written to the register specified in the transaction. Therefore, PCI data parity error recovery routines must include reinitialization of the PCI/X configuration register if the error occurred during a configuration write. A dual address cycle (DAC) transaction on the PCI bus has two address phases. The command of the first phase is DAC and the command of the second phase is the actual transaction type. If an error occurs on a DAC transaction, the DAC command is stored in command field in the PCI/X error attribute capture register. In PCI-X mode, the MPC8540 does not check the parity on the data of a split response to an outbound read transaction because no data is actually transferred. However, outbound writes that end with a split response do transfer data and thus parity is checked for these transactions. Also note that the MPC8540 does drive the correct parity for inbound read transactions that it terminates with a split response. See Section 16.4.2.13, “PCI Error Functions,” and Section 16.4.3.8, “PCI-X Error Functions,” for more details on error handling. 16.3.1.4.1 PCI/X Error Detect Register (ERR_DR) 0 1 20 21 22 23 24 25 26 27 28 29 30 31 R Multiple PCI W errors Reset Offset Trgt Mstr Trgt Addr Rcvd Mstr OWMSV ORMSV IRMSV SCM Parity SERR PERR PERR abort abort error error error error error error error error error error 0000_0000_0000_0000_0000_0000_0000_0000 0x0_8E00 TOE error Figure 16-15. PCI/X Error Detect Register (ERR_DR) Table 16-15 describes ERR_DR fields. Note that uncorrectable read errors may cause the assertion of core_fault_in, which causes the core to generate a machine check interrupt, unless it is disabled (by clearing HID1[RFXE]). If RFXE is zero and an error occurs, the appropriate parity detect and master-abort bits in ERR_DR must be cleared and the appropriate enable bits in ERR_EN must be MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 16-29 PCI/PCI-X Bus Interface set to ensure that an interrupt is generated. See Section 6.10.2, “Hardware Implementation-Dependent Register 1 (HID1).” Table 16-15. ERR_DR Field Descriptions Bits 0 1–20 21 22 23 24 25 26 27 28 29 30 31 Name Multiple PCI errors — Addr Parity error Rcvd SERR error Mstr PERR error Trgt PERR error Mstr abort error Trgt abort error OWMSV error ORMSV error IRMSV error SCM error TOE error Description 0 Multiple PCI errors of the same type were not detected (write-1-to-clear) 1 Multiple PCI errors of the same type were detected Reserved Address parity error (write-1-to-clear) Received SERR error (write-1-to-clear) Master PERR error (write-1-to-clear) Target PERR error (write-1-to-clear) Master-abort error (write-1-to-clear) Target abort error (write-1-to-clear) Outbound write memory space violation error (write-1-to-clear) Outbound read memory space violation error (write-1-to-clear) PCI/X inbound read memory space violation error (write-1-to-clear) PCI/X split completion message error (write-1-to-clear) PCI/X time out error (write-1-to-clear) 16.3.1.4.2 PCI/X Error Capture Disable Register (ERR_CAP_DR) 0 20 21 31 R W Trgt Rcvd Mstr Trgt Mstr Addr TOE OWMSV ORMSV IRMSV SCM abort parity SERR PERR PERR abort error error error error error error error error error error error capture capture capture capture capture capture capture capture capture capture capture disable disable disable disable disable disable disable disable disable disable disable 0000_0000_0000_0000_0000_0000_0000_0000 0x0_8E04 Reset Offset Figure 16-16. PCI/X Error Capture Disable Register (ERR_CAP_DR) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 16-30 Freescale Semiconductor Memory Map/Register Definitions Table 16-16. ERR_CAP_DR Field Descriptions Bits 0–20 21 22 23 24 25 26 27 28 29 30 31 Name — Reserved Description Addr parity error capture disable Disable capture for address parity errors Rcvd SERR error capture disable Disable capture for received SERR errors Mstr PERR error capture disable Disable capture for master PERR errors Trgt PERR error capture disable Disable capture for target PERR errors Mstr abort error capture disable Trgt abort error capture disable OWMSV error capture disable ORMSV error capture disable IRMSV error capture disable SCM error capture disable TOE error capture disable Disable capture for master-abort errors Disable capture for target abort errors Disable capture for outbound write memory space violation errors Disable capture for outbound read memory space violation errors Disable capture for PCI/X inbound read memory space violation errors Disable capture for PCI/X split completion message errors Disable capture for PCI/X time out errors 16.3.1.4.3 PCI/X Error Enable Register (ERR_EN) 0 20 21 22 23 24 25 26 27 28 29 30 31 R W Trgt Rcvd Mstr Trgt Mstr Addr OWMSV ORMSV IRMSV SCM TOE abort parity SERR PERR PERR abort error error error error error error error error error error error enable enable enable enable enable enable enable enable enable enable enable 0000_0000_0000_0000_0000_0000_0000_0000 0x0_8E08 Reset Offset Figure 16-17. PCI/X Error Enable Register (ERR_EN) Table 16-17 describes ERR_DR fields. Note that uncorrectable read errors may cause the assertion of core_fault_in, which causes the core to generate a machine check interrupt unless it is disabled by clearing HID1[RFXE]. If RFXE is zero and this error occurs, the appropriate parity detect and master-abort bits in ERR_DR must be cleared and the appropriate ERR_EN bits must be set to ensure that an interrupt is generated. See Section 6.10.2, “Hardware Implementation-Dependent Register 1 (HID1).” MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 16-31 PCI/PCI-X Bus Interface Table 16-17. ERR_EN Field Descriptions Bits 0–20 21 22 23 24 25 26 27 28 29 30 31 Name — Addr parity error enable Rcvd SERR error enable Mstr PERR error enable Trgt PERR error enable Mstr abort error enable Trgt abort error enable OWMSV error enable ORMSV error enable IRMSV error enable SCM error enable TOE error enable Reserved Enable reporting address parity errors Enable reporting received SERR errors Enable reporting master PERR errors Enable reporting target PERR errors Enable reporting master-abort errors Enable reporting target abort errors Enable reporting outbound write memory space violation errors Enable reporting outbound read memory space violation errors Enable reporting PCI/X inbound read memory space violation errors Enable reporting PCI/X split completion message errors Enable reporting PCI/X time out errors Description 16.3.1.4.4 PCI/X Error Attributes Capture Register (ERR_ATTRIB) 0 3 4 7 8 9 10 11 15 16 19 20 30 31 High word Low word High byte enables byte enables parity bit W Reset Offset R Low parity bit Error source Command Valid 0000_0000_0000_0000_0000_0000_0000_0000 0x0_8E0C Figure 16-18. PCI/X Error Attributes Capture Register (ERR_ATTRIB) Table 16-18. ERR_ATTRIB Field Descriptions Bits 0–3 4–7 8 9 10 Name Description High word byte enables PCI byte enables for most significant word of the double word Low word byte enables PCI byte enables for least significant word of the double word High parity bit Low parity bit — Parity bit for most significant PCI bus data word (only valid for 64-bit PCI bus) Parity bit for least significant PCI bus data word Reserved MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 16-32 Freescale Semiconductor Memory Map/Register Definitions Table 16-18. ERR_ATTRIB Field Descriptions (continued) Bits Name The source of the PCI transaction 00000 PCI/X interface 00001 Reserved 00010–00011 Reserved 00100 Local bus controller 00101–01000 Reserved 01001 Reserved 01010–01011 Reserved 01100 RapidIO interface 01101 Reserved 01110 Reserved 01111 Reserved 10000 e500 core (instruction) 10001 e500 core (data) 10010 Reserved PCI command Reserved The PCI bus capture registers contain valid information Description 11–15 Error source 10011 Reserved 10100 Reserved 10101 DMA 10110 RDC 10111 SAP 11000 TSEC1 11001 TSEC2 11010 FEC (10/100) 11011 Reserved 11100 RapidIO message units 11101 RapidIO doorbell units 11110 RapidIO port-write units 11111 Reserved 16–19 Command 20–30 — 31 Valid info 16.3.1.4.5 PCI/X Error Address Capture Register (ERR_ADDR) 0 31 R Memory Address W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x0_8E10 Figure 16-19. PCI/X Error Address Capture Register (ERR_ADDR) Table 16-19. ERR_ADDR Field Descriptions Bits 0–31 Name Memory address Memory transaction address Description MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 16-33 PCI/PCI-X Bus Interface 16.3.1.4.6 PCI/X Error Extended Address Capture Register (ERR_EXT_ADDR) 0 31 R Memory Extended Address W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x0_8E14 Figure 16-20. PCI/X Error Extended Address Capture Register (ERR_EXT_ADDR) Table 16-20. ERR_EXT_ADDR Field Descriptions Bits 0–31 Name Description Memory extended address Memory transaction extended address 16.3.1.4.7 PCI/X Error Data Low Capture Register (ERR_DL) Note that for inbound reads that have data parity errors, only the address (ERR_ADDR and ERR_EXT_ADDR) and attributes (ERR_ATTRIB) are captured. The data is not captured. Also note that PCI-X split completion error messages for outbound DWORD writes are incorrectly reported in the PCI/X error data low capture register (ERR_DL). 0 31 R Data low W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x0_8E18 Figure 16-21. PCI/X Error Data Low Capture Register (ERR_DL) Table 16-21. ERR_DL Field Description Bits 0–31 Name Data low Description Least significant PCI bus data word 16.3.1.4.8 PCI/X Error Data High Capture Register (ERR_DH) Note that for inbound reads that have data parity errors, only the address (ERR_ADDR and ERR_EXT_ADDR) and attributes (ERR_ATTRIB) are captured. The data is not captured. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 16-34 Freescale Semiconductor Memory Map/Register Definitions 0 31 R Data high W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x0_8E1C Figure 16-22. PCI/X Error Data High Capture Register (ERR_DH) Table 16-22. ERR_DH Field Description Bits 0–31 Name Data high Description Most significant PCI bus data word (only valid with 64-bit PCI bus) 16.3.1.4.9 PCI/X Gasket Timer Register (GAS_TIMR) 0 7 8 31 R0 0 0 0 0 0 0 EN W Reset Offset 0000_0001_0000_0000_0011_1111_1111_1111 (PCI mode) 0000_0000_0000_0000_0011_1111_1111_1111 (PCI-X mode) 0x0_8E20 TCNT Figure 16-23. PCI/X Gasket Timer Register (GAS_TIMR) Table 16-23. GAS_TIMR Field Descriptions Bits 0–6 7 Name — EN Reserved Gasket timer enable. 0 PCI-X default: Gasket timer is disabled. 1 PCI default: Gasket timer is enabled. Number of system clocks to purge a non-prefetchable inbound read buffer Description 8–31 TCNT MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 16-35 PCI/PCI-X Bus Interface 16.3.1.4.10PCI-X Split Completion Timer Register (PCIX_TIMR) 0 7 8 31 R0 W Reset Offset 0 0 0 000 EN TCNT 0000_0001_0000_0000_0011_1111_1111_1111 0x0_8E24 Figure 16-24. PCI-X Split Completion Timer Register (PCIX_TIMR) Table 16-24. PCIX_TIMR Field Descriptions Bits 0–6 7 Nam e — EN Reserved PCI-X timer enable 0 PCI-X timer is disabled. 1 PCI-X timer is enabled (default). Description 8–3 1 TCN Number of system clocks to purge a split completion buffer. This occurs when a split completion does T not follow a split response. 16.3.2 PCI/X Configuration Header The PCI Local Bus Specification defines the configuration registers contained within the PCI/X configuration header from 0x00 through 0x3F. The PCI-X Addendum to the PCI Local Bus Specification defines additional registers beyond 0x3F. Figure 16-25 lists the common PCI/X configuration header as implemented by the MPC8540. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 16-36 Freescale Semiconductor Memory Map/Register Definitions Reserved Device ID PCI Bus Status Bus Base Class Code BIST Control Subclass Code Header Type Vendor ID PCI Bus Command Bus Programming Interface Address Offset (Hex) 00 04 08 0C 10 14 18 1C 20 24 28 2C 30 34 38 3C 40 44 Bus Latency Timer Revision ID Bus Cache Line Size PCI Configuration and Status Register Base Address Register (PCSRBAR) 32-Bit Memory Base Address Register 64-Bit Low Memory Base Address Register 64-Bit High Memory Base Address Register 64-Bit Low Memory Base Address Register 64-Bit High Memory Base Address Register Subsystem ID Subsystem Vendor ID PCI Bus Capability Pointer PCI Bus MAX_LAT PCI Bus MIN_GNT PCI Bus Interrupt Pin PCI Bus Interrupt Line PCI Bus Arbiter Configuration PCI Bus Function Figure 16-25. MPC8540 PCI/X Configuration Header Figure 16-26 lists the additional PCI-X configuration registers. Reserved PCI-X Command PCI-X Next Capability PCI-X Status Address Offset (Hex) PCI-X Capability ID 60 64 Figure 16-26. PCI-X Additional Configuration Registers Note that software must be restricted from accessing any reserved or undefined register space on the MPC8540; doing so may hang the device. Table 16-54 in Section 16.4.2.11.1, “PCI Configuration Space Header,” provides a summary of the PCI configuration header registers. Detailed descriptions of these registers are provided in the PCI Local Bus Specification. 16.3.2.1 PCI Vendor ID Register—Offset 0x00 The PCI vendor ID register, shown in Figure 16-27, is used to identify the manufacturer of the part. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 16-37 PCI/PCI-X Bus Interface 15 0 R W Reset Offset Vendor ID 0001_0000_0101_0111 0x00 Figure 16-27. PCI Vendor ID Register Table 16-25 describes PCI vendor ID register fields. Table 16-25. PCI Vendor ID Register Field Description Bits 15–0 Name Vendor ID Description 0x1057 (Freescale Semiconductor) 16.3.2.2 PCI Device ID Register—Offset 0x02 The PCI device ID register, shown in Figure 16-28, is used to identify the device. 15 0 R W Reset Offset Device ID 0000_0000_0000_1000 0x02 Figure 16-28. PCI Device ID Register Table 16-26. PCI Device ID Register Field Description Bits 15–0 Name Device ID 0x0008 Description 16.3.2.3 PCI Bus Command Register—Offset 0x04 The 2-byte PCI bus command register provides control over the ability to generate and respond to PCI cycles. Table 16-27 describes the bits of the PCI bus command register. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 16-38 Freescale Semiconductor Memory Map/Register Definitions 15 10 9 8 7 6 5 4 3 2 1 0 R Fast back-toback SERR Parity error response Memory-writeand-invalidate Special cycles Bus master Memory space I/O space W Reset Offset 0000_0000_0000_0*00 *0 = Agent, 1 = Host 0x04 Figure 16-29. PCI Bus Command Register Table 16-27. PCI Bus Command Register Field Descriptions Bits 15–1 0 9 8 Name — Reserved Description Fast Hard-wired to 0, indicating that this PCI controller (as a master) does not run fast back-to-back back-to-back transactions SERR Controls the PCI_SERR driver of this PCI controller. This bit (and bit 6) must be set to report address parity errors. 0 Disables the PCI_SERR driver 1 Enables the PCI_SERR driver Reserved Controls whether this PCI controller responds to parity errors. 0 Parity errors are ignored and normal operation continues. 1 Parity errors cause the appropriate bit in the PCI status register to be set. However, note that errors are reported based on the values set in the PCI error enable and detection registers. Reserved 7 6 — Parity error response 5 4 — Memory-write Hard-wired to 0, indicating that this PCI controller, acting as a master, can not generate the memory-write-and-invalidate command and-invalidate Special-cycle Hard-wired to 0, indicating that this PCI controller (as a target) ignores all special-cycle s commands Bus master Indicates whether this PCI controller is configured as a master. This indicates the setting of the host/agent configuration input (LWE2) at power-on reset. 0 Disables the ability to generate PCI accesses 1 Enables this PCI controller to behave as a PCI bus master (Host) Note that the Bus master bit in the MPC8540’s PCI bus command register must be set before any outbound configuration access is attempted. Controls whether this PCI controller (as a target) responds to memory accesses. 0 This PCI controller does not respond to PCI memory space accesses. 1 This PCI controller (as a target) responds to PCI memory space accesses. Hard-wired to 0, indicating that this PCI controller (as a target) does not respond to PCI I/O space accesses 3 2 1 Memory space 0 I/O space MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 16-39 PCI/PCI-X Bus Interface 16.3.2.4 PCI Bus Status Register—Offset 0x06 The 2-byte PCI bus status register is used to record status information for PCI bus bus-related events. The definition of each bit is given in Table 16-28. Only 2-byte accesses to address offset 0x06 are allowed. Note that some errors are reported in two bits—one in the PCI bus status register and another in the PCI/X error detect register (ERR_DR). These bits must be cleared separately; that is, clearing one does not clear the other. For example, clearing the ERR_DR[Mstr abort error] does not clear the received master abort bit in the PCI bus status register. In these cases, both bits must be cleared before further error reporting can occur. Refer to Table 16-57 for PCI mode error actions and Table 16-65 for PCI-X mode error actions. Likewise, some errors are enabled by programming two bits—one in the PCI bus command register and another in the PCI/X error enable register (ERR_EN). Reads to this register behave normally. Writes are slightly different in that bits can be cleared, but not set. A bit is cleared whenever the register is written, and the data in the corresponding bit location is a 1. For example, to clear bit 14 and not affect any other bits in the register, write the value 0b0100_0000_0000_0000 to the register. 15 14 13 12 11 10 9 8 7 6 5 4 3 0 R Detected parity error Signaled system error Received master abort Received target abort Signaled target abort DEVSEL timing W Reset Offset Fast Master back-todata back parity capable error detected 66-MHz capable Capabilities list 0000_0000_1011_0000 0x06 Figure 16-30. PCI Bus Status Register . Table 16-28. PCI Bus Status Register Field Descriptions Bits 15 14 13 12 11 10– 9 Name Detected parity error Description Set whenever this PCI controller detects a PCI parity error, even if parity error handling is disabled (as controlled by bit 6 in the PCI bus command register) Signaled system Set whenever this PCI controller asserts PCI_SERR. error Received master-abort Received target-abort Signaled target-abort DEVSEL timing Set whenever this PCI controller, acting as the PCI master, terminates a transaction (except for a special-cycle) using master-abort Set whenever a PCI transaction initiated by this PCI controller (excluding a special-cycle) is terminated by a target-abort Set whenever this PCI controller, acting as the PCI target, issues a target-abort to a PCI master Hard-wired to 0b00, indicating that this PCI controller uses fast device select timing. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 16-40 Freescale Semiconductor Memory Map/Register Definitions Table 16-28. PCI Bus Status Register Field Descriptions (continued) Bits 8 Name Description Master data parity Set upon detecting a data parity error. Three conditions must be met for this bit to be set: error detected • This PCI controller detected a parity error. • This PCI controller was acting as bus master for the operation in which the error occurred. • Bit 6 in the PCI bus command register was set. Fast back-to-back Hard-wired to 1, indicating that this PCI controller (as a target) is capable of accepting fast capable back-to-back transactions — 66-MHz capable Capabilities List — Reserved Read-only bit indicates that this PCI controller is capable of 66 MHz PCI bus operation. Hard-wired to 1 Reserved 7 6 5 4 3–0 16.3.2.5 PCI Revision ID Register—Offset 0x08 The PCI Revision ID register is used to identify the revision of the part. 7 0 R W Reset Offset Revision ID revision specific 0x08 Figure 16-31. PCI Revision ID Register Table 16-29. PCI Revision ID Register Field Descriptions Bits 7–0 Name Revision ID Revision specific Description 16.3.2.6 PCI Bus Programming Interface Register—Offset 0x09 Table 16-30 describes the PCI Bus programming interface register (PIR). 7 0 R W Reset Offset Programming Interface 0000_000* *0 = Host, 1 = Agent 0x09 Figure 16-32. PCI Bus Programming Interface Register MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 16-41 PCI/PCI-X Bus Interface Table 16-30. PCI Bus Programming Interface Register Field Description Bits 7–0 Name Programming Interface Description 0x00 When the PCI controller is configured as host bridge 0x01 When the PCI controller is configured as an agent device 16.3.2.7 PCI Subclass Code Register—Offset 0x0A Table 16-32 describes the PCI subclass code register (PSCR). 7 0 R W Reset Offset Subclass Code 0010_0000 0x0A Figure 16-33. PCI Subclass Code Register Table 16-31. PCI Subclass Code Register Field Description Bits 7–0 Name Subclass Code PowerPC—0x20 Description 16.3.2.8 PCI Bus Base Class Code Register—Offset 0x0B Table 16-32 describes the PCI bus base class code register (PBCCR). 7 0 R W Reset Offset Base Class Code 0000_1011 0x0B Figure 16-34. PCI Bus Base Class Code Register Table 16-32. PCI Bus Base Class Code Register Field Description Bits 7–0 Name Base Class Code Processor—0x0B Description MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 16-42 Freescale Semiconductor Memory Map/Register Definitions 16.3.2.9 PCI Bus Cache Line Size Register—Offset 0x0C Table 16-33 describes the PCI bus cache line size register (PCLSR). 7 0 R Cache Line Size W Reset Offset 0000_0000 (PCI); 0000_1000 (PCI-X) 0x0C Figure 16-35. PCI Bus Cache Line Size Register (PCLSR) Table 16-33. PCI Bus Cache Line Size Register (PCLSR) Field Descriptions Bits 7–0 Name Cache Line Size Description Represents the cache line size of the processor in terms of 32-bit words (eight 32-bit words = 32 bytes). PCLSR is read-write; however, for PCI operation, an attempt to program this register to any value other than 0x8 results in clearing it. For PCI-X operation, this register is hardwired to 0x8. 16.3.2.10 PCI Bus Latency Timer Register—0x0D Table 16-34 describes the PCI latency timer register (PLTR). 7 3 2 0 R Latency Timer W Reset Offset 0000_0000 0x0D Latency Timer Figure 16-36. PCI Bus Latency Timer Register Table 16-34. PCI Bus Latency Timer Register Field Descriptions Bit s 7–3 Name Latency Timer Description The maximum number of PCI clocks that the device, when mastering a transaction, holds the bus after PCI bus grant has been negated. The value is in PCI clocks. The PCI 2.2 specification gives rules by which the PCI bus interface unit completes transactions when the timer has expired. Read-only bits. The minimum latency timer value when set is 8 PCI clocks. 2–0 Latency Timer 16.3.2.11 PCI Base Address Registers A PCI base address register points to the beginnings of each address range to which the device responds by assertingþ PCI_DEVSEL. The base address register (BAR) at offset 0x10 is a fixed MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 16-43 PCI/PCI-X Bus Interface 1-Mbyte window that is automatically translated to the local configuration, control, and status registers address space. The other base address registers are aliases (with differing format) of the PCI inbound ATMU windows; see Section 16.3.1.3, “PCI/X ATMU Inbound Registers.” The 32-bit base address register at offset 0x14 corresponds to inbound ATMU window 1; the 64-bit base address registers at offsets 0x18 and 0x20 correspond to inbound ATMU windows 2 and 3. If one of these registers is written, the corresponding ATMU register is also updated; if a PCI inbound ATMU register is written, the corresponding BAR is also updated. If one of these registers is read, the corresponding size of ATMU is returned on the PCI bus providing valid window size in the Inbound ATMU window attributes register. Note that PCSRBAR cannot be updated through the inbound ATMU registers. 31 20 19 4 3 2 1 0 R ADDRESS W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x10 PREF TYPE MSI Figure 16-37. PCI Configuration and Status Register Base Address Register (PCSRBAR) Table 16-35. PCSRBAR Field Descriptions Bits 31–2 0 19–4 3 2–1 0 Name Description ADDRES Indicates the base address at which the inbound configuration/run-time window resides. This S window is fixed at 1 Mbyte. — PREF TYPE MSI Reserved Prefetchable Type. 00—Locate anywhere in 32-bit address space. Memory space indicator 31 12 11 4 3 2 1 0 R ADDRESS W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x14 PREF TYPE MSI Figure 16-38. 32–Bit Memory Base Address Register MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 16-44 Freescale Semiconductor Memory Map/Register Definitions Table 16-36. 32-Bit Memory Base Address Register Field Descriptions Bits 31–1 2 11–4 3 2–1 0 Name Description ADDRES Indicates the base address at which the inbound memory window resides. The number of upper S bits that the device allows to be writable is selected through the inbound translation windows. — PREF TYPE MSI Reserved. The device allows a 4 Kbyte window minimum. Prefetchable Type: 00—Locate anywhere in 32-bit address space. Memory Space Indicator. 31 16 R ADDRESS W Reset 15 12 11 0000_0000_0000_0000 4 3 2 1 0 R ADDRESS W Reset Offset 0000_0000_0000_0100 0x18, 0x20 PRE TYPE MSI Figure 16-39. 64-Bit Low Memory Base Address Register Table 16-37. 64-Bit Low Memory Base Address Register Field Descriptions Bits 31–1 2 11–4 3 2–1 0 Name Description ADDRES Indicates the base address at which the inbound memory window resides. The number of upper S bits that the device allows to be writable is selected through the inbound translation windows. — PREF TYPE MSI Reserved. The device allows a 4 Kbyte window minimum. Prefetchable Type: 10—Locate anywhere in 64-bit address space. Memory space indicator MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 16-45 PCI/PCI-X Bus Interface 31 0 R ADDRESS W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x1C, 0x24 Figure 16-40. 64-Bit High Memory Base Address Register Table 16-38. 64-Bit High Memory Base Address Register Field Description Bits Name Description 31– ADDRESS Indicates the base address that the inbound memory window resides at. The number of upper bits 0 that the device allows to be writable is selected through the inbound translation windows. If no access to local memory is to be permitted by external masters, then all bits are programmed. 16.3.2.12 PCI Subsystem Vendor ID Register The PCI subsystem vendor ID register is used to identify the subsystem. 15 0 R W Reset Offset Subsystem Vendor ID 0000_0000_0000_0000 0x2C Figure 16-41. PCI Subsystem Vendor ID Register Table 16-39. PCI Subsystem Vendor ID Register Field Description Bits 15–0 Name Subsystem Vendor ID 0x0000 Description 16.3.2.13 PCI Subsystem ID Register The PCI Subsystem ID register is used to identify the subsystem. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 16-46 Freescale Semiconductor Memory Map/Register Definitions 15 0 R W Reset Offset Subsystem ID 0000_0000_0000_0000 0x2E Figure 16-42. PCI Subsystem ID Register Table 16-40. PCI Subsystem ID Register Field Description Bits 15–0 Name Subsystem ID 0x0000 Description 16.3.2.14 PCI Bus Capabilities Pointer Register The PCI bus capabilities pointer identifies additional functionality supported by the device. This field is set to point to the PCI-X configuration registers. The definition of each bit is given in Table 16-41. 7 0 R W Reset Offset Capabilities Pointer 0110_0000 0x34 Figure 16-43. PCI Bus Capabilities Pointer Register Table 16-41. PCI Bus Capabilities Pointer Register Field Description Bits 7–0 Name Capabilities Pointer Description Specifies the byte offset in the configuration space containing the first item in the capabilities list. The default value points to the PCI-X configuration registers. 16.3.2.15 PCI Bus Interrupt Line Register 7 0 R W Reset Offset Interrupt Line 0000_0000 0x3C Figure 16-44. PCI Bus Interrupt Line Register MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 16-47 PCI/PCI-X Bus Interface Table 16-42. PCI Bus Interrupt Line Register Field Description Bits 7–0 Name Interrupt Line Description This register is used to communicate interrupt line routing information (hard-wired to zero). 16.3.2.16 PCI Bus Interrupt Pin Register The MPC8540 does not generate specific PCI_INTA, PCI_INTB, etc. outputs, nor does this device respond to INTACK or special cycle commands on the PCI/X interfaces. The MPC8540 does have 12 general purpose interrupt request lines (IRQ[0:11]) and an interrupt output, IRQ_OUT (active low, level sensitive), to which all external and most internal interrupt sources (including PCI/X) can be routed. 7 0 R W Reset Offset Interrupt Pin 0000_0000 0x3D Figure 16-45. PCI Bus Interrupt Pin Register Table 16-43. PCI Bus Interrupt Pin Register Field Description Bits 7–0 Name Interrupt Pin No PCI_INT pin selected Description 16.3.2.17 PCI Bus Minimum Grant (MIN_GNT) Register 7 0 R W Reset Offset MIN_GNT 0000_0000 0x3E Figure 16-46. PCI Bus Minimum Grant Register Table 16-44. PCI Bus Minimum Grant Register Field Description Bits 7–0 Name Description MIN_GNT Specifies the length of the device’s burst period (0x00 indicates that this PCI controller has no major requirements for the settings of latency timers) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 16-48 Freescale Semiconductor Memory Map/Register Definitions 16.3.2.18 PCI Bus Maximum Latency (MAX_LAT) Register 7 0 R W Reset Offset MAXLAT 0000_0000 0x3F Figure 16-47. PCI Bus Maximum Latency Register Table 16-45. PCI Bus Maximum Latency Register Field Description Bits 7–0 Name MAXLAT Description Specifies how often the device needs to gain access to the PCI bus (0x00 indicates that this PCI controller has no major requirements for the settings of latency timers) 16.3.2.19 PCI Bus Function Register (PBFR) The 2-byte PCI bus function register is used to determine how different features of the PCI interface are configured. This register is at PCI configuration space at offset 0x44. 15 6 5 4 3 2 1 0 R W Reset Offset 0 0 0 0 0 0 0 0 0 0 ACL 0 P64 0 0 PAH 0000_0000_00*0_*00* *Depends on the state of the reset configuration signals at reset 0x44 Figure 16-48. PCI Bus Function Register Table 16-46. PCI Bus Function Register Field Descriptions Bits 15– 6 5 Name — ACL Reserved Agent configuration lock. Indicates to an external host whether the local processor is doing internal configuration and must be explicitly set and cleared by the local processor during this time. ACL is set during reset if the LA27 (cfg_cpu_boot) input selects the CPU as the configuration owner. This bit is only meaningful in agent mode. 0 PCI interface allows incoming PCI configuration cycles. 1 PCI interface retries all incoming PCI configuration cycles. Reserved Description 4 — MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 16-49 PCI/PCI-X Bus Interface Table 16-46. PCI Bus Function Register Field Descriptions (continued) Bits 3 Name P64 Description PCI 64-bit configuration. Read-only. Indicates the reset value of the PCI 64-bit configuration signal, PCI_REQ64. 0 64-bit interface functions as a 32-bit interface. 1 64-bit interface functions as a 64-bit interface. Reserved PCI agent/host. Read-only. Indicates the reset value of the PCI host/agent configuration signal, LWE3. 0 PCI interface is in host mode 1 PCI interface is in agent mode 2–1 0 — PAH 16.3.2.20 PCI Bus Arbiter Configuration Register (PBACR) The PCI bus arbiter configuration register is used to determine the configuration of the PCI bus arbiter. 15 14 13 12 11 10 9 8 7 6 2 1 0 R PAD PM W Reset Offset 0 PBMD 0 0 0 0 0 PBMP 0 DP 0000_0000_0000_0000 0x46 Figure 16-49. PCI Bus Arbiter Configuration Register Table 16-47. PCI Bus Arbiter Configuration Register Field Descriptions Bits 15 Name PAD Description PCI arbiter disable. Determines if the device is the PCI arbiter on the PCI bus or not. The reset state is determined by the inverse of the PCI_GNT2 configuration input signal when reset is released. 0 Device is the PCI arbiter. 1 Device is not the PCI arbiter. Device presents its request on PCI_REQ0 to the external arbiter and receives its grant on PCI_GNT0. Parking mode. controls which device receives the bus grant when there are no outstanding bus requests and the bus is idle. 0 The bus is parked on the last device to use the bus. 1 The bus is parked on this device. Reserved PCI broken master disable. Determines if the device ignores the bus requests of an initiator that requests the bus for an excessive period without using the bus. 0 An initiator that requests the bus and receives the grant must begin using the bus within 16 PCI clock periods after the bus becomes idle or else its request is subsequently ignored. 1 No requests are ignored. Reserved 14 PM 13 12 — PBMD 11–7 — MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 16-50 Freescale Semiconductor Memory Map/Register Definitions Table 16-47. PCI Bus Arbiter Configuration Register Field Descriptions (continued) Bits 6–2 Name PBMP Description PCI bus master priorities. Determines arbitration priority given to different masters on the PCI bus. Bit 6 corresponds to the priority of the master sourcing PCI_REQ0; bit 2 corresponds to the priority of the master sourcing PCI_REQ4. 0 Master n is low priority. 1 Master n is high priority. Reserved Device priority. Determines this device’s arbitration priority. 0 Device is low priority. 1 Device is high priority. 1 0 — DP 16.3.2.21 PCI-X Next Capabilities ID Register—0x60 This register is an additional standard register specified by the PCI-X Addendum to the PCI Local Bus Specification. 7 0 R W Reset Offset Capabilities ID 0000_0111 0x60 Figure 16-50. PCI-X Next Capabilities ID Register Table 16-48. PCI-X Next Capabilities ID Register Field Descriptions Bits 7–0 Name Capabilities ID Description This register identifies this item in the capabilities list as a PCI-X device having PCI-X registers. 16.3.2.22 PCI-X Capability Pointer Register—0x61 The PCI-X next capabilities pointer register is an additional standard register specified by the PCI-X Addendum to the PCI Local Bus Specification. This register identifies additional functionality supported by the device. The definition of each bit is given in Table 16-49. 7 0 R W Reset Offset Capability Pointer 0000_0000 0x61 Figure 16-51. PCI-X Capability Pointer Register MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 16-51 PCI/PCI-X Bus Interface Table 16-49. PCI-X Capability Pointer Register Field Description Bits 7–0 Name Capability Pointer Description Specifies the byte offset in the configuration space containing the first item in the capabilities list. 16.3.2.23 PCI-X Command Register—0x62 The 2-byte PCI-X command register is an additional standard register specified by the PCI-X Addendum to the PCI Local Bus Specification. This register provides control over the ability to generate and respond to PCI-X cycles. Table 16-50 describes the bits of the PCI-X command register. 15 7 6 4 3 2 1 0 R W Reset Offset 0 0 0 0 0 0 0 0 0 MOST MMRBC ERO DPRE 0000_0000_0011_0000 0x62 Figure 16-52. PCI-X Command Register Table 16-50. PCI-X Command Register Field Descriptions Bits 15–7 6-4 3-2 1 0 Name — MOST Reserved Maximum outstanding split transactions (4 supported) Description MMRBC Maximum memory read byte count (128 supported) ERO DPRE Enable relax ordering (read-only) Data parity error recovery enable (supported) 16.3.2.24 PCI-X Status Register—0x64 The 4-byte PCI-X status register is an additional standard register specified by the PCI-X Addendum to the PCI Local Bus Specification. This register is used to record additional status information for PCI-X bus-related events. In agent mode, the device and bus number are updated by the inbound configuration write transaction except for the inbound configuration write to PCI-X status register. The definition of each bit is given in Table 16-51. Only 4-byte accesses to address offset 0x64 are allowed. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 16-52 Freescale Semiconductor Functional Description 31 30 29 28 26 25 23 22 21 20 19 18 17 16 R W Reset 0 0 RSM DMCRS DMOST DMRBC DC USC SCD MC SBD 0001_0001_1000_001* *0 = 32-bit bus, 1 = 64-bit bus 15 8 7 3 2 0 R W Reset Offset DBN DDN DFN 1111_1111_1111_1010 0x64 Figure 16-53. PCI-X Status Register Table 16-51. PCI-X Status Register Field Descriptions Bits 31–30 29 28–26 25–23 22–21 20 19 18 17 16 15–8 7–3 2–0 Name — RSM DMCRS DMOST DMRBC DC USC SCD MC SBD DBN DDN DFN Reserved Receive split completion error message (write-1-to-clear) Designed maximum cumulative read size (16 Kbytes) Designed maximum outstanding split transactions (4 outstanding) Designed maximum memory read byte count (128 Bytes) Device complexity (simple device, can retry writes) Unexpected split completion (write-1-to-clear) Split completion discarded (write-1-to-clear) 133-MHz capability 64-bit device Bus number (used for diagnostic purposes) Device number (used for diagnostic purposes) Function number (used for diagnostic purposes) Description 16.4 Functional Description This section describes the functionality of the PCI/X interface. 16.4.1 PCI/X Bus Arbitration PCI/X bus arbitration is access-based. Bus masters must arbitrate for each access performed on the bus. The PCI/X bus uses a central arbitration scheme where each master has a unique request MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 16-53 PCI/PCI-X Bus Interface (REQ) output and grant (GNT) input signal. A simple request/grant handshake is used to gain access to the bus. Arbitration for the bus occurs during the previous access so that no PCI/X bus cycles are consumed due to arbitration (except when the bus is idle). The MPC8540 provides bus arbitration logic for its master interface and up to five other external PCI/X bus masters. The on-chip PCI/X arbiter functions in both host and agent modes, or it can be disabled to allow for an external PCI/X arbiter. A configuration signal (PCI_GNT2) sampled at the negation of HRESET determines if the on-chip PCI/X arbiter is enabled (high) or disabled (low). The state of the reset signal is reflected in bit 15 (read-only) of the PCI/X bus arbitration control register (PBACR[PAD]). Note that PAD is the inverse of the arbiter configuration signal; that is, when PAD = 0 the arbiter is enabled, and when PAD = 1 the arbiter is disabled. See Chapter 4, “Reset, Clocking, and Initialization,” for more information on the reset configuration signals. If the on-chip PCI/X arbiter is enabled, a request-grant pair of signals is provided for each external master (PCI_REQ[0:4] and PCI_GNT[0:4]). In addition, the internal request/grant pair for the internal master state machine of the MPC8540 governs processor, DMA, and RapidIO accesses to the PCI/X interface. If the on-chip PCI/X arbiter is disabled, the MPC8540 uses the PCI_REQ0 signal as an output to issue its request to the external arbiter and uses the PCI_GNT0 signal as an input to receive its grant from the external arbiter. The following sections describe the operation of the on-chip PCI/X arbiter that arbitrates between external PCI/X masters and the internal PCI/X bus master of the MPC8540. 16.4.1.1 PCI/X Bus Arbiter Operation The on-chip PCI/X arbiter uses a programmable two-level, round-robin arbitration algorithm. Each of the five external masters, plus the MPC8540, can be programmed for two priority levels, high or low, using the appropriate bits in the PBACR. Within each priority group, the PCI/X bus grant is asserted to the next requesting device in numerical order, with the MPC8540 positioned before device 0. Conceptually, the lowest-priority device is the master that is currently using the bus, and the highest-priority device is the device that follows the current master in numerical order and group priority. This is considered to be a fair algorithm, since a single device cannot prevent other devices from having access to the bus; it automatically becomes the lowest-priority device as soon as it begins to use the bus. If a master is not requesting the bus, then its transaction slot is given to the next requesting device within its priority group. A grant is awarded to the highest-priority requesting device as soon as the current master begins a transaction; however, the granted device must wait until the bus is relinquished by the current master before initiating a transaction. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 16-54 Freescale Semiconductor Functional Description The grant given to a particular device may be removed and awarded to another higher-priority device, whenever the higher-priority device asserts its request. If the bus is idle when a device requests the bus, then the arbiter withholds the grant for 1 clock cycle. The arbiter re-evaluates the priorities of all requesting devices and grants the bus to the highest-priority device in the following clock cycle. This allows a turnaround clock when a higher-priority device is using address stepping or when the bus is parked. The low-priority group collectively has one bus transaction request slot in the high-priority group. For N high-priority devices and M low-priority devices, each high-priority device is guaranteed at least 1 of N+1 bus transactions and each low-priority device is guaranteed at least 1 of (N+1) × M bus transactions, with one low-priority device receiving the grant in 1 of N+1 bus transactions. If all devices are programmed to the same priority level, or if the low-priority group has only one device, the algorithm defaults to give each device an equal number of bus grants in round-robin sequence. For the example in Figure 16-54, assume that several devices are requesting the bus. If two masters are in the high-priority group and three are in the low-priority group, each high-priority master is guaranteed at least one out of three transaction slots and each low-priority master is guaranteed one out of nine transaction slots. In Figure 16-54, the grant sequence (with all devices, except device 4 requesting the bus and device 3 being the current master) is 0, 2, MPC8540, 0, 2, 1, 0, 2, 3, …, and repeating. If device 2 is not requesting the bus, the grant sequence is 0, MPC8540, 0, 1, 0, 3, …, and repeating. If device 2 requests the bus when device 0 is conducting a transaction and the MPC8540 has the next grant, the MPC8540 has its grant removed and device 2 is awarded the grant since device 2 is higher priority than the MPC8540 when device 0 has the bus. High-Priority Group Device 2 (1/3) LowPriority Slot (1/3) Low-Priority Group Device 1 (1/9) Device 0 (1/3) MPC8540 (1/9) Device 3 (1/9) Figure 16-54. PCI/X Arbitration Example MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 16-55 PCI/PCI-X Bus Interface 16.4.1.2 PCI/X Bus Parking When no device is using or requesting the bus, the PCI/X arbiter grants the bus to a selected device. This is known as parking the bus on the selected device. The selected device is required to drive the PCI_AD[31:0], PCI_C/BE[0:3], and the PCI/X parity signals to a stable value, preventing these signals from floating. The parking mode parameter (PBACR[PM]) determines which device the arbiter selects for parking the PCI/X bus. If PBACR[PM] = 0 (or if the bus is not idle), then the bus is parked on the last master to use the bus. If the bus is idle and PBACR[PM] = 1, the bus is parked on the MPC8540. 16.4.1.3 Broken Master Lock-Out (PCI only) The PCI bus arbiter on the MPC8540 has a feature that allows it to lock out any masters that are broken or ill-behaved. The broken master feature is controlled by programming bit 12 of the PCI bus arbitration control register (0 = enabled, 1 = disabled). Note that the PCI-X bus arbiter does not support the broken master lockout function. When the broken master feature is enabled, a granted device that does not assert PCI_FRAME within 16 PCI clock cycles after the bus is idle, has its grant removed and subsequent requests are ignored until its REQ is negated for at least one clock cycle. This prevents ill-behaved masters from monopolizing the bus. When the broken master feature is disabled, a device that requests the bus and receives a grant never loses its grant until and unless it begins a transaction or negates its REQ signal. Note that disabling the broken master feature is not recommended. 16.4.1.4 Power-Saving Modes and the PCI Arbiter In the sleep power-saving mode, the clock signal driving SYSCLK can be disabled. If the clock is disabled, arbitration logic cannot perform its function. System programmers must park the bus with a device that can sustain the PCI_AD[31:0], PCI_C/BE[3:0], and parity signals prior to disabling the SYSCLK signal. If the bus is parked on the MPC8540 when its clocks are stopped, the MPC8540 sustains the PCI_AD[31:0], PCI_C/BE[3:0], and parity signals in their prior states. In this situation, the only way for another agent to use the PCI bus is by waking the MPC8540. In the doze or nap power-saving modes, the arbiter continues to operate allowing other PCI devices to run transactions. 16.4.2 PCI Bus Protocol This section provides a general description of the PCI bus protocol. Specific PCI bus transactions are described in Section 16.4.2.7, “PCI Bus Transactions.” Refer to Figure 16-55, Figure 16-56, Figure 16-57, and Figure 16-58 for examples of the transfer-control mechanisms described in this section. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 16-56 Freescale Semiconductor Functional Description All signals are sampled on the rising edge of the PCI bus clock (SYSCLK). Each signal has a setup and hold aperture with respect to the rising clock edge in which transitions are not allowed. Outside this aperture, signal values or transitions have no significance. See the MPC8540 Integrated Processor Hardware Specifications for specific setup and hold times. 16.4.2.1 Basic Transfer Control The basic PCI bus transfer mechanism is a burst, composed of an address phase followed by one or more data phases. Fundamentally, all PCI data transfers are controlled by three signals—PCI_FRAME (frame), PCI_IRDY (initiator ready), and PCI_TRDY (target ready). An initiator asserts PCI_FRAME to indicate the beginning of a PCI bus transaction and negates PCI_FRAME to indicate the end of a PCI bus transaction. An initiator negates PCI_IRDY to force wait cycles. A target negates PCI_TRDY to force wait cycles. The PCI bus is considered idle when both PCI_FRAME and PCI_IRDY are negated. The first clock cycle in which PCI_FRAME is asserted indicates the beginning of the address phase. The address and bus command code are transferred in that first cycle. The next cycle begins the first of one or more data phases. Data is transferred between initiator and target in each cycle that both PCI_IRDY and PCI_TRDY are asserted. Wait cycles may be inserted in a data phase by the initiator (by negating PCI_IRDY) or by the target (by negating PCI_TRDY). Once an initiator has asserted PCI_IRDY, it cannot change PCI_IRDY or PCI_FRAME until the current data phase completes regardless of the state of PCI_TRDY. Once a target has asserted PCI_TRDY or PCI_STOP, it cannot change PCI_DEVSEL, PCI_TRDY, or PCI_STOP until the current data phase completes. In simpler terms, once an initiator or target has committed to the data transfer, it cannot change its mind. When the initiator intends to complete only one more data transfer (which could be immediately after the address phase), PCI_FRAME is negated and PCI_IRDY is asserted (or kept asserted), indicating the initiator is ready. After the target indicates the final data transfer (by asserting PCI_TRDY), the PCI bus may return to the idle state (both PCI_FRAME and PCI_IRDY are negated) unless a fast back-to-back transaction is in progress. In the case of a fast back-to-back transaction, an address phase immediately follows the last data phase. 16.4.2.2 PCI Bus Commands A PCI bus command is encoded in PCI_C/BE[3:0] during the address phase of a PCI transaction. The bus command indicates to the target the type of transaction the initiator is requesting. Table 16-52 describes the PCI bus commands implemented by the MPC8540. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 16-57 PCI/PCI-X Bus Interface Table 16-52. PCI Bus Commands PCI_C/ BE[3:0] 0000 PCI Bus Command Interruptacknowledge MPC8540 MPC8540 Supports as Supports an Initiator as a Target Yes No Definition A read (implicitly addressing the system interrupt controller). Only one device on the PCI bus should respond to this command; others ignore it. See Section 16.4.2.12.1, “Interrupt-Acknowledge Transactions,” for more information. Provides a way to broadcast select messages to all devices on the PCI bus. See Section 16.4.2.12.2, “Special-Cycle Transactions,” for more information. Accesses agents mapped into the PCI I/O space. Accesses agents mapped into the PCI I/O space. — — Accesses either local memory or agents mapped into PCI memory space, depending on the address. When a PCI master issues this command to local memory, the MPC8540 (the target) fetches data from the requested address to the end of the cache line (32 bytes) from local memory, even though all of the data may not be requested by (or sent to) the initiator. Accesses either local memory or agents mapped into PCI memory space, depending on the address. — — 0001 Special cycle Yes No 0010 0011 0100 0101 0110 I/O-read I/O-write Reserved1 Reserved1 Memory-read Yes Yes No No Yes No No No No Yes 0111 1000 1001 1010 1011 1100 1101 1110 Memory-write Reserved1 Reserved 1 Yes No No Yes Yes Yes Yes Yes Yes No No Configurationread Configurationwrite Memory-readmultiple Dual-addresscycle Memory-readline Agent mode Access the 256-byte configuration space of a PCI agent. A only specific agent is selected when its IDSEL signal is asserted during the address phase. See Section 16.4.2.11, Agent mode “Configuration Cycles,” for details. only Yes Yes Yes Similar to the memory-read command, but also causes a prefetch of the next cache line (32 bytes). Used to transfer a 64-bit address (in two 32-bit address cycles) to 64-bit addressable devices. Indicates that an initiator is requesting the transfer of an entire cache line. This occurs only when the processor is performing a burst read. Note that these processors perform burst reads only when the appropriate cache is enabled and the transaction is not cache-inhibited. Indicates that an initiator is transferring an entire cache line; if this data is in any cacheable memory, that cache line needs to be invalidated. 1111 Memory-writeand-invalidate No Yes 1 Reserved command encodings are reserved for future use. The MPC8540 does not respond to these commands. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 16-58 Freescale Semiconductor Functional Description 16.4.2.3 Addressing PCI defines three physical address spaces—PCI memory space, PCI I/O space, and PCI configuration space. Access to the PCI memory and I/O space is straightforward, although one must take into account the local memory access window and address translation being used. The address translation registers are described in Section 16.3.1, “PCI/X Memory Mapped Registers.” Access to the PCI configuration space is described in Section 16.4.2.11, “Configuration Cycles.” Address decoding on the PCI bus is performed by every device for every PCI transaction. Each agent is responsible for decoding its own address. PCI supports two types of address decoding—positive decoding and subtractive decoding. For positive decoding, each device looks for accesses in the address range that the device has been assigned. For subtractive decoding, one device on the bus looks for accesses that no other device has claimed. See Section 16.4.2.4, “Device Selection,” for information about claiming transactions. The information contained in the two low-order address bits (PCI_AD[1:0]) varies by the address space (memory, I/O, or configuration). Regardless of the encoding scheme, the two low-order address bits are always included in parity calculations. 16.4.2.3.1 Memory Space Addressing For memory accesses, PCI defines two types of burst ordering controlled by the two low-order bits of the address—linear incrementing (PCI_AD[1:0] = 0b00) and cache wrap mode (PCI_AD[1:0] = 0b10), as shown in Table 16-53. The other two PCI_AD[1:0] possibilities (0b01 and 0b11) are reserved. As an initiator, the MPC8540 always encodes PCI_AD[1:0] = 00 for PCI memory space accesses. As a target, the MPC8540 executes a target disconnect after the first data phase completes if PCI_AD[1:0] = 01 or PCI_AD[1:0] = 0b11 during the address phase of a local memory access. See Section 16.4.2.8.2, “Target-Initiated Termination,” for more information on target disconnect conditions. Table 16-53. Supported Combinations of PCI_AD[1:0] MPC8540 as Target PCI_AD[1:0] Read 00 01 10 11 Linear Reserved Cache Wrap Reserved √ TD √ TD Write √ TD TD TD Read √ — — — Write √ — — — MPC8540 as Initiator For linear incrementing mode, the memory address is encoded/decoded using PCI_AD[63:2]. Thereafter, the address is incremented by 4 bytes after each data phase completes until the transaction is terminated or completed (a 4-byte data width per data phase is implied). Note that the two low-order bits on the address bus are included in all parity calculations. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 16-59 PCI/PCI-X Bus Interface For cache wrap mode (PCI_AD[1:0] = 0b10) reads, the critical memory address is decoded using PCI_AD[63:2]. The address is incremented by 4 bytes after each data phase completes until the end of the cache line is reached. For cache-wrap reads, the address wraps to the beginning of the current cache line and continues incrementing until the entire cache line (32 bytes) is read. The MPC8540 does not support cache-wrap write operations and executes a target disconnect after the data phase for the end of the cache line completes for writes with PCI_AD[1:0] = 0b10. That is, the MPC8540 does not wrap back to the beginning of the cache line. Note that the two low-order bits on the address bus are included in all parity calculations. 16.4.2.3.2 I/O Space Addressing For PCI I/O accesses, 32 address signals (PCI_AD[31:0]) are used to provide a byte address. After a target has claimed an I/O access, it must determine if it can complete the entire access as indicated by the byte enable signals. If all the selected bytes are not in the address range of the target, the entire access cannot complete. In this case, the target does not transfer any data and terminates the transaction with a target-abort error. See Section 16.4.2.8.2, “Target-Initiated Termination,” for more information. 16.4.2.3.3 Configuration Space Addressing PCI supports two types of configuration accesses that use different formats for the PCI_AD[31:0] signals during the address phase. The two low-order bits of the address indicate the format used for the configuration address phase—type 0 (PCI_AD[1:0] = 0b00) or type 1 (PCI_AD[1:0] = 0b01). Both address formats identify a specific device and a specific configuration register for that device. See Section 16.4.2.11, “Configuration Cycles,” for descriptions of the two formats. 16.4.2.4 Device Selection The PCI_DEVSEL signal is driven by the target of the current transaction. PCI_DEVSEL indicates to the other devices on the PCI bus that the target has decoded the address and claimed the transaction. PCI_DEVSEL may be driven one, two, or three clock cycles (fast, medium, or slow device select timing) following the address phase. Device select timing is encoded into the device’s PCI bus status register. If no agent asserts PCI_DEVSEL within three clock cycles of PCI_FRAME, the agent responsible for subtractive decoding may claim the transaction by asserting PCI_DEVSEL. A target must assert PCI_DEVSEL (claim the transaction) before or coincident with any other target response (assert PCI_TRDY, PCI_STOP, or data signals). In all cases except target-abort, once a target asserts PCI_DEVSEL, it must not negate PCI_DEVSEL until PCI_FRAME is negated (with PCI_IRDY asserted) and the last data phase has completed. For normal termination, negation of PCI_DEVSEL coincides with the negation of PCI_TRDY or PCI_STOP. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 16-60 Freescale Semiconductor Functional Description If the first access maps into a target’s address range, that target asserts PCI_DEVSEL to claim the access. However, if the initiator attempts to continue the burst access across the resource boundary, then the target must issue a target disconnect. The MPC8540 is hardwired for fast device select timing (PCI bus status register [10–9] = 0b00). Therefore, when the MPC8540 is the target of a transaction (local memory access or configuration register access), it asserts PCI_DEVSEL one clock cycle following the address phase. As an initiator, if the MPC8540 does not detect the assertion of PCI_DEVSEL within four clock cycles after the address phase (that is, five clock cycles after it asserts PCI_FRAME), it terminates the transaction with a master-abort termination; see Section 16.4.2.8.1, “Master-Initiated Termination.” 16.4.2.5 Byte Alignment The byte enable signals of the PCI bus (PCI_C/BE[7:0], during a data phase) are used to determine which byte lanes carry meaningful data. The byte enable signals may enable different bytes for each of the data phases. The byte enables are valid on the edge of the clock that starts each data phase and stay valid for the entire data phase. Note that parity is calculated for all bytes regardless of the state of the byte enable signals. See Section 16.4.2.13.1, “PCI Parity,” for more information. If the MPC8540, as a target, detects no byte enables asserted, it completes the current data phase with no permanent change. This implies that on a read transaction, the MPC8540 expects that the data is not changed, and on a write transaction, the data is not stored. 16.4.2.6 Bus Driving and Turnaround To avoid contention, a turnaround cycle is required on all signals that may be driven by more than one agent. The turnaround cycle occurs at different times for different signals. The PCI_IRDY, PCI_TRDY, PCI_DEVSEL, and PCI_STOP signals use the address phase as their turnaround cycle. PCI_FRAME, PCI_C/BE[7:0], and PCI_AD[63:0] signals use the idle cycle between transactions (when both PCI_FRAME and PCI_IRDY are negated) as their turnaround cycle. PCI_PERR has a turnaround cycle on the fourth clock cycle after the last data phase. The PCI address/data signals, PCI_AD[63:0], are driven to a stable condition during every address/data phase. Even when the byte enables indicate that byte lanes carry meaningless data, the signals carry stable values. Parity is calculated on all bytes regardless of the byte enables. See Section 16.4.2.13.1, “PCI Parity,” for more information. 16.4.2.7 PCI Bus Transactions This section provides descriptions of the PCI bus transactions. All bus transactions follow the protocol as described in Section 16.4.2, “PCI Bus Protocol.” Read and write transactions are MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 16-61 PCI/PCI-X Bus Interface similar for the memory and I/O spaces, so they are described as generic read transactions and generic write transactions. The timing diagrams in this section show the relationship of significant signals involved in bus transactions. When a signal is drawn as a solid line, it is actively being driven by the current master or target. When a signal is drawn as a dashed line, no agent is actively driving it. High-impedance signals are indicated to have indeterminate values when the dashed line is between the two rails. The terms ‘edge’ and ‘clock edge’ always refer to the rising edge of the clock; ‘asserted’ and ‘negated’ always refer to the globally visible state of the signal on the clock edge, and not to signal transitions. ‘ ’ represents a turnaround cycle in the timing diagrams. 16.4.2.7.1 PCI Read Transactions This section describes PCI single-beat read transactions and PCI burst read transactions. A read transaction starts with the address phase, occurring when an initiator asserts PCI_FRAME. During the address phase, PCI_AD[63:0] contains a valid address and PCI_C/BE[7:0] contains a valid bus command. The first data phase of a read transaction requires a turnaround cycle. This allows the transition from the initiator driving PCI_AD[63:0] as address signals to the target driving PCI_AD[63:0] as data signals. The turnaround cycle is enforced by the target with the TRDY signal. The target provides valid data at the earliest one cycle after the turnaround cycle. The target must drive the PCI_AD[63:0] signals when PCI_DEVSEL is asserted. During the data phase, the PCI_C/BE[7:0] signals indicate which byte lanes are involved in the current data phase. A data phase may consist of a data transfer and wait cycles. The PCI_C/BE[7:0] signals remain actively driven for both reads and writes from the first clock of the data phase through the end of the transaction. A data phase completes when data is transferred, which occurs when both PCI_IRDY and PCI_TRDY are asserted on the same clock edge. When either PCI_IRDY or PCI_TRDY is negated, a wait cycle is inserted and no data is transferred. The initiator indicates the last data phase by negating PCI_FRAME when PCI_IRDY is asserted. The transaction is considered complete when data is transferred in the last data phase. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 16-62 Freescale Semiconductor Functional Description Figure 16-55 illustrates a PCI single-beat read transaction. SYSCLK PCI_AD[63:0] PCI_C/BE[7:0] PCI_FRAME PCI_IRDY PCI_DEVSEL PCI_ TRDY ADDR CMD BE DATA Figure 16-55. PCI Single-Beat Read Transaction Figure 16-56 illustrates a PCI burst read transaction. SYSCLK PCI_AD[63:0] PCI_C/BE[7:0] PCI_FRAME PCI_IRDY PCI_DEVSEL PCI_TRDY ADDR CMD DATA1 BE1 DATA2 BE2 Figure 16-56. PCI Burst Read Transaction 16.4.2.7.2 PCI Write Transactions This section describes PCI single-beat write transactions, and PCI burst write transactions. A PCI write transaction starts with the address phase, occurring when an initiator asserts PCI_FRAME. A write transaction is similar to a read transaction except no turnaround cycle is needed following the address phase because the initiator provides both address and data. The data phases are the same for both read and write transactions. Although not shown in the figures, the initiator must MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 16-63 PCI/PCI-X Bus Interface drive the PCI_C/BE[7:0] signals, even if the initiator is not ready to provide valid data (PCI_IRDY negated). Figure 16-57 illustrates a PCI single-beat write transaction. SYSCLK PCI_AD[63:0] PCI_C/BE[7:0] PCI_FRAME PCI_IRDY PCI_D EVSEL PCI_ TRDY ADDR CMD DATA BE Figure 16-57. PCI Single-Beat Write Transaction Figure 16-58 illustrates a PCI burst write transaction. SYSCLK PCI_AD[63:0] PCI_C/BE[7:0] PCI_FRAME PCI_IRDY PCI_DEVSEL PCI_ TRDY ADDR CMD DATA1 BE1 DATA2 BE2 DATA3 BE3 DATA4 BE4 Figure 16-58. PCI Burst Write Transaction 16.4.2.8 Transaction Termination A PCI transaction may be terminated by either the initiator or the target. The initiator is ultimately responsible for concluding all transactions, regardless of the cause of the termination. All MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 16-64 Freescale Semiconductor Functional Description transactions are concluded when PCI_FRAME and PCI_IRDY are both negated, indicating the bus is idle. 16.4.2.8.1 Master-Initiated Termination Normally, a master initiates termination by negating PCI_FRAME and asserting PCI_IRDY. This indicates to the target that the final data phase is in progress. The final data transfer occurs when both PCI_TRDY and PCI_IRDY are asserted. The transaction is considered complete when data is transferred in the last data phase. After the final data phase, both PCI_FRAME and PCI_IRDY are negated (the bus becomes idle). There are three types of master-initiated termination: • • Completion—Refers to termination when the initiator has concluded its intended transaction. This is the most common reason for termination. Timeout—Refers to termination when the initiator loses its bus grant (GNTn is negated), and its internal latency timer has expired. The intended transaction is not necessarily concluded. Master-abort—An abnormal case of master-initiated termination. If no device (including the subtractive decoding agent) asserts PCI_DEVSEL to claim a transaction, the initiator terminates the transaction with a master-abort. For a master-abort termination, the initiator negates PCI_FRAME and then negates PCI_IRDY on the next clock. If a transaction is terminated by master-abort (except for a special-cycle command), the received master-abort bit (bit 13) of the PCI bus status register is set. • As an initiator, if the MPC8540 does not detect the assertion of PCI_DEVSEL within four clock cycles following the address phase (five clock cycles after asserting PCI_FRAME), it terminates the transaction with a master-abort. 16.4.2.8.2 Target-Initiated Termination By asserting PCI_STOP, a target may request that the initiator terminate the current transaction. Once asserted, the target holds PCI_STOP asserted until the initiator negates PCI_FRAME. Data may or may not be transferred during the request for termination. If PCI_TRDY and PCI_IRDY are asserted during the assertion of PCI_STOP, data is transferred. However, if PCI_TRDY is negated when PCI_STOP is asserted, it indicates that the target will not transfer any more data; therefore, the initiator does not wait for a final data transfer as it would in a completion termination. When a transaction is terminated by PCI_STOP, the initiator must negate its REQn signal for a minimum of two PCI clock cycles, (one corresponding to when the bus goes to the idle state (PCI_FRAME and PCI_IRDY negated)). If the initiator intends to complete the transaction, it can reassert its REQn immediately following the two clock cycles. If the initiator does not intend to complete the transaction, it can assert REQn whenever it needs to use the PCI bus again. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 16-65 PCI/PCI-X Bus Interface There are three types of target-initiated termination: • Disconnect—Refers to termination requested because the target is temporarily unable to continue bursting. Disconnect implies that some data has been transferred. The initiator may restart the transaction at a later time starting with the address of the next untransferred data. (That is, data transfer may resume where it left off.) Retry—Refers to termination requested because the target is currently in a state where it is unable to process the transaction. Retry implies that no data was transferred. The initiator may start the entire transaction over again at a later time. Note that the PCI Local Bus Specification, Rev. 2.2 requires that all retried transactions must be completed. Target-Abort—An abnormal case of target-initiated termination. Target-abort is used when a fatal error has occurred or when a target can never respond. It cannot respond within eight PCI clock cycles (not including the first data phase). The transaction is attempting to cross a 4-Kbyte boundary. A single beat of data has been transferred and the inbound ATMU is marked non-prefetchable. The end of a cache line has been transferred for a cache-wrap mode write transaction. See Section 16.4.2.3.1, “Memory Space Addressing.” The 16-clock latency timer has expired, and the first data phase has not begun. There is no more internal buffer space available for an inbound transaction. • • As a target, the MPC8540 terminates a transaction with a target disconnect due to the following: • • • • As a target, the MPC8540 responds to a transaction with a retry due to the following: • • Target-abort is indicated by asserting PCI_STOP and negating PCI_DEVSEL. This indicates that the target requires termination of the transaction and does not want the transaction retried. If a transaction is terminated by target-abort, the received target-abort bit (bit 12) of the initiator’s bus status register and the signaled target-abort bit (bit 11) of the target’s bus status register are set. Note that any data transferred in a target-aborted transaction may be corrupt. For PCI writes to local memory, if an address parity error or data parity error occurs, the MPC8540 aborts the transaction internally, but continues the transaction on the PCI bus. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 16-66 Freescale Semiconductor Functional Description Figure 16-59 shows several target-initiated terminations. SYSCLK PCI_FRAME PCI_IRDY PCI_DEVSEL PCI_TRDY PCI_STOP Disconnect A SYSCLK PCI_FRAME PCI_IRDY PCI_DEVSEL PCI_TRDY PCI_STOP Retry SYSCLK PCI_FRAME PCI_IRDY PCI_DEVSEL PCI_TRDY PCI_STOP Disconnect Without Data SYSCLK PCI_FRAME PCI_IRDY PCI_DEVSEL PCI_TRDY PCI_STOP Target Abort SYSCLK PCI_FRAME PCI_IRDY PCI_DEVSEL PCI_TRDY PCI_STOP Disconnect B Figure 16-59. PCI Target-Initiated Terminations MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 16-67 PCI/PCI-X Bus Interface The three disconnect terminations are unique in the data transferred at the end of the transaction. For disconnect A, the initiator is negating PCI_IRDY when the target asserts PCI_STOP and data is transferred only at the end of the current data phase. For disconnect B, the target negates PCI_TRDY one clock after it asserts PCI_STOP, indicating that the target can accept the current data, but no more data can be transferred. For disconnect-without-data, the target asserts PCI_STOP when PCI_TRDY is negated indicating that the target cannot accept any more data. 16.4.2.9 Fast Back-to-Back Transactions The PCI bus allows fast back-to-back transactions by the same master. During a fast back-to-back transaction, the initiator starts the next transaction immediately without an idle state. The last data phase completes when PCI_FRAME is negated, and PCI_IRDY and PCI_TRDY are asserted. The current master starts another transaction in the clock cycle immediately following the last data transfer for the previous transaction. Fast back-to-back transactions must avoid contention on the PCI_TRDY, PCI_DEVSEL, PCI_PERR, and PCI_STOP signals. There are two types of fast back-to-back transactions—those that access the same target and those that access multiple targets sequentially. The first type places the burden of avoiding contention on the initiator; the second type places the burden of avoiding contention on all potential targets. As an initiator, the MPC8540 does not perform any fast back-to-back transactions. As a target, the MPC8540 supports both types of fast back-to-back transactions. During fast back-to-back transactions, the MPC8540 monitors the bus states to determine if it is the target of a transaction. If the previous transaction was not directed to the MPC8540 and the current transaction is directed at the MPC8540, it delays the assertion of PCI_DEVSEL (as well as PCI_TRDY, PCI_STOP, and PCI_PERR) for one clock cycle to allow the other target to stop driving the bus. 16.4.2.10 Dual Address Cycles The MPC8540 supports dual address cycle (DAC) commands (64-bit addressing on PCI bus) as both an initiator and a target. DACs are different from single address cycles (SACs) in that the address phase takes two PCI beats instead of one PCI beat to transfer (64-bit vs. 32-bit addressing). Only PCI memory commands can use DAC cycles; I/O, configuration, interrupt acknowledge, and special cycle command cannot use DAC cycles. The MPC8540 block supports single-beat and burst DAC transactions. For the case of the local processor, DAC generation depends on the setting of the POTEARx. If the POTEARx are programmed with nonzero values and a transaction from the local processor core hits in one of the outbound windows, a DAC transaction is generated on the PCI bus with the translated lower 32-bit addresses. Refer to Section 16.3.1.2, “PCI/X ATMU Outbound Registers,” for more information. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 16-68 Freescale Semiconductor Functional Description Figure 16-60 shows the timing sequence of the PCI signals for single-beat DAC reads. PCI_CLK Lo-Addr Hi-Addr T/A BE PCI_AD[31:0] PCI_C/BE[3:0] PCI_FRAME PCI_IRDY PCI_DEVSEL PCI_TRDY 0x0D Cmd DATA T/A T/A Figure 16-60. DAC Single-Beat Read Example The timing for a DAC burst read is shown in Figure 16-61. PCI_CLK Lo-Addr Hi-Addr T/A PCI_AD[31:0] PCI_C/BE[3:0] PCI_FRAME PCI_IRDY PCI_DEVSEL PCI_TRDY DATA1 BE1 DATA2 BE2 T/A T/A 0x0D Cmd Figure 16-61. DAC Burst Read Example Figure 16-62 and Figure 16-63 show single-beat DAC writes and burst DAC writes. PCI_CLK Lo-Addr Hi-Addr DATA 0x0D Cmd BE T/A T/A PCI_AD[31:0] PCI_C/BE[3:0] PCI_FRAME PCI_IRDY PCI_DEVSEL PCI_TRDY Figure 16-62. DAC Single-Beat Write Example MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 16-69 PCI/PCI-X Bus Interface PCI_CLK Lo-Addr PCI_AD[31:0] PCI_C/BE[3:0] PCI_FRAME PCI_IRDY PCI_DEVSEL PCI_TRDY Hi-Addr DATA1 DATA2 BE2 DATA3 BE3 DATA4 T/A BE4 T/A 0x0D Cmd BE1 Figure 16-63. DAC Burst Write Example Figure 16-64 shows the timing sequence of the PCI signals for 64-bit DAC reads. Address phase 1 PCI_CLK PCI_FRAME PCI_AD[31:0] PCI_C/BE[3:0] PCI_AD[63:32] PCI_C/BE[7:4] (optional) PCI_IRDY PCI_TRDY PCI_DEVSEL Wait F Data transfer M S Wait Data transfer LO-ADDR HI-ADDR DUAL AD BUS CMD HI-ADDR BUS CMD DATA1 BE[3:0] DATA2 BE[7:4] DATA4 DATA3(2) 2 3 4 Data phase 5 6 Data phase 7 8 9 Figure 16-64. 64-Bit Dual Address Read Cycle 16.4.2.11 Configuration Cycles This section describes PCI configuration cycles used for configuring standard PCI devices. The PCI configuration space of any device is intended for configuration, initialization, and catastrophic error-handling functions only. Access to the PCI configuration space should be limited to initialization and error-handling software. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 16-70 Freescale Semiconductor Functional Description 16.4.2.11.1 PCI Configuration Space Header Figure 16-65 shows the predefined header for all PCI devices. Address Offset Device ID Status Class Code BIST Header Type Latency Timer Vendor ID Command Revision ID Cache Line Size 0x00 0x04 0x08 0x0C 0x10 0x14 Base Address Registers 0x18 0x1C 0x20 0x24 Reserved Subsystem ID Subsystem Vendor ID Expansion ROM Base Address Reserved Reserved Max_Lat Min_Gnt Interrupt Pin Interrupt Line 0x28 0x2C 0x30 0x34 0x38 0x3C Figure 16-65. Standard PCI Configuration Header The first 64 bytes of the 256-byte configuration space consists of a predefined header that every PCI device must support. The first 16 bytes of the predefined header are defined the same for all PCI devices; the remaining 48 bytes of the header may have differing layouts depending on the function of the device. Most PCI devices use the configuration header layout shown in Figure 16-65. The rest of the 256-byte configuration space is device-specific. The PCI header specific to the MPC8540 is described in Section 16.3.2, “PCI/X Configuration Header.” Table 16-54 summarizes the configuration header registers. Detailed descriptions of these registers are provided in the PCI Local Bus Specification, Rev. 2.2. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 16-71 PCI/PCI-X Bus Interface Table 16-54. PCI Configuration Space Header Summary Offset 0x00 0x02 0x04 0x06 0x08 0x09 0x0C 0x0D 0x0E Register Name Vendor ID Device ID Command Status Revision ID Class code Cache line size Latency timer Header type Description Identifies the manufacturer of the device (assigned by the PCI SIG (special-interest group) to ensure uniqueness). Identifies the particular device (assigned by the vendor). Provides coarse control over a device’s ability to generate and respond to PCI bus cycles Records status information for PCI bus-related events Specifies a device-specific revision code (assigned by vendor) Identifies the generic function of the device and (in some cases) a specific register-level programming interface Specifies the system cache line size in 32-bit units Specifies the value of the latency timer in PCI bus clock units for the device when acting as an initiator Bits 0–6 identify the layout of bytes 0x10–0x3F; bit 7 indicates a multifunction device. The most common header type (0x00) is shown in Figure 16-65 and in this table. Optional register for control and status of built-in self test (BIST) 0x0F BIST 0x10–0x27 Base address registers Address mapping information for memory and I/O space 0x28 0x2C 0x2E 0x30 — Subsystem Vendor ID Subsystem ID Expansion ROM base address Reserved for future use Identifies the subsystem vendor ID (read-only for MPC8540) Identifies the subsystem ID (read-only for MPC8540) Base address and size information for expansion ROM contained in an add-on board Reserved for future use Contains interrupt line routing information Indicates which interrupt pin the device (or function) uses Specifies the length of the device’s burst period in 0.25-µs units Specifies how often the device needs access to the bus in 0.25-µs units 0x34, 0x38 — 0x3C 0x3D 0x3E 0x3F Interrupt line Interrupt pin Min_Gnt Max_Lat 16.4.2.11.2 Accessing the PCI Configuration Space in Host Mode To access the configuration space, a 32-bit value must be written to the PCI CFG_ADDR register that specifies the target PCI bus, the target device on that bus, and the configuration register to be accessed within that device. Note that the Bus Master bit in the MPC8540’s PCI bus command register must be set before an outbound configuration access is attempted. Device 0 on PCI/X bus 0 is the MPC8540 itself; thus, device 0, bus 0 is used to access the internal PCI/X configuration header. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 16-72 Freescale Semiconductor Functional Description When the MPC8540 detects an access to PCI CFG_DATA, it checks the enable flag and the device number in the PCI CFG_ADDR register. If the enable bit is set, and the device number is not 0b1_1111, the MPC8540 performs a configuration cycle translation function and runs a configuration-read or configuration-write transaction on the PCI bus. If the bus number corresponds to the local PCI bus (bus number = 0x00), the MPC8540 performs a type 0 configuration cycle translation. If the bus number indicates a remote PCI bus (that is, nonlocal), the MPC8540 performs a type 1 configuration cycle translation. The device number 0b1_1111 is used for performing interrupt-acknowledge and special-cycle transactions. See Section 16.4.2.12, “Other Bus Transactions,” for more information. See Section 16.3.1.1.1, “PCI/X Configuration Address Register (CFG_ADDR),” for details on PCI CFG_ADDR and Section 16.3.1.1.2, “PCI/X Configuration Data Register (CFG_DATA),” for details on PCI CFG_DATA. Note that because all PCI/X registers are intrinsically little-endian, in the following examples, the data in the configuration register is shown in little-endian order. PowerPC processor accesses to the PCI CFG_DATA register should use the load/store with byte-reversed instructions. External PCI masters that use the local address map to access configuration space do not need to reverse bytes since byte lane redirection from the little-endian PCI bus is performed internally. Example: Configuration sequence, 4-byte data read from the revision ID/standard programming interface/subclass code/class code registers at address offset 0x08 of the PCI/X configuration header (device 0 on the PCI/X bus 0 is the MPC8540 itself). Initial values: r0 contains r1 contains r2 contains r3 contains Register at 0x8000_0008 CCSRBAR + 0x0_8000 (Address of PCI CFG_ADDR register) CCSRBAR + 0x0_8004 (Address of PCI CFG_DATA register) 0xFFFF_FFFF 0x08 contains 0x0B20_0002 (0x0B to 0x08) Code sequence: stw r0, 0 (r1) lwbrx r3, 0 (r2) Results: Address CCSRBAR + 0x0_8000 contains 0x8000_0008 Register r3 contains 0x0B20_0002 Example: Configuration sequence, 4-byte data write to PCI/X register at address offset 0x14 of Device 1 on PCI/X bus 0. Initial values: r0 contains r1 contains r2 contains r3 contains Register at 0x8000_0814 CCSRBAR + 0x0_8000 (Address of PCI CFG_ADDR register) CCSRBAR + 0x0_8004 (Address of PCI CFG_DATA register) 0x1122_3344 0x14 contains 0xFFFF_FFFF (0x17 to 0x14) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 16-73 PCI/PCI-X Bus Interface Code sequence: stw r0, 0 (r1) // Update PCI CFG_ADDR register to point to //register offset 0x14 of device 1. stwbrx r3, 0 (r2) Results: Address CCSRBAR + 0x0_8000 contains 0x8000_0814 Register at 0x14 contains 0x1122_3344 (0x17 to 0x14) Example: Configuration sequence, 2-byte data write to PCI register at address offset 0x1C of Device 1 on PCI/X bus 0. Initial values: r0 contains r1 contains r2 contains r3 contains Register at 0x8000_081C CCSRBAR + 0x0_8000 CCSRBAR + 0x0_8004 0xDDCC_BBAA 0x1C contains 0xFFFF_FFFF (0x1F to 0x1C) Code sequence: stw r0, 0 (r1) sthbrx r3, 0 (r2) Results: Address CCSRBAR + 0x0_8000 contains 0x8000_081C Register at 0x1C contains 0xFFFF_BBAA (0x1F to 0x1C) 16.4.2.11.3 PCI Configuration in Agent and Agent Lock Modes In general, agents should not access the configuration space of other external PCI devices. Configuration of agents is a function usually reserved for the host. When the MPC8540 is in agent mode, it responds to remote host-generated PCI/X configuration cycles. This occurs when a configuration command is decoded along with the IDSEL input signal being asserted. When the MPC8540 is in agent lock mode, it retries all externally-generated PCI/X configuration cycles until the ACL bit in the PCI bus function register (0x44) is set. See Section 16.5.1, “Power-On Reset Configuration Modes,” for more information. In either agent or agent lock mode, access to the internal PCI/X configuration header by the processor core is handled as described in Section 16.4.2.11.2, “Accessing the PCI Configuration Space in Host Mode,” using device = 0 and bus = 0 in PCI CFG_ADDR to indicate the internal PCI/X header. 16.4.2.11.4 PCI Type 0 Configuration Translation Figure 16-66 shows the PCI type 0 translation function performed on the contents of the PCI CFG_ADDR register to the PCI_AD[31:0] signals on the PCI bus during the address phase of the configuration cycle. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 16-74 Freescale Semiconductor Functional Description Contents of PCI CFG_ADDR Register Reserved Function Number E 01 000_0000 78 Bus Number Device Number 15 16 20 21 Register Number 23 24 00 29 30 31 See Table 16-55 PCI_AD[31:0] Signals During Address Phase IDSEL–Only One Signal High 31 Function/Register Number 11 10 21 00 0 Figure 16-66. PCI Type 0 Configuration Translation For PCI type 0 configuration cycles, the MPC8540 translates the device number field of the PCI CFG_ADDR register into a unique IDSEL signal for up to 21 different devices. Each device connects its IDSEL input to one of the PCI_AD[31:11] signals. For PCI type 0 configuration cycles, the MPC8540 translates the device number to ADn as shown in Table 16-55. Table 16-55. PCI Type 0 Configuration—Device Number to ADn Translation Device Number ADn Used for IIDSEL Binary 0b0_0000–0b0_1001 0b0_1010 0b0_1011 0b0_1100 0b0_1101 0b0_1110 0b0_1111 0b1_0000 0b1_0001 0b1_0010 0b1_0011 0b1_0100 1 Device Number ADn Used for IIDSEL Binary — AD31 AD11 AD12 AD13 AD14 AD15 AD16 AD17 AD18 AD19 AD20 0b1_0101 0b1_0110 0b1_0111 0b1_1000 0b1_1001 0b1_1010 0b1_1011 0b1_1100 0b1_1101 0b1_1110 0b1_1111 1 Decimal 21 22 23 24 25 26 27 28 29 30 31 AD21 AD22 AD23 AD24 AD25 AD26 AD27 AD28 AD29 AD30 — Decimal 0–9 10 11 12 13 14 15 16 17 18 19 20 A device number of all ones indicates a PCI special-cycle or interrupt-acknowledge transaction. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 16-75 PCI/PCI-X Bus Interface For PCI type 0 translations, the function number and register number fields are copied without modification onto the PCI_AD[10:2] signals during the address phase. The PCI_AD[1:0] signals are driven to 0b00 during the address phase for type 0 configuration cycles. The MPC8540 implements address stepping on configuration cycles so that the target’s IDSEL, which is connected directly to one of the PCI_AD lines, reaches a stable value. This means that a valid address and command are driven on PCI_AD[31:0] and PCI_C/BE[3:0] one clock cycle before the assertion of PCI_FRAME. 16.4.2.11.5 Type 1 Configuration Translation For type 1 translations, the MPC8540 copies the 30 high-order bits of the PCI CFG_ADDR register (without modification) onto the PCI_AD[31:2] signals during the address phase. The MPC8540 automatically translates PCI_AD[1:0] into 0b01 during the address phase to indicate a type 1 configuration cycle. 16.4.2.12 Other Bus Transactions There are two other PCI transactions that the MPC8540 supports—interrupt acknowledge and special cycles. As an initiator, the MPC8540 may initiate both interrupt acknowledge and special-cycle transactions; however, as a target, the MPC8540 ignores interrupt-acknowledge and special-cycle transactions. Both transactions make use of the PCI CFG_ADDR and PCI CFG_DATA registers described in Section 16.4.2.11.3, “PCI Configuration in Agent and Agent Lock Modes.” 16.4.2.12.1Interrupt-Acknowledge Transactions The PCI bus supports an interrupt-acknowledge transaction. The interrupt-acknowledge command is a read operation implicitly addressed to the system interrupt controller. Note that the PCI interrupt-acknowledge command does not address the MPC8540 PIC processor interrupt-acknowledge register and does not return the interrupt vector address from the PIC unit. See Chapter 10, “Programmable Interrupt Controller,” for more information about the PIC unit. When the MPC8540 detects a read to PCI CFG_DATA, it checks the enable flag and the device number in PCI CFG_ADDR. If the enable bit is set, the bus number corresponds to the local PCI bus (bus number = 0x00), the device number is all ones (0b1_1111), the function number is all ones (0b111), and the register number is zero (0b00_0000), then the MPC8540 performs an interrupt-acknowledge transaction. If the bus number indicates a nonlocal PCI bus, the MPC8540 performs a type 1 configuration cycle translation, similar to any other configuration cycle for which the bus number does not match. The address phase contains no valid information other than the interrupt-acknowledge command (PCI_C/BE[3:0] = 0b0000). Although there is no explicit address, PCI_AD[63:0] are driven to a stable state, and parity is generated. Only one device (the system interrupt controller) on the PCI bus should respond to the interrupt-acknowledge command by asserting PCI_DEVSEL. All other MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 16-76 Freescale Semiconductor Functional Description devices on the bus should ignore the interrupt-acknowledge command. As stated previously, the MPC8540 PIC unit does not respond to PCI interrupt-acknowledge commands. During the data phase, the responding device returns the interrupt vector on PCI_AD[63:0] when PCI_TRDY is asserted. The size of the interrupt vector returned is indicated by the value driven on PCI_C/BE[7:0]. The MPC8540 also can generate PCI interrupt-acknowledge transactions directly. Reads from PCI INT_ACK at offset 0x0_8008 generate PCI interrupt-acknowledge transactions. Note that processor writes to these addresses do nothing. 16.4.2.12.2Special-Cycle Transactions The special-cycle command provides a mechanism to broadcast select messages to all devices on the PCI bus. The special-cycle command contains no explicit destination address but is broadcast to all PCI agents. When the MPC8540 detects a write to PCI CFG_DATA, it checks the enable flag and the device number in PCI CFG_ADDR. If the enable bit is set, the bus number corresponds to the local PCI bus (bus number = 0x00), the device number is all ones (0b1_1111), the function number is all ones (0b111), and the register number is zero (0b00_0000), then the MPC8540 performs a special-cycle transaction on the local PCI bus. If the bus number indicates a nonlocal PCI bus, the MPC8540 performs a type 1 configuration cycle translation, similar to other configuration cycles for which the bus number does not match. Aside from the special-cycle command (PCI_C/BE[3:0] = 0b0001) the address phase contains no other valid information. Although there is no explicit address, PCI_AD[63:0] are driven to a stable state, and parity is generated. During the data phase, PCI_AD[63:0] contain the special-cycle message and an optional data field. The special-cycle message is encoded on the 16 least-significant bits (PCI_AD[15:0]); the optional data field is encoded on the most-significant 16 lines (PCI_AD[31:16]). The special-cycle message encodings are assigned by the PCI SIG steering committee. The current list of defined encodings are provided in Table 16-56. Table 16-56. Special-Cycle Message Encodings PCI_AD[15:0] 0x0000 0x0001 0x0002 0x0003–0xFFFF Message SHUTDOWN HALT x86 architecture-specific — Note that the MPC8540 does not automatically issue a special-cycle message when it enters any of its power-saving modes. It is the responsibility of software to issue the appropriate special-cycle message, if needed. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 16-77 PCI/PCI-X Bus Interface Each receiving agent must determine whether the special-cycle message is applicable to itself. It is unnecessary to assert PCI_DEVSEL in response to a special-cycle command. The initiator of the special-cycle transaction can insert wait states, but because there is no specific target, the special-cycle message and optional data field are valid on the first clock PCI_IRDY is asserted. All special-cycle transactions are terminated by master-abort; however, the master-abort bit in the initiator’s bus status register is not set for special-cycle terminations. 16.4.2.13 PCI Error Functions PCI provides for parity and other system errors to be detected and reported. This section describes generation and detection of parity and error reporting for the PCI bus. 16.4.2.13.1PCI Parity Generating parity is not optional; it must be performed by all PCI-compliant devices. All PCI transactions, regardless of type, calculate even parity; that is, the number of ones on the PCI_AD[31:0], PCI_C/BE[3:0], and PCI_PAR signals all sum to an even number and the number of ones on the PCI_AD[63:33], PCI_C/BE[7:4], and PCI_PAR64 signals all sum to an even number. Parity provides a way to determine, on each transaction, if the initiator successfully addressed the target and transferred valid data. The PCI_C/BE[7:4] and PCI_C/BE[3:0] signals are included in the parity calculation to ensure that the correct bus command is performed (during the address phase) and correct data is transferred (during the data phase). The agent responsible for driving the bus must also drive even parity on the PAR and PCI_PAR64 signals one clock cycle after a valid address phase or valid data transfer, as shown in Figure 16-67. During the address and data phases, parity covers all 64 address/data signals and 8 command/byte enable signals, regardless of whether all lines carry meaningful information. Byte lanes not actually transferring data must contain stable (albeit meaningless) data and are included in parity calculation. During configuration, special-cycle, or interrupt-acknowledge commands, some address lines are not defined, but are driven to stable values and are included in parity calculation. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 16-78 Freescale Semiconductor Functional Description SYSCLK PCI_AD[63:0] PCI_C/BE[7:0] PCI_PAR, PCI_PAR64 PCI_FRAME PCI_IRDY PCI_DEVSEL PCI_ TRDY PCI_PERR PCI_SERR ADDR CMD DATA BE ADDR CMD DATA BE Figure 16-67. PCI Parity Operation 16.4.2.13.2Error Reporting PCI provides for the detection and signaling of both parity and other system errors. Two signals are used to report these errors—PCI_PERR and PCI_SERR. The PCI_PERR signal is used exclusively to report data parity errors on all transactions except special cycles. The PCI_SERR signal is used for other error signaling including address parity errors and data parity errors on special-cycle transactions; it may also be used to signal other system errors. Note that some errors are enabled by programming two bits—one in the PCI bus command register and another in the PCI/X error enable register (ERR_EN). Likewise, some errors are reported in two bits—one in the PCI bus status register and another in the PCI/X error detect register (ERR_DR). These bits must be cleared separately; that is, clearing one does not clear the other. For example, clearing the ERR_DR[Mstr abort error] does not clear the received master abort bit in the PCI bus status register. In these cases, both bits must be cleared before further error reporting can occur. Table 16-57 shows the actions taken for each kind of error. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 16-79 PCI/PCI-X Bus Interface Table 16-57. PCI Mode Error Actions PCI Error Type Error Detect Register Bit PCI Bus Status Register Bit Comment PCI Outbound Read Received SERR at any phase Received parity error for data phase Rcvd SERR error Mstr PERR error — Detected parity error, Master data parity error detected No data transferred No data transferred Master abort Mstr abort error Received master abort No data transferred Target abort Trgt abort error Received target abort No data transferred Memory space violation ORMSV error — No data transferred. Only 8 bytes are requested in PCI bus PCI Outbound Write Received SERR related to address phase Rcvd SERR error — — Master data parity error detected Received master abort Received target abort — Only 8 bytes transferred. May float AD bus to avoid contention Received SERR related to data phase Rcvd SERR error Received PERR (data phase) Master abort Target abort Memory space violation Mstr PERR error Mstr abort error Trgt abort error OWMSV error PCI Inbound Read Detected parity error for address phase Detected parity error on upper address bus for address phase (SAC or DAC) Received SERR at any phase Received PERR (Data phase) Internal error Addr Parity error — Detected parity error, Signaled system error — Float AD bus No PAR64 check during address phase Rcvd SERR error Trgt PERR error Trgt abort error — — Signaled target abort PCI Inbound Write Detected parity error for address phase Detected parity error on upper address bus for address phase (SAC or DAC) Addr Parity error — Detected parity error, Signaled system error — Cache line purged No PAR64 check during address phase MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 16-80 Freescale Semiconductor Functional Description Table 16-57. PCI Mode Error Actions (continued) PCI Error Type Received SERR at any phase Detected parity error for data phase Error Detect Register Bit Rcvd SERR error Trgt PERR error PCI Bus Status Register Bit — Detected parity error Cache line purged Comment 16.4.3 PCI-X Bus Protocol The PCI-X bus protocol is an enhancement of the PCI bus protocol that features a backward-compatible, higher-bandwidth, 32- or 64-bit multiplexed address/data bus at frequencies up to 133 MHz. The most significant of these enhancements are the following: • • Slightly different command types define burst and DWORD transaction types A new attribute phase for each transaction provides more information about the transaction including a byte count, sequence ID, and handling instructions (for ordering and cacheability) Restricted wait state and disconnection rules A new split transaction, consisting of a split response termination and one or more split completions, replaces the delayed transaction functionality from the PCI bus protocol. • • The following sections describe the new functionality of the PCI-X bus protocol. 16.4.3.1 PCI-X Terminology The PCI-X Addendum to the PCI Local Bus Specification introduces terminology that is helpful when describing aspects of the protocol. This terminology is introduced in context in the following sections. However, three terms that are important to keep in mind throughout these sections are defined as follows: Sequence A sequence is one or more transactions associated with carrying out a single logical transfer by a requester. Each transaction in the same sequence carries the same unique sequence ID. A requester is an initiator that first introduces a transaction into the PCI-X domain.If a transaction is terminated with a split response, the requester becomes the target of the subsequent split completion. A completer is the device addressed by a transaction (other than a split completion transaction). If a target terminates a transaction with a split response, the completer becomes the initiator of the subsequent split completion. Requester Completer MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 16-81 PCI/PCI-X Bus Interface 16.4.3.2 PCI-X Command Encodings The PCI-X protocol uses a slightly different set of command encodings for PCI_C/BE[3:0] during the address phase of a PCI-X transaction. The bus command indicates to the target the type of transaction the initiator is requesting. Table 16-52 describes the PCI-X bus commands implemented by the MPC8540. Table 16-58. PCI-X Command Encodings PCI_C/ BE[3:0] 0000 PCI -X Bus Command Interruptacknowledge MPC8540 MPC8540 Length/ Byte-Enable Supports Supports as Initiator as Target Usage 4 bytes/ Attribute Yes No Definition This command is a read (implicitly addressing the system interrupt controller). Only one device on the PCI bus should respond to this command; other devices ignore it. See Section 16.4.2.12.1, “Interrupt-Acknowledge Transactions.” Provides a way to broadcast select messages to all devices on the PCI bus. See Section 16.4.2.12.2, “Special-Cycle Transactions,” for more information. Accesses agents mapped into the PCI I/O space. Accesses agents mapped into the PCI I/O space. — — Accesses 4-bytes in either local memory or agents mapped into PCI memory space, depending on the address. Accesses either local memory or agents mapped into PCI memory space, depending on the address. Reserved for use in future revisions of the PCI-X specification. Current targets must treat this command as if it were a a memory-read-block. Reserved for use in future revisions of the PCI-X specification. Current targets must treat this command as if it were a a memory-write-block. 0001 Special cycle 4 bytes/ Attribute 4 bytes/ Attribute 4 bytes/ Attribute — — 4 bytes/ Attribute Burst/ Data phase Burst/ None Yes No 0010 0011 0100 0101 0110 I/O-read I/O-write Reserved1 Reserved1 Memory-readDWORD Memory-write Reserved. (Aliased to memory-readblock) Reserved. (Aliased to memory-writeblock) Configurationread Yes Yes — — Yes No No — — Yes 0111 1000 Yes No Yes Aliased 1001 Burst/ None No Aliased 1010 4 bytes/ Attribute Yes Agent Accesses the configuration space of a PCI-X agent. mode only See Section 16.4.3.7, “PCI-X Configuration Transactions,” for more detail on PCI-X configuration cycles. Agent Accesses the configuration space of a PCI-X agent. mode only See Section 16.4.3.7, “PCI-X Configuration Transactions,” for more detail on PCI-X configuration cycles. 1011 Configurationwrite 4 bytes/ Attribute Yes MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 16-82 Freescale Semiconductor Functional Description Table 16-58. PCI-X Command Encodings (continued) PCI_C/ BE[3:0] 1100 PCI -X Bus Command Splitcompletion MPC8540 MPC8540 Length/ Byte-Enable Supports Supports as Initiator as Target Usage Burst/ None Definition Similar to the memory-write command, but the Yes Yes transaction is associated with a previous read or (as (as completer) requester) write transaction that was terminated with a split response. Yes Yes Indicates the transfer a 64-bit address (in two 32-bit address cycles) to 64-bit addressable devices. The actual transaction command is transfered either on PCI_C/BE[3:0] in the second address phase for a 32-bit bus or on PCI_C/BE[7:4] for a 64-bit bus. Indicates that an initiator is requesting the transfer of a block of data. Indicates that an initiator is transferring a block of data. 1101 Dual-addresscycle — 1110 Memory-readblock Memory-writeblock Burst/ None Burst/ None Yes Yes 1111 1 Yes Yes Reserved command encodings are reserved for future use. The MPC8540 does not respond to these commands. 16.4.3.3 PCI-X Attribute Phase PCI-X defines a new attribute phase that occurs one clock cycle after the address phase. During the attribute phase, a specially encoded attribute is driven by the initiator on PCI_C/BE[3:0] and AD[31:0]. PCI_C/BE[3:0] provide either the upper four bits of the byte count for burst transactions or byte enables for DWORD transactions. Figure 16-68 shows PCI_AD[31:0] attributes for burst and DWORD transactions. The attributes for split completion transactions are described in Section 16.4.3.6, “PCI-X Split Transactions,” and the attributes for PCI-X configuration transactions are described in Section 16.4.3.7, “PCI-X Configuration Transactions.” 0 NS RO Tag Requester Bus Number 31 30 29 28 24 23 16 Requester Device Number Requester Function Number 11 10 8 7 Lower Byte Count (Reserved for DWORD transactions) 0 15 Figure 16-68. AD[31:0] During PCI-X Burst/DWORD Attribute Phase MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 16-83 PCI/PCI-X Bus Interface Table 16-59 describes the fields of the attribute for burst and DWORD transactions. Table 16-59. Burst/DWORD Transaction Attribute Summary AD 31 30 Name — NS Reserved. Must be 0 No snoop. This attribute indicates whether the sequence should be snooped for coherency. 0 The requester cannot guarantee that the locations between the starting and ending address, inclusive, of this sequence are not stored in any cache in the system. Thus, all potentially affected caches should be snooped. 1 The requester guarantees that the locations between the starting and ending address, inclusive, of this sequence are not stored in any cache in the system. Thus, snooping is not necessary. Relaxed ordering. This attribute indicates whether the requester requires a transaction to be strictly ordered. In general, devices set the RO attribute for payload sequences and clear the RO attribute for control and status sequences. Refer to the PCI-X Addendum to the PCI Bus Specification for a complete discussion of usage models for relaxed transaction ordering. 0 The requester requires that the transaction remain in strict order. 1 The requester guarantees that the transaction is not required to remain in strict order. The tag attribute identifies up to 32 unique sequences from a single initiator. Tag is also a component of the sequence ID. Description 29 RO 28:24 23:16 15:11 Tag Requester The requester bus number attribute is supplied by the bus number field in the PCI-X status Bus Number register. The requester bus number is also a component of the sequence ID. Requester Device Number Requester Function Number Lower Byte Count The requester device number attribute is supplied by the device number field in the PCI-X status register. The requester bus number is also a component of the sequence ID. The requester function number attribute is supplied by the function number field in the PCI-X status register. The requester bus number is also a component of the sequence ID. For a burst transaction attribute phase, PCI_C/BE[3:0] is concatenated with AD[7:0] to create a 12-bit byte count. The byte count indicates the number of bytes the initiator plans to transfer in the remainder of the sequence. For DWORD transactions, the transaction has an implicit size of up to 4-bytes. Therefore, for DWORD transactions, there is no byte count and AD[7:0] is reserved. PCI_C/BE[3:0] contain the individual byte enables for DWORD transactions. 10:8 7:0 16.4.3.4 PCI-X Transactions Figure 16-69 shows a typical PCI-X write transaction. To highlight some of the difference with the PCI protocol, compare this to the PCI write transaction shown in Figure 16-58. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 16-84 Freescale Semiconductor Functional Description Address Phase Attribute Phase Target Response Phase Turn Around Data Data Data Data Data Data Phase Phase Phase Phase Phase Phase 1 PCI_CLK PCI_AD PCI_C/BE PCI_FRAME PCI_IRDY PCI_TRDY PCI_DEVSEL Bus Transaction Initiator Termination Data Transfer ADDR CMD ATTR ATTR DATA0 DATA1 DATA2 DATA3 DATA4 DATA5 BE0 BE1 BE2 BE3 BE4 BE5 2 3 4 5 6 7 8 9 10 11 12 13 Figure 16-69. Typical PCI-X Write Transaction Figure 16-70 shows a the following burst write transactions: memory write, memory write block, and split completion transactions. Signals preceded by “s1_” indicate a signal that is internal to the device after the signal has been sampled. For the memory write block and split completion transactions the PCI_C/BE[7:0] signals are driven high by the initiator on every data phase. Note that there are no target initial wait states for this example. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 16-85 PCI/PCI-X Bus Interface 1 PCI_CLK 2 3 4 5 6 7 8 PCI Bus 9 10 11 12 13 14 15 PCI_FRAME PCI_AD PCI_C/BE PCI_IRDY PCI_TRDY PCI_DEVSEL Bus Transaction Decode Speed A PCI_FRAME PCI_AD PCI_C/BE PCI_IRDY s1_TRDY s1_DEVSEL Target’s View of the PCI Bus s1_FRAME s1_AD s1_C/BE s1_IRDY PCI_TRDY PCI_DEVSEL ADDR ATTR CMD ATTR DATA0 DATA1 DATA2 DATA3 DATA4 DATA5 DATA6 DATA7 BE0 BE1 BE2 BE3 BE4 BE5 BE6 BE7 ADDR ATTR CMD ATTR DATA0 DATA1 DATA2 DATA3 DATA4 DATA5 DATA6 DATA7 BE0 BE1 BE2 BE3 BE4 BE5 BE6 BE7 Initiator Signals Termination 2 Clocks Before End of Transaction Initiator’s View of the PCI Bus ADDR ATTR CMD ATTR DATA0 DATA1 DATA2 DATA3 DATA4 DATA5 DATA6 DATA7 BE0 BE1 BE2 BE3 BE4 BE5 BE6 BE7 Data Transfer Figure 16-70. Burst Memory Write Transaction Figure 16-71 shows a memory read block transaction. Signals preceded by “s1_” indicate a signal that is internal to the device after the signal has been sampled. Note that there are no target initial wait states for this example. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 16-86 Freescale Semiconductor Functional Description 1 PCI_CLK PCI Bus PCI_FRAME PCI_AD PCI_C/BE PCI_IRDY PCI_TRDY PCI_DEVSEL Bus Transaction Decode Speed A PCI_FRAME PCI_AD PCI_C/BE PCI_IRDY s1_AD s1_TRDY s1_DEVSEL Target’s View of the PCI Bus s1_FRAME s1_AD s1_C/BE s1_IRDY PCI_AD PCI_TRDY PCI_DEVSEL DATA0 DATA1 DATA2 DATA3 DATA4 DATA5 DATA6 DATA7 ADDR ATTR CMD ATTR DATA0 DATA1 DATA2 DATA3 DATA4 DATA5 DATA6 DATA7 ADDR ATTR CMD ATTR ADDR ATTR CMD ATTR DATA0 DATA1 DATA2 DATA3 DATA4 DATA5 DATA6 DATA7 2 3 4 5 6 7 8 9 10 11 12 13 14 15 BE = FFh Data Transfer Initiator Signals Termination 2 Clocks Before End of Transaction Initiator’s View of the PCI Bus BE = FFh Figure 16-71. Memory Read Block Transaction Figure 16-72 shows a DWORD write transaction such as an I/O write. Note that there are no target initial wait states for this example. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 16-87 PCI/PCI-X Bus Interface 1 PCI_CLK PCI_AD[31:0] PCI_C/BE[3:0] PCI_FRAME PCI_IRDY PCI_TRDY PCI_DEVSEL 2 3 4 5 6 7 8 9 10 ADDR CMD ATTR ATTR DATA0 BE = Fh Figure 16-72. DWORD (4-Byte) Write Transaction Figure 16-73 shows a DWORD read transaction that can be used for a memory read DWORD or an I/O read transaction. This example has two target initial wait states before data is transfered. 1 PCI_CLK PCI_AD[31:0] PCI_C/BE[3:0] PCI_FRAME PCI_IRDY PCI_TRDY PCI_DEVSEL ADDR CMD ATTR ATTR DATA0 2 3 4 5 6 7 8 9 10 11 12 BE = Fh Figure 16-73. DWORD (4-Byte) Read Transaction 16.4.3.5 PCI-X Wait State and Termination Rules PCI-X adds the following restrictions to wait state and disconnect rules: • • Initiators are not permitted to insert wait states. Targets are not permitted to insert wait states after the initial data phase; that is, targets are only allowed to insert wait states on the initial data phase. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 16-88 Freescale Semiconductor Functional Description • Initiators and targets can end burst transactions as follows: — After a burst data transfer starts and the target signals it can accept more than one data phase, the transaction can only be stopped in one of the following ways: – Target or initiator disconnect at an allowable disconnect boundary (ADB). An ADB is a naturally-aligned, 128-byte address; that is, an address whose lower 7 bits are zeros. – The transaction byte count is satisfied – Target abort – Burst transactions can start on any byte address. Both the initiator and the target can disconnect on any ADB. A target is also permitted to disconnect a burst transaction after a single data phase. In PCI-X mode, the MPC8540 always terminates an inbound write transaction with a target-disconnect at the ADB. Also in PCI-X mode, in addition to the target-retry conditions listed in Section 16.4.2.8.2, “Target-Initiated Termination,” the MPC8540 may retry a subsequent inbound transaction if a previous outbound read transaction was terminated by a master-abort, a split-completion error message, or a target-abort. 16.4.3.6 PCI-X Split Transactions The basic sequence for a split transaction is as follows: 1. The initiator (requester) requests the bus and the arbiter grants the bus to the requester. 2. The requester initiates a transaction (other than a split completion). 3. The target (completer) communicates its intent to split the transaction by using a split response termination. 4. The completer requests the bus and the arbiter grants the bus to the completer 5. The completer initiates a split completion transaction. Split completions are routed back to original requester across bridges by using requester’s bus, device and function numbers in the split completion address. 16.4.3.6.1 Split Response Figure 16-74 shows a split response to a read transaction. Note that the read could be either a memory read block or a memory read DWORD transaction. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 16-89 PCI/PCI-X Bus Interface 1 PCI_CLK 2 3 4 5 6 FFFFFFFFh 7 8 9 10 PCI_AD PCI_C/BE PCI_FRAME PCI_IRDY PCI_TRDY PCI_STOP PCI_DEVSEL ADDR CMD ATTR ATTR BE = Fh Figure 16-74. Split Response to a Read Transaction Figure 16-74 shows a split response to a write transaction. 1 PCI_CLK PCI_AD[31:0] PCI_C/BE[3:0] PCI_FRAME PCI_IRDY PCI_TRDY PCI_STOP PCI_DEVSEL ADDR CMD ATTR ATTR DATA0 BE = Fh 2 3 4 5 6 7 8 9 10 Figure 16-75. Split Response to a DWORD (4-Byte) Write Transaction 16.4.3.6.2 Completion Address Figure 16-76 shows how AD[31:0] are driven during the address phase for split completion transactions. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 16-90 Freescale Semiconductor Functional Description 00 31 30 RO 29 28 Tag 24 23 Requester Bus Number 16 Requester Device Number 15 11 Requester Function Number 10 8 0 7 6 Lower Address [6–0] 0 Figure 16-76. Split Completion Transaction Address AD[31:0] Table 16-60 describes the fields of the attribute for Burst and DWORD transactions. Table 16-60. Split Completion Transaction Address AD 31:30 29 Name — RO Reserved. Must be 00. Relaxed ordering. The completer copies this field from the RO field of the requester attributes. Bridges optionally use this to influence transaction ordering. 0 The requester requires that the transaction remain in strict order. 1 The requester guarantees that the transaction is not required to remain in strict order. The completer copies this field from the tag field of the requester attributes. The requester uses this information to identify the appropriate split completions. The completer copies this field from the requester bus number field of the requester attributes. The requester uses this information to identify the appropriate split completions. The completer copies this field from the requester device number field of the requester attributes. The requester uses this information to identify the appropriate split completions. Description 28:24 23:16 15:11 10:8 7 6:0 Tag Requester Bus Number Requester Device Number Requester The completer copies this field from the requester function number field of the requester Function Number attributes. The requester uses this information to identify the appropriate split completions. — Lower Address Reserved. Must be 0. The completer copies the 7 lsbs of the split request address if all of the following are true: • The split request for this sequence is a burst read • This is the first split completion of the sequence • The split completion is not a split completion message. If the split completion is disconnected on an ADB, this field is cleared when the sequence resumes. If the split request is a DWORD transaction or the split completion is a split completion message, this field is cleared. 