MCF548x Reference Manual Devices Supported: MCF5485 MCF5484 MCF5483 MCF5482 MCF5481 MCF5480 Document Number: MCF5485RM Rev. 5 4/2009 How to Reach Us: Home Page: www.freescale.com Web Support: http://www.freescale.com/support USA/Europe or Locations Not Listed: Freescale Semiconductor, Inc. Technical Information Center, EL516 2100 East Elliot Road Tempe, Arizona 85284 1-800-521-6274 or +1-480-768-2130 www.freescale.com/support Europe, Middle East, and Africa: Freescale Halbleiter Deutschland GmbH Technical Information Center Schatzbogen 7 81829 Muenchen, Germany +44 1296 380 456 (English) +46 8 52200080 (English) +49 89 92103 559 (German) +33 1 69 35 48 48 (French) www.freescale.com/support Japan: Freescale Semiconductor Japan Ltd. Headquarters ARCO Tower 15F 1-8-1, Shimo-Meguro, Meguro-ku, Tokyo 153-0064 Japan 0120 191014 or +81 3 5437 9125 [email protected] Asia/Pacific: Freescale Semiconductor China Ltd. Exchange Building 23F No. 118 Jianguo Road Chaoyang District Beijing 100022 China +86 10 5879 8000 [email protected] Freescale Semiconductor Literature Distribution Center P.O. Box 5405 Denver, Colorado 80217 1-800-441-2447 or +1-303-675-2140 Fax: +1-303-675-2150 [email protected] 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 that 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. Freescale™ and the Freescale logo are trademarks of Freescale Semiconductor, Inc. All other product or service names are the property of their respective owners. © Freescale Semiconductor, Inc. 2008. All rights reserved. MCF5485RM Rev. 5 4/2009 Overview Signal Descriptions ColdFire Core Enhanced Multiply-Accumulate Unit (EMAC) Memory Management Unit (MMU) Floating-Point Unit (FPU) Local Memory Debug Support System Integration Unit (SIU) Internal Clocks and Bus Architecture General Purpose Timers (GPT) Slice Timers (SLT) Interrupt Controller (INTC) Edge Port Module (EPORT) General Purpose I/O (GPIO) System SRAM FlexBus SDRAM Controller (SDRAMC) PCI Bus Controller (PCI) PCI Bus Arbiter (PCIARB) FlexCAN Integrated Secuity Engine (SEC) IEEE 1149.1 Test Access Port (JTAG) Multichannel DMA (MCD) Comm Bus FIFO Interface Comm Timer Module (CTM) Programmable Serial Controller (PSC) DMA Serial Peripheral Interface (DSPI) I2C Interface USB 2.0 Device Controller Fast Ethernet Controller (FEC) Mechanical Data Register Memory Map Quick Reference Index 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 30 31 32 A IND 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 30 31 32 A IND Overview Signal Descriptions ColdFire Core Enhanced Multiply-Accumulate Unit (EMAC) Memory Management Unit (MMU) Floating-Point Unit (FPU) Local Memory Debug Support System Integration Unit (SIU) Internal Clocks and Bus Architecture General Purpose Timers (GPT) Slice Timers (SLT) Interrupt Controller (INTC) Edge Port Module (EPORT) General Purpose I/O (GPIO) System SRAM FlexBus SDRAM Controller (SDRAMC) PCI Bus Controller (PCI) PCI Bus Arbiter (PCIARB) FlexCAN Integrated Secuity Engine (SEC) IEEE 1149.1 Test Access Port (JTAG) Multichannel DMA (MCD) Comm Bus FIFO Interface Comm Timer Module (CTM) Programmable Serial Controller (PSC) DMA Serial Peripheral Interface (DSPI) I2C Interface USB 2.0 Device Controller Fast Ethernet Controller (FEC) Mechanical Data Register Memory Map Quick Reference Index Chapter 1 Overview 1.1 1.2 1.3 1.4 MCF548x Family Overview ......................................................................................................... 1-1 MCF548x Block Diagram ............................................................................................................. 1-2 MCF548x Family Products ........................................................................................................... 1-3 MCF548x Family Features ............................................................................................................ 1-3 1.4.1 ColdFire V4e Core Overview ....................................................................................... 1-5 1.4.2 Debug Module (BDM) ................................................................................................. 1-6 1.4.3 JTAG ............................................................................................................................. 1-6 1.4.4 On-Chip Memories ....................................................................................................... 1-7 1.4.5 PLL and Chip Clocking Options .................................................................................. 1-7 1.4.6 Communications I/O Subsystem .................................................................................. 1-8 1.4.7 DDR SDRAM Memory Controller ............................................................................ 1-10 1.4.8 Peripheral Component Interconnect (PCI) ................................................................. 1-10 1.4.9 Flexible Local Bus (FlexBus) ..................................................................................... 1-10 1.4.10 Security Encryption Controller (SEC) ........................................................................ 1-11 1.4.11 System Integration Unit (SIU) .................................................................................... 1-11 Chapter 2 Signal Descriptions 2.1 2.2 Introduction ................................................................................................................................... 2-1 2.1.1 Block Diagram .............................................................................................................. 2-1 MCF548x External Signals ......................................................................................................... 2-16 2.2.1 FlexBus Signals .......................................................................................................... 2-16 2.2.2 SDRAM Controller Signals ........................................................................................ 2-18 2.2.3 PCI Controller Signals ................................................................................................ 2-19 2.2.4 Interrupt Control Signals ............................................................................................ 2-21 2.2.5 Clock and Reset Signals ............................................................................................. 2-21 2.2.6 Reset Configuration Pins ............................................................................................ 2-22 2.2.7 Ethernet Module Signals ............................................................................................ 2-24 2.2.8 Universal Serial Bus (USB) ........................................................................................ 2-26 2.2.9 DMA Serial Peripheral Interface (DSPI) Signals ....................................................... 2-26 2.2.10 FlexCAN Signals ........................................................................................................ 2-27 2.2.11 I2C I/O Signals ........................................................................................................... 2-27 2.2.12 PSC Module Signals ................................................................................................... 2-28 2.2.13 DMA Controller Module Signals ............................................................................... 2-28 2.2.14 Timer Module Signals ................................................................................................ 2-28 2.2.15 Debug Support Signals ............................................................................................... 2-29 2.2.16 Test Signals ................................................................................................................. 2-30 2.2.17 Power and Reference Pins .......................................................................................... 2-30 MCF548x Reference Manual, Rev. 5 Freescale Semiconductor v Chapter 3 ColdFire Core 3.1 3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9 Core Overview .............................................................................................................................. 3-1 Features ......................................................................................................................................... 3-1 3.2.1 Enhanced Pipelines ....................................................................................................... 3-2 3.2.2 Debug Module Enhancements ...................................................................................... 3-6 Programming Model ..................................................................................................................... 3-7 3.3.1 User Programming Model ............................................................................................ 3-9 3.3.2 User Stack Pointer (A7) ............................................................................................... 3-9 3.3.3 EMAC Programming Model ...................................................................................... 3-10 3.3.4 FPU Programming Model .......................................................................................... 3-10 3.3.5 Supervisor Programming Model ................................................................................ 3-11 3.3.6 Programming Model Table ......................................................................................... 3-13 Data Format Summary ................................................................................................................ 3-15 3.4.1 Data Organization in Registers ................................................................................... 3-15 3.4.2 EMAC Data Representation ....................................................................................... 3-17 Addressing Mode Summary ........................................................................................................ 3-18 Instruction Set Summary ............................................................................................................. 3-19 3.6.1 Additions to the Instruction Set Architecture ............................................................. 3-19 3.6.2 Instruction Set Summary ............................................................................................ 3-22 Instruction Execution Timing ...................................................................................................... 3-27 3.7.1 MOVE Instruction Execution Timing ........................................................................ 3-28 3.7.2 One-Operand Instruction Execution Timing .............................................................. 3-30 3.7.3 Two-Operand Instruction Execution Timing .............................................................. 3-31 3.7.4 Miscellaneous Instruction Execution Timing ............................................................. 3-32 3.7.5 Branch Instruction Execution Timing ........................................................................ 3-33 3.7.6 EMAC Instruction Execution Times .......................................................................... 3-34 3.7.7 FPU Instruction Execution Times .............................................................................. 3-35 Exception Processing Overview .................................................................................................. 3-36 3.8.1 Exception Stack Frame Definition ............................................................................. 3-38 3.8.2 Processor Exceptions .................................................................................................. 3-39 Precise Faults ............................................................................................................................... 3-42 Chapter 4 Enhanced Multiply-Accumulate Unit (EMAC) 4.1 4.2 4.3 Introduction ................................................................................................................................... 4-1 4.1.1 MAC Overview ............................................................................................................ 4-2 4.1.2 General Operation ........................................................................................................ 4-2 Memory Map/Register Definition ................................................................................................. 4-5 4.2.1 MAC Status Register (MACSR) .................................................................................. 4-5 4.2.2 Mask Register (MASK) .............................................................................................. 4-10 EMAC Instruction Set Summary ................................................................................................ 4-11 4.3.1 EMAC Instruction Execution Timing ........................................................................ 4-11 4.3.2 Data Representation .................................................................................................... 4-12 4.3.3 EMAC Opcodes .......................................................................................................... 4-13 MCF548x Reference Manual, Rev. 5 vi Freescale Semiconductor Chapter 5 Memory Management Unit (MMU) 5.1 5.2 5.3 5.4 5.5 5.6 5.7 Features ......................................................................................................................................... 5-1 Virtual Memory Management Architecture .................................................................................. 5-1 5.2.1 MMU Architecture Features ......................................................................................... 5-1 5.2.2 MMU Architecture Location ........................................................................................ 5-2 5.2.3 MMU Architecture Implementation ............................................................................. 5-3 Debugging in a Virtual Environment ............................................................................................ 5-7 Virtual Memory Architecture Processor Support .......................................................................... 5-7 5.4.1 Precise Faults ................................................................................................................ 5-7 5.4.2 Supervisor/User Stack Pointers ................................................................................... 5-7 5.4.3 Access Error Stack Frame Additions ........................................................................... 5-8 MMU Definition ........................................................................................................................... 5-9 5.5.1 Effective Address Attribute Determination .................................................................. 5-9 5.5.2 MMU Functionality .................................................................................................... 5-10 5.5.3 MMU Organization .................................................................................................... 5-10 5.5.4 MMU TLB .................................................................................................................. 5-18 5.5.5 MMU Operation ......................................................................................................... 5-19 MMU Implementation ................................................................................................................. 5-20 5.6.1 TLB Address Fields .................................................................................................... 5-20 5.6.2 TLB Replacement Algorithm ..................................................................................... 5-21 5.6.3 TLB Locked Entries ................................................................................................... 5-22 MMU Instructions ....................................................................................................................... 5-23 Chapter 6 Floating-Point Unit (FPU) 6.1 6.2 6.3 6.4 6.5 6.6 Introduction ................................................................................................................................... 6-1 6.1.1 Overview ...................................................................................................................... 6-1 Operand Data Formats and Types ................................................................................................. 6-3 6.2.1 Signed-Integer Data Formats ........................................................................................ 6-3 6.2.2 Floating-Point Data Formats ........................................................................................ 6-3 6.2.3 Floating-Point Data Types ............................................................................................ 6-4 Register Definition ........................................................................................................................ 6-7 6.3.1 Floating-Point Data Registers (FP0–FP7) .................................................................... 6-7 6.3.2 Floating-Point Control Register (FPCR) ...................................................................... 6-7 6.3.3 Floating-Point Status Register (FPSR) ......................................................................... 6-9 6.3.4 Floating-Point Instruction Address Register (FPIAR) ............................................... 6-10 Floating-Point Computational Accuracy ..................................................................................... 6-11 6.4.1 Intermediate Result ..................................................................................................... 6-11 6.4.2 Rounding the Result ................................................................................................... 6-12 Floating-Point Post-Processing ................................................................................................... 6-14 6.5.1 Underflow, Round, and Overflow .............................................................................. 6-14 6.5.2 Conditional Testing .................................................................................................... 6-15 Floating-Point Exceptions ........................................................................................................... 6-17 6.6.1 Floating-Point Arithmetic Exceptions ........................................................................ 6-18 MCF548x Reference Manual, Rev. 5 Freescale Semiconductor vii 6.7 6.6.2 Floating-Point State Frames ....................................................................................... 6-23 Instructions .................................................................................................................................. 6-25 6.7.1 Floating-Point Instruction Overview .......................................................................... 6-25 6.7.2 Floating-Point Instruction Execution Timing ............................................................. 6-27 6.7.3 Key Differences between ColdFire and M68000 FPU Programming Models ........... 6-28 Chapter 7 Local Memory 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 Interactions between Local Memory Modules .............................................................................. 7-1 SRAM Overview ........................................................................................................................... 7-1 SRAM Operation ........................................................................................................................... 7-2 SRAM Register Definition ............................................................................................................ 7-2 7.4.1 SRAM Base Address Registers (RAMBAR0/RAMBAR1) ......................................... 7-2 SRAM Initialization ...................................................................................................................... 7-4 7.5.1 SRAM Initialization Code ............................................................................................ 7-5 Power Management ....................................................................................................................... 7-6 Cache Overview ............................................................................................................................ 7-6 Cache Organization ....................................................................................................................... 7-7 7.8.1 Cache Line States: Invalid, Valid-Unmodified, and Valid-Modified ........................... 7-8 7.8.2 The Cache at Start-Up .................................................................................................. 7-8 Cache Operation .......................................................................................................................... 7-10 7.9.1 Caching Modes ........................................................................................................... 7-12 7.9.2 Cache Protocol ............................................................................................................ 7-14 7.9.3 Cache Coherency (Data Cache Only) ......................................................................... 7-15 7.9.4 Memory Accesses for Cache Maintenance ................................................................ 7-15 7.9.5 Cache Locking ............................................................................................................ 7-17 Cache Register Definition ........................................................................................................... 7-19 7.10.1 Cache Control Register (CACR) ................................................................................ 7-19 7.10.2 Access Control Registers (ACR0–ACR3) .................................................................. 7-22 Cache Management ..................................................................................................................... 7-23 Cache Operation Summary ......................................................................................................... 7-26 7.12.1 Instruction Cache State Transitions ............................................................................ 7-26 7.12.2 Data Cache State Transitions ...................................................................................... 7-27 Cache Initialization Code ............................................................................................................ 7-30 Chapter 8 Debug Support 8.1 8.2 8.3 Introduction ................................................................................................................................... 8-1 8.1.1 Overview ...................................................................................................................... 8-1 Signal Descriptions ....................................................................................................................... 8-2 8.2.1 Processor Status/Debug Data (PSTDDATA[7:0]) ........................................................ 8-3 Real-Time Trace Support .............................................................................................................. 8-5 8.3.1 Begin Execution of Taken Branch (PST = 0x5) ........................................................... 8-6 8.3.2 Processor Stopped or Breakpoint State Change (PST = 0xE) ...................................... 8-7 8.3.3 Processor Halted (PST = 0xF) ...................................................................................... 8-8 MCF548x Reference Manual, Rev. 5 viii Freescale Semiconductor 8.4 8.5 8.6 8.7 8.8 8.9 Memory Map/Register Definition ................................................................................................. 8-9 8.4.1 Revision A Shared Debug Resources ......................................................................... 8-11 8.4.2 Configuration/Status Register (CSR) ......................................................................... 8-11 8.4.3 PC Breakpoint ASID Control Register (PBAC) ........................................................ 8-14 8.4.4 BDM Address Attribute Register (BAAR) ................................................................ 8-15 8.4.5 Address Attribute Trigger Registers (AATR, AATR1) .............................................. 8-16 8.4.6 Trigger Definition Register (TDR) ............................................................................. 8-17 8.4.7 Program Counter Breakpoint and Mask Registers (PBRn, PBMR) ........................... 8-20 8.4.8 Address Breakpoint Registers (ABLR/ABLR1, ABHR/ABHR1) ............................. 8-21 8.4.9 Data Breakpoint and Mask Registers (DBR/DBR1, DBMR/DBMR1) ..................... 8-22 8.4.10 PC Breakpoint ASID Register (PBASID) .................................................................. 8-24 8.4.11 Extended Trigger Definition Register (XTDR) .......................................................... 8-25 Background Debug Mode (BDM) ............................................................................................... 8-28 8.5.1 CPU Halt .................................................................................................................... 8-28 8.5.2 BDM Serial Interface ................................................................................................. 8-30 8.5.3 BDM Command Set ................................................................................................... 8-31 Real-Time Debug Support ........................................................................................................... 8-51 8.6.1 Theory of Operation ................................................................................................... 8-51 8.6.2 Concurrent BDM and Processor Operation ................................................................ 8-54 Debug C Definition of PSTDDATA Outputs ............................................................................. 8-54 8.7.1 User Instruction Set .................................................................................................... 8-54 8.7.2 Supervisor Instruction Set .......................................................................................... 8-60 ColdFire Debug History .............................................................................................................. 8-61 8.8.1 ColdFire Debug Classic: The Original Definition ...................................................... 8-61 8.8.2 ColdFire Debug Revision B ....................................................................................... 8-62 8.8.3 ColdFire Debug Revision C ....................................................................................... 8-62 Freescale-Recommended BDM Pinout ....................................................................................... 8-63 Chapter 9 System Integration Unit (SIU) 9.1 9.2 9.3 Introduction ................................................................................................................................... 9-1 Features ......................................................................................................................................... 9-1 Memory Map/Register Definition ................................................................................................. 9-1 9.3.1 Module Base Address Register (MBAR) ..................................................................... 9-2 Chapter 10 Internal Clocks and Bus Architecture 10.1 Introduction ................................................................................................................................. 10-1 10.1.1 Block Diagram ............................................................................................................ 10-1 10.1.2 Clocking Overview ..................................................................................................... 10-2 10.1.3 Internal Bus Overview ................................................................................................ 10-2 10.1.4 XL Bus Features ......................................................................................................... 10-3 10.1.5 Internal Bus Transaction Summaries .......................................................................... 10-3 10.1.6 XL Bus Interface Operations ...................................................................................... 10-3 10.2 PLL .............................................................................................................................................. 10-5 MCF548x Reference Manual, Rev. 5 Freescale Semiconductor ix 10.2.1 PLL Memory Map/Register Descriptions .................................................................. 10-5 10.2.2 System PLL Control Register (SPCR) ....................................................................... 10-5 10.3 XL Bus Arbiter ............................................................................................................................ 10-6 10.3.1 Features ....................................................................................................................... 10-6 10.3.2 Arbiter Functional Description ................................................................................... 10-6 10.3.3 XLB Arbiter Register Descriptions ............................................................................ 10-8 Chapter 11 General Purpose Timers (GPT) 11.1 Introduction ................................................................................................................................. 11-1 11.1.1 Overview .................................................................................................................... 11-1 11.1.2 Modes of Operation .................................................................................................... 11-1 11.2 External Signals ........................................................................................................................... 11-2 11.3 Memory Map/Register Definition ............................................................................................... 11-2 11.3.1 GPT Enable and Mode Select Register (GMSn) ........................................................ 11-3 11.3.2 GPT Counter Input Register (GCIRn) ........................................................................ 11-5 11.3.3 GPT PWM Configuration Register (GPWMn) .......................................................... 11-6 11.3.4 GPT Status Register (GSRn) ...................................................................................... 11-7 11.4 Functional Description ................................................................................................................ 11-8 11.4.1 Timer Configuration Method ...................................................................................... 11-8 11.4.2 Programming Notes .................................................................................................... 11-8 Chapter 12 Slice Timers (SLT) 12.1 Introduction ................................................................................................................................. 12-1 12.1.1 Overview .................................................................................................................... 12-1 12.2 Memory Map/Register Definition ............................................................................................... 12-1 12.2.1 SLT Terminal Count Register (STCNTn) ................................................................... 12-2 12.2.2 SLT Control Register (SCRn) ..................................................................................... 12-2 12.2.3 SLT Timer Count Register (SCNTn) .......................................................................... 12-3 12.2.4 SLT Status Register (SSRn) ........................................................................................ 12-4 Chapter 13 Interrupt Controller 13.1 Introduction ................................................................................................................................. 13-1 13.1.1 68K/ColdFire Interrupt Architecture Overview ......................................................... 13-1 13.2 Memory Map/Register Descriptions ........................................................................................... 13-4 13.2.1 Register Descriptions .................................................................................................. 13-6 Chapter 14 Edge Port Module (EPORT) 14.1 Introduction ................................................................................................................................. 14-1 14.2 Interrupt/General-Purpose I/O Pin Descriptions ......................................................................... 14-1 14.3 Memory Map/Register Definition ............................................................................................... 14-2 14.3.1 Memory Map .............................................................................................................. 14-2 MCF548x Reference Manual, Rev. 5 x Freescale Semiconductor 14.3.2 Register Descriptions .................................................................................................. 14-2 Chapter 15 GPIO 15.1 Introduction ................................................................................................................................. 15-1 15.1.1 Overview .................................................................................................................... 15-2 15.1.2 Features ....................................................................................................................... 15-3 15.2 External Pin Description ............................................................................................................. 15-3 15.3 Memory Map/Register Definition ............................................................................................... 