16.4.3.6.3 Completer Attributes PCI-X split completion transactions include an attribute phase one clock cycle after the address phase. The completer attribute for a split completion transaction is similar to the burst transaction attribute. For the split completion attribute, PCI_C/BE[3:0] provide the four msbs of the byte count and AD[31:0] are driven as shown in Figure 16-77. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 16-91 PCI/PCI-X Bus Interface BCM 31 SCE 30 SCM 29 28 00000 24 23 Completer Bus Number 16 Completer Device Number 15 11 Completer Function Number 10 8 7 Lower Byte Count 0 Figure 16-77. AD[31:0] During PCI-X Split Completion Attribute Phase Table 16-61 describes the fields of the completer attribute for PCI-X split completion transactions. Table 16-61. PCI-X Split Completion Transaction Attribute Summary AD 31 Name BCM Description Byte count modified. The completer must set BCM if the byte count is less than the full remaining byte count of the transaction. If the byte count is the full byte count of the split request, the completer clears BCM. Split completion error. The completer sets this bit to indicate the transaction is a split completion message that includes an error message. Split completion message. The completer clears this bit if the split completion contains read data. The completer sets this bit if the split completion contains a split completion message. Reserved. Must be 00000 The completer bus number attribute is supplied by the bus number field in the PCI-X status register. The completer device number attribute is supplied by the device number field in the PCI-X status register. 30 29 28:24 23:16 15:11 10:8 7:0 SCE SCM — Completer Bus Number Completer Device Number Completer The completer function number is assigned to the function in the completer by design. Function Number Lower Byte Count For a split completion attribute phase, PCI_C/BE[3:0] is concatenated with AD[7:0] to create a 12-bit byte count. The byte count indicates the number of bytes remaining in the sequence. 16.4.3.6.4 Split Completion Messages For PCI-X split completion transactions that include a split completion error or split completion message, the transaction has a single DWORD data phase with the format shown in Figure 16-78. Message Class 31 28 27 Message Index 20 0 19 Remaining Lower Address 18 16 Remaining Lower Address (cont.) 15 12 11 Remaining Byte Count 0 Figure 16-78. PCI-X Split Completion Message Format MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 16-92 Freescale Semiconductor Functional Description Table 16-62 describes the fields of the PCI-X split completion message. Table 16-62. PCI-X Split Completion Message Summary AD 31:28 Name Message Class Description A split completion message can be one of the following classes: 0x0 Write completion 0x1 PCI-X bridge error 0x2 Completer error 0x3–0xFReserved The message index identifies the type of message within the message class. For write completion message class: 0x00 Normal completion For PCI-X bridge error message class: 0x00 Master-abort 0x01 Target-abort 0x02 Write data parity error For completer error message class: 0x00 Byte count out-of-range error 0x01 Split write data parity error 0x8n Device-specific error Reserved If the split request was a burst memory read transaction, the remaining lower address is the least significant seven bit of the first byte of data that has not been sent. If the split request was a DWORD transaction, the completer places all zeroes in this field.1 If the split request was a burst memory read transaction, the remaining byte count is the number of bytes that have not been sent. If the split request was a DWORD transaction, the completer places 0x004 in this field. 27:20 Message Index 19 18:12 — Remaining Lower Address Remaining Byte Count 11:0 1 Note that the PCI-X 1.0 specification states that if the split request is a DWORD transaction, the completer sets the remaining lower address field (bits 18–12) of the split completion message to zero. However, the MPC8540 deviates from the specification when the split request is a DWORD read transaction to the upper 32-bits of a quad word (that is, the address ends with 1xx); in this case, bit 14 of the split completion message is set to 1. 16.4.3.7 PCI-X Configuration Transactions PCI-X configuration transactions are similar to PCI configuration transaction as described in Section 16.4.2.11, “Configuration Cycles.” The exceptions are noted in this section. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 16-93 PCI/PCI-X Bus Interface There is a rare situation where an internal access (originating from either the processor core, or other I/O ports on the device) to the 256-byte PCI-X configuration header may be corrupted by concurrent outbound PCI-X data transactions. The corruption does not occur if the PCI-X configuration transactions and the outbound PCI-X data transactions are not concurrent. The corruption does not occur with inbound PCI-X traffic. This is an issue for any system that mixes internally generated PCI-X configuration transactions with outbound PCI-X data transactions. There are two ways to avoid the situation: 1. Ensure that PCI-X configuration transactions are completed before starting PCI-X data transactions. 2. Ensure that all PCI-X outbound data traffic is quiesced before attempting PCI-X configuration transactions. The programmer must ensure all of the following: — Non-configuration transactions to PCI-X from the processor core are quiesced. — The core is not fetching instructions from PCI-X. — DMA transactions to PCI-X are quiesced. — RapidIO bridging transactions to PCI-X are quiesced. As is the case for PCI configuration accesses, the bus master bit in the MPC8540’s PCI bus command register must be set before an outbound configuration access is attempted. Figure 16-79 shows the PCI-X type 0 translation function performed on the contents of the PCI CFG_ADDR register to AD[31:0] on the PCI-X bus during the address phase of the configuration cycle. Contents of PCI CFG_ADDR Register Reserved Function Number E 01 000_0000 78 Bus Number Device Number 15 16 20 21 Register Number 23 24 00 29 30 31 See Table 16-63 PCI_AD[31:0] Signals During Address Phase IDSEL–Only One Signal High 31 16 15 Device/Function/Register Number 00 21 0 Figure 16-79. PCI-X Type 0 Configuration Translation Compare the PCI-X translation in Figure 16-79 with the PCI configuration translation shown in Figure 16-66. Note that for PCI-X mode in addition to being used to generate the IDSEL signal, the device number is repeated on AD[15:11]. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 16-94 Freescale Semiconductor Functional Description For PCI-X type 0 configuration cycles, the MPC8540 translates the device number field of PCI CFG_ADDR into a unique IDSEL signal for up to 15 different devices. Each device connects its IDSEL input to one of the AD[31:17] signals. For PCI-X type 0 configuration cycles, the MPC8540 translates the device number to ADn as shown in Table 16-63. Table 16-63. PCI-X Type 0 Configuration—Device Number to ADn Translation Device Number AD Signal Used for IDSEL Binary 0b0_0000 0b0_0001 0b0_0010 0b0_0011 0b0_0100 0b0_0101 0b0_0110 0b0_0111 0b0_1000 1 Device Number AD Signal Used for IDSEL Binary — AD17 AD18 AD19 AD20 AD21 AD22 AD23 AD24 0b0_1001 0b0_1010 0b0_1011 0b0_1100 0b0_1101 0b0_1110 0b0_1111 0b1_xxxx 0b1_1111 1 Decimal 9 10 11 12 13 14 15 — 31 AD25 AD26 AD27 AD28 AD29 AD30 AD31 — — Decimal 0 1 2 3 4 5 6 7 8 A device number of all ones indicates a PCI-X special-cycle or interrupt-acknowledge transaction. For PCI-X type 0 translations, the function number and register number fields are copied without modification onto PCI_AD[15:2] during the address phase. PCI_AD[1:0] are driven to 0b00 during the address phase for PCI-X type 0 configuration cycles. The MPC8540 implements address stepping on configuration cycles so that the target’s IDSEL, which is connected directly to one of the PCI_AD lines, reaches a stable value. This means that a valid address and command are driven on PCI_AD[31:0] and PCI_C/BE[3:0] four clocks before the assertion of PCI_FRAME. PCI-X configuration transactions include an attribute phase one clock cycle after the address phase. The the attribute for a configuration transaction is similar to the DWORD transaction attribute. For the configuration attribute, PCI_C/BE[3:0] provide the byte enables and AD[31:0] are driven as shown in Figure 16-80. 000 31 29 28 Tag 24 23 Requester Bus Number 16 Requester Device Number 15 11 Requester Function Number 10 8 7 Secondary Bus Number 0 Figure 16-80. AD[31:0] During PCI-X Configuration Attribute Phase MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 16-95 PCI/PCI-X Bus Interface Table 16-64 describes the fields of the attribute for PCI-X configuration transactions. Table 16-64. PCI-X Configuration Transaction Attribute Summary AD 31:29 28:24 Tag 23:16 Requester Bus Number 15:11 Requester Device Number 10:8 7:0 Name — Reserved. Must be 0 Identifies up to 32 unique sequences from a single initiator. Tag is also a component of the sequence ID. Supplied by the bus number field in the PCI-X status register. The requester bus number is also a component of the sequence ID. Supplied by the device number field in the PCI-X status register. The requester bus number is also a component of the sequence ID. Description Requester Supplied by the function number field in the PCI-X status register. The requester bus Function Number number is also a component of the sequence ID. Secondary Bus Number The number of the bus on which the Type 0 configuration transaction is executing 16.4.3.8 PCI-X Error Functions PCI-X provides for parity and other system errors to be detected and reported. This section describes generation and detection of parity and error reporting for the PCI-X bus. 16.4.3.8.1 Error Reporting PCI-X provides for the detection and signaling of both parity and other system errors. Two signals are used to report these errors—PCI_PERR and PCI_SERR. PCI_PERR is used exclusively to report data parity errors on all transactions except special cycles. The PCI_SERR signal is used for other error signaling including address parity errors and data parity errors on special-cycle transactions; it may also be used to signal other system errors. Note that some errors are enabled by programming two bits—one in the PCI bus command register and another in the PCI/X error enable register (ERR_EN). Likewise, some errors are reported in two bits—one in the PCI bus status register and another in the PCI/X error detect register (ERR_DR). These bits must be cleared separately; that is, clearing one does not clear the other. For example, clearing the ERR_DR[Mstr abort error] does not clear the received master abort bit in the PCI bus status register. In these cases, both bits must be cleared before further error reporting can occur. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 16-96 Freescale Semiconductor Functional Description Table 16-65 shows actions taken for each kind of error. Table 16-65. PCI-X Mode Error Actions PCI-X Error Type Error Detect Register Bit PCI Bus Status Register Bit PCI-X Status Register Bit Comment PCI-X Outbound Read Received SERR at any phase Rcvd SERR error — Detected parity error, Master data parity error detected Received master abort Received target abort — Signaled target abort Signaled target abort — — — No data transferred No data transferred Received data parity error Mstr PERR error for data phase Master abort Target abort Memory space violation Mstr abort error Trgt abort error ORMSV error — — — USC No data transferred No data transferred No data transferred. Only 8 bytes are requested in PCI bus No data transferred. Target abort SC (split completion) with SCM error byte count value different than requested SC with wrong lower SCM error address field (burst read) SC with wrong lower address field (dword read) — USC — No data transferred. Target abort Extra SC. No pending SC SCM error with that tag SC error message SC not received before SC PCI-X timer time-out Detected parity error at address or attribute phase of SC Detected parity error at data phase of SC Received SERR at any phase of SC SCM error TOE error Addr parity error — — — Detected parity error, Signaled system error Detected parity error, Master data parity error detected USC RSM — — Master abort No data transferred No data transferred No data transferred Mstr PERR error — No data transferred Rcvd SERR error — PCI-X Outbound Write — No data transferred Received SERR at any phase Received PERR (data phase) Master abort Rcvd SERR error — Rcvd PERR error Master data parity error Rcvd abort error Received master abort — — — Float AD bus MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 16-97 PCI/PCI-X Bus Interface Table 16-65. PCI-X Mode Error Actions (continued) PCI-X Error Type Target abort Memory space violation Error Detect Register Bit Trgt abort error OWMSV error PCI Bus Status Register Bit Received target abort — PCI-X Inbound Read Detected parity error for address or attribute phases Detected parity error on upper address bits for address (SAC) or attribute phases Detected parity error for address (DAC) phases Received SERR at any phase Internal error Memory space violation Received SERR at any phase of SC Received PERR at SC (data phase) Master abort for SC Target abort for SC Addr parity error Detected parity error, Signaled system error — — Split completion error message generated No PAR64 checks for SAC address or attribute phases PCI-X Status Register Bit — — No data transferred. Only 8 bytes are requested in PCI bus Comment — — Addr parity error Detected parity error, Signaled system error — Split completion error message generated Split completion error message generated Split completion error message generated Split completion error message generated Float AD bus Rcvd SERR error — — IRMSV error — — — — — — — — — PCI-X Inbound Write Rcvd SERR error — Trgt PERR error — — — — — Detected parity error for address or attribute phases Detected parity error on upper address bits for address (SAC) or attribute phases Detected parity error on upper address bits for address (DAC) phases Received SERR at any phase Detected parity error for data phase Addr parity error Detected parity error, Signaled system error — — Transaction purged — — No PAR64 checks for SAC address or attribute phases Addr parity error Detected parity error, Signaled system error — Transaction purged Rcvd SERR error — Trgt PERR error Detected parity error — — Transaction purged Transaction from the point of the error is purged MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 16-98 Freescale Semiconductor Initialization/Application Information When an outbound PCI-X read transaction has been split and the split completion transaction indicates a master-abort or a target-abort, the split-completion error message (SCEM) data field contains information for setting the received master-abort or received target-abort bit in the PCI bus status register (bits 13 and 12, respectively). However, for misaligned reads (that is, reads not aligned on a 64-bit boundary), the error is not reflected in the PCI bus status register. The error is reported in the PCI error detection register (ERR_DR) at 0x0_8E00. The master-abort or target-abort can be inferred from the error data capture register (ERR_DL). Drivers that require this information can read the ERR_DR register rather than the PCI bus status register. 16.5 Initialization/Application Information This section describes some tips for use of the PCI/X controller. 16.5.1 Power-On Reset Configuration Modes The PCI/X block can power-on in three modes: host mode, agent mode and agent configuration lock mode. Certain bits in the configuration registers are set differently according to the POR (power-on reset) mode. Also, certain configuration bits have different implications when compared with past Freescale Semiconductor parts and PCI implementations. Note that after reset, the device cannot be switched from one mode to another. The affected configuration bits are defined in Table 16-66. Table 16-66. Affected Configuration Register Bits for POR Register Bi (offset) t PCI bus command register (0x04) 2 1 Name Bus master Register Description Controls whether the device can master a transaction on the PCI/X bus. If cleared, the device can not master a transaction. This bit is independent of host or agent mode. Memory Controls acknowledgement of inbound memory transactions. If cleared, all inbound memory space accesses (including accesses to PCSRBAR space) end in a master abort. This bit is independent of host or agent mode. ACL Valid only in agent mode. Controls acknowledgement of inbound configuration accesses. If set, all inbound configuration accesses are retried. If cleared, inbound configuration accesses are acknowledged. In host mode all inbound configuration accesses end in master aborts. Determines whether the device is in agent or host mode. Zero indicates host mode. PCI bus function register (0x44) 5 0 PAH MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 16-99 PCI/PCI-X Bus Interface The POR reset values for the affected configuration bits are described in Table 16-67. Table 16-67. Power-On Reset Values for Affected Configuration Bits Configuration Bit Mode Bus Master Host Agent Agent configuration lock 1 0 0 Memory Space 0 0 0 ACL X 0 1 PAH 0 1 1 16.5.1.1 Host Mode When the device powers up in host mode, all inbound configuration accesses are ignored (and thus master aborted). The ACL bit is a don’t care. The device powers up with the ability to master transactions on the PCI bus, however in order to acknowledge memory transactions, the memory space bit must be set. 16.5.1.2 Agent Mode When the device powers up in agent mode, it acknowledges inbound configuration accesses. However the device cannot master transactions or acknowledge inbound memory accesses on the PCI bus until the appropriate configuration bits (bus master and memory space, respectively) have been set. 16.5.1.3 Agent Configuration Lock Mode Agent configuration lock mode is similar to agent mode with the added restriction that when the device powers up in agent configuration lock mode, it retries all inbound configuration accesses until the ACL bit is cleared. As in agent mode, the device in agent configuration lock mode cannot master transactions or acknowledge inbound memory accesses on the PCI bus until the appropriate configuration bits (bus master and memory space, respectively) have been set. 16.5.2 Nonposted Writes in PCI-X The PCI-X block may pipeline I/O write and configuration write accesses to a target that signals a split response. To fully serialize these non-posted writes, the processor should follow each write transaction with a read of the same target, followed by proper barrier instructions to guarantee completion of the read before continuing with a subsequent write. This method guarantees completion of the write and serialization of a series of writes. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 16-100 Freescale Semiconductor Initialization/Application Information 16.5.3 PCI-X Outbound Read Transaction Alignment The PCI-X protocol supports only read transactions with byte enables that are 32-bit aligned and are 32 bits wide. The MPC8540 requires the following: 1. Reads to PCI-X greater than 32 bits must be aligned to a 64-bit boundary and have a size that is a multiple of 8 bytes. 2. Reads to PCI-X less than or equal to 32 bits must not cross a 32-bit boundary. A read to PCI-X that is not aligned to a 32-bit boundary that crosses a 32-bit boundary will execute only to the 32-bit boundary. No mechanism is provided in hardware to split this type of read into multiple transactions with byte enables on the PCI-X bus. Care should be taken on the MPC8540 when writing software to meet these alignment restrictions. 16.5.4 PCI-X Inbound Reads that Cross Window Boundaries For Rev 1 of the MPC8540, inbound PCI-X read requests can cross a window boundary. For Rev 2, if an inbound PCI-X read request crosses a window boundary, the MPC8540 responds with a split completion message indicating an error but does not return any data. This is a deviation from the PCI-X specification which describes that data is returned up to the window boundary. 16.5.5 Bridging Between RapidIO and PCI/X The MPC8540 does not support bridging from RapidIO according to the RapidIO Interoperability Specification. If an external RapidIO device tries to target the MPC8540 with a RapidIO NWRITE-R transaction or with any RapidIO write at a priority level higher than 0, and the RapidIO inbound ATMU routes the transaction to the PCI/PCI-X interface, the MPC8540 could deadlock internally. Therefore, the use of RapidIO NWRITE-R transactions must be avoided in bridging applications to the PCI/X controller, and all RapidIO writes targeting the PCI/X controller must be of priority level 0. This means transaction ordering is relaxed because read responses will not necessarily push posted writes in the bridge, and writes will never bypass read requests. Note that this specifically precludes the topology where two MPC8540s communicating with each other via PCI or PCI-X may deadlock when two RapidIO devices (one on each MPC8540) are simultaneously performing priority-1 writes to targets on the remote MPC8540 through the PCI/X link between the devices. To support such a topology, RapidIO write transactions that bridge to the other MPC8540 must be priority 0. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 16-101 PCI/PCI-X Bus Interface MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 16-102 Freescale Semiconductor Chapter 17 RapidIO Interface The RapidIO controller conforms to Revision 1.2 of the Parallel RapidIO Interconnect Specification. This chapter describes the implementation of the parallel RapidIO controller. 17.1 Introduction The RapidIO controller is partitioned into inbound and outbound blocks, and these blocks are further partitioned into three implementation layers: physical, protocol, and logical. Note that these implementation layers do not precisely correlate with the layers described in the Parallel RapidIO Interconnect Specification. The physical and protocol layers primarily address the portion of the Parallel RapidIO Interconnect Specification titled “Physical Layer 8/16 LP-EP.” The logical layer refers primarily to the logical portions of the Parallel RapidIO Interconnect Specification. The physical layer operates at the RapidIO interface applied frequency, and data is sampled at twice this frequency. The protocol and logical layers operate at the device frequency. 17.1.1 Overview A block diagram of The RapidIO controller is shown in Figure 17-1 and includes a RapidIO interface and a system core interface; those signals are divided into several groups. These groups include: the RapidIO request, response, ack done, and data transfer interfaces, the system core request, response and data transfer interfaces, the register access interface, which is used to access the programming model registers, the interrupt interface, the systems functions interface, and the performance monitor interface. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 17-1 RapidIO Interface Core RapidIO Core Core Configuration Request ACK Done Response Data Interface Interface Interface Interface Interface Address/ Attributes Configuration Block Logical Layer Outbound Queue Response Queue RapidIO Data Interface RapidIO RapidIO Request Response Interface Interface Address/ Attributes Inbound Queue Data Data Data Buffer Misalign Block Global Arbiter Ack Generator MSG and Doorbell Unit Request Address Header Assembly Link Request Logic Ack Response Queue Header Assembly Packet and Control Symbol Assembly Protocol Layer CRC Generator Ack History Queue Asynchronous FIFO Throttle Logic Receive Registers Ack Queue Asynchronous FIFO Packet Control Symbols CRC Check Training Symbol Idle Physical Layer Transmit Registers TxData[0:7] RxData[0:7] Figure 17-1. RapidIO High Level Block Diagram The RapidIO controller also implements extra process-specific circuitry for receive clock recovery and for transmit clock and data alignment. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 17-2 Freescale Semiconductor Introduction The RapidIO inbound interface receives and latches 8 double data rate (DDR) data bits. 64 bits are accumulated in the asynchronous boundary FIFO and transferred across the asynchronous boundary into the protocol layer. At the protocol layer, the 64 bits are split into packet header, packet data, and control symbol information fields. After the packet header information arrives in the inbound packet assembly logic, packets are stripped of cyclical redundancy check (CRC) information and other transport information not required by the logical layer. The CRC is checked at the end of a packet. Packet header information and packet data are then passed to the inbound logical layer. The header is interpreted by the logical layer and the transaction is stored in the inbound queue if space is available. The data is held in the inbound data queue until it is transferred out of the RapidIO controller block and into the device. Some control symbols (acks and link responses, for example) are passed to the outbound block for processing. On the outbound side, requests are received from the device at the logical layer and queued in the outbound queue. Transactions from the outbound queue are scheduled according to priority and may be steered to the misalign block or directly to the protocol layer if global arbitration is won. Transactions whose address and data payload are not aligned according to the RapidIO alignment requirements are sent to the misalign block for processing. The response queue handles responses from the device for previously received inbound requests and schedules them as transactions out of the response queue according to priority. Of all transactions from the outbound queue, misalign block, and the response queue, one is prioritized subject to global arbitration to be sent to the protocol layer. At the protocol layer, the packet CRC is performed. The protocol layer logic chooses to transmit either control symbols or packets on the outbound RapidIO interface. The packets and control symbols are then stored in the asynchronous boundary FIFO. The physical layer logic then reads 32-bit quantities from the FIFO and transmits them on the outbound RapidIO interface as 8-bit symbols. While there exist no packet or control symbols to send, the physical layer sends idle symbols. Idle symbols may also be inserted in a packet if a throttle request is received on the inbound interface. The following sections provide greater detail on the organization and operation of RapidIO functionality. These sections are organized as follows: • • • • Section 17.2, “External Signal Descriptions,” describes the interface signal groups, their meanings, and timing relationships. Section 17.3, “Memory Map/Register Definition,” describes the configuration, runtime, and debug registers that constitute the programming model. Section 17.4.1, “RapidIO Transaction, Packet, and Control Symbol Summary,” describes the functional organization and operation of the design. Section 17.5, “RapidIO Functionality,” provides general information about implementing RapidIO features. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 17-3 RapidIO Interface 17.1.2 RapidIO Terminology Table 17-1 describes terminology used throughout this chapter. Table 17-1. RapidIO Terminology Term Ack CAR CRC CSR Domain Doorbell Double word Header ID Logical layer LP-LVDS Mailbox Physical layer Priority Protocol layer Receiver Source Symbol Target Word Acknowledge control symbol Capability register Cyclic redundancy check Command and status register Logically associated group of processing elements Port on a device that generates an interrupt to the processor An 8-byte quantity Part of a packet that contains all of the information about the packet excluding any data Identifier, the name of a processing element on the RapidIO interconnect Handles all logic transformations of packets between the protocol layer and the physical layer Low-power, low-voltage differential signaling Dedicated hardware that receives a message in the message passing unit Physical handling of the transport of packets over the RapidIO interface Relative importance of a transaction or packet; in most systems a higher priority transaction or packet is serviced or transmitted before a lower priority transaction. Handling of packets at the protocol level on the RapidIO interface Destination of a packet on the RapidIO interface, also referred to as a destination or a target Originator of a packet on the RapidIO interconnect One beat of a packet appearing on the interface pins Destination of a packet on the RapidIO interface, also referred to as a destination or a receiver A 4-byte quantity Description 17.1.3 RapidIO Interface Features This section describes the RapidIO interface features. 17.1.3.1 Supported Features RapidIO supports the following features of the Parallel RapidIO Interconnect Specification: • • 8-bit interface as defined by the 8/16 LP-EP physical layer specification Support for non-coherent I/O transactions MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 17-4 Freescale Semiconductor Introduction • • • • • • • • • A message-passing programming model Default and configurable inbound ATMU window Default outbound ATMU window CRC-based error detection (CCITT-CRC16) as described in the 8/16 LP-EP physical specification 256-byte data payloads Packet pacing/retry capability Only the small size transport information field (TT = 0b00) Link validation, error recovery, training, and time-out support as required by the physical layer specification RapidIO error injection 17.1.3.2 Features Not Implemented The following portions of the Parallel RapidIO Interconnect Specification are not supported: • • • Large size transport information field (TT = 0b01) TOD-sync control symbols (these are treated as idle control symbols) Atomic swap or atomic test and swap transactions 17.1.4 RapidIO Modes of Operation This section describes the RapidIO modes of operation. 17.1.4.1 Transmit Clock-Select Mode The RapidIO transmit clock can be chosen from either an internal source or from the RapidIO receive clock signal. The clock source is selected at power-on reset, as described in Section 4.4.3.10, “RapidIO Transmit Clock Source,” and Section 4.4.4.2, “RapidIO Clocks.” Note that the CCB clock is the clock signal for the rest of the RapidIO block. 17.1.4.2 CRC Checking Modes RapidIO offers the modes for packet CRC checking shown in Table 17-2. Note that a read that attempts to access an unmapped target causes the assertion of core_fault_in, which causes the core to generate a machine check interrupt, unless it is disabled (by clearing HID1[RFXE]). If RFXE is zero and this error occurs, PCR[CCE] must be set to ensure that an interrupt is generated. For more information, see Section 6.10.2, “Hardware Implementation-Dependent Register 1 (HID1).” MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 17-5 RapidIO Interface . Table 17-2. CRC Checking Mode PCR[CCE] 0 1 Mode CRC checking disabled CRC enabled with hardware-based recovery 17.1.4.3 Error Checking Disable Mode Error checking can be enabled or disabled by setting a bit in the port control CSR. This mode disables all RapidIO transmission error checking. Device behavior is undefined if error checking is disabled and an error condition occurs. 17.1.4.4 Output Port Enable Mode The output port can be disabled by setting a bit in the port control CSR. In this mode, RapidIO does not issue any packets of its own, except in response to maintenance packets. 17.1.4.5 Output Port Driver Disable Mode This mode, controlled by PCCSR[OPD], disables RapidIO input receivers and output drivers. 17.1.4.6 Accept All Mode This mode is controlled by a bit (CR[AA]) in the RapidIO configuration register. While this mode is enabled, all packets are accepted regardless of the target ID. Table 17-3 describes the behavior with respect to the accept all mode. Table 17-3. Accept All Mode Behavior Accept All (CR[AA]) 0 1 Device Behavior for Target ID ≠ Device ID Error reported in PNFEDR[ITE]. Accept and process the packet without regard to the target ID 17.1.4.7 Packet Response Time-Out Disable Mode The RapidIO packet response time-out counters can be disabled by setting a bit in the RapidIO port notification/fatal error disable register. For more information, see Section 17.3.2.3.2, “Port Notification/Fatal Error Detect Disable Register (PNFEDiR).” 17.1.4.8 Link Response Time-Out Disable Mode The RapidIO link time-out counters can be disabled by setting a bit in the RapidIO port notification/fatal error disable register. For more information, see Section 17.3.2.3.2, “Port Notification/Fatal Error Detect Disable Register (PNFEDiR).” MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 17-6 Freescale Semiconductor External Signal Descriptions 17.2 External Signal Descriptions This section describes the signal names of the RapidIO block, their meanings, and the timing relationships that exist among the signals. 17.2.1 Signals Overview The RapidIO block implements the functionality described in the Parallel RapidIO Interconnect Specification for an 8-bit parallel interface with separate transmit and receive ports as shown in Figure 17-2. RapidIO Interface RIO_RCLK RIO_RD[0:7] RIO_RFRAME 1 8 1 1 8 1 RIO_TCLK RIO_TD[0:7] RIO_TFRAME RIO_TX_CLK_IN Figure 17-2. RapidIO Interface Signals Model The RapidIO signals are implemented as differential pairs of signals. Table 17-4 lists the external RapidIO signals implemented on this interface. Table 17-4. RapidIO Interface Signals Summary Signal Name RIO_RCLK RIO_RCLK RIO_RD [0:7] RIO_RD [0:7] RIO_RFRAME RIO_RFRAME RIO_TCLK RIO_TCLK RIO_TD[0:7] RIO_TD[0:7] RIO_TFRAME RIO_TFRAME RIO_TX_CLK_IN RIO_TX_CLK_IN Port receive clock Port receive clock, inverted (differential pair) Port receive data Port receive data, inverted (differential pair) Port receive framing signal Port receive framing signal, inverted (differential pair) Port transmit clock Port transmit clock, inverted (differential pair) Port transmit data Port transmit data, inverted (differential pair) Port transmit framing signal Port transmit framing signal, inverted (differential pair) Port transmit clock in Port transmit clock in, inverted (differential pair) Function Pins I/O 1 1 8 8 1 1 1 1 8 8 1 1 1 1 I I I I I I O O O O O O I I MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 17-7 RapidIO Interface 17.2.2 Detailed Signal Descriptions Table 17-5 describes the RapidIO interface signals in detail. Table 17-5. RapidIO Interface Signals—Detailed Signal Descriptions Signal RIO_RCLK I/O I Description Receive clock. Free-running input clock used to capture control and data signals. RIO_RCLK is used to capture the frame signal RIO_RFRAME and the data bus RIO_RD[0:7]. RIO_RCLK is the differential pair of RIO_RCLK. State Asserted/Negated—A clock event is determined as the rising or falling edge of RIO_RCLK. Meaning Timing Assertion/Negation—Free-running clock. See the MPC8540 Integrated Processor Hardware Specification for specific timing information for this signal. RIO_RD[0:7] I Receive data. The receive data bus RIO_RD[0:7] carries packet information as well as control symbols. Data is captured on both edges of RIO_RCLK. RIO_RD [0:7] is the differential pair of RIO_RD[0:7] State Asserted/Negated—Represents the data being received on the RapidIO interface. Meaning Timing RIO_RD[0:7] data is captured on both the rising and falling edges of RIO_RCLK. RIO_RFRAME I Receive frame. Non return to zero (NRZ) input. It toggles at the beginning of a packet or control symbol. RIO_RFRAME is the differential pair of RIO_RFRAME. State Asserted/Negated—Assertion is signalled by a toggling of the signal. Toggling indicates a Meaning special packet framing event on the receive data (RIO_RD) pins. Timing RIO_RFRAME is sampled with respect to RIO_RCLK. RIO_TCLK O Transmit clock out. Free-running output clock which is launched in phase with the output data, RIO_TD, and the frame signal, RIO_TFRAME. RIO_TCLK should track the data bus RIO_TD[0:7] through all the associated routing (board, package, and so forth). RIO_TCLK is the differential pair of RIO_TCLK. See Section 17.1.4.1, “Transmit Clock-Select Mode,” for more information on the source of the transmit clock. State Asserted/Negated—A clock event in the transmit domain is determined as the rising or Meaning falling edge of this signal. Timing Assertion/Negation—RIO_TCLK is a free-running clock. Note that the RIO_TCLK is NOT phase aligned to the device clock. RIO_TD[0:7] O Transmit data. Carries packet information as well as control symbols. The data bus is switched on both edges of RIO_TCLK. RIO_TD[0:7] is the differential pair of RIO_TD[0:7]. State Asserted/Negated—Represents the data being transmitted on the RapidIO interface. Meaning Timing Assertion of RIO_TD is always performed with a fixed relationship to RIO_TCLK. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 17-8 Freescale Semiconductor Memory Map/Register Definition Table 17-5. RapidIO Interface Signals—Detailed Signal Descriptions (continued) Signal RIO_TFRAME I/O I Description Transmit frame. NRZ output. It toggles at the beginning of a packet or control symbol. RIO_TFRAME is the differential pair of RIO_TFRAME. State Asserted/Negated—Assertion is signalled by a toggling of the signal. Toggling indicates a Meaning special packet framing event on the transmit data (RIO_TD[0:7]) pins. Timing RIO_TFRAME is generated with respect to RIO_TCLK. RIO_TX_CLK_IN I Transmit clock in. Externally generated reference clock for the RapidIO transmit domain. RIO_TX_CLK_IN is the differential pair of RIO_TX_CLK_IN. 17.3 Memory Map/Register Definition The memory map for the RapidIO controller is shown in Table 17-6. Table 17-6. RapidIO Module Memory Map Offset Register Access Reset Section/Page RapidIO Architectural Registers 0xC_0000 DIDCAR—Device identity capability register 0xC_0004 DICAR—Device information capability register 0xC_0008 AIDCAR—Assembly identity capability register 0xC_000C AICAR—Assembly information capability register 0xC_0010 PEFCAR—Processing element features capability register 0xC_0014 SPICAR—Switch port information capability register 0xC_0018 SOCAR—Source operations capability register 0xC_001C DOCAR—Destination operations capability register 0xC_0040 MSR—Mailbox command and status register 0xC_0044 PWDCSR—Port-write and doorbell command and status register 0xC_004C PELLCCSR—Processing element logical layer control command and status register 0xC_0058 Reserved 0xC_005C LCSBA1CSR—Local configuration space base address 1 command and status register 0xC_0060 BDIDCSR—Base device ID command and status register 0xC_0068 HBDIDLCSR—Host base device ID lock command and status register 0xC_006C CTCSR—Component tag command and status register 0xC_0100 PMBH0CSR—8/16 LP-LVDS port maintenance block header 0 command and status register R R R/W R/W R R R R R R R R R/W R/W R/W R/W R 0x0002_0002 0x8030_0020 0x0000_0000 0x0000_0100 0xE088_0009 0x0000_0100 0x0600_FCF0 0x0600_FCF4 0x0000_0000 0x0000_0020 0x0000_0001 0x0000_0000 0x0000_0000 0x00nn_0000 0x0000_FFFF 0x0000_0000 0x0000_0002 17.3.1.1/17-13 17.3.1.2/17-14 17.3.1.3/17-14 17.3.1.4/17-15 17.3.1.5/17-15 17.3.1.6/17-17 17.3.1.7/17-17 17.3.1.8/17-19 17.3.1.9/17-22 17.3.1.10/17-22 17.3.1.11/17-24 — 17.3.1.12/17-24 17.3.1.13/17-25 17.3.1.14/17-26 17.3.1.15/17-26 17.3.1.16/17-27 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 17-9 RapidIO Interface Table 17-6. RapidIO Module Memory Map (continued) Offset Register Access R/W R/W R/W R/W R R/W R/W R/W Reset 0xFFFF_FF00 0xFFFF_FF00 0xn000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x4400_0000 Section/Page 17.3.1.17/17-27 17.3.1.18/17-28 17.3.1.19/17-29 17.3.1.20/17-29 17.3.1.21/17-30 17.3.1.22/17-31 17.3.1.23/17-31 17.3.1.24/17-33 0xC_0120 PLTOCCSR—Port link time-out control command and status register 0xC_0124 PRTOCCSR—Port response time-out control command and status register 0xC_013C PGCCSR—Port general control command and status register 0xC_0140 PLMREQCSR—Port link maintenance request command and status register 0xC_0144 PLMRESPCSR—Port link maintenance response command and status register 0xC_0148 PLASCSR—Port local ackID status command and status register 0xC_0158 PESCSR—Port error and status command and status register 0xC_015C PCCSR—Port control command and status register RapidIO Implementation Registers 0xD_0000 CR—Configuration register 0xD_0010 PCR—Port configuration register 0xD_0014 PEIR—Port error injection register RapidIO ATMU Registers 0xD_0C00 ROWTAR0—RapidIO outbound window translation address register 0 0xD_0C10 ROWAR0—RapidIO outbound window attributes register 0 0xD_0C20 ROWTAR1—RapidIO outbound window translation address register 1 0xD_0C28 ROWBAR1—RapidIO outbound window base address register 1 0xD_0C30 ROWAR1—RapidIO outbound window attributes register 1 0xD_0C40 ROWTAR2—RapidIO outbound window translation address register 2 0xD_0C48 ROWBAR2—RapidIO outbound window base address register 2 0xD_0C50 ROWAR2—RapidIO outbound window attributes register 2 0xD_0C60 ROWTAR3—RapidIO outbound window translation address register 3 0xD_0C68 ROWBAR3—RapidIO outbound window base address register 3 0xD_0C70 ROWAR3—RapidIO outbound window attributes register 3 0xD_0C80 ROWTAR4—RapidIO outbound window translation address register 4 0xD_0C88 ROWBAR4—RapidIO outbound window base address register 4 0xD_0C90 ROWAR4—RapidIO outbound window attributes register 4 R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W 0x0000_0000 0x8004_401F 0x0000_0000 0x0000_0000 0x0004_401F 0x0000_0000 0x0000_0000 0x0004_401F 0x0000_0000 0x0000_0000 0x0004_401F 0x0000_0000 0x0000_0000 0x0004_401F 17.3.2.2.1/17-37 17.3.2.2.3/17-39 17.3.2.2.1/17-37 17.3.2.2.2/17-38 17.3.2.2.3/17-39 17.3.2.2.1/17-37 17.3.2.2.2/17-38 17.3.2.2.3/17-39 17.3.2.2.1/17-37 17.3.2.2.2/17-38 17.3.2.2.3/17-39 17.3.2.2.1/17-37 17.3.2.2.2/17-38 17.3.2.2.3/17-39 R/W R/W R/W 0x0000_0001 0x0000_0010 0x0000_0000 17.3.2.1.1/17-34 17.3.2.1.2/17-35 17.3.2.1.3/17-35 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 17-10 Freescale Semiconductor Memory Map/Register Definition Table 17-6. RapidIO Module Memory Map (continued) Offset Register Access R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W Reset 0x0000_0000 0x0000_0000 0x0004_401F 0x0000_0000 0x0000_0000 0x0004_401F 0x0000_0000 0x0000_0000 0x0004_401F 0x0000_0000 0x0000_0000 0x0004_401F 0x0000_0000 0x0000_0000 0x0004_401F 0x0000_0000 0x0000_0000 0x0004_401F 0x0000_0000 0x0000_0000 0x0004_401F 0x0000_0000 0x0000_0000 0x0004_401F 0x0000_0000 0x8004_401F Section/Page 17.3.2.2.1/17-37 17.3.2.2.2/17-38 17.3.2.2.3/17-39 17.3.2.2.1/17-37 17.3.2.2.2/17-38 17.3.2.2.3/17-39 17.3.2.2.1/17-37 17.3.2.2.2/17-38 17.3.2.2.3/17-39 17.3.2.2.1/17-37 17.3.2.2.2/17-38 17.3.2.2.3/17-39 17.3.2.2.4/17-40 17.3.2.2.5/17-41 17.3.2.2.6/17-42 17.3.2.2.4/17-40 17.3.2.2.5/17-41 17.3.2.2.6/17-42 17.3.2.2.4/17-40 17.3.2.2.5/17-41 17.3.2.2.6/17-42 17.3.2.2.4/17-40 17.3.2.2.5/17-41 17.3.2.2.6/17-42 17.3.2.2.4/17-40 17.3.2.2.6/17-42 0xD_0CA0 ROWTAR5—RapidIO outbound window translation address register 5 0xD_0CA8 ROWBAR5—RapidIO outbound window base address register 5 0xD_0CB0 ROWAR5—RapidIO outbound window attributes register 5 0xD_0CC0 ROWTAR6—RapidIO outbound window translation address register 6 0xD_0CC8 ROWBAR6—RapidIO outbound window base address register 6 0xD_0CD0 ROWAR6—RapidIO outbound window attributes register 6 0xD_0CE0 ROWTAR7—RapidIO outbound window translation address register 7 0xD_0CE8 ROWBAR7—RapidIO outbound window base address register 7 0xD_0CF0 ROWAR7—RapidIO outbound window attributes register 7 0xD_0D00 ROWTAR8—RapidIO outbound window translation address register 8 0xD_0D08 ROWBAR8—RapidIO outbound window base address register 8 0xD_0D10 ROWAR8—RapidIO outbound window attributes register 8 0xD_0D60 RIWTAR4—RapidIO inbound window translation address register 4 0xD_0D68 RIWBAR4—RapidIO inbound window base address register 4 0xD_0D70 RIWAR4—RapidIO inbound window attributes register 4 0xD_0D80 RIWTAR3—RapidIO inbound window translation address register 3 0xD_0D88 RIWBAR3—RapidIO inbound window base address register 3 0xD_0D90 RIWAR3—RapidIO inbound window attributes register 3 0xD_0DA0 RIWTAR2—RapidIO inbound window translation address register 2 0xD_0DA8 RIWBAR2—RapidIO inbound window base address register 2 0xD_0DB0 RIWAR2—RapidIO inbound window attributes register 2 0xD_0DC0 RIWTAR1—RapidIO inbound window translation address register 1 0xD_0DC8 RIWBAR1—RapidIO inbound window base address register 1 0xD_0DD0 RIWAR1—RapidIO inbound window attributes register 1 0xD_0DE0 RIWTAR0—RapidIO inbound window translation address register 0 0xD_0DF0 RIWAR0—RapidIO inbound window attributes register 0 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 17-11 RapidIO Interface Table 17-6. RapidIO Module Memory Map (continued) Offset Register Access Reset Section/Page RapidIO Error Management Registers 0xD_0E00 PNFEDR—Port notification/fatal error detect register 0xD_0E04 PNFEDiR—Port notification/fatal error detect disable register 0xD_0E08 PNFEIER—Port notification/fatal error interrupt enable register 0xD_0E0C PECSR—Port error capture status register 0xD_0E10 EPCR0—Error packet capture register 0 0xD_0E14 EPCR1—Error packet capture register 1 0xD_0E18 EPCR2—Error packet capture register 2 0xD_0E20 PREDR—Port recoverable error detect register 0xD_0E28 PERTR—Port error recovery threshold register 0xD_0E2C PRTR—Port retry threshold register RapidIO Message Unit RapidIO Outbound Message Registers 0xD_1000 OMR—Outbound mode register 0xD_1004 OSR—Outbound status register 0xD_100C ODQDPAR—Outbound descriptor queue dequeue pointer address register 0xD_1014 OSAR—Outbound source address register 0xD_1018 ODPR—Outbound destination port register 0xD_101C ODATR—Outbound destination attributes register 0xD_1020 ODCR—Outbound double-word count register 0xD_1028 ODQEPAR—Outbound descriptor queue enqueue pointer address register R/W R/W R/W R/W R/W R/W R/W R/W 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0006_0000 0x0000_0000 0x0000_0000 17.3.3.1.1/17-65 17.3.3.1.2/17-67 17.3.3.1.3/17-68 17.3.3.1.4/17-69 17.3.3.1.5/17-70 17.3.3.1.6/17-70 17.3.3.1.7/17-71 17.3.3.1.8/17-72 R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x00FF_0000 0x00FF_0000 17.3.2.3.1/17-44 17.3.2.3.2/17-47 17.3.2.3.3/17-49 17.3.2.3.4/17-52 17.3.2.3.5/17-52 17.3.2.3.6/17-53 17.3.2.3.16/17-58 17.3.2.3.23/17-61 17.3.2.3.24/17-64 17.3.2.3.25/17-64 RapidIO Inbound Message Registers 0xD_1060 IMR—Inbound mailbox mode register 0xD_1064 ISR—Inbound mailbox status register 0xD_106C IFQDPAR—Inbound frame queue dequeue pointer address register 0xD_1074 IFQEPAR—Inbound frame queue enqueue pointer address register RapidIO Doorbell Registers 0xD_1460 DMR—Doorbell mode register 0xD_1464 DSR—Doorbell status register R/W R/W 0x0000_0000 0x0000_0000 17.3.3.3.1/17-77 17.3.3.3.2/17-78 R/W R/W R/W R/W 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 17.3.3.2.1/17-72 17.3.3.2.2/17-74 17.3.3.2.3/17-75 17.3.3.2.4/17-76 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 17-12 Freescale Semiconductor Memory Map/Register Definition Table 17-6. RapidIO Module Memory Map (continued) Offset Register Access R/W R/W Reset 0x0000_0000 0x0000_0000 Section/Page 17.3.3.3.3/17-79 17.3.3.3.4/17-80 0xD_146C DQDPAR—Doorbell queue dequeue pointer address register 0xD_1474 DQEPAR—Doorbell queue enqueue pointer address register RapidIO Port-Write Registers 0xD_14E0 PWMR—Port-write mode register 0xD_14E4 PWSR—Port-write status register 0xD_14EC PWQBAR—Port-write queue base address register R/W R/W R/W 0x0000_0000 0x0000_0000 0x0000_0000 17.3.3.4.1/17-81 17.3.3.4.2/17-82 17.3.3.4.3/17-83 17.3.1 Architectural Registers Following are the architectural registers of the RapidIO module. For additional details of these registers, refer to the Parallel RapidIO Interconnect Specification. 17.3.1.1 Device Identity Capability Register (DIDCAR) The DIDCAR is a read-only register that provides information regarding the type of device being used as well as the vendor who manufactured it. Figure 17-3 describes the DIDCAR. 0 15 16 31 R W Reset Offset DI DVI 0000_0000_0000_0010_0000_0000_0000_0010 0xC_0000 Figure 17-3. Device Identity Capability Register (DIDCAR) Table 17-7 describes the fields of the DIDCAR. Table 17-7. DIDCAR Field Descriptions Bits 0–15 Name DI Description Device identity field. Uniquely identifies the type of device from the vendor specified in the device vendor identity field. The values of the device identity field are assigned and managed by the vendor. 0x0002—MPC8540 Device vendor identity field. Identifies the vendor that manufactured the device. A value is uniquely assigned to a device vendor by the registration authority of the RapidIO Trade Association. 0x0002—Freescale 16–31 DVI MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 17-13 RapidIO Interface 17.3.1.2 Device Information Capability Register (DICAR) The DICAR, shown in Figure 17-4, identifies the revision level of the device. The value is assigned and managed by the vendor specified in DIDCAR[DVI]. DICAR represents a copy of the device’s system version register (SVR). 0 31 R W Reset Offset DR 1000_0000_0011_0000_0000_0000_0010_0000 0xC_0004 Figure 17-4. Device Information Capability Register (DICAR) Table 17-8 describes the field of the DICAR. Table 17-8. DICAR Field Description Bits 0–31 Name DR Description Device revision—Copy of the device’s system version register (SVR). Draco: 0x8030_0020 17.3.1.3 Assembly Identity Capability Register (AIDCAR) AIDCAR, shown in Figure 17-5, provides information regarding the type of assembly or subsystem being used as well as the vendor who manufactured it. 0 15 16 31 R W Reset Offset AI AVI 0000_0000_0000_0000_0000_0000_0000_0000 (Alternate value may be read from serial ROM) 0xC_0008 Figure 17-5. Assembly Identity Capability Register (AIDCAR) Table 17-9 describes the fields of the AIDCAR. Table 17-9. AIDCAR Field Descriptions Bits 0–15 Name AI Description Assembly identity field. The assembly field uniquely identifies the type of assembly or subsystem from the vendor specified in the assembly vendor identity field. The value of the assembly identity field by default is all zeros but may be assigned by a boot ROM device. Assembly vendor identity field. The assembly vendor identity field identifies the vendor that manufactured the assembly or subsystem. A value for the assembly vendor identity field is uniquely assigned to an assembly vendor by the registration authority of the RapidIO Trade Association. By default, this value is all zeros but may be assigned by a boot ROM device. 16–31 AVI MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 17-14 Freescale Semiconductor Memory Map/Register Definition 17.3.1.4 Assembly Information Capability Register (AICAR) The AICAR, shown in Figure 17-6, provides additional information about the assembly. 0 15 16 31 R W Reset Offset AR EFP 0000_0000_0000_0000_0000_0001_0000_0000 (Alternate value may be read from serial ROM) 0xC_000C Figure 17-6. Assembly Information Capability Register (AICAR) Table 17-10 describes the fields of the AICAR. Table 17-10. AICAR Field Descriptions Bits 0–15 16–31 Name AR EFP Description Assembly revision. This field states the assembly revision level. Extended features pointer. Pointer to the first entry in the extended features list. 17.3.1.5 Processing Element Features Capability Register (PEFCAR) The PEFCAR, shown in Figure 17-7, identifies major functionality provided by the RapidIO block. 0 1 2 3 4 7 8 9 10 11 12 13 26 27 28 29 31 RBMP S 0 W Reset Offset 0 0 0 M0 M1 M2 M3 D 0 0 0 0 0 0 0 0 0 0 0 0 0 0 CTLS EF EAS 1110_0000_1000_1000_0000_0000_0000_1001 0xC_0010 Figure 17-7. Processing Element Features Capability Register (PEFCAR) Table 17-11 describes the fields of the PEFCAR. Table 17-11. PEFCAR Field Descriptions Bits 0 Name B Description Bridge. Set for the MPC8540. 0 The RapidIO controller does not have a DRAM interface. 1 The RapidIO controller has a DRAM interface. Memory. Set for the MPC8540. 0 The RapidIO controller does not have local address space. 1 The RapidIO controller has local address space. 1 M MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 17-15 RapidIO Interface Table 17-11. PEFCAR Field Descriptions Bits 2 Name P Description Processor. Set for the MPC8540. 0 The RapidIO controller does not have a processor core. 1 The RapidIO controller has a processor core. Switch. Cleared for the MPC8540. 0 The RapidIO controller is a switch. 1 The RapidIO controller is not a switch. Reserved Mailbox 0. Set for the MPC8540. 0 The RapidIO controller does not support mailbox 0. 1 The RapidIO controller supports mailbox 0. Mailbox 1. Cleared for the MPC8540. 0 The RapidIO controller does not support mailbox 1. 1 The RapidIO controller supports mailbox 1. Mailbox 2. Cleared for the MPC8540. 0 The RapidIO controller does not support mailbox 2. 1 The RapidIO controller supports mailbox 2. Mailbox 3. Cleared for the MPC8540. 0 The RapidIO controller does not support mailbox 3. 1 The RapidIO controller supports mailbox 3. Doorbell. Set for the MPC8540. 0 The RapidIO controller does not support inbound doorbells. 1 The RapidIO controller supports inbound doorbells. Reserved 3 S 4–7 8 — M0 9 M1 10 M2 11 M3 12 D 13–26 27 — CTLS Common transport route. Cleared for the MPC8540. 0 The RapidIO controller does not support the large common transport route field. 1 The RapidIO controller supports the large common transport route field. EF Extended features. Set for the MPC8540. 0 The extended features pointer is not valid. 1 The extended features pointer is valid. Extended addressing support. Possesses a value of 001 for the MPC8540. 001 The RapidIO controller supports 34-bit local addresses. 010 Reserved 011 The RapidIO controller supports 50- and 34-bit local addresses. 100 Reserved 101 The RapidIO controller supports 66- and 34-bit local addresses. 110 Reserved 111 The RapidIO controller supports 60-, 50- and 34-bit local addresses. 28 29–31 EAS MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 17-16 Freescale Semiconductor Memory Map/Register Definition 17.3.1.6 Switch Port Information Capability Register (SPICAR) The SPICAR, shown in Figure 17-8, defines the RapidIO block’s switching capabilities. 0 15 16 23 24 31 R0 0 W Reset Offset 00 00000000000 0 PT PN 0000_0000_0000_0000_0000_0001_0000_0000 0xC_0014 Figure 17-8. Switch Port Information Capability Register (SPICAR) Table 17-12 describes the fields of the SPICAR. Table 17-12. SPICAR Field Descriptions Bits 0–15 16–23 24–31 Name — PT PN Reserved Port total. The RapidIO controller has one RapidIO port. (SPICAR[PT] = 0x01) Port number. This is the port number from which this register was read. This field is 0 since the MPC8540 has only one port. Description 17.3.1.7 Source Operations Capability Register (SOCAR) The SOCAR, shown in Figure 17-9, reflects the set of RapidIO operations that this device can initiate. See the Parallel RapidIO Interconnect Specification for details. 0 1 2 3 4 5 6 7 8 9 10 15 R W Reset 16 IR RO DCI C F IOR IKI TIE TIES 0 0 0 0 0 0 0000_0110_0000_0000 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 R W Reset Offset NR NW SW NWR M D 0 ATS AI AD AS AC 0 PW 0 0 1111_1100_1111_0000 0xC_0018 Figure 17-9. Source Operations Capability Register (SOCAR) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 17-17 RapidIO Interface Table 17-13 describes the fields of the SOCAR. Table 17-13. SOCAR Field Descriptions Bits 0 Name R Description Read operation. Cleared for the MPC8540. 0 The RapidIO controller cannot initiate a read operation. 1 The RapidIO controller can initiate a read operation. Iread operation. Cleared for the MPC8540. 0 The RapidIO controller cannot initiate an iread operation. 1 The RapidIO controller can initiate an iread operation. Read to own operation. Cleared for the MPC8540. 0 The RapidIO controller cannot initiate a read to own operation. 1 The RapidIO controller can initiate a read to own operation. Dkill operation. Cleared for the MPC8540. 0 The RapidIO controller cannot initiate a dkill operation. 1 The RapidIO controller can initiate a dkill operation. Castout operation. Cleared for the MPC8540. 0 The RapidIO controller cannot initiate a castout operation. 1 The RapidIO controller can initiate a castout operation. Flush operation. (Return globally-shared cache line to memory.) Set for the MPC8540. 0 The RapidIO controller cannot initiate a flush operation. 1 The RapidIO controller can initiate a flush operation. I/O read operation. (Read a non-cachable copy of a globally-shared cache line.) Set for the MPC8540. 0 The RapidIO controller cannot initiate an I/O read operation. 1 The RapidIO controller can initiate an I/O read operation. Ikill operation. Cleared for the MPC8540. 0 The RapidIO controller cannot initiate an ikill operation. 1 The RapidIO controller can initiate an ikill operation. TLBIE operation. Cleared for the MPC8540. 0 The RapidIO controller cannot initiate a TLBIE operation. 1 The RapidIO controller can initiate a TLBIE operation. 1 IR 2 RO 3 DCI 4 C 5 F 6 IOR 7 IKI 8 TIE 9 TIES TLBSYNC operation. Cleared for the MPC8540. 0 The RapidIO controller cannot initiate a TLBSYNC operation. 1 The RapidIO controller can initiate a TLBSYNC operation. — NR Reserved Nread operation. (Read non-sharable memory.) Set for the MPC8540. 0 The RapidIO controller cannot initiate an nread operation. 1 The RapidIO controller can initiate an nread operation. Nwrite operation. (Write non-sharable memory.) Set for the MPC8540. 0 The RapidIO controller cannot initiate an nwrite operation. 1 The RapidIO controller can initiate an nwrite operation. Swrite operation. (Write non-sharable memory.) Set for the MPC8540. 0 The RapidIO controller cannot initiate an swrite operation. 1 The RapidIO controller can initiate an swrite operation. 10–15 16 17 NW 18 SW MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 17-18 Freescale Semiconductor Memory Map/Register Definition Table 17-13. SOCAR Field Descriptions (continued) Bits 19 Name Description NWR Nwrite_r operation. (Write non-sharable memory.) Set for the MPC8540. 0 The RapidIO controller cannot initiate an nwrite_r operation. 1 The RapidIO controller can initiate an nwrite_r operation. M Message operation. (Write to port.) Set for the MPC8540. 0 The RapidIO controller cannot initiate a message operation. 1 The RapidIO controller can initiate a message operation. Doorbell operation. (Generate and interrupt.) Set for the MPC8540. 0 The RapidIO controller cannot initiate a doorbell operation. 1 The RapidIO controller can initiate a doorbell operation. Reserved Atomic test and swap operation. Cleared for the MPC8540. 0 The RapidIO controller can initiate an atomic test and swap operation. 1 The RapidIO controller cannot initiate an atomic test and swap operation. Atomic_inc operation. Set for the MPC8540. 0 The RapidIO controller cannot initiate an atomic_inc operation. 1 The RapidIO controller can initiate an atomic_inc operation. Atomic_dec operation. Set for the MPC8540. 0 The RapidIO controller cannot initiate an atomic_dec operation. 1 The RapidIO controller can initiate an atomic_dec operation. Atomic_set operation. Set for the MPC8540. 0 The RapidIO controller cannot initiate an atomic_set operation. 1 The RapidIO controller can initiate an atomic_set operation. Atomic_clr operation. Set for the MPC8540. 0 The RapidIO controller cannot initiate an atomic_clr operation. 1 The RapidIO controller can initiate an atomic_clr operation. Reserved Port-write operation. Cleared for the MPC8540. 0 The RapidIO controller cannot initiate a port-write operation. 1 The RapidIO controller can initiate a port-write operation. Reserved 20 21 D 22 23 — ATS 24 AI 25 AD 26 AS 27 AC 28 29 — PW 30–31 — 17.3.1.8 Destination Operations Capability Register (DOCAR) The DOCAR, shown in Figure 17-10, describes RapidIO I/O operations that this device can service. See the Parallel RapidIO Interconnect Specification for additional details. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 17-19 RapidIO Interface 0 1 2 3 4 5 6 7 8 9 10 15 R W Reset 16 IR RO DCI C F IOR IKI TIE TIES 0 0 0 0 0 0 0000_0110_0000_0000 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 NR W Reset Offset NW SW NWR M D 0 ATS AI AD AS AC 0 PW 0 0 1111_1100_1111_0100 0xC_001C Figure 17-10. Destination Operations Capability Register (DOCAR) Table 17-14 describes the fields of the DOCAR. Table 17-14. DOCAR Field Descriptions Bits 0 Name R Description Read operation. Cleared for the MPC8540. 0 The RapidIO controller cannot service a read operation. 1 The RapidIO controller can service a read operation. Iread operation. Cleared for the MPC8540. 0 The RapidIO controller cannot service an iread operation. 1 The RapidIO controller can service an iread operation. Read to own operation. Cleared for the MPC8540. 0 The RapidIO controller cannot service a read to own operation. 1 The RapidIO controller can service a read to own operation. Dkill operation. Cleared for the MPC8540. 0 The RapidIO controller cannot service a dkill operation. 1 The RapidIO controller can service a dkill operation. Castout operation. Cleared for the MPC8540. 0 The RapidIO controller cannot service a castout operation. 1 The RapidIO controller can service a castout operation. Flush operation. (Return globally-shared cache line to memory.) Set for the MPC8540. 0 The RapidIO controller cannot service a flush operation. 1 The RapidIO controller can service a flush operation. I/O read operation. (Read a non-cachable copy of a globally-shared cache line.) Set for the MPC8540. 0 The RapidIO controller cannot service an I/O read operation. 1 The RapidIO controller can service an I/O read operation. Ikill operation. Cleared for the MPC8540. 0 The RapidIO controller cannot service an ikill operation. 1 The RapidIO controller can service an ikill operation. TLBIE operation. Cleared for the MPC8540. 0 The RapidIO controller cannot service a TLBIE operation. 1 The RapidIO controller can service a TLBIE operation. 1 IR 2 RO 3 DCI 4 C 5 F 6 IOR 7 IKI 8 TIE MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 17-20 Freescale Semiconductor Memory Map/Register Definition Table 17-14. DOCAR Field Descriptions (continued) Bits 9 Name Description TIES TLBSYNC operation. Cleared for the MPC8540. 0 The RapidIO controller cannot service a TLBSYNC operation. 1 The RapidIO controller can service a TLBSYNC operation. — NR Reserved Nread operation. (Read non-sharable memory.) Set for the MPC8540. 0 The RapidIO controller cannot service an nread operation. 1 The RapidIO controller can service an nread operation. Nwrite operation. (Write non-sharable memory.) Set for the MPC8540. 0 The RapidIO controller cannot service an nwrite operation. 1 The RapidIO controller can service an nwrite operation. Swrite operation. (Write non-sharable memory.) Set for the MPC8540. 0 The RapidIO controller cannot service an swrite operation. 1 The RapidIO controller can service an swrite operation. 10–15 16 17 NW 18 SW 19 NWR Nwrite_r operation. (Write non-sharable memory.) Set for the MPC8540. 0 The RapidIO controller cannot service an nwrite_r operation. 1 The RapidIO controller can service an nwrite_r operation. M Message operation. (Write to port.) Set for the MPC8540. 0 The RapidIO controller cannot service a message operation. 1 The RapidIO controller can service a message operation. Doorbell operation. (Generate and interrupt.) Set for the MPC8540. 0 The RapidIO controller cannot service a doorbell operation. 1 The RapidIO controller can service a doorbell operation. Reserved Atomic test and swap operation. Cleared for the MPC8540. 0 The RapidIO controller can service an atomic test and swap operation. 1 The RapidIO controller cannot service an atomic test and swap operation. Atomic_inc operation. Set for the MPC8540. 0 The RapidIO controller cannot service an atomic_inc operation. 1 The RapidIO controller can service an atomic_inc operation. Atomic_dec operation. Set for the MPC8540. 0 The RapidIO controller cannot service an atomic_dec operation. 1 The RapidIO controller can service an atomic_dec operation. Atomic_set operation. Set for the MPC8540. 0 The RapidIO controller cannot service an atomic_set operation. 1 The RapidIO controller can service an atomic_set operation. Atomic_clr operation. Set for the MPC8540. 0 The RapidIO controller cannot service an atomic_clr operation. 1 The RapidIO controller can service an atomic_clr operation. Reserved Port-write operation. Set for the MPC8540. 0 The RapidIO controller cannot service a port-write operation. 1 The RapidIO controller can service a port-write operation. Reserved 20 21 D 22 23 — ATS 24 AI 25 AD 26 AS 27 AC 28 29 — PW 30–31 — MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 17-21 RapidIO Interface 17.3.1.9 Mailbox Command and Status Register (MSR) The MSR, shown in Figure 17-11, reflects the status of the RapidIO mailbox controller on this device. Note that the RapidIO controller only supports Mailbox 0. 0 1 2 34 5 6 31 R A FU EM B 0 ERR 0 0 0 0 0 0 0 0 0 0 0 W Reset Offset 000 000 000 000 000 0000_0000_0000_0000_0000_0000_0000_0000 0xC_0040 Figure 17-11. Mailbox Status Register (MSR) Table 17-15 describes the fields of the MSR. Table 17-15. MSR Field Definitions Bits Name 0 A Description Available 0 Mailbox 0 is not ready to accept messages. All incoming message transactions return error responses. 1 Mailbox 0 is initialized and ready to accept messages. Full 0 Mailbox 0 is not full. 1 Mailbox 0 is full. New messages are retried. Empty 0 Mailbox 0 contains outstanding messages. 1 Mailbox 0 contains no outstanding messages. Busy 0 Mailbox 0 is not busy processing a message. 1 Mailbox 0 is busy processing a message. New message operations return retry responses. Reserved Error 0 Mailbox 0 has not received an illegal message operation. 1 Mailbox 0 received an illegal message operation. All incoming message transactions return error responses. Reserved 1 FU 2 EM 3 B 4 5 — ERR 6–31 — 17.3.1.10 Port-Write and Doorbell Command and Status Register (PWDCSR) The PWDCSR, shown in Figure 17-12, reflects the status of the RapidIO doorbell and port-write hardware on this device. Additional details can be found in the Parallel RapidIO Interconnect Specification in the sections titled “Doorbell CSR” and “Write Port CSR.” MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 17-22 Freescale Semiconductor Memory Map/Register Definition 0 1 2 3 4 5 6 23 24 25 26 27 28 29 30 31 R A FU EM B FA ERR 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 W Reset Offset PA PFU PEM PB 0 PE 0 0 0000_0000_0000_0000_0000_0000_0010_0000 0xC_0044 Figure 17-12. Port-Write and Doorbell Command and Status Register (PWDCSR) Table 17-16 describes the fields of the PWDCSR. : Table 17-16. PWDCSR Field Descriptions Bits 0 Name A Description Available 0 Doorbell unit is not ready to accept doorbell messages. All incoming doorbell transactions return error responses. 1 Doorbell unit is ready to accept doorbell messages Full 0 Doorbell unit is not full 1 Doorbell unit is full. New doorbell messages are retried. Empty 0 Doorbell unit has outstanding doorbell messages 1 Doorbell unit has no outstanding doorbell messages Busy 0 Doorbell unit is not busy processing a doorbell message 1 Doorbell unit is busy processing a doorbell message. Incoming transactions are retried. Failure. (Defined by the RapidIO specification but not implemented in the MPC8540.) 0 Doorbell unit has not encountered an internal error 1 Doorbell unit has encountered an internal error Error 0 Doorbell unit has not encountered an error condition 1 Doorbell unit encountered an error condition. Incoming doorbell transactions return error responses. Reserved Port-write unit available 0 Port-write unit is not available. All incoming port-write packets are discarded. 1 Port-write unit is initialized and ready to accept a port-write packet Port-write unit full 0 Port-write unit is not full 1 Port-write unit is full. All incoming port-write packets are discarded. Port-write unit empty 0 Port-write unit contains outstanding port-write data 1 Port-write unit contains no outstanding port-write data Port-write unit busy 0 Port-write unit is not busy 1 Port-write unit is busy processing a port-write packet. Incoming port-write packets are discarded. Reserved 1 FU 2 EM 3 B 4 FA 5 ERR 6–23 24 — PA 25 PFU 26 PEM 27 PB 28 — MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 17-23 RapidIO Interface Table 17-16. PWDCSR Field Descriptions (continued) Bits 29 Name PE Description Port-write unit error 0 Received port-write packet contains no errors, for example, data is of legal size (64 bytes or less) 1 Received port-write packet contains fatal errors, for example, data is of illegal size (greater than 64 bytes). All incoming port-write transactions are discarded. Reserved 30–31 — 17.3.1.11 Processing Element Logic Layer Control Command and Status Register (PELLCCSR) The PELLCCSR, shown in Figure 17-13, is a read-only register. Additional details of this register can be found in the Parallel RapidIO Interconnect Specification. 0 28 29 31 R0 0 0 0 0 0 0 0 0 0 0 0 W Reset Offset 000000000000 0000 0 EAC 0000_0000_0000_0000_0000_0000_0000_0001 0xC_004C Figure 17-13. Processing Element Logic Layer Control Command and Status Register (PELLCCSR) Table 17-17 describes the fields of the PELLCCSR. Table 17-17. PELLCCSR Field Descriptions Bits 0–28 29–31 Name — EAC Reserved Extended addressing control. The RapidIO controller supports 34-bit addresses (read-only). Description 17.3.1.12 Local Configuration Space Base Address 1 Command and Status Register (LCSBA1CSR) The local configuration space base address 1 command and status register (LCSBA1CSR), shown in Figure 17-14, contains the 14 msbs of the double-word address, which represents the base address for the device’s register space. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 17-24 Freescale Semiconductor Memory Map/Register Definition 0 1 14 15 31 R0 LCSBA W Reset Offset 00000000000000000 0000_0000_0000_0000_0000_0000_0000_0000 0xC_005C Figure 17-14. Local Configuration Space Base Address 1 Command and Status Register (LCSBA1CSR) Table 17-18 describes the fields of the LCSBA1CSR. Table 17-18. LCSBA1CSR Field Descriptions Bits 0 1–14 15–31 Name — Reserved Description LCSBA Local configuration space base address. These bits correspond to the 14 most-significant bits of the 34-bit RapidIO address space. — Reserved 17.3.1.13 Base Device ID Command and Status Register (BDIDCSR) Figure 17-15 shows the BDIDSCR. 0 7 8 15 16 31 R0 W Reset Offset 000 0 0 00 BDID 0 0 00 0 00 00 0 00 0 00 0 0000_0000_(Value from configuration pins)_0000_0000_0000_0000 0xC_0060 Figure 17-15. Base Device ID Command and Status Register (BDIDCSR) Table 17-19 describes the fields of the BDIDCSR. Table 17-19. BDIDCSR Field Descriptions Bits 0–7 8–15 Name — BDID Reserved Base device ID (RapidIO device ID). Value read from reset configuration pins. For more information on the RapidIO device ID, see Section 4.4.3.11, “RapidIO Device ID.” Reserved Description 16–31 — MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 17-25 RapidIO Interface 17.3.1.14 Host Base Device ID Lock Command and Status Register (HBDIDLCSR) The host base device ID lock CSR, shown in Figure 17-16, contains the base device ID value for the processing element in the system that is responsible for initializing this processing element. HBDID is a write-once/resettable field that provides a lock function. After HBDID is written, subsequent writes to the field are ignored, except when the written value matches the value contained in the field. In this case, the register is reinitialized to 0xFFFF. After HBDID is written, a processing element must then read the host base device ID lock CSR to verify that it owns the lock before attempting to initialize this processing element. 0 15 16 31 R0 0 0 0 W Reset Offset 0000000 0000 0 HBDID 0000_0000_0000_0000_1111_1111_1111_1111 0xC_0068 Figure 17-16. Host Base Device ID Lock Command and Status Register (HBDIDLCSR) Table 17-20 describes the fields of the HBDIDLCSR. Table 17-20. HBDIDLCSR Field Descriptions Bits 0–15 Name — Reserved Description 16–31 HBDID Host base device ID. This is the base device ID for the processing element that is initializing this processing element. When unlocked, this field is write-able and locks immediately after being written. When locked, this field is only write-able when written with the value it already contains, at which point it becomes unlocked. 17.3.1.15 Component Tag Command and Status Register (CTCSR) The CTCSR contains a component tag value for the RapidIO block and can be assigned by software when the device is initialized. Figure 17-17 shows the CTCSR. 0 31 R CT W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0xC_006C Figure 17-17. Component Tag Command and Status Register (CTCSR) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 17-26 Freescale Semiconductor Memory Map/Register Definition Table 17-21 describes the field of the CTCSR. Table 17-21. CTCSR Field Descriptions Bits 0–31 Name CT Description Component tag. Identifying component tag for the RapidIO block. 17.3.1.16 8/16 LP-LVDS Port Maintenance Block Header 0 Command and Status Register (PMBH0CSR) The port maintenance block header 0 command and status register, shown in Figure 17-18, contains the extended features pointer to the next extended features block and the extended features ID that identifies this as the generic end point port maintenance block header. 0 15 16 31 R W Reset Offset EF_PTR EF_ID 0000_0000_0000_0000_0000_0000_0000_0010 0xC_0100 Figure 17-18. 8/16 LP-LVDS Port Maintenance Block Header 0 Command and Status Register (PMBH0CSR) Table 17-22 describes the fields of the PMBH0CSR. Table 17-22. PMBH0CSR Field Descriptions Bits 0–15 16–31 Name Description EF_PTR Extended features pointer. Points to the next extended features block. 0x0000 for the MPC8540. EF_ID Extended features ID. Identifies this as the generic end point port maintenance block header. 0x0002 for the MPC8540. 17.3.1.17 Port Link Time-Out Control Command and Status Register (PLTOCCSR) The PLTOCCSR, shown in Figure 17-19, contains the time-out timer value used to monitor link events, for example, between sending a packet and receiving the corresponding acknowledge. Alternately, it can contain the value of the timer time-out between sending a link-request and receiving the corresponding link-response. The reset value is the maximum time-out interval and represents between three and five seconds. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 17-27 RapidIO Interface 0 23 24 31 R TV W Reset Offset 1111_1111_1111_1111_1111_1111_0000_0000 0xC_0120 0 0 0 0 0 0 0 0 Figure 17-19. Port Link Time-Out Control Command and Status Register (PLTOCCSR) Table 17-23 describes the fields of the PLTOCCSR. Table 17-23. PLTOCCSR Field Descriptions Bits 0–23 24–31 Name TV — Description Time-out value. This value is expressed in platform clock cycles. Reserved 17.3.1.18 Port Response Time-Out Control Command and Status Register (PRTOCCSR) PRTOCCSR, shown in Figure 17-20, contains the time-out timer value used to monitor port events, for instance, between sending a request packet and receiving the corresponding response packet. 0 23 24 31 R TV W Reset Offset 1111_1111_1111_1111_1111_1111_0000_0000 0xC_0124 00000000 Figure 17-20. Port Response Time-Out Control Command and Status Register (PRTOCCSR) Table 17-24 describes the fields of the PRTOCCSR. Table 17-24. PRTOCCSR Field Descriptions Bits 0–23 24–31 Name TV — Description Time-out value. This value is expressed in platform clock cycles. Reserved MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 17-28 Freescale Semiconductor Memory Map/Register Definition 17.3.1.19 Port General Control Command and Status Register (PGCCSR) The PGCCSR, shown in Figure 17-21, reflects the state of the RapidIO port. 0 1 2 3 31 R HMD W Reset Offset 00 0 0 000 0 0 00 00 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 (Value from configuration pins)_0000_0000_0000_0000_0000_0000_0000 0xC_013C Figure 17-21. Port General Control Command and Status Register (PGCCSR) Table 17-25 describes the fields of the PGCCSR. Table 17-25. PGCCSR Field Descriptions Bits 0 Name H Description A host device is a device that is responsible for system exploration, initialization, and maintenance. Agent (or slave) devices are initialized by host devices. Initialized from the host/agent configuration pin; If host is sampled, H is set as host. 0 Agent device 1 Host device Master enable bit. Controls whether or not a device is allowed to issue requests into the system. If the master enable is not set, the device may only respond to requests. Initialized from the host/agent configuration pin. Out of reset, if configured as a host (PGCCSR[H] is set), M and D are also set. If configured as an agent (PGCCSR[H] is cleared), M and D are initialized by the host of the system. 0 Device is not enabled to issue requests into the system 1 Device is enabled to issue requests into the system Discovery enable bit. This device has been located by the processing element responsible for system configuration. Initialized from the host/agent configuration pin. Out of reset, if configured as a host (PGCCSR[H] is set), bits M and D are also set. If configured as an agent (PGCCSR[H] is cleared), bits M and D are initialized by the host of the system. 0 Device has not been discovered by system host 1 Device has been discovered by system host Reserved 1 M 2 D 3–31 — 17.3.1.20 Port Link Maintenance Request Command and Status Register (PLMREQCSR) A write to the PLMREQCSR, shown in Figure 17-22, generates a link-request control symbol on the RapidIO port interface. See the Parallel RapidIO Interconnect Specification for descriptions of the reset command, input-status command and send-training commands. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 17-29 RapidIO Interface 0 28 29 31 R0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 C W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0xC_0140 Figure 17-22. Port Link Maintenance Request Command and Status Register (PLMREQCSR) Table 17-26 describes the fields of the PLMREQCSR. Table 17-26. PLMREQCSR Field Descriptions Bits 0–28 29–31 Name — C Reserved Command. This is the command to be sent in the link-request control symbol. See Table 17-85 for command symbol formats. Description 17.3.1.21 Port Link Maintenance Response Command and Status Register (PLMRESPCSR) A read from the PLMRESPCSR, shown in Figure 17-23, returns the status received in a link-response control symbol. 0 1 24 25 27 28 31 R RV 0 0 W Reset Offset 000000000000 0 00 0 0 0 0 0 0 0 AS LS 0000_0000_0000_0000_0000_0000_0000_0000 0xC_0144 Figure 17-23. Port Link Maintenance Response Command and Status Register (PLMRESPCSR) Table 17-27 describes the fields of the PLMRESPCSR. Table 17-27. PLMRESPCSR Field Descriptions Bits 0 Name RV Description Response valid (clear on read) 0 If the link-request caused a link-response, the link-response has not yet been received. If the link-request did not cause a link-response, the link-request has not yet been transmitted. 1 If the link-request caused a link-response, the link-response has been received and the status fields are valid. If the link-request did not cause a link-response, the link-request has been transmitted. Reserved 1–24 — MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 17-30 Freescale Semiconductor Memory Map/Register Definition Table 17-27. PLMRESPCSR Field Descriptions (continued) Bits 25–27 28–31 Name AS LS Description AckID status. This is the ackID status field from the link-response control symbol. Link status. This is the link status field from the link-response control symbol. 17.3.1.22 Port Local AckID Status Command and Status Register (PLASCSR) A read from the PLASCSR returns the local ackID status for both the outbound and inbound RapidIO ports. Figure 17-24 shows the PLASCSR. 0 4 5 7 8 15 16 23 24 28 29 31 R0 W Reset Offset 0 0 0 0 IA 00 00 0 000 OUTA 0 0 0 0 0 OBA 0000_0000_0000_0000_0000_0000_0000_0000 0xC_0148 Figure 17-24. Port Local AckID Status Command and Status Register (PLASCSR) Table 17-28 describes the fields of the PLASCSR. Table 17-28. PLASCSR Field Descriptions Bits 0–4 5–7 8–15 Name — IA — Reserved Input port next expected ackID value Reserved Description 16–23 OUTA Outstanding unacknowledged ackIDs. A set bit indicates that the corresponding ackID value is used in a packet to an attached device but a corresponding acknowledge control symbol has not been received (read-only field). 24–28 29–31 — OBA Reserved Outbound ackID. Output port next transmitted ackID value. Software writing this value can force re-transmission of outstanding unacknowledged packets ti manually implement error recovery. OBA can only be written by software when the RapidIO port is enabled. 17.3.1.23 Port Error and Status Command and Status Register (PESCSR) The PESCSR contains port error and status information. Figure 17-25 shows the PESCSR. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 17-31 RapidIO Interface 0 10 11 12 13 14 15 16 20 21 22 23 24 27 28 29 30 31 R0000000000 0 ORE W Reset Offset OR ORS OEE OES 0 0 0 0 0 IRS IEE IES 0 0 0 0 PP PE PO PU 0000_0000_0000_0000_0000_0000_0000_0000 0xC_0158 Figure 17-25. Port Error and Status Command and Status Register (PESCSR) Table 17-29 describes the fields of the PESCSR. Table 17-29. PESCSR Field Descriptions Bits 0–10 11 Name — ORE Reserved Output port retry condition 0 Output port operating normally 1 Output port encountered a retry condition. Set when ORS is set. (Bit reset, write-one-to-clear) Output port retry (read-only field). (Bit reset, write-one-to-clear) 0 Output port received a packet-accepted or packet-not-accepted control symbol 1 Output port received a packet retry and cannot make forward progress. Set when ORS is set Output port retry stop (read-only field) 0 Output port operating normally 1 Output port is stopped due to a retry Output port error encounter. (Bit reset, write-one-to-clear) 0 Output port operating normally 1 Output port encountered an error. Set when OES is set Output port error stop (read-only field) 0 Output port operating normally 1 Output port is stopped due to a transmission error Reserved Input port retry stop (read-only field) 0 Input port operating normally 1 Input port is stopped due to a retry condition Input port error encounter. (Bit reset, write-one-to-clear) 0 Input port operating normally 1 Input port encountered a transmission error. Set when IES is set Input port error stop (read-only field) 0 Input port operating normally 1 Input port is stopped due to a transmission error Reserved Input clock toggling (read-only field) Note: Set shortly after reset, due to asynchronous interface crossing and clock detection delay 0 Input clock pin is not toggling 1 Input clock pin is toggling. (There is a device attached.) Description 12 OR 13 ORS 14 OEE 15 OES 16–20 21 — IRS 22 IEE 23 IES 24–27 28 — PP MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 17-32 Freescale Semiconductor Memory Map/Register Definition Table 17-29. PESCSR Field Descriptions (continued) Bits 29 Name PE Description Input port error. (Bit reset, write-one-to-clear) 0 Input port operating properly 1 Input port encountered an unrecoverable error or output port failed error recover Port initialization (read-only) 0 Input or output port not initialized 1 Input and output ports are initialized and can communicate with the adjacent device. Port training mode (read-only) Note: Set shortly after reset due to asynchronous interface crossing delay 0 Input and output ports initialized 1 Input and output ports are not initialized and are in training mode 30 PO 31 PU 17.3.1.24 Port Control Command and Status Register (PCCSR) The PCCSR contains bits used for port control. Figure 17-26 shows the PCCSR. 0 1 2 3 4 5 6 7 8 9 31 R OPW OPE OPD W Reset Offset 0 IPW IPE IPD 0 ECD 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0100_0100_0000_0000_0000_0000_0000_0000 0xC_015C Figure 17-26. Port Control Command and Status Register (PCCSR) Table 17-30 describes the fields of the PCCSR. Table 17-30. PCCSR Field Descriptions Bits 0 Name Description OPW Operating width of the output port (read-only field) 0 Port operating in 8-bit mode 1 Reserved OPE Output port transmit enable. Must equal the value of PCCSR[IPE] for the RapidIO controller to function properly. 0 Port is stopped and not enabled to issue any packets except to respond to maintenance packets 1 Port is enabled to issue any packets Output port drivers disable 0 Input receiver and/or output driver are enabled 1 Both the input receivers and output drivers are disabled. Reserved Operating width of the input port (read-only field) 0 Port operating in 8-bit mode 1 Reserved 1 2 OPD 3 4 — IPW MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 17-33 RapidIO Interface Table 17-30. PCCSR Field Descriptions (continued) Bits 5 Name IPE Description Input port receive enable. Must equal the value of PCCSR[OPE] for the RapidIO controller to function properly 0 Port is stopped and only enabled to receive maintenance requests packets. Other requests return packet-not-accepted control symbols with ‘General error’ cause to force an error condition to be signalled by the sending device 1 Port is enabled to issue any packets Input port receivers disable 0 Input receiver and/or output driver are enabled 1 Both the input receivers and output drivers are disabled. Reserved Error checking disable. Disables all RapidIO transmission error checking 0 Error checking and recovery is enabled. 1 Error checking and recovery is disabled. Device behavior while error checking is disabled and an error condition occurs is undefined. Reserved 6 IPD 7 8 — ECD 9–31 — 17.3.2 Implementation Registers Following are the RapidIO implementation registers. These are registers specific to this particular RapidIO implementation and, therefore, do not appear in the Parallel RapidIO Interconnect Specification. 17.3.2.1 General Registers 17.3.2.1.1 Configuration Register (CR) The CR contains configuration information regarding the behavior of the RapidIO controller. Figure 17-27 shows the CR. 0 30 31 R0 W Reset Offset 0 0 0 00 0 0 00 0 0 000 0 00 0 0 0 0 0 0 0 0 0 0 0 0 0 AA 0000_0000_0000_0000_0000_0000_0000_0001 0xD_0000 Figure 17-27. Configuration Register (CR) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 17-34 Freescale Semiconductor Memory Map/Register Definition Table 17-31 describes the fields of the CR. Table 17-31. CR Field Descriptions Bits 0–30 31 Name — AA Reserved Accept all. Note: Must initially be cleared by the user as part of initialization 0 Detect an error upon receiving a packet with a target ID not equal to its own 1 All packets are accepted without checking the target ID. Description 17.3.2.1.2 Port Configuration Register (PCR) The PCR contains configuration information regarding port behavior of the RapidIO controller. Figure 17-28 shows the PCR. 0 26 27 28 31 R0 0 0 0 0 0 0 0 0 00 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 W Reset Offset 0000_0000_0000_0000_0000_0000_0001_0000 0xD_0010 0 CCE 0 00 0 Figure 17-28. Port Configuration Register (PCR) Table 17-32 describes the fields of the PCR. Table 17-32. PCR Field Descriptions Bits 0–26 27 Name — CCE Reserved CRC checking enable. Note that a read that attempts to access an unmapped target causes the assertion of core_fault_in, which causes the core to generate a machine check interrupt, unless it is disabled (by clearing HID1[RFXE]). If RFXE is zero and this error occurs, PCR[CCE] must be set to ensure that an interrupt is generated. For more information, see Section 6.10.2, “Hardware Implementation-Dependent Register 1 (HID1).” 0 No CRC is checked on received packets. 1 CRC is checked on received packets. Reserved Description 28–31 — 17.3.2.1.3 Port Error Injection Register (PEIR) RapidIO error injection registers, shown in Figure 17-29, offer a simple way to inject errors into an outbound packet or a control symbol. The error injection register for the target port is loaded with control information for injecting an error in a packet stream or control symbol. After the register is written, the hardware starts to count up to the programmed number of words and, upon reaching that number, an error is injected in the specified byte using the byte mask. The error is MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 17-35 RapidIO Interface injected once and the enable bit is cleared if error injection is enabled and SE is not set. The error is injected continuously if both EN and SE are set. It stops if software clears the enable bit. After setting the enable bit, software can monitor the injection of the error by polling the register and observing the enable bit being cleared. 0 7 8 15 16 23 24 25 26 27 28 29 30 31 R BM W Reset Offset 00000000 WC 0000_0000_0000_0000_0000_0000_0000_0000 0xD_0014 BC 00 EM SE SS EN Figure 17-29. Port Error Injection Register (PEIR) Table 17-33 describes the fields of the PEIR. Table 17-33. PEIR Field Descriptions Bits 0–7 8–15 16–23 24–25 26–27 28 Name BM — WC BC — EM Description Byte mask. Indicates on which bit of the byte to inject error Reserved Word count. Number of words to count before injecting error Byte count. Byte within the specified word count on which to inject error Reserved Error mirror. Setting EM when BC is 00 causes the error to be mirrored to 10 and when BC is 01, the error is mirrored to 11. This gives better flexibility for error injection. For example, an EOP symbol can be converted to a STOMP symbol using this bit. Setting EM when BC is 10 or 11 causes an illegal condition. Software enable. If SE and EN are set, an error continues to be injected until EN is cleared by software. Symbol select. Packet or control symbol selection 0 Inject error in a packet. 1 Inject error in a control symbol. Enabled to start counting and inject error at the specified word count, byte within the word, and bit within the byte. Hardware injects error and clears this bit. 29 30 SE SS 31 EN 17.3.2.2 ATMU Registers Although the Parallel RapidIO Interconnect Specification allows for 48- and 64-bit addresses, the MPC8540 implementation only allows for 34-bit RapidIO addresses. ATMU window misses use the window 0 register set by default. Overlapping window hits result in the use of the lowest number window register set hit. For both inbound and outbound translation, the smallest window size is 4 Kbytes and the largest window size is 4 Gbytes. See Section 17.7, “ATMU (Address Translation and Mapping Unit),” for more information. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 17-36 Freescale Semiconductor Memory Map/Register Definition 17.3.2.2.1 RapidIO Outbound Window Translation Address Registers 0–8 (ROWTARn) The RapidIO outbound window translation address registers (ROWTARs) select the starting addresses in the external address space for window hits within the outbound translation windows. The new translated address is created by concatenating the transaction offset to this translation address. The target ID is formed from the top eight bits of the translated address. This register takes on one of two formats, determined by the RDTYP and WRTYP fields in the corresponding attributes register (ROWARn), one for standard operations and one for maintenance operations. See Section 17.7.1, “Outbound ATMU Translation,” for more information. Figure 17-30 shows the ROWTARs as used in standard operations. 0 1 2 9 10 11 12 31 R0 0 TRGTID W Reset 0000_0000_0000_0000_0000_0000_0000_0000 TREXAD TRAD Offset 0xD_0C00, 0xD_0C20, 0xD_0C40, 0xD_0C60, 0xD_0C80, 0xD_0CA0, 0xD_0CC0, 0xD_0CE0, 0xD_0D00 Figure 17-30. RapidIO Outbound Window Translation Address Registers 0–8 for Standard Transactions (ROWTARn) Table 17-34 describes the fields of the ROWTARs for standard operations. Table 17-34. ROWTARn Standard Field Descriptions Bits 0–1 2–9 10–11 Name — Reserved Description TRGTID Target ID for RapidIO packet TREXAD Translation extended address of outbound window The translation extended address correspond to bits 0–1 of the 34-bit RapidIO address. TRAD Translation address of outbound window System address that represents the starting point of the outbound translated address. The translation address must be aligned based on the size field. The translation address corresponds to bits 2–21 of the 34-bit RapidIO address. 12–31 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 17-37 RapidIO Interface Figure 17-31 shows the ROWTARs for maintenance transactions. 0 1 2 9 10 11 12 19 20 31 R0 W Reset Offset 0 TRGTID 0 0 HOP_COUNT CFG_OFFSET 0000_0000_0000_0000_0000_0000_0000_0000 0xD_0C00, 0xD_0C20, 0xD_0C40, 0xD_0C60, 0xD_0C80, 0xD_0CA0, 0xD_0CC0, 0xD_0CE0, 0xD_0D00 Figure 17-31. RapidIO Outbound Window Translation Address Registers 0–8 for Maintenance Transactions (ROWTARn) Table 17-35 describes ROWTARs fields for maintenance transactions. For maintenance transactions, the hop count is formed from ROWTAR[12–19]. Similarly, the upper 12 bits of the maintenance offset is formed from ROWTAR[20–31]. The rest of the maintenance offset is formed from the untranslated address. Table 17-35. ROWTARn Maintenance Field Descriptions Bits 0–1 2–9 10–11 12–19 20–31 Name — TRGTID — Reserved Target ID for RapidIO Packet Reserved Description HOP_COUNT Hop count of maintenance transaction CFG_OFFSET Upper 12 bits of maintenance offset. The lower 9 bits of the 21-bit RapidIO maintenance offset are formed from the untranslated address. 17.3.2.2.2 RapidIO Outbound Window Base Address Registers 1–8 (ROWBARn) The RapidIO outbound window translation address registers (ROWBARn), shown in Figure 17-32, select the base address for the windows that are translated to an alternate system address space. Addresses for outbound transactions are compared to these windows. If such a transaction does not fall within one of these spaces the transaction is forwarded through the window 0 register set. Since window 0 is the default window, and default translation does not require a base address register, ROWBAR0 is not implemented. 0 11 12 31 R0 W Reset Offset 0 0 0 00 0 00 00 0 BADD 0000_0000_0000_0000_0000_0000_0000_0000 0xD_0C028, 0xD_0C48, 0xD_0C68, 0xD_0C88, 0xD_0CA8, 0xD_0CC8, 0xD_0CEB, 0xD_0D08 Figure 17-32. RapidIO Outbound Window Base Address Registers 1–8 (ROWBARn) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 17-38 Freescale Semiconductor Memory Map/Register Definition Table 17-36 describes the fields of the ROWBARs. Table 17-36. ROWBARn Field Descriptions Bits 0–11 12–31 Name — Reserved Description BADD Base address of outbound window. Source address that is the starting point for the outbound translation window. The window must be aligned based on the size selected in the window size bits. This corresponds to bits 0–19 of a 32-bit address. 17.3.2.2.3 RapidIO Outbound Window Attributes Registers 0–8 (ROWARn) The RapidIO outbound window attributes registers (ROWARs) define the window size to translate and other attributes for the translation. The largest window size allowed is 4 Gbytes. Figure 17-33 and Figure 17-34 show the ROWARs. 0 1 3 4 5 6 7 11 12 15 16 19 20 25 26 31 R EN 0 0 0 TFLOLV PCI W Reset Offset 00 0 00 RDTYP WRTYP 000000 SIZE 1000_0000_0000_0100_0100_0000_0001_1111 0xD_0C10 Figure 17-33. RapidIO Outbound Window Attributes Register 0 (ROWAR0) 0 1 3 4 5 6 7 11 12 15 16 19 20 25 26 31 R EN W Reset Offset 000 TFLOLV PCI 00 0 00 RDTYP WRTYP 000000 SIZE 0000_0000_0000_0100_0100_0000_0001_1111 0xD_0C30, 0xD_0C50, 0xD_0C70, 0xD_0C90, 0xD_0CB0, 0xD_0CD0, 0xD_0CF0, 0xD_0D10 Figure 17-34. RapidIO Outbound Window Attributes Registers 1–8 (ROWARn) Table 17-37 describes the fields of the ROWARs. Table 17-37. ROWARn Field Descriptions Bits 0 Name EN Description Window address translation enable. Note that for ROWAR0 this bit is read-only and hardwired to 1. 0 Address translation disabled 1 Address translation enabled Reserved 1–3 — MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 17-39 RapidIO Interface Table 17-37. ROWARn Field Descriptions (continued) Bits 4–5 Name Description TFLOLV Transaction flow level priority of transaction 00 Lowest priority transaction request flow 01 Next highest priority transaction request flow 10 Highest priority level transaction request flow 11 Reserved PCI Follow PCI transaction ordering rules 0 Do not follow PCI transaction ordering rules. 1 Follow PCI transaction ordering rules (elevate write priority one level over reads). Reserved 6 7–11 12–15 — RDTYP Read transaction type. Transaction type to run on RapidIO interface if access is a read. See Section 17.7.1.2, “Outbound Special Transactions and Requirements,” for more information on maintenance transactions. 0000–0001Reserved 0010 I/O read home 0011 Reserved 0100 nread 0101–0110Reserved 0111 Maintenance read 1000–1011Reserved 1100 atomic_inc 1101 atomic_dec 1110 atomic_set 1111 atomic_clr 16–19 WRTYP Write transaction type. Transaction type to run on RapidIO interface if access is a write. See Section 17.7.1.2, “Outbound Special Transactions and Requirements,” for more information on maintenance and doorbell transactions. 0000 0001 0010 0011 0100 Reserved Flush with data Doorbell swrite nwrite 0101 nwrite_r 0110 Reserved 0111 Maintenance write 1000–1111Reserved 20–25 26–31 — SIZE Reserved Outbound window size. Outbound window size n which is the encoded 2n+1-byte window size. The smallest window size is 4 Kbytes; the largest is 4 Gbytes. 00_0000–00_1010 Reserved 00_1011 4 Kbytes 00_1100 8 Kbytes ... 01_1111 4 Gbytes 10_0000–11_1111 Reserved 17.3.2.2.4 RapidIO Inbound Window Translation Address Registers 0–4 (RIWTARn) The RapidIO inbound window translation address registers (RIWTARs), shown in Figure 17-35, select the starting addresses in the internal address space for window hits within the inbound translation windows. The new translated address is created by concatenating the transaction offset to this translation address. See Section 17.7.2, “Inbound ATMU Translation,” for more information. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 17-40 Freescale Semiconductor Memory Map/Register Definition 0 11 12 31 R0 0 0 W Reset Offset 00000 0 0 00 TRAD 0000_0000_0000_0000_0000_0000_0000_0000 0xD_0DE0, 0xD_0DC0, 0xD_0DA0, 0xD_0D80, 0xD_0D60 Figure 17-35. RapidIO Inbound Window Translation Address Registers 0–4 (RIWTARn) Table 17-38 describes the fields of the RIWTARs. Table 17-38. RIWTARn Field Descriptions Bits 0–11 Name — Reserved Description 12–31 TRAD Translation address of inbound window. System address that represents the starting point of the inbound translated address. The translation address must be aligned based on the size field. This corresponds to bits 0–19 of the 32-bit address. 17.3.2.2.5 RapidIO Inbound Window Base Address Registers 1–4 (RIWBARn) RapidIO inbound window translation address registers (RIWBARs), shown in Figure 17-36, select the base address for windows that are translated to an alternate system address space. Addresses for inbound transactions are compared to these windows. If such a transaction does not fall in the LCSBA1CSR window or one of these spaces, the transaction is forwarded through the window 0 register set. Since window 0 is the default and default translation does not require a base address register, RIWBAR0 is not implemented. 0 9 10 11 12 31 R0 0 0 0 W Reset Offset 000000 BEXAD BADD 0000_0000_0000_0000_0000_0000_0000_0000 0xD_0DC8, 0xD_0DA8, 0xD_0D88, 0xD_0D68 Figure 17-36. RapidIO Inbound Window Base Address Registers 1–4 (RIWBARn) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 17-41 RapidIO Interface Table 17-39 describes the fields of the RIWBARs. Table 17-39. RIWBARn Field Descriptions Bits 0–9 10–11 12–31 Name — Reserved Description BEXAD Base extended address of inbound window. Field represents bits 0–1 of a 34-bit address. BADD Base address of inbound window. Source address that is the starting point for the inbound translation window. The window must be aligned based on the size selected in the window size bits. This corresponds to bits 2–21 of a 34-bit address. 17.3.2.2.6 RapidIO Inbound Window Attributes Registers 0–4 (RIWARn) The RapidIO inbound window attributes registers (RIWARs) define the window size to translate and other attributes for the translation. The largest window size allowed is 4 Gbytes. Figure 17-38 and Figure 17-37 show the RIWARs. 0 1 7 8 11 12 15 16 19 20 25 26 31 R EN 0 0 W Reset Offset 00000 TGINT RDTYP WRTYP 000000 SIZE 1000_0000_0000_0100_0100_0000_0001_1111 0xD_0DF0 Figure 17-37. RapidIO Inbound Window Attributes Register 0 (RIWAR0) 0 1 7 8 11 12 15 16 19 20 25 26 31 R EN W Reset Offset 00 000 00 TGINT RDTYP WRTYP 000000 SIZE 0000_0000_0000_0100_0100_0000_0001_1111 0xD_0DD0, 0xD_0DB0, 0xD_0D90, 0xD_0D70 Figure 17-38. RapidIO Inbound Window Attributes Registers 1–4 (RIWARn) Table 17-40 describes the fields of the RIWARs. Table 17-40. RIWARn Field Descriptions Bits 0 Name EN Description Window address translation enable. Note that for RIWAR0 the enable bit is read only and hardwired to 1. 0 Address translation disabled 1 Address translation enabled Reserved 1–7 — MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 17-42 Freescale Semiconductor Memory Map/Register Definition Table 17-40. RIWARn Field Descriptions (continued) Bits 8–11 Name TGINT Description Target interface 0000 PCI/PCI-X 0001–1110 Reserved 1111 Local memory (DDR SDRAM, local bus controller (LBC), L2 cache/SRAM) Note: If TGINT is set to an I/O port (for example, PCI/PCI-X) rather than local memory space, attributes for the external I/O transaction are assigned in an outbound ATMU of that I/O controller. 12–15 RDTYP Read transaction type. Transaction type to run if access is a read. The field description depends on the target of the transaction (to I/O interface or to local memory). Following are the transaction type settings for reads from the PCI/PCI-X interface: 0000–0011 Reserved 0100 Read 0101–1111 Reserved The following are the transaction type settings for reads from local memory: 0000–0011 Reserved 0100 Received read; do not snoop processor core (e500) 0101 Received read; snoop processor core (e500) 0110 Reserved 0111 Received read; unlock L2 cache line 1000–1111 Reserved 16–19 WRTYP Write transaction type. Transaction type to run if access is a write. The field description depends on the target of the transaction (to I/O interface or to local memory). Following are the transaction type settings for writes to the PCI/PCI-X interface: 0000–0011 Reserved 0100 Write 0101–1111 Reserved Following are the transaction type settings for writes to local memory: 0000–0011 Reserved 0100 Received write; don’t snoop processor core (e500) 0101 Received write; snoop processor core (e500) 0110 Received write; allocate L2 cache line 0111 Received write; allocate and lock L2 cache line 1000–1111 Reserved 20–25 26–31 — SIZE Reserved Inbound window size. Inbound window size N which is the encoded 2n+1 bytes window size. The smallest window size is 4 Kbytes. The largest window size is 4 Gbytes. 00_0000–00_1010 Reserved 00_1011 4 Kbytes 00_1100 8 Kbytes ... 01_1111 4 Gbytes 10_0000–11_1111 Reserved 17.3.2.3 Error Management Registers RapidIO error management registers allow configuration of RapidIO error detection, capture, and reporting. Bits in the port notification/fatal error detect register (PNFEDR) report which port notification or error was detected, subject to individual detection enable bit settings in the port notification/fatal error detect disable register (PNFEDiR). Upon detecting a port notification or MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 17-43 RapidIO Interface error, an interrupt is generated subject to the interrupt enable bit settings in the port notification/fatal error interrupt enable register (PNFEIER); the port error capture status register (PECSR) indicates whether a valid packet was captured when the port notification or fatal error occurred. If a packet is captured (PECSR[V] is set), packet information may be found in the error packet capture registers (EPCRn). The packet format type, as captured in EPCR0[PFT], determines the format of the remaining packet data captured in EPCR1 and EPCR2. Bits in the port recoverable error detect register (PREDR) report which recoverable errors have been detected. The port error recovery threshold register (PERTR) and the port retry threshold register (PRTR) allow software to assign and monitor the recovery and retry thresholds, which are referenced by hardware before reporting a recovery error (PNFEDR[ETE]) or a retry error (PNFEDR[RTE]). 17.3.2.3.1 Port Notification/Fatal Error Detect Register (PNFEDR) The PNFEDR, shown in Figure 17-39, reflects individual port notifications and fatal errors that the RapidIO controller has detected. Each bit is cleared when a 1 is written to it. 0 1 2 3 5 6 7 8 9 10 11 12 13 14 15 R MFE IAXE OAXE 0 0 0 ITE MSE NTIE SE DSE MLE ETE RTE RERE LTE W Reset 16 17 18 19 20 21 0000_0000_0000_0000 22 23 24 25 26 27 28 29 30 31 R RSTE NAE IRFE URDE URE URSE IWRE IRSFE IRTE BRRE BWRE URSSE URESE URSDE MERE MDFE W Reset Offset 0000_0000_0000_0000 0xD_0E00 Figure 17-39. Port Notification/Fatal Error Detect Register (PNFEDR) Table 17-41 describes the fields of the PNFEDR. Table 17-41. PNFEDR Field Descriptions Bits 0 Name MFE Description Multiple fatal errors. (Bit reset, write-one-to-clear) 0 Multiple fatal RapidIO errors of the same type were not detected. 1 Multiple fatal RapidIO errors of the same type were detected. Inbound ATMU window crossing error. (Bit reset, write-one-to-clear) 0 No inbound window crossing error 1 This error indicates that an inbound transaction either crosses the boundary of its ATMU window or crosses into a higher priority ATMU window. 1 IAXE MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 17-44 Freescale Semiconductor Memory Map/Register Definition Table 17-41. PNFEDR Field Descriptions (continued) Bits 2 Name OAXE Description Outbound ATMU window crossing error. (Bit reset, write-one-to-clear) 0 No outbound window crossing error 1 This error indicates that an outbound transaction either crosses the boundary of its ATMU window or crosses into a higher priority ATMU window. Reserved Illegal transaction target error. (Bit reset, write-one-to-clear) 0 No illegal transaction detected 1 Received a packet whose target ID does not match the RapidIO controller deviceID determined at POR while accept_all mode is disabled Message size error. (Bit reset, write-one-to-clear) 0 No message size error detected 1 Received message packet data payload is not of the size specified in the ssize field (with the exception of the last packet which may be less) Missing idle after training. (Bit reset, write oneto clear) 0 Idle received after a requested training sequence 1 Idle not received after a requested training sequence completes Message segment error. (Bit reset, write-one-to-clear) 0 No message segment error. 1 Received a message packet with the segment field greater than the message length field (msgseg > msglen) Duplicate message segment. (Bit reset, write-one-to-clear) 0 No duplicate message segment received 1 Received a duplicate message segment from RapidIO Message length error. (Bit reset, write-one-to-clear) 0 No message length error detected 1 Received a message packet with the same mailbox, letter, and source ID as the current outstanding (not completed) message, but with a different message length. Error recovery threshold error. (Bit reset, write-one-to-clear) 0 No error recovery threshold error detected 1 Error recovery threshold count (defined in PERTR[RCTT]) exceeded Retry threshold error. (Bit reset, write-one-to-clear) 0 No retry threshold error detected 1 Consecutive ack retry control symbols received threshold count (defined in PRTR[RTT]) exceeded Received request ERROR response. (Bit reset, write-one-to-clear) 0 No ERROR response received 1 Received a response of type error for anything other than a message Link response time-out. (Bit reset, write-one-to-clear) 0 No link response time-out 1 A link response is not received within the time-out interval, as defined by PLTOCCSR[TV]. Packet response time-out. (Bit reset, write-one-to-clear) 0 No packet response time-out 1 A packet response is not received within the time-out interval, as defined by PRTOCCSR[TV]. Nonsensical ackID. (Bit reset, write-one-to-clear) 0 AckID OK 1 Unrecognized ackID received with link response 3–5 6 — ITE 7 MSE 8 NTIE 9 SE 10 DSE 11 MLE 12 ETE 13 RTE 14 RERE 15 LTE 16 RSTE 17 NAE MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 17-45 RapidIO Interface Table 17-41. PNFEDR Field Descriptions (continued) Bits 18 Name IRFE Description Illegal request fields 0 Request packet fields OK 1 Illegal combinations of fields in the request packet. (Bit reset, write-one-to-clear) Unexpected request data 0 Request data OK 1 Read type with a data payload. (Bit reset, write-one-to-clear). Unsupported request 0 Packet format type OK 1 Unsupported packet format type or transport type. (Bit reset, write-one-to-clear) Unexpected request size 0 Packet data size OK 1 Packet data is not expected size with respect to the packet size bits (it reset, write-one-to-clear) Illegal write request 0 Write request OK 1 Write type with no data payload. (Bit reset, write-one-to-clear) Illegal response fields 0 Response packet OK 1 Illegal combinations of fields in the response packet. (Bit reset, write-one-to-clear) Illegal response type 0 Response type OK 1 Illegal response for a given request type. (Bit reset, write-one-to-clear) Bad read response 0 Read response OK 1 Read data response with no data payload. (Bit reset, write-one-to-clear) 19 URDE 20 URE 21 URSE 22 IWRE 23 IRSFE 24 IRTE 25 BRRE 26 BWRE Bad write response 0 Write response OK 1 Write response with a data payload. (Bit reset, write-one-to-clear) URSSE Unexpected response data size 0 Response packet size OK 1 Response packet data is not expected size. (Bit reset, write-one-to-clear) URESE Unsolicited response without data 0 No response without data condition 1 A response without data is received if there exists no outstanding request matching that TID. (Bit reset, write-one-to-clear). URSDE Unsolicited response with data 0 No unsolicited response with data condition 1 A response with data is received if there exists no outstanding request matching that TID. (Bit reset, write-one-to-clear) MERE Message error response 0 No message error response 1 Received a response of type error for an outbound message packet. (Bit reset, write-one-to-clear) Message descriptor fetch error 0 No message descriptor fetch error 1 A message descriptor fetch from local memory results in an error. (Bit reset, write-one-to-clear) 27 28 29 30 31 MDFE MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 17-46 Freescale Semiconductor Memory Map/Register Definition 17.3.2.3.2 Port Notification/Fatal Error Detect Disable Register (PNFEDiR) The PNFEDiR individually disables the detection of each port notification or fatal error defined in the PNFEDR. Figure 17-40 shows the PNFEDiR. 