15-7 15.3.1 Register Overview ...................................................................................................... 15-7 15.3.2 Register Descriptions .................................................................................................. 15-8 15.4 Functional Description .............................................................................................................. 15-32 15.4.1 Overview .................................................................................................................. 15-32 Chapter 16 32-Kbyte System SRAM 16.1 Introduction ................................................................................................................................. 16-1 16.1.1 Block Diagram ............................................................................................................ 16-1 16.1.2 Features ....................................................................................................................... 16-2 16.1.3 Overview .................................................................................................................... 16-2 16.2 Memory Map/Register Definition ............................................................................................... 16-2 16.2.1 System SRAM Configuration Register (SSCR) ......................................................... 16-3 16.2.2 Transfer Count Configuration Register (TCCR) ....................................................... 16-4 16.2.3 Transfer Count Configuration Register—DMA Read Channel (TCCRDR) .............. 16-5 16.2.4 Transfer Count Configuration Register—DMA Write Channel (TCCRDW) ............ 16-6 16.2.5 Transfer Count Configuration Register—SEC (TCCRSEC) ..................................... 16-7 16.3 Functional Description ................................................................................................................ 16-8 Chapter 17 FlexBus 17.1 Introduction ................................................................................................................................. 17-1 17.1.1 Overview .................................................................................................................... 17-1 17.1.2 Features ....................................................................................................................... 17-1 17.1.3 Modes of Operation .................................................................................................... 17-1 17.2 Byte Lanes ................................................................................................................................... 17-2 17.3 Address Latch .............................................................................................................................. 17-2 17.4 External Signals ........................................................................................................................... 17-3 17.4.1 Chip-Select (FBCS[5:0]) ............................................................................................ 17-4 17.4.2 Address/Data Bus (AD[31:0]) .................................................................................... 17-4 17.4.3 Address Latch Enable (ALE) ..................................................................................... 17-4 17.4.4 Read/Write (R/W) ....................................................................................................... 17-4 17.4.5 Transfer Burst (TBST) ................................................................................................ 17-4 17.4.6 Transfer Size (TSIZ[1:0]) ........................................................................................... 17-4 17.4.7 Byte Selects (BE/BWE[3:0]) ...................................................................................... 17-5 17.4.8 Output Enable (OE) .................................................................................................... 17-5 MCF548x Reference Manual, Rev. 5 Freescale Semiconductor xi 17.4.9 Transfer Acknowledge (TA) ....................................................................................... 17-5 17.5 Chip-Select Operation ................................................................................................................. 17-6 17.5.1 General Chip-Select Operation ................................................................................... 17-6 17.5.2 Chip-Select Registers ................................................................................................. 17-7 17.6 Functional Description .............................................................................................................. 17-12 17.6.1 Data Transfer Operation ........................................................................................... 17-12 17.6.2 Data Byte Alignment and Physical Connections ...................................................... 17-12 17.6.3 Address/Data Bus Multiplexing ............................................................................... 17-13 17.6.4 Bus Cycle Execution ................................................................................................ 17-13 17.6.5 FlexBus Timing Examples ....................................................................................... 17-15 17.6.6 Burst Cycles .............................................................................................................. 17-26 17.6.7 Misaligned Operands ................................................................................................ 17-31 17.6.8 Bus Errors ................................................................................................................. 17-32 Chapter 18 SDRAM Controller (SDRAMC) 18.1 Introduction ................................................................................................................................. 18-1 18.2 Overview ..................................................................................................................................... 18-1 18.2.1 Features ....................................................................................................................... 18-1 18.2.2 Terminology ............................................................................................................... 18-1 18.2.3 Block Diagram ............................................................................................................ 18-2 18.3 External Signal Description ........................................................................................................ 18-2 18.3.1 SDRAM Data Bus (SDDATA[31:0]) ......................................................................... 18-2 18.3.2 SDRAM Address Bus (SDADDR[12:0]) ................................................................... 18-2 18.3.3 SDRAM Bank Addresses (SDBA[1:0]) ..................................................................... 18-2 18.3.4 SDRAM Row Address Strobe (RAS) ........................................................................ 18-3 18.3.5 SDRAM Column Address Strobe (CAS) ................................................................... 18-3 18.3.6 SDRAM Chip Selects (SDCS[3:0]) ........................................................................... 18-3 18.3.7 SDRAM Write Data Byte Mask (SDDM[3:0]) .......................................................... 18-3 18.3.8 SDRAM Data Strobe (SDDQS[3:0]) ......................................................................... 18-3 18.3.9 SDRAM Clock (SDCLK[1:0]) ................................................................................... 18-3 18.3.10 Inverted SDRAM Clock (SDCLK[1:0]) .................................................................... 18-3 18.3.11 SDRAM Write Enable (SDWE) ................................................................................. 18-3 18.3.12 SDRAM Clock Enable (SDCKE) .............................................................................. 18-4 18.3.13 SDR SDRAM Data Strobe (SDRDQS) ...................................................................... 18-4 18.3.14 SDRAM Memory Supply (SDVDD) ......................................................................... 18-4 18.3.15 SDRAM Reference Voltage (VREF) .......................................................................... 18-4 18.4 Interface Recommendations ........................................................................................................ 18-4 18.4.1 Supported Memory Configurations ............................................................................ 18-4 18.4.2 SDRAM SDR Connections ........................................................................................ 18-6 18.4.3 SDRAM DDR Component Connections .................................................................... 18-6 18.4.4 SDRAM DDR DIMM Connections ........................................................................... 18-7 18.4.5 DDR SDRAM Layout Considerations ....................................................................... 18-8 18.5 SDRAM Overview ...................................................................................................................... 18-9 18.5.1 SDRAM Commands ................................................................................................... 18-9 MCF548x Reference Manual, Rev. 5 xii Freescale Semiconductor 18.5.2 Power-Up Initialization ............................................................................................ 18-13 18.6 Functional Overview ................................................................................................................. 18-15 18.6.1 Page Management .................................................................................................... 18-15 18.6.2 Transfer Size ............................................................................................................. 18-15 18.7 Memory Map/Register Definition ............................................................................................. 18-16 18.7.1 SDRAM Drive Strength Register (SDRAMDS) ...................................................... 18-17 18.7.2 SDRAM Chip Select Configuration Registers (CSnCFG) ....................................... 18-18 18.7.3 SDRAM Mode/Extended Mode Register (SDMR) .................................................. 18-19 18.7.4 SDRAM Control Register (SDCR) .......................................................................... 18-20 18.7.5 SDRAM Configuration Register 1 (SDCFG1) ......................................................... 18-21 18.7.6 SDRAM Configuration Register 2 (SDCFG2) ......................................................... 18-23 18.8 SDRAM Example ..................................................................................................................... 18-24 18.8.1 SDRAM Signal Drive Strength Settings .................................................................. 18-25 18.8.2 SDRAM Chip Select Settings .................................................................................. 18-25 18.8.3 SDRAM Configuration 1 Register Settings ............................................................. 18-26 18.8.4 SDRAM Configuration 2 Register Settings ............................................................. 18-27 18.8.5 SDRAM Control Register Settings and PALL command ........................................ 18-27 18.8.6 Set the Extended Mode Register .............................................................................. 18-29 18.8.7 Set the Mode Register and Reset DLL ..................................................................... 18-29 18.8.8 Issue a PALL command ............................................................................................ 18-30 18.8.9 Perform Two Refresh Cycles .................................................................................... 18-31 18.8.10 Clear the Reset DLL Bit in the Mode Register ....................................................... 18-32 18.8.11 Enable Automatic Refresh and Lock Mode Register .............................................. 18-33 18.8.12 Initialization Code .................................................................................................... 18-34 Chapter 19 PCI Bus Controller 19.1 Introduction ................................................................................................................................. 19-1 19.1.1 Block Diagram ............................................................................................................ 19-1 19.1.2 Overview .................................................................................................................... 19-1 19.1.3 Features ....................................................................................................................... 19-1 19.2 External Signal Description ........................................................................................................ 19-2 19.2.1 Address/Data Bus (PCIAD[31:0]) .............................................................................. 19-2 19.2.2 Command/Byte Enables (PCICXBE[3:0]) ................................................................. 19-2 19.2.3 Device Select (PCIDEVSEL) ..................................................................................... 19-3 19.2.4 Frame (PCIFRAME) .................................................................................................. 19-3 19.2.5 Initialization Device Select (PCIIDSEL) ................................................................... 19-3 19.2.6 Initiator Ready (PCIIRDY) ........................................................................................ 19-3 19.2.7 Parity (PCIPAR) ......................................................................................................... 19-3 19.2.8 PCI Clock (CLKIN) ................................................................................................... 19-3 19.2.9 Parity Error (PCIPERR) ............................................................................................. 19-3 19.2.10 Reset (PCIRESET) .................................................................................................... 19-3 19.2.11 System Error (PCISERR) .......................................................................................... 19-3 19.2.12 Stop (PCISTOP) ........................................................................................................ 19-3 19.2.13 Target Ready (PCITRDY) ......................................................................................... 19-4 MCF548x Reference Manual, Rev. 5 Freescale Semiconductor xiii 19.3 Memory Map/Register Definition ............................................................................................... 19-4 19.3.1 PCI Type 0 Configuration Registers ........................................................................... 19-6 19.3.2 General Control/Status Registers ............................................................................. 19-13 19.3.3 Communication Subsystem Interface Registers ....................................................... 19-23 19.4 Functional Description .............................................................................................................. 19-48 19.4.1 PCI Bus Protocol ...................................................................................................... 19-48 19.4.2 Initiator Arbitration .................................................................................................. 19-55 19.4.3 Configuration Interface ............................................................................................ 19-56 19.4.4 XL Bus Initiator Interface ........................................................................................ 19-56 19.4.5 XL Bus Target Interface .......................................................................................... 19-63 19.4.6 Communication Subsystem Initiator Interface ......................................................... 19-66 19.4.7 PCI Clock Scheme .................................................................................................... 19-70 19.4.8 Interrupts ................................................................................................................... 19-70 19.5 Application Information ............................................................................................................ 19-70 19.5.1 XL Bus-Initiated Transaction Mapping .................................................................... 19-70 19.5.2 Address Maps ........................................................................................................... 19-71 19.6 XL Bus Arbitration Priority ...................................................................................................... 19-75 Chapter 20 PCI Bus Arbiter Module 20.1 Introduction ................................................................................................................................. 20-1 20.1.1 Block Diagram ............................................................................................................ 20-1 20.1.2 Overview .................................................................................................................... 20-1 20.1.3 Features ....................................................................................................................... 20-2 20.2 External Signal Description ........................................................................................................ 20-2 20.2.1 Frame (PCIFRM) ........................................................................................................ 20-2 20.2.2 Initiator Ready (PCIIRDY) ........................................................................................ 20-2 20.2.3 PCI Clock (CLKIN) ................................................................................................... 20-2 20.2.4 External Bus Grant (PCIBG[4:1]) .............................................................................. 20-2 20.2.5 External Bus Grant/Request Output (PCIBG0/PCIREQOUT) .................................. 20-3 20.2.6 External Bus Request (PCIBR[4:1]) .......................................................................... 20-3 20.2.7 External Request/Grant Input (PCIBR0/PCIGNTIN) ................................................ 20-3 20.3 Register Definition ...................................................................................................................... 20-3 20.3.1 PCI Arbiter Control Register (PACR) ........................................................................ 20-3 20.3.2 PCI Arbiter Status Register (PASR) ........................................................................... 20-5 20.4 Functional Description ................................................................................................................ 20-5 20.4.1 External PCI Requests ................................................................................................ 20-5 20.4.2 Arbitration .................................................................................................................. 20-6 20.4.3 Master Time-Out ........................................................................................................ 20-9 20.5 Reset .......................................................................................................................................... 20-10 20.6 Interrupts ................................................................................................................................... 20-10 Chapter 21 FlexCAN 21.1 Introduction ................................................................................................................................. 21-1 MCF548x Reference Manual, Rev. 5 xiv Freescale Semiconductor 21.2 21.3 21.4 21.5 21.1.1 Block Diagram ............................................................................................................ 21-1 21.1.2 The CAN System ........................................................................................................ 21-2 21.1.3 Features ....................................................................................................................... 21-3 21.1.4 Modes of Operation .................................................................................................... 21-3 External Signals ........................................................................................................................... 21-5 21.2.1 CANTX[1:0] ............................................................................................................... 21-5 21.2.2 CANRX[1:0] .............................................................................................................. 21-5 Memory Map/Register Definition ............................................................................................... 21-5 21.3.1 FlexCAN Memory Map ............................................................................................. 21-5 21.3.2 Register Descriptions .................................................................................................. 21-6 Functional Overview ................................................................................................................. 21-19 21.4.1 Message Buffer Structure ......................................................................................... 21-19 21.4.2 Message Buffer Memory Map .................................................................................. 21-22 21.4.3 Transmit Process ....................................................................................................... 21-23 21.4.4 Arbitration Process ................................................................................................... 21-24 21.4.5 Receive Process ........................................................................................................ 21-24 21.4.6 Message Buffer Handling ......................................................................................... 21-25 21.4.7 CAN Protocol Related Frames ................................................................................. 21-27 21.4.8 Time Stamp ............................................................................................................... 21-28 21.4.9 Bit Timing ................................................................................................................. 21-28 21.4.10 FlexCAN Error Counters ......................................................................................... 21-30 FlexCAN Initialization Sequence .............................................................................................. 21-31 21.5.1 Interrupts ................................................................................................................... 21-31 Chapter 22 Integrated Security Engine (SEC) 22.1 22.2 22.3 22.4 22.5 22.6 22.7 22.8 Features ....................................................................................................................................... 22-1 ColdFire Security Architecture ................................................................................................... 22-1 Block Diagram ............................................................................................................................ 22-2 Overview ..................................................................................................................................... 22-2 22.4.1 Bus Interface ............................................................................................................... 22-2 22.4.2 SEC Controller Unit ................................................................................................... 22-3 22.4.3 Crypto-Channels ......................................................................................................... 22-3 22.4.4 Execution Units (EUs) ................................................................................................ 22-4 Memory Map/Register Definition ............................................................................................... 22-8 Controller .................................................................................................................................. 22-11 22.6.1 EU Access ................................................................................................................ 22-11 22.6.2 Multiple EU Assignment .......................................................................................... 22-11 22.6.3 Multiple Channels .................................................................................................... 22-12 22.6.4 Controller Registers .................................................................................................. 22-12 Channels .................................................................................................................................... 22-18 22.7.1 Crypto-Channel Registers ........................................................................................ 22-19 ARC Four Execution Unit (AFEU) ........................................................................................... 22-28 22.8.1 AFEU Register Map ................................................................................................. 22-28 22.8.2 AFEU Reset Control Register (AFRCR) ................................................................. 22-28 MCF548x Reference Manual, Rev. 5 Freescale Semiconductor xv 22.8.3 AFEU Status Register (AFSR) ................................................................................. 22-29 22.8.4 AFEU Interrupt Status Register (AFISR) ................................................................. 22-31 22.8.5 AFEU Interrupt Mask Register (AFIMR) ................................................................ 22-32 22.9 Data Encryption Standard Execution Units (DEU) ................................................................... 22-34 22.9.1 DEU Register Map ................................................................................................... 22-34 22.9.2 DEU Reset Control Register (DRCR) ...................................................................... 22-34 22.9.3 DEU Status Register (DSR) ..................................................................................... 22-35 22.9.4 DEU Interrupt Status Register (DISR) ..................................................................... 22-37 22.9.5 DEU Interrupt Mask Register (DIMR) ..................................................................... 22-39 22.10 Message Digest Execution Unit (MDEU) ................................................................................. 22-40 22.10.1 MDEU Register Map ................................................................................................ 22-40 22.10.2 MDEU Reset Control Register (MDRCR) ............................................................... 22-41 22.10.3 MDEU Status Register (MDSR) .............................................................................. 22-41 22.10.4 MDEU Interrupt Status Register (MDISR) .............................................................. 22-43 22.10.5 MDEU Interrupt Mask Register (MDIMR) ............................................................. 22-44 22.11 RNG Execution Unit (RNG) .................................................................................................... 22-46 22.11.1 RNG Register Map ................................................................................................... 22-46 22.11.2 RNG Reset Control Register (RNGRCR) ................................................................ 22-46 22.11.3 RNG Status Register (RNGSR) ................................................................................ 22-47 22.11.4 RNG Interrupt Status Register (RNGISR) ............................................................... 22-48 22.11.5 RNG Interrupt Mask Register (RNGIMR) ............................................................... 22-49 22.12 Advanced Encryption Standard Execution Units (AESU) ...................................................... 22-50 22.12.1 AESU Register Map ................................................................................................. 22-50 22.12.2 AESU Reset Control Register (AESRCR) ............................................................... 22-50 22.12.3 AESU Status Register (AESSR) .............................................................................. 22-51 22.12.4 AESU Interrupt Status Register (AESISR) .............................................................. 22-53 22.12.5 AESU Interrupt Mask Register (AESIMR) .............................................................. 22-54 22.13 Descriptors ................................................................................................................................ 22-56 22.13.1 Descriptor Structure .................................................................................................. 22-56 22.13.2 Descriptor Chaining .................................................................................................. 22-61 22.13.3 Descriptor Type Formats ......................................................................................... 22-62 22.13.4 Descriptor Classes .................................................................................................... 22-64 22.14 EU Specific Data Packet Descriptors ....................................................................................... 22-67 22.14.1 AFEU Mode Options and Data Packet Descriptors ................................................. 22-67 22.14.2 DEU Mode Options and Data Packet Descriptors ................................................... 22-72 22.14.3 MDEU Mode Options and Data Packet Descriptors ................................................ 22-77 22.14.4 RNG Data Packet Descriptors .................................................................................. 22-82 22.14.5 AESU Mode Options and Data Packet Descriptors ................................................. 22-83 22.14.6 Multi-Function Data Packet Descriptors .................................................................. 22-90 Chapter 23 IEEE 1149.1 Test Access Port (JTAG) 23.1 Introduction ................................................................................................................................. 23-1 23.1.1 Block Diagram ............................................................................................................ 23-1 23.1.2 Features ....................................................................................................................... 23-2 MCF548x Reference Manual, Rev. 5 xvi Freescale Semiconductor 23.2 23.3 23.4 23.5 23.1.3 Modes of Operation .................................................................................................... 23-2 External Signal Description ........................................................................................................ 23-2 23.2.1 Detailed Signal Description ........................................................................................ 23-2 Memory Map/Register Definition ............................................................................................... 23-4 23.3.1 Memory Map .............................................................................................................. 23-4 23.3.2 Register Descriptions .................................................................................................. 23-4 Functional Description ................................................................................................................ 23-6 23.4.1 JTAG Module ............................................................................................................. 23-6 23.4.2 TAP Controller ........................................................................................................... 23-6 23.4.3 JTAG Instructions ....................................................................................................... 23-7 Initialization/Application Information ........................................................................................ 23-9 23.5.1 Restrictions ................................................................................................................. 23-9 23.5.2 Nonscan Chain Operation ........................................................................................... 23-9 Chapter 24 Multichannel DMA 24.1 Introduction ................................................................................................................................. 24-1 24.1.1 Block Diagram ............................................................................................................ 24-1 24.1.2 Overview .................................................................................................................... 24-2 24.1.3 Features ....................................................................................................................... 24-2 24.2 External Signals ........................................................................................................................... 24-3 24.2.1 DREQ[1:0] ................................................................................................................ 24-3 24.2.2 DACK[1:0] ................................................................................................................ 24-3 24.3 Memory Map/Register Definitions ............................................................................................. 24-3 24.3.1 DMA Task Memory ................................................................................................... 24-3 24.3.2 Memory Structure ....................................................................................................... 