0 1 2 3 5 6 7 8 9 10 11 12 13 14 15 R W Reset 0 IAXD OAXD 0 0 0 ITD MSD NTID SD DSD BMD ETD RTD RERD LTD 0000_0000_0000_0000 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 R RSTD NAD IRFD URDD URD URSD IWRD IRSFD IRTD BRRD BWRD URSSD URESD URSDD MERD MDFD W Reset Offset 0000_0000_0000_0000 0xD_0E04 Figure 17-40. Port Notification/Fatal Error Detect Disable Register (PNFEDiR) Table 17-42 describes the fields of the PNFEDiR. Table 17-42. PNFEDiR Field Descriptions Bits 0 1 Name — IAXD Reserved Inbound ATMU window crossing error detection disable 0 Error detection enabled 1 Error detection disabled Outbound ATMU window crossing error detection disable 0 Error detection enabled 1 Error detection disabled Reserved Illegal transaction target error detection disable 0 Error detection enabled 1 Error detection disabled Message size error detection disable 0 Error detection enabled 1 Error detection disabled Missing idle after training error detection disable 0 Error detection enabled 1 Error detection disabled Message segment error detection disable 0 Error detection enabled 1 Error disabled Description 2 OAXD 3–5 6 — ITD 7 MSD 8 NTID 9 SD MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 17-47 RapidIO Interface Table 17-42. PNFEDiR Field Descriptions (continued) Bits 10 Name DSD Description Duplicate message segment error detection disable 0 Error detection enabled 1 Error detection disabled Message length error detection disable 0 Error detection enabled 1 Error detection disabled Error recovery threshold error detection disable 0 Error detection enabled 1 Error detection disabled Retry threshold error detection disable 0 Error detection enabled 1 Error detection disabled Received ERROR response error detection disable 0 Error detection enabled 1 Error detection disabled Link response time-out error detection disable 0 Error detection enabled 1 Error detection disabled Packet response time-out error detection disable 0 Error detection enabled 1 Error detection disabled Nonsensical ackID error detection disable 0 Error detection enabled 1 Error detection disabled Illegal request fields error detection disable 0 Error detection enabled 1 Error detection disabled Unexpected request data error detection disable 0 Error detection enabled 1 Error detection disabled Unsupported request error detection disable 0 Error detection enabled 1 Error detection disabled Unexpected request size error detection disable 0 Error detection enabled 1 Error detection disabled Illegal write request error detection disable 0 Error detection enabled 1 Error detection disabled Illegal response fields error detection disable 0 Error detection enabled 1 Error detection disabled Illegal response type error detection disable 0 Error detection enabled 1 Error detection disabled 11 BMD 12 ETD 13 RTD 14 RERD 15 LTD 16 RSTD 17 NAD 18 IRFD 19 URDD 20 URD 21 URSD 22 IWRD 23 IRSFD 24 IRTD MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 17-48 Freescale Semiconductor Memory Map/Register Definition Table 17-42. PNFEDiR Field Descriptions (continued) Bits 25 Name BRRD Description Bad read response error detection disable 0 Error detection enabled 1 Error detection disabled Bad write response error detection disable 0 Error detection enabled 1 Error detection disabled Unexpected response data size error detection disable 0 Error detection enabled 1 Error detection disabled Unsolicited response without data error detection disable 0 Error detection enabled 1 Error detection disabled Unsolicited response with data error detection disable 0 Error detection enabled 1 Error detection disabled Message error response error detection disable 0 Error detection enabled 1 Error detection disabled Message descriptor fetch error detection disable 0 Error detection enabled 1 Error detection disabled 26 BWRD 27 URSSD 28 URESD 29 URSDD 30 MERD 31 MDFD 17.3.2.3.3 Port Notification/Fatal Error Interrupt Enable Register (PNFEIER) The PNFEIER, shown in Figure 17-41, individually enables each detected port notification or fatal error to generate an interrupt to the programmable interrupt controller (PIC). 0 1 2 3 5 6 7 8 9 10 11 12 13 14 15 R W Reset 0 0 0 0 IAXIE OAXIE ITIE MSIE NTIIE SIE DSIE BMIE ETIE RTIE RERIE LTIE 0000_0000_0000_0000 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 R RSTIE NAIE IRFIE URDIE URIE URSIE IWRIE IRSFIE IRTIE BRRIE BWRIE URSSIE URESIE URSDIE MERIE MDFIE W Reset Offset 0000_0000_0000_0000 0xD_0E08 Figure 17-41. Port Notification/Fatal Error Interrupt Enable Register (PNFEIER) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 17-49 RapidIO Interface Table 17-43 describes the fields of the PNFEIER. Table 17-43. PNFEIER Field Descriptions Bits 0 1 Name — IAXIE Reserved Inbound ATMU window crossing error interrupt enable 0 Interrupt enabled 1 Interrupt disabled Outbound ATMU window crossing error interrupt enable 0 Interrupt enabled 1 Interrupt disabled Reserved Illegal transaction target error interrupt enable 0 Interrupt enabled 1 Interrupt disabled Message size error interrupt enable 0 Interrupt enabled 1 Interrupt disabled Missing idle after training interrupt enable 0 Interrupt enabled 1 Interrupt disabled Message segment error interrupt enable 0 Interrupt enabled 1 Interrupt disabled Duplicate message segment error interrupt enable 0 Interrupt enabled 1 Interrupt disabled Message length error interrupt enable 0 Interrupt enabled 1 Interrupt disabled Error recovery threshold error interrupt enable 0 Interrupt enabled 1 Interrupt disabled Retry threshold error interrupt enable 0 Interrupt enabled 1 Interrupt disabled Received ERROR response error interrupt enable 0 Interrupt enabled 1 Interrupt disabled Link response time-out error interrupt enable 0 Interrupt enabled 1 Interrupt disabled Packet response time-out error interrupt enable 0 Interrupt enabled 1 Interrupt disabled Description 2 OAXIE 3–5 6 — ITIE 7 MSIE 8 NTIIE 9 SIE 10 DSIE 11 BMIE 12 ETIE 13 RTIE 14 RERIE 15 LTIE 16 RSTIE MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 17-50 Freescale Semiconductor Memory Map/Register Definition Table 17-43. PNFEIER Field Descriptions (continued) Bits 17 Name NAIE Description Nonsensical ackID error interrupt enable 0 Interrupt enabled 1 Interrupt disabled Illegal request fields error interrupt enable 0 Interrupt enabled 1 Interrupt disabled Unexpected request data error interrupt enable 0 Interrupt enabled 1 Interrupt disabled Unsupported request error interrupt enable 0 Interrupt enabled 1 Interrupt disabled Unexpected request size error interrupt enable 0 Interrupt enabled 1 Interrupt disabled Illegal write request error interrupt enable 0 Interrupt enabled 1 Interrupt disabled 18 IRFIE 19 URDIE 20 URIE 21 URSIE 22 IWRIE 23 IRSFIE Illegal response fields error interrupt enable 0 Interrupt enabled 1 Interrupt disabled IRTIE Illegal response type error interrupt enable 0 Interrupt enabled 1 Interrupt disabled Bad read response error interrupt enable 0 Interrupt enabled 1 Interrupt disabled 24 25 BRRIE 26 BWRIE Bad write response error interrupt enable 0 Interrupt enabled 1 Interrupt disabled URSSIE Unexpected response data size error interrupt enable 0 Interrupt enabled 1 Interrupt disabled URESIE Unsolicited response without data error interrupt enable 0 Interrupt enabled 1 Interrupt disabled URSDIE Unsolicited response with data error interrupt enable 0 Interrupt enabled 1 Interrupt disabled MERIE Message error response error interrupt enable 0 Interrupt enabled 1 Interrupt disabled Message descriptor fetch error interrupt enable 0 Interrupt enabled 1 Interrupt disabled 27 28 29 30 31 MDFIE MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 17-51 RapidIO Interface 17.3.2.3.4 Port Error Capture Status Register (PECSR) The PECSR, shown in Figure 17-42, indicates whether a valid packet has been captured when a port notification or fatal error has occurred. Hardware sets PECSR[V] and software is expected to clear it after processing the error to capture future error packets. 0 30 31 R0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 V W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0xD_0E0C Figure 17-42. Port Error Capture Status Register (PECSR) Table 17-44 describes the fields of the PECSR. Table 17-44. PECSR Field Descriptions Bits 0–30 31 Name — V Reserved Packet register information valid 0 Packet capture registers do not contain valid information. 1 Packet capture registers contain valid information. Description 17.3.2.3.5 Error Packet Capture Register 0 (EPCR0) The EPCR0, shown in Figure 17-43, captures the first 32 bits of an inbound packet that caused a port notification or fatal error. The format of the first 32 bits of an inbound packet are the same for all packet format types. 0 1 3 4 5 6 7 8 9 10 11 12 15 16 23 24 31 R S W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0xD_0E10 A R ¬S R R P TT PFT DID SID Figure 17-43. Error Packet Capture Register 0 (EPCR0) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 17-52 Freescale Semiconductor Memory Map/Register Definition Table 17-45 describes the fields of the EPCR0. Table 17-45. EPCR0 Field Descriptions Bits 0 1–3 4 5 6–7 8–9 10–11 12–15 16–23 24–31 Name S A R S-bit of the packet AckID field of a packet These bits correspond to bits 4–7 of a packet header. In the packet definition bit 4 is a reserved packet field, bit 5 is S-complement, and bits 6 and 7 are reserved packet fields. (Bits 4, 6 and 7 are reserved packet fields, not reserved register fields.) Description ¬S R P TT PFT DID SID Priority field of a packet Transport type field of a packet Packet format type field of a packet Destination ID field of a packet Source ID field of a packet 17.3.2.3.6 Error Packet Capture Register 1 (EPCR1) EPCR1 captures a portion of an inbound packet that caused a port notification or a fatal error. The format of the EPCR1 contents vary subject to the packet format type as captured in EPCR0[PFT]. The various content formats of EPCR1 are described in Section 17.3.2.3.7, “EPCR1—Type 1 Packet Format,” through Section 17.3.2.3.15, “EPCR1—Type 13 Packet Format.” 17.3.2.3.7 EPCR1—Type 1 Packet Format Figure 17-44 describes EPCR1 for a type 1 packet format (EPCR0[PFT] = 1). 0 3 4 7 8 15 16 23 24 31 R TRANS W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0xD_0E14 RDSZ STID SECID SECTID Figure 17-44. Error Packet Capture Register 1 (EPCR1)—Type 1 Packet Format MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 17-53 RapidIO Interface Table 17-46 describes the EPCR1 fields for a type 1 packet format (EPCR0[PFT] = 1). Table 17-46. EPCR1 Type 1 Field Descriptions Bits 0–3 4–7 8–15 16–23 24–31 Name TRANS Transaction field of a packet RDSZ STID SECID Read size Source transaction ID Secondary source ID Description SECTID Secondary transaction ID 17.3.2.3.8 EPCR1—Type 2 Packet Format Figure 17-45 describes the EPCR1 for a type 2 packet format (EPCR0[PFT] = 2). 0 3 4 7 8 15 16 31 R TRANS W Reset Offset RDSZ STID 0000000000000000 0000_0000_0000_0000_0000_0000_0000_0000 0xD_0E14 Figure 17-45. Error Packet Capture Register 1 (EPCR1)—Type 2 Packet Format Table 17-47 describes the EPCR1 fields for a type 2 packet format (EPCR0[PFT] = 2). Table 17-47. EPCR1 Type 2 Field Descriptions Bits 0–3 4–7 8–15 16–31 Name TRANS Transaction field of a packet RDSZ STID — Read size Source transaction ID Reserved Description 17.3.2.3.9 EPCR1—Type 5 Packet Format Figure 17-46 describes the EPCR1 for a type 5 packet format (EPCR0[PFT] = 5). 0 3 4 7 8 15 16 31 R TRANS W Reset Offset WRSZ STID 0 000 000 0000 000 00 0000_0000_0000_0000_0000_0000_0000_0000 0xD_0E14 Figure 17-46. Error Packet Capture Register 1 (EPCR1)—Type 5 Packet Format MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 17-54 Freescale Semiconductor Memory Map/Register Definition Table 17-48 describes the EPCR1 fields for a type 5 packet format (EPCR0[PFT] = 5). Table 17-48. EPCR1 Type 5 Field Descriptions Bits 0–3 4–7 8–15 16–31 Name TRANS Transaction field of a packet WRSZ Write size STID — Source transaction ID Reserved Description 17.3.2.3.10EPCR1—Type 6 Packet Format Figure 17-47 shows the EPCR1 for a type 6 packet format (EPCR0[PFT] = 6). 0 31 R0 W Reset Offset 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0000_0000_0000_0000_0000_0000_0000_0000 0xD_0E14 Figure 17-47. Error Packet Capture Register 1 (EPCR1)—Type 6 Packet Format Table 17-49 describes the field of the EPCR1 for a type 6 packet format (EPCR0[PFT] = 6). Table 17-49. EPCR1 Type 6 Field Descriptions Bits 0–31 Name — Reserved Description 17.3.2.3.11 EPCR1—Type 8 Request Packet Format Figure 17-48 shows EPCR1 for a type 8 request packet format (EPCR0[PFT] = 8 and EPCR1[TRANS] = 0 or 1). 0 3 4 7 8 15 16 31 R TRANS W Reset Offset RDSZ/WRSZ STID 0000000000000000 0000_0000_0000_0000_0000_0000_0000_0000 0xD_0E14 Figure 17-48. Error Packet Capture Register 1 (EPCR1)—Type 8 Request Packet Format MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 17-55 RapidIO Interface Table 17-50 describes EPCR1 fields for a type 8 request packet format (EPCR0[PFT] = 8 and EPCR1[TRANS] = 0 or 1). Table 17-50. EPCR1 Type 8 Request Field Descriptions Bits 0–3 4–7 8–15 16–31 Name TRANS Transaction field of a packet Description RDSZ/WTSZ Read or write size depending on whether it is a maintenance read or write request STID — Source transaction ID Reserved 17.3.2.3.12EPCR1—Type 8 Response Packet Format Figure 17-49 describes the EPCR1 for a type 8 response packet format (EPCR0[PFT] = 8 and EPCR1[TRANS] = 2 or 3). 0 3 4 7 8 15 16 31 R TRANS W Reset Offset STAT STID 000000 0000000 000 0000_0000_0000_0000_0000_0000_0000_0000 0xD_0E14 Figure 17-49. Error Packet Capture Register 1 (EPCR1)—Type 8 Response Packet Format Table 17-51 describes the fields of the EPCR1 for a type 8 response packet format (EPCR0[PFT] = 8 and EPCR1[TRANS] = 2 or 3). Table 17-51. EPCR1 Type 8 Response Field Descriptions Bits 0–3 4–7 8–15 16–31 Name TRANS Transaction field of a packet STAT STID — Status field of a packet Source transaction ID Reserved Description 17.3.2.3.13EPCR1—Type 10 Packet Format Figure 17-50 describes the EPCR1 for a type 10 packet format (EPCR0[PFT] = 10). MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 17-56 Freescale Semiconductor Memory Map/Register Definition 0 7 8 15 16 31 R0 0 0 0 0 0 0 0 STID W Reset Offset 00 00 00 00 00 00 00 00 0000_0000_0000_0000_0000_0000_0000_0000 0xD_0E14 Figure 17-50. Error Packet Capture Register 1 (EPCR1)—Type 10 Packet Format Table 17-52 describes EPCR1 fields for a type 10 packet format (EPCR0[PFT] = 10). Table 17-52. EPCR1 Type 10 Field Descriptions Bits 0–7 8–15 16–31 Name — Reserved Description STID Source transaction ID — Reserved 17.3.2.3.14EPCR1—Type 11 Packet Format Figure 17-51 describes the EPCR1 for a type 11 packet format (EPCR0[PFT] = 11). 0 3 4 7 8 9 10 11 12 15 16 31 R MSLEN W Reset Offset SSZ LET MBOX MSSEG 0000000000000000 0000_0000_0000_0000_0000_0000_0000_0000 0xD_0E14 Figure 17-51. Error Packet Capture Register 1 (EPCR1)—Type 11 Packet Format Table 17-53 describes EPCR1 fields for a type 11 packet format (EPCR0[PFT] = 11). Table 17-53. EPCR1 Type 11 Field Descriptions Bits 0–3 4–7 8–9 10–11 12–15 16–31 Name Description MSLEN Message length field of a packet SSZ LET MBOX Segment size field of a packet Letter field of a packet Mailbox field of a packet MSSEG Message segment field of a packet — Reserved MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 17-57 RapidIO Interface 17.3.2.3.15EPCR1—Type 13 Packet Format Figure 17-52 describes the EPCR1 for a type 13 packet format (EPCR0[PFT] = 13). 0 3 4 7 8 15 16 31 R TRANS W Reset Offset STAT STID 0 000 00 00 00 00 00 00 0000_0000_0000_0000_0000_0000_0000_0000 0xD_0E14 Figure 17-52. Error Packet Capture Register 1 (EPCR1)—Type 13 Packet Format Table 17-54 describes EPCR1 fields for a type 13 packet format (EPCR0[PFT] = 13). Table 17-54. EPCR1 Type 13 Field Descriptions Bits 0–3 4–7 8–15 16–31 Name TRANS Transaction field of a packet STAT STID — Status field of a packet Source transaction ID Reserved Description 17.3.2.3.16Error Packet Capture Register 2 (EPCR2) This register captures another portion of an inbound packet that caused a port notification or a fatal error. The format of the contents of this register are vary subject to the packet format type as captured in EPCR0[PFT]. The various content formats of this register are described in Section 17.3.2.3.17, “EPCR2—Type 1, 2, or 5 Packet Format,” through Section 17.3.2.3.22, “EPCR2—Type 11 or 13 Packet Format.” 17.3.2.3.17EPCR2—Type 1, 2, or 5 Packet Format Figure 17-53 shows EPCR2 for a type 1, 2, or 5 packet format (EPCR0[PFT] = 1, 2, or 5). 0 28 29 30 31 R W Reset Offset PAD WDPTR XAMSBS 0000_0000_0000_0000_0000_0000_0000_0000 0xD_0E18 Figure 17-53. Error Packet Capture Register 2 (EPCR2)—Type 1, 2, or 5 Packet Format MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 17-58 Freescale Semiconductor Memory Map/Register Definition Table 17-55 shows EPCR2 for a type 1, 2, or 5 packet format (EPCR0[PFT] = 1, 2, or 5). Table 17-55. EPCR2 Type 1, 2, or 5 Field Descriptions Bits 0–28 29 30–31 Name PAD WDPTR Address[2:30] field of a packet Word pointer field of a packet, which is used in conjunction with the data size field Description XAMSBS Address[0:1] field of a packet 17.3.2.3.18EPCR2—Type 6 Packet Format Figure 17-54 describes the EPCR2 for a type 6 packet format (EPCR0[PFT] = 6). 0 28 29 30 31 R W Reset Offset PAD 0 XAMSBS 0000_0000_0000_0000_0000_0000_0000_0000 0xD_0E18 Figure 17-54. Error Packet Capture Register 2 (EPCR2)—Type 6 Packet Format Table 17-56 describes the EPCR2 fields for a type 6 packet format (EPCR0[PFT] = 6). Table 17-56. EPCR2 Type 6 Field Descriptions Bits 0–28 29 30–31 Name PAD — Description Address[2:30] field of a packet Reserved XAMSBS Address[0:1] field of a packet 17.3.2.3.19EPCR2—Type 8 Request Packet Format Figure 17-55 shows EPCR2 for a type 8 request packet format (EPCR0[PFT] = 8 and EPCR1[TRANS] = 0 or 1). 0 7 8 28 29 30 31 R W Reset Offset HPCNT MNTOF WDPTR XAMSBS 0000_0000_0000_0000_0000_0000_0000_0000 0xD_0E18 Figure 17-55. Error Packet Capture Register 2 (EPCR2)—Type 8 Request Packet Format MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 17-59 RapidIO Interface Table 17-57 describes the fields of the EPCR2 for a type 8 request packet format (EPCR0[PFT] = 8 and EPCR1[TRANS] = 0 or 1). Table 17-57. EPCR2 Type 8 Request Field Descriptions Bits 0–7 8–28 29 30–31 Name HPCNT MNTOF WDPTR Hop count field of a packet Maintenance offset field of a packet WdPtr field of a packet Description XAMSBS Address[0:1] field of a packet 17.3.2.3.20EPCR2—Type 8 Response Packet Format Figure 17-56 shows EPCR2 for a type 8 response packet format (EPCR0[PFT] = 8 and EPCR1[TRANS] = 2 or 3). 0 7 8 31 R W Reset Offset HPCNT 0 00 0 0 00 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0000_0000_0000_0000_0000_0000_0000_0000 0xD_0E18 Figure 17-56. Error Packet Capture Register 2 (EPCR2)—Type 8 Response Packet Format Table 17-58 describes EPCR2 for a type 8 response packet format (EPCR0[PFT] = 8 and EPCR1[TRANS] = 2 or 3). Table 17-58. EPCR2 Type 8 Response Field Descriptions Bits 0–7 8–31 Name HPCNT Hop count field of a packet — Reserved Description 17.3.2.3.21EPCR2—Type 10 Packet Format Figure 17-57 describes the EPCR2 for a type 10 packet format (EPCR0[PFT] = 10). 0 15 16 31 R W Reset Offset INFO 000 000 000 000 000 0 0000_0000_0000_0000_0000_0000_0000_0000 0xD_0E18 Figure 17-57. Error Packet Capture Register 2 (EPCR2)—Type 10 Packet Format MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 17-60 Freescale Semiconductor Memory Map/Register Definition Table 17-59 describes EPCR2 for a type 10 packet format (EPCR0[PFT] = 10). Table 17-59. EPCR2 Type 10 Field Descriptions Bits 0–15 16–31 Name INFO Information field of a packet — Reserved Description 17.3.2.3.22EPCR2—Type 11 or 13 Packet Format Figure 17-58 shows EPCR2 for a type 11 or 13 packet format (EPCR0[PFT] = 11 or 13). 0 31 R00 000000000000 0 W Reset Offset 00000 00000 000000 0 0000_0000_0000_0000_0000_0000_0000_0000 0xD_0E18 Figure 17-58. Error Packet Capture Register 2 (EPCR2)—Type 11 or 13 Packet Format Table 17-60 describes EPCR2 for a type 11 or 13 packet format (EPCR0[PFT] = 11 or 13). Table 17-60. EPCR2 Type 11 or 13 Field Description Bits 0–31 Name — Reserved Description 17.3.2.3.23Port Recoverable Error Detect Register (PREDR) The PREDR, described in Figure 17-59, indicates the recoverable errors detected on RapidIO. 0 1 6 7 8 9 10 11 12 13 14 15 R MRE W Reset 16 0 0 0 0 0 0 ETE ISE ICCE IAOSE IUAE IBCE IE256E INAIE INAUA 0000_0000_0000_0000 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 R INACC INAPS INABC INASE INAGE URE UAE ABRE UTE USE W Reset Offset 0000_0000_0000_0000 0xD_0E20 ULE UEE LRE FAE FTEE ATE Figure 17-59. Port Recoverable Error Detect Register (PREDR) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 17-61 RapidIO Interface Table 17-61 describes the fields of the PREDR. Table 17-61. PREDR Field Descriptions Bits 0 Name MRE Description Multiple recoverable errors. (Bit reset, write-one-to-clear) 0 Multiple recoverable RapidIO errors of the same type were not detected. 1 Multiple recoverable RapidIO errors of the same type were detected. Reserved Embedded training. (Bit reset, write-one-to-clear) 0 No requested training pattern 1 Requested training pattern is embedded in a packet Inbound S-bit error. (Bit reset, write-one-to-clear) 0 No inbound S-bit error 1 Received a packet/control symbol with an S-bit parity error Inbound corrupt control symbol. (Bit reset, write-one-to-clear) 0 No inbound corrupt control symbol 1 Received a control symbol in which true does not match complement counterpart 1–6 7 — ETE 8 ISE 9 ICCE 10 IAOSE Inbound out-of-sequence ack symbol. (Bit reset, write-one-to-clear) 0 No inbound out-of-sequence ack symbol 1 Received an ack control symbol with an unexpected ackID IUAE Inbound packet with unexpected ackID. (Bit reset, write-one-to-clear) 0 No inbound packet with unexpected ackID 1 Received packet with unexpected ackID value—out-of-sequence ackID Inbound packet with bad CRC. (Bit reset, write-one-to-clear) 0 No inbound packet with bad CRC 1 Received packet with a bad CRC value 11 12 IBCE 13 IE256E Inbound packet exceeds 256 bytes. (Bit reset, write-one-to-clear) 0 The data payload of an inbound packet does not exceed 256 bytes. 1 The data payload of an inbound packet exceeds 256 bytes. INAIE Packet-not-accepted; encountered internal error. (Bit reset, write-one-to-clear) 0 No packet-not-accepted; internal error encountered 1 Received ack control symbol with packet-not-accepted of type ‘encountered internal error’ 14 15 INAUA Packet-not-accepted; unexpected ackID on packet. (Bit reset, write-one-to-clear) 0 No packet-not-accepted; unexpected ackID on packet condition 1 Received ack control symbol with packet-not-accepted of type ‘unexpected ackID’ INACC Packet-not-accepted; corrupt control symbol. (Bit reset, write-one-to-clear) 0 No packet-not-accepted; corrupt control symbol condition 1 Received ack control symbol with packet-not-accepted of type ‘corrupt control symbol’ INAPS Packet-not-accepted; input port stopped. (Bit reset, write-one-to-clear) 0 No packet-not-accepted; input port stopped condition 1 Received ack control symbol with packet-not-accepted of type ‘input port stopped’ INABC Packet-not-accepted; bad CRC. (Bit reset, write-one-to-clear) 0 No packet-not-accepted; bad CRC condition 1 Received ack control symbol with packet-not-accepted of type ‘bad CRC’ 16 17 18 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 17-62 Freescale Semiconductor Memory Map/Register Definition Table 17-61. PREDR Field Descriptions (continued) Bits 19 Name Description INASE Packet-not-accepted; S-bit parity error. (Bit reset, write-one-to-clear) 0 No packet-not-accepted; S-bit parity error condition 1 Received ack control symbol with packet-not-accepted of type ‘S-bit parity error’ INAGE Packet-not-accepted; general error. (Bit reset, write-one-to-clear) 0 No packet-not-accepted; general error condition 1 Received ack control symbol with packet-not-accepted of type ‘general error’ URE Unexpected restart-from-retry symbol. (Bit reset, write-one-to-clear) 0 No unexpected restart-from-retry symbol 1 Received a restart-from-retry control symbol while in the OK state Unsolicited ACK symbol. (Bit reset, write-one-to-clear) 0 No unsolicited ack symbol encountered 1 Received an ack control symbol while no packets are posted on the outstanding ack history queue 20 21 22 UAE 23 ABRE Ack before restart-from-retry. (Bit reset, write-one-to-clear) 0 No ack before restart-from-retry 1 Received an ack control symbol after receiving an ack retry and before sending a restart-from-retry UTE Unexpected training symbol. (Bit reset, write-one-to-clear) 0 No unexpected training symbol 1 Received a training symbol while in the OK state Unexpected stomp symbol. (Bit reset, write-one-to-clear) 0 No unexpected stomp symbol 1 Received a stomp control symbol while there is no packet being received Unexpected link response symbol. (Bit reset, write-one-to-clear) 0 No unexpected link response symbol 1 Received a link response control symbol while there is no outstanding request Unexpected EOP symbol. (Bit reset, write-one-to-clear) 0 No unexpected EOP symbol 1 Received an EOP control symbol while there is no packet being received Link request error. (Bit reset, write-one-to-clear) 0 No link request error 1 Received a link request control symbol before the previous link request has been serviced Frame toggle alignment. (Bit reset, write-one-to-clear) 0 No frame toggle alignment problem 1 Received FRAME signal toggles at a non 32-bit boundary Frame toggle edge. (Bit reset, write-one-to-clear) 0 No frame toggle edge problem 1 Received FRAME signal toggles on the negative edge of the clock Ack time-out. (Bit reset, write-one-to-clear) 0 No ack time-out condition 1 An ack control symbol is not received within the specified time-out interval. 24 25 USE 26 ULE 27 UEE 28 LRE 29 FAE 30 FTEE 31 ATE MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 17-63 RapidIO Interface 17.3.2.3.24Port Error Recovery Threshold Register (PERTR) Figure 17-60 shows the PERTR. 0 7 8 15 16 23 24 31 R00 00 0 0 0 0 RCTT W Reset Offset 0 0 0 0 0 0 0 0 RCC 0000_0000_1111_1111_0000_0000_0000_0000 0xD_0E28 Figure 17-60. Port Error Recovery Threshold Register (PERTR) Table 17-62 describes the fields of the PERTR. Table 17-62. PERTR Field Descriptions Bits 0–7 8–15 16–23 24–31 Name — Reserved Description RCTT Recovery threshold trigger. These bits provide the threshold value for the number of successful transmission error recoveries before reporting an error condition. — RCC Reserved Recovery counter. Maintains a count of the number of times the port has successfully recovered from a transmission error. An error is reported (PNFEDR[ETE] is set) if this value equals the value contained in the recovery threshold trigger field and if reporting is enabled (PNFEDiR[ETD] is cleared). 17.3.2.3.25Port Retry Threshold Register (PRTR) Figure 17-61 shows the PRTR. 0 7 8 15 16 23 24 31 R0 0 0 0 0 0 0 0 RTT W Reset Offset 0 0 0 0 0 0 0 0 RC 0000_0000_1111_1111_0000_0000_0000_0000 0xD_0E2C Figure 17-61. Port Retry Threshold Register (PRTR) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 17-64 Freescale Semiconductor Memory Map/Register Definition Table 17-63 describes the fields of the PRTR. Table 17-63. PRTR Field Descriptions Bits 0–7 8–15 16–23 24–31 Name — RTT — RC Reserved Retry threshold trigger. These bits provide the threshold value for the number of consecutive ack retries received for an outbound packet before reporting an error condition. Reserved Retry counter. Maintains a count of the number of consecutive times the port has received an ack retry for an outbound packet. An error is reported (PNFEDR[RTE] is set) if this value equals the value contained in the retry threshold trigger field and if reporting is enabled (PNFEDiR[RTD] is cleared). Description 17.3.3 RapidIO Message Unit Registers This section describes RapidIO outbound and inbound message registers, doorbell registers and port-write registers. Refer to Section 17.8, “RapidIO Message Unit,” for additional information about how these registers affect the message unit. 17.3.3.1 RapidIO Outbound Message Registers The registers in this section control RapidIO outbound messages. The following sections provide descriptions of these registers. Additional information may be found in the Parallel RapidIO Interconnect Specification. 17.3.3.1.1 Outbound Mode Register (OMR) The outbound mode register (OMR) allows software to start a message operation and to control various message operation characteristics. Figure 17-62 shows the OMR. 0 10 11 12 15 16 19 20 21 22 23 24 25 26 27 28 29 30 31 R0 000000000 0 DES_SEN W Reset Offset 0 00 0 CIRQ_SIZ 00 QOIE QFIE 0 QEIE EIE 00 MUTM MUI MUS 0000_0000_0000_0000_0000_0000_0000_0000 0xD_1000 Figure 17-62. Outbound Mode Register (OMR) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 17-65 RapidIO Interface Table 17-64 describes the fields of the OMR. Table 17-64. OMR Field Descriptions Bits 0–10 11 Name — Reserved Description DES_SEN Descriptor snoop enable 0 Snooping of processor core disabled 1 Snooping of the processor core when reading descriptors from memory enabled — Reserved 12–15 16–19 CIRQ_SIZ Circular descriptor queue size. Determines the number of descriptors that can be placed in the circular queue without overflow. 0110 128 (4-Kbyte page boundary) 0000 2 0111 256 0001 4 1000 512 0010 8 1001 1024 0011 16 1010 2048 0100 32 1011–1111 Reserved 0101 64 — QOIE Reserved Queue overflow interrupt enable. Only applicable to chaining mode 0 No QO interrupt generated 1 Generates an interrupt upon detection of a queue overflow (that is, the enqueue and dequeue pointers are no longer equal after being incremented by software and the queue was full) Queue full interrupt enable. Only applicable to chaining mode. 0 No QF interrupt generated 1 Generates an interrupt when the queue transitions to full (that is, the enqueue and dequeue pointers are equal after being incremented by software) Reserved Queue empty interrupt enable. Only applicable to chaining mode 0 No QE interrupt generated 1 Generates an interrupt at the completion of all outstanding message operations (that is, the enqueue and dequeue pointers are equal after an increment by the message unit controller) Error interrupt enable 0 No interrupt is generated if a programming or transfer error is detected. 1 Generates an interrupt if a programming or transfer error is detected Reserved Message unit transfer mode 0 Message unit is configured to be in chaining mode 1 Puts the message unit into direct mode. This means that software is responsible for placing all the required parameters into necessary registers to start the message transmission. 20–21 22 23 QFIE 24 25 — QEIE 26 EIE 27–28 29 — MUTM MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 17-66 Freescale Semiconductor Memory Map/Register Definition Table 17-64. OMR Field Descriptions (continued) Bits 30 31 Name MUI MUS Description Message unit increment. Software sets MUI after writing a descriptor to memory. Hardware then increments the ODQEPAR and clears this bit.Always reads as 0 if MUS is set Message unit start • Direct mode: a 0 to 1 transition while the message unit is not busy (OSR[MUB] = 0) starts the message unit. A 1 to 0 transition has no effect. • Chaining mode: a 0 to 1 transition while the message unit is not busy (OSR[MUB] = 0) initializes the message unit in chaining mode and causes the outbound descriptor queue dequeue pointer address to be stored as the base address of the circular descriptor queue. If MUS is set, the message unit starts each time the enqueue and dequeue pointers are not equal and an overflow of the circular descriptor queue has not been detected. A 1 to 0 transition while the message unit is busy halts the message unit after the current descriptor is processed. The message unit must be re-initialized after this halt condition. 17.3.3.1.2 Outbound Status Register (OSR) The outbound status register (OSR), shown in Figure 17-63, reports various message unit conditions during and after a message operation. Writing a 1 to a writable condition bit in OSR clears it. 0 10 11 12 23 24 25 26 27 28 29 30 31 R 0 0 0 0 0 0 0 0 0 0 0 QF 0 0 0 0 0 0 0 0 0 0 0 0 TE W Reset Offset 0 QOI QFI 0 MUB EOMI QEI 0000_0000_0000_0000_0000_0000_0000_0000 0xD_1004 Figure 17-63. Outbound Status Register (OSR) Table 17-65 describes the fields of the OSR. Table 17-65. OSR Field Descriptions Bits 0–10 11 Name — QF Reserved Queue full (read-only) 0 Queue not full 1 The queue has become full. Reserved Transaction error 0 No transaction error 1 There was an error condition during the message operation. The message unit has received a response error and halts. An interrupt is generated if OMR[IEI] is set (Bit reset, write-one-to-clear, resuming message unit operation) Reserved Description 12–23 24 — TE 25 — MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 17-67 RapidIO Interface Table 17-65. OSR Field Descriptions (continued) Bits 26 Name QOI Description Queue overflow interrupt. Only applicable to chaining mode. (Bit reset, write-one-to-clear) 0 No overflow interrupt condition 1 Queue overflow is detected and OMR[QOIE] is set. An interrupt is generated in this case Queue full interrupt. Only applicable to chaining mode. (Bit reset, write-one-to-clear) 0 No queue full interrupt 1 The queue became full and OMR[QFIE] is set. An interrupt is generated in this case. Reserved Message unit busy (read-only) 0 Message unit not busy. Cleared when the message unit is no longer busy 1 A message operation is currently in progress. 27 QFI 28 29 — MUB 30 EOMI End-of-message interrupt. (Bit reset, write-one-to-clear) 0 No end-of-message interrupt condition 1 After finishing this message operation, if ODATR[EOMIE] is set, EOMI is set and an interrupt is generated QEI Queue empty interrupt. Only applicable to chaining mode. (Bit reset, write-one-to-clear) 0 No queue empty interrupt condition 1 When the last message operation in the outbound descriptor queue is finished, if OMR[QEIE] is set, then this bit is set and an interrupt is generated. Otherwise, no interrupt is generated. 31 17.3.3.1.3 Outbound Descriptor Queue Dequeue Pointer Address Register (ODQDPAR) The outbound descriptor queue dequeue pointer address register (ODQDPAR), shown in Figure 17-64, contains the address of the first descriptor in memory to be processed. Software must initialize this register to point to the first descriptor in memory. After processing this descriptor, the message unit controller increments the outbound descriptor queue dequeue pointer address (ODQDPAR[DQDPA] to point to the next descriptor. If the outbound descriptor queue enqueue pointer and the outbound descriptor queue dequeue pointer are not equal (indicating that the queue is not empty), the message unit controller reads the next descriptor from memory for processing. If the pointers are equal after the dequeue pointer has been incremented, the queue is empty and the message unit halts until the enqueue pointer is incremented by software. Incrementing the pointer indicates that a new descriptor has been added to the queue and is ready for processing. If the queue becomes empty and OMR[QEIE] is set, OSR[QEI] is set and an interrupt is generated. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 17-68 Freescale Semiconductor Memory Map/Register Definition 0 26 27 31 R DQDPA W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0xD_100C 0 000 0 Figure 17-64. Outbound Descriptor Queue Dequeue Pointer Address Register (ODQDPAR) Table 17-66 describes the fields of the ODQDPAR. Table 17-66. ODQDPAR Field Descriptions Bits 0–26 27–31 Name Description DQDPA Descriptor queue dequeue pointer address. Contains the address of the first descriptor in memory to process. The descriptor must be aligned to a 32-byte boundary. — Reserved 17.3.3.1.4 Outbound Unit Source Address Register (OSAR) The outbound unit source address register (OSAR), shown in Figure 17-65, indicates the address from which the message unit controller is to read data. Software must ensure that this is a valid local memory address. The address must be aligned to a double-word boundary, so the 3 lsbs are reserved. 0 28 29 30 31 R SAD W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0xD_1014 SNEN 0 0 Figure 17-65. Outbound Unit Source Address Registers (OSAR) Table 17-67 describes the fields of the OSAR. Table 17-67. OSAR Field Descriptions Bits 0–28 29 Name SAD Description Source address. This is the source address of the message operation. SNEN Snoop enable 0 Snooping of the processor core disabled 1 Snooping of the processor core for data reads from local memory enabled — Reserved 30–31 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 17-69 RapidIO Interface 17.3.3.1.5 Outbound Destination Port Register (ODPR) The destination port register (ODPR), shown in Figure 17-66, indicates the port to which the message unit controller is to send data. Software must ensure that this is a valid port in the receiving device. 0 29 30 31 R00000000000000000000000000000 0 MAILBOX W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0xD_1018 Figure 17-66. Outbound Destination Port Registers (ODPR) Table 17-68 describes the fields of the ODPR. Table 17-68. ODPR Field Descriptions Bits 0–29 Name — Reserved Description 30–31 MAILBOX Value for mbox field in message packet 17.3.3.1.6 Outbound Destination Attributes Register (ODATR) The outbound destination attributes register (ODATR) contains the transaction attributes to be used for the message operation. Figure 17-67 shows the ODATR. 01 2 3 4 5 6 11 12 15 16 21 22 29 30 31 R00 EOMIE W Reset Offset 0 DTFLOWLVL 000000 DWRITETTYPE 000000 DTGTROUTE 00 0000_0000_0000_0110_0000_0000_0000_0000 0xD_101C Figure 17-67. Outbound Destination Attributes Register (ODATR) Table 17-69 describes the fields of the ODATR. Table 17-69. ODATR Field Descriptions Bits 0 1 2 Name — — EOMIE Reserved Reserved, hardwired to 0. End-of-message interrupt enable 0 End-of-message interrupt disabled 1 Generates an interrupt upon completion of the current message operation. Description MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 17-70 Freescale Semiconductor Memory Map/Register Definition Table 17-69. ODATR Field Descriptions (continued) Bits 3 4–5 Name — DTFLOWLVL Reserved Transaction flow priority level 00 Lowest priority transaction flow 01 Next highest priority transaction flow 10 Highest priority transaction flow 11 Reserved Reserved Description 6–11 — 12–15 DWRITETTYPE Hardwired to 0110 (message) 16–21 22–29 30–31 — DTGTROUTE — Reserved Destination target route. Contains the target route field of the transaction (Device ID of the target) Reserved 17.3.3.1.7 Outbound Double-Word Count Register (ODCR) The outbound double-word count register (ODCR), shown in Figure 17-68, contains the number of double words for the message operation. The maximum message operation size is 4 Kbytes and the minimum is 8 bytes. 0 18 19 28 29 31 R0 W Reset Offset 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 DCR 0 0 0 0000_0000_0000_0000_0000_0000_0000_0000 0xD_1020 Figure 17-68. Outbound Double-Word Count Register (ODCR) Table 17-70 describes the fields of the ODCR. Table 17-70. ODCR Field Descriptions Bits 0–18 19–28 Name — DCR Reserved Double-word count register. Contains the number of double words for the message operation 00 0000 00 0000 00 0000 00 0000 00 0000 00 0001 29–31 — 0000Reserved 00018 bytes 001016 bytes 010032 bytes 100064 bytes 0000128 bytes 00 0010 0000256 bytes 00 0100 0000512 bytes 00 1000 00001024 bytes 01 0000 00002048 bytes 10 0000 00004096 bytes All other values yield undefined behavior. Description Reserved MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 17-71 RapidIO Interface 17.3.3.1.8 Outbound Descriptor Queue Enqueue Pointer Address Register (ODQEPAR) The outbound descriptor queue enqueue pointer address register (ODQEPAR), shown in Figure 17-69, contains the address for the next descriptor in memory to be added to the queue. Software must initialize this register to match the outbound descriptor queue dequeue pointer address. If a message is ready to be sent, software writes a descriptor to the next location in the queue (indicated by the address in ODQEPAR), and then should set OMR[MUI]. The message unit then increments ODQEPAR to point to the next descriptor location in memory and clears OMR[MUI]. This can result in a number of actions: • • If the enqueue and dequeue pointers match, the queue is now full. If OMR[QFIE] is set, then OSR[QFI] is set, and an interrupt is generated. If the enqueue and dequeue pointer no longer match after the enqueue pointer has been incremented and the queue is full, then the queue overflows, and the message unit stops. If OMR[QOIE] is set, OSR[QOI] is set and an interrupt is generated. If the enqueue and dequeue pointer are the same before reading the register and no overflow condition is detected, the message unit controller starts (if enabled). 0 26 27 31 • R DQEPA W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0xD_1028 00000 Figure 17-69. Outbound Descriptor Queue Enqueue Pointer Register (ODQEPAR) Table 17-71 describes the fields of the ODQEPAR. Table 17-71. ODQEPAR Field Descriptions Bits 0–26 27–31 Name Description DQEPA Descriptor queue enqueue pointer address. Contains the address of the last descriptor in memory to process. The descriptor must be aligned to a 32-byte boundary. — Reserved 17.3.3.2 RapidIO Inbound Message Registers Registers in this section describe the RapidIO inbound message registers. 17.3.3.2.1 Inbound Mode Register (IMR) The inbound mode register (IMR) allows software to enable the mailbox controller and to control various message operation characteristics. Figure 17-70 shows the IMR. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 17-72 Freescale Semiconductor Memory Map/Register Definition 0 10 11 12 15 16 19 20 22 23 24 25 26 27 29 30 31 R00000000000 SEN FRM_SIZ W Reset Offset CIRQ_SIZ 000 QFIE 0 MIQIE EIE 000 MI ME 0000_0000_0000_0000_0000_0000_0000_0000 0xD_1060 Figure 17-70. Inbound Mode Register (IMR) Table 17-72 describes the fields of the IMR. Table 17-72. IMR Field Descriptions Bits 0–10 11 Name — SEN Reserved Snoop enable 0 Snooping of the processor core while writing messages into memory disabled 1 Snooping of the processor core while writing messages into memory enabled Description 12–15 FRM_SIZ Message frame size. Determines the maximum message size that can be accepted by this mailbox without error. This parameter combined with CIRQ_SIZ determine the maximum contiguous memory space allocated to the mailbox. 0111 256 bytes 0000–0001 Reserved 1000 512 bytes 0010 8 bytes 1001 1024 bytes 0011 16 bytes 1010 2048 bytes 0100 32 bytes 1011 4096 bytes 0101 64 bytes 1100–1111 Reserved 0110 128 bytes 16–19 CIRQ_SIZ Circular frame queue size. Determines the number of message frames that can be placed in the circular queue without overflow. This parameter combined with FRM_SIZ determines the maximum contiguous memory space allocated to the mailbox. 0110 128 (4 Kbyte page boundary) 0000 2 0111 256 0001 4 1000 512 0010 8 1001 1024 0011 16 1010 2048 0100 32 1011–1111 Reserved 0101 64 20–22 23 — QFIE Reserved Queue full interrupt enable 0 No QFI interrupt is generated 1 Generates an interrupt when the queue is full (that is, the enqueue and dequeue pointers are equal after the dequeue pointer was incremented by the mailbox controller) Reserved Message in queue interrupt enable 0 No MIQ interrupt is generated 1 Generates an interrupt when the queue transitions to not empty (that is, the enqueue and dequeue pointers are no longer equal after an increment by the mailbox controller) 24 25 — MIQIE MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 17-73 RapidIO Interface Table 17-72. IMR Field Descriptions (continued) Bits 26 Name EIE Description Error interrupt enable 0 No interrupt is generated if a programming or transfer error is detected. 1 Generates an interrupt if a programming or transfer error is detected Reserved Mailbox increment. Software sets MI after processing an inbound message. When software sets this bit, hardware increments the IFQDPAR and clears this bit. Always reads as 0 while ME is set Mailbox enable 0 Mailbox disabled 1 The mailbox has been initialized and can service incoming message operations. 27–29 30 31 — MI ME 17.3.3.2.2 Inbound Status Register (ISR) The inbound status register (ISR), shown in Figure 17-71, reports mailbox conditions during and after a message operation. Writing a 1 to a writable condition bit in ISR clears it. 0 10 11 12 14 15 16 23 24 25 26 27 28 29 30 31 R 0 0 0 0 0 0 0 0 0 0 0 QF 0 0 0 MIQ 0 0 0 0 0 0 0 0 TE W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0xD_1064 00 QFI 0 MB 0 MIQI Figure 17-71. Inbound Status Register (ISR) Table 17-73 describes the fields of the ISR. Table 17-73. ISR Field Descriptions Bits 0–10 11 Name — QF Reserved Queue full (read only) 0 Queue is not full 1 If the queue becomes full, this bit is set. Hardware clears this bit when software empties one or more entries. Reserved Message-in-queue (read only) 0 No message in queue 1 If the queue goes from empty to not empty, this bit is set. Hardware clears this bit when software has emptied the queue. Reserved Description 12–14 15 — MIQ 16–23 — MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 17-74 Freescale Semiconductor Memory Map/Register Definition Table 17-73. ISR Field Descriptions (continued) Bits 24 Name TE Description Transaction error 0 No transaction error 1 Set when there is an error condition during the message operation. An interrupt is generated if IMR[IEI] is set. (Bit reset, write-one-to-clear) Reserved Queue full interrupt. (Bit reset, write-one-to-clear) 0 No queue full interrupt condition 1 If the queue becomes full and IMR[QFIE] is set, this bit is set and an interrupt is generated. Reserved Mailbox busy (read only) 0 Cleared as a result of receiving an error or the completion of a message operation 1 When set indicates that a message operation is currently in progress. Reserved Message-in-queue interrupt. (Bit reset, write-one-to-clear) 0 No message-in-queue interrupt condition 1 If the queue goes from empty to not empty and IMR[MIQIE] is set, this bit is set and an interrupt is generated. 25–26 27 — QFI 28 29 — MB 30 31 — MIQI 17.3.3.2.3 Inbound Frame Queue Dequeue Pointer Address Register (IFQDPAR) The inbound frame queue dequeue pointer address register (IFQDPAR), shown in Figure 17-72, contains the address for the first message in memory to be processed. Software must initialize this register to the first frame location in memory. When a message has been processed, software sets IMR[MI]. The mailbox controller then increments IFQDPAR to point to the next frame in memory and clears MI. If the inbound frame queue enqueue pointer and the inbound frame queue dequeue pointer are not equal (indicating that the queue is not empty), software can read the next message frame from memory for processing. If the enqueue and dequeue pointers are equal after being incremented by software, the queue is empty and all outstanding messages have been processed. 0 28 29 31 R FQDPA W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0xD_106C 000 Figure 17-72. Inbound Frame Queue Dequeue Pointer Address Registers (IFQDPAR) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 17-75 RapidIO Interface Table 17-74 describes the fields of the IFQDPAR. Table 17-74. IFQDPAR Field Descriptions Bits 0–28 29–31 Name Description FQDPA Frame queue dequeue pointer address. Contains the address of the first message in memory to process — Reserved 17.3.3.2.4 Inbound Frame Queue Enqueue Pointer Address Register (IFQEPAR) The inbound frame queue enqueue pointer address register (IFQEPAR), shown in Figure 17-73, contains the address for the next message frame in memory to be added to the queue. Software must initialize this register to match the frame queue dequeue pointer address. When the mailbox controller receives a message, it writes the message data to the next location in the queue (indicated by the address in IFQEPAR) and then increments IFQEPAR to point to the next frame location in memory. This can result in the following: • If the enqueue and dequeue pointers match, the queue is now full and the mailbox controller does not accept more incoming messages, returning RETRY responses to the sending devices until the queue is no longer full. If IMR[QFIE] is set, then ISR[QFI] is set and an interrupt is generated. If the enqueue and dequeue pointer were the same before the register was read, the queue has transitioned from empty to not empty. If IMR[MIQIE] is set, ISR[MIQI] is set and an interrupt is generated. 0 28 29 31 • R FQEPA W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0xD_1074 000 Figure 17-73. Inbound Frame Queue Enqueue Pointer Address Registers (IFQEPAR) Table 17-75 describes the fields of the IFQEPAR. Table 17-75. IFQEPAR Field Descriptions Bits 0–28 29–31 Name Description FQEPA Frame queue enqueue pointer address. Contains the address of the next message frame to be added to the queue. — Reserved MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 17-76 Freescale Semiconductor Memory Map/Register Definition 17.3.3.3 RapidIO Doorbell Registers These registers control the RapidIO doorbell unit. The following sections provide descriptions of these registers. 17.3.3.3.1 Doorbell Mode Register (DMR) The doorbell mode register (DMR) allows software to enable the doorbell controller to control various doorbell operation characteristics. Figure 17-74 shows the DMR. 0 10 11 12 15 16 19 20 22 23 24 25 26 27 28 29 30 31 R0 000000000 0 SEN W Reset Offset 0000 CIRQ_SIZ 000 QFIE 0 DIQIE EIE 000 DI DE 0000_0000_0000_0000_0000_0000_0000_0000 0xD_1460 Figure 17-74. Doorbell Mode Register (DMR) Table 17-76 describes the fields of the DMR. Table 17-76. DMR Field Descriptions Bits 0–10 11 Name — SEN Reserved Snoop enable 0 Snooping of the processor core while writing doorbell information into memory disabled 1 Snooping of the processor core while writing doorbell information into memory enabled Reserved Description 12–15 — 16–19 CIRQ_SIZ Circular doorbell queue size. Determines the number of doorbell entries that can be place in the circular queue. CIRQ_SIZ*8 bytes determine the maximum contiguous memory space allocated to the doorbell unit. 0000 2 0001 4 0010 8 0011 16 0100 32 0101 64 0110 128 0111 256 1000 512 (4-Kbyte page boundary) 1001 1024 1010 2048 1011–1111 Reserved 20–22 — Reserved MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 17-77 RapidIO Interface Table 17-76. DMR Field Descriptions (continued) Bits 23 Name QFIE Description Queue full interrupt enable 0 No QFI interrupt generated 1 Generates an interrupt when the queue is full (that is, the enqueue and dequeue pointers are equal after the dequeue pointer was incremented by the doorbell controller) Reserved Doorbell in queue interrupt enable 0 No DIQ interrupt generated 1 Generates an interrupt when the queue transitions to not empty (that is, the enqueue and dequeue pointers are no longer equal after an increment by the doorbell controller) Error interrupt enable 0 No interrupt is generated if a programming or transfer error is detected. 1 Generates an interrupt if a programming or transfer error is detected Reserved Doorbell increment. Software sets DI after processing an inbound doorbell. When software sets DI, hardware increments the DQDPAR and clears DI. Always reads as 0 while DE is set Doorbell enable 0 Doorbell disabled 1 The doorbell has been initialized and can service incoming doorbell operations. 24 25 — DIQIE 26 EIE 27–29 30 31 — DI DE 17.3.3.3.2 Doorbell Status Register (DSR) The doorbell status register (DSR) reports various doorbell conditions after a doorbell operation. Writing a 1 to the corresponding set bit clears it. Figure 17-75 shows the DSR. 0 10 11 12 14 15 16 23 24 25 26 27 28 29 30 31 R 0 0 0 0 0 0 0 0 0 0 0 QF 0 W Reset Offset 0 0 DIQ 0 0 0 0 0 0 0 0 TE 00 QFI 0 DB 0 DIQI 0000_0000_0000_0000_0000_0000_0000_0000 0xD_1464 Figure 17-75. Doorbell Status Register (DSR) Table 17-77 describes the fields of the DSR. Table 17-77. DSR Field Descriptions Bits 0–10 11 Name — QF Reserved Queue full (read only) 0 Queue is not full 1 If the queue becomes full, this bit is set. Hardware clears this bit when software empties one or more entry. Description MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 17-78 Freescale Semiconductor Memory Map/Register Definition Table 17-77. DSR Field Descriptions (continued) Bits 12–14 15 Name — DIQ Reserved Doorbell-in-queue (read only) 0 No doorbell-in-queue conditions 1 If the queue goes from empty to not empty, this bit is set. Hardware clears this bit when software has emptied the queue. Reserved Transaction error 0 There was no error condition during the doorbell operation. 1 There was an error condition during the doorbell operation. An interrupt is generated if DMR[IEI] is set.(Bit reset, write-one-to-clear) Reserved Queue full interrupt 0 No queue full interrupt condition 1 The queue became full and the queue fill interrupt is enabled (IMR[QFIE] = 1). An interrupt is generated. (Bit reset, write-one-to-clear). Reserved Doorbell busy (read only) 0 Cleared as a result of receiving an error or the completion of a doorbell operation 1 Indicates that a doorbell operation is currently in progress Reserved Doorbell-in-queue interrupt 0 No doorbell-in-queue interrupt condition 1 The queue went from empty to not empty and OMR[DIQIE] is set. An interrupt is generated. (Bit reset, write-one-to-clear) Description 16–23 24 — TE 25–26 27 — QFI 28 29 — DB 30 31 — DIQI 17.3.3.3.3 Doorbell Queue/Dequeue Pointer Address Register (DQDPAR) The doorbell queue dequeue pointer address register (DQDPAR), shown in Figure 17-76, contains the double-word address for the first doorbell in memory to be processed. Software must initialize this register to the first doorbell entry location in memory. When a current doorbell has been processed, software sets DMR[DI]. The doorbell hardware then increments the DQDPAR to point to the next doorbell in memory and clear DMR[DI]. If the doorbell queue enqueue pointer and the doorbell queue dequeue pointer are not equal (indicating that the queue is not empty), software can read the next doorbell from memory for processing. If the enqueue and dequeue pointers are equal after being incremented by software, the queue is empty and all outstanding doorbells have been processed. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 17-79 RapidIO Interface 0 28 29 31 R DQDPA W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0xD_146C 000 Figure 17-76. Doorbell Queue Dequeue Pointer Address Registers (DQDPAR) Table 17-78 describes the fields of the DQDPAR. Table 17-78. DQDPAR Field Descriptions Bits 0–28 29–31 Name Description DQDPA Doorbell queue dequeue pointer address. Contains the double-word address of the first doorbell in memory to process. — Reserved 17.3.3.3.4 Doorbell Queue Enqueue Pointer Address Register (DQEPAR) The doorbell queue enqueue pointer address register (DQEPAR), shown in Figure 17-77, contains the double-word address for the next doorbell entry in memory to be added to the queue. Software must initialize DQEPAR to match the doorbell queue dequeue pointer address. When the doorbell controller receives a packet, it writes the doorbell information to the next location in the queue (indicated by the address in DQEPAR) and then increments DQEPAR to point to the next doorbell location. This can result in the following: • If the enqueue and dequeue pointers match, then the queue is now full and the doorbell controller does not accept any more incoming doorbell packets, returning RETRY responses to the sending devices until the queue is no longer full. If DMR[QFIE] is set, then DSR[QFI] is set and an interrupt is generated. If the enqueue and dequeue pointer were the same before reading the register, the queue has transitioned from empty to not empty. If DRM[MIQIE] is set, DSR[MIQI] is set and an interrupt is generated. 0 28 29 31 • R DQEPA W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0xD_1474 000 Figure 17-77. Doorbell Queue Enqueue Pointer Address Register (DQEPAR) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 17-80 Freescale Semiconductor Memory Map/Register Definition Table 17-79 describes the fields of the DQEPAR. Table 17-79. DQEPAR Field Descriptions Bits 0–28 29–31 Name Description DQEPA Doorbell queue enqueue pointer address. Contains the double-word address of the next doorbell location to be added to the queue — Reserved 17.3.3.4 RapidIO Port-Write Registers These registers control the RapidIO port-write unit. The following sections describe these registers. Additional information may be found in the Parallel RapidIO Interconnect Specification. 17.3.3.4.1 Port-Write Mode Register (PWMR) The port-write mode register (PWMR) allows software to enable the port-write controller and to control various port-write operation characteristics. Figure 17-78 shows the PWMR. 0 10 11 12 22 23 24 25 26 27 29 30 31 R0000000000 0 SEN W Reset Offset 00000000000 QFIE 00 EIE 000 CQ PWE 0000_0000_0000_0000_0000_0000_0000_0000 0xD_14E0 Figure 17-78. Port-Write Mode Register (PWMR) Table 17-80 describes the fields of the PWMR. Table 17-80. PWMR Field Descriptions Bits 0–10 11 Name — SEN Reserved Snoop enable 0 Snooping the processor core while writing port-write data payload into memory disabled 1 Snooping the processor core while writing port-write data payload into memory enabled Reserved Queue full interrupt enable 0 No QFI interrupt generated 1 Generates an interrupt when the queue becomes full (that is, the controller has written a port-write data payload into memory) Reserved Error interrupt enable 0 No interrupt is generated if a programming or transfer error is detected. 1 Generates an interrupt if a programming or transfer error is detected MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 17-81 Description 12–22 23 — QFIE 24–25 26 — EIE RapidIO Interface Table 17-80. PWMR Field Descriptions (continued) Bits 27–29 30 31 Name — CQ PWE Reserved Clear queue. Set by software after processing an inbound port-write operation. Hardware clears the previous port-write operation from the port-write queue and clears CC. Always reads as 0 if PWE = 1. Port-write enable 0 Port-writes are disabled. 1 The port-write controller has been initialized and can service an incoming port-write operation. Description 17.3.3.4.2 Port-Write Status Register (PWSR) The port-write status register (PWSR), shown in Figure 17-79, reports various port-write conditions after a port-write operation. Writing a 1 to a corresponding set bit clears the bit. 0 10 11 12 23 24 25 26 27 28 29 30 31 R00 0 0 00 0 0 00 0 QF W Reset Offset 000000000000 TE 0 0 QFI 0 PWB 0 0 0000_0000_0000_0000_0000_0000_0000_0000 0xD_14E4 Figure 17-79. Port-Write Status Register (PWSR) Table 17-81 describes the fields of the PWSR. Table 17-81. PWSR Field Descriptions Bits 0–10 11 Name — QF Reserved Queue full (read only) 0 The queue is not full. 1 The queue became full. Reserved Transaction error 0 No transaction error occurred 1 There was an error condition (the port-write transaction is greater than 64 bytes) during the port-write operation. An interrupt is generated if PWMR[IEI] is set. (Bit reset, write-one-to-clear) Reserved Queue full interrupt. (Bit reset, write-one-to-clear) 0 Must be cleared by software after a port-write is serviced to re-enable the port-write controller 1 The queue became full and PWMR[QFIE] is set. An interrupt is generated. Reserved Description 12–23 24 — TE 25–26 27 — QFI 28 — MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 17-82 Freescale Semiconductor Functional Description Table 17-81. PWSR Field Descriptions (continued) Bits 29 Name PWB Description Port-write busy (read only) 0 Cleared as a result of receiving an error or upon the completion of a port-write operation 1 Indicates that a port-write operation is currently in progress Reserved 30–31 — 17.3.3.4.3 Port-Write Queue Base Address Register (PWQBAR) The port-write queue base address register (PWQBAR), shown in Figure 17-80, contains the 64-byte cache line address for the port-write data payload. PWQBAR must be initialized by software. 0 25 26 31 R PWQBA W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0xD_14EC 000000 Figure 17-80. Port-Write Queue Base Address Register (PWQBAR) Table 17-82 describes the fields of the PWQBAR. Table 17-82. PWQBAR Field Descriptions Bits 0–25 26–31 Name Description PWQBA Port-write queue base address. Contains the cache line address of the port-write data payload. — Reserved 17.4 Functional Description The following sections provide summaries of RapidIO transactions, packets and control symbols. Additional information may be found in the Parallel RapidIO Interconnect Specification. 17.4.1 RapidIO Transaction, Packet, and Control Symbol Summary Table 17-83 summarizes the RapidIO transaction types supported by the RapidIO controller. The table lists the packet format type, the RapidIO transaction, the transaction programming model, the processor operations that may cause the RapidIO transaction, possible responses for the transactions, and comments that note RapidIO-specific details associated with the transaction, or response. Transactions that are not supported by the RapidIO implementation are noted in the comments section of Table 17-83. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 17-83 RapidIO Interface Table 17-83. RapidIO Transaction Summary Packet Format RapidIO Transaction Programming Model GSM Possible Processor Operations I = 1, M = 1 loads, and instruction fetches where the target of the request is a remote address space. M = 0 read type operations may also map to I/O read home. M = 0 read type operations where the target of the request is a remote address space M = 0 read type mapped operation Possible RapidIO Responses done, (data_only + done_interv) retry, error Comments I/O read home 2 Nonintervention request class None nread 2 Nonintervention request class atomic IO done, error None IO done, error Atomic transactions should map in the ATMU to the DDR memory controller, or the behavior is boundedly undefined. 5 Write class flush (with data) GSM M = 1, W = 1 or I = 1, store. M = 0 store type operation may also map to flush. N/A done, retry, error None atomic swap IO N/A The RapidIO controller does not support atomic swap transactions. None None None nwrite nwrite_r 6 Streaming write class swrite IO IO IO M = 0 store type operation M = 0 store type operation M = 0 store type operation N/A done, error N/A maintenance 8 Maintenance read class All M = 0 read type operation done, error Although possible, maintenance reads from memory locations outside of the RapidIO architectural memory space (0xC_0000– 0xC_FFFC) return unreliable data. None maintenance write 10 doorbell Doorbell class All Message passing M = 0 write type operation M = 0 write type operation done, error done, retry, error None MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 17-84 Freescale Semiconductor Functional Description Table 17-83. RapidIO Transaction Summary (continued) Packet Format 11 Message class 13 Response class RapidIO Transaction message Programming Model Message passing All Possible Processor Operations M = 0 write type operation Possible RapidIO Responses Comments done, retry, error None response N/A Depends on the request Table 17-84 summarizes packet formats. The following definitions and Figure 17-81 accompany the table. • • • • • • • • • • • • • • • • • • • • • • PMF—Physical maintenance field; prefixed to each packet format (see Figure 17-81) TYPE —Packet type (1, 2, 5, 6, 8, 10, 11, 13) TARID —target_id: the receiver’s ID within its domain SRCTID— Packet transaction ID TTYPE—Transaction type to be performed by the recipient RDSIZE—Read data size requested WRSIZE—Write data size requested SRCID—source_id: the packet sender’s ID within its domain ADDR—Address for the requested operation MSB—Most-significant byte LSB —Least-significant byte HOPCNT—hop_count selects the distance into the switch network to access CONFIG_OFFSET—Double-word offset into the CAR/CSR register file INFO—Doorbell information field; provided and used by software only MSGLEN—Total number of packets comprising the message MSGSEG—Portion of the message supplied by the packet SSIZE—Standard message packet data size MBOX—Recipient mailbox number at the target device LETTER—Identifies message within the mailbox STATUS—Packet response status field TARINFO—Target_info used for messages only. Field length is 8 bits: letter (2) + mbox (2) + msgseg (4) W—Word pointer; used in the decoding of the transaction size MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 17-85 RapidIO Interface S=0 1 AckID 3 0100 4 prio 2 TT = 00 2 Figure 17-81. Physical Maintenance Field (PMF) Definition Table 17-84. RapidIO Packet Format Summary (TT = 0b00) Type 2 Request To Home: (0010) TYPE PMF 12 4 Packet Format TARID 8 ADDRESS 29 SRCID 8 TTYPE 4 RDSIZE 4 W 1 SRCTID 8 ADDRESS 2 TTYPE:0010 (I/O read home); 0100 (nread); 1100 (atomic_inc); 1101 (atomic_dec); 1110 (atomic_set); 1111 (atomic_clr) 5 Write Class: (0101) PMF 12 TYPE 4 ADDRESS 29 TARID 8 W SRCID 8 ADDRESS 2 TTYPE 4 WRSIZE 4 SRCTID 8 MSB for DW0 --> LSB for DWn 1 TTYPE:0001 (flush); 0010–0011 (Reserved); 0100 (nwrite); 0101 (nwrite_r); 0110–0111 (Reserved); MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 17-86 Freescale Semiconductor Functional Description Table 17-84. RapidIO Packet Format Summary (TT = 0b00) (continued) Type 6 Streaming-Write Class: (0110) PMF TYPE 12 4 ADDRESS 1 2 Packet Format TARID 8 SRCID 8 ADDRESS 29 MSB for DW0 --> LSB for DWn 8 Maintenance Class Request: (1000) PMF TYPE 12 HOPCNT 8 4 TARID 8 SRCID 8 W 1 2 TTYPE 4 RDSIZE 4 SRCTID 8 CONFIG_OFFSET 21 MSB for DW0 --> LSB for DWn TTYPE: 0000 (read); 0001 (write): 0100 (port-write) Maintenance Class Response: PMF 12 TARTID 8 TYPE 4 HOPCNT 8 24 TARID 8 SRCID 8 TTYPE 4 STATUS 4 MSB for DW0 --> LSB for DWn TTYPE: 0010 (read response); 0011 (write response) STATUS: 0000 (done); 0111 (error); 10 Doorbell Class: (1010) PMF TYPE 12 4 TARID 8 SRCID 8 INFO 16 8 SRCTID 8 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 17-87 RapidIO Interface Table 17-84. RapidIO Packet Format Summary (TT = 0b00) (continued) Type 11 Message Class: (1011) PMF TYPE 12 MBOX 2 4 MSGSEG 4 Packet Format TARID 8 SRCID 8 MSGLEN 4 SSIZE 4 LETTER 2 MSB for DW0 --> LSB for DWn 13 Response Class: (1101) PMF TYPE 12 TARTID 8 4 TARID 8 SRCID 8 TTYPE 4 STATUS 4 MSB for DW0 --> LSB for DWn STATUS: 0000 (done); 0001 (data only); 0011 (retry); 0101 (done intervention); 0110 (Reserved); 0111 (error); 1000–1011 (Reserved); 1100–1111 (User) TTYPE: 0000 (request resp); 0001 (message resp); 1000 (request resp with data) Table 17-85 summarizes the defined control symbols. Implementation details for this particular RapidIO controller are noted in the comments section of the table. Table 17-85. Control Symbol Formats Symbol Type 0 (000) Packet accepted: S = 1 TgtAckID 1 1 (001) 3 00000 5 BUF_STATUS 000 4 3 None 00000 5 0000 4 001 3 Cause 000 Encountered internal error 001 Received unexpected ackID on packet 010 Uncorrectable error on control symbol 011 Input port has been stopped 100 Received bad CRC on packet 111 General error Symbol Format Comments The RapidIO controller sets buf_status to 0xF when it generates a packet accepted control symbol. Packet retry: S = 1 TgtAckID 1 3 2 (010) Packet not accepted: S = 1 TgtAckID 1 3 00000 5 1 1 CAUSE 3 010 3 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 17-88 Freescale Semiconductor RapidIO Functionality Table 17-85. Control Symbol Formats (continued) Symbol Type 4 (100) Packet control: S = 1 SUB_TYPE 1 3 00000 5 CONTENTS 4 100 3 Symbol Format Comments Sub_type (contents) 000 Idle (The RapidIO controller sets to 1111) 001 Stomp (unused) 010 EOP (The RapidIO controller sets to 1111) 011 Restart From Retry (unused) 100 Throttle (number of aligned pacing idles) 101 TOD-sync (number of max sized packets) The RapidIO controller sets buf_status to 0xF when it generates a link request command. 000 Send training 010 Override width 011 Reset 100 Input status Note that the RapidIO controller causes a checkstop if a software-initiated send training link request command collides with an inbound send training link request which is immediately followed by an inbound training pattern. Ack status 000 Expecting ackID 0 001 Expecting ackID 1 .... 111 Expecting ackID 7 Link status 0000 Un-initialized 0001 Initializing 0010 Error 0011 Stopped 0100 Retry-stopped 0101 Error-stopped 1000 Valid, ackID 0 .... 111 Valid, ackID 7 5 (101) Link request: S=1 1 CMD 3 00000 5 BUF_STATUS 100 4 3 6 (110) Link response: S = 1 ACK_STAT 1 3 00000 5 LINK_STAT 4 100 3 17.5 RapidIO Functionality This section describes RapidIO general device functionality, common transport layer functionality, and logical layer functionality. 17.5.1 General Functionality Lists The general device functionality lists are applicable to devices of all classes, including switch devices without end points. The following tables list the functions that the RapidIO controller supports. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 17-89 RapidIO Interface Table 17-86 lists general requirements for a RapidIO-compliant device. Table 17-86. General Device Functionality List Item No. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 Item Functionality All device generated packets adhere to the bit streams defined in the appropriate logical, transport, and physical layer specifications. All device-generated control symbols adhere to the bit streams defined in the physical layer specification. Device assigns reserved packet fields to logic 0s Device does not use reserved packet formats Reserved packet field contents are ignored. Device never assigns reserved encodings for packet fields Implementation-defined packet fields are ignored unless the function is understood by the receiving device. Received reserved packet field encodings are ignored if it is not necessary for the field to be defined for the requested transaction. Reads to reserved CAR bits return logic 0s. Writes to reserved CAR bits are ignored. Reads to implementation-defined CAR bits return the implementation defined value. Writes to implementation-defined CAR bits are ignored. Reads to reserved CARs do not cause an error. Reads to reserved CARs return logic 0s when read. Writes to reserved CARs do not cause an error. Writes to reserved CARs are ignored. Reads to reserved CSRs and extended features register bits return logic 0s. Writes to reserved CSRs and extended features register bits are ignored. Reads to implementation-defined CSR and extended features register bits return the implementation defined value. Writes to implementation-defined CSR and extended features register bits are determined by the implementation. Reads to reserved CSRs and extended features registers do not cause an error. Reads to reserved CSRs and extended features registers return logic 0s when read. Writes to reserved CSRs and extended features registers do not cause an error. Writes to reserved CSRs and extended features registers are ignored. Reads and writes to implementation-defined registers and register bits are implementation-defined. Meets Reqs? Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 17-90 Freescale Semiconductor RapidIO Functionality 17.5.1.1 8/16 LP-LVDS Layer Functionality Lists The physical layer 8/16 LP-LVDS (low-power, low-voltage differential signaling) layer functionality is subdivided into individual lists. Table 17-87 lists requirements for basic 8/16 LP-LVDS physical layer functionality for the device. Note that for device ID-related functions, the device ID for the RapidIO controller is set by power-on reset signals as described in Section 4.4.3.11, “RapidIO Device ID.” Table 17-87. General Device 8/16 LP-LVDS Physical Layer Basic Functionality List Item No. 1 2 3 4 5 6 7 Item Functionality CLK signal(s) rising edge is on the 32-bit boundary CLK signal(s) toggle every clock (true free running clock). RIO_TFRAME/RIO_RFRAME signal only toggles on the 32-bit boundary RIO_TFRAME/RIO_RFRAME signal only toggles to mark the first byte of a control symbol or the first byte of a new packet RIO_RFRAME/RIO_TFRAME signal is associated with RIO_RCLK/RIO_TCLK Device meets target AC specifications Device follows the link initialization and training procedure 7A ACTION: Device responds to link-request/send-training control symbol with 256 training patterns followed by at least one idle control symbol 7B DETAIL: Device port is enabled to transmit or receive packets only after device is both sending and receiving idle control symbols 7C DETAIL: Training pattern is not embedded in a packet or used to terminate a packet 8 9 10 11 12 13 Training symbol is correct for port width 16-bit ports can function as 8-bit ports Control symbols are 16 bits followed by an inverted 16-bit copy. Undefined control symbols are treated as idle control symbols. Device responds to a throttle control symbol with a pacing idle control symbol within the specified 60 data ticks time Device follows native RapidIO ordering rules at flow level 1 13A. DETAIL: Read packets are transmitted at priority level 1. 13B. DETAIL: Write packets are transmitted at priority level 1. 13C. DETAIL: Response packets are transmitted at priority level 1, 2, or 3. 13D. DETAIL: Packets of the same priority level cannot pass each other. 13E. DETAIL: Packets of higher priority may pass packets of a lower priority level. Meets Reqs? Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 17-91 RapidIO Interface Table 17-87. General Device 8/16 LP-LVDS Physical Layer Basic Functionality List (continued) Item No. 14 Item Functionality All registers in the 8/16 LP-LVDS physical layer specification extended features data structure can be read and written. 14A. DETAIL: Port maintenance block header 0 CSR; reset value is implementation-dependent (read only) 14B. DETAIL Port maintenance block header 0 CSR has the proper Extended Features block ID 14C. DETAIL: Port link time-out control CSR; reset value is 0xFFFF_FFFF 14D. DETAIL: Port response time-out control CSR; reset value is 0xFFFF_FFFF (end point devices only) 14E. DETAIL: Port general control CSR; reset value is implementation dependent 14F. DETAIL: Port 0 error and status CSR; reset value is 0x0000_0001 14G. DETAIL: Port 0 control CSR; reset value is implementation dependent 15 Device follows system initialization and exploration inter-operability procedure 15A. ACTION: End point agent device that contains a boot ROM is initially assigned base device ID = 0xFE at power-on reset 15B. ACTION: End point agent device without a boot ROM is initially assigned base device ID = 0xFF at power-on reset 15C. ACTION: System host device is initially assigned a non-reserved base device ID at power-on reset (end point device only) 15D. ACTION: Host device sets PGCCSR[D] at power-on reset (end point device only) 15E. ACTION: Agent device resets PGCCSR[D] at power-on reset (end point device only) 15F. ACTION: Host device sets the master enable bit (PGCCSR[M]) at power-on reset (end point device only) 15G. ACTION: Agent device resets the master enable bit (PGCCSR[M]) at power-on reset (end point device only) 15H. ACTION: End point device responds to all received maintenance request packets 15I. ACTION: Device with switch functionality defaults route information to allow boot ROM access by system host 15J. ACTION: Switch device sends maintenance responses that it generates to the port that the maintenance request was received on Meets Reqs? Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 17-92 Freescale Semiconductor RapidIO Functionality Table 17-88 lists requirements for link maintenance functionality and the compliance of the RapidIO controller. Table 17-88. General Device 8/16 LP-LVDS Physical Layer Link Maintenance List Item No. 1 2 Item Functionality Link-request/input-status control symbol causes a link-response control symbol Link-response control symbol link_status field indicates input port status 2A. DETAIL: Uninitialized means disabled 2B. DETAIL: Initializing means following the training sequence. 2C. DETAIL: Error means unrecoverable problem encountered; no packets are accepted. 2D. DETAIL: Retry-stopped means that input port is stopped due to a retry; no packets are accepted until receiving a restart-from-retry or a restart-from-error control symbol. 2E. DETAIL: Error-stopped means that input port is stopped due to a transmission error; no packets are accepted until receiving a restart-from-error control symbol. 2F. DETAIL: Stopped means that input is stopped for some other (not retry or transmission error) reason; no packets are accepted. 2G. DETAIL: OK means that the input port is operating and can communicate with the connected device. 3 4 5 Link-request/send-training control symbol causes 256 training patterns to be sent Link-request/reset control symbol causes the device to reset after four requests Link-request/input-status, link-response control symbol pair forces completion of all preceding activity 5A. DETAIL: All preceding packets have generated acknowledge control symbol if applicable. 5B. DETAIL: Returned link_status correctly indicates next expected ackID 6 Device does not issue more than one outstanding link-request control symbol (except in case of time-out) 6A. DETAIL: Device holds off another link-request control symbol until expected link-response control symbol is received Meets Reqs? Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Table 17-89 lists requirements for the device while transmitting a packet (a general device is required to generate response packets) and the compliance of the RapidIO controller. Table 17-89. General Device 8/16 LP-LVDS Physical Layer Packet Transmission List Item No. 1 2 3 4 Item Functionality Device cannot issue more than 8 unacknowledged packets AckIDs are always issued sequentially unless a retry or error causes the device to back up. Start with ackID = 0 after reset. Physical layer 3 reserved bits are cleared. Meets Reqs? Y Y Y Y MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 17-93 RapidIO Interface Table 17-89. General Device 8/16 LP-LVDS Physical Layer Packet Transmission List (continued) Item No. 5 6 7 8 9 10 11 12 13 Item Functionality A response packet for a request packet never transmitted before the acknowledge control symbol for the request packet. Embedded CRC properly inserted Packets that are not aligned to the 32-bit boundary are padded into alignment with 16 bits of logic 0. CRC values are correctly calculated. Switch devices preserve error coverage internally. Packets are completed normally with a new packet or EOP control symbol. Packets are canceled using stomp, restart-from-retry, or link-request control symbols. Stomp and restart-from-retry control symbols have the cause field set to all logic 0s. Device receiving packet-retry control symbol follows the defined retry behavior. 13A. ACTION: Moves from the port OK state to the output retry-stopped state when a packet-retry control symbol is received 13B. ACTION: Completes transmission of current packet if applicable; don’t transmit new packets 13C. ACTION: Sends a restart-from-retry control symbol to the receiving device 13D. ACTION: Moves from the output retry-stopped state to the port OK state upon completion of the output port internal retry recovery procedure 13E. ACTION: Begins retransmitting packets from the returned expected ackID value 13F. DETAIL: Retried packet is eventually re-transmitted (end point devices only; not true for cut-through switch devices) Meets Reqs? Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Table 17-90 lists the requirements for the device to receive a packet. Table 17-90. General Device 8/16 LP-LVDS Physical Layer Packet Reception List Item No. 1 2 3 4 5 6 7 8 AckIDs are only accepted sequentially. Starts with ackID = 0 after reset Every packet causes an acknowledge control symbol with the corresponding ackID field. Device sends acknowledge control symbols in same order as packets are received (ackIDs are issued sequentially) Device does not send a packet-accepted control symbol before the entire packet has been received and is error free (if error checking is enabled) Physical layer 3 reserved bits are ignored. Packets can be completed normally with a new packet or EOP control symbol. The cause field in stomp and restart-from-retry control symbols is ignored. Item Functionality Meets Reqs? Y Y Y Y Y Y Y Y MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 17-94 Freescale Semiconductor RapidIO Functionality Table 17-90. General Device 8/16 LP-LVDS Physical Layer Packet Reception List (continued) Item No. 9 Item Functionality Device can accept embedded control symbols 9A. DETAIL: Packet-accepted control symbol 9B. DETAIL: Packet-retry control symbol 9C. DETAIL: Packet-not-accepted control symbol 9D. DETAIL: Idle control symbol 9E. DETAIL: Throttle control symbol 9F. DETAIL: TOD-sync control symbol 9G. DETAIL: Link-response control symbol 10 11 TOD-sync control symbol is treated as idle control symbol if TOD-sync control symbol is not supported Packets can be canceled using stomp, restart-from-retry, or link-request control symbols. 11A ACTION: Device sends packet-retry control symbol for canceled packet (exception: do not send packet-retry control symbol while already in input retry-stopped or input error-stopped states) 12 Device encountering a packet retry situation follows the defined retry behavior 12A. DETAIL: Packet retry situation is an internal hazard or a canceled packet (exception: packet canceled with a link-request control symbol does not follow retry behavior) 12B. ACTION: Moves from the port OK state to the input retry-stopped state upon detection of a packet retry situation 12C. ACTION: Sends a packet-retry control symbol with the expected ackID value 12D. DETAIL: Device silently discards packets while in the input retry-stopped state 12E. DETAIL: Device detects control symbol errors while in input retry-stopped state 12F. ACTION: Moves from the input retry-stopped state to the port OK state upon reception of a restart-from-retry or a restart-from-error (link-request/input-status) control symbol 13 14 Accepting an input packet of a priority is not contingent on successfully transmitting a packet of a less than or equal priority A generated packet-not-accepted control symbol uses a defined cause field encoding. Meets Reqs? Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Recoverable errors are typically transmission errors or physical layer protocol violation errors, and the recovery algorithm can be used to maintain communications on the link without losing any data. A compliant device must recover from a single error. Recovery from multiple errors is not required. The device is always the target for error detection. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 17-95 RapidIO Interface Table 17-91 lists recoverable errors at the physical layer. Table 17-91. General Device 8/16 LP-LVDS Physical Layer Recoverable Errors List Item No. 1 Item Functionality Requested training pattern is embedded in a packet or used to terminate a packet 1A. ACTION: Moves from the port OK state to the input error-stopped state 1B. ACTION: Sends a packet-not-accepted control symbol to the sending device 1C. ACTION: Device silently discards all packets until a restart-from-error (link-request/input-status) control symbol is encountered 1D: ACTION: Moves from the input error-stopped state to the port OK state 2 Device encounters S-bit parity failure 2A. ACTION: Moves from the port OK state to the input error-stopped state 2B. ACTION: Sends a packet-not-accepted control symbol to the sending device 2C. ACTION: Device silently discards all packets until a restart-from-error (link-request/input-status) control symbol is encountered 2D: ACTION: Moves from the input error-stopped state to the port OK state 3 Device encounters incorrect CRC for a packet 3A. ACTION: Moves from the port OK state to the input error-stopped state 3B. ACTION: Sends a packet-not-accepted control symbol to the sending device 3C. ACTION: Device silently discards all packets until a restart-from-error (link-request/input-status) control symbol is encountered 3D: ACTION: Moves from the input error-stopped state to the port OK state 4 Device encounters unexpected ackID value in a packet 4A. ACTION: Moves from the port OK state to the input error-stopped state 4B. ACTION: Sends a packet-not-accepted control symbol to the sending device 4C. ACTION: Device silently discards all packets until a restart-from-error (link-request/input-status) control symbol is encountered 4D: ACTION: Moves from the input error-stopped state to the port OK state 5 Device encounters a packet that exceeds the 256 byte maximum packet size limit 5A. ACTION: Moves from the port OK state to the input error-stopped state 5B. ACTION: Sends a packet-not-accepted control symbol to the sending device 5C. ACTION: Device silently discards all packets until a restart-from-error (link-request/input-status) control symbol is encountered 5D: ACTION: Moves from the input error-stopped state to the port OK state Meets Reqs? Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 17-96 Freescale Semiconductor RapidIO Functionality Table 17-91. General Device 8/16 LP-LVDS Physical Layer Recoverable Errors List (continued) Item No. 6 Item Functionality Device encounters a corrupt control symbol 6A. ACTION: Moves from the port OK state to the input error-stopped state 6B. ACTION: Sends a packet-not-accepted control symbol to the sending device 6C. ACTION: Device silently discards all packets until a restart-from-error (link-request/input-status) control symbol is encountered 6D: ACTION: Moves from the input error-stopped state to the port OK state 7 Device detects protocol violations 7A. DETAIL: Any acknowledge control symbol with an unexpected ackID value 7B. DETAIL: Unsolicited acknowledge control symbol 7C. DETAIL: Received packet-accepted control symbol before packet transmission has completed 7D. DETAIL: Received link-request control symbol before sending link-response control symbol for previous link-response/input-status control symbol 7E. DETAIL: Unexpected link-response control symbol 7F. DETAIL: Unexpected restart-from-retry control symbol 7G. DETAIL: Unexpected training pattern 7H. DETAIL: Unexpected stomp control symbol 7I. DETAIL: Unexpected EOP control symbol 8 Device encounters a protocol violation. 8A. ACTION: Moves from the port OK state to the output error-stopped state. 8B.ACTION: Device stops transmitting new packets 8C. ACTION: Device sends link-request/input-status (restart-from-error) control symbol if no link-request control symbol is already outstanding (must wait for previous one to complete) and waits for link-response control symbol 8D. ACTION: Moves from the output error-stopped state to the port OK state 8E. ACTION: Starts re-transmitting at the ackID value returned with the received link-response control symbol 9 Device encounters a packet-not-accepted control symbol 9A. ACTION: Moves from the port OK state to the output error-stopped state 9B.ACTION: Device stops transmitting new packets 9C. ACTION: Device sends link-request/input-status (restart-from-error) control symbol if no link-request control symbol is already outstanding (must wait for previous one to complete) and waits for link-response control symbol 9D. ACTION: Moves from the output error-stopped state to the port OK state 9E. ACTION: Starts re-transmitting at the ackID value returned with the received link-response control symbol Meets Reqs? Y Y Y Y Y Y Y Y N Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 17-97 RapidIO Interface Table 17-91. General Device 8/16 LP-LVDS Physical Layer Recoverable Errors List (continued) Item No. 10 Item Functionality Device encounters an acknowledge control symbol time-out 10A. ACTION: Moves from the port OK state to the output error-stopped state 10B. ACTION: Device stops transmitting new packets 10C. ACTION: Device sends link-request/input-status (restart-from-error) control symbol if no link-request control symbol is already outstanding (must wait for previous one to complete) and waits for link-response control symbol 10D. ACTION: Moves from the output error-stopped state to the port OK state 10E. ACTION: Starts re-transmitting at the ackID value returned with the received link-response control symbol 11 Device encounters any transmission error during error recovery Meets Reqs? Y Y Y Y Y Y Y Relatively few nonrecoverable errors occur at the physical layer. Behavior of the interface is not defined for these events. Table 17-92 lists nonrecoverable errors at the physical layer. Table 17-92. General Device 8/16 LP-LVDS Physical Layer Nonrecoverable Errors List Item No. 1 2 3 4 5 6 Item Functionality Device does not respond to link-request/send-training control symbol with 256 training patterns followed by at least one idle control symbol Late or no response to throttle request Device is attempting error recovery and ackID received in link-response control symbol does not make sense (device cannot complete recovery) Time-out between link-request control symbol and link-response control symbol Device port transmits packets before device is both sending and receiving idle control symbols Any transmission error encountered during error recovery Meets Reqs? Y N Y Y Y Y 17.5.2 Common Transport Layer Functionality Lists The common transport layer functionality is subdivided into individual lists. Table 17-93 lists the basic functionality for the device at the common transport layer and the compliance of the RapidIO controller. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 17-98 Freescale Semiconductor RapidIO Functionality Table 17-93. General Device Common Transport Layer Basic Functionality List Item No. 1 Item Functionality All necessary CSRs (command and status registers) exist and can be read. 1A. DETAIL: Base device ID CSR; reset value is application and implementation dependent (end point devices only) 1B. DETAIL: Host base device ID lock CSR; Host_base_deviceID field reset value is 0xFFFF 1C. DETAIL: Component tag CSR; reset value is 0x0000_0000 2 Writable CSR fields can be written Meets Reqs? Y Y Y Y Y Table 17-94 lists the functionality for this controller in transmitting a packet (a slave device is required to generate response packets). Table 17-94. General Device Common Transport Layer Packet Transmission List Item No. 1 2 3 Item Functionality TT field is always logic 0s (small device ID fields) End point-free switch responds to maintenance requests with the c_count = 0 when received, otherwise decrement hop_count and route to proper output port Maintenance response packets are issued with the hop_count field set to 0xFF Meets Reqs? Y Y Y Table 17-95 lists the functionality for this controller in receiving a packet. Table 17-95. General Device Common Transport Layer Packet Reception List Item No. 1 Item Functionality Response packets are sent with the target and destination fields reversed from the corresponding request packet Meets Reqs? Y Logical layer errors are generally not recoverable in hardware. They may be recoverable in software. The device is always the target of a packet for error detection. Table 17-96 lists detectable errors. Table 17-96. General Device Common Transport Layer Detectable Errors List Item No. 1 Item Functionality Received reserved TT field encoding Meets Reqs? Y MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 17-99 RapidIO Interface 17.5.3 Logical Layer Functionality Lists The logical layer functionality is subdivided into individual lists. Table 17-97 lists the requirements for basic device functionality and the compliance of this controller. A class 1 device is always the target of a request transaction. Table 17-97. General Device Logical Layer Basic Functionality List Item No. 1 2 3 4 5 6 7 8 9 Item Functionality The address field of a packet is a double-word (8 byte) aligned address. Multiple double-word data payloads are linear starting at the specified address. Multiple double-word data payloads are aligned to a double-word boundary. Multiple double-word data payloads are not required to be aligned to the transfer size boundary. Sub-double-word data payloads have a defined data payload, properly aligned and padded to a double-word boundary. Response packets have the transaction ID of the associated request packet. Responses that are not expected to have a data payload must not have a data payload. Responses may not contain a data payload if the response status is ERROR. All necessary CARs (capabilities registers) exist and can be read. 9A. DETAIL: Device identity CAR; value is vendor- and implementation-dependent. 9B. DETAIL: Device information CAR; value is vendor- and implementation-dependent 9C. DETAIL: Assembly identity CAR; value is vendor- and implementation-dependent 9D. DETAIL: Assembly information CAR; value is vendor- and implementation-dependent 9E. DETAIL: Processing element features CAR; value is implementation-dependent; must indicate support for 34-bit address format packets 9F. DETAIL: Switch port information CAR; value is implementation-dependent (switch devices only) 9G. DETAIL: Source operations CAR; value is implementation-dependent (end point devices only) 9H. DETAIL: Destination operations CAR; value is implementation-dependent (end point devices only). 10 All necessary CSRs (command and status registers) exist and can be read. 10A. DETAIL: Processing element logical layer control CSR: reset value of the extended addressing control field is 0b001 (end point devices only) 11 12 13 14 Writable CSR fields can be written. All registers in the extended features data structure are double-word (8 byte) aligned (the three lsbs of the extended features pointers must be logic zeros). All extended features blocks lie within the extended features space in the register address map. Extended features list is terminated with an extended features pointer value of logic zeros Meets Reqs? Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 17-100 Freescale Semiconductor RapidIO Functionality Table 17-98 lists requirements for this controller when functioning as a transaction target. Table 17-98. General Device Logical Layer Target Transaction Support List Item No. 1 Maintenance read transaction 1A. DETAIL: Maintenance read request may be for 4 bytes 1B. DETAIL: Maintenance read request generates a maintenance read response 1C. DETAIL: Maintenance read response data payload is 4 bytes 2 Maintenance write transaction 2A. DETAIL: Maintenance write request may be for 4 bytes 2B. DETAIL: Maintenance write request generates a maintenance write responses 2C. DETAIL: Maintenance write response does not contain a data payload Item Functionality Meets Reqs? Y Y Y Y Y Y Y Y 17.5.3.1 Logical Layer Source Transaction Support List General functionality does not require a device to initiate any operations. Logical layer errors are generally not recoverable in hardware. They may be recoverable in software. The device is always the target of a packet for error detection. Table 17-99 lists the errors detected by the RapidIO controller. Table 17-99. General Device Logical Layer Detectable Errors List Item No. 1 2 3 4 5 6 Item Functionality Maintenance write packet data payload exceeds size specified in WRSIZE field Maintenance write packet data payload exceeds supported size Received maintenance read packet has a data payload Received maintenance write packet has no data payload Received maintenance packet uses a reserved field encoding for a required field Received maintenance packet uses illegal combinations of field encodings Meets Reqs? Y Y Y Y Y Y 17.5.3.2 Logical Layer Extended Functionality The RapidIO controller supports I/O read and flush with data GSM transactions, and all I/O and message transactions as described by the Parallel RapidIO Interconnect Specification except for atomic test and swap and atomic swap transactions. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 17-101 RapidIO Interface 17.6 RapidIO Errors RapidIO errors are classified under three categories: • Recoverable errors—The RapidIO controller supports hardware error detection and recovery mechanism as specified in the Parallel RapidIO Interconnect Specification. The RapidIO controller detects and attempts to recover from corrupt packet and control symbol errors and general protocol errors. In these cases, the appropriate bit is set in the RapidIO recoverable error detect register. Logging of information in capture registers may or may not take place. See Table 17-100 for specifics regarding the RapidIO controller. Notification errors—The RapidIO controller detects many hardware nonrecoverable errors that are non-fatal such as exceeded threshold counts (as defined in PERTR and PRTR). Upon detection of a notification error the appropriate bit is set in the port notification/fatal error detect register (PNFEDR) and an interrupt is generated if enabled in the port notification/fatal error interrupt enable register (PNFEIER). Hardware continues to operate after notification errors. Fatal errors—The RapidIO controller also detects a number of fatal errors that are hardware nonrecoverable, such as time-outs. Upon detection of a fatal error, relevant information may be logged (PECSR[V] indicates successful capture) in the error packet capture registers (EPCRn) if the appropriate error detect is enabled (PNFEDiR). An interrupt is generated if the appropriate fatal error interrupt is enabled in the port notification/fatal error interrupt enable register (PNFEIER). The RapidIO hardware either goes into training mode, or remains in the OK state based on the error type. • • If the valid bit of the port error capture status register (PECSR[V]) is set, the information in error packet capture registers EPCR0–EPCR3 is valid: Fatal errors caused by a bad link invoke training in an attempt to recover the link. If communication is not re-established, a RESET is required to reconcile RapidIO’s internal queues to resume normal operation. Table 17-100 lists the recoverable errors detected by the RapidIO implementation and the resulting actions taken. Table 17-100. Recoverable Errors Detected by RapidIO Controller Error Inbound corrupt control symbol Description Received a control symbol in which true does not match complement counterpart Action Sends “PACKET_NOT_ ACCEPTED” ack with “Error on control symbol” cause Sends “PACKET_NOT_ ACCEPTED” ack with “S-bit parity error” cause Capture Registers No information logged in capture register(s) No information logged in capture register(s) No information logged in capture register(s) Inbound packet/control Received a packet or a control symbol with S-bit parity error symbol with an S-bit parity error Inbound out-of-sequence ack Received an ack control symbol with Starts link recovery process symbol an unexpected ackID MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 17-102 Freescale Semiconductor RapidIO Errors Table 17-100. Recoverable Errors Detected by RapidIO Controller (continued) Error Inbound packet with unexpected ackID Description Action Capture Registers No information logged in capture register(s) Received packet with unexpected Sends “PACKET_NOT_ ackID value (out-of-sequence ackID) ACCEPTED” ack with “Received unexpected ackID with packet” cause Received packet with a bad CRC value Requested training pattern is embedded in a packet Sends “PACKET_NOT_ ACCEPTED” ack with “Bad CRC” cause Sends “PACKET_NOT_ ACCEPTED” ack with “General error” cause Inbound packet with bad CRC Embedded training No information logged in capture register(s) No information logged in capture register(s) No information logged in capture register(s) Inbound packet exceeds 256 Received packet that exceeds the Sends “PACKET_NOT_ bytes maximum allowed size by the ACCEPTED” ack with Parallel RapidIO Interconnect “General error” cause Specification. This controller performs this function by detecting a data payload greater than 256 bytes. Packet-not-accepted: Encountered internal error Received ack control symbol with packet-not-accepted of type “encountered internal error” Starts link recovery process. No information logged in capture register(s) No information logged in capture register(s) No information logged in capture register(s) No information logged in capture register(s) No information logged in capture register(s) No information logged in capture register(s) No information logged in capture register(s) Packet-not-accepted: Received ack control symbol with Unexpected ackID on packet packet-not-accepted of type “unexpected ackID” Packet-not-accepted: Corrupt control symbol Packet-not-accepted: Input port stopped Packet-not-accepted: Bad CRC Packet-not-accepted: S-bit parity error Packet-not-accepted: general error Starts link recovery process. Starts link recovery process. Received ack control symbol with packet-not-accepted of type “corrupt control symbol” Received ack control symbol with packet-not-accepted of type “input port stopped” Received ack control symbol with packet-not-accepted of type “Bad CRC” Received ack control symbol with packet-not-accepted of type “S-bit parity error” Starts link recovery process. Starts link recovery process. Starts link recovery process. Starts link recovery process. Received ack control symbol with packet-not-accepted of type “General error”. This error type could be created because of header/data problems at the receiver. Received a restart-from-retry control Starts link recovery process. symbol while in the “OK” state Starts link recovery process Received an ack control symbol while no packets are outstanding on the ack history queue Unexpected restart-from-retry symbol Unsolicited ack symbol No information logged in capture register(s) No information logged in capture register(s) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 17-103 RapidIO Interface Table 17-100. Recoverable Errors Detected by RapidIO Controller (continued) Error Ack before restart-from-retry Description Action Capture Registers No information logged in capture register(s) No information logged in capture register(s) No information logged in capture register(s) No information logged in capture register(s) No information logged in capture register(s) No information logged in capture register(s) Received an ack control symbol after Starts link recovery process receiving an ack retry and before sending a restart-from-retry Received a training symbol while in the “OK state” Received a stomp control symbol while there is no packet being received Starts link recovery process Starts link recovery process Unexpected training symbol Unexpected stomp symbol Unexpected link response symbol Unexpected EOP symbol Starts link recovery process Received a link response control symbol while there is no outstanding request. Received an EOP control symbol while there is no packet being received Received a link request control symbol before the previous link request has been serviced. Starts link recovery process Link request error Starts link recovery process Frame toggle alignment No information logged in Received FRAME signal toggles at a Sends non 32-bit boundary “PACKET_NOT_ACCEPTED capture register(s) ” ack with “General error” type Received FRAME signal toggles on the negative edge of the clock No information logged in Sends “PACKET_NOT_ACCEPTED capture register(s) ” ack with “General error” type No information logged in capture register(s) Frame toggle edge Ack time-out An ack control symbol is not received Starts link recovery process within the specified time-out interval. Table 17-101 lists the notification errors detected by the RapidIO implementation and the resulting actions taken. Table 17-101. Notification Errors Detected by RapidIO Controller Error Error recovery threshold error Retry threshold error Received request ERROR response Description Action Capture Registers No information logged in capture register(s) No information logged in capture register(s) Information logged in SYSINT capture registers. Error recovery threshold count (defined Generates interrupt if enabled in PERTR[RCTT]) exceeded (PNFEIER[ETIE] is cleared) Consecutive ack retry threshold count (defined in PRTR[RTT]) exceeded Received a response of type error for anything other than a message Generates interrupt if enabled (PNFEIER[RTIE] is cleared) Generates interrupt if enabled (PNFEIER[RERIE] is cleared) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 17-104 Freescale Semiconductor RapidIO Errors Table 17-102 lists the fatal errors detected by the RapidIO implementation and the resulting actions taken. Table 17-102. Fatal Errors Detected by RapidIO Controller Error Link response time-out Packet response time-out Missing idle after training Message segment error Duplicate message segment Message length error Nonsensical ackID Description Action Capture Registers No information logged in capture register(s) Resume training on the interface. No information logged in capture registers. No Information logged in capture register(s) No information logged in capture register(s) No information logged in capture register(s) No information logged in capture register(s) No information logged in capture register(s). Resume training on the interface. HRESET required to resume normal operation. Information logged in RapidIO capture register(s) Information logged in RapidIO capture register(s) Information logged in RapidIO capture register(s) Information logged in RapidIO capture register(s) Information logged in RapidIO capture register(s) Information logged in RapidIO capture register(s) Information logged in RapidIO capture register(s) Information logged in RapidIO capture register(s) Information logged in RapidIO capture register(s) A link response is not received within the Generates interrupt if enabled specified time-out interval. (PNFEIER[LTIE] is cleared) A packet response is not received within Generates interrupt if enabled the specified time-out interval. (PNFEIER[RSTIE] is cleared) Idle not received after a requested training sequence completes Received a message packet with the segment field greater than the message length field (msgseg > msglen) Generates interrupt if enabled (PNFEIER[NTIIIE] is cleared) Generates interrupt if enabled (PNFEIER[SIE] is cleared) Received a duplicate message segment Generates interrupt if enabled from RapidIO (PNFEIER[DSIE] is cleared) Received a message packet from RapidIO with a bad message length AckID received with link response does not make sense Generates interrupt if enabled (PNFEIER[BMIE] is cleared) Generates interrupt if enabled (PNFEIER[NAIE] is cleared) Illegal request fields Illegal combinations of fields in the request packet. Examples include: illegal transaction types, illegal sizes Unexpected request Read type with a data payload data Unsupported request Unsupported packet format type, or transport type (TT bits) Generates interrupt if enabled (PNFEIER[IRFIE] is cleared) Generates interrupt if enabled (PNFEIER[URDIE] is cleared) Generates interrupt if enabled (PNFEIER[URIE] is cleared) Generates interrupt if enabled (PNFEIER[URSIE] is cleared) Generates interrupt if enabled (PNFEIER[IWRIE] is cleared) Generates interrupt if enabled (PNFEIER[IRSFIE] is cleared) Unexpected request Packet data is not expected size size Illegal write request Illegal response fields Write types with no data payload. Illegal combinations of fields in the response packet Illegal response type Illegal response for a given request type Generates interrupt if enabled (PNFEIER[IRTIE] is cleared) Bad read response Bad write response Unexpected response data size Read “done” response with no data payload Write response with a data payload Response packet data is not expected size Generates interrupt if enabled (PNFEIER[BRRIE] is cleared) Generates interrupt if enabled (PNFEIER[BWRIE] is cleared) Generates interrupt if enabled Information logged in (PNFEIER[URSSIE] is cleared) RapidIO capture register(s) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 17-105 RapidIO Interface Table 17-102. Fatal Errors Detected by RapidIO Controller (continued) Error Unsolicited response without data Unsolicited response with data Message error response Message descriptor fetch error Message size error Description Action Capture Registers A response without data is received while Generates interrupt if enabled Information logged in (PNFEIER[URESIE] is cleared) RapidIO capture register(s) there is no corresponding request that matches that TID A response with data is received while there is no corresponding request that matches that TID Generates interrupt if enabled Information logged in (PNFEIER[URSDIE] is cleared) RapidIO capture register(s) Information logged in RapidIO capture register(s) No information logged in RapidIO capture register(s) Information logged in RapidIO capture register(s) Received a response of type error for an Generates interrupt if enabled outbound message packet (PNFEIER[MERIE] is cleared) A message descriptor fetch from local memory gets an error (for example, ECC). Generates interrupt if enabled (PNFEIER[MDFIE] is cleared) Received message packet data payload Generates interrupt if enabled (PNFEIER[MSIE] is cleared) is not of the size specified in the ssize field (with the exception of the last packet which may be less) Received a packet whose target ID does Generates interrupt if enabled (PNFEIER[ITIE] is cleared) not match RapidIO deviceID while accept_all mode is disabled Illegal transaction target error Information logged in RapidIO capture register(s) 17.7 ATMU (Address Translation and Mapping Unit) Outbound address translation refers to the translation of an address from the local address space to that of RapidIO. In the same context, inbound address translation and mapping refers to the translation of an address from the external address space of RapidIO to the local address space understood by the internal interfaces. Although, the Parallel RapidIO Interconnect Specification allows for 48- and 64-bit addresses, this implementation only supports 34-bit RapidIO addresses. Outbound and inbound window misses use the window 0 register set by default. The size field in the attributes register of window 0 for both inbound and outbound translation is used to determine how much of the translation address is used when the default window is selected. Overlapping window hits result in the use of the lowest number window register set hit. For both inbound and outbound translation, the smallest window size is 4 Kbytes and the largest is 4 Gbytes. 17.7.1 Outbound ATMU Translation The nine outbound ATMU windows (0–8) perform the mapping from the internal 32-bit address space to the 34-bit address space of RapidIO. They also map attributes such as transaction type and priority level. Window 0 is always enabled and is used as the default window if the address does not match one of the other eight windows. Overlapping outbound window hits result in the use of the lowest number window register set hit. Note that OCeaN read types can only map to MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 17-106 Freescale Semiconductor ATMU (Address Translation and Mapping Unit) RapidIO read types and OCeaN write types can only map to RapidIO write types including doorbells. 17.7.1.1 Outbound ATMU Bypass Mode Note that some transactions may not require ATMU translation because the address associated with the original transaction is already mapped to the appropriate address space. Although the DMA controller can initiate outbound RapidIO transactions that may or may not bypass ATMU translation, it is the only source of transactions that can bypass outbound ATMU translation. The DMA controller may bypass ATMU translation for any outbound transaction that it supports. 