24-4 24.3.3 DMA Registers ........................................................................................................... 24-5 24.3.4 External Request Module Registers ......................................................................... 24-20 24.4 Functional Description .............................................................................................................. 24-22 24.4.1 Tasks ......................................................................................................................... 24-22 24.4.2 Descriptors ................................................................................................................ 24-23 24.4.3 Task Initialization ..................................................................................................... 24-23 24.4.4 Initiators .................................................................................................................... 24-23 24.4.5 Prioritization ............................................................................................................. 24-24 24.4.6 Context Switch ......................................................................................................... 24-24 24.4.7 Data Movement ........................................................................................................ 24-24 24.4.8 Data Manipulation .................................................................................................... 24-24 24.4.9 Line Buffers .............................................................................................................. 24-26 24.4.10 Termination of Loop ................................................................................................. 24-27 24.4.11 Interrupts ................................................................................................................... 24-27 24.4.12 Debug Unit ............................................................................................................... 24-27 24.5 Programming Model ................................................................................................................. 24-27 24.5.1 Register Initialization ............................................................................................... 24-27 24.5.2 Task Memory ............................................................................................................ 24-28 24.6 Timing Diagrams ....................................................................................................................... 24-30 MCF548x Reference Manual, Rev. 5 Freescale Semiconductor xvii 24.6.1 24.6.2 24.6.3 Level-Triggered Requests ......................................................................................... 24-30 Edge-Triggered Requests ......................................................................................... 24-30 Pipelined Requests .................................................................................................... 24-31 Chapter 25 Comm Bus FIFO Interface 25.1 Introduction ................................................................................................................................. 25-1 25.1.1 Block Diagram ............................................................................................................ 25-1 25.1.2 Overview .................................................................................................................... 25-1 25.1.3 Features ....................................................................................................................... 25-2 25.2 Memory Map/Register Definition ............................................................................................... 25-2 25.2.1 FIFO Interface Registers ............................................................................................ 25-2 25.3 Functional Description .............................................................................................................. 25-12 25.3.1 Flow control .............................................................................................................. 25-12 25.3.2 Wait Conditions ........................................................................................................ 25-14 25.3.3 Error reporting .......................................................................................................... 25-16 25.3.4 Debug Operation ...................................................................................................... 25-17 Chapter 26 Comm Timer Module (CTM) 26.1 Introduction ................................................................................................................................. 26-1 26.1.1 Block Diagrams .......................................................................................................... 26-1 26.1.2 Overview .................................................................................................................... 26-3 26.2 External Signals ........................................................................................................................... 26-3 26.2.1 Comm Timer External Clock[7:0] .............................................................................. 26-3 26.3 Memory Map/Register Definition ............................................................................................... 26-4 26.3.1 Timer Module Register Map ...................................................................................... 26-5 26.3.2 Register Descriptions .................................................................................................. 26-5 26.4 Functional Description ................................................................................................................ 26-9 26.4.1 Fixed and Variable Timers In Baud Clock Generator Mode ...................................... 26-9 26.4.2 Fixed Timer Channel in Task Initiator Mode ............................................................. 26-9 26.4.3 Variable Timer Channel in Task Initiator Mode ....................................................... 26-11 Chapter 27 Programmable Serial Controller (PSC) 27.1 Introduction ................................................................................................................................. 27-1 27.1.1 Block Diagram ............................................................................................................ 27-1 27.1.2 Overview .................................................................................................................... 27-1 27.1.3 Features ....................................................................................................................... 27-1 27.1.4 Modes of Operation .................................................................................................... 27-1 27.2 Signal Description ....................................................................................................................... 27-2 27.2.1 PSCnCTS/PSCBCLK ................................................................................................. 27-2 27.2.2 PSCnRTS/PSCFSYNC ............................................................................................... 27-2 27.2.3 PSCnrxd ...................................................................................................................... 27-2 27.2.4 pscntxd ........................................................................................................................ 27-3 MCF548x Reference Manual, Rev. 5 xviii Freescale Semiconductor 27.3 27.4 27.5 27.6 27.7 27.2.5 Signal Properties in Each Mode ................................................................................. 27-3 Memory Map/Register Definition ............................................................................................... 27-3 27.3.1 Overview .................................................................................................................... 27-3 27.3.2 Module Memory Map ................................................................................................. 27-3 27.3.3 Register Descriptions .................................................................................................. 27-5 Functional Description .............................................................................................................. 27-37 27.4.1 UART Mode ............................................................................................................. 27-37 27.4.2 Multidrop Mode ........................................................................................................ 27-38 27.4.3 Modem8 Mode ......................................................................................................... 27-39 27.4.4 Modem16 Mode ....................................................................................................... 27-40 27.4.5 AC97 Mode .............................................................................................................. 27-41 27.4.6 SIR Mode .................................................................................................................. 27-43 27.4.7 MIR Mode ................................................................................................................ 27-43 27.4.8 FIR Mode .................................................................................................................. 27-44 27.4.9 PSC FIFO System ..................................................................................................... 27-45 27.4.10 Looping Modes ......................................................................................................... 27-48 Resets ........................................................................................................................................ 27-49 27.5.1 General ..................................................................................................................... 27-49 27.5.2 Description of Reset Operation ................................................................................ 27-49 Interrupts ................................................................................................................................... 27-50 27.6.1 Description of Interrupt Operation ........................................................................... 27-50 Software Environment ............................................................................................................... 27-50 27.7.1 General ..................................................................................................................... 27-50 27.7.2 Configuration ............................................................................................................ 27-51 27.7.3 Programming ............................................................................................................ 27-57 Chapter 28 DMA Serial Peripheral Interface (DSPI) 28.1 28.2 28.3 28.4 Overview ..................................................................................................................................... 28-1 Features ....................................................................................................................................... 28-1 Block Diagram ............................................................................................................................ 28-2 Modes of Operation ..................................................................................................................... 28-2 28.4.1 Master Mode ............................................................................................................... 28-2 28.4.2 Slave Mode ................................................................................................................. 28-2 28.5 Signal Description ....................................................................................................................... 28-3 28.5.1 Overview .................................................................................................................... 28-3 28.5.2 Detailed Signal Descriptions ...................................................................................... 28-3 28.6 Memory Map and Registers ........................................................................................................ 28-4 28.6.1 DSPI Module Configuration Register (DMCR) ......................................................... 28-5 28.6.2 DSPI Transfer Count Register (DTCR) ...................................................................... 28-7 28.6.3 DSPI Clock and Transfer Attributes Registers 0–7 (DCTARn) ................................. 28-7 28.6.4 DSPI Status Register (DSR) ..................................................................................... 28-11 28.6.5 DSPI DMA/Interrupt Request Select Register (DIRSR) .......................................... 28-13 28.6.6 DSPI Tx FIFO Register (DTFR) .............................................................................. 28-15 28.6.7 DSPI Rx FIFO Register (DRFR) .............................................................................. 28-16 MCF548x Reference Manual, Rev. 5 Freescale Semiconductor xix 28.6.8 DSPI Tx FIFO Debug Registers 0–3 (DTFDRn) ..................................................... 28-17 28.6.9 DSPI Rx FIFO Debug Registers 0–3 (DRFDRn) ..................................................... 28-17 28.7 Functional Description .............................................................................................................. 28-18 28.7.1 Start and Stop of DSPI Transfers .............................................................................. 28-19 28.7.2 Serial Peripheral Interface (SPI) .............................................................................. 28-20 28.7.3 DSPI Baud Rate and Clock Delay Generation ......................................................... 28-22 28.7.4 Transfer Formats ....................................................................................................... 28-25 28.7.5 Continuous Serial Communications Clock .............................................................. 28-30 28.7.6 Interrupts/DMA Requests ......................................................................................... 28-31 28.8 Initialization and Application Information ................................................................................ 28-33 28.8.1 How to Change Queues ............................................................................................ 28-33 28.8.2 Baud Rate Settings ................................................................................................... 28-33 28.8.3 Delay Settings ........................................................................................................... 28-34 28.8.4 Calculation of FIFO Pointer Addresses .................................................................... 28-35 Chapter 29 I C Interface 2 29.1 Introduction ................................................................................................................................. 29-1 29.1.1 Block Diagram ............................................................................................................ 29-1 29.1.2 I2C Overview ............................................................................................................. 29-2 29.1.3 Features ....................................................................................................................... 29-2 29.2 External Signals ........................................................................................................................... 29-2 29.3 Memory Map/Register Definition ............................................................................................... 29-3 29.3.1 I2C Register Map ....................................................................................................... 29-3 29.3.2 Register Descriptions .................................................................................................. 29-3 29.4 Functional Description ................................................................................................................ 29-8 29.4.1 START Signal ............................................................................................................. 29-9 29.4.2 Slave Address Transmission ....................................................................................... 29-9 29.4.3 STOP Signal ............................................................................................................... 29-9 29.4.4 Data Transfer .............................................................................................................. 29-9 29.4.5 Acknowledge ............................................................................................................ 29-10 29.4.6 Repeated Start ........................................................................................................... 29-11 29.4.7 Clock Synchronization and Arbitration .................................................................... 29-11 29.4.8 Handshaking and Clock Stretching .......................................................................... 29-12 29.5 Initialization Sequence .............................................................................................................. 29-12 29.5.1 Transfer Initiation and Interrupt ............................................................................... 29-13 29.5.2 Post-Transfer Software Response ............................................................................. 29-14 29.5.3 Generation of STOP ................................................................................................. 29-15 29.5.4 Generation of Repeated START ............................................................................... 29-16 29.5.5 Slave Mode ............................................................................................................... 29-16 29.5.6 Arbitration Lost ........................................................................................................ 29-18 29.5.7 Flow Control ............................................................................................................. 29-18 MCF548x Reference Manual, Rev. 5 xx Freescale Semiconductor Chapter 30 USB 2.0 Device Controller 30.1 Introduction ................................................................................................................................. 30-1 30.1.1 Overview .................................................................................................................... 30-1 30.1.2 Features ....................................................................................................................... 30-1 30.1.3 Block Diagram ............................................................................................................ 30-2 30.2 Memory Map/Register Definition ............................................................................................... 30-4 30.2.1 USB Memory Map ..................................................................................................... 30-4 30.2.2 USB Request, Control, and Status Registers .............................................................. 30-9 30.2.3 USB Counter Registers ............................................................................................. 30-23 30.2.4 Endpoint Context Registers ...................................................................................... 30-27 30.2.5 USB Endpoint FIFO Registers ................................................................................. 30-34 30.3 Functional Description .............................................................................................................. 30-47 30.3.1 Interrupts ................................................................................................................... 30-47 30.3.2 Device Initialization ................................................................................................. 30-47 30.3.3 Exception Handling .................................................................................................. 30-50 30.3.4 Data Transfer Operations .......................................................................................... 30-50 Chapter 31 Fast Ethernet Controller (FEC) 31.1 Introduction ................................................................................................................................. 31-1 31.1.1 MCF548x Family Products ......................................................................................... 31-1 31.1.2 Block Diagram ............................................................................................................ 31-1 31.1.3 Overview .................................................................................................................... 31-2 31.1.4 Features ....................................................................................................................... 31-3 31.1.5 Modes of Operation .................................................................................................... 31-3 31.2 External Signals ........................................................................................................................... 31-4 31.2.1 Transmit Clock (EnTXCLK) ...................................................................................... 31-4 31.2.2 Receive Clock (EnRXCLK) ....................................................................................... 31-4 31.2.3 Transmit Enable (EnTXEN) ....................................................................................... 31-4 31.2.4 Transmit Data[3:0] (EnTXD[3:0]) ............................................................................. 31-4 31.2.5 Transmit Error (EnTXER) .......................................................................................... 31-5 31.2.6 Receive Data Valid (EnRXDV) .................................................................................. 31-5 31.2.7 Receive Data[3:0] (EnRXD[3:0]) ............................................................................... 31-5 31.2.8 Receive Error (EnRXER) ........................................................................................... 31-5 31.2.9 Carrier Sense (EnCRS) ............................................................................................... 31-5 31.2.10 Collision (EnCOL) ..................................................................................................... 31-5 31.2.11 Management Data Clock (EnMDC) ........................................................................... 31-5 31.2.12 Management Data (EnMDIO) .................................................................................... 31-5 31.3 Memory Map/Register Definition ............................................................................................... 31-6 31.3.1 Top Level Module Memory Map ............................................................................... 31-6 31.3.2 Detailed Memory Map (Control/Status Registers) ..................................................... 31-7 31.3.3 MIB Block Counters Memory Map ........................................................................... 31-8 31.4 Functional Description .............................................................................................................. 31-43 31.4.1 Initialization Sequence ............................................................................................. 31-43 MCF548x Reference Manual, Rev. 5 Freescale Semiconductor xxi 31.4.2 31.4.3 31.4.4 31.4.5 31.4.6 31.4.7 31.4.8 31.4.9 31.4.10 31.4.11 31.4.12 31.4.13 31.4.14 Frame Control/Status Words .................................................................................... 31-44 Network Interface Options ....................................................................................... 31-46 FEC Frame Transmission ......................................................................................... 31-46 FEC Frame Reception .............................................................................................. 31-47 Ethernet Address Recognition .................................................................................. 31-48 Hash Algorithm ........................................................................................................ 31-49 Full Duplex Flow Control ........................................................................................ 31-52 Inter-Packet Gap (IPG) Time .................................................................................... 31-53 Collision Handling .................................................................................................... 31-53 Internal and External Loopback ............................................................................... 31-53 Ethernet Error-Handling Procedure .......................................................................... 31-54 MII Data Frame ........................................................................................................ 31-55 MII Management Frame Structure ........................................................................... 31-56 Chapter 32 Mechanical Data 32.1 Package ........................................................................................................................................ 32-1 32.2 Pinout .......................................................................................................................................... 32-1 32.3 Mechanical Diagrams .................................................................................................................. 32-8 32.3.1 MCF5485/5484 Mechanical Diagram ........................................................................ 32-8 32.3.2 MCF5483/5482 Mechanical Diagram ...................................................................... 32-12 32.4 MCF5481/5480 Mechanical Diagram ....................................................................................... 32-16 32.5 Mechanicals 388-pin PBGA Package Outline .......................................................................... 32-19 Appendix A MCF548x Memory Map MCF548x Reference Manual, Rev. 5 xxii Freescale Semiconductor About This Book The primary objective of this reference manual is to define the functionality of the MCF548x processors for use by software and hardware developers. The information in this book is subject to change without notice, as described in the disclaimers on the title page of this book. As with any technical documentation, it is the readers’ responsibility to be sure they are using the most recent version of the documentation. To locate any published errata or updates for this document, refer to the world-wide web at http://www.freescale.com/coldfire. Audience This manual is intended for system software and hardware developers and applications programmers who want to develop products for the MCF548x. It is assumed that the reader understands operating systems, microprocessor system design, basic principles of software and hardware, and basic details of the ColdFire architecture. Organization Following is a summary and a brief description of the major sections of this manual: • Chapter 1, “Overview,” includes general descriptions of the modules and features incorporated in the MCF548x, focussing in particular on new features. • Chapter 2, “Signal Descriptions,” provides an alphabetical listing of MCF548x signals, including which are inputs or outputs, how they are multiplexed, and the state of each signal at reset. • Part I, “Processor Core,” is intended for system designers who need to understand the operation of the MCF548x ColdFire core and its enhanced multiply/accumulate (EMAC) execution unit. It describes the programming and exception models, Harvard memory implementation, and debug module. Part 1 contains the following chapters: — Chapter 3, “ColdFire Core,” provides an overview of the microprocessor core of the MCF548x. The chapter begins with a description of enhancements from the V3 ColdFire core, and then fully describes the V4e programming model as it is implemented on the MCF548x. It also includes a full description of exception handling, data formats, an instruction set summary, and a table of instruction timings. — Chapter 4, “Enhanced Multiply-Accumulate Unit (EMAC),” describes the MCF548x enhanced multiply/accumulate unit, which executes integer multiply, multiply-accumulate, and miscellaneous register instructions. The EMAC is integrated into the operand execution pipeline (OEP). — Chapter 5, “Memory Management Unit (MMU),” describes describes the ColdFire virtual memory management unit (MMU), which provides virtual-to-physical address translation and memory access control. — Chapter 6, “Floating-Point Unit (FPU),” describes instructions implemented in the floating-point unit (FPU) designed for use with the ColdFire family of microprocessors. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor xxiii • • — Chapter 7, “Local Memory,” describes the MCF548x implementation of the ColdFire V4e local memory specification. — Chapter 8, “Debug Support,” describes the Revision C enhanced hardware debug support in the MCF548x. This revision of the ColdFire debug architecture encompasses earlier revisions. Part II, “System Integration Unit,” describes the system integration unit, which provides overall control of the bus and serves as the interface between the ColdFire core processor complex and internal peripheral devices. It includes a general description of the SIU and individual chapters that describe components of the SIU, such as the interrupt controller, general purpose timers, slice timers, and GPIOs. Part II contains the following chapters: — Chapter 9, “System Integration Unit (SIU),” describes the SIU programming model, bus arbitration, and system-protection functions for the MCF548x. — Chapter 10, “Internal Clocks and Bus Architecture,” describes the clocking and internal buses of the MCF548x and discusses the main functional blocks controlling the XL bus and the XL bus arbiter. — Chapter 11, “General Purpose Timers (GPT),” describes the functionality of the four general purpose timers, GPT0–GPT3. — Chapter 12, “Slice Timers (SLT),” describes the two slice timers, shorter term periodic interrupts, used in the MCF548x. — Chapter 13, “Interrupt Controller,” describes operation of the interrupt controller portion of the SIU. Includes descriptions of the registers in the interrupt controller memory map and the interrupt priority scheme. — Chapter 14, “Edge Port Module (EPORT),” describes EPORT module functionality. — Chapter 15, “GPIO,” describes the operation and programming model of the parallel port pin assignment, direction-control, and data registers. Part III, “On-Chip Integration,” describes the on-chip integration for the MCF548x device. It includes descriptions of the system SRAM, FlexBus interface, SDRAM controller, PCI, and SEC cryptography accelerator. Part III contains the following chapters: — Chapter 16, “32-Kbyte System SRAM,” describes the MCF548x on-chip system SRAM implementation. It covers general operations, configuration, and initialization. — Chapter 17, “FlexBus,” describes data transfer operations, error conditions, and reset operations. It describes transfers initiated by the MCF548x and by an external master, and includes detailed timing diagrams showing the interaction of signals in supported bus operations. — Chapter 18, “SDRAM Controller (SDRAMC),” describes configuration and operation of the synchronous DRAM controller component of the SIU. It includes a description of signals involved in DRAM operations, including chip select signals and their address, mask, and control registers. — Chapter 19, “PCI Bus Controller,” details the operation of the PCI bus controller for the MCF548x. — Chapter 20, “PCI Bus Arbiter Module,” describes the MCF548x PCI bus arbiter module, including timing for request and grant handshaking, the arbitration process, and the register in the PCI bus arbiter programing model. MCF548x Reference Manual, Rev. 5 xxiv Freescale Semiconductor Suggested Reading • • — Chapter 21, “FlexCAN,” describes the MCF548 implementation of the controller area network (CAN) protocol. This chapter describes FlexCAN module operation and provides a programming model. — Chapter 22, “Integrated Security Engine (SEC),” provides an overview of the MCF548x security encryption controller. — Chapter 23, “IEEE 1149.1 Test Access Port (JTAG),” describes configuration and operation of the MCF548x JTAG test implementation. It describes the use of JTAG instructions and provides information on how to disable JTAG functionality. Part IV, “Communications Subsystem,” contains chapters that discuss the operation and configuration of the communications I/O subsystem including the MCF548x multichannel DMA, communications timer, PSC, FEC, DSPI, and USB2, and I2C. — Chapter 24, “Multichannel DMA,” provides an overview of the multichannel DMA controller module including the operation of the external DMA request signals. — Chapter 26, “Comm Timer Module (CTM),” contains a detailed description of the communications timer module, which functions as a baud clock generator or as a DMA task initiator. — Chapter 27, “Programmable Serial Controller (PSC),” provides an overview of asynchronous, synchronous, and IrDA 1.1 compliant receiver/transmitter serial communications of the MCF548x. — Chapter 28, “DMA Serial Peripheral Interface (DSPI),” describes the use of the DMA serial peripheral interface (DSPI) implemented on the MCF548x processor, including details of the DSPI data transfers. The chapter concludes with timing diagrams and the DSPI features that support Tx and Rx FIFO queue management. — Chapter 29, “I2C Interface,” describes the MCF548x I2C module, including I2C protocol, clock synchronization, and the registers in the I2C programing model. It also provides programming examples. — Chapter 30, “USB 2.0 Device Controller,” provides an overview of the USB 2.0 device controller module used in the MCF548x. — Chapter 31, “Fast Ethernet Controller (FEC),” provides a feature-set overview, a functional block diagram, and transceiver connection information for both MII (Media Independent Interface) and 7-wire serial interfaces. It also provides describes operation and the programming model. Part V, “Mechanical,” provides a pinout and both electrical and functional descriptions of the MCF548x signals. It also describes how these signals interact to support the variety of bus operations shown in timing diagrams. — Chapter 32, “Mechanical Data,” provides a functional pin listing and package diagram for the MCF548x. Suggested Reading This section lists additional reading that provides background for the information in this manual as well as general information about the ColdFire architecture. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor xxv General Information The following documentation provides useful information about the ColdFire architecture and computer architecture in general: • ColdFire Programmers Reference Manual (CFPRM) • Using Microprocessors and Microcomputers: The Motorola Family, William C. Wray, Ross Bannatyne, Joseph D. Greenfield • Computer Architecture: A Quantitative Approach, Second Edition, by John L. Hennessy and David A. Patterson. • Computer Organization and Design: The Hardware/Software Interface, Second Edition, David A. Patterson and John L. Hennessy. ColdFire Documentation The ColdFire documentation is available from the sources listed on the back cover of this manual. Document order numbers are included in parentheses for ease in ordering. • ColdFire Programmers Reference Manual, R1.0 (CFPRM) • Reference manuals—These books provide details about individual ColdFire implementations and are intended to be used in conjunction with The ColdFire Programmers Reference Manual. These include the following: — ColdFire CF4e Core User's Manual (V4ECFUM) — MCF5475 Reference Manual (MCF5475RM) — MCF5485 Reference Manual (MCF5485RM) Additional literature on ColdFire implementations is being released as new processors become available. For a current list of ColdFire documentation, refer to the World Wide Web at http://www.freescale.com/coldfire. Conventions This document uses the following notational conventions: MNEMONICS In text, instruction mnemonics are shown in uppercase. mnemonics In code and tables, instruction mnemonics are shown in lowercase. italics Italics indicate variable command parameters. Book titles in text are set in italics. 0x0 Prefix to denote hexadecimal number 0b0 Prefix to denote binary number REG[FIELD] Abbreviations for registers are shown in uppercase. Specific bits, fields, or ranges appear in brackets. For example, RAMBAR[BA] identifies the base address field in the RAM base address register. nibble A 4-bit data unit byte An 8-bit data unit word A 16-bit data unit MCF548x Reference Manual, Rev. 5 xxvi Freescale Semiconductor Acronyms and Abbreviations longword x n ¬ & | A 32-bit data unit In some contexts, such as signal encodings, x indicates a don’t care. Used to express an undefined numerical value NOT logical operator AND logical operator OR logical operator Register Conventions This reference manual uses the register diagram format shown below. R 31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 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 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 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 R DFL W Reset Reg Addr 0 0 0 0x00C Table i. Example Register Diagram Acronyms and Abbreviations Table ii lists acronyms and abbreviations used in this document. Table ii. . Acronyms and Abbreviated Terms Term Meaning ADC Analog-to-digital conversion ALU Arithmetic logic unit AVEC Autovector BDM Background debug mode BIST Built-in self test BSDL Boundary-scan description language CODEC Code/decode comm bus Internal communications bus DAC Digital-to-analog conversion DMA Direct memory access DSP Digital signal processing MCF548x Reference Manual, Rev. 5 Freescale Semiconductor xxvii Table ii. . Acronyms and Abbreviated Terms (continued) Term Meaning EA Effective address EDO Extended data output (DRAM) FIFO First-in, first-out GPIO General-purpose I/O 2 I C Inter-integrated circuit IEEE Institute for Electrical and Electronics Engineers IFP Instruction fetch pipeline IPL Interrupt priority level JEDEC Joint Electron Device Engineering Council JTAG Joint Test Action Group LIFO Last-in, first-out LRU Least recently used LSB Least-significant byte lsb Least-significant bit MAC Multiple accumulate unit MBAR Memory base address register MSB Most-significant byte msb Most-significant bit Mux Multiplex NOP No operation OEP Operand execution pipeline PC Program counter PCLK Processor clock PLL Phase-locked loop PLRU Pseudo least recently used POR Power-on reset PQFP Plastic quad flat pack RISC Reduced instruction set computing Rx Receive SIM System integration module SOF Start of frame TAP Test access port TTL Transistor-to-transistor logic Tx Transmit MCF548x Reference Manual, Rev. 5 xxviii Freescale Semiconductor Terminology and Notational Conventions Table ii. . Acronyms and Abbreviated Terms (continued) Term Meaning UART Universal asynchronous/synchronous receiver transmitter XLB bus Internal 64-bit bus Terminology and Notational Conventions Table iii shows notational conventions used throughout this document. Table iii. Notational Conventions Instruction Operand Syntax Opcode Wildcard cc Logical condition (example: NE for not equal) Register Specifications An Ay,Ax Any address register n (example: A3 is address register 3) Source and destination address registers, respectively Dn Any data register n (example: D5 is data register 5) Dy,Dx Source and destination data registers, respectively Rc Any control register (example VBR is the vector base register) Rm MAC registers (ACC, MAC, MASK) Rn Any address or data register Rw Destination register w (used for MAC instructions only) Ry,Rx Xi Any source and destination registers, respectively index register i (can be an address or data register: Ai, Di) Register Names ACC MAC accumulator register CCR Condition code register (lower byte of SR) MACSR MAC status register MASK MAC mask register PC Program counter SR Status register Port Name PSTDDATA Processor status/debug data port Miscellaneous Operands #<data> <ea> Immediate data following the 16-bit operation word of the instruction Effective address MCF548x Reference Manual, Rev. 5 Freescale Semiconductor xxix Table iii. Notational Conventions (continued) Instruction <ea>y,<ea>x <label> <list> Operand Syntax Source and destination effective addresses, respectively Assembly language program label List of registers for MOVEM instruction (example: D3–D0) <shift> Shift operation: shift left (<<), shift right (>>) <size> Operand data size: byte (B), word (W), longword (L) bc Both instruction and data caches dc Data cache ic Instruction cache # <vector> <> <xxx> Identifies the 4-bit vector number for trap instructions identifies an indirect data address referencing memory identifies an absolute address referencing memory dn Signal displacement value, n bits wide (example: d16 is a 16-bit displacement) SF Scale factor (x1, x2, x4 for indexed addressing mode, <<1n>> for MAC operations) Operations + Arithmetic addition or postincrement indicator – Arithmetic subtraction or predecrement indicator x Arithmetic multiplication / Arithmetic division ~ Invert; operand is logically complemented & Logical AND | Logical OR ^ Logical exclusive OR << Shift left (example: D0 << 3 is shift D0 left 3 bits) >> Shift right (example: D0 >> 3 is shift D0 right 3 bits) → Source operand is moved to destination operand ←→ Two operands are exchanged sign-extended All bits of the upper portion are made equal to the high-order bit of the lower portion If <condition> then <operations> else <operations> Test the condition. If true, the operations after ‘then’ are performed. If the condition is false and the optional ‘else’ clause is present, the operations after ‘else’ are performed. If the condition is false and else is omitted, the instruction performs no operation. Refer to the Bcc instruction description as an example. Subfields and Qualifiers {} Optional operation () Identifies an indirect address dn Displacement value, n-bits wide (example: d16 is a 16-bit displacement) MCF548x Reference Manual, Rev. 5 xxx Freescale Semiconductor Terminology and Notational Conventions Table iii. Notational Conventions (continued) Instruction Address Operand Syntax Calculated effective address (pointer) Bit Bit selection (example: Bit 3 of D0) lsb Least significant bit (example: lsb of D0) LSB Least significant byte LSW Least significant word msb Most significant bit MSB Most significant byte MSW Most significant word Condition Code Register Bit Names C Carry N Negative V Overflow X Extend Z Zero Table iv. MCF548x Revision History Section/Page Substantive Changes Revision 1.0 (03/2004) Initial release. Revision 1.1 (03/2004 Figure 15-1/Page 15-2 Changed instances of FEC2 to FEC1 and FEC1 to FEC0. 31.3.1/31-6– 31.3.3.1/31-10 Changed instances of FEC2 to FEC1 and FEC1 to FEC0. Revision 1.2 (03/2004) Revision 2.0 (10/2004) Many content changes, the biggest being greatly enhancing the MC-DMA chapter and adding Clocks and Internal Buses chapter. Many editorial changes. Revision 2.1 (10/2004) Chapter 17 Took out FlexCan chapter. Fixed timing diagrams in FlexBus chapter. Revision 3 (01/2006) Throughout See revision 3 or higher of the MCF5485RMAD document for a list of all changes between the previous revision. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor xxxi Table iv. MCF548x Revision History (continued) Section/Page Substantive Changes Revision 4 (07/2006) Throughout See revision 4 or higher of the MCF5485RMAD document for a list of all changes between the previous revision. Revision 5 (4/2009) Throughout See revision 5 or higher of the MCF5485RMAD document for a list of all changes between the previous revision. MCF548x Reference Manual, Rev. 5 xxxii Freescale Semiconductor Chapter 1 Overview This chapter provides an overview of the MCF548x microprocessor features, including the major functional components. 1.1 MCF548x Family Overview The MCF548x family is based on the ColdFire V4e core, a complex which comprises the ColdFire V4 central processor unit (CPU), an enhanced multiply-accumulate unit (EMAC), a memory management unit (MMU), a double-precision floating point unit (FPU) conforming to standard IEEE-754, and controllers for caches and local data memories. The MCF548x family is capable of performing at an operating frequency of up to 200 MHz or 308 MIPS (Dhrystone 2.1). To maximize throughput, the MCF548x family incorporates three independent external bus interfaces: 1. The general-purpose local bus (FlexBus) is used for system boot memories and simple peripherals and has up to six chip selects. 2. Program code and data can be stored in SDRAM connected to a dedicated 32-bit double data rate (DDR) bus that can run at up to one-half of the CPU core frequency. The glueless DDR SDRAM controller handles all address multiplexing, input and output strobe timing, and memory bus clock generation. 3. A 32-bit PCI bus compliant with the version 2.2 specification and running at a typical frequency of 25 MHz or 50 MHz supports peripherals that require high bandwidth, the ability to arbitrate for bus mastership, and access to internal MCF548x memory resources. The MCF548x family provides substantial communications functionality by integrating the following connectivity peripherals: • Up to two 10/100 Mbps fast Ethernet controllers (FECs) • One optional USB 2.0 device (slave) module with seven endpoints and an integrated transceiver • Up to four UART/USART/IRDA/modem programmable serial controllers (PSCs) • One DMA serial peripheral interface (DSPI) • One inter-integrated circuit (I2C™) bus controller • Two controller area network 2.0B (FlexCAN) interfaces with 16 message buffers each Additionally, the MCF548x provides hardware support for a range of Internet security standards with an optional bus-mastering cryptography accelerator. This module incorporates units to speed DES/3DES and AES block ciphers, the RC4 stream cipher, bulk data hashing (MD5/SHA-1/SHA-256/HMAC), and random number generation. Hardware acceleration of these functions is critical to avoiding the throughput bottlenecks associated with software-only implementations of SSH, SSL/TLS, IPsec, SRTP, WEP, and other security standards. The incorporation of cryptography acceleration makes the MCF548x family a compelling solution for a wide range of office automation, industrial control, and SOHO networking devices that must have the ability to securely transmit critical equipment control information across typically insecure Ethernet data networks. Additional features of MCF548x products include a watchdog timer, two 32-bit slice timers for RTOS scheduling and alarm functionality, up to four 32-bit general-purpose timers with capture, compare, and pulse width modulation capability, a multisource vectored interrupt controller, a phase-locked loop (PLL) to generate the system clock, 32 Kbytes of SRAM for high-speed local data storage, and multiple general-purpose I/O ports. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 1-1 With on-chip support for multiple common communications interfaces, MCF548x products require only the addition of memories and certain physical layer transceivers to be cost-effective system solutions for many applications. Such applications include industrial routers, high-end POS terminals, building automation systems, and process control equipment. MCF548x products require four supply voltages: 1.5V for the high-performance, low power, internal core logic, 2.5V for the DDR SDRAM bus interface, 1.25V for the DDR SDRAM VREF, and 3.3V for all other I/O functionality, including the PCI and FlexBus interfaces. 1.2 MCF548x Block Diagram Figure 1-1 shows a top-level block diagram of the MCF548x products. ColdFire V4e Core FPU, MMU EMAC 32K D-cache 32K I-cache PLL DDR SDRAM Interface FlexBus Interface XL Bus Arbiter Memory Controller FlexBus Controller Cryptography Accelerator3 XL Bus Read/Write Write DMA DMA Bus Read 32K System SRAM GP Timers x 4 PCI 2.2 Controller Multichannel DMA Master Bus Interface and FIFOs FlexCAN x2 PCI Interface & FIFOs CommBus DSPI I2C PSC x 4 FEC1 FEC22 Perpheral Communications I/O Interface & Ports USB 2.0 DEVICE1 Communications I/O Subsystem Slice Timers x 2 PCI I/O Interface and Ports Watchdog Timer R/W Master/Slave Interface Crypto Interrupt Controller Slave Perpheral I/O Interface & Ports System Integration Unit XL Bus USB 2.0 PHY1 1 Available in MCF5485, MCF5484, MCF5483, and MCF5482 devices. Available in MCF5485, MCF5484, MCF5481, and MCF5480 devices. 3 Available in MCF5485, MCF5483, and MCF5481 devices. 2 Figure 1-1. MCF548x Block Diagram MCF548x Reference Manual, Rev. 5 1-2 Freescale Semiconductor MCF548x Family Products 1.3 MCF548x Family Products Table 1-1 summarizes the products available within the MCF548x product family. All products are available in pin-compatible, 388-pin PBGA packaging allowing for ease of migration between products within the family. A printed circuit board designed using the MCF5485/4 footprint is compatible with any of the MCF548x family devices. Table 1-1. MCF548x Family Products 1.4 • Product Performance Features MCF5485 308 MIPS 200 MHz Two 10/100 Ethernet Controllers Two CAN Controllers USB 2.0 Device with Integrated PHY v2.2 PCI Controller DDR Memory Controller Encryption Accelerator MCF5484 308 MIPS 200 MHz Two 10/100 Ethernet Controllers Two CAN Controllers USB 2.0 Device with Integrated PHY v2.2 PCI Controller DDR Memory Controller MCF5483 255 MIPS 166 MHz One 10/100 Ethernet Controller Two CAN Controllers USB 2.0 Device with Integrated PHY v2.2 PCI Controller DDR Memory Controller Encryption Accelerator MCF5482 255 MIPS 166 MHz One 10/100 Ethernet Controller Two CAN Controllers USB 2.0 Device with Integrated PHY v2.2 PCI Controller DDR Memory Controller MCF5481 255 MIPS 166 MHz Two 10/100 Ethernet Controllers Two CAN Controllers v2.2 PCI Controller DDR Memory Controller Encryption Accelerator MCF5480 255 MIPS 166 MHz Two 10/100 Ethernet Controllers Two CAN Controllers v2.2 PCI Controller DDR Memory Controller Temperature Range -40 to 85 ° C -40 to 85 ° C -40 to 85 ° C -40 to 85 ° C -40 to 85 ° C -40 to 85 ° C MCF548x Family Features ColdFire V4e core — Limited superscalar V4 ColdFire processor core — Up to 200 MHz peak internal core frequency (308 Dhrystone 2.1 MIPS) — Harvard architecture – 32-Kbyte instruction cache – 32-Kbyte data cache MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 1-3 • • • • • • — Memory management unit (MMU) – Separate, 32-entry, fully-associative instruction and data translation lookahead buffers — Floating point unit (FPU) – Double-precision support that conforms to IEEE-754 standard – Eight floating point registers Internal master bus (XLB) arbiter — High performance split address and data transactions — Support for various parking modes 32-bit double data rate (DDR) synchronous DRAM (SDRAM) controller — 66–133 MHz operation — Supports both DDR and SDR DRAM — Built-in initialization and refresh — Up to four chip selects enabling up to 1 GB of external memory Version 2.2 peripheral component interconnect (PCI) bus — 32-bit target and initiator operation — Support for up to five external PCI masters — 25–50 MHz operation with PCI bus to XLB divider ratios of 1:1, 1:2, and 1:4 Flexible multi-function external bus (FlexBus) — Supports operation with the following: – Non-multiplexed 32-bit address and 32-bit data (32-bit address muxed over PCI bus–PCI not usable) – Multiplexed 32-bit address and 32-bit data (PCI usable) – Multiplexed 32-bit address and 16-bit data – Multiplexed 32-bit address and 8-bit data — Provides a glueless interface to boot Flash/ROM, SRAM, and peripheral devices — Up to six chip selects — 33–50 MHz operation Communications I/O subsystem — Intelligent 16-channel DMA controller — Dedicated DMA channels for receive and transmit on all subsystem peripheral interfaces — Up to two 10/100 Mbps fast Ethernet controllers (FECs), each with separate 2-Kbyte receive and transmit FIFOs — Universal serial bus (USB) version 2.0 device controller – Support for one control and six programmable endpoints — interrupt, bulk, or isochronous – 4 Kbytes of shared endpoint FIFO RAM and 1 Kbyte of endpoint descriptor RAM – Integrated physical layer interface — Up to four programmable serial controllers (PSCs) each with separate 512-byte receive and transmit FIFOs for UART, USART, modem, codec, and IrDA 1.1 interfaces — I2C peripheral interface — Two FlexCAN controller area network 2.0B controllers each with 16 message buffers — DMA serial peripheral interface (DSPI) Optional security encryption controller (SEC) module MCF548x Reference Manual, Rev. 5 1-4 Freescale Semiconductor MCF548x Family Features • • • • • • 1.4.1 — Execution units for the following: – DES/3DES block cipher – AES block cipher – RC4 stream cipher – MD5/SHA-1/SHA-256/HMAC hashing – Random number generator compliant with FIPS 140-1 standards for randomness and non-determinism — Dual-channel architecture permits single-pass encryption and authentication 32-Kbyte system SRAM — Arbitration mechanism shares bandwidth between internal bus masters (CPU, cryptography accelerator, PCI, and DMA) System integration unit (SIU) — Interrupt controller — Watchdog timer — Two 32-bit slice timers for periodic alarm and interrupt generation — Up to four 32-bit general-purpose timers with capture, compare, and PWM capability — General-purpose I/O ports multiplexed with peripheral pins Debug and test features — Core debug support via ColdFire background debug mode (BDM) port — Chip debug support via JTAG/ IEEE 1149.1 test access port PLL and clock generator — 30–66.67 MHz input frequency range Operating Voltages — 1.5V internal logic — 2.5V DDR SDRAM bus I/O (1.25V VREF) — 3.3V PCI, FlexBus, and all other I/O Estimated power consumption — <1.5W ColdFire V4e Core Overview The ColdFire V4e core is a variable-length RISC, clock-multiplied core that includes a Harvard memory architecture, branch cache acceleration logic, and limited superscalar dual-instruction issue capabilities. The limited superscalar design approaches dual-issue performance with the cost of a scalar execution pipeline. The ColdFire V4e processor core is comprised of two separate pipelines that are decoupled by an instruction buffer. The four-stage instruction fetch pipeline (IFP) prefetches the instruction stream, examines it to predict changes of flow, partially decodes instructions, and packages fetched data into instructions for the operand execution pipeline (OEP). The IFP can prefetch instructions before the OEP needs them, minimizing the wait for instructions. The instruction buffer is a 10 instruction, first-in-first-out (FIFO) buffer that decouples the IFP and OEP by holding prefetched instructions awaiting execution in the OEP. The OEP includes five pipeline stages: the first stage decodes instructions and selects operands (DS), and the second stage generates operand addresses (OAG). The third and fourth stages fetch operands (OC1 and OC2), and the fifth stage executes instructions (EX). MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 1-5 The ColdFire V4e processor contains a double-precision floating point unit (FPU). The FPU conforms to the American National Standards Institute (ANSI)/Institute of Electrical and Electronics Engineers (IEEE) Standard for Binary Floating-Point Arithmetic (ANSI/IEEE Standard 754). The FPU operates on 64-bit, double-precision floating point data and supports single-precision and signed integer input operands. The FPU programming model is like that in the MC68060 microprocessor. The FPU is intended to accelerate the performance of certain classes of embedded applications, especially those requiring high-speed floating point arithmetic computations. The ColdFire V4e processor also incorporates the ColdFire memory management unit (MMU), which provides virtual-to-physical address translation and memory access control. The MMU consists of memory-mapped control, status, and fault registers that provide access to translation lookaside buffers (TLBs). Software can control address translation and access attributes of a virtual address by configuring MMU control registers and loading TLBs. With software support, the MMU provides demand-paged, virtual addressing. The ColdFire V4e core implements the ColdFire instruction set architecture revision B with support for floating Point instructions. Additionally, the ColdFire V4e core includes the enhanced multiply-accumulate unit (EMAC) for improved signal processing capabilities. The EMAC implements a 4-stage execution pipeline, optimized for 32 x 32-bit operations, with support for four 48-bit accumulators. Supported operands include 16- and 32-bit signed and unsigned integers, as well as signed fractional operands and a complete set of instructions to process these data types. The EMAC provides superb support for execution of DSP operations within the context of a single processor at a minimal hardware cost. Refer to Chapter 3, “ColdFire Core,” for detailed information on the ColdFire V4e core architecture. 1.4.2 Debug Module (BDM) The ColdFire processor core debug interface is provided to support system debugging in conjunction with low-cost debug and emulator development tools. Through a standard debug interface, users can access real-time trace and debug information. This allows the processor and system to be debugged at full speed without the need for costly in-circuit emulators. The MCF548x debug module provides support in three different areas: • Real-time trace support: The ability to determine the dynamic execution path through an application is fundamental for debugging. The ColdFire solution implements an 8-bit parallel output bus that reports processor execution status and data to an external BDM emulator system. • Background debug mode (BDM): Provides low-level debugging in the ColdFire processor complex. In BDM, the processor complex is halted and a variety of commands can be sent to the processor to access memory and registers. The external BDM emulator uses a three-pin, serial, full-duplex channel. • Real-time debug support: BDM requires the processor to be halted, which many real-time embedded applications cannot permit. Debug interrupts let real-time systems execute a unique service routine that can quickly save key register and variable contents and return the system to normal operation without halting. External development systems can access saved data, because the hardware supports concurrent operation of the processor and BDM-initiated commands. In addition, the option is provided to allow interrupts to occur. 1.4.3 JTAG The MCF548x family supports circuit board test strategies based on the Test Technology Committee of IEEE and the Joint Test Action Group (JTAG). The test logic includes a test access port (TAP) consisting of a 16-state controller, an instruction register, and three test registers (a 1-bit bypass register, a 256-bit MCF548x Reference Manual, Rev. 5 1-6 Freescale Semiconductor MCF548x Family Features boundary-scan register, and a 32-bit ID register). The boundary scan register links the device’s pins into one shift register. Test logic, implemented using static logic design, is independent of the device system logic. The MCF548x implementation can do the following: • Perform boundary scan operations to test circuit board electrical continuity • Sample MCF548x system pins during operation and transparently shift out the result in the boundary scan register • Bypass the MCF548x for a given circuit board test by effectively reducing the boundary-scan register to a single bit • Disable the output drive to pins during circuit-board testing • Drive output pins to stable levels 1.4.4 1.4.4.1 On-Chip Memories Caches There are two independent caches associated with the ColdFire V4e core complex: a 32-Kbyte instruction cache and a 32-Kbyte data cache. Caches improve system performance by providing single-cycle access to the instruction and data pipelines. This decouples processor performance from system memory performance, increasing bus availability for on-chip DMA or external devices. 1.4.4.2 System SRAM The SRAM module provides a general-purpose 32-Kbyte memory block that the ColdFire core can access in a single cycle. The location of the memory block can be set to any 32-Kbyte address boundary within the 4-Gbyte address space. The memory is ideal for storing critical code or data structures, for use as the system stack, or for storing FEC data buffers. Because the SRAM module is physically connected to the processor's high-speed local bus, it can quickly service core-initiated accesses or memory-referencing commands from the debug module. The SRAM module is also accessible by multiple non-core bus masters, such as the DMA controller, the encryption accelerator, and the PCI Controller. 1.4.5 PLL and Chip Clocking Options MCF548x products contain an on-chip PLL capable of accepting input frequencies from 30–66.66 MHz. Table 1-2 contains the frequencies of the system buses for the members of the MCF548x family under various core/SDRAM/PCI/Flexbus clocking options. Table 1-2. MCF548x Family Clocking Options AD[12:8]1 Clock Ratio CLKIN–PCI and FlexBus Frequency Range (MHz) Internal XLB, SDRAM bus, and PSTCLK Frequency Range (MHz) Core Frequency Range (MHz) 00011 1:2 41.67–50.0 83.33–100 166.66–200 00101 1:2 25.0–41.67 50.0–83.33 100.0–166.66 01111 1:4 25.0 100 200 1 All other values of AD[12:8] are reserved. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 1-7 1.4.6 1.4.6.1 Communications I/O Subsystem DMA Controller The communications subsystem contains an intelligent DMA unit that provides front line interrupt control and data movement interface via a separate peripheral bus to the on-chip peripheral functions, leaving the processor core free to handle higher level activities. This concurrent operation enables a significant boost in overall system performance. The communications subsystem can support up to 16 simultaneously enabled DMA tasks, with support for up to two external DMA requests. It uses internal buffers to prefetch reads and post writes such that bursting is used whenever possible. This optimizes both internal and external bus activity. The following communications and computer control peripheral functions are integrated and controlled by the communications subsystem: • Up to two 10/100 Mbps fast Ethernet controllers (FECs) • Optional universal serial bus (USB) version 2.0 device controller • Up to four programmable serial controllers (PSCs) • I2C peripheral interface • DMA serial peripheral interface (DSPI) • Two FlexCAN controller area network 2.0B controllers 1.4.6.2 10/100 Fast Ethernet Controller (FEC) The FEC supports two standard MAC/PHY interfaces: 10/100 Mbps IEEE 802.3 MII and 10Mbps 7-wire interface. The controller is full duplex, supports a programmable maximum frame length and retransmission from the transmit FIFO following a collision. Support for different Ethernet physical interfaces: — 100 Mbps IEEE 802.3 MII — 10 Mbps IEEE 802.3 MII — 10 Mbps 7-wire interface • IEEE 802.3 full-duplex flow control. • Support for full-duplex operation (200 Mbps throughput) with a minimum system clock frequency of 50 MHz. • Support for half duplex operation (100 Mbps throughput) with a minimum system clock frequency of 25 MHz. • Retransmit from transmit FIFO following collision. • Internal loopback for diagnostic purposes. 1.4.6.3 USB 2.0 Device (Universal Serial Bus) The USB module implementation on the MCF548x product family provides all the logic necessary to process the USB protocol as defined by version 2.0 specification for peripheral devices. It features the following: • High-speed operation up to 480 Mbps, full-speed operation at 12 Mbps, and low-speed operation at 1.5 Mbps • Physical interface on chip • Bulk, interrupt, and isochronous transport modes. • Six programmable in/out endpoints and one control endpoint MCF548x Reference Manual, Rev. 5 1-8 Freescale Semiconductor MCF548x Family Features • 4 Kbytes of shared endpoint FIFO RAM and 1 Kbyte of endpoint descriptor RAM 1.4.6.4 Programmable Serial Controllers (PSCs) The MCF548x product family supports four PSCs that can be independently configured to operate in the following modes: • Universal asynchronous receiver transmitter (UART) mode — 5,6,7,8 bits of data plus parity — Odd, even, none, or force parity — Stop bit width programmable in 1/16 bit increments — Parity, framing, and overrun error detection — Automatic PSCCTS and PSCRTS modem control signals • IrDA 1.0 SIR mode (SIR) — Baud rate range of 2400–115200 bps — Selectable pulse width: either 3/16 of the bit duration or 1.6 μs • IrDA 1.1 MIR mode (MIR) — Baud rate of 0.576 or 1.152 Mbps • IrDA 1.1 FIR mode (FIR) — Baud rate of 4.0 Mbps • 8-bit soft modem mode (modem8) • 16-bit soft modem mode (modem16) • AC97 soft modem mode (AC97) Each PSC supports synchronous (USART) and asynchronous (UART) protocols. The PSCs can be used to interface to external full-function modems or external codecs for soft modem support, as well as IrDA 1.1 or 1.0 interfaces. Both 8- and 16-bit data widths are supported. PSCs can be configured to support a 1200-baud plain old telephone system (POTS) modem, V.34 or V.90 protocols. The standard UART interface supports connection to an external terminal/computer for debug support. 1.4.6.5 I2C (Inter-Integrated Circuit) The MCF548x product family provides an I2C two-wire, bidirectional serial bus for on-board communication. It features the following: • Multimaster operation with arbitration and collision detection • Calling address recognition and interrupt generation • Automatic switching from master to slave on arbitration loss • Software-selectable acknowledge bit • Start and stop signal generation and detection • Bus busy status detection 1.4.6.6 DMA Serial Peripheral Interface (DSPI) The DSPI block operates as a basic SPI block with FIFOs providing support for external queue operation. Data to be transmitted and data received reside in separate FIFOs. The FIFOs can be popped and pushed by host software or by the system DMA controller. The DSPI supports these SPI features: • Full-duplex, three-wire synchronous transfers • Master and slave mode—two peripheral chip selects in master mode MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 1-9 • DMA support 1.4.6.7 Controller Area Network (CAN) The FlexCAN modules are communication controllers implementing the CAN protocol. The CAN protocol can be used as an industrial control serial data bus, meeting the specific requirements of real-time processing and reliable operation in a harsh EMI environment, while maintaining cost-effectiveness. Each of the two CAN controllers on the MCF548x family products contains sixteen message buffers. The two CAN controllers can interface to two separate 16 message buffer CAN networks or a single 32 message buffer CAN network. 1.4.7 DDR SDRAM Memory Controller The DDR SDRAM memory controller is a glueless interface to DDR memories. The module uses a 32-bit memory port and can address a maximum of 1 Gbyte of data with 16 64M x 8 (512-Mbit) devices, four per chip select. The controller supplies two clock lines and respective inverted clock lines to help minimize system complexity when using DDR. The module supports either DDR or SDR, but not both. This is due to voltage differences between the memory technologies. The supported memory clock rate is up to 100 MHz. At this memory clock rate, DDR memory can receive data at an effective rate of up to 200 MHz. • Support for up to 13 lines of row address, 11 lines of column address, two lines of bank address, and up to four chip selects • Memory bus width fixed at 32 bits • Four chip selects support up to 1 GByte of SDRAM memory • Support for page mode to maximize the data rate. Page mode remembers active pages for all four chip selects • Support for sleep mode and self refresh • Cache line reads that can use critical word first. These reads can start in the center of a burst and will wrap to the beginning. This allows the processor quicker access to a needed instruction. All on-chip bus masters have access to DRAM. This includes PCI, the ColdFire V4e core, the cryptography accelerator, and the DMA controller. 