17.7.1.2 Outbound Special Transactions and Requirements Due to requirements of the Parallel RapidIO Interconnect Specification, I/O read home and flush with data transactions require special handling. Outbound I/O read transactions cannot cross a 32-byte cache-line boundary and cannot be greater than 32 bytes in size because the RapidIO interface requires wrapping around a 32-byte cache-line for I/O read home transactions. Therefore, regardless of the size of the original request, the address for outbound I/O read transactions is double-word aligned, and the size is set to 32 bytes in order to fetch an entire cache-line. There is also a requirement for outbound requests originated by the PCI controller. These transactions must map to an ATMU window in which the ROWARn specifies a transaction flow level of zero and has the PCI bit set. This causes the outbound transaction to be sent through the RapidIO interface with a priority of one because the transaction flow level is incremented when the PCI bit is set. To comply with the Parallel RapidIO Interconnect Specification other transactions also have special requirements. Outbound maintenance transactions, for example, may not exceed 64 bytes. Also, outbound doorbells are generated through the ATMU by using write transactions whose sizes are 2, 4 or 8 bytes and whose address is aligned to a doubleword boundary. A doorbell generated using a 2-byte write gets its INFO field from the data associated with the write transaction. A doorbell generated using a 4 or 8 byte write gets its INFO field from the most-significant 2 bytes of the data associated with the write transaction. Outbound swrite transactions must be at least a double word, and atomic transactions may only request 1, 2, or 4 bytes. Furthermore, atomic transactions must meed the RapidIO alignment requirements as specified in the Parallel RapidIO Interconnect Specification. The Parallel RapidIO Interconnect Specification should be referenced for any additional requirements that are not listed in this section. 17.7.2 Inbound ATMU Translation Inbound ATMU windows perform the address translation from the 34-bit external RapidIO address space to the 32-bit internal address space. Inbound ATMU windows also attach attributes, MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 17-107 RapidIO Interface transaction type, and the target interface to the transaction. The RapidIO controller has four inbound ATMU windows plus the default window 0. The default window is always enabled. Overlapping inbound window hits result in the use of the lowest number window register set hit. 17.7.2.1 Inbound ATMU LCSBA1CSR Window There is an additional inbound window that is used for inbound translation. It is a special condition for external accesses to LCSBA1CSR (local configuration space base address register) address space. The purpose of the LCSBA1CSR window is to allow external devices to access the LCSBA1CSR address space without knowledge of the internal memory map. When a remote device attempts to access the LCSBA1CSR register set (address hitting LCSBA1CSR window), the lower 20 bits are used as the offset to access the specific register in the LCSBA1CSR address space. The LCSBA1CSR window has the highest priority for inbound translation. If the inbound transaction’s address does not hit the LCSBA1CSR window or any of the four inbound ATMU windows (1–4), the translation attributes defined by the default window (window 0) are used. The LCSBA1CSR window does not have specific RIWBAR, RIWTAR or RIWAR registers. However, for the purposes of translation, bits 1–14 of the LCSBA1CSR correspond to bits 10–23 of RIWBARn, and bits 12–23 of the CCSRBAR correspond to bits 12–23 of RIWTARn. The hard-coded attributes for this window as they correspond to the RIWAR register fields are: EN = 0b1, TGINT = 0b1111, RDTYP and WRTYP = 0b0100 and SIZE = 0b010011. 17.7.2.2 Inbound ATMU Bypass Mode Just as some outbound transactions may bypass outbound ATMU translation, some inbound transactions may also bypass ATMU translation because the address associated with the original transaction is already mapped to the appropriate address space. Inbound messages, doorbells and port-writes go through a different form of address mapping and are not mapped through the ATMU. These transactions have the stash and cache_lock attributes cleared. Similarly, the no_snoop bit is set according to the corresponding programming model of the specific transaction type regardless of the value specified in the RIWAR of the selected ATMU window. For more information about these transactions, see Section 17.8.4.1, “Data Message Controller,” Section 17.8.4.4, “Doorbell Message Controller,” and Section 17.8.4.5, “Port-Write Controller Structure.” 17.7.2.3 Inbound Special Transactions and Requirements Just as with outbound translation, some inbound transactions require special handling to comply with the Parallel RapidIO Interconnect Specification. For example, inbound I/O read home and flush with data transactions require special handling. Snooping is enabled for these transactions regardless of the no-snoop value specified in the RIWAR of the selected window. Furthermore, these transactions should target local memory. Targeting any other interface is a programming MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 17-108 Freescale Semiconductor RapidIO Message Unit error and results in undefined behavior. Also, I/O read home transactions can request less data than the maximum of 32 bytes. Inbound reads that target the PCI controller should have a priority of zero as a special requirement. Similarly, inbound writes targeting the PCI interface must have a priority of one. Inbound nwrite_r transactions that target the PCI interface are not allowed. These requirements insure proper processing through the PCI interface. Inbound swrite transactions must not be smaller than a double-word, and atomic transactions may only request one, two or four bytes. Furthermore, atomic transactions must meet the RapidIO alignment requirements as specified in the Parallel RapidIO Interconnect Specification and must target the DDR memory controller. Inbound atomic transactions targeting interfaces other than the DDR memory controller result in boundedly undefined behavior. Also, all inbound maintenance transactions must have a size of four bytes. The Parallel RapidIO Interconnect Specification should be referenced for any additional requirements that are not listed in this section. 17.7.2.4 ATMU Boundary Crossing Errors During address translation and mapping, some transactions may generate an error condition and signal an interrupt. To prevent inbound and outbound transactions from corrupting local memory space, ATMU boundary crossing errors are detected. An error is detected when an inbound or outbound transaction has a data payload that crosses its selected ATMU window boundary. An error is also detected when the given transaction’s data payload crosses into the space of a higher-priority window. The settings of the port notification/fatal error register set described in Section 17.3.2.3, “Error Management Registers,” determine if the error condition is detected and if PNFEDR[AXE] is set. The settings in this register set also determine if an interrupt is signalled upon detection of the error condition. 17.8 RapidIO Message Unit This description is an extension of Section 17.7.1, “Outbound ATMU Translation,” and Section 17.7.2, “Inbound ATMU Translation.” It describes the operation of the data message and doorbell message controllers in the message unit. The message unit is compliant with the message passing logical specification in the Parallel RapidIO Interconnect Specification. 17.8.1 Overview The RapidIO message unit supports a message passing programming model for inter-processor and inter-device communication. This model enables a producer to send a message across the interconnect fabric to a consumer’s message hardware, called a mailbox. The receiving mailbox controller places the message in a queue located in main (DDR) memory. A message may consist MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 17-109 RapidIO Interface of one or more packets depending on the size of the message. When the entire message has been received, an interrupt (if enabled) is generated for the processor to process the message. Messages can be queued for transmission in the producer’s memory and the message hardware processes them sequentially. Messages can also be queued in the consumer’s memory while software processes them sequentially. Software controls the depth of the circular message queue in the producer, or consumer. The message unit also supports another form of message called a doorbell message. doorbell messages do not have a data payload and are basically used for interrupting the processor. Like data messages, doorbell messages can be queued in a consumer’s local memory while being processed by the processor core. When a doorbell is received, an interrupt (if enabled) is communicated to the processor. The most common use of the message passing model is in systems where a processing element can only access memory that is local to itself, and communication between processing elements is achieved through message passing and communication is address independent. There are two types of RapidIO messages. A data message may consist of up to 16 packets and each packet may contain up to 256 bytes of data. A doorbell message contains a small amount of software-defined information embedded in the packet header and never has a data payload. Message and doorbell controllers are controlled through run-time registers described in Section 17.3.3.1, “RapidIO Outbound Message Registers,” Section 17.3.3.2, “RapidIO Inbound Message Registers,” and Section 17.3.3.3, “RapidIO Doorbell Registers.” 17.8.2 Message Unit Features The message unit contains the following features: • • • • • • • • One inbound data message structure (inbox) One outbound data message structure (outbox) Support for chaining and direct modes in the outbox Support for up to 16 segments per message Support for up to 256 bytes per packet, and up to 4 Kbytes of data per message Support for one inbound doorbell message structure Support for receiving messages into any mailbox, any letter Support for transmitting messages from any mailbox MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 17-110 Freescale Semiconductor RapidIO Message Unit 17.8.3 Message Unit Modes of Operation The message unit operates in the following modes: • • Direct mode—Software is expected to program all the necessary registers for sending an outbound message. Chaining mode—A transfer descriptor that describes the message information is fetched from local memory before a message is sent. 17.8.4 RapidIO Messaging Description This section describes controller operations, formats, interrupts, and response conditions. 17.8.4.1 Data Message Controller The RapidIO controller supports a data message passing programming model for inter-processor and inter-device communication. This model enables a producer to send a data message across the interconnect fabric to a consumer’s message hardware, called a mailbox. The receiving mailbox controller places the message in a queue located in its memory. A data message may consist of one or more packets depending on the size of the message. When the entire message has been received, an interrupt is generated (if enabled) for the processor core to process the message. Data messages can be queued for transmission in the producer’s memory and the message hardware processes them sequentially. Messages can also be queued in the consumer’s memory while software processes them sequentially. Software controls the depth of the circular queue in the producer, or consumer. The RapidIO data message controller contains the following features: • • • • • • • One inbound mailbox (inbox) structure One outbound mailbox (outbox) structure Support for one active letter in the inbox (letters are assigned and maintained by hardware) Support for up to one letter (0) in the outbox for message unit initiated messages Support for up to 16 segments (packets) per message Support for up to 256 bytes per segment (packet) Chaining and direct mode support for outbound messages The following sections describe the structure and operation of the outbox and inbox hardware in the data message controller. 17.8.4.2 Outbox Controller Operation The outbox controller is responsible for sending messages stored in a circular queue in local memory. The outbox controller supports two modes of operation: direct and chaining mode. In MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 17-111 RapidIO Interface direct mode, software programs the necessary registers to point to the beginning of the message in memory. In chaining mode, software programs the necessary registers to point to the beginning of the first valid descriptor in memory. The hardware then reads the descriptor to load all the necessary registers to start the message transfer. 17.8.4.2.1 Direct Mode Operation In direct mode, OMR[MUTM] is set and the outbox controller does not read descriptors from memory, but instead uses the current parameters programmed in the outbox registers to start the transfer. In direct mode, software is responsible for initializing all the parameters in all the necessary registers to start the message transmission. The message transfer is started when the outbox start bit, OMR[MUS], in the outbound mode register transitions from a 0 to 1 and the outbox is not already busy. Software is expected to program all the appropriate registers before setting OMR[MUS]. The sequence of events to start and complete a transfer in direct mode is as follows: • • • Poll the status register message unit busy bit, OSR[MUB], to make sure the message unit is not busy with a previously initiated message. Initialize the source address (OSAR), destination port (ODPR), destination attributes (ODATR) and double-word count (ODCR) registers. Initialize the outbound mode register message unit transfer mode bit, OMR[MUTM] = 1, to indicate direct mode. Other control parameters must also be initialized in the mode register. Clear, then set the mode register message unit start bit, OMR[MUS], to start the message transfer. OSR[MUB] is set by the outbox controller to indicate that the message transfer is in progress. OSR[MUB] is cleared by the outbox controller after the transfer is finished or if a transfer error occurs. An end of message interrupt is generated if ODATR[EOMIE] is set. • • • • 17.8.4.2.2 Chaining Mode Operation In chaining mode, OMR[MUTM] = 0 in the outbound mode register, and message descriptors are built in local memory in a circular queue. Two options are available to the programmer in chaining mode: normal mode and list mode. In normal mode, software can initialize the outbox controller registers and then build the descriptors in memory. ODQEPAR is maintained and incremented by hardware, but the increment is controlled by software. Software should build descriptors in memory using the enqueue pointer address and can have the hardware increment the enqueue pointer by writing OMR[MUI] to a value of 1. Software can then read the value of the enqueue pointer to write the next descriptor in memory and then increment the enqueue pointer by writing MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 17-112 Freescale Semiconductor RapidIO Message Unit OMR[MUI] = 1 and repeat this process until the descriptor queue is full. Hardware performs the enqueue pointer address calculation, queue fullness and queue wrap checks and software writes new descriptors using the computed enqueue pointer address. In list mode, software can build one or more descriptors in memory before initializing the outbox controller registers. Software is responsible for maintaining and controlling ODQEPAR. Software is also responsible for keeping track of queue fullness and wrap conditions. OMR[MUI] is not used in this mode. The outbox message controller takes descriptors off the tail of the queue and processes them, incrementing ODQDPAR to point to the next descriptor in the queue and repeats this process until the enqueue and dequeue pointers are equal. Figure 17-82 shows a sample structure of the outbox portion of the outbox message controller, a descriptor list queue, and the message data queue. In this example, the descriptor list queue has eight entries, four of which are currently valid. The processor core adds descriptors to the head of the queue and the outbox message controller takes them off from the tail. Processor Core Enqueue and Tail Pointers Packets to RapidIO Interface Local Memory Local Memory Message Data Message Data Outbound Message Controller Enqueue pointer Dequeue pointer Desc Desc Desc Desc Outbound Descriptor List Queue Message Data Message Data Message Data Queue Figure 17-82. Outbound Message Queue Structure The sequence of events to start and complete a transfer by adding descriptors after initializing the message unit in normal chaining mode is as follows: • • Initialize ODQDPAR and ODQEPAR to the same value for proper operation. Initialize the outbound mode register message unit transfer mode bit, OMR[MUTM] = 0, to indicate chaining mode and set OMR[MUS] = 0 to disable the data message controller. Other control parameters must also be initialized in the mode register. Set OMR[MUS], which enables the outbox controller and causes ODQDPAR to be saved as the base address of the circular queue. The enqueue pointer is incremented when OMR[MUI] is set by software. If the enqueue and dequeue pointers are not equal, the descriptor fetch begins immediately from the address pointed to by ODQDPAR. If they are • MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 17-113 RapidIO Interface • • • • equal, the outbox waits until the enqueue and dequeue pointers are not equal. OMR[MUI] is cleared by hardware after successfully incrementing the enqueue pointer. OSR[MUB] is set by the message unit to indicate that the message transfer is in progress while the tail and enqueue pointers are not equal. The above process continues until the dequeue pointer equals the enqueue pointer again. OSR[MUB] is cleared by the message unit after finishing the transfer of the last descriptor segment, or if an error occurs during any of the transfers. Software can continue adding descriptors as needed for new transfers by setting OMR[MUI] as long as the circular queue in memory is not full while the message unit is busy. The sequence of events to start and complete the transfer of a list of message descriptors built in memory before initializing the message unit in list chaining mode is as follows: • • Poll the status register message unit busy bit, OSR[MUB], to make sure the outbox controller is not busy with a previously initiated message. Initialize ODQDPAR to point to the first descriptor in memory and ODQEPAR and EODQEPAR to point to the circular queue entry in memory following the last descriptor in the list. The enqueue and dequeue pointers should not be the same. Initialize the outbound mode register message unit transfer mode bit, OMR[MUTM] = 0, to indicate chaining mode and set OMR[MUS] = 0 to disable the data message controller. Other control parameters must also be initialized in the mode register. Set OMR[MUS], which enables the outbox controller. The descriptor fetch begins immediately from the address pointed to by ODQDPAR. OSR[MUB] is set by the message unit to indicate that the message transfer is in progress. OMR[MUI] is not used in this mode because software directly controls the enqueue pointer; therefore, new descriptors cannot be added when the message unit is busy as indicated by OSR[MUB] = 1. OSR[MUB] is cleared by the message unit after finishing the transfer of the last descriptor segment, or if an error occurs during any of the transfers. New descriptors can be added by just updating ODQEPAR as long as the enqueue pointer address does not cause a wrap condition. If adding new descriptors causes a wrap condition, software can build descriptors in memory but not update the enqueue pointer (ODQEPAR) address registers until the message unit is idle as indicated by OSR[MUB] = 0. Before updating the enqueue and dequeue pointers, software must disable the message unit by setting OMR[MUS] = 0. It can then update the enqueue (ODQEPAR) and dequeue (ODQDPAR) pointer address registers before enabling the message unit by setting OMR[MUS] = 1 to start transferring the new list of descriptors. • • • • • • • MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 17-114 Freescale Semiconductor RapidIO Message Unit Software must guarantee that descriptors are not added to an already full queue. This can be accomplished in one of two ways: • When OMR[QFIE] is set, hardware tracks queue fullness in normal chaining mode and reports an interrupt if the queue becomes full. After the interrupt bit is set, it is the responsibility of the software to clear the interrupt condition by setting OSR[QFI]. QFI is not cleared with the write if the queue is still full. When QFI is cleared, the queue is no longer full and a descriptor may be added to the queue. When OMR[QFIE] is cleared or if operating in list chaining mode, OSR[QFI] is not set by hardware on a queue full condition, and it is the responsibility of the software to calculate queue fullness before adding a new descriptor. Software can detect queue fullness by comparing the enqueue (ODQEPAR) and the dequeue pointer (ODQDPAR) and monitoring the queue busy bit (OSR[MUB]). The queue is full if the enqueue and dequeue pointers are equal and the queue busy bit is set. Ensure that the circular queue is not full by using either of the above methods. Write the descriptor to be enqueued to the 32 bytes pointed to by the enqueue pointer (ODQEPAR). Write OMR[MUI] = 1 keeping the same values of all the other bits in the OMR. A new descriptor may be added by incrementing the enqueue pointer if the queue is not full. • The following process adds a descriptor to the circular memory queue in normal mode: • • • • 17.8.4.2.3 Switching Between Direct and Chaining Modes The message unit architecture allows switching from direct mode to chaining mode and vice-versa after all the required parameters have been initialized in the appropriate registers and when the message unit is not busy with a current transfer as indicated by OSR[MUB] being cleared. When switching from direct mode to chaining mode, if OMR[MUS] is cleared and then set, the message unit is re-initialized to chaining mode and the outbound descriptor queue dequeue pointer address is saved as the new base address of the circular queue in memory. If this is not desired, OMR[MUS] should stay set when switching from direct mode to chaining mode. When the enqueue and dequeue pointers are not equal, the message unit begins the new descriptor fetch while retaining the same circular queue parameters as the previous chaining mode transfer. When switching from chaining mode to direct mode, OMR[MUS] must be cleared and then set. This has no effect on the circular queue parameters thereby saving the queue state for the next chaining mode transfer. 17.8.4.2.4 Descriptor Format Message descriptors contain information for the message unit controller to transfer data. Software must ensure that each descriptor is aligned on a 32-byte boundary. For each descriptor in the MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 17-115 RapidIO Interface queue, the message unit controller starts a new message operation with the control parameters specified by the descriptor. Table 17-103 describes the outbound message unit descriptor. Table 17-103. Outbound Message Unit Descriptor Summary Offset 0x00 0x04 0x08 Descriptor Field Reserved Source address Destination port — Source address of the message operation. After the message controller reads the descriptor from memory, this field is loaded into the source address register. Destination port of the message operation. After the message controller reads the descriptor from memory, this field is loaded into the destination port register. Description 0x0C Destination attributes Transaction attributes of the message operation. After the message controller reads the descriptor from memory, this field is loaded into the destination attributes register. 0x10 0x14 0x18 Reserved Reserved Double-word count — — Number of double words for the message operation. After the message controller reads the descriptor from memory, this field is loaded into the double-word count register. — 0x1C Reserved Figure 17-83 shows the queue dequeue pointer and an associated descriptor. The descriptor is only valid if the enqueue and dequeue pointers are not equal. Descriptor Dequeue Pointer Address Register 0 31 Offset 0x00 0x04 0x08 0x0C 0x10 0x14 0x18 0x1C 0 Local Memory Reserved Source Address Destination Port Destination Attributes Reserved Reserved Double-Word Count Reserved Descriptor 31 Figure 17-83. Descriptor Dequeue Pointer and Descriptor 17.8.4.2.5 Outbox Controller Interrupts The outbox generates the following four interrupts, which can be individually enabled. • • Queue overflow interrupt. Generated if the enqueue pointer passes the dequeue pointer and the queue is full and the interrupt is enabled. OSR[QOI] is set if OMR[QOIE] is set. Queue full interrupt. Generated if the enqueue pointer catches up to the dequeue pointer, the queue is not empty, and the interrupt is enabled. OSR[QFI] is set if OMR[QFIE] is set. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 17-116 Freescale Semiconductor RapidIO Message Unit • • Queue empty interrupt. Generated if the dequeue pointer catches up to the enqueue pointer and the interrupt is enabled. OSR[QEI] is set if OMR[QEIE] is set. End-of-message interrupt. Generated after the completion of each message if enabled. OSR[EOMI] is set if ODATR[EOMIE] is set. 17.8.4.2.6 Special Error Case Condition If a data message is in progress and the outbox encounters an error with one of the segments, it causes an error at the recipient by placing an illegal field in the packet to cause the error. This prevents the recipient from using data in the message. For example, if the outbox controller reads message data from its DRAM memory and gets an uncorrectable ECC error. This scheme prevents the recipient from hanging or using bad data. 17.8.4.3 Inbox Controller Operation The inbox controller is responsible for receiving messages and placing them in a circular queue in local memory. Although, it only supports one mailbox and one letter, the inbox controller can receive messages with any mailbox or letter number. Furthermore, the inbox can receive segments of a message in any order. The address of where to write the message is computed as: Base address + (msgseg * ssize in double words). Unlike the outbox where the enqueue pointer is controlled by software and the dequeue pointer is controlled by hardware, the inbox controls the enqueue pointer and software controls the dequeue pointer. Figure 17-84 shows a sample structure of the inbound mailbox message frames and the frame pointers. In this example, the frame queue has eight entries, three of which are currently valid. The mailbox controller adds frames to the head of the queue and the processor core takes them off from the tail. After processing a message, the processor core writes the inbox mode register mailbox increment bit (IMR[MI]) to point to the next message frame in the queue. This process is repeated for each received message. Processor Core Read Dequeue pointer Local Memory Enqueue pointer Message Frame Message Frame Message Frame Dequeue pointer Packets from RapidIO Interface Inbound Mailbox Port Frame Queue Figure 17-84. Inbound Mailbox Structure MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 17-117 RapidIO Interface The sequence of events that takes place during the reception of a data message is as follows: • The inbox receives a message request from the RapidIO port. If the inbox is enabled (IMR[ME] = 1) and it is not busy with another message (ISR[MB] = 1), then the message is accepted. The address is computed for each segment of the message (up to 16 segments/packets per message) using the value of the inbound frame queue enqueue pointer address registers and properly adjusted based on segment number. Data is stored in the circular queue in local memory at the computed address. After the entire message is received, the enqueue pointer is incremented to point to the next message frame in local memory. An interrupt is generated if enabled (IMR[MIQIE] = 1). The interrupt is held until the dequeue pointer equals the enqueue pointer and the message queue is empty. That is, as long as messages exist in memory that have not been serviced and the interrupt is enabled, the interrupt remains asserted. • • • • 17.8.4.3.1 Retry Response Conditions The following two conditions generate a logical layer retry (response retry): • • Inbox receives a message but is busy handling another message Local memory circular queue is full and a message is received 17.8.4.3.2 Error Response Conditions Several conditions generate a logical layer error response (response error) as follows: • • • • • Inbox receives a message and it is not enabled Inbox is in an error state caused by a previous message Inbox receives a segment that exceeds the message length (msgseg > msglen) Inbox receives a message segment being processed with a bad message length Inbox receives a duplicate segment of a message being processed that was already received successfully 17.8.4.3.3 Inbox Controller Interrupts The inbox generates the following interrupts, which can be individually enabled: • • Message-in-queue interrupt. This interrupt is generated each time the circular queue becomes not empty and the interrupt is enabled. ISR[MIQI] is set if IMR[MIQIE] is set. Queue full interrupt. This interrupt is generated each time the circular queue becomes full and the interrupt is enabled. ISR[QFI] is set if IMR[QFIE] is set. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 17-118 Freescale Semiconductor RapidIO Message Unit 17.8.4.3.4 Data Message Controller Limitations and Restrictions This section describes some of the limitations and restrictions of the data message controller. This is intended to help software maximize the message passing performance and avoid programming errors. Software must guarantee that the enqueue pointer is not pointing to an invalid descriptor in chaining mode if the enqueue and dequeue pointers are not equal. 17.8.4.4 Doorbell Message Controller The RapidIO protocol supports a doorbell message type that contains no data payload. Even though the RapidIO architecture references outbound and inbound doorbell message controllers, the RapidIO controller supports only an inbound doorbell message queue. Outbound doorbells are generated outside RapidIO within the outbound ATMU. Inbound doorbell messages are handled by the doorbell message controller similar to how the data message controller handles inbound data messages. The doorbell controller receives the doorbell message and places it in a circular queue located in the local memory. Figure 17-85 shows an example of the structure of the inbound doorbell queue and its pointers. The doorbell queue of the RapidIO controller has eight entries, three of which are currently valid. The doorbell controller adds doorbell information to the head of the queue and the processor core removes it from the tail. The doorbell entry size is fixed at 64 bits because doorbell packets only pass a small amount of information, making the enqueue and dequeue pointers double-word addresses. Processor Core Read Dequeue pointer Enqueue pointer Inbound Doorbell Controller Dequeue pointer Doorbell Entry Doorbell Entry Doorbell Entry Local Memory Packets from RapidIO interface Figure 17-85. Inbound Doorbell Queue and Pointer Structure MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 17-119 RapidIO Interface 17.8.4.4.1 Inbound Doorbell Reception The following sequence of events occurs during reception of a doorbell message: • The inbox receives a doorbell message request from the RapidIO port. If the inbox is enabled (DMR[DE] = 1) and it is not busy with another doorbell message (DSR[DB] = 1), then the doorbell message is accepted. The 16-bit information field is stored in local memory using the value of DQDPAR. The enqueue pointer is incremented to point to the next doorbell queue entry in local memory. An interrupt is generated if enabled (DMR[DIQIE] = 1). The interrupt is held until the dequeue pointer equals the enqueue pointer and the doorbell queue is empty. That is, as long as doorbell messages exist in memory that have not been serviced and the interrupt is enabled, the interrupt remains asserted. • • • 17.8.4.4.2 Doorbell Queue Entry Format This section defines the format of the doorbell information written to memory by the doorbell controller. Each doorbell entry in the queue has two 32-bit offsets, one for target information and one for source information. Table 17-104 shows the target information. Table 17-104. Target Information Definition Bits 0–23 24–31 Name — TID Reserved Target ID field from the received doorbell packet Description Table 17-105 shows the source information. Table 17-105. Source Information Definition Bits 0–7 8–15 16–31 Name — SID INFO Reserved Source ID field from the received doorbell packet Information field from the received doorbell packet Description Figure 17-86 shows the doorbell queue entry fields and their related offsets. Local Memory Offset 0b000 0b100 0 Target info Source info 31 Figure 17-86. Doorbell Entry Format MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 17-120 Freescale Semiconductor RapidIO Message Unit 17.8.4.4.3 Retry Response Conditions There is one condition in which a doorbell message is logical layer retried (response retry): • Doorbell request received and the doorbell circular queue is full 17.8.4.4.4 Error Response Conditions Several conditions cause a logical layer error response (response error): • • Doorbell controller is error state Doorbell controller is not enabled 17.8.4.4.5 Doorbell Controller Interrupts The doorbell controller generates the following two different interrupts that can be individually enabled: • • Doorbell-in-queue interrupt. Generated each time the circular queue becomes not empty and the interrupt is enabled. DSR[DIQI] is set if DMR[DIQIE] is set Queue full interrupt. Generated each time the circular queue becomes full and the interrupt is enabled. DSR[QFI] is set if DMR[QFIE] is set 17.8.4.5 Port-Write Controller Structure The implementation of the port-write controller is very similar to the inbound message hardware. The port-write is intended as an error reporting mechanism from an end-point-free device to a control processor or other system host. Figure 17-87 shows an example of the structure of the inbound queue and pointer. The port-write queue only contains one entry with a fixed size of 64 bytes and aligned to a cache line boundary. The port-write controller puts an incoming port-write data payload in the location indicated by the pointer and then sets its PWSR[QFI] if PWMR[QFIE] is set. The processor core must explicitly clear PWSR[QFI] after servicing the port-write. While the full bit is set, or if it is busy handling a previous port-write, or the port-write controller is not enabled, the controller silently discards all incoming port-write packets. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 17-121 RapidIO Interface Processor Core Read Pointer Local Memory Port-write packets from RapidIO interface Inbound Port-write Port Pointer Port-write 64-byte Entry Port-write Queue Figure 17-87. Inbound Port-Write Structure Upon successfully receiving a port-write transaction, if PWMR[QFIE] is set, PWSR[QFI] is set and an interrupt is generated to the processor core. 17.9 Initialization and Application Information The following steps outline the RapidIO port initialization procedure: 1. Configure the transmit clock select pins to select either the receive clock, externally provided transmit clock or the platform clock for the RapidIO transmit clock domain. 2. Initialize the host/agent field (using configuration pins) to put the end-point in host or agent mode over RapidIO. 3. Initialize the Device ID field (using configuration pins) to assign a device identifier at power-on reset. 4. The address translation and mapping unit (ATMU) for RapidIO must be initialized for RapidIO accesses. 5. Clear the CR[AA] bit to allow the logic to detect an error when it receives packets with a target ID not equal to its own. 6. Follow the system initialization and discovery mechanism described in the inter-operability portion of the Parallel RapidIO Interconnect Specification for configuring the RapidIO system. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 17-122 Freescale Semiconductor Part IV Global Functions and Debug Part IV defines other global blocks of the MPC8540. The following chapters are included: • Chapter 18, “Global Utilities,” defines the global utilities of the MPC8540. These include power management, I/O device enabling, power-on reset (POR) configuration monitoring, general-purpose I/O signal use, and multiplexing for the interrupt and local bus chip select signals Chapter 19, “Performance Monitor,” describes the performance monitor of the MPC8540. Chapter 20, “Debug Features and Watchpoint Facility,” describes the debug features and watchpoint monitor of the MPC8540. Chapter 21, “10/100 Fast Ethernet Controller,” includes information about using the dual-speed Ethernet controller of the MPC8540 for debug purposes. • • • MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor IV-1 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 IV-2 Freescale Semiconductor Chapter 18 Global Utilities This chapter describes the global utilities of the MPC8540. It provides signal descriptions, register descriptions, and a functional description of these utilities. 18.1 Overview The global utilities block controls power management, I/O device enabling, power-on reset (POR) configuration monitoring, general-purpose I/O signal configuration, alternate function selection for multiplexed signals, and clock control. 18.2 Global Utilities Features This section provides an overview of global utilities features. 18.2.1 Power Management and Block Disables The following features affect the overall power consumption of the device: • • • • Dynamic power management mode Software-controlled power management (doze, nap, sleep) Externally controlled power management (doze, sleep) Static power management (I/O block disables) 18.2.2 Accessing Current POR Configuration Settings The POR configuration values of all device parameters sampled from pins at reset are available through memory-mapped registers in the global utilities block. 18.2.3 General-Purpose I/O The PCI and TSEC2 data bus signals can be used as general-purpose I/O signals when not used for their primary function. Memory-mapped registers in the global utilities block provide control and status for the use of these signals. A general purpose input register is loaded with the values of the local bus address/data pins at the negation of HRESET. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 18-1 Global Utilities 18.2.4 Interrupt and Local Bus Signal Multiplexing IRQ[9:11] and LCS[5:7] serve multiple functions that can be selected by configuration registers in the global utilities block. 18.2.5 Clock Control The global utilities block also selects the internal clock signal driven on CLK_OUT. 18.3 External Signal Descriptions The following subsections provide information about signals that serve as global utilities. 18.3.1 Signals Overview Table 18-1 summarizes the external signals used by the global utilities block. Table 18-1. External Signal Summary Signal Name ASLEEP CKSTP_ IN CKSTP_OUT CLK_OUT I/O O I O O Description Signals that the device has reached a sleep state Checkstop input Checkstop output Clock out. Selected by CLKOCR values Reference (Section/Page) 18.5.1.5.3/18-24 — — 18.4.1.16/18-18 18.3.2 Detailed Signal Descriptions Table 18-2 describes signals in the global utilities block in detail. Table 18-2. Detailed Signal Descriptions Signal ASLEEP I/O Description O Asleep. See Section 18.5.1.5.3, “Sleep Mode.” After negation of HRESET, ASLEEP is asserted until the device completes its power-on reset sequence and reaches its ready state. State Asserted—Indicates that the device is either still in its power-on reset sequence or has Meaning reached a sleep state after a power-down command is issued by software Negated—The device is not in sleep mode. (It has either awakened from a power-down state or completed the POR sequence.) Timing Assertion—May occur at any time; may be asserted asynchronously to the input clocks Negation—Negates synchronously with SYSCLK when leaving power-on sequence; otherwise negation is asynchronous MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 18-2 Freescale Semiconductor Memory Map/Register Definition Table 18-2. Detailed Signal Descriptions (continued) Signal CKSTP_IN I/O I Checkstop in State Asserted—Indicates that the e500 core must enter a hard stop condition. All e500 clocks Meaning are turned off. CKSTP_OUT is asserted. The rest of MPC8540 device logic, including memory controllers, internal memories and registers, and I/O interfaces, remains functional. Negated—Indicates that normal operation should proceed Timing CKSTP_OUT Assertion—May occur at any time; may be asserted asynchronously to the input clocks Negation—Must remain asserted until the MPC8540 is reset with assertion of HRESET Description O Checkstop out State Asserted—Indicates that the e500 core of the MPC8540 is in a checkstop state. The rest Meaning of the MPC8540 logic remains functional. Negated—Indicates normal operation. After CKSTP_OUT has been asserted, it is negated after the next negation (low-to-high transition) of HRESET. Timing Assertion—May occur at any time; may be asserted asynchronously to the input clocks Negation—Must remain asserted until the device has been reset with a hard reset CLK_OUT O Clock out. Reflects clock signal selected by CLKOCR (see Section 18.4.1.16, “Clock Out Control Register (CLKOCR)”) State Asserted—If CLKOCR[ENB] = 1, clock signal selected by CLKOCR[CLK_SEL] is driven Meaning High impedance—If CLKOCR[ENB] = 0 Timing Assertion/Negation—Depends on the value of CLKOCR[CLK_SEL] 18.4 Memory Map/Register Definition Table 18-3 summarizes the global utilities registers and their addresses. Table 18-3. Global Utilities Block Register Summary Offset Register Access Reset Section/Page Power-On Reset Configuration Values 0xE_0000 PORPLLSR—POR PLL ratio status register 0xE_0004 PORBMSR—POR boot mode status register 0xE_0008 PORIMPSCR—POR I/O impedance status and control register 0xE_000C PORDEVSR—POR I/O device status register 0xE_0010 PORDBGMSR—POR debug mode status register 0xE_0020 GPPORCR—General-purpose POR configuration register Signal Multiplexing and GPIO Controls 0xE_0030 GPIOCR—GPIO control register 0xE_0040 GPOUTDR—General-purpose output data register R/W R/W 0x0000_0000 18.4.1.7/18-10 0x0000_0000 18.4.1.8/18-11 R R R/W R R R 0x00nn_00nn 0xnnnn_0000 0x000n_007F See ref. See ref. See ref. 18.4.1.1/18-4 18.4.1.2/18-5 18.4.1.3/18-6 18.4.1.4/18-7 18.4.1.5/18-9 18.4.1.6/18-9 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 18-3 Global Utilities Table 18-3. Global Utilities Block Register Summary (continued) Offset Register Access R R/W Reset Section/Page 0xE_0050 GPINDR—General-purpose input data register 0xE_0060 PMUXCR—Alternate function signal multiplex control Device Disables 0xE_0070 DEVDISR—Device disable control Power Management Registers 0xE_0080 POWMGTCSR—Power management status and control register Interrupt Reporting 0xE_0090 MCPSUMR—Machine check summary register Version Registers 0xE_00A0 PVR—Processor version register 0xnnnn_0000 18.4.1.9/18-12 0x0000_0000 18.4.1.9/18-12 R/W 0x0000_0000 18.4.1.11/18-13 R/W 0x0000_0000 18.4.1.12/18-15 Read/Clear 0x0000_0000 18.4.1.13/18-16 R e500 processor version MPC8540 system version 18.4.1.14/18-17 0xE_00A4 SVR—System version register R 18.4.1.15/18-18 Debug Control 0xE_0E00 CLKOCR—Clock out select register 0xE_0E10 DDRDLLCR—DDR DLL control register 0xE_0E20 LBDLLCR—LBC DLL control register R/W R/W R/W 0x0000_0000 18.4.1.16/18-18 0x0000_0000 18.4.1.17/18-19 0x0000_0000 18.4.1.18/18-20 18.4.1 Register Descriptions This section describes the global utilities registers in detail. 18.4.1.1 POR PLL Status Register (PORPLLSR) PORPLLSR, shown in Figure 18-1, contains the settings for the PLL ratios as set by the cfg_sys_pll[0:3] and cfg_core_pll[0:1] POR configuration pins. See Section 4.4.3.1, “System PLL Ratio,” and Section 4.4.3.2, “e500 Core PLL Ratio,” for more information. 0 9 10 15 16 25 26 30 31 R0 0 0 0 0 0 0 0 0 0 W Reset Offset e500_Ratio 0000000000 Plat_Ratio 0 0000_0000_00nn_nnnn_0000_0000_00nn_nnnn 0xE_0000 Figure 18-1. POR PLL Status Register (PORPLLSR) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 18-4 Freescale Semiconductor Memory Map/Register Definition Table 18-4 describes the bit settings of PORPLLSR. Table 18-4. PORPLLSR Field Descriptions Bits 0–9 Name — Reserved Description 10–15 e500_Ratio Clock ratio between the e500 core and the CCB clock. The 2 lsbs of this field correspond to the values on cfg_core_pll[0:1] at the negation of HRESET. Patterns not shown are reserved. 0001002:1 0001015:2 0001103:1 0001117:2 16–25 — Reserved 26–30 Plat_Ratio Clock ratio between the CCB (platform) clock and SYSCLK. The 4 lsbs correspond to the values on cfg_sys_pll[0:3] at the negation of HRESET. Patterns not shown are reserved. 000102:1 010008:1 000113:1 010019:1 001004:1 0101010:1 001015:1 0110012:1 001106:1 1000016:1 31 — Reserved 18.4.1.2 POR Boot Mode Status Register (PORBMSR) The PORBMSR, shown in Figure 18-2, reports setting of the POR configuration pins that control the boot mode settings (described in Section 4.4.3.3, “Boot ROM Location,” Section 4.4.3.5, “CPU Boot Configuration,” and Section 4.4.3.6, “Boot Sequencer Configuration,”) and the default settings of RapidIO and PCI host/agent modes (described in Section 4.4.3.4, “Host/Agent Configuration”). 0 1 4 5 7 8 9 10 11 12 13 14 15 16 31 R BCFG 0 0 0 0 ROM_LOC 0 0 BSCFG 0 0 W Reset Offset HA 0000000000000000 n000_0nnn_00nn_00nn_0000_0000_0000_0000 0xE_0004 Figure 18-2. POR Boot Mode Status Register (PORBMSR) For information about how the RapidIO interface is configured, see Section 17.3.1.19, “Port General Control Command and Status Register (PGCCSR).” For more information about the PCI configurations, see Section 16.3.2.19, “PCI Bus Function Register (PBFR).” Figure 18-5 describes the bit settings of the PORBMSR. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 18-5 Global Utilities Table 18-5. PORBMSR Field Descriptions Bits 0 Name BCFG Description CPU boot configuration 0 The CPU is prevented from booting until configuration by an external master is complete. 1 The CPU is allowed to start fetching boot code. Reserved 1–4 5–7 — ROM_LOC Location of boot ROM: 000 PCI/PCI-X 001 DDR SDRAM 011 RapidIO 101 Local bus GPCM: 8-bit 110 Local bus GPCM:16-bit 111 Local bus GPCM: 32-bit — BSCFG Reserved Boot sequencer configuration 00 Reserved 01 Boot sequencer enabled with normal I2C addressing 10 Boot sequencer enabled with extended I2C addressing 11 Boot sequencer disabled Reserved Host/agent mode configuration. When the MPC8540 is an agent on an interface, it is prevented from mastering transactions on that interface until the external host configures the interface appropriately. 00 PCI and RapidIO agent mode 01 RapidIO agent mode 10 PCI/PCI-X agent mode 11 Host mode Reserved 8–9 10–11 12–13 14–15 — HA 16–31 — 18.4.1.3 POR I/O Impedance Status and Control Register (PORIMPSCR) PORIMPSCR, shown in Figure 18-3,contains the current I/O driver impedances for local bus and PCI interfaces. 0 14 15 16 24 25 26 27 28 29 30 31 R 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 PCI_Z 0 0 0 0 0 0 0 0 0 LALE_Z LADP_Z LA_CKE_Z LCS12_Z LWE_Z LGPL_Z LCLK_Z W Reset Offset 0000_0000_0000_000n_0000_0000_0111_1111 0xE_0008 Figure 18-3. POR I/O Impedance Status and Control Register (PORIMPSCR) The I/O impedance of local bus signals (including the local bus clock) is controlled through this register. The I/O impedance of PCI signals is controlled by POR configuration pins (described in MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 18-6 Freescale Semiconductor Memory Map/Register Definition Section 4.4.3.13, “PCI I/O Impedance.”) The MPC8540 Integrated Processor Hardware Specifications provides exact I/O impedances. Table 18-6 describes PORIMPSCR fields. Table 18-6. PORIMPSCR Field Descriptions Bits 0–14 15 Name — PCI_Z Reserved PCI/PCI-X I/O impedance 0 Low impedance 1 High impedance Reserved I/O impedance for local bus address latch enable 0 Low impedance 1 High impedance I/O impedance for local bus address/data and data parity (LAD[0:31] and LDP[0:7]) 0 Low impedance 1 High impedance I/O impedance for local bus address and clock enable (LA[27:31] and LCKE) 0 Low impedance 1 High impedance I/O impedance for two local bus chip selects (LCS[1] and LCS[2] only). NOTE: Other chip selects use a fixed high I/O impedance 0 Low impedance 1 High impedance I/O impedance for local bus write enables (LWE[0:3]) 0 Low impedance 1 High impedance I/O impedance for local bus general-purpose lines (LGPL[0:5]) 0 Low impedance 1 High impedance I/O impedance for local bus clocks (LCLK[0:2]) 0 Low impedance 1 High impedance Description 16–24 25 — LALE_Z 26 LADP_Z 27 LA_CKE_Z 28 LCS12_Z 29 LWE_Z 30 LGPL_Z 31 LCLK_Z 18.4.1.4 POR Device Status Register (PORDEVSR) Shown in Figure 18-4, PORDEVSR reports other POR settings for I/O devices as described in Section 4.4.3.7, “TSEC Width,” Section 4.4.3.8, “TSEC1 Protocol,” Section 4.4.3.9, “TSEC2 Protocol,” Section 4.4.3.16, “PCI-X Configuration,” Section 4.4.3.14, “PCI Arbiter Configuration,” Section 4.4.3.12, “PCI Width Configuration,” Section 4.4.3.10, “RapidIO Transmit Clock Source,” and Section 4.4.3.11, “RapidIO Device ID.” MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 18-7 Global Utilities 0 1 56 7 8 9 13 14 15 16 21 22 23 24 28 29 31 R ECW 0 0 0 0 0 ECP PCI 0 0 0 0 0 PCI_ARB PCI32 0 0 0 0 0 0 RIOCLK 0 0 0 0 0 DEV_ID W Reset Offset n000_00nn_n000_00nn_0000_00nn_0000_0nnn 0xE_000C Figure 18-4. POR Device Status Register (PORDEVSR) Table 18-7 describes the bit settings of PORDEVSR. Table 18-7. PORDEVSR Field Descriptions Bits 0 Name ECW Gigabit Ethernet controller width 0 Reduced (RGMII, RTBI) 1 Full (MII, GMII, TBI) Reserved Gigabit Ethernet controller protocol 00 Both TSEC blocks use a Media Independent Interface protocols (MII/GMII/RGMII) 01 TSEC1 uses MII, TSEC2 uses TBI 10 TSEC1 uses TBI, TSEC2 uses MII 11 Both TSEC blocks use a ten bit interface protocol (TBI or RTBI) PCI/PCI-X mode 0 PCI-X mode 1 PCI mode Reserved Description 1–5 6-7 — ECP 8 PCI 9–13 14 — PCI_ARB PCI/PCI-X arbiter enable 0 PCI/PCI-X arbiter is disabled 1 PCI/PCI-X arbiter is enabled PCI32 PCI/PCI-X interface width 0 64-bit 1 32-bit Reserved RapidIO transmit clock source 00 Reserved 01 RapidIO receive clock 10 RapidIO transmit clock input 11 CCB clock (default) Reserved RapidIO device ID [5:7] 15 16–21 22–23 — RIOCLK 24–28 29–31 — DEV_ID MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 18-8 Freescale Semiconductor Memory Map/Register Definition 18.4.1.5 POR Debug Mode Status Register (PORDBGMSR) PORDBGMSR, shown in Figure 18-5, holds debug mode settings from the POR configuration pins as described in Section 4.4.3.17, “Memory Debug Configuration,” Section 4.4.3.18, “DDR Debug Configuration,” and Section 4.4.3.15, “PCI Debug Configuration.” 0 4 5 6 7 8 31 R 0 0 0 0 0 MEM_SEL DDR_DBG PCI_DBG 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 W Reset Offset 0000_0nnn_0000_0000_0000_0000_0000_0000 0xE_0010 Figure 18-5. POR Debug Mode Status Register (PORDBGMSR) Table 18-8 describes the bit settings of PORDBGMSR. Table 18-8. PORDBGMSR Field Descriptions Bits 0–4 5 Name — Reserved Description MEM_SEL Memory select. Indicates which controller is driving MSRCID[0:4] and MDVAL 0 Local bus controller is driving debug information 1 DDR SDRAM controller is driving debug information DDR_DBG DDR debug configuration 0 SourceID and data valid information is being driven on ECC pins of DDR SDRAM interface 1 Normal mode. ECC information is being driven on ECC pins of DDR SDRAM interface PCI_DBG PCI/PCI-X debug configuration 0 PCI/PCI-X drives source ID onto PCI_AD[62:58]. 1 PCI/PCI-X drives address/data 62:58 onto PCI_AD[62:58]. — Reserved 6 7 8–31 18.4.1.6 General-Purpose POR Configuration Register (GPPORCR) Shown in Figure 18-6, GPPORCR stores the value sampled from the local bus address/data signals, LAD[0:31], during POR, as described in Section 4.4.3.21, “General-Purpose POR Configuration.” Software can use this value to inform the operating system about initial system configuration. Typical interpretations include circuit board type, board ID number, or a list of available peripherals. 0 31 R W Reset Offset POR_CFG_VEC 0xnnnn_nnnn 0xE_0020 Figure 18-6. POR Configuration Register (GPPORCR) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 18-9 Global Utilities Table 18-9 describes the bit settings of GPPORCR. Table 18-9. GPPORCR Field Descriptions Bits Name Description 0–31 POR_CFG_VEC General-purpose POR configuration vector sampled from local bus address/data signals at the negation of HRESET. Note that if nothing is driven on these signals during reset, the value of this register is indeterminate. 18.4.1.7 General-Purpose I/O Control Register (GPIOCR) Shown in Figure 18-7, GPIOCR contains the enable bits for each group of pins that may be used for general-purpose I/O. These bits have meaning only if the pins are not being used for their primary function. Note that when these signals are enabled as general-purpose I/O signals, they are read and written through GPINDR and GPOUTDR described in Section 18.4.1.9, “General-Purpose Input Data Register (GPINDR),” and Section 18.4.1.8, “General-Purpose Output Data Register (GPOUTDR).” Section 18.5.2, “General-Purpose I/O Signals,” describes the use of general-purpose I/O signals. 0 5 6 7 8 13 14 15 16 31 R00 0000 Tx2out Rx2in W Reset Offset 000000 PCIout PCIin 00 00000000000000 0000_0000_0000_0000_0000_0000_0000_0000 0xE_0030 Figure 18-7. General-Purpose I/O Control Register (GPIOCR) Table 18-10 describes the bit settings of GPIOCR. Table 18-10. GPIOCR Field Descriptions Bits 0–5 6 Name — Reserved Description Tx2out TSEC2_Tx[0:7] output enable. Used to enable TSEC2_Tx[0:7] for use as general-purpose output if the TSEC2 interface is disabled 0 TSEC2_Tx[0:7] is not used as general-purpose output. 1 TSEC2_Tx[0:7] is used as general-purpose output. Rx2in TSEC2_Rx[0:7] input enable. Used to enable TSEC2_Rx[0:7] for use as general-purpose input if the TSEC2 interface is disabled 0 TSEC2_Rx[0:7] is not used as general-purpose input. 1 TSEC2_Rx[0:7] is used as general-purpose input. — Reserved 7 8–13 14 PCIout PCI_AD[47:40] output enable. Used to enable PCI_AD[47:40] for use as general-purpose output if PCI interface is in 32-bit mode or is disabled 0 PCI_AD[47:40] is not used as general-purpose output. 1 PCI_AD[47:40] is used as general-purpose output. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 18-10 Freescale Semiconductor Memory Map/Register Definition Table 18-10. GPIOCR Field Descriptions (continued) Bits 15 Name PCIin Description PCI_AD[39:32] input enable. Used to enable PCI_AD[39:32] for use as general-purpose input if the PCI interface is in 32-bit mode or is disabled 0 PCI_AD[39:32] is not used as general-purpose input. 1 PCI_AD[39:32] is used as general-purpose input. Reserved 16–31 — 18.4.1.8 General-Purpose Output Data Register (GPOUTDR) GPOUTDR, shown in Figure 18-8, contains the data driven as general-purpose output on TSEC2_TxD[0:7] and/or PCI_AD[47:40] when either of these buses is configured as general-purpose I/O buses, as described in Section 18.4.1.7, “General-Purpose I/O Control Register (GPIOCR).” Writes to GPOUTDR affect only pins enabled as general-purpose outputs. Reads return valid data only for bits corresponding to pins enabled as general-purpose outputs. GPOUTDR may be accessed using single byte writes (using big-endian addressing) so that writes to one byte do not affect outputs controlled by others. 0 15 16 31 R GPOUT W Reset Offset 0000000000000000 0000_0000_0000_0000_0000_0000_0000_0000 0xE_0040 Figure 18-8. General-Purpose Output Data Register (GPOUTDR) Table 18-11 describes the fields of GPOUTDR. Table 18-11. GPOUTDR Field Descriptions Bits 0–15 Name Description GPOUT General-purpose output data. When the corresponding signals are configured to be general-purpose output signals, the values of the bits of GPOUT are driven onto those pins. GPOUTDR[0:7] corresponds to TSEC2_TxD[0:7]. GPOUTDR[8:15] corresponds to PCI_AD[47:40] as follows: GPOUTDR[8] ↔ PCI_AD[47] GPOUTDR[9] ↔ PCI_AD[46] GPOUTDR[10] ↔ PCI_AD[45] GPOUTDR[11] ↔ PCI_AD[44] GPOUTDR[12] ↔ PCI_AD[43] GPOUTDR[13] ↔ PCI_AD[42] GPOUTDR[14] ↔ PCI_AD[41] GPOUTDR[15] ↔ PCI_AD[40] — Reserved, should be cleared 16–31 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 18-11 Global Utilities 18.4.1.9 General-Purpose Input Data Register (GPINDR) GPINDR, shown in Figure 18-9, contains the data currently sampled as general-purpose input on TSEC2_RxD[0:7] and/or PCI_AD[39:32] when either of these buses is configured as general-purpose I/O buses in GPIOCR (see Section 18.4.1.7, “General-Purpose I/O Control Register (GPIOCR)”). GPINDR bits are updated only if the associated bits are configured as general-purpose input pins rather than their primary functions. 0 15 16 31 R W Reset Offset GPIN 0000000000000000 nnnn_nnnn_nnnn_nnnn_0000_0000_0000_0000 0xE_0050 Figure 18-9. General-Purpose Output Data Register (GPINDR) Table 18-11 describes the fields of GPINDR. Table 18-12. GPINDR Field Descriptions Bits 0–15 Name Description GPIN General-purpose input data. When the corresponding signals are configured to be general-purpose input signals, the values sampled on these signals are reflected in GPIN. GPINDR[0:7] corresponds to TSEC2_RxD[0:7]. GPINDR[8:15] corresponds to PCI_AD[39:32] as follows: GPINDR[8] ↔ PCI_AD[39] GPINDR[9] ↔ PCI_AD[38] GPINDR[10] ↔ PCI_AD[37] GPINDR[11] ↔ PCI_AD[36] GPINDR[12] ↔ PCI_AD[35] GPINDR[13] ↔ PCI_AD[34] GPINDR[14] ↔ PCI_AD[33] GPINDR[15] ↔ PCI_AD[32] — Reserved 16–31 18.4.1.10 Alternate Function Signal Multiplex Control Register (PMUXCR) Shown in Figure 18-10, PMUXCR contains bits that enable DMA channels 2 and 3 which exist as alternate functions on local bus chip select pins LCS[5:7], and interrupt input pins IRQ[9:11], respectively. Specifically, DMA request, acknowledge, and done signals comprise the secondary functions for the associated IRQ and local bus chip select signals. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 18-12 Freescale Semiconductor Memory Map/Register Definition 0 14 15 16 30 31 R000000000000000 DMA2 W Reset Offset 000000000000000 DMA3 0000_0000_0000_0000_0000_0000_0000_0000 0xE_0060 Figure 18-10. Alternate Function Pin Multiplex Control Register (PMUXCR) Table 18-13 describes the bit settings of PMUXCR. Table 18-13. PMUXCR Field Descriptions Bits 0–14 15 Name — DMA2 Reserved Enables DMA channel 2 signals 0 DMA channel 2 is disabled; the pins retain their primary function as local bus chip selects 1 DMA channel 2 is enabled and LCS5 functions as DMA_DREQ2 LCS6 functions as DMA_DACK2 LCS7 functions as DMA_DDONE2 Reserved Enables DMA channel 3 signals 0 DMA channel 3 is disabled; the pins retain their primary function as interrupt requests. 1 DMA channel 3 is enabled and IRQ9 functions as DMA_DREQ3 IRQ10 functions as DMA_DACK3 IRQ11 functions as DMA_DDONE3 Description 16–30 31 — DMA3 18.4.1.11 Device Disable Register (DEVDISR) DEVDISR, shown in Figure 18-11, contains disable bits for various MPC8540 functional blocks. 0 1 3 4 5 10 11 12 13 14 15 R PCI W Reset 16 0 0 0 LBC 0 0 0 0 0 0 L2 RIO 0 0 DDR 0000_0000_0000_0000 17 18 20 21 22 23 24 25 26 27 28 29 30 31 R E500 W Reset Offset TB 0 0 0 DMA 0 0 TSEC1 TSEC2 FEC 0 0 I2C DUART 0 0000_0000_0000_0000 0xE_0070 Figure 18-11. Device Disable Register (DEVDISR) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 18-13 Global Utilities All functional blocks are enabled after reset; unneeded blocks can be disabled to reduce power consumption or allow their signals to be used as general-purpose I/O signals. See Section 18.4.1.7, “General-Purpose I/O Control Register (GPIOCR).” Blocks disabled by DEVDISR must not be re-enabled without a hard reset. Section 18.5.1.4, “Shutting Down Unused Blocks,” has more information on the use of DEVDISR. Table 18-14 describes DEVDISR fields. . Table 18-14. DEVDISR Field Descriptions Bits 0 Name PCI Description PCI/PCI-X controller disable 0 PCI/PCI-X controller enable 1 PCI_AD[63:32] may be used as general-purpose I/O Reserved Local bus controller disable 0 Local bus controller enable 1 Local bus controller disable Reserved Level-2 cache disable 0 L2 cache enable 1 L2 cache disable RapidIO controller disable 0 RapidIO controller enable 1 RapidIO controller disable, including RapidIO messaging, doorbell, and port write controllers. Reserved DDR SDRAM controller disable 0 DDR SDRAM controller enable 1 DDR SDRAM controller disable e500 core disable 0 e500 core enable 1 e500 core disable. Places the core in the core-stopped state in which it does not respond to interrupts. Equivalent to nap mode. Instruction fetching is stopped, snooping is disabled, and clocks are shut down to all functional units of the core except for the timer facilities. For more information, see Section 18.5.1.4, “Shutting Down Unused Blocks.” Time base (timer facilities) of the e500 core disable 0 Timer facilities enabled 1 Timer facilities disabled Reserved DMA controller disabled 0 DMA controller enabled 1 DMA controller disabled Reserved 1–3 4 — LBC 5–10 11 — L2 12 RIO 13–14 15 — DDR 16 E500 17 TB 18–20 21 — DMA 22–23 24 — TSEC1 Three-speed Ethernet controller 1 disable 0 TSEC1 enabled 1 TSEC1 disabled MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 18-14 Freescale Semiconductor Memory Map/Register Definition Table 18-14. DEVDISR Field Descriptions (continued) Bits 25 Name Description TSEC2 Three-speed Ethernet controller 2 disable 0 TSEC2 enabled 1 TSEC2 disabled. RxD and TxD pins may be used for general-purpose I/O FEC Fast Ethernet controller (10/100 maintenance port) disable 0 FEC enabled 1 FEC disabled Reserved I2C controller disabled 0 I2C controller enabled 1 I2C controller disabled 26 27–28 29 — I2C 30 DUART Dual UART controller disabled 0 DUART enabled 1 DUART disabled — Reserved 31 18.4.1.12 Power Management Control and Status Register (POWMGTCSR) Shown in Figure 18-12, POWMGTCSR contains bits for placing the MPC8540 into low power states and for controlling when it wakes up. It also contains power management status bits. See Section 18.5.1.8.2, “Interrupts and Power Management Controlled by POWMGTCSR,” for more information. 0 1 2 11 12 13 14 15 27 28 29 30 31 R IRQ_MSK CI_MSK W Reset Offset 000000000 0 DOZ 0 SLP 0 0 0 0 0 0 0 0 0 0 0 0 0 DOZING NAPPING SLPING 0 0000_0000_0000_0000_0000_0000_0000_0000 0xE_0080 Figure 18-12. Power Management Control and Status Register (POWMGTCSR) Table 18-15 describes the bit settings of POWMGTCSR. Table 18-15. POWMGTCSR Field Descriptions Bits 0 Name Description IRQ_MSK Interrupt input mask 0 Interrupts cause the device to wake up from a low-power state. 1 Interrupts are masked as a wake-up condition. The device remains in a low-power state despite the presence of an interrupt request. CI_MSK Critical interrupt input mask 0 Critical interrupts cause the device to wake up from a low power state. 1 Critical interrupts are masked as a wake-up condition. The device remains in a low-power state despite the presence of a critical interrupt. 1 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 18-15 Global Utilities Table 18-15. POWMGTCSR Field Descriptions (continued) Bits 2–11 12 Name — DOZ Reserved Doze mode 0 No request to put device in doze mode. Note that this bit is automatically cleared on MCP, UDE, SRESET, core_tbint (from the core) and also int and cint if not masked. 1 Device is to be placed in doze mode. Instruction fetching is halted in the e500 core. Note that this bit is logically ORed with HID0[DOZE]. Reserved Sleep mode 0 No request to put device in sleep mode 1 Device is to be placed in sleep mode. Instruction fetching is halted, snooping of L1 caches is disabled, and most functional blocks are shut down in both the e500 core and the system logic. Reserved Doze status 0 Device is not in doze mode 1 The MPC8540 is in doze mode because POWMGTCSR[DOZ] is set or because HID0[DOZE] and MSR[WE] (in the e500 core) are set. The core has halted instruction fetching, but all other functional blocks in the core and device are running. Description 13 14 — SLP 15–27 28 — DOZING 29 NAPPING Nap status 0 Device is not in nap mode. 1 The MPC8540 is in nap mode because HID0[NAP] and MSR[WE] are set. The core has halted instruction fetching, snooping of the L1 caches is disabled, and all of the core’s functional units except the timer facilities are shut down. All functional blocks in the device are running. SLPING Sleep status 0 Device is not attempting to reach sleep mode. 1 The device is attempting to SLEEP because POWMGTCSR[SLP] is set or because HID0[SLEEP] and MSR[WE] (in the e500 core) are set. Most functional blocks in the core and device are shut down or are attempting to shut down. Reserved. Should be cleared. 30 31 — 18.4.1.13 Machine Check Summary Register (MCPSUMR) Shown in Figure 18-13, MCPSUMR contains bits summarizing some of the sources of a pending machine check interrupt. All MCPSUMR bits function as write-one-to-clear. NOTE Register fields designated as write-one-to-clear are cleared only by writing ones to them. Writing zeros to them has no effect. Note that other conditions can cause a machine check condition not summarized in MCPSUMR. For example, uncorrectable read errors cause the assertion of core_fault_in, which may directly cause a machine check (if HID1[RFXE] = 1). If RFXE = 0, the assertion of core_fault_in does not directly cause a machine check interrupt, but must be handled by the block that generated the error. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 18-16 Freescale Semiconductor Memory Map/Register Definition For more information about RFXE, see Section 6.10.2, “Hardware Implementation-Dependent Register 1 (HID1).” 0 28 29 30 31 R000000000000000000 00000000000 WRS SRESET MCP_IN W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0xE_0090 Figure 18-13. Machine Check Summary Register (MCPSUMR) Table 18-16 describes the bit settings of MCPSUMR. Table 18-16. MCPSUMR Field Descriptions Bits 0–28 29 Name — WRS Reserved Watchdog timer machine check 0 Machine check exception was not caused by watchdog timer 1 Machine check was caused by a soft reset condition from the e500 watchdog timer as configured in the core’s TSR. Specifically, TSR[WRS] = 01 and a watchdog reset condition occurred. Description 30 SRESET Soft reset machine check 0 Machine check exception was not caused by SRESET assertion 1 Machine check exception was caused by the assertion of the SRESET input signal MCP_IN MCP signal asserted 0 Machine check exception was not caused by MCP assertion 1 Machine check exception was caused by the assertion of the MCP input signal 31 18.4.1.14 Processor Version Register (PVR) Shown in Figure 18-14, the PVR contains the e500 processor version number. It is a memory-mapped copy of the PVR in the e500 core (and is therefore accessible to external devices). See Section 6.5.3, “Processor Version Register (PVR).” Section 5.2, “e500 Processor and System Version Numbers,” lists the complete values for the MPC8540. 0 15 16 31 R W Reset Offset Version Revision 0x8020_nnnn 0xE_00A0 Figure 18-14. Processor Version Register (PVR) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 18-17 Global Utilities Table 18-17 describes the fields of PVR. Table 18-17. PVR Field Descriptions Bits 0–15 Name Description Version A 16-bit number that identifies the version of the processor. Different version numbers indicate major differences between processors, such as which optional facilities and instructions are supported. 16–31 Revision A 16-bit number that distinguishes between implementations of the version. Different revision numbers indicate minor differences between processors having the same version number, such as clock rate and engineering change level. 18.4.1.15 System Version Register (SVR) Shown in Figure 18-15, the SVR contains the system version number for the MPC8540 implementation. This value can also be read though the SVR SPR of the e500 core. See Section 6.5.4, “System Version Register (SVR).” Section 5.2, “e500 Processor and System Version Numbers,” lists the complete values for the MPC8540. 0 31 R W Reset Offset System Version 0x8030_nnnn 0xE_00A4 Figure 18-15. System Version Register (SVR) Table 18-18 describes the fields of SVR. Table 18-18. SVR Field Descriptions Bits 0–31 Name SV Description System version for the MPC8540 system logic 18.4.1.16 Clock Out Control Register (CLKOCR) Shown in Figure 18-16, the CLKOCR contains control bits that select the clock sources to be placed on the clock out (CLK_OUT) signal. 0 1 29 30 31 R ENB W Reset Offset 00000000000000000000000000000 CLK_SEL 0000_0000_0000_0000_0000_0000_0000_0000 0xE_0E00 Figure 18-16. Clock Out Control Register (CLKOCR) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 18-18 Freescale Semiconductor Memory Map/Register Definition Table 18-19 describes the bit settings of CLKOCR. Table 18-19. CLKOCR Field Descriptions Bits 0 Name ENB Description Clock out enable 0 CLK_OUT signal is tri-stated 1 CLK_OUT signal is driven according to CLKOCR[CLK_SEL] Reserved 1–29 — 30–31 CLK_SEL Clock out select 00 CCB (platform) clock 01 CCB (platform) clock divided by 2 10 SYSCLK (echoes SYSCLK input) 11 SYSCLK divided by 2 (demonstrates platform PLL lock) 18.4.1.17 DDR DLL Control Register (DDRDLLCR) The DDRDLLCR, shown in Figure 18-17, contains bits that allow control and debug of the DDR SDRAM controller’s DLL. The DLL delay chain consists of 128 tap points. 0 1 4 5 7 8 15 16 20 21 23 24 31 R OVRD W Reset Offset 0000 COURSE_ADJ TAP_SEL 0 0 0 0 0 COURSE_SET TAP 0000_0000_0000_0000_0000_0000_0000_0000 0xE_0E10 Figure 18-17. DDR DLL Control Register (DDRDLLCR) Table 18-20 describes the bit settings of DDRDLLCR. Table 18-20. DDRDLLCR Field Descriptions Bits 0 Name OVRD Description Override mode 0 Override mode disabled 1 Override of current delay chain tap point with the TAPSEL tap point enabled Reserved 1–4 5–7 — COURSE_ADJ Course adjustment value to be used by the DLL when in override mode (OVRD = 1). The course adjustment is the number of CCB (platform) clock cycles of delay to inject before the delay chain. When leaving override mode (OVRD cleared), this course adjustment point serves as the starting point for a dynamic search for a lock point. TAP_SEL TAP select value to be used by the DLL when in override mode (OVRD = 1). Selects the tap point within the delay chain. When leaving override mode (OVRD cleared) this tap point serves as the starting point for a dynamic search for a lock point. Reserved 8–15 16–20 — MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 18-19 Global Utilities Table 18-20. DDRDLLCR Field Descriptions (continued) Bits Name Description 21–23 COURSE_SET Reports the current course delay setting found by the dynamic search algorithm that produced a lock. Measured in CCB clock cycles 24–31 TAP Reports the tap value found by the dynamic search algorithm that produced a lock. Measured in tap points 18.4.1.18 Local Bus DLL Control Register (LBDLLCR) Shown in Figure 18-18, the LBDLLCR contains control bits that allow debug of the local bus controller’s DLL. The delay chain of the DLL is made up of 128 tap points. 0 1 4 5 7 8 15 16 20 21 23 24 31 R OVRD W Reset Offset 0000 COURSE_ADJ TAP_SEL 0 0 0 0 0 COURSE_SET TAP 0000_0000_0000_0000_0000_0000_0000_0000 0xE_0E20 Figure 18-18. Local Bus DLL Control Register (LBDLLCR) Table 18-21 describes the bit settings of LBDLLCR. Table 18-21. LBDLLCR Field Descriptions Bits 0 Name OVRD Description Override mode 0 Override mode disabled 1 Override of current delay chain tap point with the TAPSEL tap point enabled Reserved 1–4 5–7 — COURSE_ADJ Course adjustment value to be used by the DLL when in override mode (OVRD = 1). The course adjustment is the number of CCB clock cycles of delay to inject before the delay chain. When leaving override mode (OVRD cleared) this course adjust point serves as the starting point for a dynamic search for a lock point. TAP_SEL TAP select value to be used by the DLL when in override mode (OVRD = 1). Selects the tap point within the delay chain. When leaving override mode (OVRD cleared) this tap point serves as the starting point for a dynamic search for a lock point. Reserved 8–15 16–20 — 21–23 COURSE_SET Reports the current course delay setting found by the dynamic search algorithm that produced a lock. Measured in CCB clock cycles 24–31 TAP Reports the tap value found by the dynamic search algorithm that produced a lock. Measured in tap points MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 18-20 Freescale Semiconductor Functional Description 18.5 Functional Description This section describes the global utilities from a functional perspective. 18.5.1 Power Management The MPC8540 has features to minimize power consumption at several levels. Dynamic power management locally minimizes power consumption when a block is idle. Software can also shut down clocks to individual blocks when they are not needed through a memory-mapped register (DEVDISR). Additionally, software running on the e500 core can access the core’s SPRs to put the device into doze, nap, or sleep power down state. Finally, software can access a memory-mapped register (POWMGTCR) in the global utilities block to put the device in the doze or sleep states. Note that the software that writes to either DEVDISR or POWMGTCR can be running either on the e500 core or on an external master that can write to the MPC8540 memory-mapped registers through the RapidIO or PCI interfaces. These features are described in further detail in this section. 18.5.1.1 Relationship Between Core and Device Power Management States The MPC8540 has three low-power states: doze, nap, and sleep. The mapping of core and device power management states is shown in Figure 18-19 showing state transitions from the perspective of the e500 core. Full On ¬core_halt core_halt ¬core_halt Core-Halted (Doze) core_stop & core_halt ¬core_stop core_halt & ¬core_stop Core-Stopped (Nap) ¬core_tben core_tben core_stop Core-Stopped (Sleep) ¬core_tben Figure 18-19. e500 Core Power Management State Diagram MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 18-21 Global Utilities For each operating state represented in the diagram, the core’s state is listed first, with the corresponding state of the MPC8540 shown beneath it in parenthesis. Note that there are many other variables that control the state transitions between MPC8540 power management states. These additional variables are described in more detail in Section 18.5.1.7, “Power-Down Sequence Coordination.” Table 18-22 lists basic characteristics of the low-power modes and the full on mode. Table 18-22. MPC8540 Power Management Modes—Basic Description Core Responds To M ode Description Snoop Full On Doze Nap All units operating normally Core stops dispatching new instructions (core is halted) Core is stopped with clocks off except to time base Should flush data cache before entering Core is stopped with clocks off. Clocks powered down to all blocks (including core time base) except to the interrupt controller (PIC) unit Yes Yes No Interrupts READY ASLEEP Yes Yes Yes Asserted Negated Negated Negated Negated Negated Signal States Sleep No Yes Negated Asserted 18.5.1.2 CKSTP_IN Is Not Power Management CKSTP_IN is not described here because it is not considered a power management signal, although asserting it does stop the core and a stopped core is technically in a low-power mode. CKSTP_IN is described in Section 18.3.2, “Detailed Signal Descriptions.” 18.5.1.3 Dynamic Power Management Many blocks in the MPC8540 can dynamically turn off clocks within the block when sections of the block are idle. This feature is always enabled and occurs automatically. 18.5.1.4 Shutting Down Unused Blocks As described in Section 18.4.1.11, “Device Disable Register (DEVDISR),” DEVDISR provides a way to shut down certain functional blocks within the MPC8540 when they are not needed in a particular system. DEVDISR can be written by the e500 core or by an external master. Powering down a block in this way turns off all clocks to that block. DEVDISR was designed with the expectation that, once initialized by software, it would be modified only by a hard system reset (HRESET). It is recommended that this register be written only during system initialization. Blocks disabled by DEVDISR must not be re-enabled without a hard reset. (Setting DEVDISR[TB] disables the core’s timer facilities, and setting DEVDISR[E500] places the core in the core-stopped state in which it does not respond to MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 18-22 Freescale Semiconductor Functional Description interrupts.) The results of re-enabling previously disabled blocks (by clearing the corresponding DEVDISR field) without a hard reset are boundedly undefined. NOTE Functional blocks disabled using DEVDISR cannot respond to configuration accesses. Any access to configuration, control, and status registers of a disabled block is a programming error. 18.5.1.5 Software-Controlled Power-Down States e500 software can place the device in doze, nap, or sleep power-down states by writing to HID0 in the core. In addition, external masters can write to the memory-mapped POWMGTCR in the MPC8540 to cause the device to enter doze or sleep modes. 18.5.1.5.1 Doze Mode In doze mode, the e500 core suspends instruction execution, significantly reducing the power consumption of the core. Snooping of the L1 data cache is still supported and thus the data in the data cache is kept coherent. Interrupts directed to the core as described in Section 10.1.3, “Interrupts to the Processor Core,” are monitored by the device and cause the MPC8540 to use the defined handshake mechanism to exit the core from doze mode to allow the core to recognize and process the interrupt; however, unless the interrupt subroutine turns off (or masks) the control bits that enabled doze mode (MSR[WE], and HID0[DOZE]), the device re-enters doze mode after the interrupt has been serviced. See Section 18.5.1.8, “Interrupts and Power Management,” for more information. The e500 core’s timer facilities are still enabled during doze mode, and core time base interrupts can be generated. All device logic external to the core remains fully operational in doze mode. 18.5.1.5.2 Nap Mode In nap mode all clocks internal to the e500 core are turned off except for its timer facilities clock (the core time base). The L1 caches do not respond to snoops in nap mode, so if coherency with external I/O transactions is required, the L1 cache must be flushed before entering nap mode. Similar to doze mode, interrupts occurring in nap mode cause the device to wake up the e500 core in order to service the interrupt. However, unless the interrupt service routine changes the control bits that caused the device to enter nap mode (MSR[WE], and HID0[NAP]), the MPC8540 returns to nap mode after the interrupt is serviced. See Section 18.5.1.8, “Interrupts and Power Management,” for more information. All device logic external to the e500 core remains fully operational in nap mode. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 18-23 Global Utilities 18.5.1.5.3 Sleep Mode In sleep mode, all clocks internal to the e500 core are turned off, including the timer facilities clock. All I/O interfaces in the device logic are also shut down. Only the clocks to the MPC8540 PIC are still running so that an external interrupt can wake up the device. Note that the DDR controller does not shut down unless DDR_SDRAM_INTERVAL[REFINT] is set to a non-zero value. See Section 9.4.1.7, “DDR SDRAM Interval Configuration (DDR_SDRAM_INTERVAL),” for details. After the core and I/O interfaces have shut down, ASLEEP is asserted and READY is negated. NOTE Only external interrupts can wake the MPC8540 from sleep mode. Internal interrupt sources like the core interval timer or watchdog timer depend on an active clock for their operation and these are disabled in sleep mode. 18.5.1.6 Power Management Control Fields The e500 core provides the following fields to signal power management requests to the MPC8540 device logic. • • • • MSR[WE]—Used to qualify the values of HID0[DOZE,NAP,SLEEP] in the generation of the internal doze, nap, and sleep signals HID0[DOZE]—Signals the MPC8540 to initiate doze mode HID0[NAP]—Signals the MPC8540 to initiate nap mode HID0[SLEEP]—Signals the MPC8540 to initiate sleep mode These register fields and their functional relationship are shown in Figure 18-20. The e500 Reference Manual has details on accessing these power management control bits. An external master can also initiate power management requests by setting the DOZ or SLP bits in the memory-mapped power management control and status register (POWMGTCSR). Because the core responds to snoops while dozing but not while napping, maintaining cache coherency requires significant preparation by the core before entering nap mode. For this reason only the core can initiate a nap during normal operation while other masters can initiate a doze. 18.5.1.7 Power-Down Sequence Coordination To preserve cache coherency and otherwise avoid loss of system state, the core’s transition to low-power modes is coordinated by a set of handshaking signals, shown in Figure 18-20, and protocols with all other MPC8540 functional blocks that respond to power-down requests. The mode-transition protocol is executes automatically under these conditions and is shown in Figure 18-19 and described in Table 18-23. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 18-24 Freescale Semiconductor Functional Description The column in Table 18-23 showing the global utilities block as initiating a low-power mode corresponds to the external masters that can write to the POWMGTCSR that resides in the global utilities block. For the MPC8540, these are the RapidIO and PCI/PCI-X interfaces. However, note that the core can also write to POWMGTCSR and, in this case, can initiate power management through the global utilities block. Table 18-23. Power Management Entry Protocol and Initiating Functional Units LowPower Mode Doze Nap Initiating Functional Unit Entry Protocol Global Utilities 1. Assert core_halt input to core. 2. Wait for core_halted handshake from core. 1. Follow doze protocol 2. Assert core_stop input to core. 3. Wait for core_stopped handshake from core. 1. 2. 3. 4. Follow doze protocol; send stop requests to rest of device. Follow nap protocol. Wait for all interfaces to acknowledge stop requests. Assert ASLEEP, negate READY, power down all clocks except to PIC unit √ — Core √ √ Sleep √ √ As shown in Figure 18-20, the e500 core enters low-power modes only in response to the core_halt, core_stop, or core_tben inputs from the MPC8540’s power management logic. These inputs may be prompted by the core (by setting the NAP, DOZE, or SLEEP bits in the HID0 when enabled by setting MSR[WE]) or by an external master (by setting POWMGTCSR[DOZ,SLP]. Figure 18-20 shows how all the clocking to the core timer facilities is disabled by clearing HID0[TBEN]. When enabled, (HID0[TBEN] = 1), the clock source is either the CCB clock divided by eight (the default) or a synchronized version of the RTC input. For more details, see Section 6.10.1, “Hardware Implementation-Dependent Register 0 (HID0).” MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 18-25 Global Utilities e500 Core Complex MSR[WE] sleep HID0[SLEEP] Device Logic POWMGTCSR[SLP] nap HID0[NAP] doze HID0[DOZE] POWMGTCSR[DOZ] CKSTP_IN HRESET No Instruction Execution Stopped? Yes Stop Fetching Instructions core_halt Processor interrupt sources (see Table 10-2 and Table 10-3) DEVDISR[E500] No Snoopable Traffic Stopped? Yes core_halted No Clock Distribution Stopped? Yes Stop Clock Distribution core_stop Any transactions that core_stopped could have snooped are complete DEVDISR[TB] Core Timer Facilities Core Timer Clock (Time Base) HID0[TBEN] core_tben RTC (Sampled and synchronized) core_tbint ÷8 HID0[SEL_TBCLK] CCB Clock Note: All signals shown are internal to the device. Figure 18-20. MPC8540 Power Management Handshaking Signals MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 18-26 Freescale Semiconductor Functional Description 18.5.1.8 Interrupts and Power Management Whether low-power modes are automatically re-enabled after an interrupt is processed differs depending on whether the low power mode was entered due to a write to the core MSR[WE] bit or the low power mode was entered due to a write to POWMGTCSR. 18.5.1.8.1 Interrupts and Power Management Controlled by MSR[WE] When an interrupt is asserted to the CPU, the core complex saves portions of the MSR to MCSRR1, CSRR1, or SRR1 (depending on the type of interrupt), and restores those values on return from the routine. MSR[WE], which gates the doze, nap, and sleep power management outputs (internal device signals) from the core complex, is always among the bits saved and restored; hence these outputs negate to the MPC8540 power management logic when the interrupt begins processing in the core. They return to their previous state when the core executes an rfi, rfci, or rfmci instruction. Section 10.1.3, “Interrupts to the Processor Core,” lists interrupts that cause the MPC8540 to wake up. NOTE Returning doze, nap, and sleep signals to their original state when MSR[WE] is restored differs from how power management is implemented on earlier PowerPC devices where MSR[POW], which enables power-down requests, is cleared when the processor exits a low-power state and is not automatically restored, as it is in Book E implementations. 18.5.1.8.2 Interrupts and Power Management Controlled by POWMGTCSR The IRQ_MSK and CI_MSK fields of the POWMGTCSR register prevent int interrupts or cint critical interrupts from waking the device from a low power state. This is true regardless of the method used to enter the low power state. Any unmasked interrupt (not masked by the mask bits in the POWMGTCSR register) causes the POWMGTCSR[DOZ,SLP] fields to be cleared when it occurs. When such an interrupt occurs, the device returns to the normal operating mode and does not automatically attempt to return to a low power state after the interrupt is handled. Note that interrupts caused by the unconditional debug event (UDE) and machine check (MCP) signals are not masked by the IRQ_MSK and CI_MSK fields; therefore, when these signals assert, the POWMGTCSR[DOZ,SLP] fields are cleared and the device will return to full power operation. See Section 18.4.1.12, “Power Management Control and Status Register (POWMGTCSR),” for detailed information about the bits of POWMGTCSR. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 18-27 Global Utilities Note also that unmasked interrupts that occur while the device is in the process of going into the sleep state (before sleep is completely attained) can also cause the device to clear the POWMGTCSR[DOZ,SLP] fields and return the device to full power operation. 18.5.1.9 Snooping in Power-Down Modes When the MPC8540 is in doze mode, the e500 core is in the core-halted state and it snoops its L1 caches and full coherency is maintained. In deeper power-down modes, however, the e500 core does not respond to snoops. The MPC8540 does not perform dynamic bus snooping as described in the e500 Reference Manual. That is, when the e500 core is in the core-stopped state (which is the state of the core when the MPC8540 is in either the nap or sleep state), the core is not awakened to perform snoops on global transactions. Therefore, before entering nap or sleep modes, the L1 caches should be flushed if coherency is required during these power-down modes. 18.5.1.10 Software Considerations for Power Management Setting MSR[WE] generates a request to the MPC8540 logic (external to the core complex) to enter a power saving state. It is assumed that the desired power-saving state (doze, nap, or sleep) was set up by setting the appropriate HID0 bit, typically at system start-up time. Setting WE has no direct effect on instruction execution, but is reflected on the internal doze, nap, and sleep signals, depending on the HID0 settings. To ensure a clean transition into and out of a power-saving mode, the following program sequence is recommended: sync mtmsr (WE) isync br loop loop: 18.5.1.11 Requirements for Reaching and Recovering from Sleep State In order to successfully reach the sleep state, I/O traffic to the device must be stopped. The logic that controls the power down sequence waits for all I/O interfaces to become idle. In some applications this may happen eventually without actively shutting down interfaces, but most likely, software will have to take steps to shut down the TSEC, FEC, PCI, and RapidIO interfaces before issuing the command (either the write to the core MSR[WE] as described above or writing to POWMGTCSR) to put the device into sleep state. The RapidIO and PCI interfaces will begin retrying inbound transactions before entering a power down state. Upon exiting sleep, the RapidIO interface begins its training sequence to re-establish the link. The PCI interface, however, could potentially be in an unknown state when it exits sleep if it was in the middle of a retry sequence when its internal clocks were shut down. Therefore it is strongly recommended that system software clear the memory space bit in the PCI Bus Command Register before putting the device in sleep mode. Software may also need to set the Agent Config MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 18-28 Freescale Semiconductor Functional Description Lock bit of the PCI Bus Function Register so that the device will not respond to configuration transactions. Upon exiting sleep mode, software should return these configuration bits to their normal state. 18.5.2 General-Purpose I/O Signals Several groups of signals can optionally be used as general-purpose I/O signals when not being used for their primary function. The general-purpose I/O functionality of these signals can be enabled through configuration registers in the global utilities block. These signals are the following: • PCI_AD[47:40] and PCI_AD[39:32]. When configured as general-purpose I/O, PCI_AD[47:40] function as outputs and PCI_AD[39:32] function as inputs. PCI_AD[47:32] can be used as general-purpose I/O if the PCI/PCI-X interface is disabled or if the PCI/PCI-X controller is not in 64-bit mode. TSEC2_RxD[0:7] and TSEC2_TxD[0:7]. TSEC2 pins are fixed as either inputs or outputs based on the direction of the signal’s primary function. The TSEC2_TxD pins are always outputs, so these signals may only be used as outputs when configured as general-purpose I/O. Similarly, the TSEC2_RxD pins are used as inputs when configured as general-purpose I/O. The TSEC2 TxD and RxD pins are available when the TSEC2 block is disabled. The TxD signals can then be enabled as general-purpose outputs and the RxD pins can be enabled as general-purpose inputs. • When configured as general-purpose I/O signals, software can read inputs by reading the associated GPIO data register. Output values can be set by writing the to the associated GPIO data register. For details regarding the control and status of the general-purpose I/O signals, see Section 18.4.1.7, “General-Purpose I/O Control Register (GPIOCR).” 18.5.3 Interrupt and Local Bus Signal Multiplexing The MPC8540 has very little signal multiplexing. Two sets of DMA channel triggering signals can alternately be placed on other signals as follows: • • LCS[5:7] are multiplexed with DMA channel 2 DMA_DREQ2, DMA_DACK2, and DMA_DONE2. IRQ[9:11] are multiplexed with DMA channel 3 DMA_DREQ3, DMA_DACK3, and DMA_DDONE3. For details regarding the selection of the alternate function DMA trigger, see Section 18.4.1.10, “Alternate Function Signal Multiplex Control Register (PMUXCR).” MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 18-29 Global Utilities MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 18-30 Freescale Semiconductor Chapter 19 Performance Monitor This chapter describes the MPC8540 performance monitor facility, which can be used to monitor and optimize performance. The e500 core implements a separate performance monitor for strictly core-related behavior, such as instruction timing and L1 cache operations. This is described in the e500 Reference Manual. Section 19.4.7, “Performance Monitor Events,” briefly describes the events that can be monitored. Refer to the individual chapters for a better understanding of these events. 19.1 Introduction The MPC8540 includes a performance monitor facility that can be used to monitor and record selected behaviors of the integrated device. Although the performance monitor described here is similar in many respects to the performance monitor facility implemented on the e500 core, it differs in that it is implemented using memory-mapped registers and it counts events outside the e500 core, for example, RapidIO, PCI, DDR, and L2 cache events. Performance monitor counters (PMC0–PMC8) are used to count events selected by the performance monitor local control registers. PMC0 is a 64-bit counter specifically designated to count cycles. PMC1–PMC8 are 32-bit counters that can monitor 64 counter-specific events in addition to counting 64 reference events. The benefits of the on-chip performance monitor are numerous and include the following: • Because some systems or software environments are not easily characterized by signal traces or benchmarks, the performance monitor can be used to understand the MPC8540’s behavior in any system or software environment. The performance monitor facility can be used to aid system developers when bringing up and debugging systems. System performance can be increased by monitoring memory hierarchy behavior. This can help to optimize algorithms used to schedule or partition tasks and to refine the data structures and distribution used by each task. • • 19.1.1 Overview Figure 19-1 is a high-level block diagram of the performance monitor, which consists of a global control register (PMGC0), one 64-bit counter (PMC0), eight 32-bit counters, and two control registers per counter (18 total control registers). The global control register PMGC0 affects all MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 19-1 Performance Monitor counters and takes priority over local control registers. The local control registers are divided into two groups as follows: • Local control A registers control counter freezing, overflow condition enable, event selection, and burstiness. Local control register PMLCA0, which controls counter PMC0, does not contain event selection because PMC0 counts only cycles. Local control B registers control the start and stop triggering, contain the counters’ threshold values, and the value of the threshold multiplier. Local control register PMLCB0, which controls PMC0, does not contain threshold information because PMC0 only counts cycles. Event Signals Performance Monitor Output Signals Interrupt (int) • Event 0 Control Logic Event 1 Counters and Registers Global Control Register PMGC0 Local Control Registers PMLCA0 PMLCA1 PMLCA2 PMLCA3 ... PMLCA8 Counters PMC0 (upper) PMC0 (lower) PMLCB0 PMLCB1 PMLCB2 PMLCB3 ... PMLCB8 Read Data Event 2 Event 3 Event 4 ... Event 8 PMC1 PMC2 PMC3 ... PMC8 32-bit Figure 19-1. Performance Monitor Block Diagram NOTE Note that the performance monitor implemented on the core also uses registers named PMCn and PMLCxn; however, these registers are implemented on-chip, they are accessed using the EIS-defined mtpmr and mtpmr instructions, and they do not interact with the memory-mapped registers shown in Figure 19-1. Performance monitor events are signalled by the functional blocks in the integrated device and are selectively recorded in the PMCs. Sixty-four of these events are referred to as reference events, MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 19-2 Freescale Semiconductor External Signal Descriptions which can be counted on any of the eight counters. Counter-specific events can be counted only on the counter where the event is defined. The performance monitor can generate an interrupt on overflow. Several control registers specify how a performance monitor interrupt is signalled. The PMCs can also be programmed to freeze when an interrupt is signalled. 19.1.2 Features The MPC8540 performance monitor offers a rich set of features that permits a complete performance characterization of the implementation. These features include: • • • • • • • • • One 64-bit counter exclusively dedicated to counting cycles Eight 32-bit counters that count the occurrence of selected events One global control register (affects all counters) and two local control registers per counter Ability to count up to 64 reference events that may be counted on any of the eight 32-bit counters Ability to count up to 512 counter-specific events Triggering and chaining capability Duration threshold counting Burstiness feature that permits counting of burst events with a programmable time between bursts Ability to generate an interrupt on overflow 19.2 External Signal Descriptions The performance monitor does not have any signals that are driven externally (off-chip) but it does assert the internal interrupt (int) signal on a performance monitor interrupt condition. 19.3 Memory Map and Register Definition Performance monitor registers reside in the run-time register block starting at offset 0xE_1000. This section describes the registers implemented to support the performance monitor facilities. Table 19-1 lists the performance monitor registers. These registers can be read or written only with 32-bit accesses. 19.3.1 Register Summary The performance monitor uses nine counter registers and a group of local control registers that are used to specify the method of counting. Two local control registers are associated with each counter in addition to a global control register that applies to all counters. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 19-3 Performance Monitor Table 19-1. Control Register Memory Map Address Offset (in Hex) 0xE_1000 0xE_1010 0xE_1014 0xE_1018 0xE_101C 0xE_1020 0xE_1024 0xE_1028 0xE_1030 0xE_1034 0xE_1038 0xE_1040 0xE_1044 0xE_1048 0xE_1050 0xE_1054 0xE_1058 0xE_1060 0xE_1064 0xE_1068 0xE_1070 0xE_1074 0xE_1078 0xE_1080 0xE_1084 0xE_1088 0xE_1090 0xE_1094 0xE_1098 Register PMGC0—Performance monitor global control register PMLCA0—Performance monitor local control register A0 PMLCB0—Performance monitor local control register B0 PMC0 (upper)—Performance monitor counter 0 upper PMC0 (lower)—Performance monitor counter 0 lower PMLCA1—Performance monitor local control register A1 PMLCB1—Performance monitor local control register B1 PMC1—Performance monitor counter 1 PMLCA2—Performance monitor local control register A2 PMLCB2—Performance monitor local control register B 2 PMC2—Performance monitor counter 2 PMLCA3—Performance monitor local control register A3 PMLCB3—Performance monitor local control register B3 PMC3—Performance monitor counter 3 PMLCA4—Performance monitor local control register A4 PMLCB4—Performance monitor local control register B4 PMC4—Performance monitor counter 4 PMLCA5—Performance monitor local control register A5 PMLCB5—Performance monitor local control register B 5 PMC5—Performance monitor counter 5 PMLCA6—Performance monitor local control register A6 PMLCB6—Performance monitor local control register B6 PMC6—Performance monitor counter 6 PMLCA7—Performance monitor local control register A7 PMLCB7—Performance monitor local control register B7 PMC7—Performance monitor counter 7 PMLCA8—Performance monitor local control register A8 PMLCB8—Performance monitor local control register B8 PMC8—Performance monitor counter 8 Access R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W R/W Reset 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 Section/Page 19.3.2.1/19-5 19.3.2.2/19-5 19.3.2.2/19-5 19.3.3.1/19-9 19.3.3.1/19-9 19.3.2.2/19-5 19.3.2.2/19-5 19.3.3.1/19-9 19.3.2.2/19-5 19.3.2.2/19-5 19.3.3.1/19-9 19.3.2.2/19-5 19.3.2.2/19-5 19.3.3.1/19-9 19.3.2.2/19-5 19.3.2.2/19-5 19.3.3.1/19-9 19.3.2.2/19-5 19.3.2.2/19-5 19.3.3.1/19-9 19.3.3.1/19-9 19.3.2.2/19-5 19.3.3.1/19-9 19.3.2.2/19-5 19.3.2.2/19-5 19.3.3.1/19-9 19.3.2.2/19-5 19.3.2.2/19-5 19.3.3.1/19-9 In addition to these registers, the interrupt control provides four pairs of mask registers that can be used to monitor message, interprocessor, timer, and external interrupts. See Section 10.3.4, “Performance Monitor Mask Registers (PMMRs).” MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 19-4 Freescale Semiconductor Memory Map and Register Definition 19.3.2 Control Registers This section describes the performance monitor control registers in detail. 19.3.2.1 Performance Monitor Global Control Register (PMGC0) The performance monitor global control register (PMGC0), shown in Figure 19-2, is a 32-bit register used to control all PMCs. 0 1 2 3 31 R FAC PMIE FCECE W Reset Offset 00000000000000000000000000000 0000_0000_0000_0000_0000_0000_0000_0000 0xE_1000 Figure 19-2. Performance Monitor Global Control Register (PMGC0) Table 19-2 describes PMGC0 fields. Table 19-2. PMGC0 Field Descriptions Bits 0 Name FAC Description Freeze all counters. 0 PMCs are incremented (if permitted by other PMGC0/PMLC bits). 1 PMCs are not incremented. Set by hardware when an interrupt is signalled and FCECE =1 Performance monitor interrupt enable. Interrupts are caused by PMC overflows. 0 Interrupts are disabled. 1 Interrupts are enabled and occur when an enabled condition or event occurs. 1 PMIE 2 FCECE Freeze counters on enabled condition or event. An enabled condition or event is defined as: (DISCOUNT) The msb = 1 in PMCn and PMLCAn[CE] = 1. The use of the trigger and freeze counter conditions depends on the enabled condition. 0 PMCs can be incremented (if permitted by other control bits). 1 PMCs can be incremented (if permitted by other control bits) only until an enabled condition or event occurs, at which time PMGC0[FAC] is set. It is up to software to clear FAC. — Reserved 3–31 19.3.2.2 Performance Monitor Local Control Registers (PMLCAn and PMLCBn) The performance monitor local control registers (PMLCAn and PMLCBn) are used to control the operation of the PMCs. The performance monitor local control A and B registers are paired 32-bit control registers that are associated with an individual counter to specify how the counter is used and what event is monitored on that counter. Figure 19-3 shows the performance monitor local control A0 register (PMLCA0). MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 19-5 Performance Monitor 0 1 4 5 6 31 R FC W Reset Offset 0000 CE 00000000000000000000000000 0000_0000_0000_0000_0000_0000_0000_0000 0xE_1010 Figure 19-3. Performance Monitor Local Control Register A0 (PMLCA0) Table 19-3 describes PMLCA0 fields. Table 19-3. PMLCA0 Field Descriptions Bits 0 Name FC Description Freeze counter. Basic counter enable 0 The PMCs are enabled and incremented (if permitted by other SPM control bits). 1 The PMCs are disabled–they do not increment. Reserved Condition enable. Controls counter overflow condition. Should be cleared when PMC0 is used as a trigger or is selected for chaining 0 Overflow conditions for PMC0 cannot occur (PMC0 cannot cause interrupts or freeze counters) 1 Overflow conditions occur when PMC0[msb] is set. Reserved 1–4 5 — CE 6–31 — Figure 19-4 shows the performance monitor local control registers A1–A8. 0 1 4 5 6 8 9 15 16 20 21 25 26 31 R FC W Reset Offset 0000 CE 000 EVENT BSIZE BGRAN BDIST 0000_0000_0000_0000_0000_0000_0000_0000 0xE_1020, 0xE_1030, 0xE_1040, 0xE_1050, 0xE_1060, 0xE_1070, 0xE_1080, 0xE_1090 Figure 19-4. Performance Monitor Local Control Registers A1–A8 (PMLCA1–PMLCA8) Table 19-4 describes PMLCAn fields. Table 19-4. PMLCA1–PMLC8 Field Descriptions Bits 0 Name FC Description Freeze counter 0 The PMCs are incremented (if permitted by other PMC control bits). 1 The PMCs are not incremented (if permitted by other PMC control bits). Reserved 1–4 — MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 19-6 Freescale Semiconductor Memory Map and Register Definition Table 19-4. PMLCA1–PMLC8 Field Descriptions (continued) Bits 5 Name CE Description Condition enable 0 Overflow conditions for PMC n cannot occur (PMCn cannot cause interrupts or freeze counters). Should be cleared when PMCn is used as a trigger or is selected for chaining. 1 Overflow conditions occur when PMCn[msb] is set. Reserved 6–8 9–15 16–20 — EVENT Event selector. Up to 128 events selectable. See Table 19-10 for definition of events. BSIZE Burst size. Fewest event occurrences that constitute a burst, that is, a rapid sequence of events followed by a relatively long pause. A value less than two implies regular event counting. Any non-threshold, regular event may be counted in a bursty fashion. See Section 19.4.6, “Burstiness Counting,” for more information. 21–25 26–31 BGRAN Burst granularity. The maximum number of clock cycles between events that are considered part of a single burst. See Section 19.4.6, “Burstiness Counting.” BDIST Burst distance (used with TBMULT). The number of clock cycles between bursts. Must be set to a value greater than BSIZE for proper burstiness counting behavior. 00_0000 Regular counting Figure 19-5 shows the performance monitor local control B0 register (PMLCB0). 01 2 5 67 8 11 12 13 14 15 16 31 R00 TRIGONSEL W Reset Offset 00 TRIGOFFSEL TRIGONCNTL TRIGOFFCNTL 0000000000000000 0000_0000_0000_0000_0000_0000_0000_0000 0xE_1014 Figure 19-5. Performance Monitor Local Control Register B0 (PMLCB0) Table 19-5 describes PMLCB0 fields. Table 19-5. PMLCB0 Field Descriptions Bits 0–1 2–5 Name — TRIGONSEL Reserved Trigger-on select. The number of the counter that starts event counting. When the specified counter’s TRIGONCNTL event overflows, the current counter begins counting. No triggering occurs if the value is self-referential, that is, when set to the current counter number. Reserved Trigger-off select. The number of the counter that stops event counting. When the specified counter’s TRIGONCNTL event overflows, the current counter stops counting. No triggering occurs if the value is self-referential, that is, when set to the current counter number. Description 6–7 8–11 — TRIGOFFSEL MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 19-7 Performance Monitor Table 19-5. PMLCB0 Field Descriptions (continued) Bits 12–13 Name Description TRIGONCNTL Trigger-on control. Indicates the condition under which triggering to start counting occurs 00 Trigger off (no triggering to start) 01 Trigger on change 10 Trigger on overflow 11 Reserved 14–15 TRIGOFFCNTL Trigger-off control. Indicates the condition under which triggering to stop occurs 00 Trigger off (no triggering to stop) 01 Trigger on change 10 Trigger on overflow 11 Reserved 16–31 — Reserved Figure 19-6 shows performance monitor local control registers B1–B8. 01 2 5 67 8 11 12 13 14 15 16 20 21 23 24 25 26 31 R00 TRIGONSEL W Reset Offset 00 TRIGOFFSEL TRIGONCNTL TRIGOFFCNTL 00000 TBMULT 00 THRESHOLD 0000_0000_0000_0000_0000_0000_0000_0000 0xE_1024, 0xE_1034, 0xE_1044, 0xE_1054, 0xE_1064, 0xE_1074, 0xE_1084, 0xE_1094 Figure 19-6. Performance Monitor Local Control Registers B1–B8 (PMLCB1–PMLCB8) Table 19-5 describes PMLCBn fields. Table 19-6. PMLCBn Field Descriptions Bits 0–1 2–5 Name — TRIGONSEL Reserved Trigger-on select. Set this field equal to the number of the counter that should trigger event counting to start. When the specified counter’s TRIGONCNTL event overflows, the current counter begins counting. No triggering occurs when TRIGONSEL = current counter. Reserved Trigger-off select. Set this field equal to the number of the counter that should trigger event counting to stop. When the specified counter’s TRIGONCNTL event overflows, the current counter stops counting. No triggering occurs when TRIGOFFSEL = current counter. Description 6–7 8–11 — TRIGOFFSEL 12–13 TRIGONCNTL Trigger-on control. Indicates the condition under which triggering to start counting occurs 00 Trigger off (no triggering to start) 01 Trigger on change 10 Trigger on overflow 11 Reserved MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 19-8 Freescale Semiconductor Memory Map and Register Definition Table 19-6. PMLCBn Field Descriptions (continued) Bits Name Description 14–15 TRIGOFFCNTL Trigger-off control. Indicates the condition under which triggering to stop occurs 00 Trigger off (no triggering to stop) 01 Trigger on change 10 Trigger on overflow 11 Reserved 16–20 21–23 — TBMULT Reserved Threshold and burstiness multiplier. Threshold events are counted when the event duration exceeds a specified threshold value. The threshold is scaled based on the TBMULT settings. The burst distance for burstiness counting is also scaled using the TBMULT settings. For all events that scale the threshold, the threshold field is multiplied by the factors shown below (ranging from 1 to 128). 000 1 001 2 010 4 011 8 100 16 101 32 110 64 111 128 Reserved Threshold. Only events whose (number of) occurrences exceed this value are counted. By varying the threshold value, software can characterize the events subject to the threshold. For example, if PMC2 counts TSEC BD read latencies for which the duration exceeds the threshold, software can obtain the distribution of TSEC BD read latencies for a given program by monitoring the program repeatedly using a different threshold value each time. 24–25 26–31 — THRESHOLD 19.3.3 Counter Registers This section describes the PMCs in detail. NOTE Because accessing a PMC manually has priority over incrementing it due to event counting, reading or writing a PMC while it is counting may affect the count. Likewise, accessing a performance monitor control register while its target counter is counting may also affect the count. 19.3.3.1 Performance Monitor Counters (PMC0–PMC8) PMC0–PMC8 are used to count events selected by the performance monitor local control registers. PMC0, shown in Figure 19-7, is associated with two 32-bit registers that form a 64-bit counter designated to count clock cycles. PMC0 upper represents the upper 32 bits of counter 0, and PMC0 lower represents the lower 32 bits. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 19-9 Performance Monitor 0 31 32 63 R PMC0 upper W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0xE_1018 0xE_101C PMC0 lower Figure 19-7. Performance Monitor Counter Register 0 (PMC0) Table 19-7 describes PMC0 fields. . Table 19-7. PMC0 Field Descriptions Bits 0–63 Name PMC0 Event count. Counts only clock cycles Description PMC1–PMC8, shown in Figure 19-8, are 32-bit counters that can monitor 64 unique events in addition to the 64 reference events that can be counted on all of these registers. 0 31 R PMCn W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0xE_1028, 0xE_1038, 0xE_1048, 0xE_1058, 0xE_1068, 0xE_1078, 0xE_1088, 0xE_1098 Figure 19-8. Performance Monitor Counter Registers 1–8 (PMC1–PMC8) Table 19-8 describes PMCn fields. Table 19-8. PMCn Field Descriptions Bits 0–31 Name PMCn Description Event count. An overflow is indicated when the msb = 1. Manually setting the msb can cause an immediate interrupt. 19.4 Functional Description This section describes the use of some features of the performance monitor. 19.4.1 Performance Monitor Interrupt PMCs can generate an interrupt on an overflow when the msb of a counter changes from 0 to 1. For the interrupt to be signalled, the condition enable bit (PMLCAn[CE]) and performance monitor interrupt enable bit (PMGC0[PMIE]) must be set. When an interrupt is signalled and the MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 19-10 Freescale Semiconductor Functional Description freeze-counters-on-enabled-condition-or-event bit (PMGC0[FCECE]) is set, PMGC0[FAC] is set by hardware and all of the registers are frozen. Software can clear the interrupt condition by resetting the performance monitor and clearing the most significant bit of the counter that generated the overflow. 