1.4.8 Peripheral Component Interconnect (PCI) The PCI controller is a PCI V2.2-compliant bus controller and arbiter. The PCI bus is capable of 50-MHz operation with a 32-bit address/data bus and support for five external masters. The PCI module includes an inbound FIFO to increase performance when using an external bus master. The bus can address all 4 Gbytes of PCI-addressable space. The PCI bus is also multiplexed with the flexible local bus (FlexBus) address lines. If 32-bit non-muxed local address and data is required, it can be obtained at the expense of utilizing the PCI bus. When implemented, the PCI controller acts as the central resource, bus arbiter, and configuring master on the PCI bus. 1.4.9 Flexible Local Bus (FlexBus) The FlexBus module is intended to provide the user with basic functionality required to interface to peripheral devices. The FlexBus interface is a multiplexed or non-multiplexed bus, with an operating MCF548x Reference Manual, Rev. 5 1-10 Freescale Semiconductor MCF548x Family Features frequency from 33–50 MHz. The Flexbus is targeted to support external Flash memories, boot ROMs, gate-array logic, or other simple target interfaces. Up to six chip selects are supported by the FlexBus. Possible combinations of address and data bits are the following: • Non-multiplexed 32-bit address and 32-bit data (32-bit address muxed over PCI bus–PCI not usable) • Multiplexed 32-bit address and 32-bit data (PCI usable) • Multiplexed 32-bit address and 16-bit data • Multiplexed 32-bit address and 8-bit data The non-multiplexed 32-bit address and 32-bit data mode is determined at chip reset. For all other modes, the full 32-bit address is driven during the address phase. The number of bytes used for data are determined on a chip select by chip select basis. 1.4.10 Security Encryption Controller (SEC) As consumers and businesses continue to embrace the Internet, the need for secure point-to-point communications across what is an entirely insecure network has been met by the development of a range of standard protocols. Computer cryptography fundamentally involves calculations with very large numbers. Personal computers have sufficient processing power to implement these algorithms entirely in software. When placed upon the embedded devices typically used for routing and remote access functions, this same computational burden can potentially decrease the throughput of a 100 Mbps Ethernet interface down to 10 Mbps. Hardware acceleration of common cryptography algorithms is the solution to the computational bandwidth requirements of Internet security standards. Discrete solutions currently address this problem, but the next logical step is to integrate a cryptography accelerator on an embedded processor, such as the MCF548x family. Freescale has developed the SEC on the MCF548x family for this purpose. This block accelerates the core cryptography algorithms that underlie standard Internet security protocols like SSL/TLS, IPSec, IKE, and WTLS/WAP. • The SEC includes execution units for the following: — DES/3DES block cipher — AES block cipher — RC4 stream cipher — MD5/SHA-1/SHA-256/HMAC hashing — Random number generator compliant with FIPS 140-1 standards for randomness and non-determinism • Dual-channel architecture permits single-pass encryption and authentication 1.4.11 1.4.11.1 System Integration Unit (SIU) Timers The MCF548x family integrates several timer functions required by most embedded systems. Two internal 32-bit slice timers create short cycle periodic interrupts, typically utilized for RTOS scheduling and alarm functionality. A watchdog timer resets the processor if not regularly serviced, catching software hang-ups. Four 32-bit general purpose timers can perform input capture, output compare, and PWM functionality. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 1-11 1.4.11.2 Interrupt Controller The interrupt controller on the MCF548x family can support up to 63 interrupt sources. The interrupt controller is organized as seven levels with nine interrupt sources per level. Each interrupt source has a unique interrupt vector, and 56 of the 63 sources of a given controller provide a programmable level [1-7] and priority within the level. • Support for up to 63 interrupt sources organized as follows: — 56 fully-programmable interrupt sources — 7 fixed-level interrupt sources • Seven external interrupt signals • Unique vector number for each interrupt source • Ability to mask any individual interrupt source or all interrupt sources (global mask-all) • Support for hardware and software interrupt acknowledge (IACK) cycles • Combinatorial path to provide wake-up from stop mode 1.4.11.3 General Purpose I/O All peripheral I/O pins on the MCF548x family are multiplexed with GPIO, adding flexibility and usability to all signals on the chip. MCF548x Reference Manual, Rev. 5 1-12 Freescale Semiconductor Chapter 2 Signal Descriptions 2.1 Introduction This chapter describes the MCF548x signals. NOTE The terms ‘assertion’ and ‘negation’ are used to avoid confusion when dealing with a mixture of active-low and active-high signals. The term ‘asserted’ indicates that a signal is active, independent of the voltage level. The term ‘negated’ indicates that a signal is inactive. Active-low signals, such as RAS and TA, are indicated with an overbar. 2.1.1 Block Diagram Figure 2-1 displays the signals of the MCF548x. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 2-1 FlexBus AD[31:24] AD[23:16] AD15:8] AD[7:0] FBCS[5:1] / PFBCS[5:1] FBCS0 ALE / PFBCTL0 / TBST R/W / PFBCTL2 / TBST BE/BWE3 / PFBCTL7 / TSIZ1 BE/BWE2 / PFBCTL6 / TSIZ0 BE/BWE1 / PFBCTL5 / FBADDR1 BE/BWE0 / PFBCTL4 / FBADDR0 OE / PFBCTL3 TA / PFBCTL1 SDRAM Controller SDDATA[31:24] SDDATA[23:16] SDDATA[15:8] SDDATA[7:0] SDADDR[12:0] SDBA[1:0] RAS CAS SDCS[3:0] SDDM[3:0] SDDQS[3:0] SDCLK[1:0] SDCLK[1:0] SDWE SDCKE SDRDQS VREF PCI Controller PSCs DSPI PCIAD[31:24] / FBADDR[31:24] PCIAD[23:16] / FBADDR[23:16] PCIAD[15:8] / FBADDR[15:8] PCIAD[7:0] / FBADDR[7:0] PCICXBE[3:0] PCIDEVSEL PCIFRM PCIIDSEL PCIIRDY PCIPAR PCIPERR PCIRESET PCISERR PCISTOP PCITRDY PCIBG4 / PPCIBG4 / TBST PCIBG[3:0] / PPCIBG[3:0] / TOUT[3:0] PCIBR4 / PPCIBR4 / IRQ4 PCIBR[3:0] / PPCIBR[3:0] / TIN[3:0] MCF548x PCS0TXD / PPSCL0 PSC0RXD / PPSCL1 PSC0CTS / PPSCL2 / PSC0BCLK PSC0RTS / PPSCL3 / PSC0FSYNC PSC1TXD / PPSCL4 PSC1RXD / PPSCL5 PSC1CTS / PPSCL6 / PSC1BCLK PSC1RTS / PPSCL7 / PSC1FSYNC PSC2TXD / PPSCH0 PSC2RXD / PPSCH1 PSC2CTS / PPSCH2 / PSC2BCLK / CANRX0 PSC2RTS / PPSCH3 / PSC2FSYNC / CANTX0 PSC3TXD / PPSCH4 PSC3RXD / PPSCH5 PSC3CTS / PPSCH6 / PSC3BCLK PSC3RTS / PPSCH7 / PSC3FSYNC DSPISOUT / PDSPI0 / PSC3TXD DSPISIN / PDSPI1 / PSC3RXD DSPISCK / PDSPI2 / PSC3CTS / PSC3BCLK DSPICS5/PCSS / PDSPI6 DSPICS3 / PDSPI5 / TOUT3 / CANTX1 DSPICS2 / PDSPI4 / TOUT2 / CANTX1 DSPICS0/SS / PDSPI3 / PSC3RTS / PSC3FSYNC E0MDIO / PFECI2C3 E0MDC / PFECI2C2 E0TXCLK / PFEC0H7 E0TXEN / PFEC0H6 E0TXD0 / PFEC0H5 E0COL / PFEC0H4 E0RXCLK / PFEC0H3 E0RXDV / PFEC0H2 E0RXD0 / PFEC0H1 E0CRS / PFEC0H0 E0TXD[3:1] / PFEC0L[7:5] E0TXER / PFEC0L4 E0RXD[3:1] / PFEC0L[3:1] E0RXER / PFEC0L0 Ethernet MAC 0 E1MDIO / SDA / CANRX0 E1MDC / SCL / CANTX0 E1TXCLK / PFEC1H7 E1TXEN / PFEC1H6 E1TXD0 / PFEC1H5 E1COL / PFEC1H4 E1RXCLK / PFEC1H3 E1RXDV / PFEC1H2 E1RXD0 / PFEC1H1 E1CRS / PFEC1H0 E1TXD[3:1] / PFEC1L[7:5] E1TXER / PFEC1L4 E1RXD[3:1] / PFEC1L[3:1] E1RXER / PFEC1L0 Ethernet MAC 1 USBD+ USBD– USBVBUS USBRBIAS USBCLKIN USBCLKOUT USB SDA / PFECI2C1 SCL / PFECI2C0 I2C IRQ7 / PIRQ7 IRQ[6:5] / PIRQ[6:5] / CANRX1 External Interrupts Port DREQ1 / PDMA1 / TIN1 / IRQ1 DREQ0 / PDMA0 / TIN0 DACK[1:0] / PDMA[3:2] / TOUT[1:0] DMA Controller TIN3 / PTIM7 / IRQ3 / CANRX1 TOUT3 / PTIM6 / CANTX1 TIN2 / PTIM5 / IRQ2 / CANRX1 TOUT2 / PTIM4 / CANTX1 TIN1 TOUT1 TIN0 TOUT0 Timer Module PSTCLK PSTDDATA[7:0] DSCLK / TRST BKPT / TMS DSI / TDI DSO / TDO TCK Debug & JTAG Test Port Control MTMOD[3:0] RSTI RSTO CLKIN Test / Reset & Clock EVDD IVDD VSS SDVDD PLLVDD PLLVSS USB_OSCVDD USB_PHYVDD USB_OSCAVDD USB_PLLVDD USBVDD Power Supplies Figure 2-1. MCF548x Signals MCF548x Reference Manual, Rev. 5 2-2 Freescale Semiconductor Introduction Table 2-1 lists the signals for the MCF548x in functional group order. Drive Reset State Pin Functions Pull-up Table 2-1. MCF548x Signal Description AE2, AF3, AF1, AE3, AE4, AD5, AF2, AD4 AD[31:24] — — — Multiplexed address/data bus I/O 16 Hi-Z AD3, AC3, AD2, AC2, AA4, AE1, AC1, AD1 AD[23:16] — — — Multiplexed address/data bus I/O 16 Hi-Z AB2, AA3, W4, AB1, AA2, AA1, Y1, Y2 AD[15:8] — — — Multiplexed address/data bus I/O 16 Hi-Z W3, W1, W2, V3, V1, V2, T4, U3 AD[7:0] — — — Multiplexed address/data bus I/O 16 Hi-Z R1, T2, T3, T1, U2 FBCS[5:1] PFBCS[5:1] — — Chip selects 5–1 O:I/O 24 High U1 FBCS0 — — — Chip select 0 O 24 High AD6 ALE PFBCTL0 TBST — 16 High AE5 R/W PFBCTL2 TBST — Read/write O:I/O 16 Hi-Z AF4 BE/BWE3 PFBCTL7 TSIZ1 — Byte enables O:I/O 16 High AF5 BE/BWE2 PFBCTL6 TSIZ0 — Byte enables O:I/O 16 High AC4 BE/BWE1 PFBCTL5 FBADDR1 — Byte enables O:I/O 16 High AE7 BE/BWE0 PFBCTL4 FBADDR0 — Byte enables O:I/O 16 High AE6 OE PFBCTL3 — — Output enable O:I/O 16 High AF6 TA PFBCTL1 — — Transfer acknowledge I:I/O 16 — PBGA Pin Primary GPIO Secondary Description I/O Tertiary FlexBus Address Latch Enable O:I/O SDRAM Controller C10, B9, A8, D5, A6, C8, B7, A5 SDDATA[31:24] — — — SDRAM data bus I/O 24 Hi-Z A4, C7, B6, B4, C5, B3, C4, D4 SDDATA[23:16] — — — SDRAM data bus I/O 24 Hi-Z E2, D1, G4, E1, K4, F1, G2, H3 SDDATA[15:8] — — — SDRAM data bus I/O 24 Hi-Z N4, G1, H2, J3, J1, M4, K3, K2 SDDATA[7:0] — — — SDRAM data bus I/O 24 Hi-Z A13, A12, D10, B12, C12, A11, D8, B11, C11, A10, D7, B10, A9 SDADDR[12:0] — — — SDRAM address bus O 24 Low MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 2-3 Description I/O Drive Reset State Pin Functions Pull-up Table 2-1. MCF548x Signal Description (Continued) — SDRAM bank addresses O 24 Low — — SDRAM row address strobe O 24 High — — — SDRAM column address strobe O 24 High SDCS[3:0] — — — SDRAM chip selects O 24 High B8, A3, G3, J2 SDDM[3:0] — — — SDRAM write data byte mask O 24 High A7, B5, F2, H1 SDDQS[3:0] — — — SDRAM data strobe I/O 24 High L1, N1 SDCLK[1:0] — — — SDRAM clock O 24 Low M1, N2 SDCLK[1:0] — — — Inverted SDRAM clock O 24 Low K1 SDWE — — — SDRAM write enable O 24 Low E4 SDCKE — — — SDRAM clock enable O 24 Low L2 SDRDQS — — — SDR SDRAM data strobe O 24 Low D2 VREF — — — SDRAM reference voltage I — — PBGA Pin Primary GPIO Secondary Tertiary M2, M3 SDBA[1:0] — — E3 RAS — C2 CAS R2, P2, P1, N3 PCI Controller V25, V26, U25, U26, T24, T25, T26, R24 PCIAD[31:24] — FBADDR[31:24] — PCI address/data bus I/O 16 Hi-Z R25, R26, P26, P24, P23, P25, N25, N23 PCIAD[23:16] — FBADDR[23:16] — PCI address/data bus I/O 16 Hi-Z N26, N24, M26, M25, L26, L25, K26, K25 PCIAD[15:8] — FBADDR[15:8] — PCI address/data bus I/O 16 Hi-Z J26, K24, J25, H26, J24, G26, H25, K23 PCIAD[7:0] — FBADDR[7:0] — PCI address/data bus I/O 16 Hi-Z F26, G25, E26, G24 PCICXBE[3:0] — — — PCI command/byte enables I/O 16 Hi-Z J23 PCIDEVSEL — — — PCI device select I/O 16 Hi-Z F25 PCIFRM — — — PCI frame I/O 16 Hi-Z C23 PCIIDSEL — — — PCI initialization device select I — — MCF548x Reference Manual, Rev. 5 2-4 Freescale Semiconductor Introduction Description I/O Drive Reset State Pin Functions Pull-up Table 2-1. MCF548x Signal Description (Continued) — PCI initiator ready I/O 16 Hi-Z — — PCI parity I/O 16 Hi-Z — — — PCI parity error I/O 16 Hi-Z PCIRESET — — — PCI reset O 16 Low F24 PCISERR — — — PCI system error I/O 16 Hi-Z E25 PCISTOP — — — PCI stop I/O 16 Hi-Z C26 PCITRDY — — — PCI target ready I/O 16 Hi-Z W24 PCIBG4 PPCIBG4 TBST — PCI external grant 4 O:I/O 16 GPI Y26, W25, V24, W26 PCIBG[3:0] PPCIBG[3:0] TOUT[3:0] — PCI external grant 3–0 O:I/O 16 GPI D21 PCIBR4 PPCIBR4 IRQ4 — PCI external request 4 I:I/O Y1 8 GPI B24 PCIBR3 PPCIBR3 TIN3 — PCI external request 3 I:I/O Y1 8 GPI A25, B23, A24 PCIBR[2:0] PPCIBR[2:0] TIN[2:0] — PCI external request 2–0 I:I/O 8 GPI PBGA Pin Primary GPIO Secondary Tertiary D24 PCIIRDY — — F23 PCIPAR — D26 PCIPERR G23 External Interrupts Port D14 IRQ7 PIRQ7 — — External interrupt request 7 I:I/O — — B14, A14 IRQ[6:5] PIRQ[6:5] CANRX1 — External interrupt request 6–5 I:I/O — — I/O 8 GPI 8 GPI I:I/O 8 GPI MAC transmit enable O:I/O 8 GPI Ethernet MAC 0 AF10 E0MDIO PFECI2C3 — — Management channel serial data AD11 E0MDC PFECI2C2 — — Management channel O:I/O clock AF9 E0TXCLK PFEC0H7 — — AE10 E0TXEN PFEC0H6 — — AD9 E0TXD0 PFEC0H5 — — MAC transmit data O:I/O 8 GPI AC9 E0COL PFEC0H4 — — MAC collision I:I/O 8 GPI AD14 E0RXCLK PFEC0H3 — — MAC receive clock I:I/O 8 GPI AE14 E0RXDV PFEC0H2 — — MAC receive enable I:I/O 8 GPI AD13 E0RXD0 PFEC0H1 — — MAC receive data I:I/O 8 GPI AE19 E0CRS PFEC0H0 — — MAC carrier sense I:I/O 8 GPI MAC transmit clock MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 2-5 Description I/O Drive Reset State Pin Functions Pull-up Table 2-1. MCF548x Signal Description (Continued) — MAC transmit data O:I/O 8 GPI — — MAC transmit error O:I/O 8 GPI PFEC0L[3:1] — — MAC receive data I:I/O 8 GPI PFEC0L0 — — MAC receive error I:I/O 8 GPI PBGA Pin Primary GPIO Secondary Tertiary AD8, AC6, AF7 E0TXD[3:1] PFEC0L[7:5] — AE9 E0TXER PFEC0L4 AF11, AF12, AF13 E0RXD[3:1] AC14 E0RXER Ethernet MAC 1 AE252 E1MDIO — SDA CANRX0 Management channel serial data I/O 8 — AD242 E1MDC — SCL CANTX0 Management channel clock O 8 — AE132 E1TXCLK PFEC1H7 — — MAC Transmit clock I:I/O Y1 8 GPI MAC Transmit enable O:I/O Y1 8 GPI O:I/O Y1 8 GPI 8 GPI AD252 E1TXEN PFEC1H6 — — AE122 E1TXD0 PFEC1H5 — — MAC Transmit data AF82 E1COL PFEC1H4 — — MAC Collision I:I/O Y1 B222 E1RXCLK PFEC1H3 — — MAC Receive clock I:I/O Y1 8 GPI 8 GPI B252 E1RXDV PFEC1H2 — — MAC Receive enable I:I/O Y1 AF242 E1RXD0 PFEC1H1 — — MAC Receive data I:I/O Y1 8 GPI 8 GPI AC52 E1CRS PFEC1H0 — — MAC Carrier sense I:I/O Y1 AC82, AC112, AE112 E1TXD[3:1] PFEC1L[7:5] — — MAC Transmit data O:I/O Y1 8 GPI AE242 E1TXER PFEC1L4 — — MAC Transmit error O:I/O Y1 8 GPI 8 GPI 8 GPI D252, B262, AE82 A262 E1RXD[3:1] PFEC1L[3:1] — — MAC Receive data I:I/O Y1 E1RXER PFEC1L0 — — MAC Receive error I:I/O Y1 USB AF163 USBD+ — — — USB differential data I/O 24 — AF173 USBD- — — — USB differential data I/O 24 — AC173 USBVBUS — — — USB Vbus monitor input I — — AF18 USBRBIAS — — — USB bias resistor I — — AF153 USBCLKIN — — — USB crystal input I — — USBCLKOUT — — — USB crystal output O 24 — — QSPI data out O:I/O 24 GPI 3 AF14 DSPI Y24 DSPISOUT PDSPI0 PSC3TXD MCF548x Reference Manual, Rev. 5 2-6 Freescale Semiconductor Introduction Description I/O Drive Reset State Pin Functions Pull-up Table 2-1. MCF548x Signal Description (Continued) — QSPI data in I:I/O 24 GPI PSC3CTS PSC3BCLK QSPI clock I/O 24 GPI PDSPI6 — — QSPI chip select O:I/O 24 GPI DSPICS3 PDSPI5 TOUT3 CANTX1 QSPI chip select O:I/O 24 GPI AA26 DSPICS2 PDSPI4 TOUT2 CANTX1 QSPI chip select O:I/O 24 GPI Y25 DSPICS0/SS PDSPI3 PSC3RTS PSC3FSYNC QSPI chip select O:I/O 24 GPI PBGA Pin Primary GPIO Secondary Tertiary AC24 DSPISIN PDSPI1 PSC3RXD AD22 DSPISCK PDSPI2 W23 DSPICS5/PCSS V23 I2C C24 SDA PFECI2C1 — — I2C Serial data I/O 8 GPI C25 SCL PFECI2C0 — — I2C Serial clock I/O 8 GPI PSCs AA25 PSC0TXD PPSC1PSC00 — — PSC0 transmit data O:I/O 8 GPI AC21 PSC0RXD PPSC1PSC01 — — PSC0 receive data I:I/O 8 GPI AE23 PSC0CTS PPSC1PSC03 PSC0BCLK — PSC0 clear to send I:I/O 8 GPI AB26 PSC0RTS PPSC1PSC02 PSC0FSYNC — PSC0 request to send I/O 8 GPI AB25 PSC1TXD PPSC1PSC04 — — PSC1 transmit data O:I/O 8 GPI AE22 PSC1RXD PPSC1PSC05 — — PSC1 receive data I:I/O 8 GPI AF25 PSC1CTS PPSC1PSC07 PSC1BCLK — PSC1 clear to send I:I/O 8 GPI Y23 PSC1RTS PPSC1PSC06 PSC1FSYNC — PSC1 request to send I/O 8 GPI AC26 PSC2TXD PPSC3PSC20 — — PSC2 transmit data O:I/O 8 GPI AD21 PSC2RXD PPSC3PSC21 — — PSC2 receive data I:I/O 8 GPI AC19 PSC2CTS PPSC3PSC23 PSC2BCLK CANRX0 PSC2 clear to send I:I/O 8 GPI AD26 PSC2RTS PPSC3PSC22 PSC2FSYNC CANTX0 PSC2 request to send I/O 8 GPI AE26 PSC3TXD PPSC3PSC24 — — PSC3 transmit data O:I/O 8 GPI AE21 PSC3RXD PPSC3PSC25 — — PSC3 receive data I:I/O 8 GPI AF23 PSC3CTS PPSC3PSC27 PSC3BCLK — PSC3 clear to send I:I/O 8 GPI AB23 PSC3RTS PPSC3PSC26 PSC3FSYNC — PSC3 request to send I/O 8 GPI DMA Controller AF19 DREQ1 PDMA1 TIN1 IRQ1 DMA request I:I/O 8 GPI AF20 DREQ0 PDMA0 TIN0 — DMA request I:I/O 8 GPI AC25, AB24 DACK[1:0] PDMA[3:2] TOUT[1:0] — DMA acknowledge O:I/O 8 GPI MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 2-7 Drive Reset State Pin Functions Pull-up Table 2-1. MCF548x Signal Description (Continued) AD19 TIN3 PTIM7 IRQ3 CANRX1 Timer input I:I/O 8 GPI AD23 TOUT3 PTIM6 CANTX1 — Timer output O:I/O 8 GPI AF21 TIN2 PTIM5 IRQ2 CANRX1 Timer input I:I/O 8 GPI AC22 TOUT2 PTIM4 CANTX1 — Timer output O:I/O 8 GPI AE20 TIN1 — — — Timer input I 8 GPI AC23 TOUT1 — — — Timer output O 8 GPI AF22 TIN0 — — — Timer input I 8 GPI AF26 TOUT0 — — — Timer output O 8 GPI PBGA Pin Primary GPIO Secondary Description I/O Tertiary Timer Module Debug and JTAG Test Port Control D20 PSTCLK — — — Processor clock output O 8 High A23, B21, D18, C20, A22, B20, A21, B19 PSTDDATA[7:0] — — — Processor status debug data O 8 High C15 DSCLK — TRST — Debug clock / TAP reset I Y — — B15 BKPT — TMS — Breakpoint/TAP test mode select I Y — — A15 DSI — TDI — Debug data in / TAP data in I Y — — D17 DSO — TDO — Debug data out / TAP data out O 8 High A16 TCK — — — TAP clock I — — Test, Reset, and Clock B17, C14, A18, B16 MTMOD[3:0] — — — Test mode pins I — — B13 RSTI — — — Reset input I — — A20 RSTO — — — Reset output O 8 Low A17 CLKIN — — — Clock input I — — D15 NC — — — No Connect I — — AC15 NC — — — No Connect I — — MCF548x Reference Manual, Rev. 5 2-8 Freescale Semiconductor Introduction Drive Reset State Pin Functions Pull-up Table 2-1. MCF548x Signal Description (Continued) C16, C22, E24, H24, M24, R3, U24, Y3, AA24, AB3, AD7, AD10, AD18 EVDD — — — Positive I/O supply I — — C18, D11, D12, D19, D22, H4, H23, L23, P4, R23, V4, AA23, AC12, AC20 IVDD — — — Positive core supply I — — A2, B2, C3, C17, C19, C21, D6, D9, D13, D16, D23, E23, F4, J4, L4, L11–L16, L24, M11–M16, M23, N11–N16, P11–P16, R4, R11–R16, T11–T16, T23, U4, U23, Y4, AB4, AC7, AC10, AC18, AD12, AD17, AD20, AE15–AE17 VSS — — A1, B1, C1, C6, C9, C13, D3, F3, L3, P3 SDVDD — — — Positive SDRAM supply A19 PLLVDD — — — Positive PLL analog supply B18 PLLVSS — — — PLL ground AC134 USB_OSCVDD — — — USB oscillator supply AC164 USB_PHYVDD — — — USB PHY supply USB_OSCAVDD — — — USB oscillator analog supply AD164 USB_PLLVDD — — — USB PLL supply AE184 USBVDD — — — USB supply PBGA Pin Primary GPIO Secondary Description I/O Tertiary Power Supplies AD15 4 Ground 1 Pull-up resistor when configured for general purpose input (default state after reset). This pin is a “no connect” on the MCF5483 and MCF5482 devices. 3 This pin is a “no connect” on the MCF5481 and MCF5480 devices. 4 This pin is a “no connect” on the MCF5481 and MCF5480 devices. On MCF5485, MCF5484, MCF5483, and MCF5482 device the pin should be connected to the appriopriate power rail even is USB is not being used. 2 MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 2-9 Table 2-2 lists the MCF548x signals in pin number order for the 388 PBGA package. Pin Functions Primary GPIO Secondary Tertiary A1 SDVDD — — — A2 VSS — — A3 SDDM2 — A4 SDDATA23 A5 PBGA Pin PBGA Pin Table 2-2. MCF5485/MCF5484 Signal Description by Pin Number Pin Functions Primary GPIO Secondary Tertiary P1 SDCS1 — — — — P2 SDCS2 — — — — — P3 SDVDD — — — — — — P4 IVDD — — — SDDATA24 — — — P11 VSS — — — A6 SDDATA27 — — — P12 VSS — — — A7 SDDQS3 — — — P13 VSS — — — A8 SDDATA29 — — — P14 VSS — — — A9 SDADDR0 — — — P15 VSS — — — A10 SDADDR3 — — — P16 VSS — — — A11 SDADDR7 — — — P23 PCIAD19 — FBADDR19 — A12 SDADDR11 — — — P24 PCIAD20 — FBADDR20 — A13 SDADDR12 — — — P25 PCIAD18 — FBADDR18 — A14 IRQ5 PIRQ5 CANRX1 — P26 PCIAD21 — FBADDR21 — A15 DSI — TDI — R1 FBCS5 PFBCS5 — — A16 TCK — — — R2 SDCS3 — — — A17 CLKIN — — — R3 EVDD — — — A18 MTMOD1 — — — R4 VSS — — — A19 PLLVDD — — — R11 VSS — — — A20 RSTO — — — R12 VSS — — — A21 PSTDDATA1 — — — R13 VSS — — — A22 PSTDDATA3 — — — R14 VSS — — — A23 PSTDDATA7 — — — R15 VSS — — — A24 PCIBR0 PPCIBR0 TIN0 — R16 VSS — — — A25 PCIBR2 PPCIBR2 TIN2 — R23 IVDD — — — A261 E1RXD1 PFEC1L5 — — R24 PCIAD24 — FBADDR24 — B1 SDVDD — — — R25 PCIAD23 — FBADDR23 — B2 VSS — — — R26 PCIAD22 — FBADDR22 — B3 SDDATA18 — — — T1 FBCS2 PFBCS2 — — B4 SDDATA20 — — — T2 FBCS4 PFBCS4 — — MCF548x Reference Manual, Rev. 5 2-10 Freescale Semiconductor Introduction Pin Functions Primary GPIO Secondary Tertiary B5 SDDQS2 — — — B6 SDDATA21 — — B7 SDDATA25 — B8 SDDM3 B9 PBGA Pin PBGA Pin Table 2-2. MCF5485/MCF5484 Signal Description by Pin Number (Continued) Pin Functions Primary GPIO Secondary Tertiary T3 FBCS3 PFBCS3 — — — T4 AD1 — — — — — T11 VSS — — — — — — T12 VSS — — — SDDATA30 — — — T13 VSS — — — B10 SDADDR1 — — — T14 VSS — — — B11 SDADDR5 — — — T15 VSS — — — B12 SDADDR9 — — — T16 VSS — — — B13 RSTI — — — T23 VSS — — — B14 IRQ6 PIRQ6 CANRX1 — T24 PCIAD27 — FBADDR27 — B15 BKPT — TMS — T25 PCIAD26 — FBADDR26 — B16 MTMOD0 — — — T26 PCIAD25 — FBADDR25 — B17 MTMOD3 — — — U1 FBCS0 — — — B18 PLLVSS — — — U2 FBCS1 PFBCS1 — — B19 PSTDDATA0 — — — U3 AD0 — — — B20 PSTDDATA2 — — — U4 VSS — — — B21 PSTDDATA6 — — — U23 VSS — — — B221 E1RXCLK PFEC1H3 — — U24 EVDD — — — B23 PCIBR1 PPCIBR1 TIN1 — U25 PCIAD29 — FBADDR29 — B24 PCIBR3 PPCIBR3 TIN3 — U26 PCIAD28 — FBADDR28 — B251 E1RXDV PFEC1H2 — — V1 AD3 — — — B261 E1RXD2 PFEC1L2 — — V2 AD2 — — — C1 SDVDD — — — V3 AD4 — — — C2 CAS — — — V4 IVDD — — — C3 VSS — — — V23 DSPICS3 PDSPI5 TOUT3 CANTX1 C4 SDDATA17 — — — V24 PCIBG1 PPCIBG1 TOUT1 — C5 SDDATA19 — — — V25 PCIAD31 — FBADDR31 — C6 SDVDD — — — V26 PCIAD30 — FBADDR30 — C7 SDDATA22 — — — W1 AD6 — — — C8 SDDATA26 — — — W2 AD5 — — — C9 SDVDD — — — W3 AD7 — — — MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 2-11 Pin Functions Primary GPIO Secondary Tertiary C10 SDDATA31 — — — C11 SDADDR4 — — C12 SDADDR8 — C13 SDVDD C14 PBGA Pin PBGA Pin Table 2-2. MCF5485/MCF5484 Signal Description by Pin Number (Continued) Pin Functions Primary GPIO Secondary Tertiary W4 AD13 — — — — W23 DSPICS5/PCSS PDSPI6 — — — — W24 PCIBG4 PPCIBG4 TBST — — — — W25 PCIBG2 PPCIBG2 TOUT2 — MTMOD2 — — — W26 PCIBG0 PPCIBG0 TOUT0 — C15 DSCLK — TRST — Y1 AD9 — — — C16 EVDD — — — Y2 AD8 — — — C17 VSS — — — Y3 EVDD — — — C18 IVDD — — — Y4 VSS — — — C19 VSS — — — Y23 PSC1RTS C20 PSTDDATA4 — — — Y24 DSPISOUT PDSPI0 PSC3TXD — C21 VSS — — — Y25 DSPICS0/SS PDSPI3 — — C22 EVDD — — — Y26 PCIBG3 PPCIBG3 TOUT3 — C23 PCIIDSEL — — — AA1 AD10 — — — C24 SDA PFECI2C1 — — AA2 AD11 — — — C25 SCL PFECI2C0 — — AA3 AD14 — — — C26 PCITRDY — — — AA4 AD19 — — — D1 SDDATA14 — — — AA23 IVDD — — — D2 VREF — — — AA24 EVDD — — — D3 SDVDD — — — AA25 PCS0TXD PPSC1PSC00 — — D4 SDDATA16 — — — AA26 DSPICS2 PDSPI4 TOUT2 CANTX1 D5 SDDATA28 — — — AB1 AD12 — — — D6 VSS — — — AB2 AD15 — — — D7 SDADDR2 — — — AB3 EVDD — — — D8 SDADDR6 — — — AB4 VSS — — — D9 VSS — — — AB23 PSC3RTS D10 SDADDR10 — — — AB24 DACK0 PDMA2 TOUT0 — D11 IVDD — — — AB25 PSC1TXD PPSC1PSC04 — — D12 IVDD — — — AB26 PSC0RTS PPSC1PSC02 PSC0FSYNC D13 VSS — — — AC1 AD17 — — — D14 IRQ7 PIRQ7 — — AC2 AD20 — — — PPSC1PSC06 PSC1FSYNC PPSC3PSC26 PSC3FSYNC — — — MCF548x Reference Manual, Rev. 5 2-12 Freescale Semiconductor Introduction Pin Functions Primary GPIO Secondary Tertiary D15 NC — — — D16 VSS — — PBGA Pin PBGA Pin Table 2-2. MCF5485/MCF5484 Signal Description by Pin Number (Continued) Pin Functions Primary GPIO Secondary Tertiary AC3 AD22 — — — — AC4 BE/BWE1 PFBCTL5 FBADDR1 — 1 E1CRS PFEC1H0 — — D17 DSO — TDO — AC5 D18 PSTDDATA5 — — — AC6 E0TXD2 PFEC0L6 — — D19 IVDD — — — AC7 VSS — — — D20 PSTCLK — — — AC81 E1TXD3 PFEC1L7 — — D21 PCIBR4 PPCIBR4 IRQ4 — AC9 E0COL PFEC0H4 — — D22 IVDD — — — AC10 VSS — — — E1TXD2 PFEC1L6 — — D23 VSS — — — AC111 D24 PCIIRDY — — — AC12 IVDD — — — USB_OSCVDD — — — D251 E1RXD3 PFEC1L3 — — AC132 D26 PCIPERR — — — AC14 E0RXER PFEC0L0 — — E1 SDDATA12 — — — AC15 NC — — — E2 SDDATA15 — — — AC162 USB_PHYVDD — — — USBVBUS — — — E3 RAS — — — AC172 E4 SDCKE — — — AC18 VSS — — — E23 VSS — — — AC19 PSC2CTS PPSC3PSC23 PSC2BCLK CANRX0 E24 EVDD — — — AC20 IVDD — — — E25 PCISTOP — — — AC21 PSC0RXD PPSC1PSC01 — — E26 PCICXBE1 — — — AC22 TOUT2 PTIM4 CANTX1 — F1 SDDATA10 — — — AC23 TOUT1 — — — F2 SDDQS1 — — — AC24 DSPISIN PDSPI1 PSC3RXD — F3 SDVDD — — — AC25 DACK1 PDMA3 TOUT1 — F4 VSS — — — AC26 PSC2TXD PPSC3PSC20 — — F23 PCIPAR — — — AD1 AD16 — — — F24 PCISERR — — — AD2 AD21 — — — F25 PCIFRM — — — AD3 AD23 — — — F26 PCICXBE3 — — — AD4 AD24 — — — G1 SDDATA6 — — — AD5 AD26 — — — G2 SDDATA9 — — — AD6 ALE PFBCTL0 TBST — G3 SDDM1 — — — AD7 EVDD — — — MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 2-13 Pin Functions Primary GPIO Secondary Tertiary G4 SDDATA13 — — — G23 PCIRESET — — G24 PCICXBE0 — G25 PCICXBE2 G26 PBGA Pin PBGA Pin Table 2-2. MCF5485/MCF5484 Signal Description by Pin Number (Continued) Pin Functions Primary GPIO Secondary Tertiary AD8 E0TXD3 PFEC0L7 — — — AD9 E0TXD0 PFEC0H5 — — — — AD10 EVDD — — — — — — AD11 E0MDC PFECI2C2 — — PCIAD2 — FBADDR2 — AD12 VSS — — — H1 SDDQS0 — — — AD13 E0RXD0 PFEC0H1 — — H2 SDDATA5 — — — AD14 E0RXLK PFEC0H3 — — H3 SDDATA8 — — — AD152 USB_OSCAVDD — — — USB_PLLVDD — — — H4 IVDD — — — AD162 H23 IVDD — — — AD17 VSS — — — H24 EVDD — — — AD18 EVDD — — — H25 PCIAD1 — FBADDR1 — AD19 TIN3 PTIM7 IRQ3 CANRX1 H26 PCIAD4 — FBADDR4 — AD20 VSS — — — J1 SDDATA3 — — — AD21 PSC2RXD PPSC3PSC21 — — J2 SDDM0 — — — AD22 DSPISCK PDSPI2 PSC3CTS PSC3BCLK J3 SDDATA4 — — — AD23 TOUT3 PTIM6 CANTX1 — E1MDC — SCL CANTX0 PFEC1H6 — — J4 VSS — — — AD241 J23 PCIDEVSEL — — — AD251 E1TXEN J24 PCIAD3 — FBADDR3 — AD26 PSC2RTS J25 PCIAD5 — FBADDR5 — AE1 AD18 — — — J26 PCIAD7 — FBADDR7 — AE2 AD31 — — — K1 SDWE — — — AE3 AD28 — — — K2 SDDATA0 — — — AE4 AD27 — — — K3 SDDATA1 — — — AE5 R/W PFBCTL2 TBST — K4 SDDATA11 — — — AE6 OE PFBCTL3 — — K23 PCIAD0 — FBADDR0 — AE7 BE/BWE0 PFBCTL4 FBADDR0 — E1RXER PFEC1L0 — — PPSC3PSC22 PSC2FSYNC CANTX0 K24 PCIAD6 — FBADDR6 — AE81 K25 PCIAD8 — FBADDR8 — AE9 E0TXER PFEC0L4 — — K26 PCIAD9 — FBADDR9 — AE10 E0TXEN PFEC0H6 — — L1 SDCLK1 — — — AE111 E1TXD1 PFEC1L5 — — — AE121 E1TXD0 PFEC1h5 — — L2 SDRDQS — — MCF548x Reference Manual, Rev. 5 2-14 Freescale Semiconductor Introduction Pin Functions Primary GPIO Secondary Tertiary L3 SDVDD — — — L4 VSS — — L11 VSS — L12 VSS L13 PBGA Pin PBGA Pin Table 2-2. MCF5485/MCF5484 Signal Description by Pin Number (Continued) Pin Functions Primary GPIO Secondary Tertiary AE131 E1TXCLK PFEC1H7 — — — AE14 E0RXDV PFEC1H2 — — — — AE15 VSS — — — — — — AE16 VSS — — — VSS — — — AE17 VSS — — — L14 VSS — — — AE182 USBVDD — — — L15 VSS — — — AE19 E0CRS PFEC0H0 — — L16 VSS — — — AE20 TIN1 — — — L23 IVDD — — — AE21 PSC3RXD PPSC3PSC25 — — L24 VSS — — — AE22 PSC1RXD PPSC1PSC05 — — L25 PCIAD10 — FBADDR10 — AE23 PSC0CTS PPSC1PSC03 PSC0BCLK — L26 PCIAD11 — FBADDR11 — AE241 E1TXER PFEC1L4 — — E1MDIO — SCL CANTX0 M1 SDCLK1 — — — AE251 M2 SDBA1 — — — AE26 PSC3TXD PPSC3PSC24 — — M3 SDBA0 — — — AF1 AD29 — — — M4 SDDATA2 — — — AF2 AD25 — — — M11 VSS — — — AF3 AD30 — — — M12 VSS — — — AF4 BE/BWE3 PFBCTL7 TSIZ1 — M13 VSS — — — AF5 BE/BWE2 PFBCTL6 TSIZ0 — M14 VSS — — — AF6 TA PFBCTL1 — — M15 VSS — — — AF7 E0TXD1 PFEC0L5 — — M16 VSS — — — AF81 E1COL PFEC1H4 — — M23 VSS — — — AF9 E0TXCLK PFEC0H7 — — M24 EVDD — — — AF10 E0MDIO PFECI2C3 — — M25 PCIAD12 — FBADDR12 — AF11 E0RXD3 PFEC0L3 — — M26 PCIAD13 — FBADDR13 — AF12 E0RXD2 PFEC0L2 — — N1 SDCLK0 — — — AF13 E0RXD1 PFEC0L1 — — N2 SDCLK0 — — — AF143 USBCLKOUT — — — USBCLKIN — — — N3 SDCS0 — — — AF153 N4 SDDATA7 — — — AF163 USBD+ — — — — AF173 USBD- — — — N11 VSS — — MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 2-15 PBGA Pin PBGA Pin Table 2-2. MCF5485/MCF5484 Signal Description by Pin Number (Continued) Pin Functions Primary GPIO Secondary Tertiary N12 VSS — — — N13 VSS — — N14 VSS — N15 VSS N16 N23 Pin Functions Primary GPIO Secondary Tertiary AF18 USBRBIAS — — — — AF19 DREQ1 PDMA1 TIN1 IRQ1 — — AF20 DREQ0 PDMA0 TIN0 — — — — AF21 TIN2 PTIM5 IRQ2 CANRX1 VSS — — — AF22 TIN0 — — — PCIAD16 — FBADDR16 — AF23 PSC3CTS PPSC3PSC27 PSC3BCLK — 1 N24 PCIAD14 — FBADDR14 — AF24 E1RXD0 PFEC1H1 — — N25 PCIAD17 — FBADDR17 — AF25 PSC1CTS PPSC1PSC07 PSC1BCLK — N26 PCIAD15 — FBADDR15 — AF26 TOUT0 — — — 1 This pin is a “no connect” on the MCF5483 and MCF5482 devices. This pin is a “no connect” on the MCF5481 and MCF5480 devices. On MCF5485, MCF5484, MCF5483, and MCF5482 device the pin should be connected to the appriopriate power rail even is USB is not being used. 3 This pin is a “no connect” on the MCF5481 and MCF5480 devices. 2 2.2 2.2.1 2.2.1.1 MCF548x External Signals FlexBus Signals Address/Data Bus (AD[31:0]) The AD[31:0] bus carries address and data. The full 32-bit address is always driven on the first clock of a bus cycle (address phase). The number of bytes used for data during the data phase is determined by the port size associated with the matching chip select. 2.2.1.2 Chip Select (FBCS[5:0]) FBCS[5:0] are asserted to indicate which device is being selected. A particular chip select asserts when the transfer address is within the device’s address space as defined in the base and mask address registers. Each chip select can be programmed for a base address location, masking addresses, port size, burst-capability indication, wait-state generation, and internal/external termination. Reset clears all chip select programming; FBCS0 is the only chip select initialized out of reset. FBCS0 is also unique because it can function at reset as a global chip select that allows boot ROM to be selected at any defined address space. Port size and termination (internal vs. external) for boot FBCS0 are configured by the levels on AD[2:0] on the rising edge of RSTI, as described in Section 2.2.6, “Reset Configuration Pins.” MCF548x Reference Manual, Rev. 5 2-16 Freescale Semiconductor MCF548x External Signals 2.2.1.3 Address Latch Enable (ALE) The assertion of ALE indicates that the MCF548x has begun a bus transaction and that the address and attributes are valid. ALE is asserted for one bus clock cycle. In multiplexed bus mode, ALE is used externally as an address latch enable to capture the address phase of the bus transfer. 2.2.1.4 Read/Write (R/W) The MCF548x drives the R/W signal to indicate the direction of the current bus operation. It is driven high during read bus cycles and driven low during write bus cycles. 2.2.1.5 Transfer Burst (TBST) Transfer burst indicates that a burst transfer is in progress. A burst transfer can be 2 to 16 beats depending on the size of the transfer and the port size. 2.2.1.6 Transfer Size (TSIZ[1:0]) For memory accesses, these signals along with TBST, indicate the data transfer size of the current bus operation. The FlexBus interface supports byte, word, and longword operand transfers and allows accesses to 8-, 16-, and 32-bit data ports. For misaligned transfers, TSIZ[1:0] indicates the size of each transfer. For example, if a longword access through a 32-bit port device occurs at a misaligned offset of 0x1, a byte is transferred first (TSIZ[1:0] = 01), a word is next transferred at offset 0x2 (TSIZ[1:0] = 10), then the final byte is transferred at offset 0x4 (TSIZ[1:0] = 01). For aligned transfers larger than the port size, TSIZ[1:0] behaves as follows: • If bursting is used, TSIZ[1:0] is driven to the size of transfer. • If bursting is inhibited, TSIZ[1:0] first shows the size of the entire transfer and then shows the port size. Table 2-3. Data Transfer Size TSIZ[1:0] Transfer Size 00 4 bytes (longword) 01 1 byte 10 2 bytes (word) 11 16 bytes (line) For burst-inhibited transfers, TSIZ[1:0] changes with each ALE assertion to reflect the next transfer size. For transfers to port sizes smaller than the transfer size, TSIZ[1:0] indicates the size of the entire transfer on the first access and the size of the current port transfer on subsequent transfers. For example, for a longword write to an 8-bit port, TSIZ[1:0] = 2’b00 for the first transaction and 2’b01 for the next three transactions. If bursting is used and in the case of longword write to an 8-bit port, TSIZ[1:0] is driven to 2’b00 for the entire transfer. 2.2.1.7 Byte Selects (BE/BWE[3:0]) The four byte-enables are multiplexed with the byte-write-enable signals. Each pin can be individually programmed through the chip select control registers (CSCRs). For each chip select, assertion of MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 2-17 byte-enables for reads and byte-write enables for write cycles can be programmed. Alternatively, users can program byte-write enables to assert on writes and byte-enable to not assert on reads. The byte strobe (BE/BWE[3:0]) outputs indicate that data is to be latched or driven onto a byte of the data. BE/BWE[3:0] signals are asserted only to the memory bytes used during a read or write access. 2.2.1.8 Output Enable (OE) The output enable signal is sent to the interfacing memory and/or peripheral to enable a read transfer. OE is asserted only when a chip select matches the current address decode. 2.2.1.9 Transfer Acknowledge (TA) The external system drives this input to terminate the bus transfer. For write cycles, the processor continues to drive data at least one clock after FBCSx is negated. During read cycles, the peripheral must continue to drive data until TA is recognized. The number of wait states is determined either by an internally programmed auto acknowledgement or the external TA input. If the external TA is used, the peripheral has total control over the number of wait states. 2.2.2 SDRAM Controller Signals These signals are used for SDRAM accesses. 2.2.2.1 SDRAM Data Bus (SDDATA[31:0]) SDDATA[31:0] is the bidirectional, non-multiplexed data bus used for SDRAM accesses. Data is sampled by the MCF548x on the rising edge of SDCLK when in SDR mode, and on both the rising and falling edge of SDCLK when in DDR mode. 2.2.2.2 SDRAM Address Bus (SDADDR[12:0]) The SDADDR[12:0] signals are the 13-bit address bus used for multiplexed row and column addresses during SDRAM bus cycles. The address multiplexing supports up to 256 Mbits of SDRAM per chip select. 