19.4.2 Event Counting Using the control registers described in Section 19.3.2, “Control Registers,” the nine PMCs can count the occurrences of specific events. The 64-bit PMC0 is designated to count only clock cycles. However, to provide flexibility, a total of 64 reference events can be counted on any of the 32-bit PMCs (PMC1–PMC8). Additionally, up to 64 unique events can be counted on each 32-bit counter. The performance monitor must be reset before event counting sequences. The performance monitor can be reset by first freezing one or more counters and then clearing the freeze condition to allow the counters to count according to the settings in the performance monitor registers.þCounters can be frozen individually by setting PMLCAn[FC] bits, or simultaneously by setting PMGC0[FAC]. Simply clearing these freeze bits will then allow the performance monitor to begin counting based on the register settings.þ Note that using PMLCAn[FC] to reset the performance monitor resets only the specified counter. Performance monitor registers can be configured through reads or writes while the counters are frozen as long as freeze bits are not cleared by the register accesses. 19.4.3 Threshold Events The threshold feature allows characterization of events that can take a variable number of clock cycles to occur. Threshold events are counted only if the latency is greater than the threshold value specified in PMLCBn[THRESHOLD]. For duration threshold event sequences, the PMC increments only when the duration of the event is equal to or greater than the threshold value. The threshold value is scaled by a multiple specified in PMLCBn[TBMULT]. A threshold event requires two signals: The first indicates when a threshold event sequence begins, and the second indicates when it ends. An internal counter determines when the threshold count is exceeded and when the PMC can increment. This internal counter decrements during a threshold event sequence until it reaches the value of one. A new sequence cannot begin until the current one completes. Additional threshold start signals are ignored during a sequence until a threshold stop signal occurs. If both a start and stop signal are asserted during the same cycle in a current sequence, the stop terminates the current sequence and the start signals the beginning of a new one. However, if both signals are asserted during the same cycle while not in a current event sequence, both signals are ignored. Figure 19-9 is a timing diagram for duration threshold event counting. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 19-11 Performance Monitor An illegal condition exists if the threshold value obtained from PMLCBn[THRESHOLD] and PMLCBn[TBMULT] is less than two. Under these conditions the intent of threshold counting is ambiguous. 1 2 3 4 5 6 7 8 9 10 11 12 PMLCBn[THRESHOLD] PMLCBn[TBMULT] thresh_value thresh_start thresh_stop thresh_counter thresh_exceeded perfmon_counter 1 3 0 3 3 2 1 3 2 1 3 2 1 0 1 2 For this example a threshold value of three indicates that the user wishes to count the number of times a particular event lasts three cycles or longer. Figure 19-9. Duration Threshold Event Sequence Timing Diagram 19.4.4 Chaining By configuring one counter to increment each time another counter overflows, several counters can be chained together to provide event counts larger than 32 bits. Each counter in a chain adds 32 bits to the maximum count. The register chaining sequence is not arbitrary and is specified indirectly by selecting the register overflow event to be counted. Selecting an event has the effect of selecting a source register because all available chaining events, as shown in Table 19-10, are dedicated to specific registers. Note that the chaining overflow event occurs when the counter reaches its maximum value and wraps, not when the register’s msb is set. For this overflow to occur, PMLCAn[CE] should be cleared to avoid signalling an interrupt when the counter’s most-significant bit is set. Note that several cycles may be required for the chained counters to reflect the true count because of the internal delay between when an overflow occurs and a counter increments. 19.4.5 Triggering Triggering allows one counter to start or stop counting on the change of another counter or on the overflow of another counter. More specifically, if PMC1 is set to start or stop counting as a result of a change or overflow in counter PMC2, then counter PMC2 must be identified in the local MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 19-12 Freescale Semiconductor Functional Description control register of counter PMC1. This is done by appropriately setting the trigger-on select bit or trigger-off select bit (PMLCB1[TRIGOFFSEL] or PMLCB1[TRIGONSEL]). Additionally, the condition that triggers the counter must be selected by configuring the corresponding control bits (PMLCB1[TRIGONCNTL] or PMLCB1[TRIGOFFCNTL]). Assuming the counter is enabled by other control register settings, the counter increments (or freezes) when its specified event occurs after the trigger-on (or off) condition occurs. When trigger on and trigger off are both selected, the trigger-off condition is ignored until the trigger-on condition has occurred. Furthermore, when a trigger-off condition occurs, the counter state is preserved; it is not restarted by subsequent trigger-on conditions. Triggering is disabled when the counter’s trigger-select bits specify itself as the trigger source. Similarly, triggering is disabled when the trigger control bits are cleared. 19.4.6 Burstiness Counting The burstiness counting feature makes it easier to characterize events that occur in rapid succession followed by a relatively long pause. As shown in Table 19-9, event bursts are defined by size, granularity, and distance. Table 19-9. Burst Definition Parameter Size Description The minimum number of events constituting a burst Register Field PMLCAn [BSIZE] Granularity The maximum time between individual events counted as members of the same burst PMLCAn [BGRAN] Distance The minimum time between bursts PMLCAn [BDIST] x PMLCBn[TBMULT] Figure 19-10 shows the relationships between size, granularity, and distance. Burstiness counting can be performed for all events except threshold events. Burst size Event occurrence Burst size Burst granularity Burst distance Time (cycles) Figure 19-10. Burst Size, Distance, Granularity, and Burstiness Counting The burstiness size field (PMLCAn[BSIZE]) specifies the minimum number of event occurrences that constitute a burst. A burst is identified when the number of event occurrences equals or exceeds PMLCAn[BSIZE]. Furthermore, these individual event occurrences must be separated by no more clock cycles than the value in the burstiness granularity field (PMLCAn[BGRAN]). Note MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 19-13 Performance Monitor that, although a burst is identified when the minimum number of events occurs, it is not counted until the burst sequence has ended. A burst sequence ends when the specified burstiness granularity is exceeded, at which point the last valid event has occurred for that sequence. PMLCAn[BGRAN] specifies the maximum number of cycles between individual events for them to qualify as members of the same burst sequence. The burstiness distance field (PMLCAn[BDIST]) and threshold/burstiness multiplier field (PMLCBn[TBMULT]) specify the acceptable number of cycles between the end of a burst sequence and the beginning of a new sequence for a group of event occurrences to be counted as an individual burst. The product of the burstiness distance field and the threshold/burstiness multiplier field determine the burstiness distance value used to determine when another burst sequence can begin. Note that the burst distance count begins when a new burst sequence ends and the PMC is incremented. No new burst sequence may begin until the burst distance count has reached zero. After the burst distance count reaches zero, it holds the zero value indicating that a new burst sequence can be counted. The burst distance count begins again when a new burst sequence is identified and counted. Burstiness counting is disabled when the definition of a burst is ambiguous, that is, when the burst size field is less than two, or the burst distance is zero. When burstiness counting is disabled, regular counting is allowed. Figure 19-10 shows that the burst distance is measured from the end of one burst sequence and that a new burst sequence may not begin until the burst distance count expires. Three internal counters track the different values required for burstiness counting. • Burstiness size is monitored by a counter. It is loaded with the value specified in the local control register when the burst granularity counter and the burst distance counters reach zero, and no new event is occurring. It always decrements when the following conditions occur: its value is not already zero, an event occurs, and the burst distance count equals zero. Burstiness granularity is monitored by a counter that is loaded with the specified value in the local control register on the rising edge of an event occurrence whenever the burst distance count equals zero. The granularity counter is decremented (if it has not already reached zero) when an event is not occurring and burst distance count equals zero. Burstiness distance is measured by a counter that is loaded with the product of PMLCBn[BDIST] and PMLCBn[TBMULT] when a burst sequence has been identified and counted. This counter is decremented when burstiness counting is enabled (and the counter has not already reached zero). • • A burst is counted at the end of a burst sequence when the three burst parameter counters are all equal to zero. Figure 19-11 shows a burstiness counting example. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 19-14 Freescale Semiconductor Functional Description 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 PMLCBn[BGRAN] PMLCBn[BSIZE] PMLCBn[BDIST] PMLCBn[TBMULT] 1 5 4 1 (Multiply by 2) Burst size counter Burst granularity counter Burst distance counter Burst count 1 5 1 0 0 4 1 3 0 2 1 1 01 0 5 0 876543210 43 1 2 0 54 1 For this example: count bursts of 5 event occurrences, burst granularity of 1, and acceptable distance between bursts of 8 (product of TBMULT and BDIST). Figure 19-11. Burstiness Counting Timing Diagram 19.4.7 Performance Monitor Events Table 19-10 lists performance monitor events specified in PMLCA1–PMLC8. The event assignment column indicates the event’s type and number, using the following formats: • • Ref:#—Reference events are shared across counters PMC1–PMC8. The number indicates the event. For example, Ref:6 means that PMC1–PMC8 share reference event 6. C[0–8]:#—Counter-specific events. C8 indicates an event assigned to PMC8. Thus C8:62 means PMC8 is assigned event 62 (PIC interrupt wait cycles). Note that with counter-specific events, an offset of 64 must be used when programming the field, because counter-specific events occupy the bottom 64 values of the 7-bit event field where events are numbered. For example, to specify counter-specific event 0, the event field must be programmed to 64. Counter events not specified in Table 19-10 are reserved. Table 19-10. Performance Monitor Events Event Counted Number General Events Nothing System cycles Ref:0 C0 Register counter holds current value CCB (platform) clock cycles Description of Event Counted MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 19-15 Performance Monitor Table 19-10. Performance Monitor Events (continued) Event Counted Number Description of Event Counted DDR Memory Controller Events Cycles a read is returning data from DDR SDRAM Cycles a read or write transfers data from (or to) DDR SDRAM Pipelined read misses in the row open table Pipelined read or write misses in the row open table Non-pipelined read misses in the row open table Non-pipelined read or write misses in the row open table Pipelined read hits in the row open table Pipelined read or write hits in the row open table Non-pipelined read hits in the row open table Non-pipelined read or write hits in the row open table Forced page closings not caused by a refresh Ref:10 Ref:11 C1:57 C2:0 C3:60 C4:0 C5:56 C6:0 C7:57 C8:0 C1:0 Each data beat returned to the memory controller on the DDR SDRAM interface Each data beat transferred to or from the DDR SDRAM Row open table read misses issued while a read is outstanding Row open table read or write misses issued while a read or write is outstanding Row open table read misses issued when no reads are outstanding Row open table read or write misses issued when no reads or writes are outstanding Row open table read hits issued when a read is outstanding Row open table read or write hits issued when a read or write is outstanding Row open table read hits issued when no reads are outstanding Row open table read or write hits issued when no reads or writes are outstanding Precharges issued to the DDR SDRAM for any reason except refresh. The possibilities are as follows: • A new transaction must be issued to an already active bank and sub-bank that has a different row open. • A new transaction must be issued, but the row open table is full and there is no bank/sub-bank match between the current transaction and the row open table. • The BSTOPRE interval expired for an open row. Transactions that miss in the row open table Transaction that hit in the row open table Forced page closings including those due to refreshes If ECC is enabled and a transaction requires byte enables, a read-modify-write sequence is issued on the DDR SDRAM interface. Increments if a new transaction must be issued to an active bank and sub-bank that has a different row open — — — Row open table misses Row open table hits Force page closings Read-modify-write transactions due to ECC Forced page closings due to collision with bank and sub-bank Reads or writes from core Reads or writes from TSEC 1 Reads or writes from TSEC 2 C2:1 C3:0 C4:1 C5:0 Ref:12 Ref:13 C3:1 C4:2 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 19-16 Freescale Semiconductor Functional Description Table 19-10. Performance Monitor Events (continued) Event Counted Reads or writes from RapidIO Reads or writes from PCI Reads or writes from DMA Row open table hits for reads or writes from core Row open table hits for reads or writes from TSEC 1 Row open table hits for reads or writes from TSEC 2 Row open table hits for reads or writes from RapidIO Row open table hits for reads or writes from PCI Row open table hits for reads or writes from DMA Number C3:2 C4:3 C5:2 Ref:14 C6:1 C7:0 C6:2 C7:1 C8:2 — — — — — — — — — Description of Event Counted Memory Target Queue Events MEM TQ read/write address collision C5:5 — RapidIO Controller Events Misaligned transactions on RapidIO Retried packets on RapidIO transmit due to resource limitations Flushed packets due to prior retries or error recovery on RapidIO transmit Retried or not accepted packets on RapidIO Rx Misaligned engine priority 2 occupied Misaligned engine priority 1 occupied Misaligned engine priority 0 occupied ACK history queue full Outbound RapidIO stopped for training event RapidIO outbound retry-stopped event Outbound RapidIO error-stopped event Inbound message packet protocol level retry C4:11 C2:62 C3:10 C5:7 C5:12 C6:12 C7:10 C4:61 C6:13 C7:11 C8:11 C4:18 New transactions loaded into the misaligned engine Includes only packets that caused the retry, not ignored packets that must be resent Includes the packet being acknowledge retried Includes only retried or unaccepted packets from a remote device, not ignored packets that must be resent Misaligned engine for priority 2 transactions busy — Misaligned engine for priority 0 transactions busy Unacked packets are outstanding. Event asserted during cycles when packet transmission is stopped for training. Outbound port is stopped due to an inbound ACK retry Outbound port is stopped due to an inbound ACK not accepted — MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 19-17 Performance Monitor Table 19-10. Performance Monitor Events (continued) Event Counted Inbound doorbell packet protocol level retry RapidIO outbound ACK timeout/ out-of-order event Number C6:14 C8:12 — An ACK timeout or an out of sequence ACK has been detected by the ACK history queue. Description of Event Counted DMA Controller Events Channel 0 read request Channel 1 read request Channel 2 read request Channel 3 read request Channel 0 write request Channel 1 write request Channel 2 write request Channel 3 write request Channel 0 descriptor request Channel 1 descriptor request Channel 2 descriptor request Channel 3 descriptor request Channel 0 read DW or less Channel 1 read DW or less Channel 2 read DW or less Channel 3 read DW or less Channel 0 write DW or less Channel 1 write DW or less Channel 2 write DW or less Channel 3 write DW or less C1:2 C2:5 C3:4 C4:6 C1:3 C2:6 C3:5 C4:7 C5:41 C6:44 C7:41 C8:41 DMA channel 0 read request active in the system DMA channel 1 read request active in the system DMA channel 2 read request active in the system DMA channel 3 read request active in the system DMA channel 0 write request active in the system DMA channel 1 write request active in the system DMA channel 2 write request active in the system DMA channel 3 write request active in the system DMA channel 0 descriptor request active in the system DMA channel 1 descriptor request active in the system DMA channel 2 descriptor request active in the system DMA channel 3 descriptor request active in the system C1:4 and C5:53 DMA channel 0 read double word valid C2:7 and C6:58 DMA channel 1 read double word valid C3:6 and C7:54 DMA channel 2 read double word valid C4:8 and C8:52 DMA channel 3 read double word valid C1:5 C2:8 C3:7 C4:9 DMA channel 0 write double word valid DMA channel 1 write double word valid DMA channel 2 write double word valid DMA channel 3 write double word valid e500 Coherency Module (ECM) Events ECM request wait core ECM request wait TSEC1 ECM request wait TSEC2 ECM request wait PCI/RapidIO/DMA ECM dispatch C8:13 C5:16 C6:16 C4:20 Ref:15 Asserted for every cycle core request occurs Asserted for every cycle TSEC1 request occurs Asserted for every cycle TSEC2 request occurs Asserted for every cycle PCI request occurs ECM dispatch (includes address only’s) Note: all ECM dispatch events are for committed dispatches ECM dispatch from core (includes address only’s) ECM dispatch from core C1:16 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 19-18 Freescale Semiconductor Functional Description Table 19-10. Performance Monitor Events (continued) Event Counted ECM dispatch from TSEC1 ECM dispatch from TSEC2 ECM dispatch from RapidIO ECM dispatch from PCI ECM dispatch from DMA ECM dispatch from other ECM dispatch to DDR ECM dispatch to L2/SRAM ECM dispatch to LBC ECM dispatch to RapidIO ECM dispatch to PCI ECM dispatch snoopable ECM dispatch write ECM dispatch write allocate ECM dispatch write allocate lock ECM dispatch read ECM dispatch read unlock ECM dispatch read clear atomic ECM dispatch read set atomic ECM dispatch read decrement atomic ECM dispatch read increment atomic ECM data bus grant DDR ECM data bus grant LBC ECM data bus grant PIC ECM data bus grant TSEC1 ECM data bus grant TSEC2 ECM data bus wait DDR ECM data bus wait LBC ECM data bus wait PIC ECM data bus wait TSEC1 ECM data bus wait TSEC2 ECM global data bus beat Number C3:19 C4:21 C5:17 C6:17 C7:14 C8:14 C4:22 C5:18 C6:18 C7:15 C8:15 C3:20 C1:17 C2:21 C3:21 C4:23 C5:19 C6:19 C7:16 C8:16 C7:17 C1:18 C2:22 C1:19 C3:23 C4:25 C5:20 C6:20 C5:21 C7:19 C8:18 Ref:16 — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — Description of Event Counted MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 19-19 Performance Monitor Table 19-10. Performance Monitor Events (continued) Event Counted ECM e500 direct read bus beat ECM e500 direct read bus beat forwarded ECM cancel Number Ref:17 C2:24 Ref:18 — ECM direct read bus beat forwarded directly to e500 R1 data bus — Description of Event Counted Interrupt Controller (PIC) Events PIC total interrupt count PIC interrupt wait cycles PIC interrupt service cycles PIC interrupt select 0 (duration threshold) PIC interrupt select 1 (duration threshold) PIC interrupt select 2 (duration threshold) PIC interrupt select 3 (duration threshold) Ref:26 C8:62 C2:19 C1:56 C3:59 C5:55 C6:60 PCI/PCI-X Common Events PCI/PCI-X clock cycles PCI/PCI-X inbound memory reads PCI/PCI-X inbound memory writes PCI/PCI-X inbound config reads PCI/PCI-X inbound config writes PCI/PCI-X outbound memory reads PCI outbound memory writes/PCI-X outbound memory writes attempted PCI/PCI-X outbound I/O reads PCI/PCI-X outbound I/O writes PCI outbound config reads/PCI-X outbound config reads attempted PCI/PCI-X outbound config writes PCI/PCI-X inbound total read data beats PCI/PCI-X inbound total write data beats PCI/PCI-X outbound total read data beats Ref:28 C1:62 C2:37 C3:63 C4:37 C5:30 C6:32 C3:37 C4:38 C7:26 C8:26 C5:32 C6:34 C7:28 — Includes all read types — — — Includes all read types Number of PCI outbound memory writes or number of PCI-X outbound memory writes attempted — — Number of PCI outbound config reads OR number of PCI-X outbound config reads attempted — Includes 32- and 64-bit transactions Includes 32- and 64-bit transactions Includes 32- and 64-bit transactions Total number of interrupts serviced Counts cycles when an interrupt waits to be acknowledge Number of cycles there is an interrupt currently being serviced THRESHOLD: select 0–3: interrupt count over threshold. (Note: only unmasked, nonzero priority requests are acknowledged). The four interrupts are selected through register pairs, PM0MRn–PM3MRn. See Section 10.3.4, “Performance Monitor Mask Registers (PMMRs).” MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 19-20 Freescale Semiconductor Functional Description Table 19-10. Performance Monitor Events (continued) Event Counted PCI/PCI-X outbound total write data beats PCI/PCI-X inbound 32-bit read data beats PCI/PCI-X inbound 32-bit write data beats PCI/PCI-X outbound 32-bit read data beats PCI/PCI-X outbound 32-bit write data beats PCI/PCI-X inbound 64-bit read data beats PCI/PCI-X inbound 64-bit write data beats PCI/PCI-X outbound 64-bit read data beats PCI/PCI-X outbound 64-bit write data beats PCI/PCI-X total transactions PCI/PCI-X 64-bit transactions PCI/PCI-X inbound purgeable reads PCI/PCI-X inbound (speculative reads) purgeable reads discarded PCI/PCI-X idle cycles PCI/PCI-X dual address cycles PCI/PCI-X internal cycles PCI inbound memory read PCI-X inbound memory 32-bit read PCI inbound memory readline PCI-X inbound memory alias read PCI inbound memory read multiple PCI-X inbound memory block read PCI outbound memory reads/PCI-X outbound memory DWORD reads attempted PCI outbound memory read lines/ PCI-X outbound memory burst read attempted Number C8:28 C1:30 C2:38 C3:38 C4:39 C5:31 C6:33 C7:27 C8:27 C7:29 C8:29 C2:2 C8:63 C1:31 C2:40 C3:39 C1:34 — — — — Description of Event Counted Includes 32- and 64-bit transactions — — — — — — — — Includes 32- and 64-bit transactions — — C2:44 — C3:42 — C4:43 Number of PCI outbound memory reads or number of PCI-X outbound memory DWORD reads attempted Number of PCI outbound memory read lines or number of PCI-X outbound memory burst reads attempted C5:36 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 19-21 Performance Monitor Table 19-10. Performance Monitor Events (continued) Event Counted PCI wait PCI-X initial wait Number C1:35 Description of Event Counted PCI_IRDY, PCI_TRDY not both asserted PCI Specific Events PCI cycles PCI_IRDY is asserted PCI cycles PCI_TRDY is asserted PCI cycles PCI_FRAME is asserted PCI-X 1 split transaction C6:36 C7:31 C8:31 C2:41 — — — Number of PCI-X cycles there are 1 or more outbound read transactions awaiting completion (either waiting for a split response, or a partially transacted dword disconnected by a SDPD, awaiting completion) Number of PCI-X cycles there are 2 or more outbound read transactions awaiting completion (either waiting for a split response, or a partially transacted dword disconnected by a SDPD, awaiting completion) Number of PCI-X cycles there are 3 or more outbound read transactions awaiting completion (either waiting for a split response, or a partially transacted dword disconnected by a SDPD, awaiting completion) Number of PCI-X cycles there are 4 outbound read transactions awaiting completion (either waiting for a split response, or a partially transacted dword disconnected by a SDPD, awaiting completion) Split responses received for an outbound transactions — — — — — — — PCI-X 2 split transactions C3:40 PCI-X 3 split transactions C4:41 PCI-X 4 split transactions C5:34 PCI-X split responses PCI-X ADB disconnects PCI/PCI-X snoopable PCI/PCI-X write stash PCI/PCI-X write stash with lock PCI/PCI-X read unlock PCI/PCI-X byte enable transactions PCI/PCI-X non-byte enable transactions C6:37 C7:32 C1:32 C2:42 C3:41 C4:42 C1:33 C2:43 Three-Speed Ethernet Controller (TSEC) Events TSEC1 Address Data Filtering (ADF) Events Accepted frames Individual hash table accepted frame Group hash table accepted frame Rejected frames Ref:36 C7:35 C8:35 Ref:39 — — — — MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 19-22 Freescale Semiconductor Functional Description Table 19-10. Performance Monitor Events (continued) Event Counted Dropped frames Dropped frames due to data overflow Dropped frames due to status overflow Number Ref:38 C1:41 C2:50 Description of Event Counted Dropped frames that could have been accepted (data overflow, status overflow and lack of BDs) Frames dropped because of overflow in the FIFO and 256-byte buffer Frames dropped because of inability to write status for one frame and a new frame starts TSEC1 FIFO Events Receive FIFO data valid Transmit frames without threshold Receive FIFO above 1/4 Receive FIFO above 1/2 Receive FIFO above 3/4 DMA reads BD reads RxBD reads DMA writes BD writes RxBD writes RxBD read latency (duration threshold) C1:45 C7:44 Ref:47 Ref:48 Ref:49 C1:47 C2:54 C3:51 C4:52 C5:46 C6:49 Ref:41 — Transmit frames that do not hit Tx FIFO threshold — — — Descriptor and data reads TxBD and RxBD reads RxBD reads (transmit number can be calculated by subtracting this number from total) Writes (descriptor and data) TxBDs and RxBDs closed RxBDs closed (transmit number can be calculated by subtracting this number from total) Times RxBD read latency exceeds threshold (threshold event) Start: request asserted Stop: data acknowledge received Times TxBD read latency exceeds threshold (threshold event) Start: request asserted Stop: data acknowledge received Times RxBD write latency exceeds threshold (threshold event). Only for writes that require a response Start: request asserted Stop: end of transaction received Times TxBD write response latency exceeds threshold (threshold event). Only for writes that require a response Start: request asserted Stop: end of transaction received Times data read response latency exceeds threshold (threshold event) Start: request asserted Stop: data acknowledge received TxBD read latency (duration threshold) Ref:42 RxBD write latency (duration threshold) Ref:43 TxBD write latency (duration threshold) Ref:44 Tx data read latency (duration threshold) Ref:45 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 19-23 Performance Monitor Table 19-10. Performance Monitor Events (continued) Event Counted Data beats Read data beats Rx frame interrupts Tx frame interrupts Rx frame processing (duration threshold) Number C7:48 C8:46 C1:49 C3:53 Ref:46 — — Times receive frame interrupt is set. Event occurs when the BD receive interrupt is set on a last BD Times a transmit frame interrupt is set. Event occurs when the BD transmit interrupt is set on a BD with L = 1 Times frame processing exceeds threshold (threshold event) Start: open first BD Stop: close last BD Description of Event Counted TSEC2 Address Data Filtering (ADF) Events Accepted frames Individual hash table accepted frame Group hash table accepted frame Rejected frames Dropped frames Dropped frames due to data overflow Dropped frames due to status overflow Ref:50 C7:36 C8:36 Ref:53 Ref:52 C1:42 C2:51 — — — — Dropped frames that could have been accepted (data overflow, status overflow and acknowledgement of BDs) Frames dropped because of overflow in the FIFO and 256-byte buffer Frames dropped because of inability to write status for one frame and a new frame starts TSEC2 FIFO Events Receive FIFO data valid Transmit frames without threshold Receive FIFO above 1/4 Receive FIFO above 1/2 Receive FIFO above 3/4 C1:46 C7:45 Ref:61 Ref:62 Ref:63 Receive FIFO contains receive data Transmit frames that do not hit Tx FIFO threshold — — — TSEC2 DMA Events DMA reads BD reads RxBD reads DMA writes BD writes RxBD writes C1:48 C2:55 C3:52 C4:53 C5:47 C6:50 Descriptor and data reads Transmit and RxBD reads Times that RxBD reads (transmit number can be calculated by subtracting this number from total) Descriptor and data writes Times that TxBDs and RxBDs closed Times RxBDs closed (transmit number can be calculated by subtracting this number from total) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 19-24 Freescale Semiconductor Functional Description Table 19-10. Performance Monitor Events (continued) Event Counted RxBD read latency (duration threshold) Number Ref:55 Description of Event Counted Times RxBD read latency exceeds threshold (threshold event) Start: request asserted Stop: data acknowledge received Times TxBD read latency exceeds threshold (threshold event) Start: request asserted Stop: data acknowledge received Times RxBD write response latency exceeds threshold (threshold event). Only for writes that require a response Start: request asserted Stop: End of Transaction received Times that TxBD write response latency exceeds threshold (threshold event). Only for writes that require a response Start: request asserted Stop: End of Transaction received Data read response latency exceeds threshold (threshold event) Start: request asserted Stop: Data acknowledge received — — BD receive interrupt is set on a last BD BD transmit interrupt is set on a BD with L = 1 Receive frame processing exceeds threshold (threshold event) Start: Open first BD Stop: Close last BD TxBD read latency (duration threshold) Ref:56 RxBD write latency (duration threshold) Ref:57 TxBD write latency (duration threshold) Ref:58 Tx data read latency (duration threshold) Ref:59 Data beats Read data beats Rx frame interrupts Tx frame interrupts Rx frame processing C7:49 C8:47 C1:50 C3:54 Ref:60 Local Bus Events Bank 1 hits (chip-select) Bank 2 hits (chip-select) Bank 3 hits (chip-select) Bank 4 hits (chip-select) Bank 5 hits (chip-select) Bank 6 hits (chip-select) Bank 7 hits (chip-select) Bank 8 hits (chip-select) Requests granted to ECM port Cycles atomic reservation for ECM port is enabled C1:51 C2:56 C3:55 C4:54 C5:48 C6:53 C7:50 C8:50 C2:57 C4:55 — — — — — — — — — MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 19-25 Performance Monitor Table 19-10. Performance Monitor Events (continued) Event Counted Atomic reservation time-outs for ECM port Cycles a read is taking in GPCM Cycles a read is taking in UPM Cycles a read is taking in SDRAM Cycles a write is taking in GPCM Cycles a write is taking in UPM Cycles a write is taking in SDRAM SDRAM bank misses SDRAM page misses Number C6:54 C1:53 C2:58 C3:57 C4:56 C5:50 C6:55 C7:51 C8:51 — — — — — — — — — Description of Event Counted L2 Cache/SRAM Events Core instruction accesses to L2 that hit Core instruction accesses to L2 that miss Core data accesses to L2 that hit Core data accesses to L2 that miss Non-core burst write to L2 (cache external write or SRAM) Non-core non-burst write to L2 Noncore write misses cache external write window and SRAM memory range Non-core read hit in L2 Non-core read miss in L2 L2 allocates, from any source L2 retries due to full write queue L2 retries due to address collision L2 failed lock attempts due to full set L2 victimizations of valid lines L2 invalidations of lines L2 clearing of locks Ref:22 C2:59 Ref:23 C4:57 C5:51 C6:56 C7:52 — — — — — — — Ref:24 C1:54 Ref:25 C2:60 C3:58 C4:58 C5:52 C6:57 C7:53 — — — — — — — — — Debug Events External event Watchpoint monitor hits C3:61 C2:61 Number of cycles trig_in pin is asserted — MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 19-26 Freescale Semiconductor Functional Description Table 19-10. Performance Monitor Events (continued) Event Counted Trace buffer hits Number C1:58 — Description of Event Counted DUART Events UART0 baud rate UART1 baud rate C1:63 C5:63 — — Chaining Events PMC0 carry-out PMC1 carry-out PMC2 carry-out PMC3 carry-out PMC4 carry-out PMC5 carry-out PMC6 carry-out PMC7 carry-out PMC8 carry-out Ref:1 Ref:2 Ref:3 Ref:4 Ref:5 Ref:6 Ref:7 Ref:8 Ref:9 PMC0[0] 1-to-0 transitions PMC1[0] 1-to-0 transitions. Reserved for PMC1. PMC2[0] 1-to-0 transitions. Reserved for PMC2. PMC3[0] 1-to-0 transitions. Reserved for PMC3. PMC4[0] 1-to-0 transitions. Reserved for PMC4. PMC5[0] 1-to-0 transitions. Reserved for PMC5. PMC6[0] 1-to-0 transitions. Reserved for PMC6. PMC7[0] 1-to-0 transitions. Reserved for PMC7. PMC8[0] 1-to-0 transitions. Reserved for PMC8. 19.4.8 Performance Monitor Examples Table 19-12 contains sample register settings for the four supported modes. • • • • Simple event performance monitoring example Triggering event performance monitoring example Threshold event performance monitoring example Burstiness event performance monitoring example The settings in Table 19-11 are identical for all four examples. Table 19-11. PMGC0 and PMLCAn Settings Field PMGC0[FAC] PMGC0[PMIE] PMGC0[FCECE] PMLCAn[FC] PMLCAn[CE] Setting 0 1 1 0 1 Reason Counters must not be frozen. Performance monitor interrupts are enabled. Counters should be frozen when an interrupt is signalled. Counters cannot be frozen for counting. Overflow condition enable is required to allow interrupt signalling. For simple event counting, a non-threshold event is selected in PMLCAn[EVENT] and all other features are disabled by clearing all register fields except for CE. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 19-27 Performance Monitor For the triggering example any event can be selected in PMLCAn[EVENT]. All other features are disabled by clearing these register fields except for CE to allow interrupt signalling. If PMLCBn[TRIGONSEL] is 3 and PMLCBn[TRIGOFFSEL] is 5, the counter begins and ends counting based on the conditions in counters three and five. Furthermore, if PMLCBn[TRIGONCNTL] is 1, the counter begins counting when PMC3 changes value. According to the setting in PMLCBn[TRIGOFFCNTL], the counter ends counting when PMC5 overflows. Also, although the register settings for PMC5 is not shown, PMLCAn[CE] for this counter must be cleared so that interrupt signalling is not enabled and the counter does not freeze when it overflows. For threshold counting, a threshold event must be specified in PMLCAn[EVENT]. For this example, the duration threshold value is scaled by two because PMLCBn[TBMULT] is one. All other features are disabled by clearing the appropriate fields. Any non-threshold event can use the burstiness feature. For burstiness counting, values for PMLCAn[BSIZE,BGRAN,BDIST] and PMLCBn[TBMULT] must be specified. Table 19-12. Register Settings for Counting Examples MPC8540 Register PMGC0 Register Field FAC PMIE FCECE PMLCAn FC CE EVENT BSIZE BGRAN BDIST PMLCBn TRIGONSEL TRIGOFFSEL TRIGONCNTL TRIGOFFCNTL TBMULT THRESHOLD Simple Event 0 1 1 0 1 89 0 0 0 0 0 0 0 0 0 Triggering 0 1 1 0 1 68 0 0 0 3 5 1 2 0 0 Threshold 0 1 1 0 1 39 0 0 0 0 0 0 0 0 3 Burstiness 0 1 1 0 1 2 5 1 8 0 0 0 0 0 0 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 19-28 Freescale Semiconductor Chapter 20 Debug Features and Watchpoint Facility This chapter describes all customer-visible debug modes of the MPC8540 integrated device. The debug features on the MPC8540 pertain to these three interfaces: the local bus controller (LBC), the DDR SDRAM, and the PCI/PCI-X interface. In addition to the external interfaces, the MPC8540 provides triggering capabilities based on user-programmable events. The watchpoint and trace buffer also provide some visibility to internal buses. This chapter also describes context ID registers, useful for software debug, and describes the JTAG access port signals that comply with the IEEE 1149.1 boundary-scan specification. 20.1 Introduction As shown in the block diagram of Figure 20-1, the MPC8540 device provides the following debug features (listed with references to sections of this chapter that describe them): • • • • PCI/PCI-X interface debug (Section 20.4.2, “PCI/PCI-X Interface Debug”) DDR SDRAM interface debug (Section 20.4.3, “DDR SDRAM Interface Debug”) Local bus controller (LBC) debug (Section 20.4.4, “Local Bus Interface Debug”) Watchpoint monitor and trace buffer debug (Section 20.4.5, “Watchpoint Monitor,” and Section 20.4.6, “Trace Buffer”) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 20-1 Debug Features and Watchpoint Facility LBC protocol signals PORDBGMSR[MEM_SEL] = POR value of MSRCID0 Signal name 6 6 SRCID[0:4] DVAL MSRCID[0:4], MDVAL e500 Processor Core LBC (Local Bus) e500 Coherency Module All other DDR signals 6 DDR SDRAM Controller 6 SRCID[0:4] DVAL MECC[0:5] All other PCI signals RapidIO Controller 5 PCI/PCI-X Controller SRCID[0:4] PCI_AD[62:58] Performance Monitor Trace Buffer (256 x 64) Watchpoint and trace buffer control TRIG_IN TRIG_OUT Figure 20-1. Debug and Watchpoint Monitor Block Diagram 20.1.1 Overview As shown in Figure 20-1, debug information is provided through the following interfaces: PCI/PCI-X, LBC, and DDR SDRAM. Limited visibility, through a 256 × 64 trace buffer, is also provided for the processor core interface and on an internal RapidIO outbound interface. This visibility into internal device operation is useful for debugging application software through inverse assembly and reconstruction of the fetch stream. The combination of a source ID (MSRCID[0:4]) and a data-valid signal (MDVAL) indicates that meaningful debug information is visible on either the local bus or DDR SDRAM interfaces. A logic analyzer can be programmed to capture data based on the values of MSRCID[0:4] and MDVAL. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 20-2 Freescale Semiconductor Introduction Othmer system debugging is supported by the programmable triggering of the watchpoint monitor and trace buffer. Both can be triggered from one of the following three sources: • • • Each other A performance monitor event An external source (through TRIG_IN) The watchpoint monitor can be configured to assert TRIG_OUT when a programmed event occurs. The two context ID registers, described in Section 20.3.3, “Context ID Registers,” are useful for software debug. 20.1.2 Features The principle features of the debug modes and the watchpoint monitor are as follows: • • PCI/PCI-X interface: transaction source ID driven onto PCI_AD[62:58] LBC and DDR interface source ID and data-valid indicators — LBC or DDR SDRAM source ID can be selected to be driven onto MSRCID[0:4] — Source ID and data-valid indicators can be selected to be driven onto the error correcting code (ECC) pins of the DDR interface Watchpoint monitor that supports the following: — Two-level triggering — Programmable external trigger (TRIG_OUT) — Interlocking with performance monitor to use its large number of counters Trace buffer features that supports the following: — Two-level triggering — Programmable external trigger (TRIG_OUT) — Interlocking with performance monitor to use its large number of counters — 256-entry trace buffer, 64 bits each — Programmable trace start and stop — Functionality as a second watchpoint monitor Context ID registers that can be programmed to trigger events • • • 20.1.3 Modes of Operation The PCI/PCI-X, LBC, and DDR SDRAM interfaces all have debug modes, which are controlled by values on configuration inputs during the power-on reset (POR) sequence, as shown in Table 20-3. The DDR controller can also drive debug information on either MSRCID[0:4] or MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 20-3 Debug Features and Watchpoint Facility MECC[0:5]. See Section 20.4.1, “Source and Target ID,” for additional information about the source ID information driven on the debug signals in these modes. Note that both the watchpoint monitor and trace buffer also operate in a variety of modes. Table 20-1. POR Configuration Settings and Debug Modes Configuration Signal MSRCID0 POR Value 0 1 MSRCID1 0 1 PCI_GNT3 0 Effect Local bus SDRAM information appears on MSRCID[0:4] and MDVAL. Default value (internal pull-up resistor). DDR SDRAM information appears on MSRCID[0:4] and MDVAL. MECC[0:4] operate in debug mode and provide memory debug source ID and MECC5 provides data-valid information. Default value (internal pull-up resistor). MECC[0:4] operate in normal mode and provide DDR SDRAM error correcting code information. The PCI/PCI-X interface operates in debug mode. The source ID information appears on the high-order address bits (PCI_AD[62:58]) during the bus command phase. Default value (internal pull-up resistor). The PCI/PCI-X interface operates in normal mode. 20.1.3.3/20-5 20.1.3.2/20-4 Reference 20.1.3.1/20-4 1 20.1.3.1 Local Bus (LBC) Debug Mode The LBC and the DDR SDRAM controller can drive debug information (source ID and data-valid indicator) onto MSRCID[0:4] and MDVAL As shown in Table 20-1, the MSRCID0 value during POR controls multiplexing. If MSRCID0 is low when sampled during POR, the local bus SDRAM information appears on MSRCID[0:4] and MDVAL; otherwise, the DDR SDRAM debug information is presented. 20.1.3.2 DDR SDRAM Interface Debug Modes MSRCID1 is sampled during POR to multiplex either ECC or debug information on the ECC pins of the DDR SDRAM interface. As shown in Table 20-1, if MSRCID1 is low during POR, the ECC pins operate in debug mode and provide memory debug source ID and data-valid information. MSRCID1 must be pulled low during POR to use the ECC pins in debug mode. If MSRCID1 is unconnected, an internal pull-up resistor ensures the ECC pins always source DDR SDRAM error correcting code information as their default power-on reset configuration. NOTE If the DDR ECC pins are in debug mode (configured for debug during POR), ECC checking is disabled in the memory controller. In this case, MECC[0:4] do not provide ECC information and must not be connected to SDRAM devices. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 20-4 Freescale Semiconductor Introduction 20.1.3.3 PCI/PCI-X Interface Debug Modes If PCI_GNT3 is low when sampled during POR, the PCI/PCI-X interface operates in debug mode. In this mode, the source ID information appears on the high-order address bits (PCI_AD[62:58]) during the bus command phase. See Section 20.4.2, “PCI/PCI-X Interface Debug,” for more information. 20.1.3.4 Watchpoint Monitor Modes The watchpoint monitor supports the following operating modes: • • Immediate trigger arming (one-level triggering)—The watchpoint monitor triggers as soon as the first trigger event occurs. Wait for trigger arming (two-level triggering)—The watchpoint monitor waits for a specific event before enabling (arming) the trigger logic. The monitor does not respond to trigger events until after the arming event occurs. This function is similar to two-level triggering on a logic analyzer. Assert TRIG_OUT on hit—The debug block can be programmed to assert the TRIG_OUT signal when a programmed watchpoint monitor event occurs. This signal can be used to trigger a logic analyzer. • 20.1.3.5 Trace Buffer Modes The trace buffer supports the following operating modes: • • Immediate trigger arming (one-level triggering)—The trace buffer triggers as soon as the first trigger event occurs. Wait for trigger arming (two-level triggering)—The trace buffer waits for a specific event before enabling (arming) the trigger logic. The trace buffer does not respond to trigger events until after the arming event occurs. This function is similar to two-level triggering on a logic analyzer. Specific interface selection—The trace buffer can be programmed to trace one of several internal interfaces. Specific event selection—The trace buffer can be programmed to trace on the occurrence of one or several concurrent events. Specific trace selection—To facilitate trace data filtering, the trace buffer can be configured to capture data under the following conditions: — On every cycle in which a valid transaction is present on the selected interface — Only when a watchpoint monitor event occurs — Only when the programmed trace event is detected Programmable trace stop—The trace buffer may be programmed to stop tracing when a programmed stop-tracing event occurs or when the 256-entry buffer is full. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 20-5 • • • • Debug Features and Watchpoint Facility 20.2 External Signal Descriptions This section provides information about all the external signals associated with the various MPC8540 debug functions. As shown in Table 20-1, the MPC8540 has several signals that are sampled during POR to determine the configuration of the phase-locked loop clock mode and the ROM, Flash, and dynamic memory. See Chapter 4, “Reset, Clocking, and Initialization.” To facilitate system testing, the MPC8540 provides a JTAG test access port (TAP) that complies with the IEEE 1149.1 boundary-scan specification. This section also describes JTAG TAP signals. 20.2.1 Overview All the signals associated with device debug features are summarized in Table 20-2, listed with a reference to the page number of the section with more information. The detailed descriptions are contained in Table 20-2. Some signals (the MECC bus for example) are additionally described in other chapters, but are described here also for completeness, with emphasis on their debugging utility. Table 20-2. Debug, Watchpoint and Test Signal Summary Name MDVAL MECC[0:7] Description Memory data-valid DDR error correcting code Memory source ID Trigger in Trigger out Functional Block Debug DDR SDRAM Debug Function Selectable data-valid signal from either DDR SDRAM controller or LBC In debug mode, the 6 high-order bits carry debug information (transaction source ID and data-valid indication). Reset Value 1 0x08 I/O Page # O O1 20-7 20-8 MSRCID[0:1] MSRCID[2:4] TRIG_IN TRIG_OUT Selectable transaction source ID from either DDR SDRAM Reset_cfg O controller or local bus controller 111 O Trigger for various function in the watchpoint monitor and trace buffer Can be used externally for triggering a logic analyzer. Additionally, it can be used for observing system ready indication. Functions are multiplexed onto this signal depending on TOSR[SEL] (see Table 20-25). In debug mode these pins carry the source ID of the transaction. Clock for JTAG testing. Internally pulled up Serial input for instructions and data to the JTAG test subsystem. Internally pulled up Serial data output for the JTAG test subsystem. High impedance except when scanning out data 1 1 I O 20-8 20-8 20-8 20-8 Debug Debug PCI_AD[62:58] PCI address TCK TDI TDO Test clock Test data input Test data output Debug Debug Debug Debug 0000_0 1 1 Hi-Z O I I O 20-8 20-9 20-9 20-9 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 20-6 Freescale Semiconductor External Signal Descriptions Table 20-2. Debug, Watchpoint and Test Signal Summary (continued) Name TMS TRST THERM[0:1] Description Test mode select Test reset Thermal resistor access Test select Test Test Test Functional Block Debug Debug Test Function Carries commands to the TAP controller for boundary scan operations. Internally pulled up Resets the TAP controller asynchronously These pins tie directly to an internal resistor whose value varies linearly with temperature. Factory test. Must be negated for normal operation Factory Test. Refer to the MPC8540 Integrated Processor Hardware Specifications for proper treatment. Factory Test. Refer to the MPC8540 Integrated Processor Hardware Specifications for proper treatment. Factory Test. Refer to the MPC8540 Integrated Processor Hardware Specifications for proper treatment. Reset Value 1 — — I/O Page # I I I 20-9 20-9 20-9 TEST_SEL LSSD_ MODE L1_TSTCLK L2_TSTCLK 1 Test Test Test Test — I I I I 20-9 20-9 20-9 20-9 Although these signals are normally bidirectional, when sourcing debug information they are output only. 20.2.2 Detailed Signal Descriptions This section describes the details of the debug, watchpoint monitor, and JTAG test signals 20.2.2.1 Debug Signals—Details Table 20-3 describes all signals associated with device debug modes. Table 20-3. Debug Signals—Detailed Signal Descriptions Signal MDVAL I/O O Description Memory data-valid. Indicates when valid data is available. May be used by a logic analyzer to capture the data on the data bus. State Asserted—Indicates that data is valid on the data bus during the current clock cycle. When the Meaning DDR SDRAM interface is selected to source information on MDVAL, this signal is valid for every cycle that data is driven or received on the DDR SDRAM interface. When the LBC is selected, this signal is valid for every cycle that data is driven or received on the local bus interface. The assertion of this signal may be used by a logic analyzer to capture data. Timing Asserted/Negated—Referenced to the selected interface, (DDR or local bus). Asserts when data is valid. Assertions are held for the duration of the transfer. Read data timing is similar to MA. Write data timing is similar to the output MDQ. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 20-7 Debug Features and Watchpoint Facility Table 20-3. Debug Signals—Detailed Signal Descriptions (continued) Signal MECC[0:7] I/O O Description Memory ECC. DDR error checking and correcting. The normally bidirectional operation of the memory ECC (MECC) bus is described in Section 9.5.11, “Error Checking and Correcting (ECC).” This bus is used for debug functions when MSRCID1 is sampled low during POR. In debug mode, the high-order 5 bits (MECC[0:4]) may be used to provide the transaction source ID and MECC5 can be used as the data-valid indicator. In debug mode, MECC[0:5] is constantly driven with debug information and must be disconnected from the DDR memory’s ECC pins. State Asserted/Negated—In debug mode, MECC[0:5] is always driven. The source ID values appear Meaning during RAS and CAS cycles. A value of 0x1F (all ones) is driven during cycles other than RAS and CAS. The data-valid indicator appears when data is being received or driven on the pins. Timing Driven every cycle in debug mode MSRCID[0:4] O Memory source ID. Attribute signals associated with the memory interface that indicate the source ID for a transaction on an SDRAM interface. The SDRAM interface, DDR or local bus, to which the debug information applies is specified during POR with MSRCID0 as shown in Table 20-1. Two of these signals serve as reset configuration input signals. State Asserted/Negated—In debug mode, always driven with the value of the source ID. The source Meaning ID has a value of 0x1F for cycles other than RAS and CAS. The encodings shown in Table 20-26 provide detailed information about a memory transaction. Timing Driven every cycle in debug mode. Similar timing to MA PCI_AD[62:58] O PCI Address. Provides transaction source ID for the current PCI bus transaction State Asserted/Negation—In debug mode, always driven with the value of the transaction source ID Meaning Timing Driven every cycle in debug mode 20.2.2.2 Watchpoint Monitor Trigger Signals—Details Table 20-4 shows detailed descriptions of the watchpoint monitor and trace buffer signals. Table 20-4. Watchpoint and Trigger Signals—Detailed Signal Descriptions Signal TRIG_IN I/O I Description Trigger in. Can be used to trigger the watchpoint and trace buffers. Note this is an active-high (rising-edge triggered) signal. State Asserted—Indicates that a programmed/armed external event has been detected. Assertion Meaning may be used internally to trigger trace buffers and watchpoint mechanisms. Timing Assertion/Negation—The MPC8540 interprets TRIG_IN as asserted on detection of the rising edge. It may occur at any time. Must remain asserted for at least 3 system clocks to be recognized internally MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 20-8 Freescale Semiconductor External Signal Descriptions Table 20-4. Watchpoint and Trigger Signals—Detailed Signal Descriptions (continued) Signal TRIG_OUT I/O O Description Trigger out. Function determined by TOSR[SEL]. When TOSR[SEL] is non-zero, it can be used for triggering external devices, like a logic analyzer, with either the watchpoint monitor, the trace buffer, or the performance monitor as trigger sources. When TOSR[SEL] is cleared, TRIG_OUT is multiplexed with READY, which indicates the operational readiness of the device (running or in low-power or debug modes). See Chapter 4, “Reset, Clocking, and Initialization,” and Chapter 18, “Global Utilities,” for more details about reset, low power, and debug states. State Asserted—When TOSR[SEL] is all zeros, serves as the READY signal, indicating that the Meaning device is not in a low-power or debug mode and that it has emerged from reset. SEL ≠ 0 indicates that a programmed trigger event has occurred. Negation—No final watchpoint match condition Timing Assertion may occur at any time. Remains asserted for at least 3 system clocks 20.2.2.3 Test Signals—Details Table 20-5 shows detailed descriptions of the JTAG test signals. Table 20-5. JTAG Test and Other Signals—Detailed Signal Descriptions Signal TCK I/O I JTAG test clock State Asserted/Negated—Should be driven by a free-running clock signal with a 30–70% duty cycle. Meaning Input signals to the TAP are clocked in on the rising edge. Changes to the TAP output signals occur on the falling edge. The test logic allows TCK to be stopped. An unterminated input appears as a high signal level to the test logic due to an internal pull-up resistor. Timing See IEEE 1149.1 standard for more details TDI I JTAG test data input State Asserted/Negated—The value present on the rising edge of TCK is clocked into the selected Meaning JTAG test instruction or data register. An unterminated input appears as a high signal level to the test logic due to an internal pull-up resistor. Timing TDO O See IEEE 1149.1 standard for more details Description JTAG test data output State Asserted/Negated—The contents of the selected internal instruction or data register are shifted Meaning out on this signal on the falling edge of TCK. Remains in a high-impedance state except when scanning data. Timing See IEEE 1149.1 standard for more details TMS I JTAG test mode select State Asserted/Negated—Decoded by the internal JTAG TAP controller to distinguish the primary Meaning operation of the test support circuitry. An unterminated input appears as a high signal level to the test logic due to an internal pull-up resistor. Timing See IEEE 1149.1 standard for more details MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 20-9 Debug Features and Watchpoint Facility Table 20-5. JTAG Test and Other Signals—Detailed Signal Descriptions (continued) Signal TRST I/O I JTAG test reset State Asserted—Causes asynchronous initialization of the internal JTAG TAP controller. Must be Meaning asserted during power-on reset in order to properly initialize the JTAG TAP and for normal operation of the MPC8540. An unterminated input appears as a high signal level to the test logic due to an internal pull-up resistor. Negated— Normal operation Timing See IEEE 1149.1 standard for more details LSSD_MODE L1_TSTCLK L2_TSTCLK THERM[0:1] I I I I Used for factory test. Refer to the MPC8540 Integrated Processor Hardware Specifications for proper treatment. Used for factory test. Refer to the MPC8540 Integrated Processor Hardware Specifications for proper treatment. Used for factory test. Refer to the MPC8540 Integrated Processor Hardware Specifications for proper treatment. These signals provide access to an internal resistor that has a value that varies linearly with temperature. The actual value for the resistor varies from device to device, but the linear relationship between temperature and resistance is consistent. See the MPC8540 Integrated Processor Hardware Specifications for more information on how to accurately measure the junction temperature of a device. Note that this thermal resistor is intended for engineering development only. Used for factory test. Must be negated for normal operation Description TEST_SEL I 20.3 Memory Map/Register Definition Table 20-6 shows the memory-mapped debug and watchpoint registers of the MPC8540. Table 20-6. Debug and Watchpoint Monitor Memory Map Local Memory Offset Register Acces s Reset Section/Page Watchpoint Monitor Registers 0xE_2000 0xE_2004 0xE_200C 0xE_2014 0xE_2018 0xE_201C WMCR0—Watchpoint monitor control register 0 WMCR1—Watchpoint monitor control register 1 WMAR—Watchpoint monitor address register WMAMR—Watchpoint monitor address mask register WMTMR—Watchpoint monitor transaction mask register WMSR—Watchpoint monitor status register Trace Buffer Registers 0xE_2040 0xE_2044 TBCR0—Trace buffer control register 0 TBCR1—Trace buffer control register 1 R/W R/W 0x0000_0000 0x0000_0000 20.3.2.1/20-16 20.3.2.1/20-16 R/W R/W R/W R/W R/W R/W 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 20.3.1.1/20-11 20.3.1.1/20-11 20.3.1.2/20-13 20.3.1.3/20-14 20.3.1.4/20-14 20.3.1.5/20-16 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 20-10 Freescale Semiconductor Memory Map/Register Definition Table 20-6. Debug and Watchpoint Monitor Memory Map (continued) Local Memory Offset 0xE_204C 0xE_2054 0xE_2058 0xE_205C 0xE_2060 0xE_2064 0xE_2068 Register Acces s R/W R/W R/W R/W R/W R/W R/W Reset Section/Page TBAR—Trace buffer address register TBAMR—Trace buffer address mask register TBTMR—Trace buffer transaction mask register TBSR—Trace buffer status register TBACR—Trace buffer access control register TBADHR—Trace buffer access data high register TBADR—Trace buffer access data register Context ID Registers 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 20.3.2.2/20-19 20.3.2.3/20-19 20.3.2.4/20-20 20.3.2.5/20-21 20.3.2.6/20-22 20.3.2.7/20-22 20.3.2.8/20-23 0xE_20A0 0xE_20A4 PCIDR—Programmed context ID register CCIDR—Current context ID register Other Registers R/W R/W 0x0000_0000 0x0000_0000 20.3.3.1/20-24 20.3.3.2/20-24 0xE_20B0 TOSR—Trigger output source register R/W 0x0000_0000 20.3.4.1/20-25 20.3.1 Watchpoint Monitor Register Descriptions The following sections describe the control registers for the watchpoint monitor facility. 20.3.1.1 Watchpoint Monitor Control Registers 0–1 (WMCR0, WMCR1) The watchpoint monitor control registers (WMCR0, WMCR1) shown in Figure 20-2 and Figure 20-3 control the specification of watchpoint monitor events. 0 1 2 3 4 5 6 7 20 21 23 24 31 R EN AMD TMD ECEN NECEN SIDEN TIDEN W Reset Offset 00000000000000 STRT 0 000000 0 0000_0000_0000_0000_0000_0000_0000_0000 0xE_2000 Figure 20-2. Watchpoint Monitor Control Register 0 (WMCR0) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 20-11 Debug Features and Watchpoint Facility Table 20-7 describes WMCR0 fields. Table 20-7. WMCR0 Field Descriptions Bits 0 Name EN Enable 0 Watchpoint monitor events are not flagged. 1 A watchpoint monitor event is flagged. Address match disable. Qualifies address match as a watchpoint event criterion. 0 Address matching is used to recognize a watchpoint event. 1 Address matching does not affect watchpoint event detection. Transaction match disable. Qualifies transaction type match (as defined in WMCR1[IFSEL] and WMTMR) as a watchpoint event criterion. 0 A transaction type match is used to recognize watchpoint events. 1 A transaction type match does not affect watchpoint event detection. Equal context enable. Qualifies the matching of current context with programmed context as a watchpoint event criterion, as written in the context registers described in Section 20.3.3, “Context ID Registers.” 0 Current context match does not affect watchpoint event detection 1 Watchpoint events are qualified by comparing current context with the programmed context event value. Note: ECEN and NECEN must not be enabled in the same run. If both are set, watchpoint events are inhibited (never occur). Description 1 AMD 2 TMD 3 ECEN 4 NECEN Not equal context enable. Qualifies the matching of current context with programmed context as a watchpoint event criterion, as written in the context registers described in Section 20.3.3, “Context ID Registers.” 0 The failure of a current context match does not affect watchpoint event detection 1 Watchpoint events are qualified with NOT getting a current context compare with the programmed context event value. Note: ECEN and NECEN must not be enabled in the same run. If both are set, watchpoint events are inhibited (never occur). SIDEN Source ID enable 0 Source ID does not affect watchpoint event detection 1 Watchpoint events are qualified by comparison with the programmed WMCR1(SID) value. Target ID enable 0 Target ID does not affect watchpoint event detection 1 Watchpoint events are qualified by comparison with the programmed WMCR1(TID) value. Reserved Start condition. Specifies the event that arms the watchpoint monitor to start looking for the programmed event. 000 No event. Armed immediately 001 Trace buffer event is detected 010 Performance monitor signals overflow 011 TRIG_IN transitions from 0 to 1. 100 TRIG_IN transitions from 1 to 0. 101 Current context ID equals programmed context ID 110 Current context ID is not equal to programmed context ID 111 Reserved Reserved 5 6 TIDEN 7–20 21–23 — STRT 24–31 — MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 20-12 Freescale Semiconductor Memory Map/Register Definition Figure 20-3 shows the WMCR1. 0 1 2 3 4 5 7 8 10 11 15 16 26 27 31 R0 W Reset Offset 0 0 0 0 IFSEL 000 SID 00000000000 TID 0000_0000_0000_0000_0000_0000_0000_0000 0xE_2004 Figure 20-3. Watchpoint Monitor Control Register 1 (WMCR1) Table 20-8 describes the WMCR1 fields. Table 20-8. WMCR1 Field Descriptions Bits 0–4 5–7 Name — IFSEL Reserved Interface selection. Selects the address, transaction type (as defined in WMTMR), and other attributes to be used for comparison 000 Select e500 coherency module (ECM) dispatch interface 001 Select internal DDR SDRAM interface 010 Select internal PCI outbound interface 011 Select internal RapidIO interface 1xx Reserved Reserved Source ID. Specifies the source ID associated with WMCR0[SIDEN]. For a definition of the source ID, see Table 20-26. Reserved Target ID. Specifies the target ID associated with WMCR0[TIDEN]. For a definition of the target ID, see Table 20-26. Description 8–10 11–15 16–26 27–31 — SID — TID 20.3.1.2 Watchpoint Monitor Address Register (WMAR) The watchpoint monitor address register (WMAR) shown in Figure 20-4 contains the address to match against if WMCR[AMD] is clear. Note that the transaction address is qualified with the bits described in Section 20.3.1.3, “Watchpoint Monitor Address Mask Register (WMAMR),” before being compared with WMAR. Note also that the contents of WMAR are not qualified with WMAMR. 0 31 R WMA W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0xE_200C Figure 20-4. Watchpoint Monitor Address Register (WMAR) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 20-13 Debug Features and Watchpoint Facility Table 20-9 describes the WMAR fields. Table 20-9. WMAR Field Descriptions Bits 0–31 Name WMA Watchpoint monitor address. Description 20.3.1.3 Watchpoint Monitor Address Mask Register (WMAMR) The watchpoint monitor address mask register (WMAMR) shown in Figure 20-5 contains the mask that is applied to a transaction address before the address is compared with WMAR. 0 31 R WMAM W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0xE_2014 Figure 20-5. Watchpoint Monitor Address Mask Register (WMAMR) Table 20-10 describes the WMAMR fields. Table 20-10. WMAMR Field Descriptions Bits 0–31 Name Description WMAM Watchpoint monitor address mask. A value of zero masks the address comparison for the corresponding address bit. 20.3.1.4 Watchpoint Monitor Transaction Mask Register (WMTMR) The watchpoint monitor transaction mask register (WMTMR), shown in Figure 20-6, specifies which transaction types to monitor. WMTMR allows users to qualify watchpoint events specifically with any combination of transaction types. As shown in Table 20-12, each bit represents as many as four separate transaction types; one for each interface. Setting a bit enables watchpoint monitoring for the corresponding transaction types. Because the supported transaction types vary by interface, the type designated by a WMTMR field also depends on the interface specified by WMCR1[IFSEL]. Table 20-12 lists transaction types associated with each WMTMR bit by interface. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 20-14 Freescale Semiconductor Memory Map/Register Definition 0 31 R WMTM W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0xE_2018 Figure 20-6. Watchpoint Monitor Transaction Mask Register (WMTMR) Table 20-11 describes the WMTMR fields. Table 20-11. WMTMR Field Descriptions Bits 0–31 Name WMTM Description Watchpoint monitor transaction mask. Each bit corresponds to a transaction type as defined in Table 20-12. The transaction associated with any particular bit may be different depending on the interface being monitored. A value of 1 for a given mask bit enables the matching of the transaction associated with that bit. These bits are meaningful only when WMCR0[TMD]=0. The following table defines the transactions associated with each transaction mask bit for the different interfaces supported by the watchpoint monitor. Table 20-12. Transaction Types by Interface Bit 0 1 2 3 4–7 8 9 10 11–15 16 17 18 19 e500 Coherency Module Dispatch Write with local processor snoop Write with no local processor snoop Write with allocate (L2 stashing) Write with allocate and lock (L2 stashing with locking) Reserved Read with local processor snoop Read with no local processor snoop Read with unlock Reserved Atomic clear Atomic set Atomic decrement Atomic increment ATOMIC clear ATOMIC set ATOMIC decrement ATOMIC increment — — — — ATOMIC clear ATOMIC set ATOMIC decrement ATOMIC increment Read — Read with unlock Memory Read I/O Read — NREAD IO_READ_HOME — DDR Controller Write — Write with allocate Write with allocate and lock PCI Outbound Request Memory write I/O write — — RapidIO Outbound Request NWRITE NWRITE_R SWRITE FLUSH with data 20–27 Reserved 28 — — — MAINTENANCE write MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 20-15 Debug Features and Watchpoint Facility Table 20-12. Transaction Types by Interface (continued) Bit 29 30 31 e500 Coherency Module Dispatch — — — DDR Controller — — — PCI Outbound Request — — — RapidIO Outbound Request MAINTENANCE read DOORBELL MESSAGE 20.3.1.5 Watchpoint Monitor Status Register (WMSR) The watchpoint monitor status register (WMSR) shown in Figure 20-7 indicates the state of the watchpoint monitor. 0 1 2 31 R ACT TRIG W Reset Offset 000 00000000 0000000000000000000 0000_0000_0000_0000_0000_0000_0000_0000 0xE_201C Figure 20-7. Watchpoint Monitor Status Register (WMSR) Table 20-13 describes the WMSR fields. Table 20-13. WMSR Field Descriptions Bits 0 Name ACT Description Active 0 The start triggering event has not occurred; watchpoint monitor is not armed. 1 The start triggering event has occurred; watchpoint monitor is armed. 1 TRIG Triggered 0 The programmed event in WMCR0 has not yet been triggered. 1 The programmed event in WMCR0 has been triggered at least once. — Reserved 2–31 20.3.2 Trace Buffer Register Descriptions The following sections describes the trace buffer registers. 20.3.2.1 Trace Buffer Control Registers (TBCR0, TBCR1) The trace buffer control registers (TBCR0, TBCR1), shown in Figure 20-8 and Figure 20-9, specify trace buffer events. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 20-16 Freescale Semiconductor Memory Map/Register Definition 0 1 2 3 4 5 6 7 8 13 14 15 R EN W Reset 16 20 0 AMD TMD ECEN NECEN SIDEN TIDEN HALT 0000_0000_0000_0000 21 23 24 0 0 0 0 0 MODE 28 29 31 R W Reset Offset 0 0 0 0 0 STRT 0 0 0 0 0 STOP 0000_0000_0000_0000 0xE_2040 Figure 20-8. Trace Buffer Control Register 0 (TBCR0) Table 20-14 describes the TBCR0 fields. Table 20-14. TBCR0 Field Descriptions Bits 0 Name EN Enable 0 The trace buffer facility is disabled. 1 The trace buffer facility is enabled. Address match disable 0 The address match is used to qualify a trace buffer event. 1 The address match is ignored when detecting a trace buffer event. Transaction match disable 0 The transaction type match is used to qualify a trace buffer event. 1 The transaction type match is ignored when detecting a trace buffer event. Equal context enable. Qualifies the matching of current context with programmed context as a trace buffer event criterion, as written in the context registers described in Section 20.3.3, “Context ID Registers.” 0 Current context match does not affect trace buffer event detection. 1 Trace buffer events are qualified by comparing current context with the programmed context event value. N ote: ECEN and NECEN must not be enabled in the same run. If both are set, watchpoint events are inhibited (never occur). Description 1 AMD 2 TMD 3 ECEN 4 NECEN Not equal context enable. Qualifies the matching of current context with programmed context as a trace buffer event criterion, as written in the context registers described in Section 20.3.3, “Context ID Registers.” 0 The failure of a current context match does not affect trace buffer event detection 1 trace buffer events are qualified with NOT getting a current context compare with the programmed context event value. N ote: ECEN and NECEN must not be enabled in the same run. If both are set, watchpoint events are inhibited (never occur). SIDEN Source ID enable 0 Trace buffer events ignore the programmed source ID value. 1 Trace buffer events are qualified by comparison with the programmed SID event value. TIDEN Target ID enable 0 Trace buffer events ignore the programmed TID event value. 1 Trace buffer events are qualified by comparison with the programmed TID event value. This comparison only applies when the ECM is selected for tracing (TBCR1[IFSEL] is all zeros). 5 6 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 20-17 Debug Features and Watchpoint Facility Table 20-14. TBCR0 Field Descriptions (continued) Bits 7 8–13 14–15 Name HALT — MODE Description Halt causes the trace buffer to stop tracing immediately. Reserved Trace mode. Specifies one of two trace modes 00 Trace every valid transaction 01 Reserved 10 Trace only cycles in which a trace event is detected. Note that if EN and other TBCR0 fields are not properly programmed to specify a traceable event, tracing occurs for every valid address. 11 Reserved Reserved Start condition. Specifies the event that arms the trace buffer to start looking for the programmed event. 000 No event. Armed immediately 001 Watchpoint monitor event is detected 010 Trace buffer event is detected 011 Performance monitor signals overflow 100 TRIG_IN transitions from 0 to 1. 101 TRIG_IN transitions from 1 to 0. 110 Current context ID equals programmed context ID 111 Current context ID does not equal programed context ID Reserved Trace stop mode. Specifies the event that stops the updating of the trace buffer after it has been started. Trace buffer only stops after it has been triggered at least once. 000 Buffer is full 001 Watchpoint monitor event is detected 010 Trace buffer event is detected 011 Performance monitor signals overflow 100 TRIG_IN transitions from 0 to 1. 101 TRIG_IN transitions from 1 to 0. 110 Current context ID equals programmed context ID 111 Current context ID does not equal programed context ID 16–20 21–23 — STRT 24–28 29–31 — STOP 0 4 5 7 8 9 10 11 15 16 26 27 31 R0 W Reset Offset 0 0 0 0 IFSEL 0 0 0 SID 00000000000 TID 0000_0000_0000_0000_0000_0000_0000_0000 0xE_2044 Figure 20-9. Trace Buffer Control Register 1 (TBCR1) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 20-18 Freescale Semiconductor Memory Map/Register Definition Table 20-15 describes the TBCR1 fields. Table 20-15. TBCR1 Field Descriptions Bits 0–4 5–7 Name — Reserved Description IFSEL Interface selection. Specifies the interface that sources information for both comparison/buffer control and buffer data capture 000 Selects e500 coherency module (ECM) dispatch interface 001 Selects internal DDR SDRAM interface 010 Selects internal PCI outbound interface 011 Selects internal RapidIO interface 1xx Reserved — SID — TID Reserved Source ID. Specifies the source ID associated with TBCR0[SIDEN]. The source ID is defined in Table 20-26. Reserved Target ID. Specifies the target ID associated with TBCR0[TIDEN]. The target ID is defined in Table 20-26. 8–10 11–15 16–26 27–31 20.3.2.2 Trace Buffer Address Register (TBAR) The trace buffer address register (TBAR) shown in Figure 20-10 contains the address to match against (if TBCR0[AMD] is zero). The transaction address is qualified with the bits described in Section 20.3.2.3, “Trace Buffer Address Mask Register (TBAMR),” before being compared with TBAR. Note that the contents of TBAR are not qualified with TBAMR. 0 31 R TBA W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0xE_204C Figure 20-10. Trace Buffer Address Register (TBAR) Table 20-16 describes the TBAR field. Table 20-16. TBAR Field Descriptions Bits 0–31 Name TBA Trace buffer address Description 20.3.2.3 Trace Buffer Address Mask Register (TBAMR) The trace buffer address mask register (TBAMR) shown in Figure 20-11 contains the mask that is applied to a transaction address before the address is compared with TBAR. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 20-19 Debug Features and Watchpoint Facility 0 31 R TBAM W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0xE_2054 Figure 20-11. Trace Buffer Address Mask Register (TBAMR) Table 20-17 describes the TBAMR field. Table 20-17. TBAMR Field Descriptions Bits 0–31 Name Description TBAM Trace buffer address mask.A value of zero masks the address comparison for the corresponding address bit. 20.3.2.4 Trace Buffer Transaction Mask Register (TBTMR) The trace buffer transaction mask register (TBTMR) shown in Figure 20-12 specifies which transaction types to monitor. Each bit in the TBTMR represents a transaction type on the selected interface. The transaction associated with any particular bit depends on the interface being monitored as specified by TBCR1[IFSEL]. Note that the transactions used for defining trace buffer events are the same as those defined for watchpoint monitor events. Thus, Table 20-12 defines the transaction types associated with each interface. Setting a bit enables a hit when this transaction is matched (provided all other match criteria are met and TBCR0[TMD] is clear). Different interfaces support different transaction types, and the same bit may represent different transaction types depending on the interface. 0 31 R TBTM W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0xE_2058 Figure 20-12. Trace Buffer Transaction Mask Register (TBTMR) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 20-20 Freescale Semiconductor Memory Map/Register Definition Table 20-18 describes the TBTMR field. Table 20-18. TBTMR Field Descriptions Bits 0–31 Name Description TBTM Trace buffer transaction mask. Each bit corresponds to a transaction type as defined in Table 20-12. The transaction associated with a bit depends on the interface being monitored. A value of 1 for a given mask bit enables the matching of the transaction associated with that bit. These bits are meaningful only when TBCR0[TMD]=0. 20.3.2.5 Trace Buffer Status Register (TBSR) The trace buffer status register (TBSR) shown in Figure 20-13 indicates the operational state of the trace buffer. 0 1 2 3 4 23 24 31 R ACT TRIG STP WRAP W Reset Offset 0 00000000 00000000000 C_INDX 0000_0000_0000_0000_0000_0000_0000_0000 0xE_205C Figure 20-13. Trace Buffer Status Register (TBSR) Table 20-19 describes the TBSR fields. Table 20-19. TBSR Field Descriptions B its 0 Name ACT Description Active. Indicates trace buffer activity 0 The start triggering event has not yet occurred. Trace buffer is not armed 1 The start triggering event has occurred. Trace buffer is armed. Triggered. Indicates whether or not a programmed event has been triggered 0 The programmed event in TBCR0 has not yet been triggered. 1 The programmed event in TBCR0 has been triggered at least once. Stopped. Indicates whether or not a trace buffer stop condition has been detected. 0 No stop condition yet detected 1 The trace buffer has detected a stop condition and is no longer capturing events. Wrapped. Indicates that the trace buffer write pointer has wrapped to the beginning of the buffer at least once. Set when the last entry of the trace buffer is written 0 Pointer has not yet wrapped 1 Pointer has wrapped to the beginning at least once Reserved 1 TRIG 2 STP 3 WRAP 4–23 — 24–31 C_INDX Current index. Represents the current value of the write pointer at the time TBSR was read. This value may be written by software to initialize the write pointer; however, software is not allowed to write the write pointer while the trace buffer is active. Writes are ignored while the trace buffer is active. It is recommended to write the status register before enabling the trace buffer in order to zero out any bits that might have been set during a prior run and to initialize the write pointer to zero. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 20-21 Debug Features and Watchpoint Facility 20.3.2.6 Trace Buffer Access Control Register (TBACR) The trace buffer access control register (TBACR), shown in Figure 20-14, enables software to read or write the trace buffer. Each entry is 64 bits; therefore, it takes one write of TBACR and two reads of the access data register (TBADR and TBADHR) to read one 256-entry array entry. Similarly, it takes one write of TBACR and two writes of TBADR and TBADHR to write one array entry. Software can access any entry by writing the appropriate index into TBACR[INDX]. To read or write the buffer sequentially, starting with entry 0, the index must start with a value of 0 and increment every time a new entry is accessed. 0 1 2 23 24 31 R RD WR W Reset Offset 0000000000000000000000 INDX 0000_0000_0000_0000_0000_0000_0000_0000 0xE_2060 Figure 20-14. Trace Buffer Access Control Register (TBACR) Table 20-20 describes the TBACR fields. Table 20-20. TBACR Field Descriptions Bits 0 1 Name RD WR Description Read command. When set, a trace buffer read is performed using the value of TBACR[INDX]. This bit is automatically cleared when the read is performed. Write command. When set, a trace buffer write is performed using the value of TBACR[INDX]. This bit is automatically cleared when the write is performed. A write occurs only if the trace buffer is not active: write requests are ignored while the buffer is active. Reserved Buffer index to read from or write into (0–255). Used in conjunction with TBACR[RD] and TBACR[WR] 2–23 24–31 — INDX 20.3.2.7 Trace Buffer Access Data High Register (TBADHR) The trace buffer access data high register (TBADHR), shown in Figure 20-15, contains the high-order 32 bits of the data read from the trace buffer during a software-initiated read command (TBACR[RD]), or the write data to be written into the trace buffer during a software-initiated write command (TBACR[WR]). TBACR must be configured to perform a read before this register contains valid data. This register must be initialized by software before configuring the TBACR to perform a write command. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 20-22 Freescale Semiconductor Memory Map/Register Definition 0 31 R TBADH W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0xE_2064 Figure 20-15. Trace Buffer Read High Register (TBADHR) Table 20-21 describes TBADHR. Table 20-21. TBADHR Field Descriptions Bits 0–31 Name Description TBADH Trace buffer access data high. The higher 32 bits of the data read from or to be written into the trace buffer, depending on whether the array is accessed with a read or a write. 20.3.2.8 Trace Buffer Access Data Register (TBADR) The trace buffer access data register (TBADR), shown in Figure 20-16, contains the low-order 32 bits of the data read from the trace buffer during a software-initiated read command (TBACR[RD]) or the write data to be written into the trace buffer during a software-initiated write command (TBACR[WR]). TBACR must be configured to perform a read before this register contains valid data. This register must be initialized by software before configuring the TBACR to perform a write command. 0 31 R TBAD W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0xE_2068 Figure 20-16. Trace Buffer Access Data Register (TBADR) Table 20-22 describes the TBADR field. Table 20-22. TBADR Field Descriptions Bits 0–31 Name Description TBAD Trace buffer access data. Corresponds to the lower 32 bits of the data read from the trace buffer or to be written into the trace buffer, depending on whether software is accessing the array with a read or a write. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 20-23 Debug Features and Watchpoint Facility 20.3.3 Context ID Registers This section describes the context ID registers. The current context ID register (CCIDR) and programmed context ID registers (PCIDR) are set by software and facilitate debugging complex software. 20.3.3.1 Programmed Context ID Register (PCIDR) The programmed context ID register (PCIDR), shown in Figure 20-17, contains the user-programmed context ID. This register can be configured to trigger watchpoint events when its value matches the current context ID register (CCIDR), as controlled by WMCR0[ECEN] and WMCR0[NECEN]. See Section 20.3.1.1, “Watchpoint Monitor Control Registers 0–1 (WMCR0, WMCR1),” for more information. 0 31 R PCID W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0xE_20A0 Figure 20-17. Programmed Context ID Register (PCIDR) Table 20-23 describes the PCIDR field. Table 20-23. PCIDR Field Descriptions Bits 0–31 Name Description PCID Programmed context ID. Contains the user-programmed context ID. Compared with current context ID for context-sensitive event triggering. 20.3.3.2 Current Context ID Register (CCIDR) The current context ID register (CCIDR) shown in Figure 20-18 contains the current context ID. This register is written by software after a context switch and can be used to trigger events when compared with the programmed context ID register (PCIDR). 0 31 R CCID W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0xE_20A4 Figure 20-18. Current Context ID Register (CCIDR) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 20-24 Freescale Semiconductor Memory Map/Register Definition Table 20-24 describes the CCIDR field. Table 20-24. CCIDR Field Descriptions Bits 0–31 Name Description CCID Current context ID. Set by user software. Typically loaded immediately following a context switch. Compared with user-programmed context ID for context-sensitive event triggering. 20.3.4 Trigger Out Function TRIG_OUT provides a convenient mechanism for triggering external system monitors and diagnostic equipment such as logic analyzers. Note that READY is multiplexed with TRIG_OUT. See the last paragraph of Section 4.4.2, “Power-On Reset Sequence,” for more information about READY functionality. When the trace buffer hit is selected by TOSR[SEL], TRIG_OUT is only meaningful if the trace buffer control register 0 (TBCR0) is properly configured to hit on a traceable event. The same holds true for the watchpoint monitor when the watchpoint monitor is selected by TOSR[SEL]. 20.3.4.1 Trigger Out Source Register (TOSR) The trigger out source register (TOSR) shown in Figure 20-19 specifies the source for TRIG_OUT. The three event-trigger sources are the following: • • • The watchpoint monitor The trace buffer The performance monitor 0 4 5 7 8 30 31 R0 0 0 0 0 SEL W Reset Offset 00000000000000000000000 SYNC 0000_0000_0000_0000_0000_0000_0000_0000 0xE_20B0 Figure 20-19. Trigger Out Source Register (TOSR) MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 20-25 Debug Features and Watchpoint Facility Table 20-25 describes the TOSR fields. Table 20-25. TOSR Field Descriptions Bits 0–4 5–7 Name — SEL Reserved Select. Selects the source for TRIG_OUT 000 READY signal. Multiplexed with TRIG_OUT. Basic device state indicator. READY asserts whenever the device is not in reset or not asleep. See Chapter 4, “Reset, Clocking, and Initialization,” for more details about the reset sequence, and Chapter 18, “Global Utilities,” for more information about power management states. 001 Selects the watchpoint monitor hit indication 010 Selects the trace buffer hit indication 011 Selects the performance monitor overflow indication Reserved Description 8–31 — 20.4 Functional Description The debug features on the MPC8540 use the LBC interfaces, DDR SDRAM interface, and the PCI interface. 20.4.1 Source and Target ID Debug information that is common to all the interfaces is the source ID (SID). The transaction source ID provides enough information to determine which block or port originated a transaction including the distinction between instruction and data fetches from the processor core. Table 20-26 shows the values and interpretation for the 5-bit SID field. Note that the table also includes ports that are only slaves, such as local memory. These ports are always targets. As such, the value shown represents a target ID (TID) and not a source ID. For ports that can function in both capacities, the value indicates source ID when mastering transactions, and target ID when responding as slave. The TID field is only meaningful when one of the following participates in the transaction: • • • The e500 coherency module (ECM) dispatch bus The watchpoint monitor (WMCR1[IFSEL] = 000) The trace buffer (TBCR1[IFSEL] = 000) Table 20-26. Source and Target ID Values Value (Hex) 00 01 02 Source (or Target) Port PCI/PCI-X Reserved Reserved Value (Hex) 10 11 12 Source (or Target) Port Local processor (instruction fetch) Local processor (data fetch) Reserved MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 20-26 Freescale Semiconductor Functional Description Table 20-26. Source and Target ID Values (continued) Value (Hex) 03 04 05 06 07 08 09 0A 0B 0C 0D 0E 0F Reserved Local bus controller Reserved Reserved Reserved Configuration space Reserved Boot sequencer Reserved RapidIO Reserved Reserved Local space (DDR) Source (or Target) Port Value (Hex) 13 14 15 16 17 18 19 1A 1B 1C 1D 1E 1F Reserved Reserved DMA Reserved System access port (SAP) TSEC1 TSEC2 FEC Reserved RapidIO message unit RapidIO doorbell unit RapidIO port-write unit Non valid port indicator (reserved for debug info) Source (or Target) Port 20.4.2 PCI/PCI-X Interface Debug If PCI_GNT3 is low when sampled during POR, the PCI/PCI-X interface operates in debug mode. In debug mode the source ID appears on the high-order address bits (PCI_AD[62:58]) during the bus command phase of a PCI/PCI-X transaction. The bus command phase occurs either during the first cycle that PCI_FRAME is asserted, or, in the case of addresses greater than 32 bits, after a dual-address cycle phase. In either case, the debug information appears on the highest order address bits while PCI_FRAME is asserted and both PCI_IRDY and PCI_TRDY are negated. When accessing the low 4 Gbytes of PCI address space for which no dual-address cycle is needed, the debug information appears during the first (and only) address phase on PCI_AD[62:58]. Whenever a dual-address cycle must be run, (addresses above 4 Gbytes) the debug information appears on PCI_AD[62:58] during the second address cycle. In either case a logic analyzer should be configured to sample information on the first cycle of the assertion of PCI_FRAME and the cycle following a dual-address cycle command. NOTE Because they share the same pins, an entire 64-bit address and the debug information cannot be captured in a single cycle. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 20-27 Debug Features and Watchpoint Facility 20.4.3 DDR SDRAM Interface Debug The DDR interface has two debug modes distinguished by which pins drive the debug information. In one mode, debug information (source ID, data valid) is multiplexed onto the ECC pins; the other mode uses the debug pins. 20.4.3.1 Debug Information on Debug Pins If MSRCID0 is high when sampled during POR, the debug information from the DDR SDRAM interface is driven on MSRCID[0:4] and MDVAL. This POR value is captured in PORDBGMSR[MEM_SEL] as described in Section 18.4.1.5, “POR Debug Mode Status Register (PORDBGMSR).” In this mode, the source ID appears on MSRCID[0:4] during a RAS or CAS cycle. During any other cycle, the value of MSRCID[0:4] is all ones, which indicates idle cycles on the address/command interface. Similarly, MDVAL is asserted during valid data cycles on the DDR interface. 20.4.3.2 Debug Information on ECC Pins If MSRCID1 is low when sampled during POR, debug information from the DDR SDRAM interface is selected to appear on MECC[0:5] as shown in Figure 20-1. In this mode, the ID value of the source port, (the source ID), appears on MECC[0:4] during a RAS or CAS cycle. During any other cycle the value of MECC[0:4] is all ones. A data-valid signal (DVAL) is driven on MECC5 during valid DDR SDRAM data cycles. NOTE In this mode, MECC[0:5] must be disconnected from all SDRAM devices to prevent contention on those lines. 20.4.4 Local Bus Interface Debug If MSRCID0 is low when sampled during POR, the LBC is selected as the source for the debug information appearing on MSRCID[0:4] and MDVAL. For more information on this mode, see Section 13.1.3.2, “Source ID Debug Mode.” 20.4.5 Watchpoint Monitor The watchpoint monitor (WM) can be programmed to arm and trigger on many different events including any of the following: • • • • External event (through TRIG_IN) A trace buffer event A performance monitor overflow event A comparison of the current and programmed context ID registers MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 20-28 Freescale Semiconductor Functional Description A watchpoint event can be used in the following ways: • • • Trigger a logic analyzer (using TRIG_OUT) Arm or trigger the trace buffer Trigger a performance monitor event The large counters available in the performance monitor block and the interlock between it and the watchpoint monitor support sophisticated debug scenarios. A WM trigger event may be composed of several events programmed in the watchpoint monitor control registers (WMCR0–WMCR1). Because the watchpoint monitor is disabled by default during POR, these registers must be initialized to make use of this debug feature. Note that the WM address mask register (WMAMR) and the type mask register (WMTMR) are cleared during POR. This means that the watchpoint monitor’s default behavior following a power-on reset is to trigger on any address and no transaction type. The reset value of WMCR0[TMD] is 0 which means transaction matching is enabled but since no transaction is selected (WMTMR=0), a match will never occur. Either the transaction matching must be disabled by setting WMCR0[TMD] to a value of 1, or valid transactions must be selected by setting one or more of the WMTMR bits to a value of 1. 20.4.5.1 Watchpoint Monitor Performance Monitor Events The WM can produce a performance monitor (PM) event with every trigger. This is accomplished by configuring the performance monitor to count WM events. For more information on this configuration see the events named “Number of watchpoint monitor hits,” and “Number of trace buffer hits,” in Table 19-10. Multi-level triggers can be created using the watchpoint monitor, the performance monitor, and the trace buffer combined. For example, the WM can be programmed to trigger on events that also increment a PM counter (the performance monitor must also be programmed to respond to this event), the output of which (perfmon_overflow) could trigger the start of tracing in the trace buffer. 20.4.6 Trace Buffer The trace buffer is a 256 × 64 array that can capture information about the internal processing of transactions to selected interfaces. The trace buffer controls are a superset of those for the watchpoint monitor. Close inspection of the trace buffer control registers (TBCRn) and the WM control registers (WMCRn) shows that trace buffer controls not needed for the WM are marked reserved in WMCRn. This permits using the trace buffer as a second watchpoint monitor by simply ignoring the trace options. The trace buffer provides great flexibility about when to start tracing, when to stop tracing, and what to trace. The trace mode field, TBCR0[MODE], indicates when to trace: on every valid cycle, on a watchpoint monitor event, or when all the programmed events in the TBCR are met. This MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 20-29 Debug Features and Watchpoint Facility permits a user to program the trace condition in the watchpoint monitor and to program a start or stop condition in the trace buffer control register. The user can also program the TBCR with the conditions in which to stop tracing: on an event, or when the buffer is full. TBCR0[IFSEL] specifies which interface transactions are being captured. The trace buffer can be programmed to trace the dispatch bus from any of the following: • • • • e500 coherency module (ECM) Outbound host interface to the RapidIO controller Outbound host interface to the PCI controller Host interface to the DDR controller Transactions come into the ECM, arbitrate for common resources, and get dispatched to the target port. Information such as transaction types, source ID, and other attributes can be captured in any of the selected interfaces. The trace buffer can be used to distinguish between requests due to an instruction fetch versus a data fetch on the RapidIO interface, where external I/O alone does not allow for easy discrimination. 20.4.6.1 Traced Data Formats (as a Function of TBCR1[IFSEL]) Figure 20-20 shows the trace buffer entry format for an ECM dispatch (CMD) transaction that is specified when TBCR1[IFSEL] = 000. 0 4 5 9 10 13 14 18 19 31 32 63 R CMDTT W Reset 0000_0000_0000_0000_0000_0000_0000_0000 CMDSID CMDTID CMDBC — CMDADDR Figure 20-20. e500 Coherency Module Dispatch (CMD) Trace Buffer Entry Table 20-27 describes the fields of CMD trace buffer entries. Table 20-27. CMD Trace Buffer Entry Field Descriptions (TBCR1[IFSEL] = 000) Bits 0–4 5–9 10–13 Name CMDTT CMDSID CMDTID Function Transaction type. Specifies the transaction type as shown in Table 20-12. For example, a value of zero indicates a write with local processor snoop condition. Source ID. Identifies the source of the transaction as shown in Table 20-26. For example, a value of 0b01100 indicates that RapidIO is the transaction source. Target ID. Identifies the target of the transaction as shown in Table 20-26. For example, a value of 0b1100 indicates that RapidIO is the transaction target. Note that for a target ID the topmost bit of the 5-bit field is always 0 and, therefore, only 4 bits are required to define a target interface. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 20-30 Freescale Semiconductor Functional Description Table 20-27. CMD Trace Buffer Entry Field Descriptions (TBCR1[IFSEL] = 000) (continued) Bits 14–18 Name CMDBC Function Byte count. Range: 32 to 1 where a value of 0 indicates 32 bytes. 00000 32 bytes 00001 1 byte 00010 2 bytes ··· 11110 30 bytes 11111 31 bytes Reserved Address bits 0–31 19–31 32–63 — CMDADDR Figure 20-21 shows the trace buffer entry format for the DDR SDRAM interface, TBCR1[IFSEL] = 001. 0 4 5 9 10 13 14 18 19 31 32 63 R DDRTT W Reset 0000_0000_0000_0000_0000_0000_0000_0000 DDRSID — DDRBC — DDRADDR Figure 20-21. DDR Trace Buffer Entry Table 20-28 describes the fields of DDR SDRAM trace buffer entries when TBCR1[IFSEL] = 001. Table 20-28. DDR Trace Buffer Entry Field Descriptions (TBCR1[IFSEL] = 001) Bits 0–4 5–9 10–13 14–18 19–31 32–63 Name DDRTT DDRSID — DDRBC — DDRADDR Function Transaction type. Specifies the transaction type as shown in Table 20-12. For example, a value of all zeros maps to write. Source ID. Specifies the source of the transaction as shown in Table 20-26. For example, a value of 01100 indicates that RapidIO is the transaction source, and so on. Reserved Byte count Reserved Address bits 0–31 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 20-31 Debug Features and Watchpoint Facility Figure 20-22 shows the PCI trace buffer entry format when TBCR1[IFSEL] = 010. 0 4 5 9 10 11 12 31 32 63 R PCITT W Reset 0000_0000_0000_0000_0000_0000_0000_0000 PCISID PCIBC — PCIADDR Figure 20-22. PCI Trace Buffer Entry Table 20-29 describes the fields of PCI trace buffer entries when TBCR1[IFSEL] = 010. Table 20-29. PCI Trace Buffer Entry Field Descriptions (TBCR1[IFSEL] = 010) Bits 0–4 5–9 10–11 Name PCITT PCISID PCIBC Function Transaction type. Specifies the transaction type as shown in Table 20-12. For example, a value of all zeros maps to write. Source ID. Identifies the source of the transaction as shown in Table 20-26. For example, a value of 01100 identifies RapidIO as the transaction source. Byte count. The size of the transaction. 00 32 bytes 01 8 bytes 10 16 bytes 11 24 bytes Reserved 12–31 — 32–63 PCIADDR Address bits 0–31 Figure 20-23 shows the format for RapidIO interface trace buffer entries when TBCR1[IFSEL] = 011. 0 4 5 9 10 13 14 18 19 32 33 63 R RIOTT W Reset RIOSID — RIODWC — RIOADDR 0000_0000_0000_0000_0000_0000_0000_0000 Figure 20-23. RapidIO Trace Buffer Entry Table 20-30 describes the fields of the RapidIO interface trace buffer entries. Table 20-30. RapidIO Trace Buffer Entry Field Descriptions (TBCR1[IFSEL] = 011) Bits 0–4 5–9 Name RIOTT RIOSID Function Transaction type. Specifies the transaction type as shown in Table 20-12. For example, a value of all zeros indicates an NWRITE transaction. Source ID. Identifies the source of the transaction as shown in Table 20-26. For example, a value of 01100 identifies RapidIO as the transaction source. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 20-32 Freescale Semiconductor Initialization Table 20-30. RapidIO Trace Buffer Entry Field Descriptions (TBCR1[IFSEL] = 011) (continued) Bits 10–13 14–18 19–32 33–63 Name — RIODWC — RIOADDR Reserved Double-word count. This number times eight equals the byte count. Reserved Address bits 0–30. Represents a double-word address on RapidIO Function 20.5 Initialization Configuring the appropriate control register must be the last step in the initialization sequence for either the watchpoint or trace buffer. That is, all required registers except the corresponding control register must be configured before any control register bits that enable watchpoint or trace events are set. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 20-33 Debug Features and Watchpoint Facility MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 20-34 Freescale Semiconductor Chapter 21 10/100 Fast Ethernet Controller The MPC8540 provides a 10/100 fast Ethernet controller (FEC) operating at 10 to 100 megabits per second (Mbps) that can be used as a device debug and maintenance port. The three-speed (10/100/1Gb) Ethernet controller (TSEC) is separate from this 10/100 Ethernet controller, and is described in Chapter 14, “Three-Speed Ethernet Controllers.” 21.1 Introduction The 10/100 Ethernet standard supports the IEEE 802.3 Ethernet frame format, backward compatibility for installed media, and the use of full- or half-duplex CSMA/CD. The Ethernet protocol implements the bottom two layers of the open systems interconnection (OSI) 7-layer model, that is, the data link and physical layers. Figure 21-1 depicts the typical Ethernet protocol stack and the relationship to the OSI model. OSI Reference Model Layers LAN CSMA/CD Layers Higher layers Logical link control (LLC) Media access control (MAC) Reconciliation MII Physical PCS PMA PMD MDI Physical medium 10/100 Mbps 10/100 Ethernet Data Link Figure 21-1. 10/100 Fast Ethernet Controller in Relation to the OSI Protocol Stack MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 21-1 10/100 Fast Ethernet Controller Ethernet provides the following sublayers: • Media access control (MAC) sublayer—The MAC sublayer provides a logical connection between the MAC and its peer station. Its primary responsibility is to initialize, control, and manage the connection with the peer station. Reconciliation sublayer—The reconciliation sublayer acts as a command translator. It maps the terminology and commands used in the MAC layer into electrical formats appropriate for the physical layer entities. MII (media-independent interface) sublayer—The MII sublayer provides a standard interface between the MAC layer and the physical layer for 10/100 Mbps operation. It isolates the MAC layer and the physical layer, enabling the MAC layer to be used with various implementations of the physical layer. PCS (physical coding sublayer)—The PCS sublayer is responsible for encoding and decoding data stream to and from the MAC sublayer. PMA (physical medium attachment) sublayer—The PMA sublayer is responsible for serializing code groups into a bit stream suitable for serial bit-oriented physical devices(SERDES) and vice versa. Synchronization is also performed for proper data decoding in this sublayer. The PMA sits between the PCS and the PMD sublayers. PMD (physical medium dependent) sublayer—The PMD sublayer is responsible for signal transmission. The typical PMD functionality includes amplifier, modulation, and wave shaping. Different PMD devices may support different media. MDI (medium-dependent interface) sublayer—MDI is a connector. It defines different connector types for different physical media and PMD devices. • • • • • • Ethernet/IEEE 802.3 frames are based on the frame structure shown in Figure 21-2. The term “packet” is sometimes used to refer to the frame plus the preamble and start frame delimiter (SFD). Frame length is 64–1518 bytes Preamble 7 bytes Start frame delimiter 1 byte Destination address 6 bytes Source address 6 bytes Type/ Length 2 bytes Data 46–1500 bytes Frame check sequence 4 bytes Figure 21-2. Ethernet/IEEE 802.3 Frame Structure The elements of an Ethernet frame are as follows: • • Preamble—The 7-byte preamble of alternating ones and zeros used for receiver timing synchronization (each byte containing the value 0x55) Start frame delimiter (SFD)—A sequence of 0xD5 (10101011 because the bit ordering is lsb first) indicates the beginning of the frame MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 21-2 Freescale Semiconductor Introduction • • • • • 48-bit destination address (DA)—The first bit identifies the address as an individual address (0) or a group address (1). The second bit is used to indicate whether the address is locally-defined (1) or globally-defined (0) 48-bit source address (SA)—(Original versions of the IEEE 802.3 specification allowed 16-bit addressing, which has never been used widely.) Ethernet type field/IEEE 802.3 length field—The type field signifies the protocol (for example, TCP/IP) used in the rest of the frame. The length field specifies the length of the data portion of the frame. For both Ethernet and IEEE 802.3 frames to exist on the same LAN, the length field must be unique from any type fields used in Ethernet. This limitation requires that a type field be identified by a decimal number equal to or greater than 1536 (0x0600) but less than 65535 (0xFFFF). If the number, however, is between 0 and 1,500 (0x0000 through 0x05DC) then this field indicates the length of the MAC client data. The range from 1,501 to 1,535 (0x5DD through 0x5FF) was intentionally left undefined. Data and padding—Padding is optional. It is only needed if the data is smaller than 46 octets (one octet = one byte) to ensure the minimum frame size of 64 octets as specified in the IEEE 802.3 standard. In 802.3x the first two octets of the data field are used as opcode (OP) (pause = 0x0001) and the second two octets are used to transmit a pause time (PT) parameter (pausetime = 0x0000 for on and 0xFFFF for off). In addition, a third two-octet field can be used for an extended pause control parameter (PTE). Because the use of these fields varies with the protocol used, the ability to examine them and report their content can significantly accelerate Ethernet frame processing. Frame-check sequence (FCS)—Specifies the standard 32-bit cyclic redundancy check (CRC) obtained using the standard CCITT-CRC polynomial on all fields except the preamble, SFD and CRC. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 21-3 10/100 Fast Ethernet Controller Figure 21-3 details the Ethernet/IEEE 802.3 frame structure. Frame length is 64–1518 bytes Preamble 7 bytes Start frame delimiter 1 byte Destination address 6 bytes Source address 6 bytes Type/ Length 2 bytes Data 46–1500 bytes Frame check sequence 4 bytes Vendor 3 bytes Bit 0 (LSB) yzxx xxxx I/G U/L xxxx xxxx UserValue DSAP SSAP Control Information LLC frame Destination address xxxx xxxx xxxx xxxx xxxx xxxx Bit 47 xxxx xxxx LLD destination address Example: 02608C : 876543 (02608C = 3COM, 876543 = UserValue assigned by 3COM) Note: The lsb of each octet is transmitted first. Figure 21-3. Ethernet/IEEE 802.3 Frame Structure With More Details Relative to Figure 21-3, the IEEE 802.3 section 3.11 (MAC frame format) defines the frame format such that the octets of a frame are transmitted from left to right (preamble first, the FCS last), and the bits of each octet are transmitted least-significant bit (lsb) first. The destination address example shown in Figure 21-3 (02608C:876543) which would normally be written as 0000 0010 0110 0000 1000 1100 1000 0111 0110 0101 0100 0011 is transmitted bit by bit as 0100 0000 0000 0110 0011 0001 1110 0001 1010 0110 1100 0010 which is an individual address (because the lsb is cleared) but locally-defined (because the second least-significant bit is set). When first originated, a type field was used for protocol identification. The IEEE 802.3 specification eliminated the type field, replacing it with the length field. The length field is used to identify the length, in bytes, of the data field. The protocol type in 802.3 frames is held within the data portion of the packet. The logical link control (LLC) is responsible for providing services to the network layer regardless of media type, such as FDDI, Ethernet, token ring, and others. The LLC layer makes use of LLC protocol data units (PDUs) in order to communicate between the media access control (MAC) layer and the upper layers of the protocol stack. Three variables determine access into the upper layers via the LLC-PDU. The variables include the destination service access point (DSAP), the source service access point (SSAP), and a control variable. The DSAP address specifies a unique identifier within the station providing protocol information for the upper layer. The SSAP provides the same information for the source address. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 21-4 Freescale Semiconductor Introduction The LLC defines service access for protocols that conform to the open system interconnection (OSI) model for network protocols. However, many protocols do not obey the rules for those layers and additional information must be added to the LLC in order to provide information regarding those protocols. Protocols that fall into this category include IP and IPX. The method used to provide this additional protocol information is called a subnetwork access protocol (SNAP) frame. A SNAP encapsulation is indicated by the DSAP and SSAP addresses being set to 0xAA and the LLC control field being set to 0x03. If that address is seen, a SNAP header follows. The SNAP header is five bytes long. The first three bytes consist of the organization code (SNAP OUI), which is assigned by the IEEE. The last two bytes become the type value set from the original Ethernet specifications if SNAP OUI = 0, or they become a SNAP protocol identifier if SNAP OUI is nonzero. The FEC allows the flexibility to accelerate the identification and retrieval of all the standard and non-standard protocols mentioned above. Figure 21-4 depicts the block diagram of the FEC. Tx/Rx descriptors Tx/Rx data Tx/Rx status DMA Block interface Rx FIFO 1 kbyte FIFO control Tx FIFO 1 kbyte Address/Data filtering Pack words Unpack words Tx system CLK data path MAC Layer Clocks To PHY MII Mgmt Figure 21-4. FEC Block Diagram 21.1.1 10/100 Fast Ethernet Controller Overview The FEC is designed to support 10/100 Mbps Ethernet/802.3 networks and contains the following components: • • • Ethernet media access controller (MAC) First-in first-out (FIFO) controller Direct memory access (DMA) controller MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 21-5 10/100 Fast Ethernet Controller The most-significant byte of data in a receive or transmit data buffer corresponds to the most-significant byte of a frame, respectively. The complete FEC is designed for single MAC applications. The FEC supports the 10/100 Mbps MII interface to connect to an external Ethernet transceiver. The FEC software programming model is similar to the MPC8260 (PowerQUICC II) device. Hence, it enables Freescale Semiconductor customers to leverage already implemented Ethernet drivers, reducing the software development cycle. 21.2 Features The MPC8540 FEC includes these distinctive features: • • IEEE 802.3, 802.3u, 802.3x, 802.3ac compliant Support for different Ethernet physical interfaces: — 10/100 Mbps IEEE 802.3 MII — 10 Mbps IEEE 802.3 MII Full- and half-duplex support IEEE 802.3 full-duplex flow control (automatic PAUSE frame generation or software programmed PAUSE frame generation and recognition) Support for out-of-sequence transmit queue (for initiating flow control) Programmable maximum frame length supports jumbo frames (up to 9.6 Kbytes) and IEEE 802.1 virtual local area network (VLAN) frames Retransmission from transmit FIFO following a collision Support for CRC generation and verification of inbound/outbound packets Address recognition — Each exact match can be programmed to be accepted or rejected — Broadcast address (accept/reject) — Exact match 48-bit individual (unicast) address — Hash (256-bit hash) check of individual (unicast) addresses — Hash (256-bit hash) check of group (multicast) addresses — Promiscuous mode Extraction data and its associated buffer descriptors can be directed to processor’s L2 cache to reduce access latency • • • • • • • • MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 21-6 Freescale Semiconductor Modes of Operation 21.3 Modes of Operation The FEC’s primary operational modes are the following: • Full- and half-duplex operation. Determined by MACCFG2 Full_Duplex bit. Full-duplex mode is intended for use on point to point links between switches or end node to switch. Half-duplex mode is used in connections between an end node and a repeater or between repeaters. If configured in half-duplex mode (only 10 and 100 Mbps operation, MACCFG2 register’s Full_Duplex bit is cleared), the MAC complies with the IEEE CSMA/CD access method. If configured in full-duplex mode (10/100 Mbps operation; MACCFG2 register’s Full_Duplex bit is set), the MAC supports flow control. If flow control is enabled, it allows the MAC to receive or send PAUSE frames. 10 Mbps and 100 Mbps MII interface operation. The MAC-PHY interface operates in MII mode by configuring bits 22–23 of MACCFG2 (MACCFG2[I/F_Mode] = 01). The MII is the media-independent interface defined by the 802.3 standard for 10/100 Mbps operation. Operation speed is determined by the FEC_TX_CLK and FEC_RX_CLK signals which are driven by the transceiver. The transceiver auto negotiates the speed. Address recognition options The options supported are promiscuous, broadcast, exact individual address hash or exact match, and multicast hash match. For detailed descriptions refer to Section 21.5.3.7.1, “Individual Address Registers 0–7 (IADDRn),” Section 21.5.3.7.2, “Group Address Registers 0–7 (GADDRn),” Section 21.5.3.4.1, “Receive Control Register (RCTRL),” and Section 21.6.2.5, “Frame Recognition.” Internal loop back Internal loop back mode is selected via the Loop_Back bit in the MACCFG1 register. See Section 21.6.2.9, “Internal and External Loop Back,” for details. • • • 21.4 External Signal Descriptions This section defines the FEC-to-chip signals. The buses are described using the bus convention used in IEEE 802.3 because the PHY follows this same convention (that is, TxD[3:0] means 0 is the lsb). Notice that except for external physical interfaces the buses and registers follow a big-endian format. The FEC network interface supports the MII option that requires 15 I/O signals. The MII option supports both 10- and 100-Mbps Ethernet rates. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 21-7 10/100 Fast Ethernet Controller 21.4.1 Detailed Signal Descriptions Table 21-1 provides a detailed description of the FEC interface signals. The modes follow the IEEE 802.3 standard, 2000 Edition. Input signals not used are internally disabled. Output signals not used are driven low. Table 21-1. FEC Signals—Detailed Signal Descriptions Signal FEC_COL I/O I Description Collision. The behavior of this signal is not specified when in full-duplex mode. State Asserted/Negated—Asserted upon detection of a collision. It must remain asserted Meaning while the collision persists. Timing Asserted/Negated—Not required to transition synchronously with FEC_TX_CLK or FEC_RX_CLK. FEC_CRS I Carrier sense. Used as SDET (signal detect), an optional signal that some PHYs generate. State Asserted/Negated—Asserted when the transmit or receive medium is not idle. In case Meaning of a collision, FEC_CRS must stay asserted through the duration of the collision. Timing Asserted/Negated—Not required to transition synchronously with FEC_TX_CLK or FEC_RX_CLK. FEC_RX_CLK FEC_RX_DV FEC_RXD[3:0 ] FEC_RX_ER I I I Receive clock. This is a continuous clock (2.5, or 25 MHz) that provides a timing reference for FEC_RX_DV, FEC_RXD, and FEC_RX_ER. Receive data valid. If FEC_RX_DV is asserted, the PHY is indicating that valid data is present on the MII interfaces. Receive data in. Represents a nibble of data to be transferred from the PHY to the MAC when FEC_RX_DV is asserted. A completely formed SFD must be passed across the MII. When FEC_RX_DV is not asserted, FEC_RXD has no meaning. Receive error State Asserted—When FEC_RX_ER and FEC_RX_DV are asserted, the PHY has detected Meaning an error in the current frame. Negated—The PHY has not detected an error in the current frame. FEC_TX_CLK I Transmit clock in. FEC_TX_CLK is a continuous clock (with a frequency of 2.5 or 25 MHz) that provides a timing reference for FEC_TX_EN, FEC_TXD, and FEC_TX_ER. FEC_TX_CLK is generated in the PHY and provided to the MAC. Transmit data out State Asserted/Negated—Represents a nibble of data to be sent from the MAC to the PHY Meaning when FEC_TX_EN is asserted; has no meaning if FEC_TX_EN is negated. FEC_TX_EN O Transmit data valid State Asserted—The MAC is indicating that valid data is present on the MII. Meaning Negated—Valid data is not guaranteed to be present on the MII. Timing Asserted/Negated—This signal transitions synchronously with respect to FEC_TX_CLK. I FEC_TXD[3:0] O MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 21-8 Freescale Semiconductor Memory Map/Register Definition Table 21-1. FEC Signals—Detailed Signal Descriptions (continued) Signal FEC_TX_ER I/O O Transmit error State Asserted/Negated—Assertion of this signal for one or more clock cycles while Meaning FEC_TX_EN is asserted causes the PHY to transmit one or more illegal symbols. Asserting FEC_TX_ER has no effect when operating at 10 Mbps or when FEC_TX_EN is negated. Timing Asserted/Negated—This signal transitions synchronously with respect to FEC_TX_CLK. Description 21.5 Memory Map/Register Definition The FEC uses a software model similar to that employed by the Fast Ethernet function supported on the Freescale Semiconductor MPC8260 CPM FCC and in the 10/100 controller of the MPC860T. The FEC is programmed by a combination of control/status registers (CSRs) and buffer descriptors. The CSRs are used for mode control, interrupts, and to access status information. The descriptors are used to pass data buffers and related buffer status or frame information between the hardware and software. All accesses to and from the registers must be made with 32-bit accesses. There is no support for accesses of sizes other than 32 bits. This section of the document defines the memory map and describes the registers in detail. The buffer descriptors are described in Section 21.6.3, “Buffer Descriptors.” The MII registers are described in this section. 21.5.1 Top-Level Module Memory Map The FEC implementation requires 4 Kbytes of memory-mapped space, of which more than 1 Kbyte is reserved for future expansion. The space is divided into the following sections: • • • • • General control/status registers Transmit-specific control/status registers Receive specific control/status registers MAC registers Hash function registers MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 21-9 10/100 Fast Ethernet Controller Table 21-2 defines the top-level memory map. Table 21-2. Module Memory Map Summary Address 000–0FF 100–2FF 300–4FF 500–5FF 600–7FF 800–8FF 900–AFF B00–BFF C00–FFF Function FEC general control/status registers FEC transmit control/status registers FEC receive control/status registers FEC MAC registers Reserved FEC HASH function registers Reserved FEC attribute registers Future expansion space 21.5.2 Detailed Memory Map Table 21-3 shows the address of each register, the name of the register, and a cross-reference to the complete description of the register. Table 21-3. Module Memory Map Offset Name Access 1 Reset Section/Page FEC General Control and Status Registers 0x2_6000– Reserved 0x2_600C 0x2_6010 0x2_6014 0x2_6018 IEVENT—Interrupt event register IMASK—Interrupt mask register EDIS—Error disabled register R R R/W R/W R R/W R/W R/W R 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0040 0x0000_0000 0x0000_0000 0x0000_0000 — 21.5.3.1.1/21-13 21.5.3.1.2/21-16 21.5.3.1.3/21-18 — 21.5.3.1.4/21-18 21.5.3.1.5/21-19 21.5.3.1.6/21-20 — 0x2_601C– Reserved 0x2_6020 0x2_6024 0x2_6028 0x2_602C MINFLR—Minimum frame length register PTV—Pause time value register DMACTRL—DMA control register 0x2_6030– Reserved 0x2_6088 FEC FIFO Control and Status Registers 0x2_604C 0x2_608C FIFO_PAUSE_CTRL—FIFO pause control register FIFO_TX_THR—FIFO transmit threshold register R/W R/W 0x0000_0000 0x0000_0080 21.5.3.2.1/21-22 21.5.3.2.2/21-23 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 21-10 Freescale Semiconductor Memory Map/Register Definition Table 21-3. Module Memory Map (continued) Offset 0x2_6090– Reserved 0x2_6094 0x2_6098 0x2_609C FIFO_TX_STARVE—FIFO transmit starve register FIFO_TX_STARVE_SHUTOFF—FIFO transmit starve shutoff register Name Access 1 R R/W R/W R Reset 0x0000_0000 0x0000_0020 0x0000_0080 0x0000_0000 Section/Page — 21.5.3.3.4/21-27 21.5.3.3.5/21-28 — 0x2_60A0– Reserved 0x2_60FC FEC Transmit Control and Status Registers 0x2_6100 0x2_6104 0x2_6108 0x2_610C TCTRL—Transmit control register TSTAT—Transmit status register Reserved TBDLEN—TxBD data length R/W R/W R R R R R R/W R R/W R R/W R/W R 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 21.5.3.3.1/21-25 21.5.3.3.2/21-26 — 21.5.3.3.3/21-27 — 21.5.3.3.4/21-27 — 21.5.3.3.5/21-28 — 21.5.3.3.6/21-28 — 21.5.3.3.7/21-29 21.5.3.3.8/21-31 — 0x2_6110– Reserved 0x2_6120 0x2_6124 CTBPTR—Current TxBD pointer register 0x2_6128– Reserved 0x2_6180 0x2_6184 TBPTR—TxBD pointer register 0x2_6188– Reserved 0x2_6200 0x2_6204 TBASE—TxBD base address register 0x2_6208– Reserved 0x2_62AC 0x2_62B0 0x2_62B4 OSTBD—Out-of-sequence TxBD register OSTBDP—Out-of-sequence TxBD pointer register 0x2_62B8– Reserved 0x2_62FC FEC Receive Control and Status Registers 0x2_6300 0x2_6304 0x2_6308 0x2_630C RCTRL—Receive control register RSTAT—Receive status register Reserved RBDLEN—RxBD data length register R/W R/W R R R R 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 21.5.3.4.1/21-31 21.5.3.4.2/21-32 — 21.5.3.4.3/21-33 — 21.5.3.4.4/21-33 0x2_6310– Reserved 0x2_6320 0x2_6324 CRBPTR—Current RxBD pointer register MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 21-11 10/100 Fast Ethernet Controller Table 21-3. Module Memory Map (continued) Offset 0x2_6328– Reserved 0x2_633C 0x2_6340 MRBLR—Maximum receive buffer length register Name Access 1 R R/W R R/W R R/W R Reset 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 Section/Page — 21.5.3.4.5/21-34 — 21.5.3.4.6/21-35 — 21.5.3.4.7/21-35 — 0x2_6344– Reserved 0x2_6380 0x2_6384 RBPTR—RxBD pointer register 0x2_6388– Reserved 0x2_6400 0x2_6404 RBASE—RxBD base address register 0x2_6408– Reserved 0x2_64FC FEC MAC Registers 0x2_6500 0x2_6504 0x2_6508 0x2_650C 0x2_6510 MACCFG1—MAC configuration register 1 MACCFG2—MAC configuration register 2 IPGIFG—Inter-packet gap/inter-frame gap register HAFDUP—Half-duplex register MAXFRM—Maximum frame length register R/W R/W R/W R/W R/W R R R/W R/W R 0x0000_0000 0x0000_7000 0x4060_5060 0x00A0_F037 0x0000_0600 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 21.5.3.6.1/21-38 21.5.3.6.2/21-40 21.5.3.6.3/21-41 21.5.3.6.4/21-41 21.5.3.6.5/21-42 — 21.5.3.6.6/21-43 21.5.3.6.7/21-43 21.5.3.6.8/21-44 — 0x2_6514– Reserved 0x2_6538 0x2_653C 0x2_6540 0x2_6544 IFSTAT—Interface status register MACSTNADDR1—Station address register, part 1 MACSTNADDR2—Station address register, part 2 0x2_6548– Reserved 0x2_67FC FEC Hash Function Registers 0x2_6800 0x2_6804 0x2_6808 0x2_680C 0x2_6810 0x2_6814 0x2_6818 0x2_681C IADDR0—Individual address register 0 IADDR1—Individual address register 1 IADDR2—Individual address register 2 IADDR3—Individual address register 3 IADDR4—Individual address register 4 IADDR5—Individual address register 5 IADDR6—Individual address register 6 IADDR7—Individual address register 7 R/W R/W R/W R/W R/W R/W R/W R/W R 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 21.5.3.7.1/21-45 0x2_6820– Reserved 0x2_687C — MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 21-12 Freescale Semiconductor Memory Map/Register Definition Table 21-3. Module Memory Map (continued) Offset 0x2_6880 0x2_6884 0x2_6888 0x2_688C 0x2_6890 0x2_6894 0x2_6898 0x2_689C Name GADDR0—Group address register 0 GADDR1—Group address register 1 GADDR2—Group address register 2 GADDR3—Group address register 3 GADDR4—Group address register 4 GADDR5—Group address register 5 GADDR6—Group address register 6 GADDR7—Group address register 7 Access 1 R/W R/W R/W R/W R/W R/W R/W R/W R Reset 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 0x0000_0000 — Section/Page 21.5.3.7.2/21-46 0x2_68A0– Reserved 0x2_6AFF FEC Attribute Registers 0x2_6B00– Reserved 0x2_6BF4 0x2_6BF8 0x2_6BFC ATTR—Attribute register ATTRELI—Attribute EL & EI register R R/W R/W 0x0000_0000 0x0000_0000 0x0000_0000 — 21.5.3.8.1/21-46 21.5.3.8.2/21-47 FEC Future Expansion Space 0x2_6C00– Reserved 0x2_6FFF 1 R 0x0000_0000 — R = Read Only, W = Write Only, R/W = Read and Write, LH = Latches High, SC = Self-clearing. 21.5.3 Memory-Mapped Register Descriptions This section provides a detailed description of the FEC registers. Because all of these registers are 32 bits wide, only 32-bit register accesses are supported. 21.5.3.1 FEC General Control and Status Registers This section describes general control and status registers used for both transmitting and receiving Ethernet frames. All of the registers are 32 bits wide. 21.5.3.1.1 Interrupt Event Register (IEVENT) If an event occurs that sets a bit in the interrupt event (IEVENT) register, shown in Figure 21-5, an interrupt is generated if the corresponding bit in the interrupt enable register (IMASK) is also set. The bit in the interrupt event register is cleared if a 1 is written to that bit position. A write of 0 has no effect. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 21-13 10/100 Fast Ethernet Controller These interrupts can be divided into operational interrupts, transceiver/network error interrupts, and internal error interrupts. Interrupts that may occur in normal operation are: • • • GTSC, GRSC, TXF, TXB, TXC, RXF, RXB, RXC, and MSRO BABR, BABT, LC, and CRL/XDA EBERR, XFUN, and BSY 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Interrupts resulting from errors/problems detected in the network or transceiver are: Interrupts resulting from internal errors are: R BABR RXC W Reset 16 17 0 BSY EBERR MSRO GTSC BABT TXC TXE TXB TXF 0 LC CRL/ XFUN XDA 0000_0000_0000_0000 22 23 24 25 31 R RXB W Reset Offset 0 0 0 0 0 0 GRSC RXF 0000_0000_0000_0000 0x2_6010 0 0 0 0 0 0 0 Figure 21-5. IEVENT Register Definition Table 21-4 describes the fields of the IEVENT register. Table 21-4. IEVENT Field Descriptions Bits 0 Name BABR Description Babbling receive error. This bit indicates that a frame was received with length in excess of the MAC’s maximum frame length register while MACCFG2[Huge Frame] is set. 0 Excessive frame not received 1 Excessive frame received Receive control interrupt. A control frame was received. If MACCFG1[Rx_Flow] is set, a pause operation is performed lasting for the duration specified in the received pause control frame and beginning when the frame was received. 0 Control frame not received 1 Control frame received Busy condition interrupt. Indicates that a frame was received and discarded due to a lack of buffers. When IEVENT[BSY] is set RSTAT[QHLT] is also set. In order to begin receiving packets again, the user must clear RSTAT[QHLT]. This bit and RSTAT[QHLT] are set whenever the FEC reads an RxBD with its Empty field cleared. 0 No frame received and discarded. 1 Frame received and discarded. 1 RXC 2 BSY MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 21-14 Freescale Semiconductor Memory Map/Register Definition Table 21-4. IEVENT Field Descriptions (continued) Bits 3 Name Description EBERR Ethernet bus error. This bit indicates that a system bus error occurred while a DMA transaction was underway. If the EBERR is set while transmission is in progress, the DMA stops sending data to the Tx FIFO which eventually causes an underrun error (XFUN) and TSTAT[THLT] is set. If the EBERR is set while receiving a frame, the DMA discards the frame and RSTAT[QHLT] is set. 0 No system bus error occurred. 1 System bus error occurred. — MSRO Reserved MSTAT register overflow. This interrupt is asserted if the count for one of the MSTAT registers has exceeded the size of the register. 0 MSTAT count not exceeding register size. 1 MSTAT count exceeds register size. Graceful transmit stop complete. This interrupt is asserted following the completion of a graceful stop, which was intitiated by setting either DMACTRL[GTS] or TCTRL[TFC_PAUSE]. Graceful stop means that the transmitter is put into a pause state after completion of the frame currently being transmitted. 0 No graceful stop interrupt. 1 Graceful stop requested. Babbling transmit error. This bit indicates that the transmitted frame length has exceeded the value in the MAC’s Maximum Frame Length register and MACCFG2[Huge Frame] is cleared. Frame truncation occurs when this condition occurs. TxBD[TXTRUNC] is set in the last TxBD (TxBD[L] is set) of the frame. 0 Transmitted frame length not exceeding maximum frame length. 1 Transmitted frame length exceeding maximum frame length. Transmit control interrupt. This bit indicates that a control frame was transmitted. 0 Control frame not transmitted. 1 Control frame transmitted. Transmit error. Indicates that an error occurred on the transmitted channel that caused TSTAT[THLT] to be set by the FEC. This bit is set whenever any transmit error occurs that causes the transmitter to halt (EBERR, LC, CRL/XDA, XFUN). It is not set if DMACTRL[WOP] is set and the FEC runs out of TxBDs to process. In order to begin transmitting packets again, the user must clear TSTAT[THLT]. 0 No transmit channel error occurred. 1 Transmit channel error occurred. Transmit buffer. Indicates that a transmit buffer descriptor was updated whose I (Interrupt) bit was set in its status word and was not the last buffer descriptor of the frame. 0 No transmit buffer descriptor updated. 1 Transmit buffer descriptor updated. Transmit frame interrupt. Indicates that a frame was transmitted and that the last corresponding transmit buffer descriptor (TxBD) was updated. This only occurs if the I (Interrupt) bit in the status word of the buffer descriptor is set. 0 No frame transmitted/TxBD not updated. 1 Frame transmitted/TxBD updated. Reserved Late collision. This bit indicates that a collision occurred beyond the collision window (slot time) in half-duplex mode. The frame is truncated with a bad CRC and the remainder of the frame is discarded. 0 No late collision occurred. 1 Late collision occurred. 4 5 6 GTSC 7 BABT 8 TXC 9 TXE 10 TXB 11 TXF 12 13 — LC MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 21-15 10/100 Fast Ethernet Controller Table 21-4. IEVENT Field Descriptions (continued) Bits 14 Name CRL/ XDA Description Collision retry limit/excessive defer abort. Indicates either one of two conditions occurred while attempting to transmit a frame: 1) the number of successive transmission collisions has exceeded the MAC’s HAFDUP[Retransmission Maximum] count or 2) an excessive defer abort condition has occurred. An excessive defer abort condition occurs when the TSEC waits more than 3036 bytes while attempting to send a frame and HAFDUP[EXCESS_DEFER] is 0. The TxBD[DEF] or OSTBD[DEF] is also set. In either case the frame is discarded without being transmitted and the TSEC is halted (TSTAT[THLT] is set). The CRL or XDA condition can only occur while in half-duplex mode. 0 Successive transmission collisions do not exceed maximum and no excessive defer abort condition has occurred. 1 Successive transmission collisions exceed maximum or an excessive defer abort condition has occurred. Transmit FIFO underrun. Indicates that the transmit FIFO became empty before the complete frame was transmitted. 0 Transmit FIFO not underrun. 1 Transmit FIFO underrun. Receive buffer. Indicates that a receive buffer descriptor was updated which had the I (Interrupt) bit set in its status word and was not the last buffer descriptor of the frame. 0 Receive buffer descriptor not updated. 1 Receiver buffer descriptor updated. Reserved Graceful receive stop complete. This interrupt is asserted if a graceful receive stop is completed. It allows the user to know if the system has completed the stop and it is safe to write to receive registers (status, control or configuration registers) that are used by the system during normal operation. 0 Graceful stop not completed. 1 Graceful stop completed. Receive frame interrupt. The last receive buffer descriptor (RxBD) of a frame was updated. This occurs only if the I (interrupt) bit in the buffer descriptor status word is set. 0 Frame not received. 1 Frame received. Reserved 15 XFUN 16 RXB 17–22 23 — GRSC 24 RXF 25–31 — 21.5.3.1.2 Interrupt Mask Register (IMASK) The interrupt mask register, shown in Figure 21-6, provides control over which possible interrupt events are allowed to generate an actual interrupt. All implemented bits in this CSR are R/W. This register is cleared upon a hardware reset. If the corresponding bits in both the IEVENT and IMASK registers are set, an interrupt is generated. The interrupt signal can be cleared by clearing the corresponding IEVENT bit. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 21-16 Freescale Semiconductor Memory Map/Register Definition 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 R BREN RXCEN BSYEN EBERREN 0 MSROEN GTSCEN BTEN TXCEN TXEEN TXBEN TXFEN 0 LCEN W Reset 16 17 CRL/ XFUNEN XDAEN 0000_0000_0000_0000 22 23 24 25 31 R RXBEN 0 0 0 0 0 0 GRSCEN RXFEN 0 0 0 0 0 0 0 W Reset Offset 0000_0000_0000_0000 0x2_6014 Figure 21-6. IMASK Register Definition Table 21-5 describes the fields of the IMASK register. Table 21-5. IMASK Field Descriptions Bits 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17–22 23 24 25–31 Name BREN RXCEN BSYEN EBERREN — MSROEN GTSCEN BTEN TXCEN TXEEN TXBEN TXFEN — LCEN CRL/XDAEN XFUNEN RXBEN — GRSCEN RXFEN — Description Babbling receiver interrupt enable Receive control interrupt enable Busy interrupt enable Ethernet controller bus error enable Reserved MSTAT register overflow interrupt enable Graceful transmit stop complete interrupt enable Babbling transmitter interrupt enable Transmit control interrupt enable Transmit error interrupt enable Transmit buffer interrupt enable Transmit frame interrupt enable Reserved Late collision enable Collision retry limit / excessive defer abort enable Transmit FIFO underrun enable Receive buffer interrupt enable Reserved Graceful receive stop complete interrupt enable Receive frame interrupt enable Reserved MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 21-17 10/100 Fast Ethernet Controller 21.5.3.1.3 Error Disabled Register (EDIS) The error disabled register allows the user to control the reporting of errors by the FEC. The IEVENT bit corresponding to an error is not set if the error is disabled in the EDIS register. Figure 21-7 shows the EDIS register. 0 1 2 3 4 12 13 14 15 R W Reset 0 0 BSYDIS EBERRDIS 0 0 0 0 0 0 0 0 0 LCDIS 0 0 0000_0000_0000_0000 16 31 R W Reset Offset 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0000_0000_0000_0000 0x2_6018 Figure 21-7. EDIS Register Definition Table 21-6 describes the fields of the EDIS register. Table 21-6. EDIS Field Descriptions Bits 0–1 2 3 4–12 13 14–31 Name — BSYDIS EBERRDIS — LCDIS — Reserved Busy disabled Ethernet controller bus error disabled Reserved Late collision disabled Reserved Description 21.5.3.1.4 Minimum Frame Length Register (MINFLR) MINFLR, shown in Figure 21-8, is written by the user and tells the FEC the smallest packet to accept and place in a receive buffer pointed to by the RxBD. 0 24 25 31 R0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 MINFLR W Reset Offset 0000_0000_0000_0000_0000_0000_0100_0000 0x2_6024 Figure 21-8. MINFLR Register Definition MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 21-18 Freescale Semiconductor Memory Map/Register Definition Table 21-7 describes the fields of the MINFLR register. Table 21-7. MINFLR Field Descriptions Bits 0–24 Name — Reserved Description 25–31 MINFLR Minimum receive frame length (typically 64 decimal, which is the maximum). Ethernet discards incoming frames shorter than MINFLR unless RCTRL[RSF] (receive short frames) is set, in which case RxBD[SH] (frame too short) is set in the last RxBD. Unlike the MPC8260, in which pads are added to make the transmit frame equal to MINFLR bytes, if padding is requested, the FEC always pads transmit frames to 64 bytes ignoring MINFLR. MINFLR is only used to determine the minimum size of acceptable receive frames. 21.5.3.1.5 Pause Time Value Register (PTV) PTV, shown in Figure 21-9, is written by the user to store the pause duration used when the FEC initiates a pause frame through TCTRL[TFC_PAUSE]. The low-order 16 bits (PT) represent the pause time and the high-order 16 bits (PTE) represent the extended pause control parameter. The pause time is measured in units of pause_quanta, equal to 512 bit times. The pause time can range from 0 to 65,535 pause_quanta, or 0 to 33,553,920 bit times. See Section 21.6.2.6, “Flow Control,” for additional details. 0 15 16 31 R PTE W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x2_6028 PT Figure 21-9. PTV Register Definition MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 21-19 10/100 Fast Ethernet Controller Table 21-8 describes the fields of the PTV register. Table 21-8. PTV Field Descriptions Bits 0–15 Name PTE Description Extended pause control. This field allows software to add a 16-bit additional control parameter into the pause frame to be sent when TCTRL[TFC_PAUSE] is set. Note that current IEEE 802.3 pause frame format requires this parameter to be set to 0. Pause time value. Represents the 16-bit pause quanta (that is, 512 bit times). This pause value is used as part of the pause frame to be sent when TCTRL[TFC_PAUSE] is set. See Section 21.6.2.6, “Flow Control,” for more information. 16–31 PT 21.5.3.1.6 DMA Control Register (DMACTRL) DMACTRL is writable by the user to configure the DMA block. Figure 21-10 describes the definition for the DMACTRL register. 0 23 24 25 26 27 28 29 30 31 R000000000000000000000000 TDSEN TBDSEN W Reset Offset 0 GRS GTS 0 WWR WOP 0000_0000_0000_0000_0000_0000_0000_0000 0x2_602C Figure 21-10. DMACTRL Register Definition Table 21-9 describes the fields of the DMACTRL register. Table 21-9. DMACTRL Field Descriptions Bits 0–23 24 Name — TDSEN Reserved Tx data snoop enable 0 Disables snooping of all transmit frames from memory. 1 Enables snooping of all transmit frames from memory. Description 25 TBDSEN TxBD snoop enable 0 Disables snooping of all transmit BD memory accesses. 1 Enables snooping of all transmit BD memory accesses. — Reserved 26 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 21-20 Freescale Semiconductor Memory Map/Register Definition Table 21-9. DMACTRL Field Descriptions (continued) Bits 27 Name GRS Description Graceful receive stop. If this bit is set, the Ethernet controller stops receiving frames following completion of the frame currently being received. (That is, after a valid end of frame was received). The buffer of the receive frame associated with the EOF is closed and the IEVENT[GRSC] is set. Because the MAC’s receive enable bit (MACCFG[Rx_EN]) may still be set, the MAC may continue to receive but the FEC ignores the receive data until GRS is cleared. If this bit is cleared, the FEC scans the input data stream for the start of a new frame (preamble sequence and start of frame delimiter) and the first valid frame received uses the next RxBD. If GRS is set, the user must monitor the graceful receive stop complete (GRSC) bit in the IEVENT register to insure that the graceful receive stop was completed. The user can then clear IEVENT[GRSC] and can write to receive registers that are accessible to both user and the FEC hardware without fear of conflict. 0 The FEC scans the input data stream for valid frames. 1 The FEC stops receiving frames following completion of current frame. Graceful transmit stop. If this bit is set, the Ethernet controller stops transmission after the completion of any frame that is currently being transmitted and the GTSC interrupt in the IEVENT register is asserted. If frame transmission is not currently underway, the GTSC interrupt is asserted immediately. Once transmission has completed, a ‘restart’ can be accomplished by clearing GTS. 0 The FEC continues. 1 The FEC stops transmission after completion of current frame. Reserved Write with response. This bit gives the user the assurance that a BD was updated in memory before it receives an interrupt concerning a transmit or receive frame. 0 Do not wait for acknowledgement from system for BD writes before setting IEVENT bits. 1 Before setting IEVENT bits TXB, TXF, TXE, IE, XFUN, LC, CRL/XDA, RXB, RXF, the FEC waits for system acknowledgement that the transmit or receive BD being updated was stored in memory. Wait or poll. This bit, which is applicable only to the transmitter, provides the user the option for the FEC to periodically poll TxBD or to wait for software to tell FEC to fetch a buffer descriptor. While operating in the wait mode, the FEC allows two additional reads of a descriptor which is not ready before entering a halt state. No interrupt is driven (IEVENT[TXE] is clear). To resume transmission, software must clear TSTAT[THLT]. Note that if this bit is set, the user must ensure that all TxBDs involved with sending a frame have their ready bits set before any transmission begins. Otherwise, FEC behavior is boundedly undefined. In wait mode (DMACTRL[WOP] is set) when FEC is processing a frame and an intermediate TxBD’s ready bit is cleared, FEC does not halt, but instead continuously polls the same TxBD until the TxBD becomes ready or an Ethernet interface error or a memory error is encountered. If the TxBD becomes ready (TxBD[Ready] is set), FEC continues to process the frame. If an error occurs before the TxBD becomes ready, FEC reports the error in both the IEVENT register as well as the TxBD for Ethernet interface errors, or the IEVENT[EBERR] for a memory error and sets TSTAT[THLT]. Note: Software must eventually set all of its TxBDs for a frame, because FEC continuously reads an intermediate TxBD until it becomes ready if insufficient data has been read to surpass the FIFO transmit threshold (FIFO_TX_THR) register value. 0 Poll TxBD every 512 clocks. 1 Do not poll, but wait for a write to TSTAT[THLT]. 28 GTS 29 30 — WWR 31 WOP 21.5.3.2 FEC FIFO Control and Status Registers The following registers allow the user to change some of the default settings in the FIFO that can be used to optimize operation for performance or for safety. They must be set carefully in order to MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 21-21 10/100 Fast Ethernet Controller avoid an underrun condition. Underrun is an error condition in which data is not retrieved from external memory quickly enough, leaving the TX FIFO empty before the complete frame is transmitted. Because different combinations of events, several of which are determined by the user, can lead to underrun, the FEC provides several FIFO registers that allow the user to select the proper setting to be able to tune the system and obtain the maximum performance with minimal chance of underrun. The principal causes for underrun in the FEC are: • • • Misaligned data buffer addresses Small data buffer sizes Combinations of the above It is recommended that the minimum size data buffers be 64 bytes and that data buffers be 64-byte aligned. The user can deviate from these recommended values to try to increase performance or to use less memory, but unless the default values of some of the FIFO registers are adjusted, the probability of an underrun may also increase. The FIFO_TX_THR (default is 512 entries) indicates the amount of data required to be in the FIFO before starting the transmission of a frame. The FIFO_TX_STARVE (default is 128 entries) is used to indicate that the amount of data in the FIFO is so low that the risk of underrun is extremely high. The FIFO_TX_STARVE_SHUTOFF (default is 512 entries) contains the watermark level to be used for exiting the starve state. These registers are intended to allow the user to make the proper trade-off. If triggered, the starve mode, for instance, automatically raises the priority of FEC fetches from memory. 21.5.3.2.1 FIFO Pause Control Register (FIFO_PAUSE_CTRL) FIFO_PAUSE_CTRL, shown in Figure 21-11, is writable by the user to configure the properties of the FEC FIFO. 0 15 R W Reset 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0000_0000_0000_0000 16 29 30 31 R W Reset Offset 0 0 0 0 0 0 0 0 0 0 0 0 0 0 TFC_PAUSE_EN 0 0000_0000_0000_0000 0x2_604C Figure 21-11. FIFO_PAUSE_CTRL Register Definition Table 21-10 describes the fields of the FIFO_PAUSE_CTRL register. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 21-22 Freescale Semiconductor Memory Map/Register Definition Table 21-10. FIFO_PAUSE_CTRL Field Descriptions Bits 0–29 30 Name — TFC_PAUSE_EN Reserved. TFC_PAUSE enable. This bit enables the ability to transmit a pause control frame by setting the TCTRL[TFC_PAUSE] bit. This bit is cleared at reset. 0 Pause control frame transmission disabled. 1 Pause control frame transmission enabled. Reserved. Description 31 — 21.5.3.2.2 FIFO Transmit Threshold Register (FIFO_TX_THR) The purpose of the threshold register is to trigger the unloading of FIFO data to the PHY. It represents the numerical SRAM entry (0-255 for a 1-Kbyte FIFO) to trigger the threshold function. If the used entry count is equal to or greater than the threshold register, then transmission can begin. This register is read/write by software. This register is initialized to 0000_0000_0000_0000_0000_0000_1000_0000 at system reset. Figure 21-12 describes the definition for the FIFO_TX_THR register. 0 15 16 23 24 31 R0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 FIFO_TX_THR W Reset Offset 0000_0000_0000_0000_0000_0000_1000_0000 0x2_608C Figure 21-12. FIFO_TX_THR Register Definition MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 21-23 10/100 Fast Ethernet Controller Table 21-11 describes the fields of the FIFO_TX_THR register. Table 21-11. FIFO_TX_THR Field Descriptions Bits 0–23 Name — Reserved Description 24–31 FIFO_TX_THR FIFO transmit threshold. Marks the number of entries in the transmit FIFO that when reached triggers the unloading of frame data into the MAC. If the FIFO Tx used entry count is equal to or greater than FIFO_TX_THR, frame transmission request is sent to the MAC. 21.5.3.2.3 FIFO Transmit Starve Register (FIFO_TX_STARVE) The purpose of the starve register is to inform the system of extremely imminent underrun conditions. It represents the numerical SRAM entry (0–255 for 1-Kbyte FIFO) to trigger the starve function. If the used entry count in the FIFO is less than or equal to the starve register, a starve alert is triggered. This triggered condition is used to throttle back the OCN2CU interface at the e500 coherency module (ECM), to change the Tx transaction priority. This register is read/write by software. Figure 21-13 describes the definition for the FIFO_TX_STARVE register. 0 23 24 31 R0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 FIFO_TX_STARVE W Reset Offset 0000_0000_0000_0000_0000_0000_0010_0000 0x2_6098 Figure 21-13. FIFO_TX_STARVE Register Definition Table 21-12 describes the fields of the FIFO_TX_STARVE register. Table 21-12. FIFO_TX_STARVE Field Descriptions Bits 0–23 24–31 Name — FIFO_TX_STARVE Reserved FIFO transmit starve. Indicates the value to trigger the transmit starve function. It triggers once the FIFO transmit used entry count is less than or equal to the FIFO Tx starve. The starve state turns off when the FIFO Tx used entry count becomes greater than or equal to the FIFO Tx starve shutoff register. Description 21.5.3.2.4 FIFO Transmit Starve Shutoff Register (FIFO_TX_STARVE_SHUTOFF) The starve shutoff registers contains the watermark level to be used for coming out of the starve state. If the starve state is in effect and the number of valid entries in the FIFO becomes greater than or equal to the value in the FIFO_TX_STARVE_SHUTOFF register, the starve condition MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 21-24 Freescale Semiconductor Memory Map/Register Definition ends. This register is read/write by software. Figure 21-14 describes the definition for the FIFO_TX_STARVE_SHUTOFF register. 0 23 24 31 R00000000000000000000000 0 FIFO_TX_STARVE_SHUTOFF W Reset Offset 0000_0000_0000_0000_0000_0000_1000_0000 0x2_609C Figure 21-14. FIFO_TX_STARVE_SHUTOFF Register Definition Table 21-13 describes the fields of the FIFO_TX_STARVE_SHUTOFF register. Table 21-13. FIFO_TX_STARVE_SHUTOFF Field Descriptions Bits 0–23 24–31 Name — FIFO_TX_STARVE_SHUTOFF Reserved FIFO transmit starve shutoff. Indicates the value to use to turn off the starve state. The starve state turns off when the FIFO Rx used entry count becomes greater than or equal to the FIFO Rx starve shutoff register. Description 21.5.3.3 FEC Transmit Control and Status Registers This section describes the control and status registers that are used specifically for transmitting Ethernet frames. All of the registers are 32 bits wide. 21.5.3.3.1 Transmit Control Register (TCTRL) This register is writable by the user to configure the transmit block. Figure 21-15 describes the definition for the TCTRL register. 0 19 20 21 26 27 28 29 31 R000000000000000 0 0 00 0 THDF W Reset Offset 0 0 0 0 0 0 RFC_PAUSE TFC_PAUSE 000 0000_0000_0000_0000_0000_0000_0000_0000 0x2_6100 Figure 21-15. TCTRL Register Definition MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 21-25 10/100 Fast Ethernet Controller Table 21-14 describes the fields of the TCTRL register. Table 21-14. TCTRL Field Descriptions Bits 0–19 20 Name — THDF Reserved Transmit half-duplex flow control. Written by user. This bit is not self-resetting. 0 Disable back pressure. 1 Back pressure is applied to media. Reserved Description 21–26 27 — RFC_PAUSE Receive flow control pause frame. Written by the FEC. This read-only status bit is set if a flow control pause frame was received and the transmitter is paused for the duration defined in the received pause frame. This bit automatically clears after the pause duration is complete. 0 Pause duration complete. 1 Pause duration in progress. TFC_PAUSE Transmit flow control pause frame. Use this bit to transmit a PAUSE frame. To transmit a flow control pause frame, first set FIFO_PAUSE_CTRL[TFC_PAUSE_EN]. Next, set MACCFG1[GTS]. If TFC_PAUSE is then set, the MAC stops transmission of data frames after the current transmission completes. The GTSC interrupt in the IEVENT register is asserted. With transmission of data frames stopped, the MAC transmits a MAC control PAUSE frame with the duration value obtained from the PTV register. The TXC interrupt occurs after sending the control pause frame. Next, the MAC clears TFC_PAUSE and resumes transmitting data frames. Note that if the transmitter is paused due to user assertion of GTS or reception of a PAUSE frame, the MAC may still transmit a MAC control PAUSE frame. 0 No outstanding pause frame transmission request. 1 Pause frame transmission requested. — Reserved 28 29–31 21.5.3.3.2 Transmit Status Register (TSTAT) This register is written by the FEC to convey DMA status information. Figure 21-16 describes the definition for the TSTAT register. 0 1 31 R THLT W Reset: Offset 0000000000000000000000000000000 0000_0000_0000_0000_0000_0000_0000_0000 0x2_6104 Figure 21-16. TSTAT Register Definition MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 21-26 Freescale Semiconductor Memory Map/Register Definition Table 21-15 describes the fields of the TSTAT register. Table 21-15. TSTAT Field Descriptions Bit 0 Name THLT Description Transmit halt 0 No hardware initiated transmission halt. 1 FEC transmission function halted. Set by the FEC to inform the user that it is no longer processing transmit frames and the transmit DMA function is disabled by hardware. To re-start the transmission function, the user must clear this bit by writing a one. Reserved 1–31 — 21.5.3.3.3 TxBD Data Length Register (TBDLEN) TBDLEN is a DMA register that contains the number of bytes remaining in the current transmit buffer. Figure 21-17 describes the definition for the TBDLEN register. 0 15 16 31 R0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 W Reset Offset TBDLEN 0000_0000_0000_0000_0000_0000_0000_0000 0x2_610C Figure 21-17. TBDLEN Register Definition Table 21-16 describes the fields of the TBDLEN register. Table 21-16. TBDLEN Field Descriptions Bits 0–15 Name — Reserved Description 16–31 TBDLEN Internally written by the DMA module. The transmit channel remains active until TBDLEN is 0. 21.5.3.3.4 Current Transmit Buffer Descriptor Pointer Register (CTBPTR) CTBPTR, shown in Figure 21-18, contains the address of the transmit buffer descriptor either currently being processed, or processed most recently. 0 28 29 31 R W Reset Offset CTBPTR 00 0 0000_0000_0000_0000_0000_0000_0000_0000 0x2_6124 Figure 21-18. CTBPTR Register Definition MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 21-27 10/100 Fast Ethernet Controller Table 21-17 describes the fields of the CTBPTR register. Table 21-17. CTBPTR Field Descriptions Bits 0–28 29–31 Name Description CTBPTR Internally written by the DMA module. The value increments by eight (bytes) each time a descriptor is read from memory. — Reserved 21.5.3.3.5 Transmit Buffer Descriptor Pointer Register (TBPTR) TBPTR contains the low-order 32 bits of the next transmit buffer descriptor address. Figure 21-19 describes the definition for the TBPTR register. This register takes on the value of TBASE when the TBASE register is written by software. 0 28 29 31 R TBPTR W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x2_6184 000 Figure 21-19. TBPTR Register Definition Table 21-18 describes the fields of the TBPTR register. Table 21-18. TBPTR Field Descriptions Bit 0–28 29–31 Name TBPTR — Description Internally written by the DMA module. The value increments by eight (bytes) each time a descriptor is read from memory. Reserved 21.5.3.3.6 Transmit Descriptor Base Address Register (TBASE) The TBASE register is written by the user with the TxDB base address. The value must be divisible by eight for the 8-byte data buffer descriptors. Figure 21-20 describes the definition for the TBASE register. 0 28 29 31 R TBASE W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x2_6204 000 Figure 21-20. TBASE Register Definition MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 21-28 Freescale Semiconductor Memory Map/Register Definition Table 21-19 describes the fields of the TBASE register. Table 21-19. TBASE Field Descriptions Bits 0–28 Name TBASE Description Transmit base. Defines the starting location in the memory map for the FEC TxBDs. This field must be 8-byte aligned. Together with setting the W (wrap) bit in the last BD, the user can select how many BDs to allocate for the transmit packets. The user must initialize TBASE before enabling the FEC transmit function. Reserved 29–31 — 21.5.3.3.7 Out-of-Sequence TxBD Register (OSTBD) The out-of-sequence TxBD register (OSTBD), shown in Figure 21-21, includes the status/control and data length in the same format as a regular TxBD. It is useful for sending flow control frames. The OSTBD[R] is always checked between frames. If it is not ready, a regular frame is sent. If a flow control frame is sent and OSTBD[I] is set, a TXC event is generated after frame transmission. This area must be cleared while not in use. Once the TSEC is in paused mode the out-of-sequence buffer descriptor cannot be used to send another flow control frame because the MAC regards it as a regular TxBD. 0 1 2 34 5 6 7 8 9 10 13 14 15 16 31 R R PAD/CRC W I W Reset Offset L TC DEF TO1 HFE/LC RL RC UN 0 OSTBDLEN 0000_0000_0000_0000_0000_0000_0000_0000 0x2_62B0 Figure 21-21. OSTBD Register Definition Table 21-20 describes the fields of the OSTBD register. Table 21-20. OSTBD Field Descriptions Bits 0 Name R Description Ready. Written by both the FEC and the user. 0 The data buffer associated with this BD is not ready for transmission. The user is free to manipulate this BD or its associated data buffer. The FEC clears this bit after the buffer was transmitted or after an error condition is encountered. 1 The data buffer, prepared for transmission by the user, was not transmitted or is currently being transmitted. No fields of this BD may be written by the user once this bit is set. Padding and CRC attachment for short frames. (Valid only when MACCFG2[PAD/CRC] is cleared, and MACCFG2[CRC EN] is cleared.) If MACCFG2[PAD/CRC] is set, pads are added to all short frames, however, this bit is ignored. 0 Do not add pads to short frames unless OSTBD[TC] is set. 1 Add pads to short frames. Pad bytes are inserted until the length of the transmitted frame equals 64 bytes. Unlike the MPC8260 which pads up to MINFLR value, FEC always pads up to the IEEE minimum frame length of 64 bytes. 1 PAD/CRC MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 21-29 10/100 Fast Ethernet Controller Table 21-20. OSTBD Field Descriptions (continued) Bits 2 3 Name W I Description Wrap. Written by user. This bit is ignored by FEC. Interrupt. Written by user. 0 No interrupt is generated after this buffer is serviced. 1 IEVENT[TXF] is set after this buffer is serviced. Setting this bit can cause an interrupt if IMASK[TXFEN] is enabled. Last in frame. (The OSTBD is always the last in the frame, so L is always set). Hardwired to a value of 1. Tx CRC. Written by user. (Valid only while it is set in the first BD and OSTBD[PAD/CRC] is cleared, MACCFG2[PAD/CRC] is cleared, and MACCFG2[CRC EN] is cleared.) If MACCFG2[PAD/CRC] is set or MACCFG2[CRC EN] is set, a CRC is added to all frames and this bit is ignored. 0 End transmission immediately after the last data byte, unless TxBD[PAD/CRC] is set. 1 Transmit the CRC sequence after the last data byte. Defer indication. Written by the FEC. Hardware updates this bit after transmitting a frame when used as a ‘defer indicator.’ Software/user updates this bit while building a transmit buffer descriptor if used as a ‘hardware event indicator.’ 0 This frame was not deferred. 1 This frame did not have a collision before it was sent but it was sent late because of deferring. Transmit software ownership. This read/write bit may be utilized by software, as necessary. Its state does not affect the hardware nor is it affected by the hardware. Huge frame enable (written by user)/Late collision (written by TSEC) Huge frame enable. Written by user. Valid only while it is set in first BD and the MACCFG2[Huge Frame] is cleared. If MACCFG2[Huge Frame] is set, this bit is ignored. 0 Truncate transmit frame if its length is greater than the MAC’s Maximum Frame Length register. 1 Do not truncate the transmit frame. Late collision. Written by TSEC. 0 No late collision. 1 A collision occurred after 64 bytes are sent. The TSEC terminates the transmission and updates LC. 9 RL Retransmission limit. Written by the FEC. 0 Transmission before maximum retry limit is hit. 1 The transmitter failed (maximum retry limit + 1 attempts to successfully send a message due to repeated collisions). The FEC terminates the transmission and updates RL. Retry count. Written by the FEC. 0 The frame is sent correctly the first time. 1 More than zero attempts where needed to send the transmit frame. When this field is 15, 15 or more retries were needed. The Ethernet controller updates RC after sending the buffer. Underrun. Written by the FEC. 0 No underrun encountered (data was retrieved from external memory in time to send a complete frame). 1 The Ethernet controller encountered a transmitter underrun condition while sending the associated buffer. The FEC terminates the transmission and updates UN. Reserved 4 5 L TC 6 DEF 7 8 TO1 HFE/LC 10–13 RC 14 UN 15 — 16–31 OSTBDLEN Out-of-sequence TxBD data length. Written by user. Data length is the number of octets the FEC transmits from this BD’s data buffer. It is never modified by the FEC. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 21-30 Freescale Semiconductor Memory Map/Register Definition 21.5.3.3.8 Out-of-Sequence Tx Data Buffer Pointer Register (OSTBDP) The out-of-sequence Tx data buffer pointer register (OSTBDP), shown in Figure 21-22, is written by the user and contains the data buffer pointer fields in the same format as a regular TxBD and, together with OSTBD, provides the complete 8-byte descriptor. This area must be cleared while not in use. 0 15 R OSTBDP W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x2_62B4 Figure 21-22. OSTBDP Register Definition Table 21-21 describes the field of the OSTBDP register. Table 21-21. OSTBDP Field Descriptions Bita 0–31 Name OSTDBP Description Out-of-sequence Tx data buffer pointer. Written by user. The Tx data buffer pointer contains the address of the associated data buffer. There are no alignment requirements for this address. 21.5.3.4 FEC Receive Control and Status Registers This section describes the control and status registers that are used specifically for receiving Ethernet frames. All of the registers are 32 bits wide. 21.5.3.4.1 Receive Control Register (RCTRL) The RCTRL, shown in Figure 21-23, controls the operational mode of the receive block. It must be written only after a system reset (at initialization) or if DMACTRL[GRS] is zero. 0 15 16 26 27 28 29 30 31 R000000000000000 0 0 00 0 W Reset Offset 0 0000 0 0 BC_REJ PROM RSF 0 0 0000_0000_0000_0000_0000_0000_0000_0000 0x2_6300 Figure 21-23. RCTRL Register Definition MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 21-31 10/100 Fast Ethernet Controller Table 21-22 describes the fields of the RCTRL register. Table 21-22. RCTRL Field Descriptions Bits 0–26 27 Name — BC_REJ Reserved Broadcast frame reject. If this bit is set, frames with DA (destination address) = 0xFFFF_FFFF_FFFF are rejected unless RCTRL[PROM] is set. If both BC_REJ and PROM are set, then frames with broadcast DA are accepted and the M (MISS) bit is set in the receive BD. 0 Broadcast frame reject is disabled. 1 Broadcast frame reject is enabled. Promiscuous mode. When set, all frames except pause frames are accepted. 0 Promiscuous mode is disabled. 1 Promiscuous mode is enabled. Receive short frame mode. When set, enables the reception of frames shorter than MINFLR bytes. Note that in order for short frames to be received when RSF is set, a DA hit must occur. When RSF is cleared, all frames shorter than MINFLR are automatically rejected. 0 Receive short frame mode is disabled. 1 Receive short frame mode is enabled. Reserved Description 28 PROM 29 RSF 30–31 — 21.5.3.4.2 Receive Status Register (RSTAT) The FEC writes to this register under the following conditions: • • The receiver runs out of descriptors. The receiver was halted because an error condition was encountered while receiving a frame. Software must clear the QHLT bit to take the FEC’s receiver function out of a halt state. Figure 21-24 describes the definition for the RSTAT register. 0 7 8 9 31 R0 W Reset Offset 0000000 QHLT 0 0 0 00000000000000000000 0000_0000_0000_0000_0000_0000_0000_0000 0x2_6304 Figure 21-24. RSTAT Register Definition MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 21-32 Freescale Semiconductor Memory Map/Register Definition Table 21-23 describes the fields of the RSTAT register. Table 21-23. RSTAT Field Descriptions Bits 0–7 8 Name — Reserved Description QHLT RxBD queue is halted. When IEVENT[BSY] or IEVENT[EBERR] is set during reception of a packet, RSTAT[QHLT] is also set. In order to begin receiving packets again, the user must clear RSTAT[QHLT]. This bit is set whenever the FEC reads an RxBD with its Empty field cleared or encounters a system bus error while processing an RX packet. It is a hardware-initiated stop indication (DMA_CTRL[GRS] being set by the user does not cause this bit to be set). The current frame and all other frames directed to the halted queue are discarded. A write with a value of 1 re-enables the queue for receiving. 0 RxBD queue is enabled for Ethernet reception. (That is, it is not halted.) 1 All Ethernet controller receive activity to RxBD queue is halted. — Reserved 9–31 21.5.3.4.3 RxBD Data Length Register (RBDLEN) RBDLEN is a DMA register that contains the number of bytes remaining in the current receiver buffer. Figure 21-25 describes the definition for the RBDLEN register. 0 15 16 31 R00 00 00 00 0 00 00 00 0 W Reset Offset RBDLEN 0000_0000_0000_0000_0000_0000_0000_0000 0x2_630C Figure 21-25. RBDLEN Register Definition Table 21-24 describes the fields of the RBDLEN register. Table 21-24. RBDLEN Field Descriptions Bits 0–15 Name — Reserved Description 16–31 RBDLEN The RBDLEN is internally written by the DMA module. If RBDLEN is cleared, all activity in the receive channel stops. 21.5.3.4.4 Current Receive Buffer Descriptor Pointer Register (CRBPTR) CRBPTR, shown in Figure 21-26, contains the address of the receive buffer descriptor either currently being processed, or processed most recently. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 21-33 10/100 Fast Ethernet Controller 0 28 29 31 R W Reset Offset CRBPTR 00 0 0000_0000_0000_0000_0000_0000_0000_0000 0x2_6324 Figure 21-26. CRBPTR Register Definition Table 21-25 describes the fields of the CRBPTR register. \ Table 21-25. CRBPTR Field Descriptions Bits 0–28 29–31 Name CRBPTR — Description The CRBPTR register is internally written by the DMA module. The value increments by eight (bytes) each time a descriptor is read from memory. Reserved 21.5.3.4.5 Maximum Receive Buffer Length Register (MRBLR) The MRBLR, shown in Figure 21-27, is written by the user. It informs the FEC how much space is in the receive buffer pointed to by the RxBD. 0 15 16 25 26 31 R0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 MRBL W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x2_6340 000000 Figure 21-27. MRBLR Register Definition Table 21-26 describes the fields of the MRBLR register. \ Table 21-26. MRBLR Field Descriptions Bits 0–15 16–25 Name — MRBL Description To ensure that MRBL is a multiple of 64 these bits are reserved and must be cleared. Maximum receive buffer length. The number of bytes that the FEC receiver writes to the receive buffer. The MRBL register is written by the user with a multiple of 64 for all modes. The FEC can write fewer bytes to the buffer than the value set in MRBL if a condition such as an error or end-of-frame condition occurs, but it never exceeds the MRBL value; therefore, user-supplied buffers must be at least as large as the MRBL. Note that transmit buffers can have varying lengths by programming TxBD[Data Length], as needed, and are unaffected by the MRBLR value. MRBLR is not intended to be changed dynamically while FEC is operating. Change MRBLR only when the FEC receive function is disabled. To ensure that MRBL is a multiple of 64 these bits are reserved and must be cleared. 26–31 — MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 21-34 Freescale Semiconductor Memory Map/Register Definition 21.5.3.4.6 Receive Buffer Descriptor Pointer Register (RBPTR) RBPTR, shown in Figure 21-28, contains the receive buffer descriptor address. It takes on the value of RBASE when the RBASE register is written by software. 0 28 29 31 R RBPTR W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x2_6384 000 Figure 21-28. Receive Buffer Descriptor Pointer Definition Table 21-27 describes the fields of the RBPTR register. \ Table 21-27. RBPTR Field Descriptions Bits 0–28 29–31 Name RBPTR — Description The RBPTR register is internally written by the DMA module. The value increments by eight (bytes) each time a descriptor is read from memory. Reserved 21.5.3.4.7 Receive Descriptor Base Address Register (RBASE) RBASE, shown in Figure 21-29, is written by the user with the RxBD base address and must be divisible eight for the 8-byte buffer descriptors. 0 28 29 31 R RBASE W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x2_6404 000 Figure 21-29. RBASE Register Definition Table 21-28 describes the fields of the RBASE register. \ Table 21-28. RBASE Field Descriptions Bits 0–28 29–31 Name RBASE — Description Receive base. RBASE defines the starting location in the memory map for the FEC RxBD. This field must be 8-byte aligned. Reserved MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 21-35 10/100 Fast Ethernet Controller 21.5.3.5 MAC Functionality This section describes the MAC registers and provides a brief overview of some of the functionality that can be exercised through the use of these registers, particularly those that provide functionality not explicitly required by the IEEE 802.3 standard. All of the MAC registers are 32 bits wide. 21.5.3.5.1 Configuring the MAC MAC configuration registers 1 and 2 provide a way to configure the MAC as follows: • • Adjusting the preamble length – The length of the preamble can be adjusted from the nominal seven bytes to some other(non-zero) value. Varying pad/CRC combinations – Two different pad/CRC combinations are provided to handle a variety of system requirements. The most simple are frames that already have a valid FCS field. In this case, the CRC is checked and reported via the transmit statistics vector (TSV[51:0]). These options include appending a valid CRC, or padding and then appending a valid CRC, resulting in a minimum frame of 64 octets. In addition to the programmable register set, the pad/CRC behavior can be dynamically adjusted on a per-packet basis. 21.5.3.5.2 Controlling CSMA/CD The half-duplex register (HAFDUP) controls the carrier-sense multiple access/collision detection (CSMA/CD) logic. After a packet is sent, the device begins timing the inter-packet gap (IPG) as programmed in the back-to-back IPG configuration register. The system is then free to begin another frame transfer. In full-duplex mode, the carrier sense (FEC_CRS) and collision (FEC_COL) indications from the PHY are ignored, but in half-duplex mode, the FEC defers to FEC_CRS and, following a carrier event, times the IPG using the non–back-to-back IPG configuration values that include support for the optional two-thirds, one-third FEC_CRS deferral process. This optional IPG mechanism enhances system robustness and ensures fair access to the medium. During the first two-thirds of the IPG, the IPG timer is cleared if FEC_CRS is sensed. During the final one-third of the IPG, FEC_CRS is ignored and the transmission begins when IPG is timed. The two-thirds/one-third ratio is recommended. 21.5.3.5.3 Handling Packet Collisions When transmitting a packet in half-duplex mode, the FEC is sensitive to FEC_COL. If a collision occurs, it aborts the packet and outputs the 32-bit jam sequence. The jam sequence is made up of several bits of the CRC, inverted to guarantee an invalid CRC upon reception. A signal is sent to the system indicating a collision occurred and the start of the frame is needed for retransmission. The FEC then backs-off of the medium for a time determined by the truncated binary exponential MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 21-36 Freescale Semiconductor Memory Map/Register Definition back-off (BEB) algorithm. Following this back-off time, the packet is retried. The back-off time can be skipped if configured via the half-duplex register. However, this is non-standard behavior and its use must be carefully applied. Should any one packet experience excessive collisions, the packet is aborted. The system flushes the frame and move to the next one in line. If the system requests to send a packet while the FEC is deferring to a carrier, the FEC simply waits until the end of the carrier event and the timing of IPG before it honors the request. When packet transmission attempts experience collisions, the FEC outputs the jam sequence and waits some amount of time before retrying the packet. This amount of time is determined by a controlled randomization process called truncated binary exponential back-off. The amount of time is an integer number of slot times. The number of slot times to delay before the nth retransmission attempt is chosen as a uniformly distributed random integer r in the range: 0 ≤ r ≤ 2k, where k = min(n,10). So, after the first collision, FEC backs-off either 0 or 1 slot times. After the fifth collision, the FEC backs-off between 0 and 32 slot times. After the tenth collision, the maximum number of slot times to back-off is 1024. This can be adjusted via the half-duplex register. An alternate truncation point, such as 7 for instance, can be programmed. On average, the MAC is be more aggressive after 7 collisions than other stations on the network. 21.5.3.5.4 Controlling Packet Flow Packet flow can be dealt with in a number of ways with the FEC. A default retransmit attempt limit of 15 can be reduced using the half-duplex register. The slot time or collision window can be used to gate the retry window and possibly reduce the amount of transmit buffering within the system. The slot time for 10/100 Mbps is 512 bit times. Since the slot time begins at the beginning of the packet, the end occurs around the 56th byte of the frame data. Full-duplex flow control is provided for in IEEE 802.3x. Currently the standard does not address flow control in half-duplex environments. Common in the industry, however, is the concept of back pressure. The FEC implements the optional back pressure mechanism using the raise carrier method. When the system receive logic wants to stop the reception of packets in a network-friendly way, transmit half-duplex flow control (THDF) is asserted (TCTRL[THDF]). If the medium is idle, the FEC raises carrier by transmitting preamble. Other stations on the half-duplex network then defer to the carrier. In the event the preamble transmission happens to cause a collision, the FEC ensures the minimum 96-bit presence on the wire, then drops preamble and waits a back-off time depending on the value of the configuration bit back pressure no back-off (half-duplex) [BPNB]. These transmitting-preamble-for-back pressure collisions are not counted. When BPBN is asserted, the FEC waits an inter-packet gap before resuming the transmission of preamble following the collision and does not defer. If deasserted, the FEC adheres to the truncated BEB algorithm that allows packets to be received. This also can be detrimental in that packets can now experience excessive collisions causing them to be dropped in the stations from which they originate. To MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 21-37 10/100 Fast Ethernet Controller reduce the likelihood of lost packets and packets leaking through the back pressure mechanism, BPNB must be set. The FEC drops carrier (ceases transmitting preamble) periodically to avoid excessive defer conditions in other stations on the shared network. If while applying back pressure, the FEC is requested to send a packet, it stops sending preamble, and waits one IPG before sending the packet. BPNB applies for any collision that occurs during the sending of this packet. Collisions for packets while THBP is asserted are counted. The FEC does not defer while attempting to send packets while in back pressure. Again, back pressure is nonstandard, but can reduce the flow of receive packets. 21.5.3.6 MAC Registers This section contains descriptions of the MAC registers. 21.5.3.6.1 MAC Configuration Register 1 (MACCFG1) The MACCFG1, shown in Figure 21-30, is written by the user. 0 1 2 11 12 13 14 15 R Soft_Reset W Reset 16 0 0 0 0 0 0 0 0 0 0 0 Reset_ Reset_ Reset_ Reset_ Rx_MC Tx_MC Rx_Fun Tx_Fun 0000_0000_0000_0000 22 23 24 25 26 27 28 29 30 31 R W Reset Offset 0 0 0 0 0 0 0 Loop_ 0 Back 0 Sync’d_ Sync’d_ Rx_Flow Tx_Flow Rx_EN Rx_EN Tx_EN Tx_EN 0000_0000_0000_0000 0x2_6500 Figure 21-30. MACCFG1 Register Definition Table 21-29 describes the fields of the MACCFG1 register. \ Table 21-29. MACCFG1 Field Descriptions Bits 0 1–11 12 Name Soft_Reset — Reset_Rx_MC Description 0 Default 1 Setting this bit puts all modules within the MAC in reset. Reserved 0 Default 1 Setting this bit puts the receive MAC control block in reset. This block detects control frames and contains the pause timers. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 21-38 Freescale Semiconductor Memory Map/Register Definition Table 21-29. MACCFG1 Field Descriptions (continued) Bits 13 Name Reset_Tx_MC Description 0 Default 1 Setting this bit puts the PETMC transmit MAC control block in reset. This block multiplexes data and control frame transfers. It also responds to XOFF PAUSE control frames. 14 Reset_Rx_Fun 0 Default 1 Setting this bit puts the receive function block in reset. This block performs the receive frame protocol. Reset_Tx_Fun 0 Default 1 Setting this bit puts the transmit function block in reset. This block performs the frame transmission protocol. — Loop_Back — Rx_Flow Tx_Flow Reserved 0 Default. 1 Causes the MAC transmit outputs to be looped back to the MAC receive inputs. Reserved 0 Default. The receive MAC control ignores PAUSE flow control frames. 1 The receive MAC control can detect and act on PAUSE flow control frames. 0 Default. The transmit MAC control cannot send flow control frames. 1 The transmit MAC control can send PAUSE flow control frames when requested by the system. 15 16–22 23 24–25 26 27 28 Sync’d_Rx_EN Receive enable synchronized to the receive stream. (Read only) 0 Frame reception is not enabled. 1 Frame reception is enabled. Rx_EN Receive enable. This bit is cleared by default. If set, prior to clearing this bit, set DMACTRL[GRS] then confirm subsequent occurrence of the graceful receive stop interrupt (IEVENT[GRSC] is set). 0 Default. The MAC cannot receive frames from the PHY. 1 The MAC can receive frames from the PHY. 29 30 Sync’d_Tx_EN Transmit enable synchronized to the transmit stream. (Read only) 0 Frame transmission is not enabled. 1 Frame transmission is enabled. Tx_EN Transmit enable. This bit is cleared by default. If set, prior to clearing this bit, set DMACTRL[GTS] then confirm subsequent occurrence of the graceful transmit stop interrupt (IEVENT[GTSC] is set). 0 Default. Clearing this bit prevents the transmission of frames. 1 Setting this bit allows the MAC to transmit frames from the system. 31 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 21-39 10/100 Fast Ethernet Controller 21.5.3.6.2 MAC Configuration Register 2 (MACCFG2) The MACCFG2, shown in Figure 21-31, is written by the user. 0 15 R W Reset 0 0 0 0 00 0 0 00 0 0 0 0 0 0 0000_0000_0000_0000 16 19 20 21 22 23 24 25 26 27 28 29 30 31 R Preamble_Length W Reset Offset 00 I/F_Mode 00 Huge Frame Length_Check 0111_0000_0000_0000 0x2_6504 0 PAD/CRC CRC EN Full_Duplex Figure 21-31. MACCFG2 Register Definition Table 21-30 describes the fields of the MACCFG2 register. Table 21-30. MACCFG2 Field Descriptions Bits 0–15 16–19 20–21 22–23 Name — Reserved Description Preamble_Length This field determines the length in bytes of the preamble field of the packet. Its default is 0x7. A preamble length of 0 is not supported. — I/F_Mode Reserved Determines the type of interface the MAC is connected to. 00 Reserved 01 Nibble mode (MII) 10 Reserved 11 Reserved Reserved 0 Default. The MAC limits the length of frames at the MAXIMUM FRAME LENGTH value. 1 Frames longer than the MAXIMUM FRAME LENGTH can be transmitted and received. 0 Default. Clear this bit if no length field checking is desired. 1 The MAC checks the frame’s length field to ensure it matches the actual data field length. Reserved 0 Default. Transmitted short frames are not padded. 1 The MAC pads all transmitted short frames and appends a CRC to every frame whether or not padding was required. CRC enable 0 Default. Clear if frames presented to the MAC have a valid length and contain a valid CRC. 1 Causes the MAC to append a CRC on all frames. If the configuration bit PAD/CRC or the per-packet PAD/CRC is set, CRC EN is ignored. 0 Default. Clearing this bit configures the MAC to operate in half-duplex mode only. 1 Configures the MAC to operate in full-duplex mode. 24–25 26 27 — Huge Frame Length_Check 28 29 — PAD/CRC 30 CRC EN 31 Full_Duplex MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 21-40 Freescale Semiconductor Memory Map/Register Definition 21.5.3.6.3 Inter-Packet Gap/Inter-Frame Gap Register (IPGIFG) The IPGIFG, shown in Figure 21-32, is written by the user. 0 1 7 8 9 15 16 23 24 25 31 R0 W Reset Offset 0 Non Back-to-Back Non Back-to-Back Inter-Packet-Gap, Part 1 Inter-Packet-Gap, Part 2 Minimum IFG Enforcement 0 Back-to-Back Inter-Packet-Gap 0100_0000_0110_0000_0101_0000_0110_0000 0x2_6508 Figure 21-32. IPGIFG Register Definition Table 21-31 describes the fields of the IPGIFG register. Table 21-31. IPGIFG Field Descriptions Bits 0 1–7 Name — Reserved Description Non Back-to-Back Optional carrier sense window referenced in IEEE 802.3/4.2.3.2.1 carrier deference. Inter-Packet-Gap, Part 1 If carrier is detected during the timing of IPGR1, the MAC defers to carrier. But if carrier becomes active after IPGR1, the MAC continues timing IPGR2 and transmits, knowingly causing a collision and ensuring fair access to medium. Its range of values is 0x00 to IPGR2. The default, 0x40, follows the two-thirds/one-thirds guideline. — Reserved 8 9–15 16–23 Non Back-to-Back This is a programmable field representing the non-back-to-back inter-packet-gap in Inter-Packet-Gap, Part 2 bits. The default, 0x60, represents the minimum IPG of 96 bits. Minimum IFG Enforcement — Back-to-Back Inter-Packet-Gap Minimum number of bits of IFG to enforce between frames. A frame whose IFG is less than that programmed is dropped. The default setting of 0x50 (80d) represents half of the nominal minimum IFG, which is 160 bits. Reserved IPG between back-to-back packets. This IPG parameter is used exclusively in full-duplex mode and in half-duplex mode when two transmit packets are sent back-to-back. Set this field to the number of bits of IPG desired. The default (minimum) of 0x60 (96d) represents the IPG of 96 bits. 24 25–31 21.5.3.6.4 Half-Duplex Register (HAFDUP) The HAFDUP register, shown in Figure 21-33, is written by the user. 0 7 8 11 12 13 14 15 16 19 20 25 26 31 000000 R 0 0 0 0 0 0 0 0 Alternate BP No No Excess Retransmission BEB Alt BEB BackOff BackOff Defer Maximum W Truncation Reset Offset 0000_0000_1010_0000_1111_0000_0011_0111 0x2_650C Collision Window Figure 21-33. Half-Duplex Register Definition MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 21-41 10/100 Fast Ethernet Controller Table 21-32 describes the fields of the HAFDUP register. Table 21-32. HAFDUP Field Descriptions Bits 0–7 8–11 12 Name — Reserved Description Alternate BEB Used when alternate binary exponential back-off enable is set. The value programmed is Truncation substituted for the Ethernet standard value of ten. Alt BEB Setting this bit configures the Tx MAC to use the alternate binary exponential back-off truncation setting instead of the 802.3 standard tenth collision. The standard specifies that any collision after the tenth uses 210 – 1 as the maximum back-off time. Clearing this bit causes the Tx MAC to follow the standard binary exponential back-off rule. 13 BP No BackOff Setting this bit configures the Tx MAC to immediately retransmit, following a collision, during back pressure operation. Clearing this bit causes the Tx MAC to follow the binary exponential back-off rule. No BackOff Excess Defer Setting this bit configures the Tx MAC to immediately retransmit following a collision. Clearing this bit causes the Tx MAC to follow the binary exponential back-off rule. Setting this bit configures the Tx MAC to allow the transmission of a packet that was excessively deferred. Clearing this bit causes the Tx MAC to abort the transmission of a packet that was excessively deferred. 14 15 16–19 Retransmission Specifies the number of retransmission attempts following a collision before aborting the Maximum packet due to excessive collisions. The standard specifies the attempt limit to be 0xF (15d), which is the default. 20–25 26–31 — Collision Window Reserved The slot time or collision window during which collisions occur in properly configured networks. Because the window starts at the start of transmission, the preamble and SFD are included. Its default of 0x37 (55d) corresponds to the count of frame bytes at the end of the window. 21.5.3.6.5 Maximum Frame Length Register (MAXFRM) The MAXFRM register is written by the user. Figure 21-34 describes the MAXFRM register. 0 15 16 31 R0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Maximum Frame W Reset Offset 0000_0000_0000_0000_0000_0110_0000_0000 0x2_6510 Figure 21-34. Maximum Frame Length Register Definition MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 21-42 Freescale Semiconductor Memory Map/Register Definition Table 21-33 describes the fields of the MAXFRM register. Table 21-33. MAXFRM Descriptions Bit 0–15 16–31 Name — Reserved Description Maximum Frame Sets the maximum frame size in both the transmit and receive directions. (Refer to MACCFG2[Huge Frame].) 21.5.3.6.6 Interface Status Register (IFSTAT) Figure 21-35 shows the IFSTAT register. 0 21 22 23 27 28 29 31 R W Reset Offset 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Excess Defer 0 0 0 0 0 Link Fail 0 00 0000_0000_0000_0000_0000_0000_0000_0000 0x2_653C Figure 21-35. Interface Status Register Definition Table 21-34 describes the fields of the IFSTAT register. Table 21-34. IFSTAT Field Descriptions Bit 0–21 22 Name — Excess Defer Reserved Excessive transmission defer. This bit latches high and is cleared when read. This bit is cleared by default. 0 Normal operation 1 The MAC excessively defers a transmission. Reserved Link Fail. This bit indicates the status of signal detection. 0 The100X module has detected a “signal detect” for longer than 330 mS. 1 The100X module has detected a “signal detect” for less than 330 mS or not at all. Reserved Description 23–27 28 — Link Fail 29–31 — 21.5.3.6.7 Station Address Register Part 1 (MACSTNADDR1) The MACSTNADDR1 register is written by the user. Figure 21-36 describes the MACSTNADDR1 register. The value of the station address written by the user into MACSTNADDR1 and MACSTNADDR2 is byte-reversed from how it would appear in the DA field of a frame in memory. For example, for a station address of 0x12345678ABCD, perform a write to MACSTNADDR1 of 0xCDAB7856, and to MACSTNADDR2 of 0x34120000. When the MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 21-43 10/100 Fast Ethernet Controller user reads MACSTNADDR1, 0xCDAB7856 is returned. A read of MACSTNADDR2 returns a value of 0x34120000. Note that the I/G and U/L bits of the frame’s DA field are located at the lsbs of the 1st octet stored in MACSTNADDR2, where the I/G bit is bit 15, and the U/L bit is bit 14. 0 7 8 15 16 23 24 31 R Station Address, 6th Octet Station Address, 5th Octet Station Address, 4th Octet Station Address, 3rd Octet W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x2_6540 Figure 21-36. Station Address Part 1 Register Definition Table 21-35 describes the fields of the MACSTNADDR1 register. Table 21-35. MACSTNADDR1 Field Descriptions Bit 0–7 Name Description Station Address, 6th Octet This field holds the sixth octet of the station address. The sixth octet (station address bits 40–47) defaults to a value of 0x00. 8–15 Station Address, 5th Octet This field holds the fifth octet of the station address. The fifth octet (station address bits 32–39) defaults to a value of 0x00. 16–23 Station Address, 4th Octet This field holds the fourth octet of the station address. The fourth octet (station address bits 24–31) defaults to a value of 0x00. 24–31 Station Address, 3rd Octet This field holds the third octet of the station address. The third octet (station address bits 16–23) defaults to a value of 0x00. 21.5.3.6.8 Station Address Register Part 2 (MACSTNADDR2) The MACSTNADDR2 register is written by the user. Figure 21-37 describes the MACSTNADDR2 register. Note that the I/G and U/L bits of the frame’s DA field are located at the lsbs of the 1st octet stored in MACSTNADDR2, where the I/G bit is bit 15, and the U/L bit is bit 14. 0 7 8 15 16 31 R W Reset Offset Station Address, 2nd Octet 0000000000000000 Station Address, 1st Octet 0000_0000_0000_0000_0000_0000_0000_0000 0x2_6544 Figure 21-37. Station Address Part 2 Register Definition MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 21-44 Freescale Semiconductor Memory Map/Register Definition Table 21-36 describes the fields of the MACSTNADDR2 register. Table 21-36. MACSTNADDR2 Field Descriptions Bit 0–7 8–15 16–3 1 Name Description Station Address, 2nd Octet This field holds the second octet of the station address. The second octet (station address bits 8–15) defaults to a value of 0x00. Station Address, 1st Octet This field holds the first octet of the station address. The first octet (station address bits 0–7) defaults to a value of 0x00. — Reserved 21.5.3.7 Hash Function Registers This section provides detailed descriptions of the registers used for hash functions. All of the registers are 32 bits wide. When the DA field of a receive frame is processed through a 32-bit CRC generator, the 8-bit MSB of the CRC remainder is mapped to a hash table entry. The user can enable a hash entry by setting the appropriate bit. A hash entry usually represents a set of addresses. A hash table hit occurs when the DA CRC result points to an enabled hash entry. The user must further filter the address. See Section 21.6.2.5.2, “Hash Table Algorithm,” for more information on the hash algorithm. 21.5.3.7.1 Individual Address Registers 0–7 (IADDRn) The IADDRn registers, shown in Figure 21-38, represent 256 entries of the individual (unicast) address hash table used in the address recognition process. When the DA field of a receive frame is processed through a 32-bit CRC generator, the 8 high-order bits (0–7) of the CRC remainder are mapped to one of the 256 entries. The user can enable a hash entry by setting the appropriate bit. A hash table hit occurs when the DA CRC result points to an enabled hash entry. 0 31 R IADDRn W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x2_6800, 0x2_6804, 0x2_6808, 0x2_680C, 0x2_6810, 0x2_6814, 0x2_6818, 0x2_681C Figure 21-38. IADDRn Register Definition Table 21-37 describes the field of the IADDRn registers. Table 21-37. IADDRn Field Description Bits 0–31 Name Description IADDRn Represents the 32-bit value associated with the corresponding register. IADDR0 contains the high-order 32 bits of the 256-entry hash table and IADDR7 represents the low-order 32 bits. For instance, the MSB of IADDR0 correlates to entry 0 and the LSB of IADDR7 correlates to entry 255. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 21-45 10/100 Fast Ethernet Controller 21.5.3.7.2 Group Address Registers 0–7 (GADDRn) The GADDRn registers, shown in Figure 21-39, represent 256 entries of the group (multicast) address hash table used in the address recognition process. When the DA field of a receive frame is processed through a 32-bit CRC generator, the 8-bit MSB of the CRC remainder is mapped to one of the 256 entries. The user can enable a hash entry by setting the appropriate bit. A hash table hit occurs when the DA CRC result points to an enabled hash entry. 0 15 R GADDR n W Reset Offset 0000_0000_0000_0000_0000_0000_0000_0000 0x2_6880, 0x2_6884, 0x2_6888, 0x2_688C, 0x2_6890, 0x2_6894, 0x2_6898, 0x2_689C Figure 21-39. GADDRn Register Definition Table 21-38 describes the field of the GADDRn registers. Table 21-38. GADDRn Field Description Bits 0–31 Name Description GADDRn Represents the 32-bit value associated with the corresponding register. GADDR0 contains the high-order 32 bits of the 256-entry group hash table and GADDR7 represents the low-order 32 bits. For instance, the MSB of GADDR0 correlates to entry 0 and the LSB of GADDR7 correlates to entry 255. 21.5.3.8 Attribute Registers This section describes the two FEC frame attribute registers. 21.5.3.8.1 Attribute Register (ATTR) The ATTR, shown in Figure 21-40, defines attributes and transaction types used to access buffer descriptors and to write receive data and to read transmit data. Snoop enable attributes may be set for reading buffer descriptors and for reading transmit data. Buffer descriptors may be written with attributes that cause allocation into the L2 cache with or without line locking. Similarly, sections of a receive frame header may have attributes attached that cause allocation and locking in the L2 cache. This process of specifying a region of each frame to stash into the L2 cache is referred to as extraction. Extraction is specified in conjunction with ATTRELI. ATTR[ELCWT] only has meaning if the ATTRELI[EL] field is non-zero. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 21-46 Freescale Semiconductor Memory Map/Register Definition 0 15 16 17 18 19 20 21 22 23 24 25 26 31 R W Reset Offset 0000_0000_0000_0000 0 ELCWT 0 BDLWT 00 RDSEN RBDSEN 000000 0000_0000_0000_0000_0000_0000_0000_0000 0x2_6BF8 Figure 21-40. ATTR Register Definition Table 21-39 describes the fields of the ATTR register. Table 21-39. ATTR Field Descriptions Bits 0–16 17–18 Name — ELCWT Reserved Extracted L2 cache write type. Specifies the write transaction type to perform for the extracted data. This occurs only if ATTRELI[EL] is non-zero. For maximum performance, it is recommended that if ELCWT is set to allocate, BDLWT also be set to allocate. Writes to cache are always performed with snoop. This setting overrides the RDSEN bit setting. 00 No allocation performed. 01 Reserved, no extraction occurs. 10 Allocate L2 cache line. 11 Allocate and lock L2 cache line. Reserved Buffer descriptor L2 cache write type. Specifies the write transaction type to perform for the buffer descriptor for a receive frame. Writes to cache are always performed with snoop. 00 No allocation performed. This setting overrides the RBDSEN bit setting. 01 Reserved 10 Allocate L2 cache line. 11 Allocate and lock L2 cache line. Reserved Rx data snoop enable. This bit is superseded by the ELCWT settings. 0 Disables snooping of all receive frames data to memory unless ELCWT specifies L2 allocation. 1 Enables snooping of all receive frames data to memory. Description 19 20–21 — BDLWT 22–23 24 — RDSEN 25 RBDSEN RxBD snoop enable. This bit is superceded by the BDLWT settings. 0 Disables snooping of all receive BD memory accesses unless BDLWT specifies L2 allocation. 1 Enables snooping of all receive BD memory accesses. — Reserved 26–31 21.5.3.8.2 Attribute Extract Length and Extract Index Register (ATTRELI) The ATTRELI registers are written by the user to specify the extract index and extract length. Figure 21-41 describes the definition for ATTRELI. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 21-47 10/100 Fast Ethernet Controller 0 1 2 15 16 17 18 31 R W RESET Offset 00 EL 00 EI 0000_0000_0000_0000_0000_0000_0000_0000 0x2_6BFC Figure 21-41. ATTRELI Register Definition Table 21-40 describes the fields of the ATTRELI register. Table 21-40. ATTRELI Field Descriptions Bits 0–1 2–15 16–17 18–31 Name — EL — EI Reserved Extracted length. Specifies the number of bytes to extract from the receive frame. The DMA controller uses this field to perform extraction. If set to zero, no extraction is performed. Reserved Extracted index. Points to the first byte within the receive frame from which to begin extracting data. Description 21.6 Functional Description This section describes many of the functions of the 10/100 fast Ethernet controller (FEC). 21.6.1 Connecting to the Physical Interface This section describes how to connect the FEC to the MII interface. The bus follows the bus conventions used in the IEEE 802.3 specification because the PHY follows the same conventions (that is, in the bus TXD[3:0], bit 3 is the MSB and bit 0 the LSB). 21.6.1.1 Media-Independent Interface (MII) This section describes the media independent interface (MII) intended to be used between the PHYs and the FEC. Figure 21-42 depicts the basic components of the MII including the signals required to establish the FEC module connection with a PHY. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 21-48 Freescale Semiconductor Functional Description Transmit Error (FEC_TX_ER) Transmit Data (FEC_TXD[3:0]) Transmit Enable (FEC_TX_EN) Transmit Clock (FEC_TX_CLK) Collision Detect (FEC_COL) 10/100 Controller Receive Data (RXD[3:0]) Receive Error (FEC_RX_ER) Receive Clock (FEC_RX_CLK) Receive Data Valid (FEC_RX_DV) Carrier Sense Output (FEC_CRS) 10/100 Controller PHY M edium Figure 21-42. FEC–MII Connection An MII interface has 16 signals, as defined by the IEEE 802.3u standard, for connecting to an Ethernet PHY. Table 21-41 lists the MII interface signals. Table 21-41. MII Interface Signals Frequency [MHz] 25 Voltage [V] 3.3 Signals FEC_TX_CLK FEC_TxD[0] FEC_TxD[1] FEC_TxD[2] FEC_TxD[3] FEC_TX_EN FEC_TX_ER FEC_RX_CLK I/O I O O O O O O I No. of Signals 1 1 1 1 1 1 1 1 Signals FEC_RxD[0] FEC_RxD[1] FEC_RxD[2] FEC_RxD[3] FEC_RX_DV FEC_RX_ER FEC_COL FEC_CRS I/O I I I I I I I I No. of Signals 1 1 1 1 1 1 1 1 21.6.2 FEC Channel Operation This section describes the operation of the FEC. First the software initialization sequence is described. Next, the software (Ethernet driver) interface for transmitting and receiving frames is described. Address recognition and hash table algorithm features are also discussed. The section concludes with interrupt handling, inter-packet gap time, and loop-back descriptions. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 21-49 10/100 Fast Ethernet Controller 21.6.2.1 Initialization Sequence This section describes which registers are reset due to a hard or software reset and what registers the user must initialize before the FEC is enabled. 21.6.2.1.1 Hardware Controlled Initialization A hard reset occurs when the system powers up. All of the FEC’s registers and control logic are reset to their default states after a hard reset has occurred. 21.6.2.1.2 User Initialization After the system has undergone a hard reset, software must initialize certain basic FEC registers. Other registers can also be initialized during this time, but they are optional and must be determined based on the requirements of the system. See Table 21-2 for the register list. Table 21-42 describes the minimum steps for register initialization. Table 21-42. Steps of Minimum Register Initialization Description 1. Set, then clear MACCFG1 [Soft_Reset] 2. Initialize MACCFG2 3. Initialize MAC station address 4. Set up the PHY using the MII Mgmt Interface 5. Clear IEVENT 6. Initialize IMASK 7. Initialize IADDR n 8. Initialize GADDRn 9. Initialize RCTRL 10. Initialize DMACTRL After the registers are initialized, the user must execute the following steps in the order described below to bring the FEC into a functional state (out of reset): 1. For the transmission of Ethernet frames, TxBDs have to be first built in memory, linked together as a ring, and pointed to by the TBASE register. At least two buffer descriptors per ring are required. Setting the ring size to one causes the same frame to be sent twice. 2. Likewise, for the reception of Ethernet frames, the receive queue must be ready, with its RxBD pointed to by the RBASE register. Both transmit and receive can be gracefully stopped after transmission and reception begins. 3. Write to the MACCFG1 register and set the appropriate bits. These must include RX_EN = 1, and TX_EN = 1. To enable flow control, Rx_Flow and Tx_Flow must also be set. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 21-50 Freescale Semiconductor Functional Description 4. Clearing DMACTRL[GTS] triggers the transmission of frame data if the transmitter had been previously stopped. DMACTRL[GRS] must be cleared if the receiver was previously stopped. See Section 21.5.3.1.6, “DMA Control Register (DMACTRL),” and Section 21.6.3.1, “Transmit Data Buffer Descriptor (TxBD).” 21.6.2.2 Soft Reset and Reconfiguring Procedure Before issuing a soft-reset to and/or reconfiguring the MAC with new parameters, the user must properly shutdown the DMA and make sure it is in an idle state for the entire duration. The user must gracefully stop the DMA by setting both GRS and GTS bits in the DMACTRL register, then wait for both GRSC and GTSC bits to be set in the IEVENT register before resetting the MAC or changing parameters. Both GRS and GTS bits must be cleared before re-enabling the MAC to resume the DMA. During the MAC configuration, if a new set of Tx buffer descriptors are used, the user must load the pointers into the TBASE register. Likewise if a new set of Rx buffer descriptors are used, the RBASE register must be written with the new pointer. Following is a procedure to gracefully reset and reconfigure the MAC: 1. Set GTS bit in DMACTRL register. 2. Poll GTSC bit in IEVENT register until detected as set. 3. Clear both Rx_EN and Tx_EN bits in MACCFG1. 4. Wait for a period of 9.6 Kbytes worth of data on the interface (~8ms worst case). 5. Set GRS bit in DMACTRL register. 6. Poll GRSC bit in IEVENT register until detected as set. 7. Set Soft_Reset bit in MACCFG1 register. 8. Clear Soft_Reset bit in MACCFG1 register. 9. Load TBASE with new TxBD pointer. 10. Load RBASE with new RxBD pointer. 11. Set up other MAC registers (MACCFG2, MAXFRM, and so on). 12. Set WWR and WOP bits in DMACTRL register. 13. Clear THLT bit in TSTAT register and QHLT bit in RSTAT register by writing 1 to these bits. 14. Clear GRS/GTS bits in DMACTRL. (Do not change other bits.) 15. Enable Tx_EN/Rx_EN in MACCFG1 register. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 21-51 10/100 Fast Ethernet Controller 21.6.2.3 FEC Frame Transmission The Ethernet transmitter requires almost no core intervention. After the software driver initializes the system, the FEC begins to poll the first TxBD in the TxBD ring every 512 transmit clocks. If TxBD[R] is set, the FEC begins moving transmit buffer from memory to its Tx FIFO. The transmitter takes data from the Tx FIFO and transmits data to the MAC. The MAC transmits the data through the MII interface to the physical media. The transmitter, once initialized, runs until the end-of-frame (EOF) condition is detected unless a collision within the collision window occurs (half-duplex mode) or an abort condition is encountered. If the user has a frame ready to transmit, a transmit-on-demand function may be emulated while in polling mode by using the graceful-transmit-stop feature. First, clear the IMASK[GTSCEN] bit to mask the graceful-transmit-stop complete interrupt. Next set, then immediately clear the DMACTRL[GTS] bit. Clear the resulting IEVENT[GTSC] bit. Finally, the IMASK[GTSCEN] bit may be set once again. There is one internal buffer for out-of-sequence flow control frames. When the FEC is between frames, this buffer is polled. The buffer must contain the whole frame. Once the FEC is in paused mode, the out-of-sequence buffer descriptor cannot be used to send another flow control frame because the MAC treats it like a regular TxBD. In half-duplex mode (MACCFG2[Full_Duplex] = 0) the MAC defers transmission if the line is busy (FEC_CRS asserted). Before transmitting, the MAC waits for carrier sense to become inactive, at which point it then determines if FEC_CRS remains negated for 60 clocks. If so, transmission begins after an additional 36 bit times (96 bit times after FEC_CRS originally became negated). If FEC_CRS continues to be asserted, the MAC follows a specified back-off procedure and tries to retransmit the frame until the retry limit is reached. Data stored in the Tx FIFO is retransmitted in case of a collision. This improves bus usage and latency. The transmitter also monitors for an abort condition and terminates the current frame if an abort condition is encountered. In full-duplex mode the protocol is independent of network activity, and only the transmit inter-frame gap needs to be enforced. The transmit block also implements full-duplex flow control. If a flow control frame is received, the MAC does not service the transmitter’s request to send data until the pause duration is over. If the MAC is currently sending data while a pause frame is received, the MAC finishes sending the current frame, then suspends subsequent frames (except a pause frame) until the pause duration is over. The pause duration is defined by the received pause control frame and begins when the frame was first received. In addition, the transmitter supports transmission of flow control frames via TCTRL[TFC_PAUSE]. The transmit pause frame is generated internally based on the PAUSE register that defines the pause value to be sent. Note that it is possible to send a pause frame while the pause timer has not expired. The MAC automatically appends FCS (32-bit CRC) bytes to the frame if any of the following values are set: MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 21-52 Freescale Semiconductor Functional Description • • • • TxBD[PAD/CRC] is set in first TxBD TxBD[TC] is set in first TxBD MACCFG2[PAD/CRC] is set MACCFG2[CRC EN] is set After the FCS is sent, the Ethernet controller writes the frame status bits into the BD and clears TxBD[R]. When the end of the current buffer is reached and TxBD[L] = 0 (a frame is comprised of multiple buffer descriptors), only TxBD[R] is cleared. For both half- and full-duplex modes, an interrupt can be issued depending on TxBD[I]. The Ethernet controller then proceeds to the next TxBD in the table. In this way, the core can be interrupted after each frame, after each buffer, or after a specific buffer is sent. If TxBD[PAD/CRC] = 1, the Ethernet controller pads any frame shorter than 64 bytes. To pause transmission, or rearrange the transmit queue, set DMACTRL[GTS]. This can be useful for transmitting expedited data ahead of previously linked buffers or for error situations. When this bit is set the FEC transmitter performs a graceful transmit stop. The Ethernet controller stops immediately if no transmission is in progress or continues transmission until the current frame either finishes or terminates with an error. The IEVENT[GTSC] interrupt occurs when the graceful transmit stop operation is completed. When DMACTRL[GTS] is cleared, the FEC resumes transmission with the next frame. The FEC sends bytes least-significant nibble first and each nibble is sent least-significant bit first. 21.6.2.4 FEC Frame Reception The FEC Ethernet receiver is designed to work with almost no core intervention and can perform address recognition, CRC checking, short frame checking, and maximum frame-length checking. The receiver can also force frame headers and buffer descriptors to be allocated into the L2 cache. See Section 21.6.4, “Data Extraction to the L2 Cache” for additional information. After a hardware reset, the software driver clears the RSTAT register and sets MACCFG1[RX_EN]; the Ethernet receiver is enabled and immediately starts processing receive frames. The MAC checks for when FEC_RX_DV is asserted and as long as FEC_COL remains negated (full-duplex mode ignores FEC_COL) the MAC looks for the start of a frame by searching for a valid preamble/SFD (start of frame delimiter) header, which is stripped, and the frame begins to be processed. If a valid header is not found, the frame is ignored. If the receiver detects the first bytes of a frame, the FEC begins to perform the frame recognition function through destination address (DA) recognition. (See Section 21.6.2.5, “Frame Recognition” for additional information.) Based on this match the frame can be accepted or rejected. Once accepted, the FEC processes the frame based on user-defined attributes. The receiver can also filter frames based on physical (individual), group (multicast), and broadcast addresses. Because Ethernet receive frame data is not written to memory until the internal frame MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 21-53 10/100 Fast Ethernet Controller recognition algorithm is complete, system bus usage is not wasted on frames not wanted by this station. If a frame is accepted, the Ethernet controller fetches the receive buffer descriptor (RxBD) from the queue. If the RxBD is not being used by software (RxBD[E] is set), the FEC begins transferring the incoming frame. RxBD[F] is set for the first RxBD used for any receive frame. If the buffer is filled, the FEC clears RxBD[E] and generates an interrupt if RxBD[I] is set. If the incoming frame is larger than the buffer, the Ethernet controller fetches the next RxBD in the table; if it is empty, it continues receiving the rest of the frame. In half-duplex mode, if a collision is detected during the frame, no RxBDs are used. Thus, no collision frames are presented to the user except late collisions, which indicate serious LAN problems. The RxBD length is determined by the MRBL field in the maximum receive buffer length register (MRBLR). The smallest valid value is 64 bytes. During reception, the Ethernet controller checks for frames that are too short or too long. When the frame ends (FEC_CRS is negated), the receive CRC field is checked and written to the data buffer. The data length written to the last RxBD in the Ethernet frame is the length of the entire frame, which enables the software to recognize a frame-too-long condition. Receive frames are not truncated if they exceed maximum frame bytes in the MAC’s maximum frame register if MACCFG2[Huge Frame] is set, however the babbling receiver error interrupt occurs (IEVENT[BABR] is set) and RxBD[LG] is set. When the receive frame is complete, the Ethernet controller sets RxBD[L], updates the frame status bits in the RxBD, and clears RxBD[E]. If RxBD[I] is set, the Ethernet controller next generates a maskable interrupt, indicating that a frame was received and is in memory. The Ethernet controller then waits for a new frame. To interrupt reception, or rearrange the receive queue, DMACTRL[GRS] must be set. When this bit is set the FEC receiver does a graceful receive stop. The Ethernet controller stops immediately if no frames are being received or continues receiving until the current frame either finishes or an error condition occurs. The IEVENT[GRSC] interrupt event is signalled when the graceful receive stop operation is completed. While in this mode the user can then clear IEVENT[GRSC] and can write to registers that are accessible to both user and the FEC hardware without fear of conflict. When DMACTRL[GRS] is cleared, the FEC scans the input data stream for the start of a new frame (preamble sequence and start of frame delimiter), it resumes receiving, and the first valid frame received is placed in the next available RxBD. 21.6.2.5 Frame Recognition The Ethernet controller performs frame recognition using destination address (DA) recognition. A frame can be rejected or accepted based on the outcome. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 21-54 Freescale Semiconductor Functional Description 21.6.2.5.1 Destination Address Recognition The Ethernet controller can also perform the frame filtering using the traditional destination address (DA) recognition methods. Figure 21-43 depicts a flowchart for address recognition on received frames that is used to explain the concept. In the actual implementation most of the decision points shown in the figure actually occur simultaneously. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 21-55 10/100 Fast Ethernet Controller Incoming Frame G I/G Address I Broadcast Address T F T Station Address Match? F Hash Search (Use Group Table) T Hash Match? Hash Search (Use Individual Table) F Broadcast Reject T F F Promiscuous? T Discard Frame Receive Frame Figure 21-43. Ethernet Address Recognition Flowchart MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 21-56 Freescale Semiconductor Functional Description The Ethernet controller compares the destination address field of the received frame with the physical address that the user programs in the station address registers (MACSTNADDR1 and MACSTNADDR2). If the DA does not match the station address, then the FEC performs address recognition on multiple individual addresses using the IADDRn hash table. The user must write zeros to the hash in order to avoid a hash match, and ones to station address in order to avoid individual address match, or the user can turn on promiscuous mode (See Section 21.5.3.4.1, “Receive Control Register (RCTRL).”) In the group type of address recognition, the Ethernet controller determines whether the group address is a broadcast address. If it is a broadcast, and broadcast addresses are enabled, the frame is accepted. If the group address is not a broadcast address, the user can perform address recognition on multiple group addresses using the GADDRn hash table. In promiscuous mode, the Ethernet controller receives all of the incoming frames regardless of their address. 21.6.2.5.2 Hash Table Algorithm The hash table process used in the individual and group hash filtering operates as follows. The Ethernet controller maps any 48-bit destination address into one of 256 bins, represented by the 256 bits in GADDR0–7 or IADDR0–7. The eight high-order bits of a cyclic redundancy check (CRC) checksum are used to index into the hash table. The high-order three bits of this 8-bit field are used to select one of the eight registers in either the individual or group hash table. The low-order five bits select a bit within the 32-bit register. A value of 0 in the high-order three bits selects IADDR0/GADDR0. The same process is used if the Ethernet controller receives a frame. If the CRC checksum selects a bit that is set in the group/individual hash table, the frame is accepted. If 32 group addresses are stored in the hash table and random group addresses are received, the hash table prevents roughly 224/256 (87.5%) of the group address frames from reaching memory. Software must further filter those that reach memory to determine if they contain the correct addresses. Better performance is achieved by using the group and individual hash tables in combination. For instance, if 32 group and 32 physical addresses are stored in their respective hash tables, because 87.5% of all group addresses and 87.5% of all individual address are rejected, then 87.5% of all frames are prevented from reaching memory. The effectiveness of the hash table declines as the number of addresses increases. For instance, as the number of addresses stored in the 256-bin hash table increases, the vast majority of the hash table bits are set, preventing only a small fraction of frames from reaching memory. 21.6.2.5.3 CRC Computation Examples There are many algorithms for calculating the CRC value of a number. Refer to the RFC 3309 standard, which can be found at http://www.faqs.org/rfcs/rfc3309.html, to compute the CRC value MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 21-57 10/100 Fast Ethernet Controller for the purposes of FEC. The RFC 3309 algorithm uses the following polynomial to calculate the CRC value: x32+x26+x23+x22+x16+x12+x11+x10+x8+x7+x5+x4+x2+x1+x0 or 0x04c11db7. Given a destination MAC address of DA=01000CCCCCCC, the algorithm results in a CRC remainder value of 0xA29F4BBC. Bit-reversing the low-order byte of the CRC value (0xBC) yields BR_CRC = 0x3D = 0b00111101 The high-order 3-bits of the new BR_CRC value are used to select which 32-bit register (of the 8) to use. This example maps the DA to register 1. High-order 3 bits of BR_CRC: HO_CRC = 0b001 = 1 The low-order 5 bits are used to select which bit to set in the given register (with a value of 0 setting 0x8000_0000 and 31 setting 0x0000_0001). Therefore, the example DA maps to bit 29 of register 1. Low-order 5 bits of BR_CRC: LO_CRC = 0b11101 = 29 Therefore, GADDR1 is ORed with the value 0x0000_0004. Additional calculated examples follow: Example 1: • • • • • • • • • • • • Destination MAC address: DA = 01005E000128 CRC remainder value: CRC = 0x821D6CD3 Bit-reversed least-significant byte of CRC value: BR_CRC = 0xCB = 0b11001011 High-order 3 bits of BR_CRC: HO_CRC = 0b110 = 6 Low-order 5 bits of BR_CRC: LO_CRC = 0b01011 = 11 GADDR6 = 0x0010_0000 Destination MAC address: DA = 0004F0604F10 CRC remainder value: CRC = 0x1F5A66B5 Bit-reversed least-significant byte of CRC value: BR_CRC = 0xAD = 0b10101101 High-order 3 bits of BR_CRC: HO_CRC = 0b101 = 5 Low-order 5 bits of BR_CRC: LO_CRC = 0b01101 = 13 GADDR5 = 0x0004_0000 Example 2: MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 21-58 Freescale Semiconductor Functional Description 21.6.2.6 Flow Control Because collisions cannot occur in full-duplex mode, the FEC can operate at the maximum rate. When the rate becomes too fast for a station’s receiver, the station’s transmitter can send flow-control frames to reduce the rate. Flow-control instructions are transferred by special frames of minimum frame size. The length/type fields of these frames have a special value. Table 21-43 shows the flow control frame structure. Table 21-43. Flow Control Frame Structure Size [Octets ] 7 1 6 6 2 2 2 Description Value Comment Preamble SFD Destination address Source address Length/type MAC opcode MAC parameter 88-08 00-01 Control frame type Pause command Pause period measured in slot times, most-significant octet first with a two time-slot resolution. Note: Because the pause period has a resolution of two time slots, the value programmed in this field is rounded up to the nearest even number before being used, as follows: MAC Parameter Value 0 1 or 2 3 or 4 … Pause Period None 2 × slot time 4 × slot time … Start frame delimiter 01-80C2-00-00-01 Multicast address reserved for use in MAC frames 42 4 Reserved FCS — Frame check sequence (CRC) When flow-control mode is enabled (MACCFG1[Rx_Flow] is set) and the receiver identifies a pause-flow control frame, transmission stops for the time specified in the control frame. During this pause, only a control frame can be sent (TCTRL[TFC_PAUSE] is set). Normal transmission resumes after the pause timer stops counting. If another pause-control frame is received during the pause, the period changes to the new value received. 21.6.2.7 Interrupt Handling The following describes what usually occurs within a FEC interrupt handler: • When an interrupt occurs, read IEVENT to determine interrupt sources. IEVENT bits to be handled in this interrupt handler are normally cleared at this time. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 21-59 10/100 Fast Ethernet Controller • • • • Process the TxBDs to reuse them if the IEVENT[TXB] or IEVENT[TXE] were set. If the transmit speed is fast or the interrupt delay is long, more than one transmit buffer may have been sent by the FEC. Thus, it is important to check more than just one TxBD during the interrupt handler. One common practice is to process all TxBDs in the interrupt handler until one is found with R set. See Table 21-44. Obtain data from the RxBD if IEVENT[RXC], IEVENT[RXB], or IEVENT[RXF] is set. If the receive speed is fast or the interrupt delay is long, the FEC may have received more than one receive buffer. Thus, it is important to check more than just one RxBD during interrupt handling. Typically, all RxBDs in the interrupt handler are processed until one is found with bit E set. Because the FEC pre-fetches BDs, the BD table must be big enough so that there is always another empty BD to pre-fetch. See Table 21-45. Clear any set halt bits in TSTAT and RSTAT registers, or DMACTRL[GTS] and DMACTRL[GRS]. Continue normal execution. Table 21-44. Non-Error Transmit Interrupts Interrupt GTSC Description Graceful transmit stop complete: Transmitter is put into a pause state after completion of the frame currently being transmitted. Transmit control: Instead of the next transmit frame, a control frame was sent. Transmit buffer: A TxBD, that is not the last one in the frame, was updated. None Action taken by FEC TXC TXB TXF None Programmable write-with-response TxBD to memory before setting IEVENT[TXB]. Transmit frame: A frame was transmitted and the last TxBD of Programmable write-with-response to that frame was updated. memory on the last TxBD before setting IEVENT[TXF]. Table 21-45. Non-Error Receive Interrupts Interrupt GRSC RXC Description Graceful receive stop complete: Receiver is put into a pause None state after completion of the frame currently being received. Receive control: A control frame was received. As soon as the transmitter finishes sending the current frame, a pause operation is performed lasting for the duration specified in the received pause control frame and beginning when the frame was first received. Receive buffer: An RxBD, that is not the last one of the frame, was updated. Receive frame: A frame was received and the last RxBD of that frame was updated. None Action taken by FEC RXB RXF Programmable write-with-response RxBD to memory before setting IEVENT[RXB]. Programmable write-with-response to memory on the last RxBD before setting IEVENT[RXF]. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 21-60 Freescale Semiconductor Functional Description 21.6.2.8 Inter-Packet Gap Time When a station needs to transmit, it waits until the LAN becomes silent for a specified period (inter-packet gap). When a station starts sending, it continually checks for collisions on the LAN. If a collision is detected, the station forces a jam signal (all ones) on its frame and stops transmitting. Collisions usually occur close to the beginning of a packet. The station then waits a random time period (back-off) before attempting to send again. When the back-off completes, the station waits for silence on the LAN and then begins retransmission on the LAN. This process is called a retry. If the packet is not successfully sent within specified number of retries, an error is indicated. The minimum inter-packet gap time for back-to-back transmission is 96 serial clocks. The receiver receives back-to-back packets with this minimum spacing. In addition, after waiting a required number of clocks (based on the back-off algorithm), the transmitter waits for carrier sense to be negated before retransmitting the packet. Retransmission begins 36 serial clocks after carrier sense is negated for at least 60 serial clocks. 21.6.2.9 Internal and External Loop Back Setting MACCFG1[Loop_Back] causes MAC transmit outputs to be looped back to the MAC receive inputs. Clearing it results in normal operation. This bit is cleared by default. 21.6.2.10 Error-Handling Procedure The Ethernet controller reports frame reception and transmission error conditions using the channel BDs, the error counters, and the IEVENT register. Transmission errors are described in Table 21-46. Table 21-46. Transmission Errors Error Transmitter underrun Response The FEC sends 32 bits that ensure a CRC error, terminates buffer transmission, sets TxBD[UN], closes the buffer, sets IEVENT[XFUN] and IEVENT[TXE]. The controller resumes transmission after TSTAT[THLT] is cleared (and DMACTRL[GTS] is cleared). The FEC terminates buffer transmission, sets TxBD[RL], closes the buffer, sets IEVENT[CRL/XDA] and IEVENT[TXE]. Transmission resumes after TSTAT[THLT] is cleared (and DMACTRL[GTS] is cleared). The FEC terminates buffer transmission, sets TxBD[LC], closes the buffer, sets IEVENT[LC] and IEVENT[TXE]. The controller resumes transmission after TSTAT[THLT] is cleared (and DMACTRL[GTS] is cleared). A system bus error occurred during a DMA transaction. The FEC sets IEVENT[EBERR], DMA stops sending data to the FIFO which causes an underrun error but IEVENT[XFUN] is not set. The TSTAT[THLT] is set. Transmits are continued once TSTAT[THLT] is cleared. Retransmission attempts limit expired Late collision Memory read error MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 21-61 10/100 Fast Ethernet Controller Table 21-46. Transmission Errors (continued) Error Babbling transmit error Response A frame is transmitted which exceeds the MAC’s Maximum Frame Length and MACCFG2[Huge Frame] is a 0. The FEC sets IEVENT[BABT] and continues without interruption. TxBD[TXTRUNC] is set in the last TxBD (TxBD[L] is set) of the frame. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 21-62 Freescale Semiconductor Functional Description Reception errors are described in Table 21-47. Table 21-47. Reception Errors Error Description O verrun error The FEC maintains an internal FIFO buffer for receiving data. If a receiver FIFO buffer overrun occurs, the controller sets RxBD[OV], sets RxBD[L], closes the buffer, and sets IEVENT[RXF], The receiver then enters hunt mode (seeking start of a new frame). Busy error Non-octet error (dribbling bits) CRC error A frame is received and discarded due to a lack of buffers. The FEC sets IEVENT[BSY]. In addition, RSTAT[RHLT] is set. The halted queue resumes reception once the user clears RSTAT[RHLT]. The Ethernet controller handles a nibble of dribbling bits if the receive frame terminates as non-octet aligned and it checks the CRC of the frame on the last octet boundary. If there is a CRC error, the frame non-octet aligned (RxBD[NO]) error is reported, IEVENT[RXF] is set, and the alignment error counter increments. The FEC relies on the statistics collector block to increment the receive alignment error counter (RALN). If there is no CRC error, no error is reported. If a CRC error occurs, the FEC sets RxBD[CR], closes the buffer, and sets IEVENT[RXF]. This FEC relies on the statistics collector block to record the event. After receiving a frame with a CRC error, the receiver then enters hunt mode. A system bus error occurred during a DMA transaction. The FEC sets IEVENT[EBERR] and discards the frame. Also, RSTAT[RHLT] is set. The halted queue resumes reception once RSTAT[RHLT] is cleared. Memory Read Error Babbling A frame is received that exceeds the MAC’s maximum frame length. The FEC sets IEVENT[BABR] and Receive Error continues. 