2.2.2.3 SDRAM Bank Addresses (SDBA[1:0]) Each SDRAM module has four internal row banks. The SDBA[1:0] signals are used to select the row bank. It is also used to select the SDRAM internal mode register during power-up initialization. 2.2.2.4 SDRAM Row Address Strobe (RAS) This output is the SDRAM synchronous row address strobe. 2.2.2.5 SDRAM Column Address Strobe (CAS) This output is the SDRAM synchronous column address strobe. 2.2.2.6 SDRAM Chip Selects (SDCS[3:0]) These signals interface to the chip select lines of the SDRAMs within a memory block. Thus, there is one SDCS line for each memory block (the MCF548x supports up to four SDRAM memory blocks). MCF548x Reference Manual, Rev. 5 2-18 Freescale Semiconductor MCF548x External Signals 2.2.2.7 SDRAM Write Data Byte Mask (SDDM[3:0]) These output signals are sampled by the SDRAM on both edges of SDDQS to determine which byte lanes of the SDRAM data bus should be latched during a write cycle. In DDR mode, these bits are ignored during read operations. 2.2.2.8 SDRAM Data Strobe (SDDQS[3:0]) These bidirectional signals indicate when valid data is on the SDRAM data bus when in DDR mode. 2.2.2.9 SDRAM Clock (SDCLK[1:0]) These signals are the output clock for SDRAM cycles. 2.2.2.10 Inverted SDRAM Clock (SDCLK[1:0]) These signals are the inverted version of the SDRAM clock. They are used with SDCLK to provide the differential clocks for DDR SDRAM. 2.2.2.11 SDRAM Write Enable (SDWE) The SDRAM write enable (SDWE) is asserted to signify that an SDRAM write cycle is underway. A read cycle is indicated by the negation of SDWE. 2.2.2.12 SDRAM Clock Enable (SDCKE) This output is the SDRAM clock enable. SDCKE is negated to put the SDRAM into low-power, self-refresh mode. 2.2.2.13 SDR SDRAM Data Strobe (SDRDQS) This signal is connected to SDDQS inputs. It is used in SDR mode only. 2.2.2.14 SDRAM Reference Voltage (VREF) This is the input reference voltage for differential SSTL_2 inputs. It is used in both DDR and SDR modes. 2.2.3 2.2.3.1 PCI Controller Signals PCI Address/Data Bus (PCIAD[31:0]) The PCIAD[31:0] lines are a time-multiplexed address data bus. The address is presented on the bus during the address phase while the data is presented on the bus during one or more data phases. If the FlexBus is used in 32-bit address or 32-bit data non-multiplexed mode, PCIAD[31:0] are used as a 32-bit address for FlexBus transfers. 2.2.3.2 Command/Byte Enables (PCICXBE[3:0]) The PCICXBE[3:0] lines are time-multiplexed. The PCI command is presented during the address phase, and the byte enables are presented during the data phase. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 2-19 2.2.3.3 Device Select (PCIDEVSEL) The PCIDEVSEL signal is asserted active low when the MCF548x decodes that it is the target of a PCI transaction from the address presented on the PCI bus during the address phase. 2.2.3.4 Frame (PCIFRM) The PCIFRM signal is asserted by a PCI initiator to indicate the beginning of a transaction. It is negated when the initiator is ready to complete the final data phase. 2.2.3.5 Initialization Device Select (PCIIDSEL) The PCIIDSEL signal is asserted during a PCI type-0 configuration cycle to address the PCI configuration header. 2.2.3.6 Initiator Ready (PCIIRDY) The PCIIRDY signal is asserted to indicate that the PCI initiator is ready to transfer data. During a write operation, assertion indicates that the master is driving valid data on the bus. During a read operation, assertion indicates that the master is ready to accept data. 2.2.3.7 Parity (PCIPAR) The PCIPAR signal indicates the parity of data on the PCIAD[31:0] and PCICXBE[3:0] lines. 2.2.3.8 Parity Error (PCIPERR) The PCIPERR signal is asserted when a data phase parity error is detected if enabled. 2.2.3.9 Reset (PCIRESET) The PCIRESET signal is asserted active low by MCF548x to reset the PCI bus. This signal is asserted after the MCF548x is reset and must be negated to enable usage of the PCI bus. 2.2.3.10 System Error (PCISERR) The PCISERR signal, if enabled, is asserted when an address phase parity error is detected. 2.2.3.11 Stop (PCISTOP) The PCISTOP signal is asserted by the currently addressed target to indicate that it wishes to stop the current transaction. 2.2.3.12 Target Ready (PCITRDY) The PCITRDY signal is asserted by the currently addressed target to indicate that it is ready to complete the current data phase. MCF548x Reference Manual, Rev. 5 2-20 Freescale Semiconductor MCF548x External Signals 2.2.3.13 External Bus Grant (PCIBG[4:1]) The PCIBG signal is asserted to an external master to give it control of the PCI bus. If the internal PCI arbiter is enabled, it asserts one of the PCIBG[4:1] lines to grant ownership of the PCI bus to an external master. When the PCI arbiter module is disabled, PCIBG[4:1] is driven high and should be ignored. 2.2.3.14 External Bus Grant/Request Output (PCIBG0/PCIREQOUT) The PCIBG0 signal is asserted to external master device 0 to give it control of the PCI bus. When the PCI arbiter module is disabled, the signal operates as the PCIREQOUT output. It is asserted when the MCF548x needs to initiate a PCI transaction. 2.2.3.15 External Bus Request (PCIBR[4:0]) The PCIBR signal is asserted by an external PCI master when it requires access to the PCI bus. 2.2.3.16 External Request/Grant Input (PCIBR0/PCIGNTIN) The PCIBR0 signal is asserted by external PCI master device 0 when it requires access to the PCI bus. When the internal PCI arbiter module is disabled, this signal is used as a grant input for the PCI bus, PCIGNTIN. It is driven by an external PCI arbiter. 2.2.4 Interrupt Control Signals The interrupt control signals supply the external interrupt level to the MCF548x device. 2.2.4.1 Interrupt Request (IRQ[7:1]) The IRQ[7:1] signals are the external interrupt inputs. 2.2.5 Clock and Reset Signals The clock and reset signals configure the MCF548x and provide interface signals to the external system. 2.2.5.1 Reset In (RSTI) Asserting RSTI causes the MCF548x to enter reset exception processing. RSTO is asserted automatically when RSTI is asserted. 2.2.5.2 Reset Out (RSTO) After RSTI is asserted, the PLL temporarily loses its lock, during which time RSTO is asserted. When the PLL regains its lock, RSTO negates again. This signal can be used to reset external devices. 2.2.5.3 Clock In (CLKIN) CLKIN is the MCF548x input clock frequency to the on-board, phase-locked loop (PLL) clock generator. CLKIN is used to internally clock or sequence the MCF548x internal bus interface at a selected multiple of the input frequency used for internal module logic. CLKIN is used as the clock reference for PCI and FlexBus transfers. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 2-21 2.2.6 Reset Configuration Pins This section describes address/data pins, AD[12:0], that are read at reset to configure the MCF548x. 2.2.6.1 AD[12:8]—CLKIN to SDCLK Ratio (CLKCONFIG[4:0]) The clock configuration inputs, CLKCONFIG[4:0], indicate the CLKIN to SDCLK ratio. CLKIN is used as the external reference for both PCI and FlexBus cycles. The CLKIN to SDCLK ratio is selectable, where SDCLK is the clock frequency used for SDRAM accesses and the internal XLB bus. The core is always clocked at twice the SDCLK frequency. These signals are sampled on the rising edge of RSTI. Table 2-4 shows how the logic levels of AD[12:8] correspond to the selected clock ratio. Table 2-4. MCF548x Divide Ratio Encodings 1 AD[12:8]1 Clock Ratio CLKIN–PCI and FlexBus Frequency Range (MHz) Internal XLB, SDRAM bus, and PSTCLK Frequency Range (MHz) Core Frequency Range (MHz) 00011 1:2 41.67–50.0 83.33–100 166.66–200 00101 1:2 25.0–41.67 50.0–83.33 100.0–166.66 01111 1:4 25.0 100 200 All other values of AD[12:8] are reserved. Figure 2-2 correlates CLKIN, internal bus, and core clock frequenciesi for the 2x–4x multipliers. CLKIN Internal Clock Core Clock 2x 25.0 2x 50.0 50.0 100.0 100.0 4x 2x 25.0 25 200.0 100.0 50 70 CLKIN (MHz) 30 50 70 90 110 200.0 130 60 80 100 120 140 Internal Clock (MHz) 160 180 200 220 240 260 Core Clock (MHz) Figure 2-2. CLKIN, Internal Bus, and Core Clock Ratios 2.2.6.2 AD5—FlexBus Size Configuration (FBSIZE) At reset, the enabling and disabling of BE/BWE[3:0] versus TSIZ[1:0] and ADDR[1:0] is determined by the logic level driven on AD5 at the rising edge of RSTI. FBSIZE is multiplexed with AD5 and sampled only at reset. Table 2-5 shows how the AD5 logic level corresponds to the BE/BWE[3:0] function. MCF548x Reference Manual, Rev. 5 2-22 Freescale Semiconductor MCF548x External Signals Table 2-5. AD5/FBSIZE Selection of BE/BWE[3:0] Signals AD5 2.2.6.3 FlexBus Byte Enable Mode 0 BE/BWE[3:0] used as byte/byte write enables. 1 BE/BWE[3:2] configured as TSIZ[1:0]. BE/BWE[1:0] configured as FBADDR[1:0]. AD4—32-bit FlexBus Configuration (FBMODE) During reset, the FlexBus can be configured to operate in a non-multiplexed 32-bit address with 32-bit data mode. In this mode, the 32-bit FlexBus AD[31:0] is used for the data bus, and the PCI bus PCIAD[31:0] is used as the address bus. The FlexBus operating mode is determined by the logic level driven on AD4 at the rising edge of RSTI. Table 2-6 shows how the logic level of AD4 corresponds to the FlexBus mode. Table 2-6. AD4/FBMODE Selection of Non-Multiplexed 32-bit Address/32-bit Data Mode AD4 1 2.2.6.4 FlexBus Operating Mode 0 AD[31:0] used for data. PCIAD[31:0] used for address1 1 PCIAD[31:0] used for PCI bus. AD[31:0] used for both address and data. If the non-multiplexed 32-bit address/32-bit data mode is selected, the PCI bus cannot be used. AD3—Byte Enable Configuration (BECONFIG) The default byte enable mode of the boot FBCS0 is determined by the logic level driven on AD3 at the rising edge of RSTI. This logic level is reflected as the reset value of CSCR0[BEM]. Table 2-7 shows how the logic level of AD3 corresponds to the byte enable mode for FBCS0 at reset. Table 2-7. AD3/BECONFIG, BE/BWE[3:0] Boot Configuration AD3 2.2.6.5 Boot FBCS0 Byte Strobe Configuration 0 BWE[3:0] are not asserted for reads; BWE[3:0] only assert for write cycles 1 BE[3:0] can assert for both read and write cycles. AD2—Auto Acknowledge Configuration (AACONFIG) At reset, the enabling and disabling of auto acknowledge for boot FBCS0 is determined by the logic level driven on AD2 at the rising edge of RSTI. AACONFIG is multiplexed with AD2 and sampled only at reset. The AD2 logic level is reflected as the reset value of CSCR0[AA]. Table 2-8 shows how the AD2 logic level corresponds to the auto acknowledge timing for FBCS0 at reset. Auto acknowledge can be disabled by driving a logic 0 on AD2 at reset. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 2-23 Table 2-8. AD2/AA_CONFIG Selection of FBCS0 Automatic Acknowledge AD2 2.2.6.6 Boot FBCS0 AA Configuration at Reset 0 Disabled 1 Enabled with 63 wait states AD[1:0]—Port Size Configuration (PSCONFIG) The default port size value of the boot FBCS0 is determined by the logic levels driven on AD[1:0] at the rising edge of RSTI, which are reflected as the reset value of CSCR0[PS]. Table 2-9 shows how the logic levels of AD[1:0] correspond to the FBCS0 port size at reset. Table 2-9. AD[1:0]/PSCONFIG[1:0] Selection of FBCS0 Port Size 2.2.7 AD[1:0] Boot FBCS0 Port Size 00 32-bit port 01 8-bit port 1X 16-bit port Ethernet Module Signals The following signals are used by the Ethernet module for data and clock signals. 2.2.7.1 Management Data (E0MDIO, E1MDIO) The bidirectional EMDIO signals transfer control information between the external PHY and the media-access controller. Data is synchronous to EMDC and applies to MII mode operation. This signal is an input after reset. When the FEC operates in 10 Mbps 7-wire interface mode, this signal should be connected to VSS. 2.2.7.2 Management Data Clock (E0MDC, E1MDC) EMDC is an output clock that provides a timing reference to the PHY for data transfers on the EMDIO signal; it applies to MII mode operation. 2.2.7.3 Transmit Clock (E0TXCLK, E1TXCLK) This is an input clock that provides a timing reference for ETXEN, ETXD[3:0], and ETXER. 2.2.7.4 Transmit Enable (E0TXEN, E1TXEN) The transmit enable (ETXEN) output indicates when valid nibbles are present on the MII. This signal is asserted with the first nibble of a preamble and is negated before the first ETXCLK following the final nibble of the frame. MCF548x Reference Manual, Rev. 5 2-24 Freescale Semiconductor MCF548x External Signals 2.2.7.5 Transmit Data 0 (E0TXD0, E1TXD0) ETXD0 is the serial output Ethernet data and is only valid during the assertion of ETXEN. This signal is used for 10 Mbps Ethernet data. This signal is also used for MII mode data in conjunction with ETXD[3:1]. 2.2.7.6 Collision (E0COL, E1COL) The ECOL input is asserted upon detection of a collision and remains asserted while the collision persists. This signal is not defined for full-duplex mode. 2.2.7.7 Receive Clock (E0RXCLK, E1RXCLK) The receive clock (ERXCLK) input provides a timing reference for ERXDV, ERXD[3:0], and ERXER. 2.2.7.8 Receive Data Valid (E0RXDV, E1RXDV) Asserting the receive data valid (ERXDV) input indicates that the PHY has valid nibbles present on the MII. ERXDV should remain asserted from the first recovered nibble of the frame through to the last nibble. Assertion of ERXDV must start no later than the SFD and exclude any EOF. 2.2.7.9 Receive Data 0 (E0RXD0, E1RXD0) ERXD0 is the Ethernet input data transferred from the PHY to the media-access controller when ERXDV is asserted. This signal is used for 10 Mbps Ethernet data. This signal is also used for MII mode Ethernet data in conjunction with ERXD[3:1]. 2.2.7.10 Carrier Receive Sense (E0CRS, E1CRS) ECRS is an input signal that, when asserted, signals that transmit or receive medium is not idle, and applies to MII mode operation. 2.2.7.11 Transmit Data 1–3 (E0TXD[3:1], E1TXD[3:1]) These pins contain the serial output Ethernet data and are valid only during assertion of ETXEN in MII mode. 2.2.7.12 Transmit Error (E0TXER, E1TXER) When the ETXER output is asserted for one or more clock cycles while ETXEN is also asserted, the PHY sends one or more illegal symbols. ETXER has no effect at 10 Mbps or when ETXEN is negated, and applies to MII mode operation. 2.2.7.13 Receive Data 1–3 (E0RXD[3:1], E1RXD[3:1]) These pins contain the Ethernet input data transferred from the PHY to the media-access controller when ERXDV is asserted in MII mode operation. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 2-25 2.2.7.14 Receive Error (E0RXER, E1RXER) ERXER is an input signal that, when asserted along with ERXDV, signals that the PHY has detected an error in the current frame. When ERXDV is not asserted, ERXER has no effect and applies to MII mode operation. 2.2.8 2.2.8.1 Universal Serial Bus (USB) USB Differential Data (USBD+, USBD–) USBD+ and USBD– are the outputs of the on-chip USB 2.0 transceiver. They provide differential data for the USB 2.0 bus. 2.2.8.2 USBVBUS This is the USB cable Vbus monitor input, which is 5 V tolerant. 2.2.8.3 USBRBIAS This is the connection for external current setting resistor. It should be connected to a 9.1kΩ +/– 1% pull-down resistor. For the MCF5481 and MCF5480 devices this pin should be connected to a 9.1kΩ +/– 20% pull-down resistor. 2.2.8.4 USBCLKIN This is the input pin for 12-MHz USB crystal circuit. 2.2.8.5 USBCLKOUT This is the output pin for 12-MHz USB crystal circuit. 2.2.9 2.2.9.1 DMA Serial Peripheral Interface (DSPI) Signals DSPI Synchronous Serial Data Output (DSPISOUT) The DSPISOUT output provides the serial data from the DSPI and can be programmed to be driven on the rising or falling edge of DSPISCK. 2.2.9.2 DSPI Synchronous Serial Data Input (DSPISIN) The DSPISIN input provides the serial data to the DSPI and can be programmed to be sampled on the rising or falling edge of DSPISCK. 2.2.9.3 DSPI Serial Clock (DSPISCK) DSPISCK is a serial communication clock signal. In master mode, the DSPI generates the DSPISCK. In slave mode, DSPISCK is an input from an external bus master. MCF548x Reference Manual, Rev. 5 2-26 Freescale Semiconductor MCF548x External Signals 2.2.9.4 DSPI Peripheral Chip Select/Slave Select (DSPICS0/SS) In master mode, the DSPICS0 signal is a peripheral chip select output that selects which slave device the current transmission is intended for. In slave mode, the SS signal is a slave select input signal that allows an SPI master to select the DSPI as the target for transmission. 2.2.9.5 DSPI Chip Selects (DSPICS[2:3]) The synchronous peripheral chip selects (DSPICS[2:3]) outputs provide DSPI peripheral chip selects that can be programmed to be active high or low. 2.2.9.6 DSPI Peripheral Chip Select 5/Peripheral Chip Select Strobe (DSPICS5/PCSS) DSPICS5 is a peripheral chip select output signal. When the DSPI is in master mode and the DMCR[PCSSE] bit is cleared, this signal is used to select which slave device the current transfer is intended for. PCSS provides a strobe signal that can be used with an external demultiplexer for deglitching of the DSPICSn signals. When the DSPI is in master mode and DMCR[PCSSE] is set, the PCSS provides the appropriate timing for the decoding of the DSPICS[0,2,3] signals which prevents glitches from occurring. This signal is not used in slave mode. 2.2.10 2.2.10.1 FlexCAN Signals FlexCAN Transmit (CANTX0, CANTX1) Controller area network transmit data output. 2.2.10.2 FlexCAN Receive (CANRX0, CANRX1) Controller area network receive data input. 2.2.11 I2C I/O Signals The I2C serial interface module uses the signals in this section. 2.2.11.1 Serial Clock (SCL) This bidirectional open-drain signal is the clock signal for the I2C interface. It is either driven by the I2C module when the bus is in master mode, or it becomes the clock input when the I2C is in slave mode. 2.2.11.2 Serial Data (SDA) This bidirectional open-drain signal is the data input/output for the I2C interface. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 2-27 2.2.12 PSC Module Signals The PSC modules use the signals in this section. The baud rate clock inputs are not supported. 2.2.12.1 Transmit Serial Data Output (PSC0TXD, PSC1TXD, PSC2TXD, PSC3TXD) PSCnTXD are the transmitter serial data outputs for the PSC modules. The output is held high (mark condition) when the transmitter is disabled, idle, or in the local loopback mode. The PSCxTXD pins can be programmed to be driven low (break status) by a command. 2.2.12.2 Receive Serial Data Input (PSC0RXD, PSC1RXD, PSC2RXD, PSC3RXD) PSCnRXD are the receiver serial data inputs for the PSC modules. When the PSC clock is stopped for power-down mode, any transition on the pins restarts them. 2.2.12.3 Clear-to-Send (PSCnCTS/PSCBCLK) These signals either operate as the clear-to-send input signals in UART mode or the bit clock input signals in modem modes and IrDA modes. In MIR and FIR mode, the frequency is a multiple of the input bit clock frequency, and the bit clock frequency should be within +/-0.1% and +/-0.01% of the ideal one, respectively. 2.2.12.4 Request-to-Send (PSCnRTS/PSCFSYNC) The PSCnRTS signals act as transmitter request-to-send (RTS) outputs in UART mode, the frame sync input in modem8 and modem16 modes, or the RTS output (which acts as frame sync) in AC97 modem mode. 2.2.13 DMA Controller Module Signals The DMA controller module uses the signals in the following subsections to provide external requests for either a source or destination. 2.2.13.1 DMA Request (DREQ[1:0]) These inputs are asserted by a peripheral device to request an operand transfer between that peripheral and memory by either channel 0 or 1 of the on-chip DMA module. 2.2.13.2 DMA Acknowledge (DACK[1:0]) These outputs are asserted to acknowledge that a DMA request has been recognized. 2.2.14 Timer Module Signals The signals in the following sections are external interfaces to the four general-purpose MCF548x timers. These 32-bit timers can capture timer values, trigger external events or internal interrupts, or count external events. MCF548x Reference Manual, Rev. 5 2-28 Freescale Semiconductor MCF548x External Signals 2.2.14.1 Timer Inputs (TIN[3:0]) TINn can be programmed as clocks that cause events in the counter and prescalers. They can also cause captures on the rising edge, falling edge, or both edges. 2.2.14.2 Timer Outputs (TOUT[3:0]) The programmable timer outputs, TOUTn, pulse or toggle on various timer events. 2.2.15 Debug Support Signals The MCF548x complies with the IEEE 1149.1a JTAG testing standard. JTAG test pins are multiplexed with background debug pins. Except for TCK, these signals are selected by the value of MTMOD0. If MTMOD0 is high, JTAG signals are chosen; if it is low, debug module signals are chosen. MTMOD0 should be changed only while RSTI is asserted. 2.2.15.1 Processor Clock Output (PSTCLK) The internal PLL generates this output signal, and is the processor clock output that is used as the timing reference for the debug bus timing (PSTDDATA[7:0]). PSTCLK is at the same frequency as the internal XLB and SDRAM bus frequency. The frequency is one-half the core frequency. 2.2.15.2 Processor Status Debug Data (PSTDDATA[7:0]) Processor status data outputs indicate both processor status and captured address/data values. They operate at half the processor’s frequency, using PSTCLK. Given that real-time trace information appears as a sequence of 4-bit data values, there are no alignment restrictions; that is, PST values and operands may appear on either PSTDDATA[7:0] nibble. The upper nibble, PSTDDATA[7:4], is most significant. 2.2.15.3 Development Serial Clock/Test Reset (DSCLK/TRST) If MTMOD0 is low, DSCLK is selected. DSCLK is the development serial clock for the serial interface to the debug module. The maximum DSCLK frequency is 1/5 CLKIN. If MTMOD0 is high, TRST is selected. TRST asynchronously resets the internal JTAG controller to the test logic reset state, causing the JTAG instruction register to choose the bypass instruction. When this occurs, JTAG logic is benign and does not interfere with normal MCF548x functionality. Although TRST is asynchronous, Freescale recommends that it makes an asserted-to-negated transition only while TMS is held high. TRST has an internal pull-up resistor so if it is not driven low, it defaults to a logic level of 1. If TRST is not used, it can be tied to ground or, if TCK is clocked, to EVDD. Tying TRST to ground places the JTAG controller in test logic reset state immediately. Tying it to EVDD causes the JTAG controller (if TMS is a logic level of 1) to eventually enter test logic reset state after 5 TCK clocks. 2.2.15.4 Breakpoint/Test Mode Select (BKPT/TMS) If MTMOD0 is low, BKPT is selected. BKPT signals a hardware breakpoint to the processor in debug mode. If MTMOD0 is high, TMS is selected. The TMS input provides information to determine the JTAG test operation mode. The state of TMS and the internal 16-state JTAG controller state machine at the rising edge of TCK determine whether the JTAG controller holds its current state or advances to the next state. This directly controls whether JTAG data or instruction operations occur. TMS has an internal pull-up MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 2-29 resistor so that if it is not driven low, it defaults to a logic level of 1. But if TMS is not used, it should be tied to VDD. 2.2.15.5 Development Serial Input/Test Data Input (DSI/TDI) If MTMOD0 is low, DSI is selected. DSI provides the single-bit communication for debug module commands. If MTMOD0 is high, TDI is selected. TDI provides the serial data port for loading the various JTAG boundary scan, bypass, and instruction registers. Shifting in data depends on the state of the JTAG controller state machine and the instruction in the instruction register. Shifts occur on the TCK rising edge. TDI has an internal pull-up resistor, so when not driven low it defaults to high. But if TDI is not used, it should be tied to EVDD. 2.2.15.6 Development Serial Output/Test Data Output (DSO/TDO) If MTMOD0 is low, DSO is selected. DSO provides single-bit communication for debug module responses. If MTMOD0 is high, TDO is selected. The TDO output provides the serial data port for outputting data from JTAG logic. Shifting out data depends on the JTAG controller state machine and the instruction in the instruction register. Data shifting occurs on the falling edge of TCK. When TDO is not outputting test data, it is three-stated. TDO can be three-stated to allow bused or parallel connections to other devices having a JTAG port. 2.2.15.7 Test Clock (TCK) TCK is the dedicated JTAG test logic clock independent of the MCF548x processor clock. Various JTAG operations occur on the rising or falling edge of TCK. Holding TCK high or low for an indefinite period does not cause JTAG test logic to lose state information. If TCK is not used, it must be tied to ground. 2.2.16 2.2.16.1 Test Signals Test Mode (MTMOD[3:0]) The test mode signals choose between multiplexed debug module and JTAG signals. If MTMOD0 is low, the part is in normal and background debug mode (BDM); if it is high, it is in normal and JTAG mode. All other MTMOD values are reserved; MTMOD[3:1] should be tied to ground and MTMOD[3:0] should not be changed while RSTI is negated 2.2.17 Power and Reference Pins These pins provide system power, ground, and references to the device. Multiple pins are provided for adequate current capability. All power supply pins must have adequate bypass capacitance for high-frequency noise suppression. 2.2.17.1 Positive Pad Supply (EVDD) This pin supplies positive power to the I/O pads. MCF548x Reference Manual, Rev. 5 2-30 Freescale Semiconductor MCF548x External Signals 2.2.17.2 Positive Core Supply (IVDD) This pin supplies positive power to the core logic. 2.2.17.3 Ground (VSS) This pin is the negative supply (ground) to the chip. 2.2.17.4 USB Power (USBVDD) This pin supplies positive power to the USB module’s digital logic. 2.2.17.5 USB Oscillator Power (USB_OSCVDD) This pin supplies positive power to the USB oscillator’s digital logic. 2.2.17.6 USB PHY Power (USB_PHYVDD) This pin supplies positive power to the USB PHY’s digital logic. 2.2.17.7 USB Oscillator Analog Power (USB_OSCAVDD) This pin supplies positive power to the USB oscillator’s analog circuits. 2.2.17.8 USB PLL Analog Power (USB_PLLVDD) This pin supplies positive power to the USB PLL’s circuits. 2.2.17.9 SDRAM Memory Supply (SDVDD) This pin supplies positive power to the SDRAM module. 2.2.17.10 PLL Analog Power (PLLVDD) This pin supplies the positive power for the PLL. 2.2.17.11 PLL Analog Ground (PLLVSS) This pin is the negative supply (ground) to the PLL. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 2-31 MCF548x Reference Manual, Rev. 5 2-32 Freescale Semiconductor Part I Processor Core Part I is intended for system designers who need to understand the operation of the MCF548x ColdFire core and its enhanced multiply/accumulate (EMAC) execution unit. It describes the programming and exception models, Harvard memory implementation, and debug module. Contents Part 1 contains the following chapters: • Chapter 3, “ColdFire Core,” provides an overview of the microprocessor core of the MCF548x. The chapter begins with a description of enhancements from the V3 ColdFire core, and then fully describes the V4e programming model as it is implemented on the MCF548x. It also includes a full description of exception handling, data formats, an instruction set summary, and a table of instruction timings. • Chapter 4, “Enhanced Multiply-Accumulate Unit (EMAC),” describes the MCF548x enhanced multiply/accumulate unit, which executes integer multiply, multiply-accumulate, and miscellaneous register instructions. The EMAC is integrated into the operand execution pipeline (OEP). • Chapter 5, “Memory Management Unit (MMU),” describes the ColdFire virtual memory management unit (MMU), which provides virtual-to-physical address translation and memory access control. • Chapter 6, “Floating-Point Unit (FPU),” describes instructions implemented in the floating-point unit (FPU) designed for use with the ColdFire family of microprocessors. • Chapter 7, “Local Memory,” describes the MCF548x implementation of the ColdFire V4e local memory specification. • Chapter 8, “Debug Support,” describes the Revision C enhanced hardware debug support in the MCF548x. This revision of the ColdFire debug architecture encompasses earlier revisions. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor i MCF548x Reference Manual, Rev. 5 ii Freescale Semiconductor Chapter 3 ColdFire Core This chapter provides an overview of the microprocessor core of the MCF548x. The CF4e implementation of the Version 4 (V4) core includes the floating-point unit (FPU), enhanced multiply-accumulate unit (EMAC), and memory management unit (MMU); all are defined as optional in the V4 architecture. This chapter also includes a full description of exception handling, data formats, an instruction set summary, and a table of instruction timings. 3.1 Core Overview The MCF548x is the first standard product to contain a Version 4e ColdFire microprocessor core. To create this next-generation, high-performance core, many advanced microarchitectural techniques were implemented. Most notable are a Harvard memory architecture, branch cache acceleration logic, and limited superscalar dual-instruction issue capabilities, which together provide 308 (Dhrystone 2.1) MIPS performance at 200 MHz. The MCF548x core design emphasizes performance and backward compatibility, and represents the next step on the ColdFire performance roadmap. 3.2 Features The CF4e includes the following features defined as optional in the V4 core architecture: • Floating-point unit (FPU) • Virtual memory management unit (MMU) • Enhanced multiply-accumulate unit (EMAC) for increased signal processing functionality plus backward code compatibility with the MAC unit of previous ColdFire processors V4 architecture features are defined as follows: • Variable-length RISC, clock-multiplied core • Revision B of the ColdFire instruction set architecture (ISA_B), providing new instructions to improve performance and code density • Two independent, decoupled pipelines—four-stage instruction fetch pipeline (IFP) and five-stage operand execution pipeline (OEP) for increased performance • Ten-instruction, FIFO buffer that decouples the IFP and OEP • Limited superscalar design approaches dual-issue performance with the cost of a scalar execution pipeline • Two-level branch acceleration mechanism with a branch cache, plus a prediction table for increased performance of conditional Bcc instructions • 32-bit address bus supporting 4 Gbytes of linear address space • 32-bit data bus • 16 user-accessible, 32-bit-wide, general-purpose registers • Supervisor/user modes for system protection • Two separate stack pointer (A7) registers—the supervisor stack pointer (SSP) and the user stack pointer (USP)—that provide the required isolation between operating modes to support the MMU. • Vector base register to relocate the exception-vector table • Optimized for high-level language constructs MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 3-1 3.2.1 Enhanced Pipelines The IFP prefetches instructions. The OEP decodes instructions, fetches required operands, then executes the specified function. The two independent, decoupled pipeline structures maximize performance while minimizing core size. Pipeline stages are shown in Figure 3-1 and are summarized as follows: • Four-stage IFP (plus optional instruction buffer stage) — Instruction address generation (IAG) calculates the next prefetch address. — Instruction fetch cycle 1 (IC1) initiates prefetch on the processor’s local instruction bus. — Instruction fetch cycle 2 (IC2) completes prefetch on the processor’s local instruction bus. — Instruction early decode (IED) generates time-critical decode signals needed for the OEP. — Instruction buffer (IB) stage uses FIFO queue to minimize effects of fetch latency. • Five-stage OEP with two optional processor bus write cycles — Decode stage (DS/secDS) decodes and selects for two sequential instructions. — Operand address generation (OAG) generates the address for the data operand. — Operand fetch cycle 1 and 2 (OC1 and OC2) fetch data operands. — Execute (EX) performs prescribed operations on previously fetched data operands. — Write data available (DA) makes data available for operand write operations only. — Store data (ST) updates memory element for operand write operations only. MCF548x Reference Manual, Rev. 5 3-2 Freescale Semiconductor Features Instruction Fetch Pipeline IAG Branch Cache Instruction Memory IC1 IC2 Branch Accel. IED IB Operand Execution Pipeline DS Internal Bus secDS OAG Data (Operand) Memory OC1 OC2 Misalignment Module EX DA Debug DSCLK DSI DSO DDATA PSTDDATA PSTCLK Figure 3-1. ColdFire Enhanced Pipeline 3.2.1.1 Instruction Fetch Pipeline (IFP) Because the fetch and execution pipelines are decoupled by a ten-instruction FIFO buffer, the IFP can prefetch instructions before the OEP needs them, minimizing stalls. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 3-3 3.2.1.1.1 Branch Acceleration To maximize the performance of conditional branch instructions, the IFP implements a sophisticated two-level acceleration mechanism. The first level is an 8-entry, direct-mapped branch cache with 2 bits for indicating four prediction states (strongly or weakly; taken or not-taken) for each entry. The branch cache also provides the association between instruction addresses and the corresponding target address. In the event of a branch cache hit, if the branch is predicted as taken, the branch cache sources the target address from the IC1 stage back into the IAG to redirect the prefetch stream to the new location. The branch cache implements instruction folding, so conditional branch instructions correctly predicted as taken can execute in zero cycles. For conditional branches with no information in the branch cache, a second-level, direct-mapped prediction table is accessed. Each of its 128 entries uses the same 2-bit prediction mechanism as the branch cache. If a branch is predicted as taken, branch acceleration logic in the IED stage generates the target address. Other change-of-flow instructions, including unconditional branches, jumps, and subroutine calls, use a similar mechanism where the IFP calculates the target address. The performance of subroutine return instruction (RTS) is improved through the use of a four-entry, LIFO hardware return stack. In all cases, these mechanisms allow the IFP to redirect the fetch stream down the predicted path well ahead of instruction execution. 3.2.1.2 Operand Execution Pipeline (OEP) The two instruction registers in the decode stage (DS) of the OEP are loaded from the FIFO instruction buffer or are bypassed directly from the instruction early decode (IED). The OEP consists of two traditional, two-stage RISC compute engines with a dual-ported register file access feeding an arithmetic logic unit (ALU). The compute engine at the top of the OEP (the address ALU) is used typically for operand address calculations; the execution ALU at the bottom is used for instruction execution. The resulting structure provides 4 Gbytes/S operand bandwidth (at 162 MHz) to the two compute engines and supports single-cycle execution speeds for most instructions, including all load and store operations and most embedded-load operations. The V4 OEP supports the ColdFire Revision B instruction set, which adds a few new instructions to improve performance and code density. The OEP also implements the following advanced performance features: • Stalls are minimized by dynamically basing the choice between the address ALU or execution ALU for instruction execution on the pipeline state. • The address ALU and register renaming resources together can execute heavily used opcodes and forward results to subsequent instructions with no pipeline stalls. • Instruction folding involving MOVE instructions allows two instructions to be issued in one cycle. The resulting microarchitecture approaches full superscalar performance at a much lower silicon cost. 3.2.1.2.1 Illegal Opcode Handling To aid in conversion from M68000 code, every 16-bit operation word is decoded to ensure that each instruction is valid. If the processor attempts execution of an illegal or unsupported instruction, an illegal instruction exception (vector 4) is taken. 3.2.1.2.2 Enhanced Multiply/Accumulate (EMAC) Unit The EMAC unit in the Version 4e provides hardware support for a limited set of digital signal processing (DSP) operations used in embedded code, while supporting the integer multiply instructions in the MCF548x Reference Manual, Rev. 5 3-4 Freescale Semiconductor Features ColdFire microprocessor family. The MAC features a four-stage execution pipeline, optimized for 32 × 32 multiplies. It is tightly coupled to the OEP, which can issue a 32 x 32 multiply with a 32-bit accumulation and fetch a 32-bit operand in a single cycle. A 32 x 32 multiply with a 32-bit accumulation requires four cycles before the next instruction can be issued. Figure 3-2 shows basic functionality of the EMAC. A full set of instructions are provided for signed and unsigned integers plus signed, fixed-point fractional input operands. Operand Y Operand X X Shift 0,1,-1 +/- Accumulator Figure 3-2. ColdFire Multiply-Accumulate Functionality Diagram The EMAC provides functionality in the following three related areas, which are described in detail in Chapter 4, “Enhanced Multiply-Accumulate Unit (EMAC):” • Signed and unsigned integer multiplies • Multiply-accumulate operations with signed and unsigned fractional operands • Miscellaneous register operations 3.2.1.2.3 Memory Management Unit (MMU) The ColdFire memory management architecture provides a demand-paged, virtual-address environment with hardware address translation acceleration. It supports supervisor/user, read, write, and execute permission checking on a per-memory request basis. The architecture defines the MMU TLB, associated control logic, TLB hit/miss logic, address translation based on the TLB contents, and access faults due to TLB misses and access violations. It intentionally leaves some virtual environment details undefined to maximize the software-defined flexibility. These include the exact structure of the memory-resident pointer descriptor/page descriptor tables, the base registers for these tables, the exact information stored in the tables, the methodology (if any) for maintenance of access, and written information on a per-page basis. 3.2.1.2.4 Floating Point Unit (FPU) The floating-point unit (FPU) provides hardware support for floating point math operations. The FPU conforms to the American National Standards Institute (ANSI)/Institute of Electrical and Electronics Engineers (IEEE) Standard for Binary Floating-Point Arithmetic (ANSI/IEEE Standard 754). MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 3-5 The hardware unit is optimized for real-time execution with exceptions disabled and default results provided for specific operations, operands, and number types. The FPU does not support all IEEE-754 number types and operations in hardware. Exceptions can be enabled to support these cases in software. 3.2.1.2.5 Hardware Divide Unit The hardware divide unit performs the following integer division operations: • 32-bit operand/16-bit operand producing a 16-bit quotient and a 16-bit remainder • 32-bit operand/32-bit operand producing a 32-bit quotient • 32-bit operand/32-bit operand producing a 32-bit remainder 3.2.1.3 Harvard Memory Architecture A Harvard memory architecture supports the increased bandwidth requirements of the CF4e processor pipelines by providing separate configuration, access control, and protection resources for data (operand) and instruction memory. The CF4e has separate instruction and data buses to processor-local memories, eliminating conflicts between instruction fetches and operand accesses. 3.2.2 Debug Module Enhancements The ColdFire processor core debug interface supports system integration in conjunction with low-cost development tools. Real-time trace and debug information can be accessed through a standard interface, which allows the processor and system to be debugged at full speed without costly in-circuit emulators. The CF4e debug unit is a compatible upgrade to MCF52xx and MCF53xx debug modules with added support for the CF4e MMU module. The Version 2 ColdFire core implemented the original debug architecture, now called Revision A. Based on feedback from customers and third-party developers, enhancements have been added to succeeding generations of ColdFire cores. For Revision A, CSR[HRL] is 0. See Section 8.4.2, “Configuration/Status Register (CSR).” The Version 3 core implements Revision B of the debug architecture, offering more flexibility for configuring the hardware breakpoint trigger registers and removing the restrictions involving concurrent BDM processing while hardware breakpoint registers are active. For Revision B, CSR[HRL] is 1. Revision C of the debug architecture more than doubles the on-chip breakpoint registers and provides an ability to interrupt debug service routines. For Revision C, CSR[HRL] is 2. Differences between Revision B and C are summarized as follows: • Debug Revision B has separate PST[3:0] and DDATA[3:0] signals. • Debug Revision C adds breakpoint registers and supports normal interrupt request service during debug. It combines debug signals into PSTDDATA[7:0]. The addition of the memory management unit (MMU) to the baseline architecture requires corresponding enhancements to the ColdFire debug functionality, resulting in Revision D. For Revision D, the revision level bit, CSR[HRL], is 3. With software support, the MMU can provide a demand-paged, virtual address environment. To support debugging in this virtual environment, the debug enhancements are primarily related to the expansion of the virtual address to include the 8-bit address space identifier (ASID). Conceptually, the virtual address is expanded to a 40-bit value: the 8-bit ASID plus the 32-bit address. The expansion of the virtual address affects the following two major debug functions: MCF548x Reference Manual, Rev. 5 3-6 Freescale Semiconductor Programming Model • • The ASID is optionally included in the specification of the hardware breakpoint registers. As an example, the four PC breakpoint registers are each expanded by 8 bits, so that a specific ASID value may be programmed as part of the breakpoint instruction address. Likewise, each operand address/data breakpoint register is expanded to include an ASID value. Finally, new control registers define if and how the ASID is to be included in the breakpoint comparison trigger logic. The debug module implements the concept of ownership trace in which the ASID value may be optionally displayed as part of the real-time trace functionality. When enabled, real-time trace displays instruction addresses on every change-of-flow instruction that is not absolute or PC-relative. For Revision D, this instruction address display optionally includes the contents of the ASID, thus providing the complete instruction virtual address on these instructions. Additionally when a Sync_PC serial BDM command is loaded from the external development system, the processor optionally displays the complete virtual instruction address, including the 8-bit ASID value. In addition to these ASID-related changes, the new MMU control registers are accessible by using serial BDM commands. The same BDM access capabilities are also provided for the EMAC and FPU programming models. Finally, a new serial BDM command is implemented to assist debugging when a software error generates an incorrect memory address that hangs the external bus. The new BDM command attempts to break this condition by forcing a bus termination. 3.3 Programming Model The MCF548x programming model consists of two instruction and register groups—user and supervisor, shown in Figure 3-3. User mode programs are restricted to user, EMAC, and floating point instructions and programming models. Supervisor-mode system software can reference all user-mode, EMAC, and floating point instructions and registers and additional supervisor instructions and control registers. The user or supervisor programming model is selected based on SR[S]. The following sections describe the registers in the user, EMAC, floating point, and supervisor programming models. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 3-7 31 0 31 Data registers A0 A1 A2 A3 A4 A5 A6 A7 PC CCR Address registers User Registers 0 63 User stack pointer Program counter Condition code register 0 31 FP0 FP1 FP2 FP3 FP4 FP5 FP6 FP7 FPCR FPSR FPIAR Floating-point data registers MACSR ACC0 ACC1 ACC2 ACC3 ACCext01 ACCext23 MASK MAC status register MAC accumulator 0 MAC accumulator 1 (EMAC only) MAC accumulator 2 (EMAC only) MAC accumulator 3 (EMAC only) ACC0 and ACC1 extensions ACC2 and ACC3 extensions MAC mask register Floating-point control register Floating-point status register Floating-point instruction address register 0 15 31 Supervisor Registers D0 D1 D2 D3 D4 D5 D6 D7 0 (CCR) SR OTHER_A7 Must be zeros VBR CACR ASID ACR0 ACR1 ACR2 ACR3 MMUBAR ROMBAR0 ROMBAR1 RAMBAR0 RAMBAR1 MBAR 19 Status register Supervisor A7 stack pointer Vector base register Cache control register Address space ID register Access control register 0 (data) Access control register 1 (data) Access control register 2 (instruction) Access control register 3 (instruction) MMU base address register ROM base address register 0 ROM base address register 1 RAM base address register0 RAM base address register 1 Module base address register Figure 3-3. ColdFire Programming Model MCF548x Reference Manual, Rev. 5 3-8 Freescale Semiconductor Programming Model 3.3.1 User Programming Model The user programming model, shown in Figure 3-3, consists of the following registers: • 16 general-purpose, 32-bit registers (D7–D0 and A7–A0); A7 is a user stack pointer • 32-bit program counter • 8-bit condition code register • Registers to support the EMAC • Register to support the floating-point unit (FPU) 3.3.1.1 Data Registers (D0–D7) Registers D0–D7 are used as data registers for bit, byte (8-bit), word (16-bit), and longword (32-bit) operations. They may also be used as index registers. 3.3.1.2 Address Registers (A0–A6) The address registers (A0–A6) can be used as software stack pointers, index registers, or base address registers, and may be used for word and longword operations. 3.3.2 User Stack Pointer (A7) The CF4e architecture supports two unique stack pointer (A7) registers—the supervisor stack pointer (SSP) and the user stack pointer (USP). This support provides the required isolation between operating modes as dictated by the virtual memory management scheme provided by the memory management unit (MMU). The SSP is described in Section 5.4.2, “Supervisor/User Stack Pointers.” 3.3.2.1 Program Counter (PC) The PC holds the address of the executing instruction. For sequential instructions, the processor automatically increments PC. When program flow changes, the PC is updated with the target instruction. For some instructions, the PC specifies the base address for PC-relative operand addressing modes. 3.3.2.2 Condition Code Register (CCR) The CCR, Figure 3-4, occupies SR[7–0], as shown in Figure 3-3. The CCR[4–0] bits are indicator flags based on results generated by arithmetic operations. R 7 6 5 4 3 2 1 0 0 0 0 X N Z V C 0 0 0 0 0 0 0 0 W Reset Reg Addr Accessed using R/W commands for the status register Figure 3-4. Condition Code Register (CCR) MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 3-9 Table 3-1. CCR Field Descriptions 3.3.3 Bits Name Description 7–5 — Reserved, should be cleared. 4 X Extend condition code bit. Assigned the value of the carry bit for arithmetic operations; otherwise not affected or set to a specified result. Also used as an input operand for multiple-precision arithmetic. 3 N Negative condition code bit. Set if the msb of the result is set; otherwise cleared. 2 Z Zero condition code bit. Set if the result equals zero; otherwise cleared. 1 V Overflow condition code bit. Set if an arithmetic overflow occurs, implying that the result cannot be represented in the operand size; otherwise cleared. 0 C Carry condition code bit. Set if a carry-out of the data operand msb occurs for an addition or if a borrow occurs in a subtraction; otherwise cleared. EMAC Programming Model The registers in the EMAC portion of the user programming model are described in Chapter 4, “Enhanced Multiply-Accumulate Unit (EMAC),” and include the following registers: • Four 48-bit accumulator registers partitioned as follows: — Four 32-bit accumulators (ACC0–ACC3) — Eight 8-bit accumulator extension bytes (two per accumulator). These are grouped into two 32-bit values for load and store operations (ACCEXT01 and ACCEXT23). Accumulators and extension bytes can be loaded, copied, and stored, and results from EMAC arithmetic operations generally affect the entire 48-bit destination. • Eight 8-bit accumulator extensions (two per accumulator), packaged as two 32-bit values for load and store operations (ACCext01 and ACCext23) • One 16-bit mask register (MASK) • One 32-bit status register (MACSR), including four indicator bits signaling product or accumulation overflow (one for each accumulator: PAV0–PAV3). These registers are shown in Figure 3-5. 31 0 MACSR ACC0 ACC1 ACC2 ACC3 ACCext01 ACCext23 MASK MAC status register MAC accumulator 0 MAC accumulator 1 MAC accumulator 2 MAC accumulator 3 Extensions for ACC0 and ACC1 Extensions for ACC2 and ACC3 MAC mask register Figure 3-5. EMAC Register Set 3.3.4 FPU Programming Model The registers in the FPU portion of the programming model are described in Chapter 6, “Floating-Point Unit (FPU),” and include the folllowing registers: MCF548x Reference Manual, Rev. 5 3-10 Freescale Semiconductor Programming Model • • • • Eight 64-bit floating-point data registers (FP0–FP7) One 32-bit floating-point control register (FPCR) One 32-bit floating-point status register (FPSR) One 32-bit floating-point instruction address register (FPIAR) Figure 3-6 shows the FPU programming model. 63 31 0 FP0 FP1 FP2 FP3 FP4 FP5 FP6 FP7 FPCR FPSR FPIAR Floating-point data registers Floating-point control register Floating-point status register Floating-point instruction address register Figure 3-6. Floating-Point Programmer’s Model 3.3.5 Supervisor Programming Model The MCF548x supervisor programming model is shown in Figure 3-3. Typically, system programmers use the supervisor programming model to implement operating system functions and provide memory and I/O control. The supervisor programming model provides access to the user registers and additional supervisor registers, which include the upper byte of the status register (SR), the vector base register (VBR), and registers for configuring attributes of the address space connected to the Version 4 processor core. Most supervisor-level registers are accessed by using the MOVEC instruction with the control register definitions in Table 3-2. Table 3-2. MOVEC Register Map Rc[11–0] Register Definition 0x002 Cache control register (CACR) 0x004 Access control register 0 (ACR0) 0x005 Access control register 1 (ACR1) 0x006 Access control register 2 (ACR2) 0x007 Access control register 3 (ACR3) 0x801 Vector base register (VBR) 0xC04 RAM base address register 0 (RAMBAR0) 0xC05 RAM base address register 1 (RAMBAR1) 0xC0F Module base address register (MBAR) MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 3-11 3.3.5.1 Status Register (SR) The SR stores the processor status, the interrupt priority mask, and other control bits. Supervisor software can read or write the entire SR; user software can read or write only SR[7–0], described in Section 3.3.2.2, “Condition Code Register (CCR).” The control bits indicate processor states—trace mode (T), supervisor or user mode (S), and master or interrupt state (M). SR is set to 0x27xx after reset. 15 14 13 12 11 10 9 8 7 6 System byte R T 0 S M 0 0 0 1 0 0 5 4 3 2 1 0 Condition code register (CCR) I 0 0 0 X N Z V C 0 0 0 — — — — — W Reset 1 Reg Addr 1 1 0x27xx Figure 3-7. Status Register (SR) Table 3-3 describes SR fields. Table 3-3. SR Field Descriptions Bits Name 15 T Trace enable. When T is set, the processor performs a trace exception after every instruction. 13 S Supervisor/user state. Indicates whether the processor is in supervisor or user mode 0 User mode 1 Supervisor mode 12 M Master/interrupt state. Cleared by an interrupt exception. It can be set by software during execution of the RTE or move to SR instructions so the OS can emulate an interrupt stack pointer. 10–8 I Interrupt priority mask. Defines the current interrupt priority. Interrupt requests are inhibited for all priority levels less than or equal to the current priority, except the edge-sensitive level-7 request, which cannot be masked. 7–0 CCR 3.3.5.2 Description Condition code register. See Table 3-1. Vector Base Register (VBR) The VBR holds the base address of the exception vector table in memory. The displacement of an exception vector is added to the value in this register to access the vector table. The VBR[19–0] bits are not implemented and are assumed to be zero, forcing the vector table to be aligned on a 0-modulo-1-Mbyte boundary. MCF548x Reference Manual, Rev. 5 3-12 Freescale Semiconductor Programming Model 31 30 R 29 28 27 26 25 24 Exception vector table base address 23 22 21 20 1 19 18 17 16 0 0 0 0 W Reset R 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 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 0 0 W Reset Reg Addr 0x801 1 Written from a BDM serial command or from the CPU using the MOVEC instruction. VBR can be read from the debug module only. The upper 12 bits are returned, the low-order 20 bits are undefined. Figure 3-8. Vector Base Register (VBR) 3.3.5.3 Cache Control Register (CACR) The CACR controls operation of both the instruction and data cache memory. It includes bits for enabling, freezing, and invalidating cache contents. It also includes bits for defining the default cache mode and write-protect fields. See Section 7.10.1, “Cache Control Register (CACR).” 3.3.5.4 Access Control Registers (ACR0–ACR3) The access control registers (ACR0–ACR3) define attributes for four user-defined memory regions: ACR0 and ACR1 control data memory space, and ACR2 and ACR3 control instruction memory space. Attributes include definition of cache mode, write protect and buffer write enables. See Section 7.10.2, “Access Control Registers (ACR0–ACR3).” 3.3.5.5 RAM Base Address Registers (RAMBAR0 and RAMBAR1) The RAMBAR registers determine the base address location of the internal SRAM modules and indicate the types of references mapped to each. Each RAMBAR includes a base address, write-protect bit, address space mask bits, and an enable. The RAM base address must be aligned on a 0-module-2-Kbyte boundary. See Section 7.4.1, “SRAM Base Address Registers (RAMBAR0/RAMBAR1).” 3.3.5.6 Module Base Address Register (MBAR) The module base address register (MBAR) defines the logical base address for the memory-mapped space containing the control registers for the on-chip peripherals. See Section 9.3.1, “Module Base Address Register (MBAR).” 3.3.6 Programming Model Table Table 3-4 lists register names, the CPU space location, whether the register is written from the processor using the MOVEC instruction, and the complete register name. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 3-13 Table 3-4. ColdFire CPU Registers Name CPU Space (Rc) Written with MOVEC Register Name Memory Management Control Registers CACR 0x002 Yes Cache control register ASID 0x003 Yes Address space identifier ACR0–ACR3 0x004–0x007 Yes Access control registers 0–3 MMUBAR Yes MMU base address register 0x008 Processor General-Purpose Registers D0–D7 0x(0,1)80–0x(0,1 )87 No Data registers 0–7 (0 = load, 1 = store) A0–A7 0x(0,1)88–0x(0,1 )8F No Address registers 0–7 (0 = load, 1 = store) A7 is user stack pointer Processor Miscellaneous Registers OTHER_A7 0x800 No Other stack pointer VBR 0x801 Yes Vector base register MACSR 0x804 No MAC status register MASK 0x805 No MAC address mask register ACC0–ACC3 0x806–0x80B No MAC accumulators 0–3 ACCext01 0x807 No MAC accumulator 0, 1 extension bytes ACCext23 0x808 No MAC accumulator 2, 3 extension bytes SR 0x80E No Status register PC 0x80F Yes Program counter Processor Floating-Point Registers FPU0 0x810 No 32 msbs of floating-point data register 0 FPL0 0x811 No 32 lsbs of floating-point data register 0 FPU1 0x812 No 32 msbs of floating-point data register 1 FPL1 0x813 No 32 lsbs of floating-point data register 1 FPU2 0x814 No 32 msbs of floating-point data register 2 FPL2 0x815 No 32 lsbs of floating-point data register 2 FPU3 0x816 No 32 msbs of floating-point data register 3 FPL3 0x817 No 32 lsbs of floating-point data register 3 FPU4 0x818 No 32 msbs of floating-point data register 4 FPL4 0x819 No 32 lsbs of floating-point data register 4 FPU5 0x81A No 32 msbs of floating-point data register 5 FPL5 0x81B No 32 lsbs of floating-point data register 5 FPU6 0x81C No 32 msbs of floating-point data register 6 MCF548x Reference Manual, Rev. 5 3-14 Freescale Semiconductor Data Format Summary Table 3-4. ColdFire CPU Registers (Continued) Name CPU Space (Rc) Written with MOVEC Register Name FPL6 0x81D No 32 lsbs of floating-point data register 6 FPU7 0x81E No 32 msbs of floating-point data register 7 FPL7 0x81F No 32 lsbs of floating-point data register 7 FPIAR 0x821 No Floating-point instruction address register FPSR 0x822 No Floating-point status register FPCR 0x824 No Floating-point control register Local Memory and Module Control Registers RAMBAR0 0xC04 Yes RAM base address register 0 RAMBAR1 0xC05 Yes RAM base address register 1 MBAR 0xC0F Yes Primary module base address register (not a core register) 3.4 Data Format Summary Table 3-5 lists the operand data formats. Integer operands can reside in registers, memory, or instructions. The operand size is either explicitly encoded in the instruction or implicitly defined by the instruction operation. Table 3-5. Integer Data Formats Operand Data Format 3.4.1 Size Bit 1 bit Byte integer 8 bits Word integer 16 bits Longword integer 32 bits Data Organization in Registers The following sections describe data organization in data, address, and control registers. Section 6.2.2, “Floating-Point Data Formats,” describes floating-point formatting. 3.4.1.1 Integer Data Format Organization in Registers Figure 3-9 shows the integer format for data registers. Each integer data register is 32 bits wide. Byte and word operands occupy the lower 8- and 16-bit portions of integer data registers, respectively. Longword operands occupy the entire 32 bits of integer data registers. A data register that is either a source or destination operand only uses or changes the appropriate lower 8 or 16 bits in byte or word operations, respectively. The remaining high-order portion does not change. Note that the least-significant bit is bit 0 for all data types, whereas the msbs for longword integer is bit 31, the msb of a word integer is bit 15, and the msb of a byte integer is bit 7. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 3-15 31 30 1 0 msb lsb 31 8 Not used 31 msb 6 1 0 msb Lower-order byte lsb 16 Not used 31 7 15 msb 14 1 Byte (8 bits) 0 Lower-order word 30 Bit (0 bit number 31) lsb 1 Word (16 bits) 0 Longword lsb Longword (32 bits) Figure 3-9. Organization of Integer Data Format in Data Registers Instruction encodings disallow use of address registers for byte operands. When an address register is a source operand, either the low-order word or the entire longword operand is used, depending on the operation size. Word-length source operands are sign-extended to 32 bits and then used in the operation with an address register destination. When an address register is a destination, the entire register is affected, regardless of the operation size. Figure 3-10 shows integer formats for address registers. 31 16 Sign-Extended 15 0 16-Bit Address Operand 31 0 Full 32-Bit Address Operand Figure 3-10. Organization of Integer Data Formats in Address Registers The size of control registers varies according to function. Some have undefined bits reserved for future definition by Freescale. Those bits read as zeros and must be written as zeros for future compatibility. Operations to the SR and CCR are word-sized. The upper CCR byte is read as all zeros and is ignored when written, regardless of privilege mode. 3.4.1.2 Integer Data Format Organization in Memory ColdFire processors use big-endian addressing. Byte-addressable memory organization allows lower addresses to correspond to higher-order bytes. The address N of a longword data item corresponds to the address of the high-order word. The lower-order word is at address N + 2. The address of a word data item corresponds to the address of the high-order byte. The lower-order byte is at address N + 1. This organization is shown in Figure 3-11. MCF548x Reference Manual, Rev. 5 3-16 Freescale Semiconductor Data Format Summary 31 24 23 16 15 8 7 0 Longword 0x0000_0000 . . . Word 0x0000_0000 Word 0x0000_0002 Byte 0x0000_0000 Byte 0x0000_0001 Byte 0x0000_0002 Byte 0x0000_0003 Longword 0x0000_0004 Word 0x0000_0004 Word 0x0000_0006 Byte 0x0000_0004 Byte 0x0000_0005 Byte 0x0000_0006 Byte 0x0000_0007 . . . Longword 0xFFFF_FFFC Word 0xFFFF_FFFC Word 0xFFFF_FFFE Byte 0xFFFF_FFFC Byte 0xFFFF_FFFD Byte 0xFFFF_FFFE Byte 0xFFFF_FFFF . . . Figure 3-11. Memory Operand Addressing 3.4.2 EMAC Data Representation The EMAC supports the following three modes, where each mode defines a unique operand type. • Two’s complement signed integer: In this format, an N-bit operand value lies in the range -2(N-1) < operand < 2(N-1) - 1. The binary point is right of the lsb. • Unsigned integer: In this format, an N-bit operand value lies in the range 0 < operand < 2N - 1. The binary point is right of the lsb. • Two’s complement, signed fractional: In an N-bit number, the first bit is the sign bit. The remaining bits signify the first N-1 bits after the binary point. Given an N-bit number, aN-1aN-2aN-3... a2a1a0, its value is given by the equation in Figure 3-12. N–2 value = – ( 1 ⋅ a N – 1 ) + ∑ 2 (i + 1 – N) ⋅ ai i=0 Figure 3-12. Two’s Complement, Signed Fractional Equation This format can represent numbers in the range -1 < operand < 1 - 2(N-1). For words and longwords, the largest negative number that can be represented is -1, whose internal representation is 0x8000 and 0x8000_0000, respectively. The largest positive word is 0x7FFF or (1 – 2-15); the most positive longword is 0x7FFF_FFFF or (1 – 2-31). For more information, see Chapter 4, “Enhanced Multiply-Accumulate Unit (EMAC).” 3.4.2.1 Floating-Point Data Formats and Types The FPU supports signed byte, word, and longword integer formats, which are identical to those supported by the integer unit. The FPU also supports single- and double-precision binary floating-point formats that fully comply with the IEEE-754 standard. For more information, see Chapter 6, “Floating-Point Unit (FPU).” MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 3-17 3.4.2.1.1 Signed-Integer Data Formats The FPU supports 8-bit byte (B), 16-bit word (W), and 32-bit longword (L) integer data formats. 3.4.2.1.2 Floating-Point Data Formats Figure 3-13 shows the two binary floating-point data formats. 31 S 63 S 62 30 8-Bit Exponent Sign of Mantissa 51 11-Bit Exponent 0 22 52-Bit Fraction 23-Bit Fraction Single 0 Double Sign of Mantissa Figure 3-13. Floating-Point Data Formats Note that, throughout this chapter, a mantissa is defined as the concatenation of an integer bit, the binary point, and a fraction. A fraction is the term designating the bits to the right of the binary point in the mantissa. Mantissa (integer bit).(fraction) Figure 3-14. Mantissa The integer bit is implied to be set for normalized numbers and infinities, clear for zeros and denormalized numbers. For not-a-numbers (NANs), the integer bit is ignored. The exponent in both floating-point formats is an unsigned binary integer with an implied bias added to it. Subtracting the bias from exponent yields a signed, two’s complement power of two. This represents the magnitude of a normalized floating-point number when multiplied by the mantissa. By definition, a normalized mantissa always takes values starting from 1.0 and going up to, but not including, 2.0; that is, [1.0...2.0). 3.5 Addressing Mode Summary Addressing modes are categorized by how they are used. Data addressing modes refer to data operands. Memory addressing modes refer to memory operands. Alterable addressing modes refer to alterable (writable) data operands. Control addressing modes refer to memory operands without an associated size. These categories sometimes combine to form more restrictive categories. Two combined classifications are alterable memory (both alterable and memory) and data alterable (both alterable and data). Twelve of the most commonly used effective addressing modes from the M68000 Family are available on ColdFire microprocessors. Table 3-6 summarizes these modes and their categories. MCF548x Reference Manual, Rev. 5 3-18 Freescale Semiconductor Instruction Set Summary Table 3-6. ColdFire Effective Addressing Modes Addressing Modes Mode Field Reg. Field Dn An 000 001 (An) (An)+ –(An) (d16, An) Syntax Category Data Memory Control Alterable reg. no. reg. no. X — — — — — X X 010 011 100 101 reg. no. reg. no. reg. no. reg. no. X X X X X X X X X — — X X X X X (d8, An, Xi*SF) 110 reg. no. X X X X Program counter indirect with displacement (d16, PC) 111 010 X X X — Program counter indirect with scaled index 8-bit displacement (d8, PC, Xi*SF) 111 011 X X X — Absolute data addressing Short Long (xxx).W (xxx).L 111 111 000 001 X X X X X X — — Immediate #<xxx> 111 100 X X — — Register direct Data Address Register indirect Address Address with Postincrement Address with Predecrement Address with Displacement Address register indirect with scaled index 8-bit displacement 3.6 Instruction Set Summary The ColdFire instruction set is a simplified version of the M68000 instruction set. The removed instructions include BCD, bit field, logical rotate, decrement and branch, and integer multiply with a 64-bit result. “About This Book” lists notational conventions used throughout this manual. 3.6.1 Additions to the Instruction Set Architecture The original ColdFire ISA was derived from M68000 Family opcodes based on extensive analysis of embedded application code. After the first ColdFire compilers were created, developers identified ISA additions that would enhance both code density and overall performance. Additionally, as users implemented ColdFire-based designs into a wide range of embedded systems, they identified frequently used instruction sequences that could be improved by creating new instructions. This observation was especially prevalent in environments that used substantial amounts of assembly language code. The original ISA minimized support for instructions referencing byte and word operands. MOVE.B and MOVE.W were fully supported; otherwise, only CLR (clear) and TST (test) supported these data types. Based on input from compiler writers and system users, a set of instruction enhancements was proposed to address the following: MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 3-19 • • Enhanced support for byte and word-sized operands through new move operations Enhanced support for position-independent code For descriptions of the ColdFire instruction set, see the latest version of the ColdFire Programmer’s Reference Manual. The following list summarizes new and enhanced instructions of ISA_B: • New instructions: — INTOUCH loads blocks of instructions to be locked in the instruction cache. — MOV3Q.L moves 3-bit immediate data to the destination location. — MOVE to/from USP loads and stores user stack pointer. — MVS.{B,W} sign-extends the source operand and moves it to the destination register. — MVZ.