21.6.3 Buffer Descriptors The FEC buffer descriptor (BD) is modeled after the MPC8260 Fast Ethernet controller BD for ease of reuse. Drawing from the MPC8260 FEC BD programming model, the FEC descriptor base registers point to the beginning of BD rings. There is an 8-byte data BD format designed to be similar to the existing MPC8260 BD model. The FEC is capable of placing extracted data directly into the L2 cache memory. This allows the processor to quickly access critical frame information as soon as the processor is ready without having to first fetch the data from main memory. This results in substantial improvement in throughput and hence reduction in latency. Data buffers are used in the transmission and reception of Ethernet frames (see Figure 21-44). Data BDs encapsulate all information necessary for the FEC to transmit or receive an Ethernet frame. Within each data BD there is a status field, a data length field, and a data pointer. The BD completely describes an Ethernet packet by centralizing status information for the data packet in the status field of the BD and by containing a data BD pointer to the location of the data buffer. Software is responsible for setting up the BDs in memory. Because of prefetching, at least two buffer descriptors per ring are required. This applies to both the transmit and the receive descriptor rings. Software also must have the data pointer pointing to memory. An ownership bit in the status field defines the current state of the buffer (pointed to by the data pointer). Other status field bits MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 21-63 10/100 Fast Ethernet Controller in the buffer descriptor communicate status/control information between the FEC and the software driver. Because there is no next BD pointer in the transmit/receive BD (see Figure 21-45), all BDs must reside sequentially in memory. The FEC increments the current BD location appropriately to the next BD location to be processed. There is a wrap bit in the last BD that informs the FEC to loop back to the beginning of the BD chain. Software must initialize the TBASE and RBASE registers that point to the beginning transmit and receive BDs for the FEC. Memory Map System Memory Tx Buffer Descriptors Status & Control TxBD Table Pointer (TBASE) RxBD Table Pointer (RBASE) TxBD Table Data Length Buffer Pointer Tx Buffer Rx Buffer Descriptors RxBD Table Status & Control Data Length Buffer Pointer Rx Buffer Figure 21-44. Example of FEC Memory Structure for BD Beginning BD pointer 0 1 4 W=1 3 2 Figure 21-45. Buffer Descriptor Ring MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 21-64 Freescale Semiconductor Functional Description 21.6.3.1 Transmit Data Buffer Descriptor (TxBD) Data is presented to the FEC for transmission by arranging it in memory buffers referenced by the TxBDs. In the TxBD the user initializes the R, PAD/CRC, W, L, and TC bits and the length (in bytes) in the first word, and the buffer pointer in the second word. The FEC clears the R bit in the first word of the BD after it finishes using the data buffer. The transfer status bits are then updated. Additional transmit frame status can be found in statistic counters in the MIB block. Figure 21-46 describes the TxBD. 0 1 2 34 5 6 7 8 9 10 11 12 13 14 15 Offset + 0 Offset + 2 Offset + 4 R PAD/CRC W I L T C DEF TO 1 HFE/LC RL RC U N TXTRUNC DATA LENGTH TX DATA BUFFER POINTER Offset + 6 Figure 21-46. Transmit Buffer Descriptor The TxBD fields are detailed in Table 21-48. Table 21-48. Transmit Data Buffer Descriptor (TxBD) Field Descriptions Offset Offset + 0 Bits 0 Name R Description Ready. Written by the FEC and user. 0 The data buffer associated with this BD is not ready for transmission. The user is free to manipulate this BD or its associated data buffer. The FEC clears this bit after the buffer is transmitted or after an error condition is encountered. 1 The data buffer, which is prepared for transmission by the user, was not transmitted or is currently being transmitted. No fields of this BD may be written by the user once this bit is set. PAD/CRC. Padding and CRC attachment for frames. (Valid only while it is set in the first BD and MACCFG2[PAD/CRC] is cleared.) If MACCFG2[PAD/CRC] is set, this bit is ignored. 0 Do not add padding to short frames. No CRC is appended unless TxBD[TC] is set. 1 Add PAD/CRCs to frames. Pad bytes are inserted until the length of the transmitted frame equals 64 bytes. Unlike the MPC8260 which pads up to MINFLR value, the FEC pads always up to the IEEE minimum frame length of 64 bytes. CRC is always appended to frames. Wrap. Written by user. 0 The next buffer descriptor is found in the consecutive location. 1 The next buffer descriptor is found at the location defined in TBASE. Interrupt. Written by user. 0 No interrupt is generated after this buffer is serviced. 1 IEVENT[TXB] or IEVENT[TXF] are set after this buffer is serviced. These bits can cause an interrupt if they are enabled (That is, IEVENT[TXBEN] or IEVENT[TXFEN] are set). Offset + 0 1 PAD/CRC Offset + 0 2 W Offset + 0 3 I MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 21-65 10/100 Fast Ethernet Controller Table 21-48. Transmit Data Buffer Descriptor (TxBD) Field Descriptions (continued) Offset Offset + 0 Bits 4 Name L Description Last in frame. Written by user. 0 The buffer is not the last in the transmit frame. 1 The buffer is the last in the transmit frame. Tx CRC. Written by user. (Valid only while it is set in first BD and TxBD[PAD/CRC] is cleared and MACCFG2[PAD/CRC] is cleared and MACCFG2[CRC EN] is cleared.) If MACCFG2[PAD/CRC] is set or MACCFG2[CRC EN] is set, this bit is ignored. 0 End transmission immediately after the last data byte with no hardware generated CRC appended, unless TxBD[PAD/CRC] is set. 1 Transmit the CRC sequence after the last data byte. Defer indication. Hardware updates this bit if an excessive defer condition occurs. 0 This frame was not deferred. 1 If HAFDUP[EXCESS_DEFER]=1, this frame did not have a collision before it was sent but it was sent late because of deferring. If HAFDUP[EXCESS_DEFER]=0, this frame was aborted and not sent. Transmit software ownership. This read/write bit may be utilized by software, as necessary. Its state does not affect the hardware nor is it affected by the hardware. Huge frame enable (written by user)/Late collision (written by the FEC) Huge frame enable. Written by user. Valid only while it is set in first BD and the MACCFG2[Huge Frame] is cleared. If MACCFG2[Huge Frame] is set, this bit is ignored. 0 Truncate transmit frame if its length is greater than the MAC’s Maximum Frame Length register. 1 Do not truncate the transmit frame. Late collision. Written by the FEC. 0 No late collision. 1 A collision occurred after 64 bytes are sent. The FEC terminates the transmission and updates LC. Offset + 0 9 RL Retransmission Limit. Written by the FEC. 0 Transmission before maximum retry limit is hit. 1 The transmitter failed (maximum retry limit + 1) attempts to successfully send a message due to repeated collisions. The FEC terminates the transmission and updates RL. Retry Count. Written by the FEC. 0 The frame is sent correctly the first time or if RL is set then the retry limit has been reached x One or more attempts where needed to send the transmit frame. If this field is 15, then 15 or more retries were needed. The Ethernet controller updates RC after sending the buffer. Underrun. Written by the FEC. 0 No underrun encountered (data was retrieved from external memory in time to send a complete frame). 1 The Ethernet controller encountered a transmitter underrun condition while sending the associated buffer. The FEC terminates the transmission and updates UN. Offset + 0 5 TC Offset + 0 6 DEF Offset + 0 Offset + 0 7 8 TO1 HFE/LC Offset + 0 10–13 RC Offset + 0 14 UN MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 21-66 Freescale Semiconductor Functional Description Table 21-48. Transmit Data Buffer Descriptor (TxBD) Field Descriptions (continued) Offset Offset + 0 Offset + 2 Bits 15 0–15 Name Description TXTRUNC TX truncation. Set in the last TxBD (TxBD[L] is set) when IEVENT[BABT] occurs for the frame. Data Length TX Data Buffer Pointer Data length is the number of octets the FEC transmits from this BD’s data buffer. It is never modified by the FEC. This field must be greater than zero. This field must be greater than zero. The transmit buffer pointer contains the address of the associated data buffer. There are no alignment requirements for this address. Offset + 4 0–31 21.6.3.2 Receive Buffer Descriptor (RxBD) In the RxBD the user initializes the E, I, and W bits in the first word and the pointer in second word. If the data buffer is used, the FEC modifies the E, L, F, M, BC, MC, LG, NO, SH, CR, OV, and TR bits and writes the length of the used portion of the buffer in the first word. The M, BC, MC, LG, NO, SH, CR, OV, and TR bits in the first word of the buffer descriptor are modified by the FEC only if the L bit is set. The first word of the RxBD contains control and status bits. Figure 21-47 describes the RxBD. 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Offset + 0 Offset + 2 Offset + 4 E RO1 W I L F 0 M BC MC LG NO SH CR OV TR DATA LENGTH RX DATA BUFFER POINTER Offset + 6 Figure 21-47. Receive Buffer Descriptor Table 21-49 describes the fields of the RxBD. Table 21-49. Receive Buffer Descriptor Field Descriptions Offset Offset + 0 Bits 0 Name E Description Empty. Written by the FEC (when cleared) and by the user (when set). 0 The data buffer associated with this BD is filled with received data, or data reception is aborted due to an error condition. Status and length fields have been updated as required. 1 The data buffer associated with this BD is empty, or reception is currently in progress. Receive software ownership bit. This field is reserved for use by software. This read/write bit is not modified by hardware, nor does its value affect hardware. Wrap. Written by user. 0 The next buffer descriptor is found in the consecutive location. 1 The next buffer descriptor is found at the location defined in RBASE. Offset + 0 1 RO1 Offset + 0 2 W MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 21-67 10/100 Fast Ethernet Controller Table 21-49. Receive Buffer Descriptor Field Descriptions (continued) Offset Offset + 0 Bits 3 Name I Description Interrupt. Written by user. 0 No interrupt is generated after this buffer is serviced. 1 IEVENT[RXB] or IEVENT[RXF] are set after this buffer is serviced. This bit can cause an interrupt if enabled (IMASK[RXBEN] or IMASK[RXFEN]). If the user wants to be interrupted only if RXF occurs, then the user must disable RXB (IMASK[RXBEN] is cleared) and enable RXF (IMASK[RXFEN] is set). Last in frame. Written by the FEC. 0 The buffer is not the last in a frame. 1 The buffer is the last in a frame. First in frame. Written by the FEC. 0 The buffer is not the first in a frame. 1 The buffer is the first in a frame. Reserved Miss. Written by the FEC. (This bit is valid only if L is set and the FEC is in promiscuous mode.) This bit is set by the FEC for frames that were accepted in promiscuous mode, but were flagged as a ‘miss’ by the internal address recognition; thus, while in promiscuous mode, the user can use bit M to quickly determine whether the frame was destined to this station. 0 The frame was received because of an address recognition hit. 1 The frame was received because of promiscuous mode. Broadcast. Written by the FEC. (Only valid if L is set.) Is set if the DA is broadcast (FF-FF-FF-FF-FF-FF). Multicast. Written by the FEC. (Only valid if L is set.) Is set if the DA is multicast and not BC. Rx frame length violation. Written by the FEC (only valid if L is set). A frame length greater than maximum frame length was recognized while MACCFG2[Huge Frame] was set. Note, if MACCFG2[Huge Frame] is cleared, the frame is truncated to the value programmed in the maximum frame length register. Rx non–octet-aligned frame. Written by the FEC (only valid if L is set). A frame that contained a number of bits not divisible by eight was received. Short frame. Written by the FEC (only valid if L is set). A frame length that was less than the minimum length defined for this channel (MINFLR) was recognized, provided RCTRL[RSF] is set. Rx CRC error. Written by the FEC (only valid if L is set). This frame contains a CRC error and is an integral number of octets in length.This bit is also set if a receive code group error is detected. Overrun. Written by the FEC (only valid if L is set). A receive FIFO overrun occurred during frame reception. If this bit is set, the other status bits, M, LG, NO, SH, CR, and CL lose their normal meaning and are zero. Truncation. Written by the FEC (only valid if L is set). Is set if the receive frame is truncated. This can happen if a frame length greater than maximum frame length was received and the MACCFG2[Huge Frame] is cleared. If this bit is set, the frame must be discarded and the other error bits must be ignored as they may be incorrect. Offset + 0 4 L Offset + 0 5 F Offset + 0 Offset + 0 6 7 — M Offset + 0 Offset + 0 Offset + 0 8 9 10 BC MC LG Offset + 0 Offset + 0 11 12 NO SH Offset + 0 13 CR Offset + 0 14 OV Offset + 0 15 TR MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 21-68 Freescale Semiconductor Application Information Table 21-49. Receive Buffer Descriptor Field Descriptions (continued) Offset Bits Name Data Length Rx Data Buffer Pointer Description Data length. Written by the FEC. Data length is the number of octets written by the FEC into this BD’s data buffer if L is cleared (the value is equal to MRBL), or the length of the frame including CRC, if L is set. Receive buffer pointer. Written by user. The receive buffer pointer, which always points to the first location of the associated data buffer, must be 64-byte aligned. The buffer must reside in memory external to the FEC. Offset + 2 0–15 Offset + 4 0–31 21.6.4 Data Extraction to the L2 Cache Some applications require the ability to identify selected portions of data within a frame’s data payload. This process is called extraction although the data is not truly removed. Rather than literarily extracting a section of the data and copying it into a new memory location, the data is placed in the L2 cache. This allows the processor to quickly access critical frame information as soon as the processor is ready without having to first fetch the data from main memory. This reduction in latency can result in a substantial improvement in packet processing throughput. Extraction functionality is controlled and configured with ATTR and ATTRELI. See Section 21.5.3.8.1, “Attribute Register (ATTR),” and Section 21.5.3.8.2, “Attribute Extract Length and Extract Index Register (ATTRELI),” for specific register information. 21.7 Application Information 21.7.1 Interface Mode Configuration This section describes how to configure the FEC in the MII mode. The pinout, the data registers that must be initialized, as well as speed selection options are described. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 21-69 10/100 Fast Ethernet Controller 21.7.1.1 MII Interface Mode Table 21-50 shows the signal configuration needed for MII interface mode Table 21-50. MII Interface Mode Signal Configuration M II INTERFACE Frequency [MHz] 25 Voltage[V] 3.3 Signals FEC_TX_CLK FEC_TxD[0] FEC_TxD[1] FEC_TxD[2] FEC_TxD[3] FEC_TX_EN FEC_TX_ER FEC_RX_CLK FEC_RxD[0] FEC_RxD[1] FEC_RxD[2] FEC_RxD[3] FEC_RX_DV FEC_RX_ER FEC_COL FEC_CRS I/O I O O O O O O I I I I I I I I I # of signals 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 sum 16 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 21-70 Freescale Semiconductor Application Information Table 21-51 describes the register initializations required to reconfigure the PHY to MII mode following intial auto-negotiation. Table 21-51. MII Mode Register Initialization Steps Set Soft_Reset, MACCFG1[1000_0000_0000_0000_0000_0000_0000_0000] Clear Soft_Reset, MACCFG1[0000_0000_0000_0000_0000_0000_0000_0000] Initialize MACCFG2, for MII, half duplex operation. Set I/F Mode bit to nibble mode, MACCFG2[0000_0000_0000_0000_0111_0001_0000_0100] (This example has Full Duplex = 0, Preamble count = 7, PAD/CRC append = 1) Initialize MAC Station Address, MACSTNADDR2[0110_0000_0000_0010_0000_0000_0000_0000] Set station address to 02_60_8C_87_65_43, for example. (Note that PHY configuration register addresses are not necessarily consistent in all PHYs.) Initialize MAC Station Address, MACSTNADDR1[0100_0011_0110_0101_1000_0111_1000_1100] Set station address to 02_60_8C_87_65_43, for example. (Note that PHY configuration register addresses are not necessarily consistent in all PHYs.) Check auto-negotiation attributes in the PHY as necessary. Clear IEVENT register, IEVENT[0000_0000_0000_0000_0000_0000_0000_0000] Initialize IMASK, (Optional) IMASK[0000_0000_0000_0000_0000_0000_0000_0000] Initialize IADDRn, (Optional) IADDRn[0000_0000_0000_0000_0000_0000_0000_0000] Initialize GADDRn, (Optional) GADDRn[0000_0000_0000_0000_0000_0000_0000_0000] Initialize RCTRL, (Optional) RCTRL[0000_0000_0000_0000_0000_0000_0000_0000] Initialize DMACTRL, (Optional) DMACTRL[0000_0000_0000_0000_0000_0000_0000_0000] Initialize FIFO_PAUSE_CTRL, FIFO_PAUSE_CTRL[0000_0000_0000_0000_0000_0000_0000_0010] Initialize (Empty) Transmit Descriptor ring and fill buffers with Data. Initialize TBASE, TBASE[LLLL_LLLL_LLLL_LLLL_LLLL_LLLL_LLLL_L000] Initialize (Empty) Receive Descriptor ring and fill with empty buffers. Initialize RBASE, RBASE[LLLL_LLLL_LLLL_LLLL_LLLL_LLLL_LLLL_L000] Enable Rx and Tx, MACCFG1[0000_0000_0000_0000_0000_0000_0000_0101] MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor 21-71 10/100 Fast Ethernet Controller MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 21-72 Freescale Semiconductor Appendix A Revision History This appendix provides a list of the major differences between revisions of the MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Revision 0 and MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Revision 1. A.1 Changes From Revision 0 to Revision 1 Major changes to the MPC8540 PowerQUICC III Integrated Host Processor Reference Manual between Revision 0 to Revision 1 are as follows: Section, Page Chapters 4, 16, 18, 20 Changes Throughout all applicable chapters, correctly state the debug mode source ID (formerly documented to be visible on signals PCI_AD[63:59]) to be visible on signals PCI_AD[62:58]. 2.4, 2-18 2.4, 2-27 0x2_404C 0x2_4050– 0x2_4088 In Table 2-9, correct the reset value for the I2CDFSRR register to be 0x10. In Table 2-9, add the following rows under the TSEC1 FIFO Control and Status Registers section: FIFO_PAUSE_CTRL—FIFO pause control register Reserved R/W R 0x0000_0000 0x0000_0000 14.5.3.2.1/14-30 — 2.4, 2-27 2.4, 2-28 2.4, 2-28 2.4, 2-31 In Table 2-9, correct the reset value for the OSTBD register to be 0x0800_0000. In Table 2-9, correct the MACCFG1 register from having R/W, R access to R/W access. In Table 2-9, correct the reset value for the IFSTAT register to be 0x0000_0000. In Table 2-9, correct the CAR1 and CAR2 registers from having R access to R/W access. Also, remove the “Cleared on read” footnote designation from the CAR1 and CAR2 registers. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor A-1 Revision History 2.4, 2-32 0x2_604C 0x2_6050– 0x2_6088 In Table 2-9, add the following two rows under “FIFO Control and Status Registers”: FIFO_PAUSE_CTRL—FIFO pause control register Reserved R/W R 0x0000_0000 0x0000_0000 21.5.3.2.1/21-22 — 2.4, 2-35 2.4, 2-35 2.4, 2-47 2.4, 2-48 2.4, 2-42–46 4.3.1.3, 4-7 In Table 2-9, correct the ATTR and ATTRELI registers from having R access to R/W access. In Table 2-9, correct the offset for the Reserved addresses after the EOI register to be 0x4_00C0–0x4_0FF0. In Table 2-9, correct the reset value for the GPINDR register to be 0xnnnn_0000. In Table 2-9, correct the name of the register with offset 0xE_1094 from PMLCB5 to PMLCB8. In Table 2-9, correct the access designation for all RapidIO Registers listed as “Special” to read “R/W”. Replace the second sentence of the first paragraph to read: “The first instruction executed by the e500 core is always address 0xFFFF_FFFC, which must be a branch to an address within the 4-Kbyte boot page.” 4.4.3.3, 4-13 Insert the following sentence at the start of the first paragraph: “The first instruction executed by the e500 core is always address 0xFFFF_FFFC, which must be a branch to an address within the 4-Kbyte boot page.” 4.4.3.3, 4-14 4.4.4.1, 4-22 Add the following sentence to the end of the last paragraph: “If enabled, this translation only affects CPU accesses to 0xFFFF_Fnnn.” Append the end of the first paragraph of Section 4.4.4.1, with the following sentence: “Note that the divide-by-two CCB clock divider and the divide-by-n CCB clock divider, shown in Figure 4-6, are located in the DDR and local bus blocks, respectively.” MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 A-2 Freescale Semiconductor Changes From Revision 0 to Revision 1 4.4.4.1, 4-23 Replace the existing Figure 4-6 with the following figure (Note the addition of the CCB clock dividers prior to the DDR DLL and the LBC DLL): MPC8540 e500 Core cfg_core_pll[0:1] 2 Core PLL core_clk MSYNC_IN ÷2 cfg_sys_pll[0:3] 4 Device PLL CCB_clk DLL MSYNC_OUT 6 6 MCK[0:5] MCK[0:5] LSYNC_IN DDR Controller SYSCLK/ PCI_CLK ÷n CCB_clk to Rest of the Device LSYNC_OUT DLL LCLK0 LCLK1 LCLK2 LBC 4.4.4.4, 4-25 In Figure 4-8, replace the sentence: “Watchdog timer events based on one of the 64 TB bits selected by TCR[WP] concatenated with the EIS-defined TCR[WPEXT] (WP||WPEXT).” with the following sentence: “Watchdog timer events based on one of the 64 TB bits selected by concatenating TCR[WPEXT] with the EIS-defined TCR[WP] (WPEXT||WP).” 4.4.4.4, 4-25 In Figure 4-8, replace the sentence: “Fixed-interval timer events based on one of the 64 TB bits selected by TCR[FP] concatenated with the EIS-defined TCR[FPEXT] (FP||FPEXT).” with the following sentence: “Fixed-interval timer events based on one of the 64 TB bits selected by concatenating TCR[FPEXT] with the EIS-defined TCR[FP] (FPEXT||FP).” MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor A-3 Revision History 5.5, 5-13 Prior to Section 5.5.1, insert a new section, “Initial Instruction Fetch,” consisting of the following text (former Section 5.5.1 now becomes Section 5.5.2.): “The e500 core begins execution at fixed virtual address 0xFFFF_FFFC. The MMU has a default page translation which maps this to the identical physical address. So, the instruction at physical address 0xFFFF_FFFC must be a branch to another address within the 4-Kbyte boot page.” 5.6, 5-18 5.14, 5-33 Remove the registers IAC3 and IAC4 from Figure 5-6. In Table 5-8, replace the HID1 implementation description of HID1[RFXE] with the following: HID1[RFXE] controls whether assertion of core_fault_in causes a machine check interrupt. Assertion of core_fault_in can result from uncorrectable data error, such as an L2 multibit ECC error. It can also occur for a system error if logic on the integrated device signals a fault for nonfatal errors (read data is corrupt (or zero), but the transaction can complete without corrupting other system state, making it unnecessary to take a machine check (for example, a master abort of a PCI transaction). If RFXE is 0, and core_fault_in is asserted, any data on the CCB is dropped, either stalling the load/store unit and causing the e500 pipeline to stall until an interrupt occurs (typically generated by the programmable interrupt controller (PIC) in response to the fault) or allowing processingþto continue with the bad data until the interrupt occurs. Because core_fault_in cannot cause a machine check if RFXE is 0, it is critical that the system be configured to generate the appropriate interrupt, as described below. It is also possible to hang the processor (requiring a hard reset to recover) if a guardedþload hits in the L2 cache and gets an uncorrectable ECC error. Because of this, avoid defining memory as cacheable but guarded. If this combination is required, RFXE must be enabled, in which case an error causes both a machine check interrupt and an external interrupt when a bus fault condition is detected unless interrupts are masked for all sources of bus faults. If RFXE is 0, conditions that cause the assertion of core_fault_in cannot directly cause the e500 to generate a machine check; however, PowerQUICC III devices must be configured to detect and enable such conditions. The following describes how error bits should be configured: • ECM mapping errors: EEER[LAEE] must be set. See Section 8.2.1.4, “ECM Error Enable Register (EEER).” • L2 multiple-bit ECC errors: L2ERRDIS[MBECCDIS] must be cleared to ensure that error can be detected. L2ERRINTEN[MBECCINTEN] must be set. SeeSection 7.3.1.5, “L2 Error Registers.” • DDR multiple-bit ECC errors. ERR_DISABLE[MBED] and ERR_INT_EN[MBEE] must be zero and DDR_SDRAM_CFG[ECC_EN] must be one to ensure that an interrupt is generated. See Section 9.4.1, “Register Descriptions.” • PCI. The appropriate parity detect and master-abort bits in ERR_DR must be cleared and the corresponding enable bits in ERR_EN must be set to ensure that an interrupt is generated. Section 16.3.1.4, “PCI/X Error Management Registers.” • Local bus controller parity errors. LTEDR[PARD] must be cleared and LTEIR[PARI] must be set to ensure that an parity errors can generate an interrupt. See Section 13.3.1.11, “Transfer Error Check Disable Register (LTEDR),” and Section 13.3.1.12, “Transfer Error Interrupt Enable Register (LTEIR).” • RapidIO. PCR[CCE] must be set to ensure that an interrupt is generated due to a CRC error. See Section 17.3.2.1.2, “Port Configuration Register (PCR).” RFXE must also be set if software requires that code execution stop immediately when a bus fault occurs rather than continuing with the bad data until the interrupt arrives. Again, this results in both a machine check interrupt and an external interrupt when a bus fault is detected, unless all possible sources for bus fault have their interrupts masked.þThe machine check interrupt can then reenable normal interrupts and wait for the interrupt due to the fault to be received before returning from the machine check. 6.1.1, 6-2 IAC3 and IAC4 are not supported and should be removed from Figure 6-1. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 A-4 Freescale Semiconductor Changes From Revision 0 to Revision 1 6.6.1, 6-14 In Table 6-8, change the second sentence of the description for bits 32–33 of the TCR register to the following: WPEXT[0–3] || WP[0–1] = 0b00_0000 selects TBU[32] (the msb of the TB) WPEXT[0–3] || WP[0–1] = 0b11_1111 selects TBL[63] (the lsb of the TB) 6.6.1, 6-14 6.6.1, 6-15 In Table 6-8, change the description of TCR[WRC] as follows: 01 A second timeout is ignored, regardless of the value of MSR[ME]. In Table 6-8, change the second sentence of the description for bits 38–39 of the TCR register to the following: FPEXT[0–3] || FP[0–1] = 0b00_0000 selects TBU[32] (the msb of the TB) FPEXT[0–3] || FP[0–1] = 0b11_1111 selects TBL[63] (the lsb of the TB) 6.7.2.4, 6-22 6.10.2, 6-27 In Table 6-12, change the bit range “43–46” to “36–46”. In Table 6-17, replace the description of HID1[RFXE] with the following: Read fault exception enable. Controls whether assertion of core_fault_in causes a machine check interrupt. The assertion of core_fault_in can result from an L2 multibit ECC error. It can also occur for a system error if logic on the integrated device signals a fault for nonfatal errors (read data is corrupt (or zero), but the bus transaction can complete without corrupting other system state, making it unnecessary to take a machine check (for example, a master abort of a PCI transaction). 0 Assertion of core_fault_in cannot cause a machine check. RFXE should be left clear if an interrupt is to be reported by the integrated device through int or c_int for this condition. If RFXE = 0, it is important that the integrated device generates an interrupt if core_fault_in is asserted. If core_fault_in is asserted, any data on the CCB is dropped, stalling the load/store unit and eventually causing the e500 pipeline either to stall until an interrupt occurs (typically generated by the programmable interrupt controller (PIC) in response to the fault) or to continue processing with bad data untill the interrupt occurs. Because core_fault_in cannot cause a machine check, if RFXE is 0, it is critical that the system be configured to generate the appropriate interrupt. It is also possible to hang the processor (requiring a hard reset to recover) if a guarded load hits in the L2 cache and gets an uncorrectable ECC error. Because of this, avoid defining memory as cacheable but guarded. If this combination is required, RFXE must be enabled, in which case an error causes both a machine check interrupt and an external interrupt when a bus fault condition is detected unless interrupts are masked for all sources of bus faults, such as DRAM ECC errors, PCI parity errors, local bus parity errors, and others. RFXE must also be set if software requires that code execution stop immediately when a bus fault occurs rather than continuing with the bad data until the interrupt arrives. Again, this results in both a machine check interrupt and an external interrupt when a bus fault is detected, unless all possible sources for bus fault have their interrupts masked.þThe machine check interrupt can then reenable normal interrupts and wait for the interrupt due to the fault to be received before returning from the machine check. 1 A machine check can occur due to assertion of core_fault_in. If MSR[ME] = 1 and a fault is signaled, a machine check interrupt occurs. If MSR[ME] = 0 and a fault is signaled, a checkstop occurs. Note that if RFXE is set and another mechanism is configured to generate an interrupt in response to assertion of core_fault_in, the same event causes two interrupts, the machine check enabled by setting RFXE and the interrupt triggered by the on-chip peripheral or other block; therefore, RFXE should be set only if no other mechanism is configured to generate an interrupt for this case. Note that the L2 cache detects any assertion of core_fault_in and ensures that the L2 cache is not corrupted when data is dropped for this type of transaction. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor A-5 Revision History 6.11.3, 6-30 In Table 6-20, describe bits 34–38 as ‘Reserved’ as is properly expressed in Figure 6-35. Disregard descriptions associated with bits 34, 35, or 36 in Table 6-20. In Figure 6-38 and Table 6-22, describe bits 59 and 60 as “Reserved”. In Table 6-28, describe bits 52–56 as “Reserved for implementation-specific use”. Remove all references to IAC3 and IAC4 from this section. In Table 7-2, change the description for bit 1 from “Accesses to memory-mapped SRAM are unaffected...” to “Data in memory-mapped SRAM regions is unaffected...” In Figure 7-7 and Table 7-2, show bit 11 to be reserved. 6.12.2, 6-32 6.12.5.3, 6-37 6.13.3, 6-45 7.3.1.1, 7-8 7.3.1.1, 7-8 and 7-9 7.3.1.5.2, 7-16 and 7-17 In Table 7-12, change the bit descriptions for bits 27 and 31 of the L2ERRDET register as follows: Bit 27, TPARERR: Add the following text: “Note that if an L2 cache tag parity error occurs on an attempt to write a new line, the L2 cache must be flash invalidated. L2 functionality is not guaranteed if flash invalidation is not performed after a tag parity error.” Bit 31, L2CFGERR: Add the following text to the 1st line of description, “Reports inconsistencies between the L2SIZ, L2BLKSZ, and L2SRAM settings of the L2 control register (L2CTL).” 7.7.4, 7-25 Add the following note after the second paragraph: NOTE There is a scenario in which a lock clear operation appears to fail to clear a lock in the L2 cache. This occurs only when the attempt to set the lock results in a bus error (for example, PCI returns an error condition). Assume the following scenario: 1. The e500 attempts to set a lock in the L2 cache (by executing a dcbtls or icbtls instruction with CT = 1). The line is not already present in the cache, so it must be read from external memory. This read encounters an error which, depending on the chip configuration, will be reported to the core (probably as an interrupt). 2. At (or near) the same time, a cache external write to the same cache line is being mastered by the ECM. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 A-6 Freescale Semiconductor Changes From Revision 0 to Revision 1 3. Very soon after the cache external write, a transaction to clear the lock occurs. This can be caused by the processor executing a dcblc or icblc instruction with CT = 1, or by the ECM mastering a lock clear transaction. If this scenario occurs within a tight timing window, the cache line may unexpectedly remain locked at the end of the sequence. The interrupt handler may want to clear the erroneously remaining lock in this case. 7.9, 7-28 Add new Section 7.9.2, “Flash Invalidation of L2 Cache” as follows: “The L2 cache may be completely invalidated by setting the L2I bit of the L2 control register (L2CTL). Note that no data is lost in this process because the L2 cache is a write-through cache and contains no modified data. Flash invalidation of the cache is necessary when the cache is initially enabled and may be necessary to recover from some error conditions such as a tag parity error. The invalidation process requires several cycles to complete. The L2I bit remains set during this procedure and is then cleared automatically when the procedure is complete. The L2 cache controller issues retries for all transactions on the e500 core complex bus (CCB) while the flash invalidation process is in progress. Note that the contents of memory-mapped SRAM regions of the data array are unaffected by a flash invalidation of the L2 cache regions fo the array. 7.10, 7-32 Add new Section 7.10, “Initialization/Application Information,” with the following sections: Section 7.10.1, “Initialization,” with new subsections Section 7.9.1.1, “L2 Cache Initialization,” and Section 7.9.1.2, “Memory-Mapped SRAM Initialization.” 7.9.1.1 L2 Cache Initialization After power-on reset, the valid bits in the L2 cache status array are in random states. Therefore, it is necessary to perform a flash invalidate operation before using the array as an L2 cache. This is done by writing a 1 to the L2I bit of the L2 control register (L2CTL). This can be done before or simultaneously with the write that enables the L2 cache. That is, the L2E and L2I bits of L2CTL can be set simultaneously. The L2I bit clears automatically, so no further writes are necessary. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor A-7 Revision History 7.9.1.2 Memory-Mapped SRAM Initialization After power-on reset, the contents of the data and ECC arrays and are random, so all SRAM data must be initialized before it is read. If the cache is initialized by the core or any other device that uses sub-cacheline transactions, ECC error checking should be disabled during the initialization process to avoid false ECC errors generated during the read-modify-write process used for sub-cacheline writes to the SRAM array. This is done by setting the multi- and single-bit ECC error disable bits of the L2 error disable register (L2ERRDIS[MBECCDIS, SBECCDIS]). See Section 7.3.1.9.2, “Error Control and Capture Registers.” If the array is initialized by a DMA engine using cache-line writes, then ECC checking can remain enabled during the initialization process. Also, add new Section 7.10.2, “Managing Errors,” with subsections Section 7.10.2.1, “ECC Errors,” and Section 7.10.2.2, “Tag Parity Errors.” 7.10.2.1 ECC Errors An individual soft error that causes a single- or mulit-bit ECC error can be cleared from the L2 array simply by executing a dcbf instruction for the address captured in the L2ERRADDR register. This will invalidate the line in the L2 cache. When the load that caused the ECC error is performed again, the data will be re-allocated into the L2 with ECC bits set properly again. If the threshold for single bit errors set in the L2ERRCTL register is exceeded, then the L2 cache should be flash invalidated to clear out all single-bit errors. Note that no data is lost by executing dcbf instructions or flash invalidate operations because the L2 cache is write-through and contains no modified data. 7.10.2.2 Tag Parity Errors A tag parity error must be fixed by flash invalidating the L2 cache. Note that executing a dcbf instruction for the address that caused the error to be reported is not sufficient because a tag parity error is seen as an L2 miss and does not cause invalidation of the bad tag. Proper L2 operation cannot be MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 A-8 Freescale Semiconductor Changes From Revision 0 to Revision 1 guaranteed if an L2 tag parity error is not repaired by a flash invalidation of the entire array. 9.3.2.1, 9-6 In Table 9-3, change the Timing description for the MA[0:14] signals to read: “Assertion/Negation—The address is always driven when the memory controller is active. It is valid when a transaction is driven to DRAM (when MCSn is negated). High-impedance—When the memory controller is idle.” 9.3.2.1, 9-6 In Table 9-3, change the Timing description for the MCAS signal to read: “Assertion/Negation—Assertion and negation timing is directed by the values described in Section 9.4.1.3, “DDR SDRAM Timing Configuration 1 (TIMING_CFG_1).” High-impedance—MCAS is always driven unless the memory controller is idle.” 9.3.2.1, 9-7 In Table 9-3, change the Timing description for the MRAS signal to read: “Assertion/Negation—Timing is directed by the values described in Section 9.4.1.3, “DDR SDRAM Timing Configuration 1 (TIMING_CFG_1).” High-impedance—MRAS is always driven unless the memory controller is idle.” 9.3.2.1, 9-7 In Table 9-3, change the Timing description for the MWE signal to read: “Assertion/Negation—Similar timing as MRAS and MCAS. Used for write commands. High-impedance—MWE is always driven unless the memory controller is idle.” 9.3.2.2, 9-8 In Table 9-4, correct the “MCKE” signal name to read “MCKE[0:1]”. Also, change the first full sentence of the MCKE[0:1] signal description to read: “Two identical output signals (each hereafter referred to simply as MCKE) used as the clock enable to one or more SDRAMs.” 9.3.2.2, 9-8 In Table 9-4, change the Timing description for the MCKE[0:1] signal to read: “Assertion/Negation—Similar timing to MAn. High-impedance—Always driven.” MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor A-9 Revision History 9.4.1.1, 9-10 In Table 9-6, change the CSn_BNDS[EAn] field description to read: “Ending address for chip select (bank) n. This value is compared against the 8 msbs of the address.” 9.4.1.4, 9-13 9.4.1.5, 9-14 In Table 9-9, replace all instances of |CASLAT| with CASLAT . In Table 9-10, change the setting of the MBEE bit in the DDR_SDRAM_CFG[ECC_EN] description to read as follows: “ECC enable. Note that uncorrectable read errors may cause the assertion of core_fault_in, which causes the core to generate a machine check interrupt unless it is disabled (by clearing HID1[RFXE]). If RFXE is zero and this error occurs, ERR_DISABLE[MBED] must be zero and ERR_INT_EN[MBEE] and ECC_EN must be one to ensure an interrupt is generated. See Section 6.10.2, “Hardware Implementation-Dependent Register 1 (HID1). 0 1 No ECC errors are reported. No ECC interrupts are generated. ECC is enabled.” 9.4.1.7, 9-16 In Table 9-12, add the following last sentence to the DDR_SDRAM_INTERVAL[REFINT] field description: “Note that REFINT must be set to a non-zero value in order for the DDR to enter sleep mode. See Section 18.5.1.5.3, “Sleep Mode,” for additional details.” 9.4.1.15, 9-21 Change the setting of MBEE bit in the ERR_DISABLE[MBED] description to read as follows: Multiple-bit ECC error disable 0 Multiple-bit ECC errors are detected if DDR_SDRAM_CFG[ECC_EN] is set. They are reported if ERR_INT_EN[MBEE] is set. Note that uncorrectable read errors cause the assertion of core_fault_in, which causes the core to generate a machine check interrupt, unless it is disabled (by clearing HID1[RFXE]). If RFXE is zero and this error occurs, MBED must be zero and ERR_INT_EN[MBEE] and ECC_EN must be one to ensure that an interrupt is generated. Multiple-bit ECC errors are not detected or reported. 1 9.4.1.16, 9-21 Change the setting of the MBEE bit in the ERR_INT_EN[MBEE] description to read as follows: “Multiple-bit ECC error interrupt enable. Note that uncorrectable read errors may cause the assertion of core_fault_in, which causes the core to MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 A-10 Freescale Semiconductor Changes From Revision 0 to Revision 1 generate a machine check interrupt, unless it is disabled (by clearing HID1[RFXE]). If RFXE is zero and this error occurs, ERR_DISABLE[MBED] must be zero and MBEE and DDR_SDRAM_CFG[ECC_EN] must be set to ensure that an interrupt is generated. For more information, see Section 6.10.2, “Hardware Implementation-Dependent Register 1 (HID1).” 0 1 9.5, 9-25 Request from Master Address from Master Multiple-bit ECC errors cannot generate interrupts. Multiple-bit ECC errors generate interrupts.” Replace Figure 9-21 with the figure below: DDR SDRAM Memory Array: MA[0:14] MBAA[0:1] Debug Signals: MSRCID[0:4] MDVAL DDR SDRAM Memory Control: MCS[0:3] MCAS MRAS MWE MDM[0:8] MCKE[0:1] Data Strobes: MDQS[0:8] Address Decode Input Staging Queue Address Control Physical Bank, Logical Bank, Row ECM Row Open Table Row Open SDRAM Control EN To Error Management Error Signals ECC DQ Delay Chain Data from SDRAM DQ POS FIFO NEG FIFO RMW ECC DQ EN Clocks: MCK[0:5] MCK[0:5] MSYNC_OUT MSYNC_IN Data Signals: MDQ[0:63] MECC[0:7] Data from Master DLL 9.5, 9-25 9.5.1.1, 9-30 9.5.1.1, 9-30 Remove the last sentence of the fourth paragraph. In the first two columns of Table 9-26, change “Mbytes” to “Mbits” and “Gbytes” to “Gbits”. Delete the second-to-last paragraph of this sentence that starts with the words, “If a disabled bank...” MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor A-11 Revision History 9.5.4–9.5.7, 9-35–9-39 In all burst examples, re-number the data beats in the figures to be D0, D1, D2, D3, D0, D1, D2, D3. 9.5.4, 9-35 In the last paragraph, change the phrase, “see Figure 9-26 for a single-beat read operation,” to read “see Figure 9-26 for a back-to-back burst read operation,”. Change the title of Figure 9-28 to “DDR SDRAM Burst Write Timing—ACTTORW = 4”. Change the last sentence of this section to read “Figure 9-31 shows the registered DDR SDRAM DIMM back-to-back burst write timing.” Replace the existing Figure 9-32 with the figure below: 0 1 2 3 4 5 6 7 8 9 10 11 12 9.5.4, 9-36 9.5.6, 9-38 9.5.7, 9-39 SDRAM Clock MCS ACTTORW MRAS MCAS MA[13:0] ROW COL COL MWE MDQ[0:63] MDQS MDM[0:7] MDQ[0:63] MDQS MDM[0:7] 00 00 D0 D1 D2 D3 D4 D5 D6 D7 D0 D1 D2 D3 D4 D5 D6 D7 1/4 Delay 1/2 Delay 9.5.8, 9-40 In the paragraph preceding Section 9.5.8.1, change the reference to “Two sets of auto refresh commands...” to read “Three sets of auto refresh commands...”. Also, change the reference to “...commands are also staggered in two groups...” to read “... commands are also staggered in three groups...” MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 A-12 Freescale Semiconductor Changes From Revision 0 to Revision 1 9.5.8.1, 9-40 9.5.8.2, 9-41 In Figure 9-33, remove the annotation of “0 or 3” in clock 12. After the second paragraph, add the following paragraph: “All open pages are precharged before self refresh mode is entered.” Also, correct the last sentence of the fifth paragraph to read: “This mode is controlled with DDR_SDRAM_CFG[DYN_PWR].” 9.5.9, 9-43 Replace the third and fourth sentences of this section with the following: “If ECC is enabled for a sub-doubleword write transaction, a full read-modify-write is performed to properly update ECC bits. If ECC is disabled then no read-modify-write is required for sub-doubleword writes, and the data masks (MDM[0:8]) are used to prevent writing unwanted data to SDRAM.” 9.5.11, 9-44 Add the following sentences before the second-to-last sentence of the second paragraph: “This read-modify-write operation is performed as an atomic transaction in the DDR controller. The write command is then issued 3-5 memory clocks after the completion of the read, depending on various system parameters.” 9.5.12, 9-46 9.6.1, 9-48 In the last paragraph, delete the three sentences that start and end with, “For all memory select errors...” and “...to show the transaction is not real.” In the second sentence, replace “after the DLL has locked” with the following: “after the memory clocks are stable (that is, the DLL has locked or initialization is complete of all clock related configuration registers).” Replace the last sentence of the first paragraph with the following: “After MEM_EN has been set, the DDR memory controller automatically performs the JEDEC-compliant initialization sequence to initialize memories according to the information in the SDMODE and ESDMODE fields of the DDR_SDRAM_MODE register. The initialization sequence is as follows:” After the numbered list, add the following sentence: “Note that the BA0 and BA1 bits are automatically driven appropriately during the MODE REGISTER SET commands.” Then the final sentence should be changed to, “After this automatic initialization is complete the memory array is ready for access and the memory controller begins processing memory transactions as they arrive.” MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor A-13 Revision History 10.2.2, 10-7 10.3.5.3, 10-29 In Table 10-5, delete the last sentence in the Timing/Negation section for IRQ[0:11] signals about edge-sensitive interrupts. Append the last sentence of the first paragraph to read: “Reading the corresponding message register or writing a 1 to a status bit clears the corresponding message interrupt and the status bit.” 10.3.7, 10-36 10.4.1, 10-41 In Table 10-35, correct the who am I register name from “WHOAMI0” to “WHOAMI.” Add the following new paragraph after the first paragraph: “Note that the IPR, IS, and IRR are internal registers that are not accessible to the programmer.” MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 A-14 Freescale Semiconductor Changes From Revision 0 to Revision 1 10.4.1, 10-41 Replace the existing Figure 10-37, with the figure below: int To Processor Core Interrupt Router Processor’s Task Priority Register Pipeline Latch Interrupt Request Register (IRR) Highest Priority Interrupt IACK Interrupt Selector (IS) PIC Configuration and Status Registers EOI End of Interrupt In-Service Register (ISR) Activity Bit Mask Register Interrupt Pending Register (IPR) Global Timers (4) Critical Interrupt xIDR[CI] Internal, External, and Message Interrupts Inter-Processor cint To Processor Core External Pin xIDR[EP] IRQ_OUT Note: All signal lines represent buses except for int, cint, and IRQ_OUT. The behavior of the PIC unit is not defined if both the EP and CI bits of the same interrupt destination register are set. 11.3, 11-4 In Table 11-3, correct the I2CDFSRR reset value to be 0x10. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor A-15 Revision History 11.4.5.2, 11-18 Insert the following two sentences after the eighth sentence of the first paragraph: “The boot sequencer expects the address offset to be a 32-bit (word) offset, that is, the 2 low-order bits are not included in the boot sequencer command. For example, to access LAWBAR0 (byte offset of 0x00C08), the boot sequencer ADDR[0:17] should be set to 0x00302.” 11.4.5.2, 11-18 Insert the following sentence after the polynomial used for CRC calculation in the fourth paragraph: “CRC values are calculated using the above polynomial with a start value of 0xFFFF_FFFF and an XOR with 0x0000_0000.” 12.3.1.9, 12-15 In Table 12-16, change the second sentence of the asserted (“1”) description for the ULSR[BI] to read: “A break condition is expected to last at least two character lengths and a new character is not loaded until SIN returns to the mark state (logic 1) and a valid START is detected.” Chapter 13 13.1.3.1, 13-3 Correct all instances of “UPWAIT” to “LUPWAIT.” Change the third sentence of the first paragraph to read: “In addition to establishing the frequency of the external local bus clock, CLKDIV also affects the resolution of signal timing shifts in GPCM mode, and the interpretation of UPM array words in UPM mode.” 13.1.3.1, 13-4 13.2, 13-5 In the last sentence, change “LCLK[0:3]” to “LCLK[0:2].” Delete the Reset State (outputs) column (and references to it) of Table 13-1. 13.2, 13-6, 13-7, and 13-9 In Table 13-2, change all instances of “RAS” to “RAS” and all instances of CAS to CAS. 13.3.1.2.3, 13-16 and 17 In Table 13-7, Table 13-8, Figure 13-4, and Figure 13-5, change bits 17–18 to reserved. 13.4.4, 13-61 Insert the following footnote associated with the word ‘signals’ in the second sentence of the section: “If the LGPL4/LGTA/LUPWAIT/LPBSE signal is used as both an input and an output, a weak pullup is required. Refer to the hardware specification for details regarding termination options.” MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 A-16 Freescale Semiconductor Changes From Revision 0 to Revision 1 13.4.4.2, 13-65 Add the following third paragraph: “Note that the UPM memory region must be cache-inhibited or write-through (the MMU page must have the I or W bit set) during the time that the UPM array is being written. If the memory is to be cacheable and/or copyback, the MMU must be set accordingly after the UPM array is initialized.” 13.4.7.3, 13-55 Chapter 14, 21 14.2, 14-7 14.2, 14-7 In the section title, change “CAS” to “CAS”. For consistency, change all instances of “HFE”, “Huge_Frame”, and “HUGE FRAME” to “Huge Frame”. In the first bullet item of the TSEC features list, change “820.3x” to “802.3x.” Change the fifth element of the second bullet item in the TSEC features list to read: “– 10/100 Mbps RMII” In Table 14-3, add the following as the first row under the TSEC1 FIFO Control and Status Registers section: FIFO_PAUSE_CTRL—FIFO pause control register R/W 0x0000_0000 14.5.3.2.1/A-19 14.5.2, 14-14 0x2_404C 14.5.2, 14-15 14.5.2, 14-16 14.5.2, 14-16 14.5.2, 14-16 14.5.2, 14-18 14.5.2, 14-18 14.5.2, 14-19 14.5.3.1.1, 14-20 In Table 14-3, correct the reset value for the OSTBD register to 0x0800_0000. In Table 14-3, correct the MACCFG1 register from having R/W, R access to R/W access. In Table 14-3, correct the IFSTAT register from having R access to R/W access. In Table 14-3, change the reset value for the TSEC IFSTAT register to be 0x0000_0000. In Table 14-3, correct the CAR1 and CAR2 registers from having R access to R/W access. In Table 14-3, remove the incorrect "Cleared on read" footnote designation from the CAR1 and CAR2 registers. In Table 14-3, correct the ATTR and ATTRELI registers from having R access to R/W access. In Table 14-4, correct the IEVENT[RXC] bitfield description to read as follows: “Receive control interrupt. A control frame was received. If MACCFG1[Rx_Flow] is set, a pause operation is performed lasting for the MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor A-17 Revision History duration specified in the received pause control frame and beginning when the frame was received. 0Control frame not received 1 14.5.3.1.1, 14-20 Control frame received” Replace the last two sentences of the first paragraph with: “Clearing the IEVENT bit clears the interrupt signal. The bit in the IEVENT register is cleared if a 1 is written to that bit position. A write of 0 has no effect.” 14.5.3.1.1, 14-21 In Table 14-4, replace the EBERR field description with: “Ethernet bus error. This bit indicates that a system bus error for a memory read ocurred while a DMA transaction was underway. If the EBERR is set while transmission is in progress, the DMA stops sending data to the Tx FIFO which eventually causes an underrun error (XFUN) and TSTAT[THLT] is set. If the EBERR is set while receiving a frame, the DMA discards the frame and RSTAT[QHLT] is set. 0 1 14.5.3.1.2, 14-22 No system bus error occurred. System bus error occurred.” Replace the last sentence in the first paragraph with: “The interrupt signal can be cleared by clearing the corresponding IEVENT bit.” 14.5.3.1.3, 14-24 Undocument bitfields TXEDIS, CRL/XDADIS, and XFUNDIS. 14.5.3.1.7, 14-27 and 28 Remove bitfield DMACTRL[TOD] from Figure 14-12 and Table 14-10. This bit position (29) is now reserved. 14.5.3.1.7, 14-28 In Table 14-10, add the following paragraph after the second paragraph of the WOP field description: “A buffer descriptor is considered not ready when its TxBD[Ready] field is clear or its TxBD[Data Length] field is zero. It is considered ready when both TxBD[Ready] is set and TxBD[Data Length] is non-zero.” 14.5.3.1.7, 14-28 In Table 14-10, change the second sentence of the third paragraph (which becomes the fourth paragraph after incorporating of the above errata) of the WOP field description to read: “If the TxBD becomes ready, TSEC continues to process the frame.” MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 A-18 Freescale Semiconductor Changes From Revision 0 to Revision 1 14.5.3.2.1, 14-30 Insert the following section as the first subsection of Section 14.5.3.2: 14.5.3.2.1 FIFO Pause Control Register (FIFO_PAUSE_CTRL) FIFO_PAUSE_CTRL, shown in Figure 14-20, is writable by the user to configure the properties of the TSEC FIFO. 0 15 R W Reset 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0000_0000_0000_0000 16 29 30 31 R W Reset Offset 0 0 0 0 0 0 0 0 0 0 0 0 0 0 TFC_PAUSE_EN 0 0000_0000_0000_0000 TSEC1:0x2_404C; TSEC2:0x2_504C Figure 14-14 FIFO_PAUSE_CTRL Register Definition Table 14-20 describes the fields of the FIFO_PAUSE_CTRL register. Table 14-12 FIFO_PAUSE_CTRL Field Descriptions Bits 0–29 30 Name — Reserved Description TFC_PAUSE_EN TFC_PAUSE enable. This bit enables the ability to transmit a pause control frame by setting the TCTRL[TFC_PAUSE] bit. This bit is cleared at reset. 0 Pause control frame transmission disabled. 1 Pause control frame transmission enabled. — Reserved 31 14.5.3.3.1, 14-32 In Table 14-15, replace the bitfield description of TCTRL[TFC_PAUSE] with the following: “Transmit flow control pause frame. Use this bit to transmit a PAUSE frame. To transmit a flow control pause frame, first set FIFO_PAUSE_CTRL[TFC_PAUSE_EN]. Next, set MACCFG1[GTS]. If TFC_PAUSE is then set, the MAC stops transmission of data frames after the current transmission completes. The GTSC interrupt in the IEVENT register is asserted. With transmission of data frames stopped, the MAC transmits a MAC control PAUSE frame with the duration value obtained from the PTV register. The TXC interrupt occurs after sending the control MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor A-19 Revision History pause frame. Next, the MAC clears TFC_PAUSE and resumes transmitting data frames. Note that if the transmitter is paused due to user assertion of GTS or reception of a PAUSE frame, the MAC may still transmit a MAC control PAUSE frame. 0 1 14.5.3.3.6, 14-35 No outstanding pause frame transmission request. Pause frame transmission requested.” Change the last sentence in the first paragraph to read: “Although not necessary in most applications, the user can modify this register when the transmitter has been gracefully stopped or halted, as indicated by IEVENT[GTSC] or TSTAT[THLT].” 14.5.3.3.8, 14-38 In Table 14-22, add the following sentence to the OSTBDLEN field description: “This field must be greater than zero in order for a transfer to take place.” In Table 14-25, change the description for the RSTAT[QHLT] field to read: “RxBD queue is halted. When IEVENT[BSY] or IEVENT[EBERR] is set during reception of a packet, RSTAT[QHLT] is also set. In order to begin receiving packets again, the user must clear RSTAT[QHLT]. This bit is set whenever the TSEC reads an RxBD with its Empty field cleared or encounters a system bus error while processing an RX packet. It is a hardware-initiated stop indication (DMA_CTRL[GRS] being set by the user does not cause this bit to be set.). The current frame and all other frames directed to the halted queue are discarded. A write with a value of 1 re-enables the queue for receiving. 0 1 RxBD queue is enabled for Ethernet reception. (That is, it is not halted.) All Ethernet controller receive activity to RxBD queue is halted.” 14.5.3.4.2, 14-40 14.5.3.4.2, 14-39 0 Replace Figure 14-27 with the following figure: 7 8 9 31 R00000000 QHLT W Reset Offset 0000000 000 00 000 000 00 000 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_4304; TSEC2:0x2_5304 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 A-20 Freescale Semiconductor Changes From Revision 0 to Revision 1 14.5.3.6.1, 14-47 0 1 Replace Figure 14-34 with the following figure: 11 12 13 14 15 R Soft_Reset W Reset 16 000000 0 00 0 0 Reset Rx MC Reset Tx MC Reset Rx Fun Reset Tx Fun 0000_0000_0000_0000 22 23 24 25 26 27 28 29 30 31 R W Reset Offset 0 000000 Loop Back 00 Rx_Flow Tx_Flow Sync’d Rx EN Rx_EN Sync’d Tx EN Tx_EN 0000_0000_0000_0000 TSEC1:0x2_4500; TSEC2:0x2_5500 14.5.3.6.1, 14-48 In Table 14-32, add the following sentence to the MACCFG1[Rx_EN] field description: “If set, prior to clearing this bit, set DMACTRL[GRS] then confirm subsequent occurrence of the graceful receive stop interrupt (IEVENT[GRSC] is set).” 14.5.3.6.1, 14-48 In Table 14-32, add the following sentence to the MACCFG1[Tx_EN] field description: “If set, prior to clearing this bit, set DMACTRL[GTS] then confirm subsequent occurrence of the graceful transmit stop interrupt (IEVENT[GTSC] is set).” 14.5.3.6.6, 14-52 In Table 14-37, clarify the bit description of the Mgmt Clock Select field of the MIIMCFG register so that instead of concluding with “... divided by eight.” the third sentence concludes with “... divided first by eight and then futher divided by the following value:”. Delete the first sentence and replace Figure 14-45 with the following figure (exposing the Link Fail status field): 14.5.3.6.12, 14-55 0 21 22 23 28 31 R W Reset Offset 000000000000000000000 0 Excess Defer 0 0000 Link 00 Fail 0 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_453C; TSEC2:0x2_553C MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor A-21 Revision History 14.5.3.6.12, 14-56 23–27 28 — Link Fail In Table 14-43, replace the last row with the following rows: Reserved Link Fail. This bit indicates the status of signal detection. 0 The100X module has detected a “signal detect” for longer than 330 mS. 1 The100X module has detected a “signal detect” for less than 330 mS or not at all. Reserved 29–31 — 14.5.3.6.12, 14-55 14.5.3.7.44, 14-79 14.5.3.7.45, 14-80 14.5.3.7.47, 14-83 14.5.3.8.1, 14-84 In Figure 14-45, correct the reset value for the IFSTAT register to be 0x0000_0000. Correct the first sentence to read: "Carry register bits are cleared when written with a one." Correct the first sentence to read: "Carry register bits are cleared when written with a one." In Figure 14-94, change bit 30 to be “Reserved”. Change the third sentence of the first paragraph to read: “When the DA field of a receive frame is processed through a 32-bit CRC generator, the 8 high-order bits (0–7) of the CRC remainder are mapped to one of the 256 entries.” 14.5.3.8.1, 14-85 In Table 14-93, add the following sentence to the IADDRn field description: “For instance, the MSB of IADDR0 correlates to entry 0 and the LSB of IADDR7 correlates to entry 255.” 14.5.3.8.2, 14-85 In Table 14-94, add the following sentence to the GADDRn field description: “For instance, the MSB of GADDR0 correlates to entry 0 and the LSB of GADDR7 correlates to entry 255.” 14.5.3.9.2, 14-87 Replace Figure 14-98 with the following figure to correspond to Table 14-96: 15 16 17 18 31 0 1 2 R0 0 EL W RESET Offset 00 EI 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_4BFC; TSEC2:0x2_5BFC MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 A-22 Freescale Semiconductor Changes From Revision 0 to Revision 1 14.6.2.2, 14-108 Replace the existing procedure list with the following list: “Following is a procedure to gracefully reset and reconfigure the MAC: 1. Set GTS bit in DMACTRL register 2. Poll GTSC bit in IEVENT register until detected as set 3. Clear both Rx_EN and Tx_EN bits in MACCFG1 4. Set GRS bit in DMACTRL register 5. Poll GRSC bit in IEVENT register until detected as set 6. Set Soft_Reset bit in MACCFG1 register 7. Clear Soft_Reset bit in MACCFG1 register 8. Load TBASE with new TxBD pointer 9. Load RBASE with new RxBD pointer 10. Set up other MAC registers (MACCFG1, MAXFRM, and so on) 11. Set WWR, WOP, TOD bits in DMACTRL register 12. Clear THLT bit in TSTAT register and QHLT bit in RSTAT register by writing 1 to these bits 13. Clear GRS/GTS bits in DMACTRL (do not change other bits) 14. Enable Tx_EN/Rx_EN in MACCFG1 register” 14.6.2.2, 14-108 Insert the following step between steps 3 and (formerly) 4: "4. Wait for a period of 9.6 Kbytes worth of data on the interface (~8ms worst case)." Correct step 11 of the graceful reset procedure to read: "11. Set WWR and WOP bits in DMACTRL register" Replace the second paragraph with the following: "If the user has a frame ready to transmit, a transmit-on-demand function may be emulated while in polling mode by using the graceful-transmit-stop feature. First, clear the IMASK[GTSCEN] bit to mask the graceful-transmit-stop complete interrupt. Next set, then immediately clear the DMACTRL[GTS] bit. Clear the resulting IEVENT[GTSC] bit. Finally, the IMASK[GTSCEN] bit may be set once again." Insert the following sentence between the third and (formerly) fourth sentences of the sixth paragraph: "The pause duration is defined by the received pause control frame and begins when the frame was first received." 14.6.2.2, 14-109 14.6.2.3, 14-109 14.6.2.3, 14-110 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor A-23 Revision History 14.6.2.6.2, 14-114 In the first paragraph, replace the third and fourth sentences with the following: “The eight high-order bits of a cyclic redundancy check (CRC) checksum are used to index into the hash table.” 14.6.2.6.3, 14-114 Add the following section after Section 14.6.2.6.2: 14.6.2.6.3 CRC Computation Examples There are many algorithms for calculating the CRC value of a number. Refer to the RFC 3309 standard, which can be found at http://www.faqs.org/rfcs/rfc3309.html, to compute the CRC value for the purposes of TSEC. The RFC 3309 algorithm uses the following polynomial to calculate the CRC value: x32+x26+x23+x22+x16+x12+x11+x10+x8+x7+x5+x4+x2+x1+x0 or 0x04c11db7. Given a destination MAC address of DA=01000CCCCCCC, the algorithm results in a CRC remainder value of 0xA29F4BBC. Bit-reversing the low-order byte of the CRC value (0xBC) yields BR_CRC = 0x3D = 0b00111101 The high-order 3-bits of the new BR_CRC value are used to select which 32-bit register (of the 8) to use. This example maps the DA to register 1. High-order 3 bits of BR_CRC: HO_CRC = 0b001 = 1 The low-order 5 bits are used to select which bit to set in the given register (with a value of 0 setting 0x8000_0000 and 31 setting 0x0000_0001). Therefore, the example DA maps to bit 29 of register 1. Low-order 5 bits of BR_CRC: LO_CRC = 0b11101 = 29 Therefore, GADDR1 is ORed with the value 0x0000_0004. Additional calculated examples follow: Example 1: •Destination MAC address: DA = 01005E000128 •CRC remainder value: CRC = 0x821D6CD3 •Bit-reversed least-significant byte of CRC value: BR_CRC = 0xCB = 0b11001011 •High-order 3 bits of BR_CRC: HO_CRC = 0b110 = 6 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 A-24 Freescale Semiconductor Changes From Revision 0 to Revision 1 •Low-order 5 bits of BR_CRC: LO_CRC = 0b01011 = 11 •GADDR6 = 0x0010_0000 Example 2: •Destination MAC address: DA = 0004F0604F10 •CRC remainder value: CRC = 0x1F5A66B5 •Bit-reversed least-significant byte of CRC value: BR_CRC = 0xAD = 0b10101101 •High-order 3 bits of BR_CRC: HO_CRC = 0b101 = 5 •Low-order 5 bits of BR_CRC: LO_CRC = 0b01101 = 13 •GADDR5 = 0x0004_0000 In Table 14-115, corrected the RXC description to read as follows: “Receive control: A control frame was received. As soon as the transmitter finishes sending the current frame, a pause operation is performed lasting for the duration specified in the received pause control frame and beginning when the frame was first received.” 14.6.2.11, 14-118 Insert the following paragraph after the first paragraph: “Programming note: When the TSEC encounters a halt condition (TSTAT[THLT] is set), it stops processing the frame at the current TxBD. The TSEC relies on the user to manage the buffer descriptor pointer, TBPTR, or the buffer descriptor queue before resuming transmissions. Once the TSEC resumes, it fetches the TxBD pointed to by TBPTR.” 14.6.3, 14-120 Between the second and third paragraph, add the following: “The status field of the BD is 16-bit field, as is the length field. The data buffer pointer is a 32-bit field. Therefore, the BDs should be accessed with the following C structure: typedef unsigned short uint_16; /* choose 16-bit native type */ typedef unsigned int uint_32; /* choose 32-bit native type */ typedef struct bd_struct { uint_16 flags; uint_16 length; uint_32 bufptr; };” 14.6.2.8, 14-116 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor A-25 Revision History 14.6.3.2, 14-124 In Table 14-120, change the LG bitfield description to read as follows: “Rx frame length violation. Written by TSEC. (Only valid if L is set.) A frame length greater than maximum frame length was recognized while MACCFG2[HUGE FRAME] was set. Note, if MACCFG2[HUGE FRAME] is cleared, the frame is truncated to the value programmed in the Maximum Frame Length register.” 14.6.4, 14-125 Change the first paragraph to read: “Some applications require the ability to identify selected portions of data within a frame’s data payload. This process is called extraction, although the data is not truly removed. Rather than literarily extracting a section of the data and copying it into a new memory location, the data is placed in the L2 cache. This allows the processor to quickly access critical frame information as soon as the processor is ready without having to first fetch the data from main memory. This results in substantial improvement in throughput and hence reduction in latency.” 14.7.1.1, 14-129 In Table 14-123, add the following step immediately following the optional DMACTRL initialization step: Initialize FIFO_PAUSE_CTRL, FIFO_PAUSE_CTRL[0000_0000_0000_0000_0000_0000_0000_0010] 14.7.1.2, 14-129 14.7.1.2, 14-129 14.7.1.2, 14-132 Change the title of Table 14-124 to read “GMII Interface Mode Signal Configuration.” In Table 14-124, remove the TX_CLK signal from the list of GMII interface signals and reduce the total number of signals listed at the bottom to 22. In Table 14-126, add the following step immediately following the optional DMACTRL initialization step: Initialize FIFO_PAUSE_CTRL, FIFO_PAUSE_CTRL[0000_0000_0000_0000_0000_0000_0000_0010] 14.7.1.3, 14-136 In Table 14-129, add the following step immediately following the optional DMACTRL initialization step: Initialize FIFO_PAUSE_CTRL, FIFO_PAUSE_CTRL[0000_0000_0000_0000_0000_0000_0000_0010] 14.7.1.4, 14-137 Change the title of Table 14–130 to read “RGMII Interface Mode Signal Configuration.” MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 A-26 Freescale Semiconductor Changes From Revision 0 to Revision 1 14.7.1.4, 14-140 In Table 14-132, add the following step immediately following the optional DMACTRL initialization step: Initialize FIFO_PAUSE_CTRL, FIFO_PAUSE_CTRL[0000_0000_0000_0000_0000_0000_0000_0010] 14.7.1.5, 14-144 In Table 14-135, add the following step immediately following the optional DMACTRL initialization step: Initialize FIFO_PAUSE_CTRL, FIFO_PAUSE_CTRL[0000_0000_0000_0000_0000_0000_0000_0010] Chapter 15 Correct the DMA signal enumeration for consistency as in the following example: “DMA0_DREQ” to “DMA_DREQ0” 15.2.2, 15-5 and 15.4.1.3, 15-32 In the third bullet of the list of signals defined for the external control interface, clarify that the falling edge of DMA_DREQ sets MRn[CS] and for DMA_DDONE with the following: “DMA_DDONE assertion indicates that the DMA engine has completed the transfer. SRn[CB] is clear. Note, however, that write data may still be queued at the target interface or in the process of transfer on an external interface.” In Table 15-4, designate the following memory locations as “Reserved”: 0x2_112C 0x2_1148–0x2_117C 0x2_11AC 0x2_11C8–0x2_11FC 0x2_122C 0x2_1248–0x2_127C 0x2_12AC 0x2_12C8–0x2_12FC 15.3.2.5.1, 15-17 Change the designation for the source address register when used for RapidIO maintenance transactions from “SARMn” to “SARn for RapidIO maintenance transactions,” reflecting the fact that any single SARn register can assume one of two formats. Change the designation for the destination address register when used for RapidIO maintenance transactions from “DARMn” to “DARn for RapidIO 15.3.1, 15-7 15.3.2.7.1, 15-20 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor A-27 Revision History maintenance transactions,” reflecting the fact that any single DARn register can assume one of two formats. 15.3.2.10, 15-22 15.3.2.11, 15-23 15.5.1.1.1, 15-27 In Figure15-17, correct the offset value for CLSDAR2 to 0x2_1234. In Figure15-18, correct the offset value for the NLSDARn–DMA 2 next list descriptor address register to 0x2_123C. Change step 1 in the sequence to read, “Poll the channel state (see Table 15-21 "Channel State Table") to confirm that the specific DMA channel is idle.” Change step 1 in the sequence to read, “Poll the channel state (see Table 15-21 "Channel State Table") to confirm that the specific DMA channel is idle.” Change step 2 in the sequence to read, “Poll the channel state (see Table 15-21 "Channel State Table") to confirm that the specific DMA channel is idle.” Change step 3 in the sequence to read, “Poll the channel state (see Table 15-21 "Channel State Table") to confirm that the specific DMA channel is idle.” Change step 2 in the sequence to read, “Poll the channel state (see Table 15-21 "Channel State Table") to confirm that the specific DMA channel is idle.” Change step 3 in the sequence to read, “Poll the channel state (see Table 15-21 "Channel State Table") to confirm that the specific DMA channel is idle.” Replace Figure 15-22 with the following, showing more flexible negation of DREQ: 15.5.1.1.2, 15-28 15.5.1.1.3, 15-29 15.5.1.1.4, 15-29 15.5.1.2.3, 15-30 15.5.1.2.4, 15-31 15.4.1.3, 15-32 CLOCK DREQ DACK DDONE EMP_EN Transfer Start Transfer Done Transfer Start Transfer Pause Transfer Restart Transfer In Progress MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 A-28 Freescale Semiconductor Changes From Revision 0 to Revision 1 15.5.1.2, 15-42 16.2, 16-12 Remove the last sentence of the second paragraph. In Table 16-2, add the following to the timing description for the PCI_PERR signal: “Note: If a parity error occurs on the last data beat of a PCI-X transaction with inbound data (outbound read with split completion data or inbound write), the MPC8540 asserts PERR longer than permitted by the PCI-X specification. This condition may cause a subsequent transaction to erroneously detect a parity error. The condition may be remedied by placing a stronger pull-up resistors on the PERR signal. For example: •1 Kohm for 133 MHz point-to-point operation •2 Kohm for 66 MHz operation •4 Kohm for 33 MHz operation. These values are stronger than the specification allows (5 Kohm minimum for pull-up resistors).” 16.3.1.2, 16-20 Add the following sentence to the end of the third paragraph: “Note that outbound translation windows must not overlap the configuration access registers.” 16.3.1.2.4, 16-22 In Figure 16–9, add a footnote to EN field (bit 0) that states, “ For POWAR0, translation is always enabled. The enable field (EN) may be read and written, but the value is ignored.” In Table 16–10, add the following to the description for bit 0 (EN), “Note that for POWAR0, translation is always enabled. The enable field (EN) may be read and written, but the value is ignored.” In Table 16–10, add the following to the description for bits 26–31 (OWS), “Also note that for POWAR0, setting OWS to less than 4 Gbytes causes addresses that miss in the other outbound windows to be aliased to the smaller address range defined by POWAR0[OWS] and POTAR0.” Insert the following between the second and third paragraphs: “Note that some errors are reported in two bits—one in the PCI/X error detect register (ERR_DR) and another in the PCI bus status register in the PCI/X configuration header. These bits must be cleared separately; that is, clearing one does not clear the other. For example, clearing the ERR_DR[Mstr abort error] does not clear the received master abort bit in the PCI bus status register. In these cases, both bits must be cleared before further error reporting can occur. Refer to Table 16-57 for PCI mode error actions and Table 16-64 for PCI-X mode error actions. Likewise, some 16.3.1.2.4, 16-22 16.3.1.2.4, 16–23 16.3.1.4, 16–27 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor A-29 Revision History errors are enabled by programming two bits—one in the PCI/X error enable register and another in the PCI bus command register in the PCI/X configuration header.” 16.3.1.4, 16–28 Insert the following between the fourth and fifth paragraphs: “If a data parity error occurs during an inbound configuration write access, the error is reported and captured. However, the erroneous data is written to the register specified in the transaction. Therefore, PCI data parity error recovery routines must include reinitialization of the PCI/X configuration register if the error occurred during a configuration write.” 16.3.1.4.7, 16–32 Add the following sentence before Figure 16-21: “Note that for inbound reads that have data parity errors, only the address (ERR_ADDR and ERR_EXT_ADDR) and attributes (ERR_ATTRIB) are captured. The data is not captured.” 16.3.1.4.7, 16–32 Add the following immediately before Figure 16-21: “Also note that PCI-X split completion error messages for outbound DWORD writes are incorrectly reported in the PCI/X error data low capture register (ERR_DL).” 16.3.1.4.8, 16-33 Add the following sentence before Figure 16-22: “Note that for inbound reads that have data parity errors, only the address (ERR_ADDR and ERR_EXT_ADDR) and attributes (ERR_ATTRIB) are captured. The data is not captured.” 16.3.2, 16-35 Insert the following after Figure 16-26: “Note that software must be restricted from accessing any reserved or undefined register space on the MPC8540; doing so may hang the device.” 16.3.2.3, 16-37 In Table 16-27, add the following sentence to the description of the Bus master bit field: “Note that the Bus master bit in the MPC8540’s PCI bus command register must be set before any outbound configuration access is attempted.” 16.3.2.4, 16–37 Insert the following between the first and second paragraphs: “Note that some errors are reported in two bits—one in the PCI bus status register and another in the PCI/X error detect register (ERR_DR). These bits must be cleared separately; that is, clearing one does not clear the other. For example, clearing the ERR_DR[Mstr abort error] does not clear the received master abort bit in the PCI bus status register. In these cases, both bits must be cleared before further error reporting can occur. Refer to Table 16-57 for PCI mode error actions and Table 16-64 for PCI-X mode MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 A-30 Freescale Semiconductor Changes From Revision 0 to Revision 1 error actions. Likewise, some errors are enabled by programming two bits—one in the PCI bus command register and another in the PCI/X error enable register (ERR_EN).” 16.3.2.9, 16-40 16.3.2.19, 16-47 In Figure 16-35, change the reset value from “0000_0000” to “0000_0000 (PCI); 0000_1000 (PCI-X)”. In Table 16-46, change the description for bit 0 to read as follows: “PCI agent/host. Read-only. Indicates the reset value of the PCI host/agent configuration signal, LWE3. 0 1 16.4.1, 16-51 PCI interface is in host mode PCI interface is in agent mode” Replace the second and third sentences of the third paragraph with the following: “The state of the reset signal is reflected in bit 15 (read-only) of the PCI bus arbitration control register (PBACR[PAD]). Note that PAD is the inverse of the arbiter configuration signal; that is, when PAD = 0 the arbiter is enabled, and when PAD = 1 the arbiter is disabled.” 16.4.1.3, 16-53 Add the following to the end of the first paragraph: “Note that the PCI-X bus arbiter does not support the broken master lockout function.” 16.4.2.11.1, 16-69 Insert the following between the first and second sentences in the first paragraph: “Note that the Bus Master bit in the MPC8540’s PCI bus command register must be set before an outbound configuration access is attempted.” 16.4.2.11.2–16.4.2.11.3, 16-69–16-71 Replace the existing sections with the following: 16.4.2.11.2 Accessing the PCI Configuration Space in Host Mode To access the configuration space, a 32-bit value must be written to the PCI CFG_ADDR register that specifies the target PCI bus, the target device on that bus, and the configuration register to be accessed within that device. Note that the Bus Master bit in the MPC8540’s PCI bus command register must be set before an outbound configuration access is attempted. Device 0 on PCI/X bus 0 is the MPC8540 itself; thus, device 0, bus 0 is used to access the internal PCI/X configuration header. When the MPC8540 detects an access to PCI CFG_DATA, it checks the enable flag and the device number in the PCI CFG_ADDR register. If the MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor A-31 Revision History enable bit is set, and the device number is not 0b1_1111, the MPC8540 performs a configuration cycle translation function and runs a configuration-read or configuration-write transaction on the PCI bus. If the bus number corresponds to the local PCI bus (bus number = 0x00), the MPC8540 performs a type 0 configuration cycle translation. If the bus number indicates a remote PCI bus (that is, nonlocal), the MPC8540 performs a type 1 configuration cycle translation. The device number 0b1_1111 is used for performing interrupt-acknowledge and special-cycle transactions. See Section 16.4.2.12, “Other Bus Transactions,” for more information. See Section 16.3.1.1, “PCI/X Configuration Address Register (CFG_ADDR),” for details on PCI CFG_ADDR and Section 16.3.1.1.2, “PCI/X Configuration Data Register (CFG_DATA),” for details on PCI CFG_DATA. Note that because all PCI/X registers are intrinsically little-endian, in the following examples, the data in the configuration register is shown in little-endian order. PowerPC processor accesses to the PCI CFG_DATA register should use the load/store with byte-reversed instructions. External PCI masters that use the local address map to access configuration space do not need to reverse bytes since byte lane redirection from the little-endian PCI bus is performed internally. Example: Configuration sequence, 4-byte data read from the revision ID/standard programming interface/subclass code/class code registers at address offset 0x08 of the PCI/X configuration header (device 0 on the PCI/X bus 0 is the MPC8540 itself). Initial values: r0 contains 0x8000_0008 r1 contains CCSRBAR + 0x0_8000 (Address of PCI register) r2 contains CCSRBAR + 0x0_8004 (Address of PCI register) r3 contains 0xFFFF_FFFF Register at 0x08 contains 0x0B20_0002 (0x0B to 0x08) Code sequence: stw r0, 0 (r1) lwbrx r3, 0 (r2) Results: Address CCSRBAR + 0x0_8000 contains 0x8000_0008 Register r3 contains 0x0B20_0002 CFG_ADDR CFG_DATA MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 A-32 Freescale Semiconductor Changes From Revision 0 to Revision 1 Example: Configuration sequence, 4-byte data write to PCI/X register at address offset 0x14 of Device 1 on PCI/X bus 0. Initial values: r0 contains 0x8000_0814 r1 contains CCSRBAR + 0x0_8000 (Address of PCI register) r2 contains CCSRBAR + 0x0_8004 (Address of PCI register) r3 contains 0x1122_3344 Register at 0x14 contains 0xFFFF_FFFF (0x17 to 0x14) CFG_ADDR CFG_DATA Code sequence: stw r0, 0 (r1)// Update PCI CFG_ADDR register to point to //register offset 0x14 of device 1. stwbrx r3, 0 (r2) Results: Address CCSRBAR + 0x0_8000 contains 0x8000_0814 Register at 0x14 contains 0x1122_3344 (0x17 to 0x14) Example: Configuration sequence, 2-byte data write to PCI register at address offset 0x1C of Device 1 on PCI/X bus 0. Initial values: r0 contains 0x8000_081C r1 contains CCSRBAR + 0x0_8000 r2 contains CCSRBAR + 0x0_8004 r3 contains 0xDDCC_BBAA Register at 0x1C contains 0xFFFF_FFFF (0x1F to 0x1C) Code sequence: stw r0, 0 (r1) sthbrx r3, 0 (r2) Results: Address CCSRBAR + 0x0_8000 contains 0x8000_081C Register at 0x1C contains 0xFFFF_BBAA (0x1F to 0x1C) 16.4.2.11.3 PCI Configuration in Agent and Agent Lock Modes In general, agents should not access the configuration space of other external PCI devices. Configuration of agents is a function usually reserved for the host. When the MPC8540 is in agent mode, it responds to remote host-generated PCI/X configuration cycles. This occurs when a configuration command is decoded along with the IDSEL input signal being asserted. When the MPC8540 is in agent lock mode, it retries all externally-generated PCI/X configuration cycles until the ACL bit in the PCI bus function register (0x44) is set. See Section 16.5.1, “Power-On Reset Configuration Modes,” for more information. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor A-33 Revision History In either agent or agent lock mode, access to the internal PCI/X configuration header by the processor core is handled as described in Section 16.4.2.11.2, “Accessing the PCI Configuration Space in Host Mode,” using device = 0 and bus = 0 in PCI CFG_ADDR to indicate the internal PCI/X header. 16.4.2.13, 16-75 16.4.2.13.1, 16-75 16.4.2.13.2, 16-76 Delete the second and third sentences in the first paragraph. Delete the last paragraph. Add the following between the first and second paragraphs: “Note that some errors are enabled by programming two bits—one in the PCI bus command register and another in the PCI/X error enable register (ERR_EN). Likewise, some errors are reported in two bits—one in the PCI bus status register and another in the PCI/X error detect register (ERR_DR). These bits must be cleared separately; that is, clearing one does not clear the other. For example, clearing the ERR_DR[Mstr abort error] does not clear the received master abort bit in the PCI bus status register. In these cases, both bits must be cleared before further error reporting can occur.” 16.4.3.6.4, 16-88 Add the following new section that describes the PCI-X split completion message: 16.4.3.6.4 Split Completion Messages For PCI-X split completion transactions that include a split completion error or split completion message, the transaction has a single DWORD data phase with the format shown in Figure 16-78. Message Class 31 28 27 Message Index 20 0 19 Remaining Lower Address 18 16 Remaining Lower Address (cont.) 15 12 11 Remaining Byte Count 0 Figure 16-78. PCI-X Split Completion Message Format Table 16-62 describes the fields of the PCI-X split completion message. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 A-34 Freescale Semiconductor Changes From Revision 0 to Revision 1 Table 16-62. PCI-X Split Completion Message Summary AD 31:28 Name Message Class Description A split completion message can be one of the following classes: 0x0 Write completion 0x1 PCI-X bridge error 0x2 Completer error 0x3–0xFReserved The message index identifies the type of message within the message class. For write completion message class: 0x00 Normal completion For PCI-X bridge error message class: 0x00 Master-abort 0x01 Target-abort 0x02 Write data parity error For completer error message class: 0x00 Byte count out-of-range error 0x01 Split write data parity error 0x8n Device-specific error Reserved If the split request was a burst memory read transaction, the remaining lower address is the least significant seven bits of the first byte of data that has not been sent. If the split request was a DWORD transaction, the completer places all zeroes in this field.1 If the split request was a burst memory read transaction, the remaining byte count is the number of bytes that have not been sent. If the split request was a DWORD transaction, the completer places 0x004 in this field. 27:20 Message Index 19 18:12 — Remaining Lower Address Remaining Byte Count 11:0 1 Note that the PCI-X 1.0 specification states that if the split request is a DWORD transaction, the completer sets the remaining lower address field (bits 18–12) of the split completion message to zero. However, the MPC8540 deviates from the specification when the split request is a DWORD read transaction to the upper 32 bits of a quad word (that is, the address ends with 1xx); in this case, bit 14 of the split completion message is set to 1. 16.4.3.7, 16-88 Add the following two paragraphs at the beginning of the section: “PCI-X configuration transactions are similar to PCI configuration transaction as described in Section 16.4.2.11, “Configuration Cycles.” The exceptions are noted in this section. There is a rare situation where an internal access (originating from either the processor core, or other I/O ports on the device) to the 256-byte PCI-X configuration header may be corrupted by concurrent outbound PCI-X data transactions. The corruption does not occur if the PCI-X configuration transactions and the outbound PCI-X data transactions are not concurrent. The corruption does not occur with inbound PCI-X traffic. This is an issue for any system that mixes internally generated PCI-X configuration transactions with outbound PCI-X data transactions. There are two ways to avoid the situation: MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor A-35 Revision History Ensure that PCI-X configuration transactions are completed before starting PCI-X data transactions. Ensure that all PCI-X outbound data traffic is quiesced before attempting PCI-X configuration transactions. The programmer must ensure all of the following: Non-configuration transactions to PCI-X from the processor core are quiesced. The core is not fetching instructions from PCI-X. DMA transactions to PCI-X are quiesced. RapidIO bridging transactions to PCI-X are quiesced.” 16.4.3.7, 16-88 Add the following before the first sentence: “As is the case for PCI configuration accesses, the Bus master bit in the MPC8540’s PCI bus command register must be set before an outbound configuration access is attempted.” 16.4.3.8.1, 16-91 Add the following between the first and second paragraphs: “Note that some errors are enabled by programming two bits—one in the PCI bus command register and another in the PCI/X error enable register (ERR_EN). Likewise, some errors are reported in two bits—one in the PCI bus status register and another in the PCI/X error detect register (ERR_DR). These bits must be cleared separately; that is, clearing one does not clear the other. For example, clearing the ERR_DR[Mstr abort error] does not clear the received master abort bit in the PCI bus status register. In these cases, both bits must be cleared before further error reporting can occur.” 16.4.3.8.1, 16–93 Add the following after Table 16-64: “When an outbound PCI-X read transaction has been split and the split completion transaction indicates a master-abort or a target-abort, the split-completion error message (SCEM) data field contains information for setting the received master-abort or received target-abort bit in the PCI bus status register (bits 13 and 12, respectively). However, for misaligned reads (that is, reads not aligned on a 64-bit boundary), the error is not reflected in the PCI Status register. The error is reported in the PCI error detection register (ERR_DR) at 0x0_8E00. The master-abort or target-abort can be inferred from the error data capture register (ERR_DL). Drivers that require this information can read the ERR_DR register rather than the PCI bus status register.” MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 A-36 Freescale Semiconductor Changes From Revision 0 to Revision 1 16.5.4, 16–95 Add the following new section: 16.5.4 PCI-X Inbound Reads that Cross Window Boundaries For Rev. 1 of the MPC8540, inbound PCI-X read requests can cross a window boundary. For Rev. 2, if an inbound PCI-X read request crosses a window boundary, the MPC8540 responds with a split completion message indicating an error but does not return any data. This is a deviation from the PCI-X specification which describes that data is returned up to the window boundary. 16.5.5, 16–95 Add the following new section: 16.5.5 Bridging Between RapidIO and PCI/X The MPC8540 does not support bridging from RapidIO according to the RapidIO Interoperability Specification. If an external RapidIO device tries to target the MPC8540 with a RapidIO NWRITE-R transaction or with any RapidIO write at a priority level higher than 0, and the RapidIO inbound ATMU routes the transaction to the PCI/PCI-X interface, the MPC8540 could deadlock internally. Therefore, the use of RapidIO NWRITE-R transactions must be avoided in bridging applications to the PCI/X controller, and all RapidIO writes targeting the PCI/X controller must be of priority level 0. This means transaction ordering is relaxed because read responses will not necessarily push posted writes in the bridge, and writes will never bypass read requests. Note that this specifically precludes the topology where two MPC8540s communicating with each other via PCI or PCI-X may deadlock when two RapidIO devices (one on each MPC8540) are simultaneously performing priority-1 writes to targets on the remote MPC8540 through the PCI/X link between the devices. To support such a topology, RapidIO write transactions that bridge to the other MPC8540 must be priority 0. 17.3, 17-9–2-13 In Table 17-6, correct the access designation for all RapidIO Registers listed as “Special” to “R/W”. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor A-37 Revision History 17.3.1.5, 17-16 In Table 17-11, correct the PEFCAR[CTLS] bitfield state description to read as follows: “0 1 The RapidIO controller does not support the large common transport route field. The RapidIO controller supports the large common transport route field." 17.3.1.20, 17-30 In Table 17-26, add the following cross-reference to the field description for PLMREQCSR[C]: “See Table 17-85 for command symbol formats.” In Table 17-83, add the following to the "Comments" field for "Maintenance Class" reads: "Although possible to execute, RapidIO maintenance reads from memory locations outside of the RapidIO architectural memory space (0xC_0000–0xC_FFFC) return unreliable data." In Table 17-85, add the following to the "Comments" field of the symbol type 5 (Link Request) control symbol format description: "Note that the RapidIO controller causes a checkstop if a software-initiated send training link request command collides with an inbound send training link request which is immediately followed by an inbound training pattern." In Table 18-3, change the reset value of the GPINDR register to be 0xnnnn_0000. In Figure 18-9, change the reset value of the GPINDR register to be nnnn_nnnn_nnnn_nnnn_0000_0000_0000_0000. In Table 18-17, change the bit ranges “32–47” to “0–15” and “48–63” to “16–31”. Add the following sentence to the end of the first paragraph: “Note that the DDR controller does not shut down unless DDR_SDRAM_INTERVAL[REFINT] is set to a non-zero value. See Section 9.4.1.7, “DDR SDRAM Interval Configuration (DDR_SDRAM_INTERVAL),” for details.” 17.4.1, 17-83 17.4.1, 17-86 18.4, 18-4 18.4.1.9, 18-11 18.4.1.14, 18-17 18.5.1.5.3, 18-23 19.3.1, 19-4 In Table 19-1, correct the name of the register with offset 0xE_1094 from PMLCB5 to PMLCB8. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 A-38 Freescale Semiconductor Changes From Revision 0 to Revision 1 20.2.1, 20-7 LSSD_ MODE L1_TSTCLK In Table 20-2, add the following signals: Test Test Factory test. Refer to the MPC8540 Integrated Processor Hardware Specifications for proper treatment. Factory test. Refer to the MPC8540 Integrated Processor Hardware Specifications for proper treatment. Factory test. Refer to the MPC8540 Integrated Processor Hardware Specifications for proper treatment. I 20-9 Test Test I 20-9 L2_TSTCLK Test Test I 20-9 20.2.2.3, 20-9 LSSD_MODE L1_TSTCLK L2_TSTCLK I I I In Table 20-5, add the following signal descriptions: Used for factory test. Refer to the M PC8540 Integrated Processor Hardware Specifications for proper treatment. Used for factory test. Refer to the M PC8540 Integrated Processor Hardware Specifications for proper treatment. Used for factory test. Refer to the M PC8540 Integrated Processor Hardware Specifications for proper treatment. 20.3.1.2, 20-13 Change the last sentence in the first paragraph to, “Note that the transaction address is qualified with the bits described in Section 20.3.1.3, “Watchpoint Monitor Address Mask Register (WMAMR),” before being compared with WMAR. Note also that the contents of WMAR are not qualified with WMAMR.” Change the last part of the first sentence to be “...contains the mask that is applied to a transaction address before the address is compared with WMAR.” In the first sentence, replace “TBCR[AMD]” with “TBCR0[AMD]”. Also, replace the last sentence of the first paragraph with the following, “The transaction address is qualified with the bits described in Section 20.3.2.3, “Trace Buffer Address Mask Register (TBAMR), ” before being compared with TBAR. Note that the contents of TBAR are not qualified with TBAMR.” Change the last part of the first sentence to be “...contains the mask that is applied to a transaction address before the address is compared with TBAR.” In the last sentence of the first paragraph, replace “TBCR[TMD]” with “TBCR0[TMD]”. 20.3.1.3, 20-13 20.3.2.2, 20-18 20.3.2.3, 20-18 20.3.2.4, 20-19 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor A-39 Revision History 20.4.6.1, 20-29 In Table 20-27, replace the “Function” description of the CMDBC field with the following: “Byte count. Range: 32 to 1 where a value of 0 indicates 32 bytes. 00000 = 32 bytes 00001 = 1 byte 00010 = 2 bytes ... 11110 = 30 bytes 11111 = 31 bytes” 20.4.6.1, 20-30 and 20-31 Change all instances of “TRSEL” and “TRSEL1” to “IFSEL” and all instances of “TBCR” to “TBCR1”. 21.2, 21-5 21.5.2, 21-10 0x2_604C In the first bullet item of the FEC features list, change “820.3x” to “802.3x.” In Table 21-3, add the following row: FIFO_PAUSE_CTRL—FIFO pause control register R/W 0x0000_0000 14.5.3.2.1/A-19 21.5.2, 21-11 21.5.3.1.1, 21-14 In Table 21-3, change the reset value for the FEC IFSTAT register to be 0x0000_0000. In Table 21-4, correct the IEVENT[RXC] bitfield description to read as follows: “Receive control interrupt. A control frame was received. If MACCFG1[Rx_Flow] is set, a pause operation is performed lasting for the duration specified in the received pause control frame and beginning when the frame was received. 0 1 21.5.3.1.1, 21-14 Control frame not received Control frame received” In Table 21-4, replace the EBERR field description with: “Ethernet bus error. This bit indicates that a system bus error for a memory read ocurred while a DMA transaction was underway. If the EBERR is set while transmission is in progress, the DMA stops sending data to the Tx FIFO which eventually causes an underrun error (XFUN) and TSTAT[THLT] is set. If the EBERR is set while receiving a frame, the DMA discards the frame and RSTAT[QHLT] is set. 0 1 No system bus error occurred. System bus error occurred.” MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 A-40 Freescale Semiconductor Changes From Revision 0 to Revision 1 21.5.3.1.2, 21-16 Replace the last sentence in the first paragraph with: “The interrupt signal can be cleared by clearing the corresponding IEVENT bit.” 21.5.3.1.3, 21-17 21.5.3.1.6, 21-20 21.5.3.2.1, 21-21 Undocument bitfields TXEDIS, CRL/XDADIS, and XFUNDIS. Remove all references to bitfield DMACTRL[TOD]. This bit position (29) is now reserved. Insert the following section as the first subsection of Section 21.5.3.2: 21.5.3.2.1 FIFO Pause Control Register (FIFO_PAUSE_CTRL) FIFO_PAUSE_CTRL, shown in Figure 21-11, is writable by the user to configure the properties of the FEC FIFO. 0 15 R W Reset 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0000_0000_0000_0000 16 29 30 31 R W Reset Offset 0 0 0 0 0 0 0 0 0 0 0 0 0 0 TFC_PAUSE_EN 0 0000_0000_0000_0000 0x2_604C Figure 21-11FIFO_PAUSE_CTRL Register Definition Table 21-10 describes the fields of the FIFO_PAUSE_CTRL register. Table21-10FIFO_PAUSE_CTRL Field Descriptions Bits 0–29 30 Name — TFC_PAUSE_EN Reserved. TFC_PAUSE enable. This bit enables the ability to transmit a pause control frame by setting the TCTRL[TFC_PAUSE] bit. This bit is cleared at reset. 0 Pause control frame transmission disabled. 1 Pause control frame transmission enabled. Reserved. Description 31 — 21.5.3.3.1, 21-24 In Table 21-13, replace the bitfield description of TCTRL[TFC_PAUSE] with the following: “Transmit flow control pause frame. Use this bit to transmit a PAUSE frame. To transmit a flow control pause frame, first set FIFO_PAUSE_CTRL[TFC_PAUSE_EN]. Next, set MACCFG1[GTS]. If MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor A-41 Revision History TFC_PAUSE is then set, the MAC stops transmission of data frames after the current transmission completes. The GTSC interrupt in the IEVENT register is asserted. With transmission of data frames stopped, the MAC transmits a MAC control PAUSE frame with the duration value obtained from the PTV register. The TXC interrupt occurs after sending the control pause frame. Next, the MAC clears TFC_PAUSE and resumes transmitting data frames. Note that if the transmitter is paused due to user assertion of GTS or reception of a PAUSE frame, the MAC may still transmit a MAC control PAUSE frame. 0 1 21.5.3.4.2, 21-31 No outstanding pause frame transmission request. Pause frame transmission requested.” In Table 21-22, replace the description for the RSTAT[QHLT] field with the following: “RxBD queue is halted. When IEVENT[BSY] or IEVENT[EBERR] is set during reception of a packet, RSTAT[QHLT] is also set. In order to begin receiving packets again, the user must clear RSTAT[QHLT]. This bit is set whenever the FEC reads an RxBD with its Empty field cleared or encounters a system bus error while processing an RX packet. It is a hardware-initiated stop indication (DMA_CTRL[GRS] being set by the user does not cause this bit to be set.). The current frame and all other frames directed to the halted queue are discarded. A write with a value of 1 re-enables the queue for receiving. 0 1 RxBD queue is enabled for Ethernet reception. (That is, it is not halted.) All Ethernet controller receive activity to RxBD queue is halted.” 21.5.3.4.2, 21-30 0 Replace Figure 21-23 with the following figure: 7 8 9 31 R0 W Reset Offset 0000000 QHLT 0 0 0 00000000000000000000 0000_0000_0000_0000_0000_0000_0000_0000 0x2_6304 21.5.3.6.1, 21-37 In Table 21-28, add the following sentence to the MACCFG1[Rx_EN] field description: “If set, prior to clearing this bit, set DMACTRL[GRS] then confirm subsequent occurrence of the graceful receive stop interrupt (IEVENT[GRSC] is set).” MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 A-42 Freescale Semiconductor Changes From Revision 0 to Revision 1 21.5.3.6.1, 21-37 In Table 21-28, add the following sentence to the MACCFG1[Tx_EN] field description: “If set, prior to clearing this bit, set DMACTRL[GTS] then confirm subsequent occurrence of the graceful transmit stop interrupt (IEVENT[GTSC] is set).” 21.5.3.6.6, 21-41 In Figure 21-34, change the reset value for the FEC IFSTAT register to be 0x0000_0000. 21.5.3.6.6, 21-40 and 21-41 Delete the first sentence and replace Figure 21-35 with the following figure (exposing the Link Fail status field): 0 21 22 23 28 31 R W Reset Offset 000000000000000000000 0 Excess Defer 0 0000 Link 00 Fail 0 0000_0000_0000_0000_0000_0000_0000_0000 TSEC1:0x2_453C; TSEC2:0x2_553C 21.5.3.6.6, 21-41 23–27 28 — Link Fail In Table 21-34, replace the last row with the following rows: Reserved Link Fail. This bit indicates the status of signal detection. 0 The100X module has detected a “signal detect” for longer than 330 mS. 1 The100X module has detected a “signal detect” for less than 330 mS or not at all. Reserved 29–31 — 21.5.3.7.1, 21-43 Change the second sentence of the first paragraph to read: “When the DA field of a receive frame is processed through a 32-bit CRC generator, the 8 high-order bits (0–7) of the CRC remainder are mapped to one of the 256 entries.” 21.5.3.7.1, 21-43 In Table 21-36, add the following sentence to the IADDRn field description: “For instance, the MSB of IADDR0 correlates to entry 0 and the LSB of IADDR7 correlates to entry 255.” 21.5.3.7.2, 21-44 In Table 21-37, add the following sentence to the GADDRn field description: “For instance, the MSB of GADDR0 correlates to entry 0 and the LSB of GADDR7 correlates to entry 255.” MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor A-43 Revision History 21.6.2.2, 21-48 Replace the soft reset steps with the following: 1.Set GTS bit in DMACTRL register. 2.Poll GTSC bit in IEVENT register until detected as set. 3.Clear both Rx_EN and Tx_EN bits in MACCFG1. 4.Wait for a period of 9.6 Kbytes worth of data on the interface (~8ms worst case). 5.Set GRS bit in DMACTRL register. 6.Poll GRSC bit in IEVENT register until detected as set. 7.Set Soft_Reset bit in MACCFG1 register. 8.Clear Soft_Reset bit in MACCFG1 register. 9.Load TBASE with new TxBD pointer. 10.Load RBASE with new RxBD pointer. 11.Set up other MAC registers (MACCFG2, MAXFRM, and so on). 12.Set WWR and WOP bits in DMACTRL register. 13.Clear THLT bit in TSTAT register and QHLT bit in RSTAT register by writing 1 to these bits. 14.Clear GRS/GTS bits in DMACTRL. (Do not change other bits.) 15.Enable Tx_EN/Rx_EN in MACCFG1 register. Replace the second paragraph with the following: "If the user has a frame ready to transmit, a transmit-on-demand function may be emulated while in polling mode by using the graceful-transmit-stop feature. First, clear the IMASK[GTSCEN] bit to mask the graceful-transmit-stop complete interrupt. Next set, then immediately clear the DMACTRL[GTS] bit. Clear the resulting IEVENT[GTSC] bit. Finally, the IMASK[GTSCEN] bit may be set once again." Insert the following sentence between the third and (formerly) fourth sentences of the sixth paragraph: "The pause duration is defined by the received pause control frame and begins when the frame was first received." Replace the third and fourth sentences of the first paragraph with the following: “The eight high-order bits of a cyclic redundancy check (CRC) checksum are used to index into the hash table.” 21.6.2.3, 21-49 21.6.2.3, 21-50 21.6.2.5.2, 21-54 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 A-44 Freescale Semiconductor Changes From Revision 0 to Revision 1 21.6.2.5.3, 21-54 Add the following section after Section 21.6.2.5.2: There are many algorithms for calculating the CRC value of a number. Refer to the RFC 3309 standard, which can be found at http://www.faqs.org/rfcs/rfc3309.html, to compute the CRC value for the purposes of TSEC. The RFC 3309 algorithm uses the following polynomial to calculate the CRC value: x32+x26+x23+x22+x16+x12+x11+x10+x8+x7+x5+x4+x2+x1+x0 or 0x04c11db7. Given a destination MAC address of DA=01000CCCCCCC, the algorithm results in a CRC remainder value of 0xA29F4BBC. Bit-reversing the low-order byte of the CRC value (0xBC) yields BR_CRC = 0x3D = 0b00111101 The high-order 3-bits of the new BR_CRC value are used to select which 32-bit register (of the 8) to use. This example maps the DA to register 1. High-order 3 bits of BR_CRC: HO_CRC = 0b001 = 1 The low-order 5 bits are used to select which bit to set in the given register (with a value of 0 setting 0x8000_0000 and 31 setting 0x0000_0001). Therefore, the example DA maps to bit 29 of register 1. Low-order 5 bits of BR_CRC: LO_CRC = 0b11101 = 29 Therefore, GADDR1 is ORed with the value 0x0000_0004. Additional calculated examples follow: Example 1: •Destination MAC address: DA = 01005E000128 •CRC remainder value: CRC = 0x821D6CD3 •Bit-reversed least-significant byte of CRC value: BR_CRC = 0xCB = 0b11001011 •High-order 3 bits of BR_CRC: HO_CRC = 0b110 = 6 •Low-order 5 bits of BR_CRC: LO_CRC = 0b01011 = 11 •GADDR6 = 0x0010_0000 Example 2: •Destination MAC address: DA = 0004F0604F10 •CRC remainder value: CRC = 0x1F5A66B5 21.6.2.5.3 CRC Computation Examples MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor A-45 Revision History 21.6.2.7, 21-56 •Bit-reversed least-significant byte of CRC value: BR_CRC = 0xAD = 0b10101101 •High-order 3 bits of BR_CRC: HO_CRC = 0b101 = 5 •Low-order 5 bits of BR_CRC: LO_CRC = 0b01101 = 13 •GADDR5 = 0x0004_0000 In Table 21-44, corrected the RXC description to read as follows: “Receive control: A control frame was received. As soon as the transmitter finishes sending the current frame, a pause operation is performed lasting for the duration specified in the received pause control frame and beginning when the frame was first received.” 21.6.3.2, 21-63 In Table 21-48, change the LG bitfield description to read: “Rx frame length violation. Written by TSEC. (Only valid if L is set.) A frame length greater than maximum frame length was recognized while MACCFG2[HUGE FRAME] was set. Note, if MACCFG2[HUGE FRAME] is cleared, the frame is truncated to the value programmed in the maximum frame length register.” 21.6.4, 21-64 Replace the first paragraph with the following: “Some applications require the ability to identify selected portions of data within a frame’s data payload. This process is called extraction although the data is not truly removed. Rather than literarily extracting a section of the data and copying it into a new memory location, the data is placed in the L2 cache. This allows the processor to quickly access critical frame information as soon as the processor is ready without having to first fetch the data from main memory. This results in substantial improvement in throughput and hence reduction in latency.” 21.7.1.1, 21-66 21.7.1.1, 21-66 In Table 21-50, remove the sixth initialization step. In Table 21-50, add the following step immediately following the optional DMACTRL initialization step: Initialize FIFO_PAUSE_CTRL, FIFO_PAUSE_CTRL[0000_0000_0000_0000_0000_0000_0000_0010] MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 A-46 Freescale Semiconductor Glossary of Terms and Abbreviations The glossary contains an alphabetical list of terms, phrases, and abbreviations used in this reference manual. A Architecture. A detailed specification of requirements for a processor or computer system. It does not specify details of how the processor or computer system must be implemented; instead it provides a template for a family of compatible implementations. Atomic access. A bus access that attempts to be part of a read-write operation to the same address uninterrupted by any other access to that address (the term refers to the fact that the transactions are indivisible). The PowerPC architecture implements atomic accesses through the lwarx/stwcx. instruction pair. Autobaud. The process of determining a serial data rate by timing the width of a single bit. B Beat. A single state on the bus interface that may extend across multiple bus cycles. A transaction can be composed of multiple address or data beats. Big endian. A byte-ordering method in memory where the address n of a word corresponds to the most-significant byte. In an addressed memory word, the bytes are ordered (left to right) 0, 1, 2, 3, with 0 being the most-significant byte. See Little endian. Boundedly undefined. A characteristic of certain operation results that are not rigidly prescribed by the PowerPC architecture. Boundedly- undefined results for a given operation may vary among implementations and between execution attempts in the same implementation. Although the architecture does not prescribe the exact behavior for when results are allowed to be boundedly undefined, the results of executing instructions in contexts where results are allowed to be boundedly undefined are constrained to ones that could have been achieved by executing an arbitrary sequence of defined instructions, in valid form, starting in the state the machine was in before attempting to execute the given instruction. Breakpoint. A programmable event that forces the core to take a breakpoint exception. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor Glossary-1 Burst. A multiple-beat data transfer whose total size is typically equal to a cache block. Bus clock. Clock that causes the bus state transitions. Bus master. The owner of the address or data bus; the device that initiates or requests the transaction. C Cache. High-speed memory containing recently accessed data or instructions (subset of main memory). Cache block. A small region of contiguous memory that is copied from memory into a cache. The size of a cache block may vary among processors; the maximum block size is one page. In PowerPC processors, cache coherency is maintained on a cache-block basis. Note that the term ‘cache block’ is often used interchangeably with ‘cache line.’ Cache coherency. An attribute wherein an accurate and common view of memory is provided to all devices that share the same memory system. Caches are coherent if a processor performing a read from its cache is supplied with data corresponding to the most recent value written to memory or to another processor’s cache. Cache flush. An operation that removes from a cache any data from a specified address range. This operation ensures that any modified data within the specified address range is written back to main memory. This operation is generated typically by a Data Cache Block Flush (dcbf) instruction. Caching-inhibited. A memory update policy in which the cache is bypassed and the load or store is performed to or from main memory. Cast out. A cache block that must be written to memory when a cache miss causes a cache block to be replaced. Changed bit. One of two page history bits found in each page table entry (PTE). The processor sets the changed bit if any store is performed into the page. See also Page access history bits and Referenced. Clean. An operation that causes a cache block to be written to memory, if modified, and then left in a valid, unmodified state in the cache. Clear. To cause a bit or bit field to register a value of zero. See also Set. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Glossary-2 Freescale Semiconductor Completer. In PCI-X, a completer is the device addressed by a transaction (other than a split completion transaction). If a target terminates a transaction with a split response, the completer becomes the initiator of the subsequent split completion. Context synchronization. An operation that ensures that all instructions in execution complete past the point where they can produce an exception, that all instructions in execution complete in the context in which they began execution, and that all subsequent instructions are fetched and executed in the new context. Context synchronization may result from executing specific instructions (such as isync or rfi) or when certain events occur (such as an exception). Copy-back operation. A cache operation in which a cache line is copied back to memory to enforce cache coherency. Copy-back operations consist of snoop push-out operations and cache cast-out operations. D Direct-mapped cache. A cache in which each main memory address can appear in only one location within the cache; operates more quickly when the memory request is a cache hit. Double data rate. Memory that allows data transfers at the start and end of a clock cycle. thereby doubling the data rate. E Effective address (EA). The 32-bit address specified for a load, store, or an instruction fetch. This address is then submitted to the MMU for translation to either a physical memory or an I/O address. Exclusive state. MEI state (E) in which only one caching device contains data that is also in system memory. F Frame-check sequence (FCS). Specifies the standard 32-bit cyclic redundancy check (CRC) obtained using the standard CCITT-CRC polynomial on all fields except the preamble, SFD, and CRC. Fetch. Retrieving instructions from either the cache or main memory and placing them into the instruction queue. Flush. An operation that causes a cache block to be invalidated and the data, if modified, to be written to memory. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor Glossary-3 G General-purpose register (GPR). Any of the 32 registers in the general-purpose register file. These registers provide the source operands and destination results for all integer data manipulation instructions. Integer load instructions move data from memory to GPRs and store instructions move data from GPRs to memory. Guarded. The guarded attribute pertains to out-of-order execution. When a page is designated as guarded, instructions and data cannot be accessed out-of-order. H I Harvard architecture. An architectural model featuring separate caches and other memory management resources for instructions and data. IEEE 754. A standard written by the Institute of Electrical and Electronics Engineers that defines operations and representations of binary floating-point numbers. Illegal instructions. A class of instructions that are not implemented for a particular PowerPC processor. These include instructions not defined by the PowerPC architecture. In addition, for 32-bit implementations, instructions that are defined only for 64-bit implementations are considered to be illegal instructions. For 64-bit implementations instructions that are defined only for 32-bit implementations are considered to be illegal instructions. Implementation. A particular processor that conforms to the PowerPC architecture, but may differ from other architecture-compliant implementations for example in design, feature set, and implementation of optional features. The PowerPC architecture has many different implementations. Imprecise exception. A type of synchronous exception that is allowed not to adhere to the precise exception model (see Precise exceptions). The PowerPC architecture allows only floating-point exceptions to be handled imprecisely. Inbound ATMU windows. Mappings that perform address translation from the external address space to the local address space, attach attributes and transaction types to the transaction, and map the transaction to its target interface. Inter-packet gap. The gap between the end of one Ethernet packet and the beginning of the next transmitted packet. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Glossary-4 Freescale Semiconductor Integer unit. An execution unit in the core responsible for executing integer instructions. In-order. An aspect of an operation that adheres to a sequential model. An operation is said to be performed in-order if, at the time that it is performed, it is known to be required by the sequential execution model. Instruction latency. The total number of clock cycles necessary to execute an instruction and make ready the results of that instruction. K L Kill. An operation that causes a cache block to be invalidated without writing any modified data to memory. Latency. The number of clock cycles necessary to execute an instruction and make ready the results of that execution for a subsequent instruction. L2 cache. Level-2 cache. See Secondary cache. Least-significant bit (lsb). The bit of least value in an address, register, field, data element, or instruction encoding. Least-significant byte (LSB). The byte of least value in an address, register, data element, or instruction encoding. Little endian. A byte-ordering method in memory where the address n of a word corresponds to the least-significant byte. In an addressed memory word, the bytes are ordered (left to right) 3, 2, 1, 0, with 3 being the most-significant byte. See Big endian. Local access window. Mapping used to translate a region of memory to a particular target interface, such as the DDR SDRAM controller or the PCI controller. The local memory map is defined by a set of eight local access windows. The size of each window can be configured from 4 Kbytes to 2 Gbytes. M Media access control (MAC) sublayer. Sublayer that provides a logical connection between the MAC and its peer station. Its primary responsibility is to initialize, control, and manage the connection with the peer station. Medium-dependent interface (MDI) sublayer—Sublayer that defines different connector types for different physical media and PMD devices. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor Glossary-5 Media-independent interface (MII) sublayer. Sublayer that provides a standard interface between the MAC layer and the physical layer for 10/100-Mbps operations. It isolates the MAC layer and the physical layer, enabling the MAC layer to be used with various implementations of the physical layer. MEI (modified/exclusive/invalid). Cache coherency protocol used to manage caches on different devices that share a memory system. Note that the PowerPC architecture does not specify the implementation of a MEI protocol to ensure cache coherency. Memory access ordering. The specific order in which the processor performs load and store memory accesses and the order in which those accesses complete. Memory-mapped accesses. Accesses whose addresses use the page or block address translation mechanisms provided by the MMU and that occur externally with the bus protocol defined for memory. Memory coherency. An aspect of caching in which it is ensured that an accurate view of memory is provided to all devices that share system memory. Memory consistency. Refers to agreement of levels of memory with respect to a single processor and system memory (for example, on-chip cache, secondary cache, and system memory). Memory management unit (MMU). The functional unit that is capable of translating an effective (logical) address to a physical address, providing protection mechanisms, and defining caching methods. Modified state. MEI state (M) in which one, and only one, caching device has the valid data for that address. The data at this address in external memory is not valid. Most-significant bit (msb). The highest-order bit in an address, registers, data element, or instruction encoding. Most-significant byte (MSB). The highest-order byte in an address, registers, data element, or instruction encoding. N NaN. An abbreviation for not a number; a symbolic entity encoded in floating-point format. There are two types of NaNs—signaling NaNs and quiet NaNs. No-op. No-operation. A single-cycle operation that does not affect registers or generate bus activity. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Glossary-6 Freescale Semiconductor O OCeaN (On-chip network). Non-blocking crossbar switch fabric. Enables full duplex port connections at 128Gb/s concurrent throughput and independent per port transaction queuing and flow control. Permits high bandwidth, high performance, as well as the execution of multiple data transactions. Outbound ATMU windows. Mappings that perform address translations from local 32-bit address space to the address spaces of RapidIO or PCI/PCI-X, which may be much larger than the local space. Outbound ATMU windows also map attributes such as transaction type or priority level. P Packet. A unit of binary data that can be routed through a network. Sometimes packet is used to refer to the frame plus the preamble and start frame delimiter (SFD). Page. A region in memory. The OEA defines a page as a 4-Kbyte area of memory aligned on a 4-Kbyte boundary. Page access history bits. The changed and referenced bits in the PTE keep track of the access history within the page. The referenced bit is set by the MMU whenever the page is accessed for a read or write operation. The changed bit is set when the page is stored into. See Changed bit and Referenced bit. Page fault. A page fault is a condition that occurs when the processor attempts to access a memory location that does not reside within a page not currently resident in physical memory. On PowerPC processors, a page fault exception condition occurs when a matching, valid page table entry (PTE[V] = 1) cannot be located. Page table. A table in memory is comprised of page table entries, or PTEs. It is further organized into eight PTEs per PTEG (page table entry group). The number of PTEGs in the page table depends on the size of the page table (as specified in the SDR1 register). Page table entry (PTE). Data structures containing information used to translate effective address to physical address on a 4-Kbyte page basis. A PTE consists of 8 bytes of information in a 32-bit processor and 16 bytes of information in a 64-bit processor. Physical coding sublayer (PCS). Sublayer responsible for encoding and decoding data stream to and from the MAC sublayer. Medium (1000BASEX) 8B/10B coding is used for fiber. Medium (1000BASET) 8B1Q coding is used for unshielded twisted pair (UTP). MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor Glossary-7 Physical medium attachment (PMA) sublayer. Sublayer responsible for serializing code groups into a bit stream suitable for serial bit-oriented physical devices (SERDES) and vice versa. Synchronization is also performed for proper data decoding in this sublayer. The PMA sits between the PCS and the PMD sublayers. Physical medium dependent (PMD) sublayer. Sublayer responsible for signal transmission. The typical PMD functionality includes amplifier, modulation, and wave shaping. Different PMD devices may support different media. Physical memory. The actual memory that can be accessed through the system’s memory bus. Pipelining. A technique that breaks operations, such as instruction processing or bus transactions, into smaller distinct stages or tenures (respectively) so that a subsequent operation can begin before the previous one has completed. Precise exceptions. A category of exception for which the pipeline can be stopped so instructions that preceded the faulting instruction can complete and subsequent instructions can be flushed and redispatched after exception handling has completed. See Imprecise exceptions. Primary opcode. The most-significant 6 bits (bits 0–5) of the instruction encoding that identifies the type of instruction. Program order. The order of instructions in an executing program. More specifically, this term is used to refer to the original order in which program instructions are fetched into the instruction queue from the cache. Protection boundary. A boundary between protection domains. Protection domain. A protection domain is a segment, a virtual page, a BAT area, or a range of unmapped effective addresses. It is defined only when the appropriate relocate bit in the MSR (IR or DR) is 1. Q Quad word. A group of 16 contiguous locations starting at an address divisible by 16. Quiesce. To come to rest. The processor is said to quiesce when an exception is taken or a sync instruction is executed. The instruction stream is stopped at the decode stage and executing instructions are allowed to complete to create a controlled context for instructions that may be affected by out-of-order, parallel execution. See Context synchronization. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Glossary-8 Freescale Semiconductor R rA. The rA instruction field is used to specify a GPR to be used as a source or destination. rB. The rB instruction field is used to specify a GPR to be used as a source. rD. The rD instruction field is used to specify a GPR to be used as a destination. rS. The rS instruction field is used to specify a GPR to be used as a source. RapidIO. High-performance, packet-switched, interconnect architecture that provides reliability, increased bandwidth, and faster bus speeds in an intra-system interconnect. Designed to be compatible with integrated communications processors, host processors, and networking digital signal processors, Record bit. Bit 31 (or the Rc bit) in the instruction encoding. When it is set, updates the condition register (CR) to reflect the result of the operation. Reconciliation sublayer. Sublayer that maps the terminology and commands used in the MAC layer into electrical formats appropriate for the physical layer entities. Referenced bit. One of two page history bits found in each page table entry. The processor sets the referenced bit whenever the page is accessed for a read or write. See also Page access history bits. Requester. In PCI-X, a requester is an initiator that first introduces a transaction into the PCI-X domain. If a transaction is terminated with a split response, the requester becomes the target of the subsequent split completion. Reservation. The processor establishes a reservation on a cache block of memory space when it executes an lwarx instruction to read a memory semaphore into a GPR. Reservation station. A buffer between the dispatch and execute stages that allows instructions to be dispatched even though the results of instructions on which the dispatched instruction may depend are not available. RISC (reduced instruction set computing). An architecture characterized by fixed-length instructions with nonoverlapping functionality and by a separate set of load and store instructions that perform memory accesses. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor Glossary-9 S Secondary cache. A cache memory that is typically larger and has a longer access time than the primary cache. A secondary cache may be shared by multiple devices. Also referred to as L2, or level-2, cache. Sequence. In PCI-X, a sequence is one or more transactions associated with carrying out a single logical transfer by a requester. Each transaction in the same sequence carries the same unique sequence ID. Set (v). To write a nonzero value to a bit or bit field; the opposite of clear. The term ‘set’ may also be used to generally describe the updating of a bit or bit field. Set (n). A subdivision of a cache. Cacheable data can be stored in a given location in one of the sets, typically corresponding to its lower-order address bits. Because several memory locations can map to the same location, cached data is typically placed in the set whose cache block corresponding to that address was used least recently. See Set associative. Set associative. Aspect of cache organization in which the cache space is divided into sections, called sets. The cache controller associates a particular main memory address with the contents of a particular set, or region, within the cache. Slave. The device addressed by a master device. The slave is identified in the address tenure and is responsible for supplying or latching the requested data for the master during the data tenure. Snooping. Monitoring addresses driven by a bus master to detect the need for coherency actions. Snoop push. Response to a snooped transaction that hits a modified cache block. The cache block is written to memory and made available to the snooping device. Stall. An occurrence when an instruction cannot proceed to the next stage. Sticky bit. A bit that when set must be cleared explicitly. Superscalar machine. A machine that can issue multiple instructions concurrently from a conventional linear instruction stream. Supervisor mode. The privileged operation state of a processor. In supervisor mode, software, typically the operating system, can access all control registers and can access the supervisor memory space, among other privileged operations. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Glossary-10 Freescale Semiconductor Synchronization. A process to ensure that operations occur strictly in order. See Context synchronization. Synchronous exception. An exception that is generated by the execution of a particular instruction or instruction sequence. There are two types of synchronous exceptions, precise and imprecise. System memory. The physical memory available to a processor. T Time-division multiplex (TDM). A single serial channel used by several channels taking turns. Tenure. The period of bus mastership. There can be separate address bus tenures and data bus tenures. TLB (translation lookaside buffer). A cache that holds recently-used page table entries. Throughput. The measure of the number of instructions that are processed per clock cycle. Transaction. A complete exchange between two bus devices. A transaction is typically comprised of an address tenure and one or more data tenures, which may overlap or occur separately from the address tenure. A transaction may be minimally comprised of an address tenure only. Transfer termination. Signal that refers to both signals that acknowledge the transfer of individual beats (of both single-beat transfer and individual beats of a burst transfer) and to signals that mark the end of the tenure. U User mode. The operating state of a processor used typically by application software. In user mode, software can access only certain control registers and can access only user memory space. No privileged operations can be performed. Also referred to as problem state. Virtual address. An intermediate address used in the translation of an effective address to a physical address. Virtual memory. The address space created using the memory management facilities of the processor. Program access to virtual memory is possible only when it coincides with physical memory. V MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor Glossary-11 W Way. A location in the cache that holds a cache block, its tags, and status bits. Word. A 32-bit data element. Write-back. A cache memory update policy in which processor write cycles are directly written only to the cache. External memory is updated only indirectly, for example, when a modified cache block is cast out to make room for newer data. Write-through. A cache memory update policy in which all processor write cycles are written to both the cache and memory. MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Glossary-12 Freescale Semiconductor Index 1 Register Index (Memory-Mapped Registers) A AICAR (RapidIO assembly information capability register), 17-15 AIDCAR (RapidIO assembly identity capability register), 17-14 ALTCAR (alternate configuration attribute address register), 4-6 ALTCBAR (alternate configuration base address register), 4-6 ANA (TSEC AN advertisement register), 14-94 ANLPANP (TSEC AN link partner ability next page register), 14-99 ANLPBPA (TSEC AN link partner base page ability register), 14-96 ANNPT (TSEC AN next page transmit register), 14-98 ATTR (FEC attribute register), 21-46 ATTR (TSEC attribute register), 14-89 ATTRELI (FEC attribute extract length and extract index register), 21-47 ATTRELI (TSEC attribute extract length and extract index register), 14-90 CAR2 (TSEC carry register 2), 14-84 CCIDR (current context ID register), 20-24 CCSRBAR (configuration, control, and status registers base address register), 4-4-4-5 CFG_ADDR (PCI/X configuration address register), 16-18 CFG_DATA (PCI/X configuration data register), 16-19 CISR0 (PIC critical interrupt summary register 0), 10-27 CISR1 (PIC critical interrupt summary register 1), 10-28 CLKOCR (clock out select register), 18-18 CLNDARn (DMA 0–3 current link descriptor address registers), 15-14 CLSDARn (DMA 0–3 current list-alternate base descriptor address registers), 15-24 CR (RapidIO configuration register), 17-34 CR (TSEC control register), 14-92 CRBPTR (FEC current receive buffer descriptor pointer register), 21-33 CRBPTR (TSEC current receive buffer descriptor pointer register), 14-44 CSnBNDS (DDR chip select 0–3 bounds registers), 9-10 CSnCONFIG (DDR chip select 0–3 configuration registers), 9-10 CTBPTR (FEC current transmit buffer descriptor pointer register), 21-27 CTBPTR (TSEC current transmit buffer descriptor pointer register), 14-36 CTCSR (RapidIO component tag command and status register), 17-26 CTPR0 (PIC per-CPU processor current task priority register), also mapped as global CTPR, 10-40 B BCRn (DMA 0–3 byte count registers), 15-23 BDIDCSR (RapidIO base device ID command and status register), 17-25 BPTR (boot page translation register), 4-8 BRn (LBC base registers 0–7), 13-11 C CAM1 (TSEC carry mask register 1), 14-85 CAM2 (TSEC carry mask register 2), 14-86 CAPTURE_ADDRESS (DDR memory error address capture register), 9-24 CAPTURE_ATTRIBUTES (DDR memory error attributes capture register), 9-23 CAPTURE_DATA_HI (DDR memory data path read capture high register), 9-19 CAPTURE_DATA_LO (DDR memory data path read capture low register), 9-20 CAPTURE_ECC (DDR memory data path read capture ECC register), 9-20 CAR1 (TSEC carry register 1), 14-82 D DARn (DMA 0–3 destination address registers), 15-21 DARn (DMA 0–3 destination address registers for RapidIO maintenance writes), 15-22 DATA_ERR_INJECT_HI (DDR memory data path error injection mask high register), 9-17 DATA_ERR_INJECT_LO (DDR memory data path error injection mask low register), 9-18 DATRn (DMA 0–3 destination attributes registers), 15-19 DDR_SDRAM_CFG (DDR SDRAM control configuration register), 9-14 DDR_SDRAM_INTERVAL (DDR SDRAM interval configuration register), 9-16 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor Index 1-1 E–I DDR_SDRAM_MODE (DDR SDRAM mode configuration register), 9-16 DDRDLLCR (DDR DLL control register), 18-19 DEVDISR (device disable control register), 18-13 DGSR (DMA general status register), 15-26 DICAR (RapidIO device information capability register), 17-14 DIDCAR (RapidIO device identity capability register), 17-13 DMACTRL (FEC DMA control register), 21-20 DMACTRL (TSEC DMA control register), 14-28 DMR (RapidIO doorbell mode register), 17-77 DOCAR (RapidIO destination operations capability register), 17-19 DQDPAR (RapidIO doorbell queue dequeue pointer address register), 17-79 DQEPAR (RapidIO doorbell queue enqueue pointer address register), 17-80 DSR (RapidIO doorbell status register), 17-78 DSRn (DMA 0–3 destination stride registers), 15-26 ERR_SBE (DDR single-bit ECC memory error management register), 9-25 EXST (TSEC extended status register), 14-100 F FIFO_TX_STARVE (FEC FIFO transmit starve register), 21-24 FIFO_TX_STARVE (TSEC FIFO transmit starve register), 14-32 FIFO_TX_STARVE_SHUTOFF (FEC FIFO transmit starve shutoff register), 21-24 FIFO_TX_STARVE_SHUTOFF (TSEC FIFO transmit starve shutoff register), 14-32 FIFO_TX_THR (FEC FIFO transmit threshold register), 21-23 FIFO_TX_THR (TSEC FIFO transmit threshold register), 14-31 FRR (PIC feature reporting register), 10-16 E ECC_ERR_INJECT (DDR memory data path error injection mask ECC register), 9-18 ECNTRL (TSEC Ethernet control register), 14-25 EDIS (FEC error disabled register), 21-18 EDIS (TSEC error disabled register), 14-24 EEADR (ECM error address capture register), 8-8 EEATR (ECM error attributes capture register), 8-7 EEBACR (ECM CCB address configuration register), 8-3 EEBPCR (ECM CCB port configuration register), 8-4 EEDR (ECM error detect register), 8-5 EEER (ECM error enable register), 8-6 EIDRn (PIC external interrupt destination registers 0–11), 10-33 EIVPRn (PIC external interrupt vector/priority registers 0–11), 10-32 EOI0 (PIC per-CPU processor end of interrupt register) also mapped as global EOI, 10-42 ERR_ADDR (PCI/X error address capture register), 16-33 ERR_ATTRIB (PCI/X error attributes register), 16-32 ERR_CAP_DR (PCI/X error capture disable register), 16-30 ERR_DETECT (DDR memory error detect register), 9-21 ERR_DH (PCI/X error data high capture register), 16-34 ERR_DISABLE (DDR memory error disable register), 9-21 ERR_DL (PCI/X error data low capture register), 16-34 ERR_DR (PCI/X error detect register), 16-29 ERR_EN (PCI/X error enable register), 16-31 ERR_EXT_ADDR (PCI/X error extended address capture register), 16-34 ERR_INT_EN (DDR memory error interrupt enable register), 9-22 G GADDRn (FEC group address registers 0–7), 21-46 GADDRn (TSEC group address registers 0–7), 14-88 GAS_TIMR (PCI/X gasket timer register), 16-35 GCR (PIC global configuration register), 10-17 GPINDR (general-purpose input data register), 18-12 GPIOCR (GPIO control register), 18-10 GPOUTDR (general-purpose output data register), 18-11 GPPORCR (general-purpose POR configuration register), 18-9 GTBCRn (PIC global timer 0–3 base count registers), 10-21 GTCCRn (PIC global timer 0–3 current count registers), 10-21 GTDRn (PIC global timer 0–3 destination registers), 10-23 GTVPRn (PIC global timer 0–3 vector/priority registers), 10-22 H HAFDUP (FEC half-duplex register), 21-41 HAFDUP (TSEC half-duplex register), 14-53 HBDIDLCSR (RapidIO host base device ID lock command and status register), 17-26 I I2CADR (I2C address register), 11-5 I2CCR (I2C control register), 11-6 I2CDFSRR (I2C digital filter sampling rate register), 11-10 I2CDR (I2C data register), 11-9 I2CFDR (I2C frequency divider register), 11-5 I2CSR (I2C status register), 11-7 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Index 1-2 Freescale Semiconductor J–M IACK0 (PIC per-CPU processor interrupt acknowledge register), also mapped as global IACK, 10-41 IADDRn (FEC individual address registers 0–7), 21-45 IADDRn (TSEC individual address registers 0–7), 14-88 IEVENT (FEC interrupt event register), 21-13 IEVENT (TSEC interrupt event register), 14-20 IFQDPAR (RapidIO inbound frame queue dequeue pointer address register), 17-75 IFQEPAR (RapidIO inbound frame queue enqueue pointer address register), 17-76 IFSTAT (FEC interface status register), 21-43 IFSTAT (TSEC interface status register), 14-58 IIDPRn (PIC internal interrupt destination registers 0–31), 10-35 IIVPRn (PIC internal interrupt vector/priority registers 0–31), 10-34 IMASK (FEC interrupt mask register), 21-16 IMASK (TSEC interrupt mask register), 14-23 IMR (RapidIO inbound mailbox mode register), 17-72 INT_ACK (PCI/X interrupt acknowledge register), 16-20 IPGIFG (FEC inter-packet gap/inter-frame gap register), 21-41 IPGIFG (TSEC inter-packet gap/inter-frame gap register), 14-52 IPIDRn (PIC per-CPU interprocessor interrupt dispatch registers 0–3), 10-39 IPIVPRn (PIC IPI vector/priority registers 0–3), 10-19 IRQSR0 (PIC IRQ_OUT summary register 0), 10-26 IRQSR1 (PIC IRQ_OUT summary register 1), 10-27 ISR (RapidIO inbound mailbox status register), 17-74 L2ERRINJHI (L2 error injection mask high register), 7-13 L2ERRINJLO (L2 error injection mask low register), 7-13, 7-14 L2ERRINTEN (L2 error interrupt enable register), 7-15, 7-18 L2SRBARn (L2 memory-mapped SRAM base address registers 0–1), 7-11 LAWARn (Local access window 0–7 attributes registers), 2-6 LAWBARn (Local access window 0–7 base address registers), 2-6 LBCR (LBC configuration register), 13-31 LBDLLCR (LBC DLL control register), 18-20 LCRR (LBC clock ratio register), 13-32 LCSBA1CSR (RapidIO local configuration space base address 1 command and status register), 17-24 LSDMR (LBC SDRAM mode register), 13-22 LSRT (LBC SDRAM refresh timer), 13-25 LTEAR (LBC transfer error address register), 13-30 LTEATR (LBC transfer error attributes register), 13-29 LTEDR (LBC transfer error disable register), 13-27 LTEIR (LBC transfer error interrupt register), 13-28 LTESR (LBC transfer error status register), 13-26 LURT (UPM refresh timer), 13-24 M MACCFG1 (FEC MAC configuration register 1), 21-38 MACCFG1 (TSEC MAC configuration register 1), 14-49 MACCFG2 (FEC MAC configuration register 2), 21-40 MACCFG2 (TSEC MAC configuration register 2), 14-51 MACSTNADDR1 (FEC station address part 1 register), 21-43 MACSTNADDR1 (TSEC station address part 1 register), 14-59 MACSTNADDR2 (FEC station address part 2 register), 21-44 MACSTNADDR2 (TSEC station address part 2 register), 14-60 MAR (UPM address register), 13-18 MAXFRM (FEC maximum frame length register), 21-42 MAXFRM (TSEC maximum frame length register), 14-54 MCPSUMR (machine check summary register), 18-16 MDR (UPM data register), 13-22 MER (PIC message enable register), 10-31 MIDRn (PIC messaging interrupt destination registers 0–3), 10-37 MIIMADD (TSEC MII management address register), 14-56 MIIMCFG (TSEC MII management configuration register), 14-54 MIIMCOM (TSEC MII management command register), 14-55 MIIMCON (TSEC MII management control register), 14-57 MIIMIND (TSEC MII management indicator register), 14-58 MIIMSTAT (TSEC MII management status register), 14-57 J JD (TSEC jitter diagnostics register), 14-101 L L2CAPTDATAHI (L2 error capture data high register), 7-15 L2CAPTDATALO (L2 error capture data low register), 7-15 L2CAPTECC (L2 error syndrome register), 7-13, 7-16 L2CEWARn (L2 cache external write address registers 0–3), 7-10 L2CEWCRn (L2 cache external write control registers 0–3), 7-10 L2CTL (L2 control register), 7-7 L2ERRADDR (L2 error address capture register), 7-15, 7-20 L2ERRATTR (L2 error attributes capture register), 7-15, 7-19 L2ERRCTL (L2 error control register), 7-15, 7-20 L2ERRDET (L2 error detect register), 7-15, 7-16 L2ERRDIS (L2 error disable register), 7-15, 7-17 L2ERRINJCTL (L2 error injection mask control register), 7-13, 7-14 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor Index 1-3 N–P MINFLR (FEC minimum frame length register), 21-18 MINFLR (TSEC minimum frame length register), 14-26 MIVPRn (PIC messaging interrupt vector/priority registers 0–3), 10-36 MRBLR (FEC maximum receive buffer length register), 21-34 MRBLR (TSEC maximum receive buffer length register), 14-44 MRn (DMA 0–3 mode registers), 15-10 MRTPR (LBC memory refresh timer prescaler register), 13-21 MSGRn (PIC message registers 0–3), 10-30 MSR (PIC message status register), 10-31 MSR (RapidIO mailbox command and status register), 17-22 MxMR (LBC UPM A–C mode registers), 13-19-13-21 N NLNDARn (DMA 0–3 next link descriptor address registers), 15-23 NLSDARn (DMA 0-3 next list descriptor address register), 15-25 O ODATR (RapidIO outbound destination attributes register), 17-70 ODCR (RapidIO outbound double-word count register), 17-71 ODPR (RapidIO outbound destination port register), 17-70 ODQDPAR (RapidIO outbound descriptor queue dequeue pointer address register), 17-68 ODQEPAR (RapidIO outbound descriptor queue enqueue pointer address register), 17-72 OMR (RapidIO outbound mode register), 17-65 ORn (LBC options registers 0–7), 13-12-13-18 OSAR (RapidIO outbound source address register), 17-69 OSR (RapidIO outbound status register), 17-67 OSTBD (FEC out-of-sequence TxBD register), 21-29 OSTBD (TSEC out-of-sequence TxBD register), 14-38 OSTBDP (FEC out-of-sequence Tx data buffer pointer register), 21-31 OSTBDP (TSEC out-of-sequence Tx data buffer pointer register), 14-40 P PCCSR (RapidIO port control command and status register), 17-33 PCI configuration header registers, 16-36-16-53 see also General Index PCIDR (programmed context ID register), 20-24 PCIX_TIMR (PCI-X split completion timer register), 16-36 PCR (RapidIO port configuration register), 17-35 PECR (RapidIO port error control register), 17-52 PECRx (RapidIO packet error registers) see PEPR1x and PEPR2x PEFCAR (RapidIO processing element features capability register), 17-15 PEIR (RapidIO port error injection register), 17-35 PELLCCSR (RapidIO processing element logical layer control command and status register), 17-24 PEPCSR0 (RapidIO port error packet/control symbol register 0), 17-52 PEPR1 (RapidIO port error packet register 1), 17-53 PECR1T1 (packet error capture register 1, type 1), 17-53 PECR1T10 (packet error capture register 1, type 10), 17-56 PECR1T11 (packet error capture register 1, type 11), 17-57 PECR1T13 (packet error capture register 1, type 13), 17-58 PECR1T2 (packet error capture register 1, type 2), 17-54 PECR1T5 (packet error capture register 1, type 5), 17-54 PECR1T6 (packet error capture register 1, type 6), 17-55 PECR1T8R (packet error capture register 1, type 8 response), 17-56 PECR1T8RQ (packet error capture register 1, type 8 request), 17-55 PEPR2 (RapidIO port error packet register 2), 17-58 PECR2T10 (packet error capture register 2, type 10), 17-60 PECR2T11 (packet error capture register 2, types 11 and 13), 17-61 PECR2T125 (packet error capture register 2, types 1, 2, and 5), 17-58 PECR2T6 (packet error capture register 2, type 6), 17-59 PECR2T8R (packet error capture register 2, type 8 response), 17-60 PERTR (RapidIO port error recovery threshold register), 17-64 PESCSR (RapidIO port error and status command and status register), 17-31 PGCCSR (RapidIO port general control command and status register), 17-29 PIR (PIC processor initialization register), 10-18 PITARn (PCI/X inbound translation address registers 1–3), 16-25 PIWARn (PCI/X inbound window attributes registers 1–3), 16-26 PIWBARn (PCI/X inbound window base address registers 1–3), 16-25 PIWBEARn (PCI/X inbound window base extended address registers 1–3), 16-26 PLASCSR (RapidIO port local ackID status command and status register), 17-31 PLMREQCSR (RapidIO port link maintenance request command and status register), 17-29 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Index 1-4 Freescale Semiconductor R–R PLMRESPCSR (RapidIO port link maintenance response command and status register), 17-30 PLTOCCSR (RapidIO port link time-out control command and status register), 17-27 PMBH0CSR (RapidIO 8/16 LP-LVDS port maintenance block header 0 command and status register), 17-27 PMCn (performance monitor counters 0–8), 19-9 PMGC0 (performance monitor global control register 0), 19-5 PMLCA0 (performance monitor local control register A0), 19-6 PMLCAn (performance monitor local control registers A1–A8), 19-6 PMLCB0 (performance monitor local control register B0), 19-7 PMLCBn (performance monitor local control registers B1–B8), 19-8 PMnMR0 (PIC performance monitor 0–3 mask registers (lower)), 10-29 PMnMR1 (PIC performance monitor 0–3 mask registers (upper)), 10-30 PMUXCR (alternate function signal multiplex control), 18-12 PNFEDiR (RapidIO port notification/fatal error disable register), 17-47 PNFEDR (RapidIO port notification/fatal error detect register), 17-44 PNFEIER (RapidIO port notification/fatal error interrupt enable register), 17-49 PORBMSR (POR boot mode status register), 4-14, 4-15, 4-16, 18-5 PORDBGMSR (POR debug mode status register), 4-20, 4-21, 18-9 PORDEVSR (POR I/O device status register), 4-16, 4-17, 4-18, 4-19, 4-20, 18-7 PORIMPSCR (POR I/O impedance status and control register), 4-19, 18-6 PORPLLSR (POR PLL ratio status register), 4-12, 4-13, 18-4 POTARn (PCI/X outbound translation address registers 0–4), 16-21 POTEARn (PCI/X outbound translation extended address registers 0–4), 16-21 POWARn (PCI/X outbound window attributes registers 0–4), 16-22 POWBARn (PCI/X outbound window base address registers 0–4), 16-22 POWMGTCSR (power management status and control register), 18-15 PREDR (RapidIO port recoverable error detect register), 17-61 PRTOCCSR (RapidIO port response time-out control command and status register), 17-28 PRTR (RapidIO port retry threshold register), 17-64 PTV (FEC pause time value register), 21-19 PTV (TSEC pause time value register), 14-27 PVR (processor version register), 18-17 PWDCSR (RapidIO port-write and doorbell command and status register), 17-22 PWMR (RapidIO port-write mode register), 17-81 PWQBAR (RapidIO port-write queue base address register), 17-83 PWSR (RapidIO port-write status register), 17-82 R RALN (TSEC receive alignment error counter register), 14-68 RBASE (FEC receive descriptor base address register), 21-35 RBASE (TSEC receive descriptor base address register), 14-45 RBCA (TSEC receive broadcast packet counter register), 14-66 RBDLEN (FEC RxBD data length register), 21-33 RBDLEN (TSEC RxBD data length register), 14-42 RBPTR (FEC receive buffer descriptor pointer register), 21-35 RBPTR (TSEC receive buffer descriptor pointer register), 14-45 RBYT (TSEC receive byte counter register), 14-64 RCDE (TSEC receive code error counter register), 14-69 RCSE (TSEC receive carrier sense error counter register), 14-70 RCTRL (FEC receive control register), 21-31 RCTRL (TSEC receive control register), 14-41 RDRP (TSEC receive dropped packet counter register), 14-72 RFLR (TSEC receive frame length error counter register), 14-69 RFRG (TSEC receive fragments counter register), 14-71 RIWARn (RapidIO inbound window attributes registers 0–4), 17-42 RIWBARn (RapidIO inbound window base address registers 1–4), 17-41 RIWTARn (RapidIO inbound window translation address registers 0–4), 17-40 RJBR (TSEC receive jabber counter register), 14-72 RMCA (TSEC receive multicast packet counter register), 14-66 ROVR (TSEC receive oversize packet counter register), 14-71 ROWARn (RapidIO outbound window attributes registers 0–8), 17-39 ROWBARn (RapidIO outbound window base address registers 1–8), 17-38 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor Index 1-5 S–T ROWTARn (RapidIO outbound window translation address registers 0–8) for maintenance transactions, 17-38 ROWTARn (RapidIO outbound window translation address registers 0–8) for standard transactions, 17-37 RPKT (TSEC receive packet counter register), 14-65 RSTAT (FEC receive status register), 21-32 RSTAT (TSEC receive status register), 14-41 RUND (TSEC receive undersize packet counter register), 14-70 RXCF (TSEC receive control frame packet counter register), 14-67 RXIC (TSEC receive interrupt coalescing configuration register), 14-43 RXPF (TSEC receive pause frame packet counter register), 14-67 RXUO (TSEC receive unknown opcode packet counter register), 14-68 S SARn (DMA 0–3 source address registers), 15-18 SARn (DMA 0–3 source address registers for RapidIO maintenance reads), 15-19 SATRn (DMA 0–3 source attributes registers), 15-16 SOCAR (RapidIO source operations capability register), 17-17 SPICAR (RapidIO switch port information capability register), 17-17 SR (TSEC status register), 14-93 SRn (DMA 0–3 status registers), 15-13 SSRn (DMA 0–3 source stride registers), 15-25 SVR (PIC spurious vector register), 10-19 SVR (system version register), 18-18 T TBACR (trace buffer access control register), 20-22 TBADHR (trace buffer access data high register), 20-22 TBADR (trace buffer access data register), 20-23 TBAMR (trace buffer address mask register), 20-19 TBAR (trace buffer address register), 20-19 TBASE (FEC transmit descriptor base address register), 21-28 TBASE (TSEC transmit descriptor base address register), 14-37 TBCA (TSEC transmit broadcast packet counter register), 14-74 TBCRn (trace buffer control registers 0–1), 20-16 TBDLEN (FEC TxBD data length register), 21-27 TBDLEN (TSEC TxBD data length register), 14-35 TBICON (TSEC TBI control register), 14-102 TBIPA (TSEC TBI physical address register), 14-29 TBPTR (FEC transmit buffer descriptor pointer register), 21-28 TBPTR (TSEC transmit buffer descriptor pointer register), 14-37 TBSR (trace buffer status register), 20-21 TBTMR (trace buffer transaction mask register), 20-20 TBYT (TSEC transmit byte counter register), 14-73 TCR (PIC timer control register), 10-24 TCTRL (FEC transmit control register), 21-25 TCTRL (TSEC transmit control register), 14-33 TDFR (TSEC transmit deferral packet counter register), 14-75 TDRP (TSEC drop frame counter register), 14-79 TEDF (TSEC transmit excessive deferral packet counter register), 14-76 TFCS (TSEC transmit FCS error counter register), 14-80 TFRG (TSEC transmit fragment counter register), 14-82 TFRR (PIC timer frequency reporting register), 10-20 TIMING_CFG_1 (DDR SDRAM timing configuration register 1), 9-11 TIMING_CFG_2 (DDR SDRAM timing configuration register 2), 9-13 TJBR (TSEC jabber frame counter register), 14-79 TMCA (TSEC transmit multicast packet counter register), 14-74 TNCL (TSEC transmit total collision counter register), 14-78 TOSR (trigger output source register), 20-25 TOVR (TSEC transmit oversize frame counter register), 14-81 TPKT (TSEC transmit packet counter register), 14-73 TR127 (TSEC transmit and receive 65- to 127-byte frame counter register), 14-61 TR1K (TSEC transmit and receive 512- to 1023-byte frame counter register), 14-63 TR255 (TSEC transmit and receive 128- to 255-byte frame counter register), 14-62 TR511 (TSEC transmit and receive 256- to 511-byte frame counter register), 14-62 TR64 (TSEC transmit and receive 64-byte frame counter register), 14-61 TRMAX (TSEC transmit and receive 1024- to 1518-byte frame counter register), 14-63 TRMGV (TSEC transmit and receive 1519- to 1522-byte VLAN frame counter register), 14-64 TSCL (TSEC transmit single collision packet counter register), 14-76 TSTAT (FEC transmit status register), 21-26 TSTAT (TSEC transmit status register), 14-34 TUND (TSEC transmit undersize frame counter register), 14-81 TXCF (TSEC transmit control frame counter register), 14-80 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Index 1-6 Freescale Semiconductor U–W TXCL (TSEC excessive collision packet counter register), 14-78 TXIC (TSEC transmit interrupt coalescing configuration register), 14-35 TXPF (TSEC transmit pause control frame counter register), 14-75 U UAFRn (DUART alternate function registers 0–1), 12-17 UDLBn (DUART divisor least significant byte registers 0–1), 12-7 UDMBn (DUART divisor most significant byte registers 0–1), 12-7 UDSRn (DUART DMA status registers 0–1), 12-6, 12-18 UFCRn (DUART FIFO control registers 0–1), 12-11 UIER (DUART interrupt enable register), 12-9 UIIRn (DUART interrupt ID registers 0–1), 12-10 ULCRn (DUART line control registers 0–1), 12-12 ULSRn (DUART line status registers 0–1), 12-15 UMCRn (DUART MODEM control registers 0–1), 12-14 UMSRn (DUART MODEM status registers 0–1), 12-16 URBRn (DUART receiver buffer registers 0–1), 12-3, 12-6 USCRn (DUART scratch registers 0–1), 12-17 UTHRn (DUART transmitter holding registers 0–1), 12-3, 12-6 V VIR (PIC vendor ID register), 10-17 W WHOAMI0 (PIC per-CPU who am I register—P0) also mapped as global WHOAMI, 10-41 WMAMR (watchpoint monitor address mask register), 20-14 WMAR (watchpoint monitor address register), 20-13 WMCRn (watchpoint monitor control registers 0–1), 20-11 WMSR (watchpoint monitor status register), 20-16 WMTMR (watchpoint monitor transaction mask register), 20-14 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor Index 1-7 W–W MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Index 1-8 Freescale Semiconductor Index 2 General Index A Accumulator (ACC), 6-50 Acronyms and abbreviated terms, list, lxxxviii Address maps addressing on PCI/PCI-X bus, 16-59 device address map overview, 1-20 Address mask (LBC), 13-13 Address multiplexing (LBC SDRAM), 13-54 Address translation and mapping units (ATMUs) inbound windows, 2-9 illegal interactions between inbound ATMUs and local access windows, 2-9 PCI/PCI-X—4 windows, 2-9, 16-24 RapidIO—4 windows plus default, 2-9, 17-40-17-43, 17-107-17-109 local access windows, 2-3-2-9 see also Local access windows outbound windows, 2-8 PCI/PCI-X—4 windows, 16-20 RapidIO—8 windows plus default, 17-37-17-40, 17-106-17-107 processing across the OCeaN fabric, 1-20 Address translation, see e500 core, memory management unit (MMU) Addressing PCI bus addressing, 16-60 configuration space, 16-60 I/O space, 16-60 memory space, 16-59 Alignment, byte (PCI/PCI-X), 16-61 Application examples, 1-21-1-24 control processing, 1-21 APUs accumulator, 6-50 Arbitration I2C interface arbitration control, 11-14 loss of arbitration—forcing of slave mode, 11-23 procedure for arbitration, 11-14 PCI/PCI-X, 16-5, 16-53 Architecture, overview of device, 1-8 ASLEEP (global utilities asleep) signal, 18-2, 18-24 ATMUs, see Address translation and mapping units B BBEAR (branch buffer address register), see e500 core, registers BBTAR (branch buffer target address register), see e500 core, registers Block diagrams DDR controller, 9-1, 9-25 debug modes, watchpoint monitor, and trace buffer, 20-1 DMA controller, 15-1 DUART, 12-2 e500 coherency module (ECM), 8-1 e500 core complex, 5-1 FEC, 21-5 I2C interface, 11-1 interrupt controller (PIC), 10-1, 10-43 L2 cache/SRAM, 7-1 local bus controller (LBC), 13-1 PCI/PCI-X controller, 16-1 performance monitor, 19-2 TSEC, 14-6 Book E architecture, see e500 core, Book E architecture Boot mode CPU holdoff (POR), 4-15, 8-4 POR status register (PORBMSR), 18-5 Boot page translation, 4-7 Boot ROM location (POR), 4-14 Boot sequencer boot holdoff mode (POR), 4-16, 8-4 boot page translation, 4-7 I2C interface, 11-2, 11-17-11-19 overview, 1-16, 4-8 POR configuration, 4-16 BUCSR (branch unit control and status register), see e500 core, registers Buffer descriptors see FEC controller, buffer descriptors see TSEC, buffer descriptors Burst operations (PCI) see PCI/PCI-X controller, bus protocol Bus operations PCI/PCI-X, see PCI/PCI-X controller, bus protocol Byte alignment (PCI/PCI-X), 16-61 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor Index 2-1 C–D C Chaining performance monitor events, 19-27 CKSTP_IN (global utilities checkstop in) signal, 18-3 CKSTP_OUT (global utilities checkstop out) signal, 18-3 CLK_OUT (global utilities clock out) signal, 18-3, 18-18 Clocks clock out (CLK_OUT), see Global utilities DDR clock distribution, 9-27, 9-38 device clock signals summary, 4-3 see also Signals, clock device clocking operation, 4-23-4-26 CCB (platform) clock, 4-23 Ethernet clocks, 4-25 overview, 1-20 RapidIO clocks, 4-24 system clock/PCI clock, 4-23 e500 core clock multipliers, 5-11 see also Clocks, POR settings I2 C clock stretching, 11-17 clock synchronization, 11-16 input synchronization and digital filter, 11-16 LBC bus clocks and clock ratios, 13-4 clock ratio register (LCRR), 13-32 PCI/PCI-X clocking, 16-57, 16-62 POR settings e500 core PLL ratio, 4-13 RapidIO transmit clock source, 4-18 system/CCB PLL ratio, 4-12 TSEC inputs and outputs, 14-11 management clock out (EC_MDC), 14-55 Coherency rules L1 caches, 5-28 L2 cache, 7-22 Commands PCI, see PCI/PCI-X controller Configuration DDR, 9-10-9-17, 9-28 ECM CCB address configuration register (EEBACR), 8-3 CCB port configuration register (EEBPCR), 8-4 LBC configuration register (LBCR), 13-31 SDRAM configurations supported, 13-51 PCI configuration cycles, 16-70 configuration space header, 16-70 host accessing PCI configuration space, 16-72 type 0 configuration translation, 16-74 type 1 configuration translation, 16-76 PCI-X configuration transactions, 16-93 PIC global configuration register, 10-17 POR, see Power-on-reset (POR) RapidIO configuration/error injection registers, 17-34-17-36 TSEC interfaces, 14-131-14-151 Configuration space PCI/PCI-X addressing, 16-60 Configuration, control, and status accessing CCSR memory from external masters, 2-11 accessing CCSRs, 4-4 alternate configuration space (ALTBAR and ALTCAR), 4-5 boot page translation, 4-7 CCSR and RapidIO registers, 2-15 CCSR memory map, 2-10-2-16 CCSRBAR update guidelines, 4-4 memory map/register definition, 4-3 organization of CCSR memory, 2-11 Context ID registers, 20-24-20-25 Conventions notational, lxxxvii Core complex bus (CCB), see e500 core, core complex bus (CCB) CR (condition register), see e500 core, registers Critical interrupts see also Interrupt controller (PIC), critical interrupts see e500 core, critical interrupts CSRR0 (critical save/restore register 0), see e500 core, registers CSRR1 (critical save/restore register 1), see e500 core, registers CTR (count register), see e500 core, registers CTS, see DUART_CTS[0:1] D DACn (data address compare registers 1–2), see e500 core, registers Data cache see also e500 core, L1 caches Data cache, see L2 cache/SRAM Data processing with the e500 coherency module (ECM), 1-21 DBCRn (debug control register 0–2), see e500 core, registers DBSR (debug status register), see e500 core, registers DDR controller address signal mappings, 9-4 block diagram, 9-1, 9-25 clock distribution, 9-38 DLL operation, 9-27 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Index 2-2 Freescale Semiconductor D–D configuration, example, 9-28 data beat ordering, 9-45 debug mode signal selection (POR), 4-20, 4-21 source and target ID, 20-4, 20-26 DLL control (global utilities), 18-19 error checking and correcting (ECC), 9-46 testing ECC with error injection, 9-17-9-19 error handling, 9-19, 9-48 features, 9-2 functional description, 9-25 initialization/application information, 9-49 interrupts, 9-22 memory map, 9-9 modes of operation, 9-3 overview, 1-14 page mode and logical bank retention, 9-45 performance monitor events, 19-16 register descriptions, 9-9 by acronym, see Register Index configuration registers, 9-10-9-17 error handling registers, 9-19-9-25 error injection registers, 9-17-9-19 SDRAM operation, 9-30 address multiplexing, 9-31 initialization sequence, 9-50 JEDEC standard interface commands, 9-32 mode-set command timing, 9-38 organizations supported, 9-31 refresh operation, 9-41 power-saving modes, 9-43 timing, 9-42 registered DIMM mode, 9-39 timing, 9-34 write timing adjustments, 9-40 self-refresh in sleep mode, 9-44 signals summary, 9-3 see also Signals, DDR DEAR (data exception address register), see e500 core, registers Debug modes and watchpoint monitor signals summary, 20-6 see also Signals, debug and watchpoint monitor/trace buffer block diagram, 20-1 DDR signal selection (POR) ECC pins used for debug, 4-21 DDR source ID debug modes, 20-4, 20-26 source ID on debug signals, 20-28 source ID on ECC pins, 20-28 DDR/LBC signal selection (POR), 4-20 e500 core registers, 6-42-6-48 features, 20-3 functional description, 20-26 LBC source ID debug mode, 13-4, 20-4, 20-26 memory map/register definition, 20-10 modes of operation (set at POR), 20-3 overview, 20-2 PCI/PCI-X debug configuration (POR), 4-20 source ID debug mode, 20-5, 20-26 performance monitor events, 19-26 POR status (global utilities), 18-9 READY negation, 4-2 software debug context ID registers, 20-24 trace buffer, see Trace buffer watchpoint, see Watchpoint monitor DEC (decrementer register), see e500 core, registers DECAR (decrementer auto-reload register), see e500 core, registers Delay-locked loops (DLLs) DDR DLL control, 18-19 LBC DLL control, 18-20 DMA channel 2 and 3 signal select, 18-12 DMA controller block diagram, 15-1 channel operation, 15-28 bandwidth control, 15-36 channel abort, 15-35 channel state, 15-36 stride size and distance, 15-36 descriptor formats, 15-37 error handling, 15-37 features, 15-2 functional description, 15-28 interrupts, 15-10-15-14, 15-16, 15-24, 15-27, 15-37 limitations and restrictions, 15-40 memory map/register definition, 15-6 modes of operation, 15-2 basic mode transfer, 15-29 basic chaining mode, 15-31 basic chaining single-write start mode, 15-31 basic direct mode, 15-29 basic direct single-write start mode, 15-30 channel continue mode for cascading transfer chains, 15-34 basic channel continue mode, 15-35 extended mode, 15-35 extended DMA mode transfer, 15-32 extended chaining mode, 15-32 extended chaining single-write start mode, 15-33 extended direct mode, 15-32 extended direct single-write start mode, 15-32 external control mode transfer, 15-33 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor Index 2-3 E–E overview, 1-18, 15-1 performance monitor events, 19-18 register descriptions, 15-10-15-28 by acronym, see Register Index signal select—channel 2 and 3, 18-12, 18-29 signals summary, 15-5 see also Signals, DMA controller system considerations, 15-41 unusual scenarios, 15-43 DMA to configuration and control registers, 15-44 DMA to DUART, 15-44 DMA to e500 core, 15-43 DMA to Ethernet, 15-44 DMA to I2C, 15-44 transfer interfaces, 15-37 DMA_DACK[0:3] (DMA acknowledge) signals, 15-6 DMA_DDONE[0:3] (DMA done) signals, 15-6 DMA_DREQ[0:3] (DMA request) signals, 15-6 Doorbell message controller see RapidIO controller, message unit Doze mode, 1-19, 18-23 see also Global utilities, power management DUART asynchronous communication bits, 12-2 parity bit, 12-22 START bit, 12-21 STOP bit, 12-22 baud-rate generator logic, 12-22 block diagram, 12-2 divisor latch access bit (ULCRn[DLAB]), 12-4, 12-12 error handling, 12-23 framing error, 12-9, 12-16, 12-21, 12-22, 12-23 overrun error, 12-23 parity error, 12-23 errors detected, 12-3 features, 12-2 functional description, 12-20 initialization/application information, 12-25 interrupts interrupt control logic, 12-25 interrupt enable and control registers, 12-9-12-11 memory map/register definition, 12-4 modes of operation, 12-3 DMA mode selection, 12-24 FIFO mode, 12-24 interrupts, 12-24 local loop-back mode, 12-22 overview, 1-16, 12-1 PC16450 UART compatibility, 12-2 performance monitor events, 19-27 register descriptions, 12-4, 12-6-12-20 by acronym, see Register Index UART0 register offsets, 12-4 UART1 register offsets, 12-4 serial interface data format, 12-2 serial interface operation, 12-20-12-22 data transfer, 12-21 START bit, 12-21 STOP bit, 12-22 transaction protocol example, 12-21 signals summary, 12-3 see also Signals, DUART E e500 coherency module (ECM) block diagram, 8-1 CCB arbiter, 8-9 CCB interface, 8-10 configuration CCB address configuration register (EEBACR), 8-3 CCB port configuration register (EEBPCR), 8-4 error handling error handling registers, 8-5-8-8 features, 8-2 functional description, 8-9 global data multiplexor, 8-10 I/O arbiter, 8-9 initialization/application information, 8-10-8-11 interrupts ECM error enable register (EEER), 8-6 memory map/register definition, 8-2 overview, 1-14, 8-1 performance monitor events, 19-18 register descriptions, 8-3 by acronym, see Register Index transaction queue, 8-9 e500 core block diagram, 5-1 Book E architecture auxiliary processing units (APUs), 5-3 branch target buffer (BTB) locking, 5-5, 5-13 cache line lock and unlock, 5-5, 5-12 integer select (isel), 5-5 machine check, 5-5, 5-12 performance monitor, 5-5, 5-12, 5-29 single-precision floating-point (SPFP), 5-12 future upward compatibility and SPE APU, 5-3 legacy support of PowerPC architecture, 5-31 exception handling, 5-32 instruction set compatibility, 5-31 little-endian mode, 5-33 memory management unit (MMU) and TLBs, 5-32 reset operation, 5-32 boot mode (POR), 4-15 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Index 2-4 Freescale Semiconductor E–E branch operations registers, 6-9-6-11 branch target buffer (BTB) registers, 6-25-6-27 branch target buffer (BTB) locking, 5-5 instructions, 5-13 cache line lock and unlock APU, 5-5 instructions, 5-12 clock multipliers, 5-11 see also Clocks, POR settings computational operations registers, 6-8-6-9 core complex bus (CCB), 5-9, 5-29 critical interrupts, 5-22 data cache, see e500 core, L1 caches debug registers, 6-42-6-48 device implementation details, 5-33 exceptions and interrupt handling, 5-20 critical interrupts, 5-22 interrupt classes, 5-21 interrupt latencies (upper bound), 5-22 interrupt registers, 5-22 interrupt types, 5-21 execution units load/store unit, 5-8 multiple-cycle unit (MU), 5-7 features, 5-5 Freescale Semiconductor Book E implementation standards (EIS), 5-3 future compatibility warning and SPE APU, 1-9, 5-3, 5-33, 6-12 hardware implementation-dependent registers (HID0–1), 6-27-6-30 instruction flow, 5-14 branch detection and prediction, 5-14 execution pipeline, 5-15 instruction pipeline stages complete and write-back, 5-17 decode/dispatch, 5-16 execute, 5-17 instruction fetch, 5-16 issue queues (BIQ, GIQ), 5-16 interrupts registers, 6-18-6-24 sources, 10-4 L1 caches, 5-20 cache coherency, 5-9, 5-28 cache control instructions, 5-28 registers, 6-30-6-34 structure, 5-6 machine check APU, 5-5 rfmci instruction, 5-12 memory coherency atomic update memory references, 5-28 memory access ordering, 5-28 memory management unit (MMU), 5-24 address translation, 5-26 registers, 6-34-6-42 MMU assist registers (MAS1–MAS4, MAS6), 5-27 process ID registers (PID0–PID2), 5-27 TLB coherency, 5-27 TLB instructions, 5-26, 5-27 two-level MMU structure, 5-25 memory/cache attributes (WIMGE), 5-29 overview, 1-8, 5-1 performance monitor, 5-29 instructions, 5-12 purposes, 5-5 registers, 6-51-6-54 power management, 5-10 processor control registers, 6-11-6-15 programming model instruction set, 5-12 registers diagram, 5-18 registers BBEAR (branch buffer address register), 6-25 BBTAR (branch buffer target address register), 6-25 BUCSR (branch unit control and status register), 6-26 CR (condition register), 6-9 CSRR0 (critical save/restore register 0), 6-18 CSRR1 (critical save/restore register 1), 6-19 CTR (count register), 6-11 DACn (data address compare registers 1–2), 6-48 DBCRn (debug control register 0–2), 6-42-6-46 DBSR (debug status register), 6-46 DEAR (data exception address register 0), 6-19 DEC (decrementer register), 6-17 DECAR (decrementer auto-reload register), 6-17 ESR (exception syndrome register), 6-20 GPRs (general-purpose registers), 6-8 HID0 (hardware implementation-dependent register 0), 6-27 HID1 (hardware implementation-dependent register 1), 6-28 IACn (instruction address compare registers 1–2), 6-47 IVORn (interrupt vector offset registers), 6-19 IVPR (interrupt vector prefix register), 6-19 L1CFG0 (L1 cache configuration register 0), 6-32 L1CFG1 (L1 cache configuration register 1), 6-33 L1CSR0 (L1 cache status and control register 0), 6-30 L1CSR1 (L1 cache status and control register 1), 6-31 LR (link register), 6-11 MAS0–MAS6 (MMU assist registers 0–6), 5-27, 6-37-6-42 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor Index 2-5 F–F MCAR (machine check address register), 6-22 MCSR (machine check syndrome register), 6-23 MCSRR0 (machine check save/restore register 0), 6-21 MCSRR1 (machine check save/restore register 1), 6-22 MMUCFG (MMU configuration register), 6-35 MMUCSR0 (MMU control and status register 0), 6-34 MSR (machine state register), 6-12 PIDn (process ID registers 0–2), 5-27, 6-34 PIR (processor ID register), 6-14 PMCn (performance monitor counter registers 0–3), 5-30, 6-54 PMGC0 (performance monitor global control register 0), 5-30, 6-52 PMLCan (performance monitor local control registers a0–a3), 5-30, 6-52 PMLCbn (performance monitor local control registers b0–b3), 5-30, 6-53 PVR (processor version register), 5-4, 6-14 SPEFSCR (signal processing and embedded floating-point status and control register), 6-48 SPRGn (software-use registers 0–7), 6-24 SRR0 (save/restore register 0), 6-18 SRR1 (save/restore register 1), 6-18 SVR (system version register), 5-4, 6-14 TBL (time base lower register), 6-17 TBU (time base upper register), 6-17 TCR (timer control register), 6-15 TLB0CFG (TLB0 configuration register), 6-35 TLB1CFG (TLB1 configuration register), 6-36 TSR (timer status register), 6-16 UPMCn (user performance monitor counter registers 0–3), 5-30, 6-54 UPMGC0 (user performance monitor global control register 0), 5-30, 6-52 UPMLCan (user performance monitor local control registers a0–a3), 5-30, 6-52 UPMLCbn (user performance monitor local control registers b0–b3), 5-30, 6-53 USPRG0 (user software-use register 0), 6-24 XER (integer exception register), 6-8 signal processing engine (SPE) registers, 6-48 single-precision floating-point (SPFP) instructions, 5-12 software-use SPRs, 6-24 time base RTC (real time clock) signal options, 4-3, 4-25 timer registers, 6-15-6-18 translation lookaside buffers (TLBs), see e500 core, memory management unit (MMU) EC_GTX_CLK125 (TSEC gigabit transmit 125 MHz source) signals, 14-10 EC_MDC (TSEC management data clock) signals, 14-11 EC_MDIO (TSEC management data input/output) signals, 14-11 Error handling DDR, 9-19-9-25, 9-48 DMA, 15-37 DUART, 12-3, 12-23 framing error, 12-9, 12-16, 12-21, 12-22, 12-23 overrun error, 12-23 parity error, 12-23 ECM error handling registers, 8-5-8-8 FEC, 21-61-21-63 I2C interface boot sequencer mode, 11-17, 11-18 L2 cache/SRAM, 7-35 error handling registers, 7-12 error injection, 7-13 LBC transfer error registers, 13-26-13-30 PCI/PCI-X address/data parity, 16-66, 16-78, 16-79 error management registers, 16-28-16-35 reporting, 16-79 PERR and SERR signals, 16-79, 16-96 target-initiated termination, 16-65 retry transactions, 16-66 target-abort, 16-66 target-disconnect, 16-66 PCI-X, 16-96 address/data parity, 16-96 reporting, 16-96 RapidIO see RapidIO controller, error handling TSEC, 14-123-14-124 ESR (exception syndrome register), see e500 core, registers External system configuration POR (LAD[0:31]) status, 4-23, 18-9 External writes, see L2 cache/SRAM, stashing F Fast Ethernet controller, see FEC Features, overview of device features, 1-2 FEC controller block diagram, 21-5 buffer descriptors, 21-63 receive buffer descriptor (RxBD), 21-67 transmit data buffer descriptor (TxBD), 21-65 channel operation, 21-49 flow control, 21-59 frame reception, 21-53 frame recognition, 21-54 destination address recognition, 21-55 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Index 2-6 Freescale Semiconductor G–G frame transmission, 21-52 initialization sequence, 21-50 hardware controlled initialization, 21-50 user initialization, 21-50 inter-packet gap time, 21-61 interrupt handling, 21-59-21-60 clocks operation, 4-25 error handling, 21-61-21-63 features, 21-6 functional description, 21-48 hash function hash table algorithm, 21-57 hash table effectiveness, 21-57 registers, 21-45 interrupts, 21-59 interrupt registers, 21-13-21-18 MAC functionality, 21-36-21-45 configuration, 21-36 CSMA/CD controlling, 21-36 packet collisions, 21-36 packet flow, 21-37 registers, 21-38 memory map/register definition, 21-9 detailed memory map, 21-10-21-13 top level module map, 21-9 modes of operation, 21-7 10 Mbps and 100 Mbps MII operation, 21-7 address recognition options, 21-7 full and half-duplex operation, 21-7 internal and external loop back, 21-61 overview, 1-16, 21-5 physical interface connections, 21-48 media-independent interface (MII), 21-48 register descriptions, 21-13 by acronym, see Register Index FIFO control and status registers, 21-21-21-25 general control and status registers, 21-13-21-21 hash function registers, 21-45-21-46 MAC registers, 21-36-21-45 receive control and status registers, 21-31-21-36 transmit control and status registers, 21-25-21-31 signals, 21-7-21-9 see also Signals, FEC soft reset and reconfiguring procedure, 21-51 FEC_COL (FEC collision input) signals, 21-8 FEC_CRS (FEC carrier sense input) signals, 21-8 FEC_RX_CLK (FEC receive clock) signals, 21-8 FEC_RX_DV (FEC receive data valid) signals, 21-8 FEC_RX_ER (FEC receive error) signals, 21-8 FEC_RXD[3:0] (FEC receive data in) signals, 21-8 FEC_TX_CLK (FEC transmit clock in) signals, 21-8 FEC_TX_EN (FEC transmit data valid in) signals, 21-8 FEC_TX_ER (FEC transmit error in) signals, 21-9 FEC_TXD[3:0] (FEC transmit data out) signals, 21-8 G General-purpose I/O (PCI and TSEC2) see Global utilities Global utilities clock out CLK_OUT signal, 18-3, 18-18 clock out control register (CLKOCR), 18-18 overview, 18-2 DDR DLL control register (DDRDLLCR), 18-19 DMA signal multiplex control register (PMUXCR), 18-12, 18-29 features, 18-1 functional description, 18-21 general-purpose I/O signals (PCI and TSEC2), 18-1 control register (GPIOCR), 18-10 input data register (GPINDR), 18-12 operation of, 18-29 output data register (GPOUTDR), 18-11 interrupt and local bus signal multiplexing, 18-2, 18-12 operation, 18-29 interrupts and power management, 18-27 LBC DLL control register (LBDLLCR), 18-20 machine check summary sources of mcp (MCPSUMR), 18-16 memory map/register definition, 18-3 overview, 18-1 POR configuration boot mode status register (PORBMSR), 18-5 debug mode status register (PORDBGMSR), 18-9 device status register (PORDEVSR), 18-7 I/O impedance status register (PORIMPSCR), 18-6 LAD[0:31] external system configuration (GPPORCR), 4-23, 18-9 PLL status register (PORPLLSR), 18-4 see also Power-on reset (POR) power management and interrupts, 18-27 and snooping, 18-28 block disable (DEVDISR), 18-13, 18-22 CKSTP_IN and core-stopped mode, 18-22 core and device control bits, 18-24 core and device modes, 18-21 device mode control and status register (POWMGTCSR), 18-15 doze mode, 18-23 dynamic power management, 18-22 features, 18-1 functional description, 18-21 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor Index 2-7 H–I nap mode, 18-23 power-down sequence, 18-24 sleep mode, 18-24, 18-28 software considerations, 18-28 processor version register (PVR), 18-17 register descriptions, 18-4 by acronym, see Register Index signals summary, 18-2 see also Signals, global utilities snooping and power management, 18-28 system version register (SVR), 18-18 GPCM (LBC general-purpose chip-select machine), 13-38 see also Local bus controller (LBC) GPRs (general-purpose registers), see e500 core, registers H Hash table algorithm, see FEC controller, hash function Hash table algorithm, see TSEC, hash function HID0 (hardware implementation-dependent register 0), see e500 core, registers HID1 (hardware implementation-dependent register 1), see e500 core, registers Host/agent configuration (POR), PCI and RapidIO, 4-14 HRESET (hard reset) signal, 4-2, 4-9 HRESET_REQ (hard reset request) signal, 4-2, 11-17, 11-18 I I/O impedance LBC and PCI/PCI-X signals control and status register (global utilities), 18-6 PCI/PCI-X interface (POR), 4-19 I/O requirements, 18-28 I/O space PCI/PCI-X addressing, 16-60 I2C controller overview, 1-15 I2C interface arbitration arbitration control, 11-14 loss of arbitration—forcing of slave mode, 11-23 procedure for arbitration, 11-14 block diagram, 11-1 boot sequencer POR configuration, 4-16 boot sequencer mode, 11-2, 11-17-11-19 error condition behavior, 11-17, 11-18 calling address match condition, 11-5 clock control, 11-15 clock stretching, 11-17 clock synchronization, 11-16 input synchronization and digital filter, 11-16 master mode, 11-16 slave mode, 11-16 data transfer, 11-12 error handling boot sequencer mode, 11-17, 11-18 features, 11-2 frequency divider frequency divider register (I2CFDR), 11-5 functional description, 11-10 handshaking, 11-15 implementation details, 11-13 address compare, 11-14 control transfer, 11-13 transaction monitoring, 11-13 initialization/application information, 11-20-11-24 boot sequencer mode, see I2C interface, boot sequencer mode generation of SCL when SDA low, 11-22 initialization sequence, 11-20 post-transfer software response, 11-21 repeated START generation, 11-22 START generation, 11-11, 11-21 STOP generation, 11-12, 11-22 interrupts calling address match condition, 11-5 flowchart for interrupt service routine, 11-23 interrupt after transfer, 11-21 interrupt enable bit (I2CCR[MIEN]), 11-7 interrupt on START, 11-21 interrupt pending status bit (I2CSR[MIF]), 11-9 interrupt-driven byte-to-byte transfers, 11-2 read of last byte, 11-22 slave mode interrupt service routine guidelines, 11-23 for slave transmitter routine, 11-23 loss of arbitration, 11-23 memory map/register definition, 11-4 modes of operation, 11-2 boot sequencer mode, 11-2, 11-17-11-19 interrupt-driven byte-to-byte data transfer, 11-2 master mode, 11-2 slave mode, 11-2 overview, 11-2 register descriptions, 11-4 by acronym, see Register Index signals summary, 11-3 see also Signals, I2C transaction protocol, 11-10 handshaking, 11-15 repeated START condition, 11-3, 11-12 slave address transmission, 11-11 START condition, 11-3, 11-11, 11-21 STOP condition, 11-3, 11-12, 11-22 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Index 2-8 Freescale Semiconductor I–I IACn (instruction address compare registers 1–2), see e500 core, registers Initialization DDR (initialization and application information), 9-49 ECM (initialization and application information), 8-10-8-11 I2C interface (initialization and application information), 11-20-11-24 boot sequencer mode, see I2C interface, boot sequencer mode generation of SCL when SDA low, 11-22 initialization sequence, 11-20 post-transfer software response, 11-21 repeated START generation, 11-22 START generation, 11-11, 11-21 STOP generation, 11-12, 11-22 L2 cache/SRAM, 7-34-7-36 LBC (initialization and application information), 13-84-13-121 LBC SDRAM power-on initialization, 13-52 PCI/PCI-X (initialization and application information), 16-99-16-101 PIC (initialization and application information), 10-48 RapidIO (initialization and application information), 17-122 TSEC (initialization and application information), 14-131-14-151 see also TSEC, configuration watchpoint monitor and trace buffer, 20-33 Intel PC133 SDRAM commands (LBC), 13-52 Interrupt controller (PIC) block diagram, 10-1, 10-43 configuration (global), 10-17 critical interrupts, 10-7, 10-27, 10-33 destination (interrupt routing), 10-33 critical interrupt to core, 10-7 external interrupt to core, 10-7 IRQ_OUT, 10-7 see also e500 core, critical interrupts end of interrupt (EOI), 10-42, 10-46 external interrupts routed to critical interrupt (cint), 10-28 routed to IRQ_OUT, 10-26 features, 10-3 flow (interrupt processing), 10-43 functional description, 10-43 global timers, 10-20, 10-47 cascading of timers, 10-24, 10-25 clocking of timers, 10-21, 10-25 RTC (real time clock) signal options, 4-3, 4-25, 10-24, 10-25 initialization/application information, 10-48 interrupt acknowledge (IACK) signaling, 10-41, 10-46 interrupt routing (mixed mode), 10-7, 10-43 interrupt source priorities, 10-45 memory map/register definition, 10-9 messaging interrupts, 10-26, 10-28, 10-47 modes of operation, 10-5, 10-17 mixed mode, 10-5 pass-through mode (to support external interrupt controllers), 10-6 nesting of interrupts, 10-46 overview, 1-15, 10-2 performance monitor events, 19-20 power management (wake up conditions), 10-4 processor core interrupt sources, 10-4 critical interrupt (cint) sources, 10-27 processor current task priority, 10-45 programming guidelines, 10-48 changing interrupt source configuration, 10-50 register descriptions, 10-16 by acronym, see Register Index, 10-16 global registers, 10-16-10-20 global timer registers, 10-20-10-25 interrupt source configuration registers, 10-20-10-23, 10-32-10-37 message registers, 10-30-10-32 non-accessible registers in-service register (ISR), 10-45 interrupt pending register (IPR), 10-43 interrupt request register (IRR), 10-43 per-CPU registers, 10-38-10-42 performance monitor mask registers, 10-28-10-30 summary registers, 10-26-10-28 reset of PIC, 10-17, 10-47 reset processor from software, 10-18, 10-46 signals summary, 10-8 see also Signals, PIC simultaneous interrupts, priorities, 10-45 sources of interrupts, 10-6 internal (to PIC) interrupt destinations, 10-27, 10-28 internal (to PIC) interrupt sources, 10-7 spurious vector generation, 10-19, 10-46 vendor identification, 10-17 Interrupts DDR, 9-22 DMA, 15-10-15-14, 15-16, 15-24, 15-27, 15-37 DUART interrupt control logic, 12-25 interrupt enable and control registers, 12-9-12-11 e500 core registers, 6-18-6-24 ECM interrupt register (ECM error enable register—EEER), 8-6 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor Index 2-9 J–L FEC, 21-59-21-60 interrupt registers, 21-13-21-18 I2C interface calling address match condition, 11-5 flowchart for interrupt service routine, 11-23 interrupt after transfer, 11-21 interrupt enable bit (I2CCR[MIEN]), 11-7 interrupt on START, 11-21 interrupt pending status bit (I2CSR[MIF]), 11-9 interrupt-driven byte-to-byte transfers, 11-2 read of last byte, 11-22 slave mode interrupt service routine guidelines, 11-23 for slave transmitter routine, 11-23 loss of arbitration, 11-23 IRQ[9:11] signal select, 18-12, 18-29 LBC interrupt register, 13-28 PCI/PCI-X error enable register, 16-31 performance monitor (PIC), 19-20 power management and interrupts (global utilities), 18-27 RapidIO message unit doorbell controller interrupts, 17-121 inbox controller interrupts, 17-118 outbox controller interrupts, 17-116 port notification/fatal error interrupt enable register (PNFEIER), 17-49 see also Interrupt controller (PIC) TSEC, 14-120-14-123 interrupt registers, 14-20-14-25 IRQ[0:11] (interrupt request 0–11) signals, 10-8 IRQ[9:11] signal select global utilities, 18-12 IRQ_OUT (interrupt request out) signal, 10-9, 10-26 IVORn (interrupt vector offset registers), see e500 core, registers IVPR (interrupt vector prefix register), see e500 core, registers J JEDEC SDRAM commands (LBC), 13-52 JTAG test access port signals summary, 20-6 see also Signals, JTAG, 20-6 L L1CFG0 (L1 cache configuration register 0), see e500 core, registers L1CFG1 (L1 cache configuration register 1), see e500 core, registers L1CSR0 (L1 cache status and control register 0), see e500 core, registers L1CSR1 (L1 cache status and control register 1), see e500 core, registers L2 cache/SRAM allocation of lines, 7-29 block diagram, 7-1 coherency rules, 7-22 configuration and organization, 7-3 error handling ECC errors, 7-35 tag parity errors, 7-36 error handling registers, 7-12 error injection, 7-13 external writes, see stashing flash clearing, instruction and data locks, 7-27 initialization/application information, 7-34-7-36 invalidation, 7-30 locking clearing locks on selected lines, 7-26 entire, 7-25 programmed memory ranges, 7-25 selected lines, 7-25 with stale data, 7-27 memory map/register definition, 7-6 memory-mapped SRAM coherency rules, 7-24 memory-mapped windows, 2-4 operation, 7-29 organization, 7-3 overview, 1-12, 7-1 performance monitor events, 19-26 PLRU bit update considerations, 7-28 register descriptions, 7-7-7-20 replacement policy, 7-27 SRAM features, 7-2 stashing, 7-21 state transitions, 7-30 due to core-initiated transactions, 7-31 due to system-initiated transactions, 7-33 timing, 7-21 LA[27:31] (LBC non-multiplexed address) signals, 13-7 LAD[0:31] (LBC multiplexed address/data) signals, 13-8 LALE (LBC external address latch enable) signal, 13-6, 13-35 LBCTL (LBC data buffer control) signal, 13-7, 13-37 LBS[0:3] (LBC UPM byte select) signals, 13-6 LCK[0:2] (LBC clock) signals, 13-8 LCKE (LBC clock enable) signal, 13-8 LCS[0:7] (LBC chip select) signals, 13-6 LCS[5:7] signal select global utilities, 18-12 LCS0 (LBC chip select 0) signal, 13-50 LDP[0:3] (LBC data parity) signals, 13-8, 13-38 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Index 2-10 Freescale Semiconductor L–L LGPL0 (LBC GP line 0) signal, 13-6 LGPL1 (LBC GP line 1) signal, 13-6 LGPL2 (LBC GP line 2) signal, 13-6 LGPL3 (LBC GP line 3) signal, 13-7 LGPL4 (LBC GP line 4) signal, 13-7 LGPL5 (LBC GP line 5) signal, 13-7 LGTA (LBC GPCM transfer acknowledge) signal, 13-7, 13-49 Local access windows, 2-3-2-9 ATMUs, see Address translation and mapping units (ATMUs) configuring local access windows, 2-7 distinguishing local access windows from other mapping functions, 2-8 illegal interactions between inbound ATMUs and local access windows, 2-9 between local access windows and DDR SDRAM chip selects, 2-8 L2 cache/SRAM window interactions, 2-4 precedence if overlapping among themselves, 2-7 precedence if overlapping with L2 cache/SRAM windows, 2-4 registers, 2-6-2-7 by acronym, see Register Index Local address map, 1-20 see also Local access windows Local bus controller (LBC) address and address space checking, 13-34 address mask field—option registers, 13-13 atomic bus operations, 13-37 block diagram, 13-1 boot chip-select operation, 13-50 bus monitor, 13-38 bus turnaround, 13-87 additional address phases (UPM cycles), 13-88 address following read, 13-87 read data following address, 13-88 read-modify-write cycle (parity), 13-88 clocks and clock ratios, 13-4 clock ratio register (LCRR), 13-32 configuration LBC configuration register (LBCR), 13-31 debug mode signal selection (POR), 4-20 source and target ID, 20-4, 20-26 DLL control (global utilities), 18-20 DSP hosts (interface to), 13-103 MSC8101 HDI16 interface, 13-103 MSC8102 DSI interface, 13-107 TI TMS320Cxxxx interface, 13-118 error handling transfer error registers, 13-26-13-30 external access termination (LGTA), 13-49 features, 13-2 functional description, 13-33 general-purpose chip-select machine (GPCM), 13-38 chip-select and write enable negation timing, 13-44 chip-select assertion timing, 13-44 extended hold time on read accesses, 13-48 GPCM mode registers, 13-14 output enable timing, 13-48 programmable wait state configuration, 13-44 relaxed timing, 13-45 timing configuration, 13-39 initialization/application information, 13-84-13-121 interrupts transfer error interrupt enable register (LTEIR), 13-28 LCS[5:7] signal select, 18-12, 18-29 memory map/register definition, 13-9 memory refresh timer prescaler, 13-21 modes of operation, 13-3 bus clock and clock ratios, 13-4 GPCM mode, registers, 13-14 power-down mode, 13-4 SDRAM mode, registers, 13-17 source ID debug mode, 13-4 UPM mode, registers, 13-16 output hold configuration (POR), 4-22 overview, 1-17, 13-2 parity generation and checking, 13-38, 13-101 performance monitor events, 19-25 peripherals, 13-84 GPCM timing, 13-86 hierarchy for very high speeds, 13-85 hierarchy on the local bus, 13-84 multiplexed address/data, 13-84 port sizes, 13-88 references, other, 13-4 register descriptions, 13-10 by acronym, see Register Index SDRAM interface, 13-51-13-61, 13-90 address multiplexing, 13-54 basic capabilities, 13-90 commands (Intel PC133 and JEDEC), 13-52 configurations supported, 13-51 device-specific parameters, 13-55 limitations, 13-92-13-100 maximum SDRAM supported, 13-91 page hit checking, 13-53 page management, 13-54 parity support, 13-101 power-on initialization, 13-52 refresh, 13-61 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor Index 2-11 M–M SDRAM mode registers, 13-17, 13-22 timing, 13-58 activate-to-read/write interval, 13-56 CAS latency, 13-56 external buffers, 13-57 MODE-SET commands, 13-60 precharge-to-activate interval, 13-55 refresh recovery, 13-57 refresh timing, 13-61 write recovery, 13-56 transactions, 13-60 signals summary, 13-5 see also Signals, LBC UPM interfaces, 13-62-13-83 block diagram, 13-62 example interface, 13-78 extended hold time (reads), 13-77 programming the UPMs, 13-66 RAM array, 13-67 address multiplexing, 13-74 byte select signal timing, 13-72 chip select signal timing, 13-72 data timing, 13-75 general purpose signal timing, 13-73 LGPL[0:5] timing (LAST), 13-76 loop control, 13-73 RAM word definition, 13-68 REDO, 13-74 wait mechanism (WAEN), 13-76 signal timing, 13-66 synchronous UPWAIT (early transfer acknowledge), 13-77 UPM mode registers, 13-16, 13-18 UPM requests, 13-63 exception requests, 13-66 memory access requests, 13-64 refresh timer requests, 13-65 software requests, 13-65 ZBT SRAM interface, 13-101 LOE (LBC GPCM output enable) signal, 13-6 Low-voltage differential signaling see RapidIO, functionality lists, 8/16 LP-LVDS LPBSE (LBC parity byte select) signal, 13-7 LR (link register), see e500 core, registers LSDA10 (LBC SDRAM A10) signal, 13-6 LSDCAS (LBC SDRAM CAS) signal, 13-7 LSDDQM[0:3] (LBC SDRAM data mask) signal, 13-6 LSDRAS (LBC SDRAM RAS) signal, 13-6 LSDWE (LBC SDRAM write enable) signal, 13-6 LSYNC_IN (LBC DLL synchronization in) signal, 13-8 LSYNC_OUT (LBC DLL synchronization out) signal, 13-8 LWE[0:3] (LBC GPCM write enable) signals, 13-6 M MA[0:14] (DDR address bus) signals, 9-6 MAC functionality see FEC controller, MAC functionality see TSEC, MAC functionality Machine check MCP (processor machine check) signal, 10-9 mcp summary register (MCPSUMR), 18-16 SRESET (soft reset) signal, 4-9 MAS0-MAS6 (MMU assist registers 0–6), see e500 core, registers MBA[0:1] (DDR logical bank address) signals, 9-6 MCAR (machine check address register), see e500 core, registers MCAS (DDR column address strobe) signal, 9-6 MCK[0:5] (DDR clock output complement) signals, 9-8 MCK[0:5] (DDR clock output) signals, 9-8 MCKE (DDR clock enable) signal, 9-8 MCP (processor machine check) signal, 10-9 MCS[0:3] (DDR chip select) signals, 9-7 MCSR (machine check syndrome register), see e500 core, registers MCSRR0 (machine check save/restore register 0), see e500 core, registers MCSRR1 (machine check save/restore register 1), see e500 core, registers MDM[0:8] (DDR SDRAM data output mask) signals, 9-7 MDQ[0:8] (DDR data bus strobe) signals, 9-5, 9-27 MDVAL (DDR/LBC debug mode data valid) signal, 4-20, 13-9, 20-4, 20-7 MECC[0:5] (DDR error correcting code) signals as debug, 20-4, 20-8 MECC[0:7] (DDR error correcting code) signals, 4-21, 9-6 Memory management unit (MMU), see e500 core, memory management unit (MMU) Memory maps CCSR memory, 2-4 accessing CCSR memory from external masters, 2-11 CCSR and RapidIO registers, 2-15 CCSR map, complete list of memory-mapped registers (by offset), 2-16 CCSR organization, 2-11 CCSR registers, 2-10-2-16 device-specific utilities, 2-15 general utilities registers, 2-13 programmable interrupt controller (PIC) space, 2-14 configuration, control, and status registers, 4-3 DDR controller, 9-9 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Index 2-12 Freescale Semiconductor N–P illegal interaction between local access windows and DDR SDRAM chip selects, 2-8 debug, watchpoint, and trace buffer registers, 20-10 device memory map address translation and mapping, 2-3 overview and example, 2-1 DMA, 15-6 DUART, 12-4 ECM, 8-2 FEC, 21-9 global utilities, 18-3 I2C, 11-4 interrupt controller (PIC), 10-9 L2 cache/SRAM, 7-6 LBC, 13-9 PCI/PCI-X, 16-15 performance monitor, 19-3 RapidIO, 17-9 TSEC, 14-13 Memory space PCI/PCI-X addressing, 16-59 Memory target queue performance monitor events, 19-17 Memory unit, see L2 cache/SRAM Message interrupts, see Interrupt controller (PIC), message interrupts Message unit see RapidIO controller, message unit, 17-109 MMUCFG (MMU configuration register), see e500 core, registers MMUCSR0 (MMU control and status register 0), see e500 core, registers MRAS (DDR row address strobe) signal, 9-7 MSR (machine state register), see e500 core, registers MSRCID[0:4] (DDR/LBC debug source ID) signals, 4-20, 13-9, 20-4, 20-8 MSYNC_IN (DDR DRAM DLL synchronization input) signal, 9-8 MSYNC_OUT (DDR DRAM DLL synchronization output) signal, 9-8 MWE (DDR write enable) signal, 9-7 see also L2 cache/SRAM Output hold see Power-on reset (POR), configuration P Page hit checking (LBC SDRAM), 13-53 Page management (LBC SDRAM), 13-54 PCI Local Bus Specification configuration registers see PCI/PCI-X controller, register descriptions PCI/PCI-X controller 64-bit/32-bit bus, 16-6 address bus decoding, 16-59 address translation and mapping unit (ATMU) inbound windows (4), 2-9, 16-24 outbound windows (4), 16-20 arbiter configuration (POR), 4-19, 16-5, 16-6 block diagram, 16-1 burst operations cache wrap mode, 16-59 linear incrementing, 16-59 bus arbitration, 16-5, 16-53 bus protocol, 16-56 burst operation, 16-57 command encodings, 16-57 PCI-X-specific protocol, see PCI/PCI-X controller, PCI-X operation clocking, 16-57, 16-62 commands command register, 16-38, 16-72 encodings, 16-57 interrupt-acknowledge transactions, 16-76 PCI-X encodings, 16-82 special-cycle, 16-77 configuration cycles, 16-70, 16-71 configuration space addressing, 16-60 host access example, 16-72 type 0 configuration translation, 16-74 type 1 configuration translation, 16-76 data bus width (POR), 4-19 debug configuration (POR), 4-20 debug mode source and target ID (PCI_AD[62:58]), 20-5, 20-26 error handling, 16-79 address/data parity, 16-66, 16-78, 16-79 detection and reporting, 16-78, 16-96 error management registers, 16-28-16-35 PCI-X, 16-96 address/data parity, 16-96 reporting PERR and SERR signals, 16-79, 16-96 target-initiated termination, 16-65 retry transactions, 16-66 N Nap mode, 1-19, 18-23 see also Global utilities, power management O OCeaN switch fabric, 1-20 On-chip memory as L2 cache, 1-13 as memory-mapped SRAM, 1-13 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor Index 2-13 P–P target-abort, 16-66 target-disconnect, 16-66 features, 16-4 functional description, 16-53 host/agent configuration (POR), 4-14 I/O impedance (POR), 4-19 I/O space addressing, 16-60 initialization/application information, 16-99-16-101 initiator/master operation, 16-3 interrupts error enable register, 16-31 latency timer, 16-43, 16-66, 16-72 memory map/register definition, 16-15 memory space addressing, 16-59 modes of operation, 16-5 agent configuration lock mode, 16-100 agent mode, 16-100 cache wrap mode, 16-59 host mode, 16-100 linear incrementing, 16-59 PCI-X mode selection (POR), 4-20 output hold configuration (POR), 4-21 overview, 1-18, 16-2 PCI-X operation attribute phase, 16-83 bus protocol, 16-81-16-98 command encodings, 16-82 configuration, 16-93 nonposted writes, 16-100 outbound read transaction alignment, 16-101 terminology, 16-81 transactions, 16-84-16-88, 16-89-16-92 completer attributes, 16-91 completion address, 16-90 split response, 16-89 split transactions, 16-89-16-92 wait state and terminations rules, 16-88 performance monitor events common events, 19-20 PCI-specific events, 19-22 POR configuration, 16-99 power management special-cycle operations, 16-77 register descriptions configuration header registers, 16-36-16-53, 16-72 32-bit memory base address register, 16-44 64-bit high memory base address register, 16-46 64-bit low memory base address register, 16-45 arbiter configuration register (PBACR), 16-50 base address registers, 16-43-16-46 base class code register, 16-42 bus function register (PBFR), 16-49 bus status register, 16-40, 16-61, 16-65 cache line size register, 16-43 capabilities pointer register, 16-47 command register, 16-38, 16-72 configuration and status register base address (PCSRBAR), 16-44 device ID register, 16-38 interrupt line register, 16-47 interrupt pin register, 16-48 latency timer register, 16-43 maximum grant (MAX GNT) register, 16-48 maximum latency (MAX LAT) register, 16-49 PCI-X capability pointer register, 16-51 PCI-X command register, 16-52 PCI-X next capabilities ID register, 16-51 PCI-X status register, 16-52 programming interface register, 16-41 revision ID register, 16-41 subclass code register, 16-42 subsystem ID register, 16-46 subsystem vendor ID register, 16-46 vendor ID register, 16-37 memory-mapped registers, 16-15 ATMU inbound registers, 16-24-16-28 ATMU outbound registers, 16-20-16-24 by acronym, see Register Index configuration access registers, 16-18-16-20 error management registers, 16-28-16-35 signals summary, 16-7 see also Signals, PCI/PCI-X target/slave operation, 16-4 target-abort termination, 16-66 target-disconnect cycles, 16-4, 16-66 target-initiated termination target-abort error, 16-66 target-disconnect, 16-4, 16-66 transactions dual address cycles (DACs), 16-68 fast back-to-back transactions, 16-68 interrupt-acknowledge transactions, 16-76 PCI-X, see PCI/PCI-X controller, PCI-X operation read transactions, 16-62 retry transactions, 16-66 special-cycle transactions, 16-77 timing diagrams, 16-61 transaction termination, 16-64 bus status register, termination status, 16-66 completion, 16-65 master-abort termination, 16-65 master-initiated, 16-65 target-initiated, 16-65, 16-66 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Index 2-14 Freescale Semiconductor P–P timeout, 16-65 write transactions, 16-61, 16-63 turnaround cycle, 16-61 PCI_AD[47:40] signals as GP I/O, see Global utilities, general-purpose I/O signals PCI_AD[62:58] (high-order PCI address) signals as debug, 20-4, 20-8 PCI_AD[63:0] (PCI address/data bus) signals, 16-8, 16-59 PCI_C/BE[7:0] (PCI command/byte enable) signals, 16-9, 16-57, 16-60, 16-61, 16-78 PCI_DEVSEL (PCI device select) signal, 16-60 PCI_FRAME (PCI frame) signal, 16-57 PCI_GNT[4:0] (PCI bus grant) signals, 16-10, 16-54 PCI_IDSEL (PCI initialization device) signal, 16-10 PCI_IRDY (PCI intitiator ready) signal, 16-11, 16-57 PCI_PAR (PCI parity) signal, 16-11, 16-78 PCI_PAR64 (PCI upper DWORD parity) signal, 16-12 PCI_PERR (PCI parity error) signal, 16-12, 16-79, 16-96 PCI_REQ[4:0] (PCI bus request) signals, 16-13, 16-54 PCI_REQ64 (PCI 64-bit transaction request) signal, 16-13 PCI_SERR (PCI system error) signal, 16-13, 16-79, 16-96 PCI_STOP (PCI stop) signal, 16-14, 16-61 PCI_TRDY (PCI target ready) signal, 16-14, 16-57 PCI_TRDY (target ready) signal, 16-65 Performance monitor (device) block diagram, 19-2 burstiness, 19-13, 19-28 control registers, 19-5-19-9 counters (PMCn) chaining, 19-12 registers, 19-9 triggering, 19-12 event counting, 19-11 events, 19-15-19-27 chaining, 19-27 DDR controller, 19-16 debug, 19-26 DMA controller, 19-18 DUART, 19-27 e500 coherency module (ECM), 19-18 interrupt controller (PIC), 19-20 L2 cache/SRAM, 19-26 local bus controller (LBC), 19-25 memory target queue, 19-17 PCI/PCI-X common events, 19-20 PCI-specific events, 19-22 RapidIO controller, 19-17 TSEC1, 19-22, 19-23 TSEC2, 19-24 events triggered by watchpoint monitor, 20-29 examples, 19-27 burstiness event, 19-13 burstiness event counting, 19-28 simple event counting, 19-27, 19-28 threshold event counting, 19-28 triggering event counting, 19-28 external signals, 19-3 features, 19-3 functional description, 19-10 interrupts, 19-10 interrupts (from PIC) to generate events, 10-28 masking interrupts (from PIC), 10-28 memory map/register definition, 19-3 overflow indication on TRIG_OUT, 20-26 overview, 19-1 threshold events, 19-11 Performance monitor (e500 core) counter registers, 5-30 Phase-locked loops (PLLs) POR status (global utilities), 18-4 PID n (process ID registers 0–2), see e500 core, registers PIR (processor ID register), see e500 core, registers PMCn (performance monitor counter registers 0–3), see e500 core, registers PMGC0 (performance monitor global control register 0), see e500 core, registers PMLCan (performance monitor local control registers a0–a3), see e500 core, registers PMLCbn (performance monitor local control registers b0–b3), see e500 core, registers Power management block disable block disable control (DEVDISR), 18-13, 18-22 LBC, 13-4 DDR interface, 9-43 device low-power modes, 18-21-18-28 control and status register (POWMGTCSR), 18-15 READY negation, 4-2 interrupts that cause wake-up, 10-4 overview, 1-19 PCI special-cycle operations, 16-77 see also Global utilities, power management Power-on reset (POR) configuration boot ROM location, 4-14 boot sequencer configuration, 4-16 clock e500 core PLL ratio, 4-13 system/CCB PLL ratio, 4-12 CPU boot configuration, 4-15 DDR debug mode (ECC pins used for debug), 4-21, 20-3 general-purpose (external system) configuration—LAD[0:31] (GPPORCR), 4-23 host/agent configuration (PCI and RapidIO), 4-14 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor Index 2-15 R–R memory debug select (DDR or LBC), 4-20, 20-3 output hold LBC output hold, 4-22 PCI/PCI-X output hold, 4-21 PCI data bus width, 4-19 PCI debug configuration, 4-20, 20-3 PCI I/O impedance, 4-19 PCI/PCI-X arbiter configuration, 4-19 PCI/PCI-X, modes of operation, 16-99 PCI-X mode selection, 4-20 RapidIO device ID, 4-18 RapidIO transmit clock source, 4-18 TSEC data width, 4-16 TSEC1 protocol, 4-17 TSEC2 protocol, 4-17 configuration reporting global utilities, 18-4, 18-5, 18-6, 18-7, 18-9 debug modes summary, 20-3 hard reset, 4-9 output signal states during reset, 3-17 reset configuration signals, 3-15 sequence of events, 4-9 and READY signal, 4-2, 4-10 Processor version (PVR), 18-17 Protocols PCI, see PCI/PCI-X controller, bus protocol PVR (processor version register), see e500 core, registers R RapidIO controller accept-all mode, 17-6 address translation and mapping unit (ATMU), 17-106 inbound ATMU translation, 2-9, 17-107 bypass mode, 17-108 errors, boundary crossing, 17-109 inbound windows (4 plus default), 17-40-17-43 LCSRBAR window, 17-108 special transactions and requirements, 17-108 outbound ATMU translation, 17-106 bypass mode, 17-107 outbound windows (8 plus default), 17-37-17-40 special transactions and requirements, 17-107 assembly identity capability, 17-14 clocks clock selection, transmit clock source (POR), 4-18, 17-8 operation, 4-24 command and status registers, 17-22-17-34 configuration configuration/error injection registers, 17-34-17-36 CRC checking modes, 17-5 device ID (POR), 4-18 error handling, 17-102 error checking disable mode, 17-6 error handling registers, 17-43-17-65 message unit doorbell error response conditions, 17-121 inbox error response conditions, 17-118 special outbox error response conditions, 17-117 packet error packet register 1 (PEPR1) packet error capture registers (9 types) (PECR1Tx), 17-53-17-58 packet error packet register 2 (PEPR2) packet error capture registers (9 types) (PECR2Tx), 17-58-17-61 features, 17-4 features not implemented, 17-5 functional description, 17-83 functionality of RapidIO interface, 17-89 functionality lists 8/16 LP-LVDS layer, 17-91 logical layer, 17-100 source transaction support list, 17-101 host/agent configuration (POR), 4-14 initialization/application information, 17-122 interrupts message unit doorbell controller interrupts, 17-121 inbox controller interrupts, 17-118 outbox controller interrupts, 17-116 port notification/fatal error interrupt enable register (PNFEIER), 17-49 low-voltage differential signaling, see RapidIO, functionality lists, 8/16 LP-LVDS mailbox command and status, 17-22 memory map/register definition, 17-9 message unit data message controller, 17-111 doorbell message controller, 17-119 doorbell controller interrupts, 17-121 doorbell queue entry format, 17-120 error response conditions, 17-121 inbound doorbell reception, 17-120 retry response conditions, 17-121 features, 17-110 inbox controller operation, 17-117 data message controller limitations and restrictions, 17-119 error response conditions, 17-118 inbox controller interrupts, 17-118 retry response conditions, 17-118 messaging, description, 17-111 modes of operation, 17-111 outbound descriptors, 17-72 outbox controller operation, 17-111 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Index 2-16 Freescale Semiconductor R–R chaining mode operation, 17-112 descriptor format, 17-115 direct mode operation, 17-112 interrupts, 17-116 special error case condition, 17-117 switching between modes, 17-115 overview, 1-19, 17-109 port-write controller structure, 17-121 registers, see RapidIO controller, registers modes of operation, 17-5 accept-all mode, 17-6 CRC checking modes, 17-5 error checking disable mode, 17-6 link response time-out disable mode, 17-6 output port driver disable mode, 17-6 output port enable mode, 17-6 packet response time-out disable mode, 17-6 transmit clock-select mode, 17-5 see also RapidIO controller, clocks, 17-5 overview, 1-18, 17-7 performance monitor events, 19-17 registers architectural registers, 17-13-17-34 block header command and status registers, 17-27-17-34 command and status registers, 17-22-17-34 implementation registers, 17-34-17-65 ATMU, 17-36-17-43 configuration/error injection, 17-34-17-36 error handling, 17-43-17-65 message unit registers, 17-65-17-83 doorbell registers, 17-77-17-81 inbound message registers, 17-72-17-76 outbound descriptors, 17-72 outbound message registers, 17-65-17-72 port-write registers, 17-81-17-83 signals summary, 17-7 see also Signals, RapidIO terminology, 17-4 transaction, packet, and control symbols, 17-83 READY signal, 4-2, 4-10, 20-25, 20-26 Registers by acronym (memory-mapped registers) see Register Index configuration, control, and status, 2-10-2-16, 4-3 device-specific utilities, 2-15 general utilities, 2-13 programmable interrupt controller (PIC) space, 2-14 context ID, 20-24-20-25 DDR configuration registers, 9-10-9-17 DLL control, 18-19 error handling registers, 9-19-9-25 error injection registers, 9-17-9-19 e500 core, see e500 core, registers ECM, 8-3 FEC FIFO control and status registers, 21-21-21-25 general control and status registers, 21-13-21-21 hash function registers, 21-45-21-46 MAC registers, 21-36-21-45 receive control and status registers, 21-31-21-36 transmit control and status registers, 21-25-21-31 global utilities, 18-4 POR boot mode status, 18-5 POR debug mode status, 18-9 POR device status, 18-7 POR external system configuration, 18-9 POR I/O impedance status, 18-6 POR PLL status, 18-4 I2C interface, 11-4 L2 cache/SRAM registers, 7-6-7-20 LBC, 13-10 DLL control, 18-20 local access window registers attributes registers (LAWAR0–LAWAR7), 2-6 base address registers (LAWBAR0–LAWBAR7), 2-6 PCI/PCI-X configuration header registers, 16-36-16-53, 16-72 see also PCI/PCI-X controller, registers memory-mapped registers ATMU inbound registers, 16-24-16-28 ATMU outbound registers, 16-20-16-24 configuration access registers, 16-18-16-20 error management registers, 16-28-16-35 performance monitor (e500 core) counter registers, 5-30 performance monitor, descriptions, 19-3 PIC, 10-16 global registers, 10-16-10-20 global timer registers, 10-20-10-25 interrupt source configuration registers, 10-20-10-23, 10-32-10-37 message registers, 10-30-10-32 non-accessible registers in-service register (ISR), 10-45 interrupt pending register (IPR), 10-43 interrupt request register (IRR), 10-43 per-CPU registers, 10-38-10-42 performance monitor mask registers, 10-28-10-30 summary registers, 10-26-10-28 processor version register (PVR), 18-17 RapidIO architectural registers, 17-13-17-34 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor Index 2-17 S–S block header command and status registers, 17-27-17-34 command and status registers, 17-22-17-34 implementation registers, 17-34-17-65 ATMU, 17-36-17-43 configuration/error injection, 17-34-17-36 error handling, 17-43-17-65 message unit registers, 17-65-17-83 doorbell registers, 17-77-17-81 inbound message registers, 17-72-17-76 outbound descriptors, 17-72 outbound message registers, 17-65-17-72 port-write registers, 17-81-17-83 system version register (SVR), 18-18 trace buffer, 20-16-20-23 trigger out source register, 20-25 TSEC FIFO control and status registers, 14-30-14-33 general control and status registers, 14-20-14-30 hash function registers, 14-87-14-89 MAC registers, 14-46-14-60 MIB registers, 14-60-14-87 receive control and status registers, 14-40-14-46 ten-bit interface (TBI) registers, 14-91-14-103 transmit control and status registers, 14-33-14-40 watchpoint monitor, 20-11-20-16 Reset core reset through PIC register, 10-18, 10-46 hard reset actions, 4-9 operations, 4-8 power-on reset (POR) configuration, see Power-on reset (POR), configuration sequence of events, 4-9 signals summary, 4-2 see also Signals, reset soft reset actions, 4-9 RIO_RCLK and RIO_RCLK (RapidIO receive clock and complement) signals, 17-8 RIO_RD[0:7] and RIO_RD [0:7] (RapidIO receive data and complement) signals, 17-8 RIO_RFRAME and RIO_RFRAME (RapidIO receive frame and complement) signals, 17-8 RIO_TCLK and RIO_TCLK (RapidIO transmit clock out and complement) signals, 17-8 RIO_TD[0:7] and RIO_TD[0:7] (RapidIO transmit data and complement) signals, 17-8 RIO_TFRAME and RIO_TFRAME (RapidIO transmit frame and complement) signals, 17-9 RIO_TX_CLK_IN and RIO_TX_CLK_IN (RapidIO transmit clock in and complement) signals, 17-9 RTC (real time clock) signal, 4-3, 4-25, 10-24, 10-25, 18-25 RTS, see DUART_RTS[0:1] S SCL (I2C serial clock) signal, 11-3, 11-4 SDA (I2C serial data) signal, 11-3, 11-4 SDRAM interface (LBC), 13-51-13-61 see also Local bus controller (LBC), SDRAM interface Serial data/clock, see I2C interface, 11-1 Signals clock RTC (real time clock), 4-3, 4-25, 10-24, 10-25, 18-25 SYSCLK (system clock input), 4-3 complete signal listing alphabetical reference, 3-9 configuration signals, sampled at POR, 3-15 see also Power-on reset (POR) figure showing groupings, 3-1 output signal states at power-on reset, 3-17 reference by functional block, 3-4 DDR MA[0:14] (address bus), 9-6 MBA[0:1] (logical bank address), 9-6 MCAS (column address strobe), 9-6 MCK[0:5] (DDR clock output complements), 9-8 MCK[0:5] (DDR clock outputs), 9-8 MCKE (DDR clock enable), 9-8 MCS[0:3] (chip selects), 9-7 MDM[0:8] (SDRAM data output mask), 9-7 MDQS[0:8] (data bus strobes), 9-5, 9-27 MDVAL (debug mode data valid), 4-20, 20-4, 20-7 MECC[0:5] (error correcting code) as debug signals, 20-4, 20-8 MECC[0:7] (error correcting code), 4-21 MECC[0:7] (error correcting codes), 9-6 MRAS (row address strobe), 9-7 MSRCID[0:4] (debug source ID), 4-20, 20-4, 20-8 MSYNC_IN (DRAM DLL synchronization input), 9-8 MSYNC_OUT (DRAM DLL synchronization output), 9-8 MWE (write enable), 9-7 DMA DMA_DACK[0:3] (DMA acknowledge), 15-6 DMA_DDONE[0:3] (DMA done), 15-6 DMA_DREQ[0:3] (DMA request), 15-6 DUART UART_CTS[0:1] (DUART clear to send), 12-1, 12-3, 12-4 UART_RTS[0:1] (DUART request to send), 12-1, 12-3, 12-4 UART_SIN [0:1] (DUART transmitter serial data in), 12-3, 12-4 UART_SOUT [0:1] (DUART transmitter serial data out), 12-3, 12-4 FEC MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Index 2-18 Freescale Semiconductor S–S FEC_COL (FEC collision input), 21-8 FEC_CRS (FEC carrier sense input), 21-8 FEC_RX_CLK (FEC receive clock), 21-8 FEC_RX_DV (FEC receive data valid), 21-8 FEC_RX_ER (FEC receive error), 21-8 FEC_RXD[3:0] (FEC receive data in), 21-8 FEC_TX_CLK (FEC transmit clock in), 21-8 FEC_TX_EN (FEC transmit data valid in), 21-8 FEC_TX_ER (FEC transmit error in), 21-9 FEC_TXD[3:0] (FEC transmit data out), 21-8 global utilities ASLEEP, 18-2, 18-24 CKSTP_IN (checkstop in), 18-3 CKSTP_OUT (checkstop out), 18-3 CLK_OUT, 18-3, 18-18 I2 C SCL (serial clock), 11-3, 11-4 SDA (serial data), 11-3, 11-4 JTAG TCK (JTAG test clock), 20-9 TDI (JTAG test data input), 20-9 TDO (JTAG test data output), 20-9 TMS (JTAG test mode select), 20-9 TRST (JTAG test reset), 20-10 LBC LA[27:31] (non-multiplexed address), 13-7 LAD[0:31] (multiplexed address/data), 13-8 LALE (external address latch enable), 13-6, 13-35 LBCTL (data buffer control), 13-7, 13-37 LBS[0:3] (UPM byte select), 13-6 LCK[0:2] (clock), 13-8 LCKE (clock enable), 13-8 LCS[0:7] (chip select), 13-6 LCS0 (LBC chip select 0), 13-50 LDP[0:3] (data parity), 13-8, 13-38 LGPL0 (GP line 0), 13-6 LGPL1 (GP line 1), 13-6 LGPL2 (GP line 2), 13-6 LGPL3 (GP line 3), 13-7 LGPL4 (GP line 4), 13-7 LGPL5 (GP line 5), 13-7 LGTA (GPCM transfer acknowledge), 13-7, 13-49 LOE (GPCM output enable), 13-6 LPBSE (parity byte select), 13-7 LSDA10 (SDRAM A10), 13-6 LSDCAS (SDRAM CAS), 13-7 LSDDQM[0:3] (SDRAM data mask), 13-6 LSDRAS (SDRAM RAS), 13-6 LSDWE (SDRAM write enable), 13-6 LSYNC_IN (DLL synchronization in), 13-8 LSYNC_OUT (DLL synchronization out), 13-8 LWE[0:3] (GPCM write enable), 13-6 MDVAL (debug mode data valid), 4-20, 13-9, 20-4, 20-7 MSRCID[0:4] (debug source ID), 4-20, 13-9, 20-4, 20-8 TA (data transfer acknowledge), 13-36 UPWAIT (UPM wait), 13-7, 13-63 other TEST_SEL (factory test), 20-10 THERM[0:1] (thermal resistor access), 20-10 PCI/PCI-X PCI_AD[62:58] (high-order PCI address) as debug signals, 20-4, 20-8 PCI_AD[63:0] (address/data bus), 16-8, 16-59 PCI_C/BE[7:0] (command/byte enable), 16-9, 16-57, 16-60, 16-61, 16-78 PCI_DEVSEL (device select), 16-60 PCI_FRAME (frame), 16-57 PCI_GNT[4:0] (bus grant), 16-10, 16-54 PCI_IDSEL (initialization device), 16-10 PCI_IRDY (intitiator ready), 16-11, 16-57 PCI_PAR (parity), 16-11, 16-78 PCI_PAR64 (upper DWORD parity), 16-12 PCI_PERR (parity error), 16-12, 16-79, 16-96 PCI_REQ[4:0] (bus request), 16-13, 16-54 PCI_REQ64[4:0] (64-bit transaction request), 16-13 PCI_SERR (system error), 16-13, 16-79, 16-96 PCI_STOP (stop), 16-14, 16-61 PCI_TRDY (target ready), 16-14, 16-57 PIC IRQ[0:11], 10-8 IRQ_OUT, 10-9, 10-26 MCP, 10-9 UDE, 10-9 RapidIO RIO_RCLK and RIO_RCLK (receive clock and complement), 17-8 RIO_RD[0:7] and RIO_RD[0:7] (receive data and complements), 17-8 RIO_RFRAME and RIO_RFRAME (receive frame and complement), 17-8 RIO_TCLK and RIO_TCLK (transmit clock out and complement), 17-8 RIO_TD[0:7] and RIO_TD [0:7] (transmit data and complement), 17-8 RIO_TFRAME and RIO_TFRAME (transmit frame and complement), 17-9 RIO_TX_CLK_IN and RIO_TX_CLK_IN (transmit clock in and complement), 17-9 reset HRESET (hard reset), 4-2, 4-9 HRESET_REQ (hard reset request), 4-2, 11-17, 11-18 READY, 4-2, 20-25, 20-26 SRESET (soft reset), 4-2, 4-9 TSEC MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor Index 2-19 T–T EC_GTX_CLK125 (TSEC gigabit transmit 125 MHz source), 14-10 EC_MDC (TSEC management data clock), 14-11 EC_MDIO (TSEC management data input/output), 14-11 TSECn_COL (TSEC 1–2 collision input), 14-10 TSECn_CRS (TSEC 1–2 carrier sense input), 14-10 TSECn_GTX_CLK (TSEC 1–2 gigabit transmit clock), 14-10 TSECn_RX_CLK (TSEC 1–2 receive clock), 14-11 TSECn_RX_DV (TSEC 1–2 receive data valid), 14-11 TSECn_RX_ER (TSEC 1–2 receive error), 14-12 TSECn_RXD[7:0] (TSEC 1–2 receive data in), 14-11 TSECn_TX_CLK (TSEC 1–2 transmit clock in), 14-12 TSECn_TX_EN (TSEC 1–2 transmit data valid in), 14-12 TSECn_TX_ER (TSEC 1–2 transmit error in), 14-12 TSECn_TXD[7:0] (TSEC 1–2 transmit data out), 14-12 watchpoint monitor TRIG_IN (watchpoint trigger in), 20-8, 20-12, 20-18 TRIG_OUT (watchpoint trigger out), 20-9, 20-25 Sleep mode, 1-19, 18-24, 18-28 see also Global utilities, power management Snooping power management and snooping (global utilities), 18-28 Soft reset and reconfiguring for FEC, 21-51 Soft reset and reconfiguring for TSEC, 14-112 SPEFSCR (signal processing and embedded floating-point status and control register), see e500 core, registers SPRGn (software-use registers 0–7), see e500 core, registers SRAM, see L2 cache/SRAM, 7-22 SRESET (soft reset) signal, 4-2, 4-9 SRR0 (save/restore register 0), see e500 core, registers SRR1 (save/restore register 1), see e500 core, registers Stashing, see L2 cache/SRAM, stashing, 7-21 SVR (system version register), see e500 core, registers SYSCLK (system clock input) signal, 4-3 System version (SVR), 18-18 T TA (LBC data transfer acknowledge) signal, 13-36 Target-disconnect, see PCI/PCI-X controller TBL (time base lower register), see e500 core, registers TBU (time base upper register), see e500 core, registers TCK (JTAG test clock) signal, 20-9 TCR (timer control register), see e500 core, registers TDI (JTAG test data input) signal, 20-9 TDO (JTAG test data output) signal, 20-9 Termination PCI/PCI-X, termination of PCI transactions, 16-64 Test interface, see JTAG test access port TEST_SEL (factory test) signal, 20-10 THERM[0:1] (thermal resistor access) signals, 20-10 Three-speed Ethernet controller, see TSEC Timing diagrams PCI/PCI-X transactions TLB0CFG (TLB0 configuration register), see e500 core, registers TLB1CFG (TLB1 configuration register), see e500 core, registers TMS (JTAG test mode select) signal, 20-9 Trace buffer and watchpoint monitor, block diagram, 20-1 as a second watchpoint monitor, 20-29 functional description, 20-29-20-32 initialization, 20-33 modes of triggering and arming, 20-5 overview, 20-2 register descriptions, 20-16-20-23 by acronym, see Register Index see also Watchpoint monitor, 20-5 traced data formats relative to TBCR1[IFSEL] DDR trace buffer entry, 20-31 ECM trace buffer entry, 20-30 PCI trace buffer entry, 20-32 RapidIO trace buffer entry, 20-32 Transactions PCI/PCI-X, see PCI/PCI-X controller, transactions TRIG_IN (watchpoint trigger in) signal, 20-8, 20-12, 20-18 TRIG_OUT (watchpoint trigger out) signal, 20-9, 20-25 TRST (JTAG test reset) signal, 20-10 TSEC block diagram, 14-6 buffer descriptors, 14-125 receive buffer descriptor (RxBD), 14-128 transmit data buffer descriptor (TxBD), 14-126 clocks inputs and outputs, 14-11 management clock out (EC_MDC), 14-55 operation, 4-25 configuration of interfaces, 14-131 GMII interface mode, 14-135 MII interface mode, 14-131 RGMII interface mode, 14-143 RTBI interface mode, 14-147 TBI interface mode, 14-138 data width (POR), 4-16 error handling, 14-123-14-124 features, 14-7 functional description, 14-103 gigabit Ethernet channel operation, 14-111 flow control, 14-119 frame reception, 14-115 frame recognition, 14-116 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Index 2-20 Freescale Semiconductor U–U destination address recognition, 14-116 frame transmission, 14-113 initialization sequence, 14-111 hardware controlled initialization, 14-111 user initialization, 14-111 inter-packet gap time, 14-123 interrupt handling, 14-120-14-123 hash function hash table algorithm, 14-118 hash table effectiveness, 14-118 registers, 14-87 initialization/application information, 14-131-14-151 see also TSEC, configuration interrupts, 14-120 interrupt coalescing, 14-121 by frame count threshold, 14-122 by timer threshold, 14-122 interrupt registers, 14-20-14-25 MAC functionality, 14-46-14-60 configuration, 14-46 CSMA/CD controlling, 14-46 packet collisions, 14-47 packet flow, 14-47 PHY link control, 14-48 registers, 14-49 memory map/register definition, 14-13 detailed memory map, 14-14-14-20 top level module map, 14-13 TSEC2 controller offsets, 14-20 modes of operation, 14-8 10 Mbps and 100 Mbps MII operation, 14-9 1000 Mbps GMII and TBI operation, 14-9 address recognition options, 14-9 full and half-duplex operation, 14-8 internal and external loop back, 14-123 RMON support, 14-116 overview, 1-17, 14-7 performance monitor events TSEC1, 19-22, 19-23 TSEC2, 19-24 physical interface connections, 14-103 gigabit media-independent interface (GMII), 14-104 media-independent interface (MII), 14-104 reduced gigabit media-independent interface (RGMII), 14-105 reduced ten-bit interface (RTBI), 14-107 ten-bit interface (TBI), 14-106 register descriptions, 14-20 by acronym, see Register Index, 14-20 FIFO control and status registers, 14-30-14-33 general control and status registers, 14-20-14-30 hash function registers, 14-87-14-89 MAC registers, 14-46-14-60 MIB registers, 14-60-14-87 receive control and status registers, 14-40-14-46 ten-bit interface (TBI) registers, 14-91-14-103 transmit control and status registers, 14-33-14-40 signals, 14-9-14-12 see also Signals, TSEC soft reset and reconfiguring procedure, 14-112 TBI MII registers, see TSEC, register descriptions TSEC1 protocol (POR), 4-17 TSEC2 protocol (POR), 4-17 TSEC2 signals as GP I/O, see Global utilities, general-purpose I/O signals TSECn_COL (TSEC 1–2 collision input) signals, 14-10 TSECn_CRS (TSEC 1–2 carrier sense input) signals, 14-10 TSECn_GTX_CLK (TSEC 1–2 gigabit transmit clock) signals, 14-10 TSECn_RX_CLK (TSEC 1–2 receive clock) signals, 14-11 TSECn_RX_DV (TSEC 1–2 receive data valid) signals, 14-11 TSECn_RX_ER (TSEC 1–2 receive error) signals, 14-12 TSECn_RXD[7:0] (TSEC 1–2 receive data in) signals, 14-11 TSECn_TX_CLK (TSEC 1–2 transmit clock in) signals, 14-12 TSECn_TX_EN (TSEC 1–2 transmit data valid in) signals, 14-12 TSECn_TX_ER (TSEC 1–2 transmit error in) signals, 14-12 TSECn_TXD[7:0] (TSEC 1–2 transmit data out) signals, 14-12 TSR (timer status register), see e500 core, registers U UART_CTS[0:1] (DUART clear to send) signals, 12-1, 12-3, 12-4 UART_RTS[0:1] (DUART request to send) signals, 12-1, 12-3, 12-4 UART_SIN [0:1] (DUART transmitter serial data in) signals, 12-3, 12-4 UART_SOUT [0:1] (DUART transmitter serial data out) signals, 12-3, 12-4 UDE (unconditional debug event) signal, 10-9 Universal asynchronous receiver/transmitter, see DUART UPMCn (user performance monitor counter registers 0–3), see e500 core, registers UPMGC0 (user performance monitor global control register 0), see e500 core, registers UPMLCan (user performance monitor local control registers a0–a3), see e500 core, registers UPMLCbn (user performance monitor local control registers b0–b3), see e500 core, registers UPWAIT (LBC UPM wait) signal, 13-7, 13-63 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Freescale Semiconductor Index 2-21 W–Z USPRG0 (user software-use register 0), see e500 core, registers W Watchpoint monitor and trace buffer, block diagram, 20-1 functional description, 20-28-20-29 initialization, 20-33 modes of triggering and arming, 20-5 overview, 20-2 performance monitor events, 20-29 register descriptions, 20-11-20-16 by acronym, see Register Index second WM by using trace buffer, 20-29 see also Trace buffer, 20-5 signals summary, 20-6 see also Signals, watchpoint, 20-6 X XER (integer exception register), see e500 core, registers Z ZBT SRAM interface (LBC), 13-101 MPC8540 PowerQUICC III Integrated Host Processor Reference Manual, Rev. 1 Index 2-22 Freescale Semiconductor Part I—Overview Overview Memory Map Signal Descriptions Reset, Clocking, and Initialization Part II—e500 Core Complex and L2 Cache e500 Core Complex Overview e500 Register Summary L2 Look-Aside Cache/SRAM Part III—Memory and I/O Interfaces e500 Coherency Module DDR Memory Controller Programmable Interrupt Controller I2C Interface DUART Local Bus Controller Three-Speed Ethernet Controllers DMA Controller PCI/PCI-X Bus Interface RapidIO Interface Part IV—Global Functions and Debug Global Utilities Performance Monitor Debug Features and Watchpoint Facility 10/100 Fast Ethernet Controller Appendix A—Revision History Glossary of Terms and Abbreviations Register Index (Memory-Mapped Registers) General Index I 1 2 3 4 II 5 6 7 III 8 9 10 11 12 13 14 15 16 17 IV 18 19 20 21 A GLO REG IND I 1 2 3 4 II 5 6 7 III 8 9 10 11 12 13 14 15 16 17 IV 18 19 20 21 A GLO REG IND Part I—Overview Overview Memory Map Signal Descriptions Reset, Clocking, and Initialization Part II—e500 Core Complex and L2 Cache e500 Core Complex Overview e500 Register Summary L2 Look-Aside Cache/SRAM Part III—Memory and I/O Interfaces e500 Coherency Module DDR Memory Controller Programmable Interrupt Controller I2C Interface DUART Local Bus Controller Three-Speed Ethernet Controllers DMA Controller PCI/PCI-X Bus Interface RapidIO Interface Part IV—Global Functions and Debug Global Utilities Performance Monitor Debug Features and Watchpoint Facility 10/100 Fast Ethernet Controller Appendix A—Revision History Glossary of Terms and Abbreviations Register Index (Memory-Mapped Registers) General Index