{B,W} zero-fills the source operand and moves it to the destination register. — SATS.L performs a saturation operation for signed arithmetic and updates the destination register depending on CCR[V] and bit 31 of the register. — TAS.B performs an indivisible read-modify-write cycle to test and set the addressed memory byte. • Enhancements to existing Revision_A instructions: — Longword support for branch instructions (Bcc, BRA, BSR) — Byte and word support for compare instructions (CMP, CMPI) — Word support for the compare address register instruction (CMPA) — Byte and longword support for MOVE.x,where the source is immediate data and the destination is specified by d16(Ax); that is, MOVE.{B,W} #<data>, d16(Ax) • Floating-point instructions. See Chapter 6, “Floating-Point Unit (FPU).” • EMAC instructions. See Chapter 4, “Enhanced Multiply-Accumulate Unit (EMAC),” for more information. Table 3-7 shows the syntax for the new and enhanced instructions. As Table 3-7 shows, some ISA_B opcodes were defined in the M68000 family and others are new. Table 3-7. V4 New Instruction Summary Instruction Mnemonic1 Source Destination M68000 ISA_B Extensions Branch Always bra.l <label> Yes Branch Conditionally bcc.l <label> Yes Branch to Subroutine bsr.l <label> Yes Compare cmp.{b,w,l} <ea>y Dx Yes cmpa.w <ea>y Ax Yes cmpi.{b,w} #<data> Dx Yes Instruction Fetch Touch intouch <Ay> Move 3-Bit Data Quick mov3q.l #<data> <ea>x move.{b,w} #<data> d16(Ax) Yes move.l USP Ax Yes Compare Address Compare Immediate Move Data Source to Destination Move from USP MCF548x Reference Manual, Rev. 5 3-20 Freescale Semiconductor Instruction Set Summary Table 3-7. V4 New Instruction Summary (Continued) Mnemonic1 Source Destination M68000 move.l Ay USP Yes Move with Sign Extend mvs.{b,w} <ea>y Dx Move with Zero-Fill mvz.{b,w} <ea>y Dx Instruction Move to USP Signed Saturate sats.l Dx Test and Set an Operand tas.b <ea>x Yes EMAC Extensions Move from an Accumulator and Clear movclr.l ACCx Rx No Copy an Accumulator move.l ACCy ACCx No Move from Accumulator 0 and 1 Extensions move.l ACCext01 Rx No Move from Accumulator 2 and 3 Extensions move.l ACCext23 Rx No Move to Accumulator 0 and 1 Extensions move.l Ry ACCext01 No Move to Accumulator 2 and 2 Extensions move.l Ry ACCext23 No FPU Instructions Floating-Point Absolute Value fabs.{b,w,l,s,d} <ea>y FPx Yes Floating-Point Add fadd.{b,w,l,s,d} <ea>y FPx Yes <label> Yes Floating-Point Branch Conditionally Floating-Point Compare fbcc.{w,l} fcmp.{b,w,l,s,d} <ea>y FPx Yes Floating-Point Divide fdiv.{b,w,l,s,d} <ea>y FPx Yes Floating-Point Integer fint.{b,w,l,s,d} <ea>y FPx Yes Floating-Point Integer Round-to-Zero fintrz.{b,w,l,s,d} <ea>y FPx Yes Move Floating-Point Data Register fmove.{b,w,l,s,d} <ea>y FPx Yes Move from FPCR fmove.l FPCR <ea>x Yes Move from FPIAR fmove.l FPIAR <ea>x Yes Move from FPSR fmove.l FPSR <ea>x Yes Move from FPCR fmove.l <ea>y FPCR Yes Move from FPIAR fmove.l <ea>y FPIAR Yes Move from FPSR fmove.l <ea>y FPSR Yes fmovem.d #list <ea>y <ea>x #list Yes Floating-Point Multiply fmul.{b,w,l,s,d} <ea>y FPx Yes Floating-Point Negate fneg.{b,w,l,s,d} <ea>y FPx Yes Move Multiple Floating Point Data Registers Floating-Point No Operation fnop Restore Internal Floating Point State frestore Yes <ea>y Yes MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 3-21 Table 3-7. V4 New Instruction Summary (Continued) Mnemonic1 Instruction Save Internal Floating Point State Source Destination M68000 <ea>x Yes fsave Floating-Point Square Root fsqrt.{b,w,l,s,d} <ea>y FPx Yes Floating-Point Subtract fsub.{b,w,l,s,d} <ea>y FPx Yes Test Floating-Point Operand ftst.{b,w,l,s,d} <ea>y 1 Yes Operand sizes in this column reflect only newly supported operand sizes for existing instructions (Bcc, BRA, BSR, CMP, CMPA, CMPI, and MOVE) 3.6.2 Instruction Set Summary Table 3-8 lists user-mode instructions by opcode. Table 3-8. User-Mode Instruction Set Summary Instruction Operand Syntax Operand Size Operation ADD L L L Source + Destination → Destination ADDA Dy,<ea>x <ea>y,Dx <ea>y,Ax ADDI ADDQ #<data>,Dx #<data>,<ea>x L L Immediate Data + Destination → Destination ADDX Dy,Dx L Source + Destination + CCR[X] → Destination AND <ea>y,Dx Dy,<ea>x L L Source & Destination → Destination ANDI #<data>, Dx L Immediate Data & Destination → Destination ASL Dy,Dx #<data>,Dx L L CCR[X,C] ← (Dx << Dy) ← 0 CCR[X,C] ← (Dx << #<data>) ← 0 ASR Dy,Dx #<data>,Dx L L msb → (Dx >> Dy) → CCR[X,C] msb → (Dx >> #<data>) → CCR[X,C Bcc <label> B, W, L If Condition True, Then PC + dn → PC BCHG Dy,<ea>x #<data>,<ea>x B, L B, L ~ (<bit number> of Destination) → CCR[Z] → <bit number> of Destination BCLR Dy,<ea>x #<data>,<ea>x B, L B, L ~ (<bit number> of Destination) → CCR[Z]; 0 →<bit number> of Destination BRA <label> B, W, L BSET Dy,<ea>x #<data>,<ea>x B, L B, L BSR <label> B, W, L BTST Dy,<ea>x #<data>,<ea>x B, L B, L CLR <ea>x B, W, L PC + dn → PC ~ (<bit number> of Destination) → CCR[Z]; 1 → <bit number> of Destination SP – 4 → SP; nextPC → (SP); PC + dn → PC ~ (<bit number> of Destination) → CCR[Z] 0 → Destination MCF548x Reference Manual, Rev. 5 3-22 Freescale Semiconductor Instruction Set Summary Table 3-8. User-Mode Instruction Set Summary (Continued) Instruction Operand Syntax Operand Size CMP CMPA <ea>y,Dx <ea>y,Ax B, W, L W, L Destination – Source → CCR CMPI #<data>,Dx B, W, L Destination – Immediate Data → CCR DIVS/DIVU <ea>y,Dx W, L Destination / Source → Destination (Signed or Unsigned) EOR Dy,<ea>x L Source ^ Destination → Destination EORI #<data>,Dx L Immediate Data ^ Destination → Destination EXT Dx Dx Dx B→W W→L B→L <ea>y,FPx FPy,FPx FPx B,W,L,S,D D D Absolute Value of Source → FPx FADD <ea>y,FPx FPy,FPx B,W,L,S,D D Source + FPx → FPx FBcc <label> W, L FCMP <ea>y,FPx FPy,FPx B,W,L,S,D D FPx - Source FDABS <ea>y,FPx FPy,FPx FPx B,W,L,S,D D D Absolute Value of Source → FPx; round destination to double Absolute Value of FPx → FPx; round destination to double FDADD <ea>y,FPx FPy,FPx B,W,L,S,D D Source + FPx → FPx; round destination to double FDDIV <ea>y,FPx FPy,FPx B,W,L,S,D D FPx / Source → FPx; round destination to double FDIV <ea>y,FPx FPy,FPx B,W,L,S,D D FPx / Source → FPx FDMOVE FPy,FPx D Source → Destination; round destination to double FDMUL <ea>y,FPx FPy,FPx B,W,L,S,D D Source * FPx → FPx; round destination to double FDNEG <ea>y,FPx FPy,FPx FPx B,W,L,S,D D D - (Source) → FPx; round destination to double FDSQRT <ea>y,FPx FPy,FPx FPx B,W,L,S,D D D Square Root of Source → FPx; round destination to double Square Root of FPx → FPx; round destination to double FDSUB <ea>y,FPx FPy,FPx B,W,L,S,D D FPx - Source → FPx; round destination to double EXTB FABS Operation Sign-Extended Destination → Destination Absolute Value of FPx → FPx If Condition True, Then PC + dn → PC - (FPx) → FPx; round destination to double MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 3-23 Table 3-8. User-Mode Instruction Set Summary (Continued) Instruction Operand Syntax Operand Size FINT <ea>y,FPx FPy,FPx FPx B,W,L,S,D D D Integer Part of Source → FPx <ea>y,FPx FPy,FPx FPx B,W,L,S,D D D Integer Part of Source → FPx; round to zero <ea>y,FPx FPy,<ea>x FPy,FPx FPcr,<ea>x <ea>y,FPcr B,W,L,S,D B,W,L,S,D D L L Source → Destination FMOVEM #list,<ea>x <ea>y,#list D FMUL <ea>y,FPx FPy,FPx B,W,L,S,D D Source * FPx → FPx FNEG <ea>y,FPx FPy,FPx FPx B,W,L,S,D D D - (Source) → FPx FNOP none none FSABS <ea>y,FPx FPy,FPx FPx B,W,L,S,D D D Absolute Value of Source → FPx; round destination to single Absolute Value of FPx → FPx; round destination to single FSADD <ea>y,FPx FPy,FPx B,W,L,S,D Source + FPx → FPx; round destination to single FSDIV <ea>y,FPx FPy,FPx B,W,L,S,D D FPx / Source → FPx; round destination to single FSMOVE <ea>y,FPx B,W,L,S,D Source → Destination; round destination to single FSMUL <ea>y,FPx FPy,FPx B,W,L,S,D D Source * FPx → FPx; round destination to single FSNEG <ea>y,FPx FPy,FPx FPx B,W,L,S,D D D - (Source) → FPx; round destination to single <ea>y,FPx FPy,FPx FPx B,W,L,S,D D D Square Root of Source → FPx FSSQRT <ea>y,FPx FPy,FPx FPx B,W,L,S,D D D Square Root of Source → FPx; round destination to single Square Root of FPx → FPx; round destination to single FSSUB <ea>y,FPx FPy,FPx B,W,L,S,D D FPx - Source → FPx; round destination to single FINTRZ FMOVE FSQRT Operation Integer Part of FPx → FPx Integer Part of FPx → FPx; round to zero FPcr can be any floating point control register: FPCR, FPIAR, FPSR Listed registers → Destination Source → Listed registers - (FPx) → FPx PC + 2 → PC (FPU Pipeline Synchronized) - (FPx) → FPx; round destination to single Square Root of FPx → FPx MCF548x Reference Manual, Rev. 5 3-24 Freescale Semiconductor Instruction Set Summary Table 3-8. User-Mode Instruction Set Summary (Continued) Instruction Operand Syntax Operand Size FSUB <ea>y,FPx FPy,FPx B,W,L,S,D D FTST <ea>y B, W, L, S, D ILLEGAL none none SP – 4 → SP; PC → (SP) → PC; SP – 2 → SP; SR → (SP); SP – 2 → SP; Vector Offset → (SP); (VBR + 0x10) → PC JMP <ea>y none Source Address → PC JSR <ea>y none SP – 4 → SP; nextPC → (SP); Source → PC LEA <ea>y,Ax L <ea>y → Ax LINK Ay,#<displacement> W SP – 4 → SP; Ay → (SP); SP → Ay, SP + dn → SP LSL Dy,Dx #<data>,Dx L L CCR[X,C] ← (Dx << Dy) ← 0 CCR[X,C] ← (Dx << #<data>) ← 0 LSR Dy,Dx #<data>,Dx L L 0 → (Dx >> Dy) → CCR[X,C] 0 → (Dx >> #<data>) → CCR[X,C] MAC Ry,RxSF,ACCx Ry,RxSF,<ea>y,Rw,ACCx W, L W, L MOV3Q #<data>,<ea>x L Immediate Data → Destination MOVCLR ACCy,Rx L Accumulator → Destination, 0 → Accumulator MOVE <ea>y,<ea>x MACcr,Dx <ea>y,MACcr CCR,Dx <ea>y,CCR B,W,L L L W W Source → Destination where MACcr can be any MAC control register: ACCx, ACCext01, ACCext23, MACSR, MASK MOVEA <ea>y,Ax W,L → L MOVEM #list,<ea>x <ea>y,#list L Listed Registers → Destination Source → Listed Registers MOVEQ #<data>,Dx B→L Immediate Data → Destination MSAC Ry,RxSF,ACCx Ry,RxSF,<ea>y,Rw,ACCx W, L W, L MULS/MULU <ea>y,Dx W*W→L L*L→L MVS <ea>y,Dx B,W Source with sign extension → Destination MVZ <ea>y,Dx B,W Source with zero fill → Destination NEG Dx L 0 – Destination → Destination NEGX Dx L 0 – Destination – CCR[X] → Destination NOP none none MOVE from CCR MOVE to CCR Operation FPx - Source → FPx Source Operand Tested → FPCC ACCx + (Ry * Rx){<<|>>}SF → ACCx ACCx + (Ry * Rx){<<|>>}SF → ACCx; (<ea>y(&MASK)) → Rw Source → Destination ACCx - (Ry * Rx){<<|>>}SF → ACCx ACCx - (Ry * Rx){<<|>>}SF → ACCx; (<ea>y(&MASK)) → Rw Source * Destination → Destination (Signed or Unsigned) PC + 2 → PC (Integer Pipeline Synchronized) MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 3-25 Table 3-8. User-Mode Instruction Set Summary (Continued) Instruction Operand Syntax Operand Size Operation NOT Dx L ~ Destination → Destination OR <ea>y,Dx Dy,<ea>x L L Source | Destination → Destination ORI #<data>,Dx L Immediate Data | Destination → Destination PEA <ea>y L SP – 4 → SP; <ea>y → (SP) PULSE none none REMS/REMU <ea>y,Dw:Dx L RTS none none SATS Dx L If CCR[V] == 1; then if Dx[31] == 0; then Dx[31:0] = 0x80000000; else Dx[31:0] = 0x7FFFFFFF; else Dx[31:0] is unchanged Scc Dx B If Condition True, Then 1s → Destination; Else 0s → Destination SUB L L L Destination - Source → Destination SUBA <ea>y,Dx Dy,<ea>x <ea>y,Ax SUBI SUBQ #<data>,Dx #<data>,<ea>x L L Destination – Immediate Data → Destination SUBX Dy,Dx L Destination – Source – CCR[X] → Destination SWAP Dx W MSW of Dx ↔ LSW of Dx TAS <ea>x B Destination Tested → CCR; 1 → bit 7 of Destination TPF none #<data> #<data> none W L PC + 2→ PC PC + 4 → PC PC + 6→ PC TRAP #<vector> none 1 → S Bit of SR; SP – 4 → SP; nextPC → (SP); SP – 2 → SP; SR → (SP) SP – 2 → SP; Format/Offset → (SP) (VBR + 0x80 +4*n) → PC, where n is the TRAP number TST <ea>y B, W, L UNLK Ax none WDDATA <ea>y B, W, L Set PST = 0x4 Destination / Source → Remainder (Signed or Unsigned) (SP) → PC; SP + 4 → SP Source Operand Tested → CCR Ax → SP; (SP) → Ax; SP + 4 → SP Source → DDATA port Table 3-9 describes supervisor-mode instructions. MCF548x Reference Manual, Rev. 5 3-26 Freescale Semiconductor Instruction Execution Timing Table 3-9. Supervisor-Mode Instruction Set Summary 3.7 Instruction Operand Syntax Operand Size Operation CPUSHL ic,(Ax) dc,(Ax) bc,(Ax) none If data is valid and modified, push cache line; invalidate line if programmed in CACR (synchronizes pipeline) FRESTORE <ea>y none FPU State Frame → Internal FPU State FSAVE <ea>x none Internal FPU State → FPU State Frame HALT none none Halt processor core INTOUCH Ay none Instruction fetch touch at (Ay) MOVE from SR SR,Dx W SR → Destination MOVE from USP USP,Dx L USP → Destination MOVE to SR <ea>y,SR W Source → SR; Dy or #<data> source only MOVE to USP Ay,USP L Source → USP MOVEC Ry,Rc L Ry → Rc RTE none none 2 (SP) → SR; 4 (SP) → PC; SP + 8 →SP Adjust stack according to format STOP #<data> none Immediate Data → SR; STOP WDEBUG <ea>y L Addressed Debug WDMREG Command Executed Instruction Execution Timing The timing data in this section assumes the following: • Execution times for individual instructions make no assumptions concerning the OEP’s ability to dispatch multiple instructions in one machine cycle. For sequences where instruction pairs are issued, the execution time of the first instruction defines the execution time of pair; the second instruction effectively executes in zero cycles. • The OEP is loaded with the opword and all required extension words at the beginning of each instruction execution. This implies that the OEP spends no time waiting for the IFP to supply opwords or extension words. • The OEP experiences no sequence-related pipeline stalls. For the V4, the most common example of this type of stall occurs when a register is modified in the EX engine and a subsequent instruction generates an address that uses the previously modified register. The second instruction stalls in the OEP until the previous instruction updates the register. For example: muls.l move.l #<data>,d0 (a0,d0.l*4),d1 move.l waits 3 cycles for the muls.l to update d0. If consecutive instructions update a register and use that register as a base of index value with a scale factor of 1 (Xi.l*1) in an address calculation, a 2-cycle pipeline stall occurs. If the destination register is used as an index register with any other scale factor (Xi.l*2, Xi.l*4), a 3-cycle stall occurs. NOTE Address register results from postincrement and predecrement modes are available to subsequent instructions without stalls. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 3-27 • The OEP can complete all memory accesses without memory causing any stalls. Thus, these timings assume an infinite, zero-wait state memory attached to the core. Operand accesses are assumed to be aligned as follows: — 16-bit operands are aligned on 0-modulo-2 addresses — 32-bit operands are aligned on 0-modulo-4 addresses Operands that do not meet these guidelines are misaligned. Table 3-10 shows how the core decomposes a misaligned operand reference into a series of aligned accesses. • Table 3-10. Misaligned Operand References 1 Additional C(R/W)1 A[1:0] Size Bus Operations x1 Word Byte, Byte 2(1/0) if read 1(0/1) if write x1 Long Byte, Word, Byte 3(2/0) if read 2(0/2) if write 10 Long Word, Word 2(1/0) if read 1(0/1) if write Each timing entry is presented as C(r/w), described as follows: C is the number of processor clock cycles, including all applicable operand fetches and writes, as well as all internal core cycles required to complete the instruction execution. r/w is the number of operand reads (r) and writes (w) required by the instruction. An operation performing a read-modify write function is denoted as (1/1). 3.7.1 MOVE Instruction Execution Timing The following tables show execution times for the MOVE.{B,W,L} instructions. Table 3-13 shows the timing for the other generic move operations. NOTE In these tables, times using PC-relative effective addressing modes are the same as using An-relative mode. ET with {<ea> = (d16,PC)} equals ET with {<ea> = (d16,An)} ET with {<ea> = (d8,PC,Xi*SF)} equals ET with {<ea> = (d8,An,Xi*SF)} The (xxx).wl nomenclature refers to both forms of absolute addressing, (xxx).w and (xxx).l. Table 3-11 lists execution times for MOVE.{B,W} instructions. Table 3-11. Move Byte and Word Execution Times Destination Source Rx (Ax) (Ax)+ –(Ax) (d16,Ax) (d8,Ax,Xi*SF) (xxx).wl Dy 1(0/0) 1(0/1) 1(0/1) 1(0/1) 1(0/1) 2(0/1) 1(0/1) Ay 1(0/0) 1(0/1) 1(0/1) 1(0/1) 1(0/1) 2(0/1) 1(0/1) (Ay) 1(1/0) 2(1/1) 2(1/1) 2(1/1) 2(1/1) 3(1/1) 2(1/1) MCF548x Reference Manual, Rev. 5 3-28 Freescale Semiconductor Instruction Execution Timing Table 3-11. Move Byte and Word Execution Times (Continued) Destination Source Rx (Ax) (Ax)+ –(Ax) (d16,Ax) (d8,Ax,Xi*SF) (xxx).wl (Ay)+ 1(1/0) 2(1/1) 2(1/1) 2(1/1) 2(1/1) 3(1/1) 2(1/1) -(Ay) 1(1/0) 21/1) 2(1/1) 2(1/1) 2(1/1) 3(1/1) 2(1/1) (d16,Ay) 1(1/0) 2(1/1) 2(1/1) 2(1/1) 2(1/1) — — (d8,Ay,Xi*SF) 2(1/0) 3(1/1) 3(1/1) 3(1/1) — — — (xxx).w 1(1/0) 2(1/1) 2(1/1) 2(1/1) — — — (xxx).l 1(1/0) 2(1/1) 2(1/1) 2(1/1) — — — (d16,PC) 1(1/0) 2(1/1) 2(1/1) 2(1/1) 2(1/1) — — (d8,PC,Xi*SF) 2(1/0) 3(1/1) 3(1/1) 3(1/1) — — — #<xxx> 1(0/0) 1(0/1) 1(0/1) 1(0/1) 1(0/1) — — Table 3-12 lists timings for MOVE.L. Table 3-12. Move Long Execution Times Destination Source Rx (Ax) (Ax)+ –(Ax) (d16,Ax) (d8,Ax,Xi*SF) (xxx).wl Dy 1(0/0) 1(0/1) 1(0/1) 1(0/1) 1(0/1) 2(0/1) 1(0/1) Ay 1(0/0) 1(0/1) 1(0/1) 1(0/1) 1(0/1) 2(0/1) 1(0/1) (Ay) 1(1/0) 2(1/1) 2(1/1) 2(1/1) 2(1/1) 3(1/1) 2(1/1) (Ay)+ 1(1/0) 2(1/1) 2(1/1) 2(1/1) 2(1/1) 3(1/1) 2(1/1) -(Ay) 1(1/0) 2(1/1) 2(1/1) 2(1/1) 2(1/1) 3(1/1) 2(1/1) (d16,Ay) 1(1/0) 2(1/1) 2(1/1) 2(1/1) 2(1/1) — — (d8,Ay,Xi*SF) 2(1/0) 3(1/1) 3(1/1) 3(1/1) — — — (xxx).w 1(1/0) 2(1/1) 2(1/1) 2(1/1) — — — (xxx).l 1(1/0) 2(1/1) 2(1/1) 2(1/1) — — — (d16,PC) 1(1/0) 2(1/1) 2(1/1) 2(1/1) 2(1/1) — — (d8,PC,Xi*SF) 2(1/0) 3(1/1) 3(1/1) 3(1/1) — — — #<xxx> 1(0/0) 1(0/1) 1(0/1) 1(0/1) — — — Table 3-13 gives timings for MOVE.L instructions accessing program-visible EMAC registers, along with other MOVE.L timings. Execution times for moving ACC or MACSR contents into a destination location represent the best-case scenario when the store instruction is executed and no load, MAC, or MSAC instructions are in the EMAC execution pipeline. In general, these store operations take only 1 cycle to execute, but if preceded immediately by a load, MAC, or MSAC instruction, the EMAC pipeline depth is exposed and execution time is 3 cycles. Table 3-19 lists EMAC execution times. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 3-29 Table 3-13. MAC and Miscellaneous Move Execution Times Effective Address Opcode <ea> Rn (An) (An)+ –(An) (d16,An) (d8,An,Xi*SF) (xxx).wl #<xxx> move.l <ea>,ACC 1(0/0) — — — — — — 1(0/0) move.l <ea>,MACSR 6(0/0) — — — — — — 6(0/0) move.l <ea>,MASK 5(0/0) — — — — — — 5(0/0) move.l ACC,Rx 1(0/0) — — — — — — — move.l MACSR,CCR 1(0/0) — — — — — — — move.l MACSR,Rx 1(0/0) — — — — — — — move.l MASK,Rx 1(0/0) — — — — — — — moveq #imm,Dx — — — — — — — 1(0/0) mov3q #imm,<ea> 1(0/0) 1(1/0) 1(1/0) 1(1/0) 1(1/0) 2(1/0) 1(1/0) — mvs <ea>,Dx 1(0/0) 1(1/0) 1(1/0) 1(1/0) 1(1/0) 2(1/0) 1(1/0) 1(0/0) mvz <ea>,Dx 1(0/0) 1(1/0) 1(1/0) 1(1/0) 1(1/0) 2(1/0) 1(1/0) 1(0/0) 3.7.2 One-Operand Instruction Execution Timing Table 3-14 shows standard timings for single-operand instructions. Table 3-14. One-Operand Instruction Execution Times Effective Address Opcode <ea> Rn (An) (An)+ –(An) (d16,An) (d8,An,Xi*SF) (xxx).wl #xxx clr.b <ea> 1(0/0) 1(0/1) 1(0/1) 1(0/1) 1(0/1) 2(0/1) 1(0/1) — clr.w <ea> 1(0/0) 1(0/1) 1(0/1) 1(0/1) 1(0/1) 2(0/1) 1(0/1) — clr.l <ea> 1(0/0) 1(0/1) 1(0/1) 1(0/1) 1(0/1) 2(0/1) 1(0/1) — ext.w Dx 1(0/0) — — — — — — — ext.l Dx 1(0/0) — — — — — — — extb.l Dx 1(0/0) — — — — — — — neg.l Dx 1(0/0) — — — — — — — negx.l Dx 1(0/0) — — — — — — — not.l Dx 1(0/0) — — — — — — — sats.l Dx 1(0/0) — — — — — — — scc Dx 1(0/0) — — — — — — — swap Dx 1(0/0) — — — — — — — tas <ea> 1(1/1) 1(1/1) 1(1/1) 1(1/1) 1(1/1) 2(1/1) 1(1/1) — tst.b <ea> 1(0/0) 1(1/0) 1(1/0) 1(1/0) 1(1/0) 2(1/0) 1(1/0) 1(0/0) tst.w <ea> 1(0/0) 1(1/0) 1(1/0) 1(1/0) 1(1/0) 2(1/0) 1(1/0) 1(0/0) tst.l <ea> 1(0/0) 1(1/0) 1(1/0) 1(1/0) 1(1/0) 2(1/0) 1(1/0) 1(0/0) MCF548x Reference Manual, Rev. 5 3-30 Freescale Semiconductor Instruction Execution Timing 3.7.3 Two-Operand Instruction Execution Timing Table 3-15 shows standard timings for double operand instructions. Table 3-15. Two-Operand Instruction Execution Times Effective Address Opcode <ea> Rn (An) (An)+ –(An) (d16,An) (d8,An,Xi*SF) (xxx).wl #<xxx> add.l <ea>,Rx 1(0/0) 1(1/0) 1(1/0) 1(1/0) 1(1/0) 2(1/0) 1(1/0) 1(0/0) add.l Dy,<ea> — 1(1/1) 1(1/1) 1(1/1) 1(1/1) 2(1/1) 1(1/1) — addi.l #imm,Dx 1(0/0) — — — — — — — addq.l #imm,<ea> 1(0/0) 1(1/1) 1(1/1) 1(1/1) 1(1/1) 2(1/1) 1(1/1) — addx.l Dy,Dx 1(0/0) — — — — — — — and.l <ea>,Rx 1(0/0) 1(1/0) 1(1/0) 1(1/0) 1(1/0) 2(1/0) 1(1/0) 1(0/0) and.l Dy,<ea> — 1(1/1) 1(1/1) 1(1/1) 1(1/1) 2(1/1) 1(1/1) — andi.l #imm,Dx 1(0/0) — — — — — — — asl.l <ea>,Dx 1(0/0) — — — — — — 1(0/0) asr.l <ea>,Dx 1(0/0) — — — — — — 1(0/0) bchg Dy,<ea> 2(0/0) 2(1/1) 2(1/1) 2(1/1) 2(1/1) 3(1/1) 2(1/1) — bchg #imm,<ea> 2(0/0) 2(1/1) 2(1/1) 2(1/1) 2(1/1) — — — bclr Dy,<ea> 2(0/0) 2(1/1) 2(1/1) 2(1/1) 2(1/1) 3(1/1) 2(1/1) — bclr #imm,<ea> 2(0/0) 2(1/1) 2(1/1) 2(1/1) 2(1/1) — — — bset Dy,<ea> 2(0/0) 2(1/1) 2(1/1) 2(1/1) 2(1/1) 3(1/1) 2(1/1) — bset #imm,<ea> 2(0/0) 2(1/1) 2(1/1) 2(1/1) 2(1/1) — — — btst Dy,<ea> 1(0/0) 1(1/0) 1(1/0) 1(1/0) 1(1/0) 2(1/0) 1(1/0) — btst #imm,<ea> 1(0/0) 1(1/0) 1(1/0) 1(1/0) 1(1/0) — — — cmp.b <ea>,Rx 1(0/0) 1(1/0) 1(1/0) 1(1/0) 1(1/0) 2(1/0) 1(1/0) 1(0/0) cmp.w <ea>,Rx 1(0/0) 1(1/0) 1(1/0) 1(1/0) 1(1/0) 2(1/0) 1(1/0) 1(0/0) cmp.l <ea>,Rx 1(0/0) 1(1/0) 1(1/0) 1(1/0) 1(1/0) 2(1/0) 1(1/0) 1(0/0) cmpi.b #imm,Dx 1(0/0) — — — — — — — cmpi.w #imm,Dx 1(0/0) — — — — — — — cmpi.l #imm,Dx 1(0/0) — — — — — — — divs.w <ea>,Dx 20(0/0) 20(1/0) 20(1/0) 20(1/0) 20(1/0) 21(1/0) 20(1/0) 20(0/0) divu.w <ea>,Dx 20(0/0) 20(1/0) 20(1/0) 20(1/0) 20(1/0) 21(1/0) 20(1/0) 20(0/0) divs.l <ea>,Dx 35(0/0) 35(1/0) 35(1/0) 35(1/0) 35(1/0) — — — divu.l <ea>,Dx 35(0/0) 35(1/0) 35(1/0) 35(1/0) 35(1/0) — — — eor.l Dy,<ea> 1(0/0) 1(1/1) 1(1/1) 1(1/1) 1(1/1) 2(1/1) 1(1/1) — eori.l #imm,Dx 1(0/0) — — — — — — — lea <ea>,Ax — 1(0/0) — — 1(0/0) 2(0/0) 1(0/0) — lsl.l <ea>,Dx 1(0/0) — — — — — — 1(0/0) MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 3-31 Table 3-15. Two-Operand Instruction Execution Times (Continued) Effective Address Opcode <ea> Rn (An) (An)+ –(An) (d16,An) (d8,An,Xi*SF) (xxx).wl #<xxx> lsr.l <ea>,Dx 1(0/0) — — — — — — 1(0/0) mac.w Ry,Rx 1(0/0) — — — — — — — mac.l Ry,Rx 3(0/0) — — — — — — — msac.w Ry,Rx 1(0/0) — — — — — — — msac.l Ry,Rx 3(0/0) — — — — — — — mac.w Ry,Rx,ea,Rw — 1(1/0) 1(1/0) 1(1/0) 1(1/0) — — — mac.l Ry,Rx,ea,Rw — 3(1/0) 3(1/0) 3(1/0) 3(1/0) — — — msac.w Ry,Rx,ea,Rw — 1(1/0) 1(1/0) 1(1/0) 1(1/0) — — — msac.l Ry,Rx,ea,Rw — 3(1/0) 3(1/0) 3(1/0) 3(1/0) — — — muls.w <ea>,Dx 3(0/0) 3(1/0) 3(1/0) 3(1/0) 3(1/0) 4(1/0) 3(1/0) 3(0/0) mulu.w <ea>,Dx 3(0/0) 3(1/0) 3(1/0) 3(1/0) 3(1/0) 4(1/0) 3(1/0) 3(0/0) muls.l <ea>,Dx 5(0/0) 5(1/0) 5(1/0) 5(1/0) 5(1/0) — — — mulu.l <ea>,Dx 5(0/0) 5(1/0) 5(1/0) 5(1/0) 5(1/0) — — — or.l <ea>,Rx 1(0/0) 1(1/0) 1(1/0) 1(1/0) 1(1/0) 2(1/0) 1(1/0) 1(0/0) or.l Dy,<ea> — 1(1/1) 1(1/1) 1(1/1) 1(1/1) 2(1/1) 1(1/1) — or.l #imm,Dx 1(0/0) — — — — — — — rems.l <ea>,Dx 35(0/0) 35(1/0) 35(1/0) 35(1/0) 35(1/0) — — — remu.l <ea>,Dx 35(0/0) 35(1/0) 35(1/0) 35(1/0) 35(1/0) — — — sub.l <ea>,Rx 1(0/0) 1(1/0) 1(1/0) 1(1/0) 1(1/0) 2(1/0) 1(1/0) 1(0/0) sub.l Dy,<ea> — 1(1/1) 1(1/1) 1(1/1) 1(1/1) 2(1/1) 1(1/1) — subi.l #imm,Dx 1(0/0) — — — — — — — subq.l #imm,<ea> 1(0/0) 1(1/1) 1(1/1) 1(1/1) 1(1/1) 2(1/1) 1(1/1) — subx.l Dy,Dx 1(0/0) — — — — — — — 3.7.4 Miscellaneous Instruction Execution Timing Table 3-16 lists timings for miscellaneous instructions. Table 3-16. Miscellaneous Instruction Execution Times Effective Address Opcode <ea> Rn (An) (An)+ –(An) (d16,An) (d8,An,Xi*SF) (xxx).wl #<xxx> — — — — — — cpushl (Ax) — 9(0/1) intouch (Ay) — 19(1/0) link.w Ay,#imm 2(0/1) — — — — — — — move.w CCR,Dx 1(0/0) — — — — — — — move.w <ea>,CCR 1(0/0) — — — — — — 1(0/0) MCF548x Reference Manual, Rev. 5 3-32 Freescale Semiconductor Instruction Execution Timing Table 3-16. Miscellaneous Instruction Execution Times (Continued) Effective Address Opcode <ea> Rn (An) (An)+ –(An) (d16,An) (d8,An,Xi*SF) (xxx).wl #<xxx> move.w SR,Dx 1(0/0) — — — — — — — move.w <ea>,SR 4(0/0) — — — — — — 4(0/0) movec Ry,Rc 20(0/1) — — — — — — — movem.l 1 <ea>,&list — n(n/0) — — n(n/0) — — — movem.l — n(0/n) — — n(0/n) — — — 6(0/0) — — — — — — — 2(0/1) 1(0/1) — &list,<ea> nop pea <ea> pulse — 1(0/1) — — 1(0/1)2 3 1(0/0) — — — — — — — stop #imm — — — — — — — 6(0/0)4 trap #imm — — — — — — — 18(1/2) tpf 1(0/0) — — — — — — — tpf.w 1(0/0) — — — — — — — tpf.l 1(0/0) — — — — — — — 1(1/0) — — — — — — — unlk Ax wddata.l <ea> — 1(1/0) 1(1/0) 1(1/0) 1(1/0) 2(1/0) 1(1/0) — wdebug.l <ea> — 3(2/0) — — 3(2/0) — — — 1 n is the number of registers moved by the MOVEM opcode. PEA execution times are the same for (d16,PC). 3 PEA execution times are the same for (d8,PC,Xi*SF). 4 The execution time for STOP is the time required until the processor begins sampling continuously for interrupts. 2 3.7.5 Branch Instruction Execution Timing Table 3-17 shows general branch instruction timing. Table 3-17. General Branch Instruction Execution Times Effective Address Opcode <ea> Rn (An) (An)+ –(An) (d16,An) (d8,An,Xi*SF) (xxx).wl #<xxx> — — — — — bra — — — — 1(0/1)1 bsr — — — — 1(0/1)1 — 5(0/0)1 jmp jsr <ea> <ea> rte rts — 5(0/0) — — 6(0/0) 1(0/0) 1 — — — 5(0/1) — — 5(0/1) 6(0/1) 1(0/1)1 — — 15(2/0) — — — — — — 2(1/0)2 — — — — — — 9(1/0)3 8(1/0)4 1 Assumes branch acceleration. Depending on the pipeline status, execution times may vary from 1 to 3 cycles. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 3-33 2 If predicted correctly by the hardware return stack. If mispredicted by the hardware return stack. 4 If not predicted by the hardware return stack. 3 Table 3-18 shows timing for Bcc instructions. Table 3-18. Bcc Instruction Execution Times Opcode Branch Cache Correctly Predicts Taken Prediction Table Correctly Predicts Taken Predicted Correctly as Not Taken 0(0/0) 1(0/0) 1(0/0) bcc 3.7.6 Predicted Incorrectly 8(0/0) EMAC Instruction Execution Times Table 3-19 specifies instruction execution times associated with the enhanced multiply-accumulate (EMAC) execute engine. Table 3-19. EMAC Instruction Execution Times Effective Address Opcode mac.l <ea>y Ry,Rx,ACCx Rn (An) (An)+ –(An) (d16,An) (d16,PC) (d8,An,Xi*SF) (d8,PC,Xi*SF) xxx.wl #xxx 1(0/0) — — — — — — — 1 — 1(1/0) 1(1/0) 1(1/0) 1(1/0) — — — 1(0/0) — — — — — — — — 1(1/0) 1(1/0) 1(1/0) 1(1/0)1 — — — <ea>y,ACCx 1(0/0) — — — — — — 1(0/0) mov.l ACCy,ACCx 1(0/0) — — — — — — — mov.l <ea>y,MACSR 8(0/0) — — — — — — 8(0/0) mov.l <ea>y,MASK 7(0/0) — — — — — — 7(0/0) mov.l <ea>y,ACCext01 1(0/0) — — — — — — 1(0/0) mov.l <ea>y,ACCext23 1(0/0) — — — — — — 1(0/0) mov.l ACCx,<ea>x 1(0/0)2 — — — — — — — mov.l MACSR,<ea>x 1(0/0) — — — — — — — mov.l MASK,<ea>x 1(0/0) — — — — — — — mov.l ACCext01,<ea>x 1(0/0) — — — — — — — mov.l ACCext23,<ea>x 1(0/0) — — — — — — — msac.l Ry,Rx,ACCx 1(0/0) — — — — — — — — — — mac.l Ry,Rx,<ea>,Rw,ACCx mac.w Ry,Rx,ACCx mac.w Ry,Rx,<ea>,Rw,ACCx mov.l msac.l Ry,Rx,<ea>,Rw,ACCx msac.w Ry,Rx,ACCx — 1(1/0) 1(1/0) 1(1/0) 1(1/0)1 1(0/0) — — — — — — — — — — — 1(1/0) 1(1/0) 1(1/0) 1(1/0)1 <ea>y,Dx 4(0/0) 4(1/0) 4(1/0) 4(1/0) 4(1/0) — — — <ea>y,Dx 4(0/0) 4(1/0) 4(1/0) 4(1/0) 4(1/0) 5(1/0) 4(1/0) 4(0/0) msac.w Ry,Rx,<ea>,Rw,ACCx muls.l muls.w MCF548x Reference Manual, Rev. 5 3-34 Freescale Semiconductor Instruction Execution Timing Table 3-19. EMAC Instruction Execution Times (Continued) Effective Address Opcode <ea>y Rn (An) (An)+ –(An) (d16,An) (d16,PC) (d8,An,Xi*SF) (d8,PC,Xi*SF) xxx.wl #xxx mulu.l <ea>y,Dx 4(0/0) 4(1/0) 4(1/0) 4(1/0) 4(1/0) — — — mulu.w <ea>y,Dx 4(0/0) 4(1/0) 4(1/0) 4(1/0) 4(1/0) 5(1/0) 4(1/0) 4(0/0) 1 2 Effective address of (d16,PC) not supported. Storing the accumulator requires 1 additional clock cycle when saturation is enabled, or fractional rounding is performed (MACSR[7:4] = 1---, -11-, --11). Execution times for moving the contents of the ACC, ACCext[01,23], MACSR, or MASK into a destination location <ea>x in this table represent the best-case scenario when the store is executed and no load, copy, MAC, or MSAC instructions are in the EMAC execution pipeline. In general, these store operations require only a single cycle for execution, but if preceded immediately by a load, copy, MAC, or MSAC instruction, the depth of the EMAC pipeline is exposed and the execution time is 4 cycles. 3.7.7 FPU Instruction Execution Times Table 3-20 specifies the instruction execution times associated with the FPU execute engine. Table 3-20. FPU Instruction Execution Times1, 2 Effective Address <ea> Opcode Format FPn Dn (An) (An)+ –(An) (d16,An) (d16,PC) fabs <ea>y,FPx 1(0/0) 1(0/0) 1(1/0) 1(1/0) 1(1/0) 1(1/0) 1(1/0) fadd <ea>y,FPx 4(0/0) 4(0/0) 4(1/0) 4(1/0) 4(1/0) 4(1/0) 4(1/0) fbcc <label> — — — — — — 2(0/0) if correct, 9(0/0) if incorrect fcmp <ea>y,FPx 4(0/0) 4(0/0) 4(1/0) 4(1/0) 4(1/0) 4(1/0) 4(1/0) fdiv <ea>y,FPx 23(0/0) 23(0/0) 23(1/0) 23(1/0) 23(1/0) 23(1/0) 23(1/0) fint <ea>y,FPx 4(0/0) 4(0/0) 4(1/0) 4(1/0) 4(1/0) 4(1/0) 4(1/0) fintrz <ea>y,FPx 4(0/0) 4(0/0) 4(1/0) 4(1/0) 4(1/0) 4(1/0) 4(1/0) fmove <ea>y,FPx 1(0/0) 1(0/0) 1(1/0) 1(1/0) 1(1/0) 1(1/0) 1(1/0) fmove FPy,<ea>x — 2(0/1) 2(0/1) 2(0/1) 2(0/1) 2(0/1) — fmove <ea>y,FP*R — 6(0/0) 6(1/0) 6(1/0) 6(1/0) 6(1/0) 6(1/0) fmove FP*R,<ea>x — 1(0/0) 1(0/1) 1(0/1) 1(0/1) 1(0/1) — fmovem3 <ea>y,#list — — 2n(2n/0) — — 2n(2n/0) 2n(2n/0) fmovem3, 4 #list,<ea>x — — 1+2n(0/2n) — — 1+2n(0/2n) — fmul <ea>y,FPx 4(0/0) 4(0/0) 4(1/0) 4(1/0) 4(1/0) 4(1/0) 4(1/0) fneg <ea>y,FPx 1(0/0) 1(0/0) 1(1/0) 1(1/0) 1(1/0) 1(1/0) 1(1/0) — — — — — — 2(0/0) — — 6(4/0) — — 6(4/0) 6(4/0) fnop frestore <ea>y MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 3-35 Table 3-20. FPU Instruction Execution Times1, 2 (Continued) Effective Address <ea> Opcode Format FPn Dn (An) (An)+ –(An) (d16,An) (d16,PC) — — 7(0/3) — — 7(0/3) — fsave <ea>x fsqrt <ea>y,FPx 56(0/0) 56(0/0) 56(1/0) 56(1/0) 56(1/0) 56(1/0) 56(1/0) fsub <ea>y,FPx 4(0/0) 4(0/0) 4(1/0) 4(1/0) 4(1/0) 4(1/0) 4(1/0) ftst <ea>y,FPx 1(0/0) 1(0/0) 1(1/0) 1(1/0) 1(1/0) 1(1/0) 1(1/0) 1 Add 1(1/0) for an external read operand of double-precision format for all instructions except FMOVEM, and 1(0/1) for FMOVE FPy,<ea>x when the destination is double-precision. 2 If the external operand is an integer format (byte, word, or longword), there is a 4-cycle conversion time that must be added to the basic execution time. 3 For FMOVEM, n refers to the number of registers being moved. 4 If any exceptions are enabled, the execution time for FMOVE FPy,<ea>x increases by 1 cycle. If the BSUN exception is enabled, the execution time for FBcc increases by one cycle. 3.8 Exception Processing Overview Exception processing for ColdFire processors is streamlined for performance. Differences from previous ColdFire Family processors include the following: • An instruction restart model for translation (TLB miss) and access faults. This new functionality extends the existing ColdFire access error fault vector and exception stack frames. • Use of separate system stack pointers for user and supervisor modes. Previous ColdFire processors use an instruction restart exception model but require additional software support to recover from certain access errors. Exception processing can be defined as the time from the detection of the fault condition until the fetch of the first handler instruction has been initiated. It consists of the following four major steps: 1. The processor makes an internal copy of the status register (SR) and then enters supervisor mode by setting SR[S] and disabling trace mode by clearing SR[T]. The occurrence of an interrupt exception also clears SR[M] and sets the interrupt priority mask, SR[I] to the level of the current interrupt request. 2. The processor determines the exception vector number. For all faults except interrupts, the processor bases this calculation on exception type. For interrupts, the processor performs an interrupt acknowledge (IACK) bus cycle to obtain the vector number from peripheral. The IACK cycle is mapped to a special acknowledge address space with the interrupt level encoded in the address. The processor saves the current context by creating an exception stack frame on the system stack. As a result, the exception stack frame is created at a 0-modulo-4 address on top of the system stack pointed to by the supervisor stack pointer (SSP). As shown in Figure 3-15, the CF4e processor uses the same fixed-length stack frame as previous ColdFire Versions with additional fault status (FS) encodings to support the MMU. In some exception types, the program counter (PC) in the exception stack frame contains the address of the faulting instruction (fault); in others the PC contains the next instruction to be executed (next). (Note that previous ColdFire processors support a single stack pointer in the A7 address register.) MCF548x Reference Manual, Rev. 5 3-36 Freescale Semiconductor Exception Processing Overview If the exception is caused by an FPU instruction, the PC contains the address of either the next floating-point instruction (nextFP) if the exception is pre-instruction, or the faulting instruction (fault) if the exception is post-instruction. 3. The processor acquires the address of the first instruction of the exception handler. The instruction address is obtained by fetching a value from the exception table at the address in the vector base register. The index into the table is calculated as 4 x vector_number. When the index value is generated, the vector table contents determine the address of the first instruction of the desired handler. After the fetch of the first opcode of the handler is initiated, exception processing terminates and normal instruction processing continues in the handler. The vector base register described in the ColdFire Programmers Reference Manual, holds the base address of the exception vector table in memory. The displacement of an exception vector is added to the value in this register to access the vector table. VBR[19–0] are not implemented and are assumed to be zero, forcing the vector table to be aligned on a 0-modulo-1-Mbyte boundary. ColdFire processors support a 1,024-byte vector table aligned on any 0-modulo-1 Mbyte address boundary; see Table 3-21. The table contains 256 exception vectors, the first 64 of which are defined by Freescale. The rest are user-defined interrupt vectors. Table 3-21. Exception Vector Assignments Vector Numbers Vector Offset (Hex) Stacked Program Counter1 Assignment 0 000 — Initial supervisor stack pointer 1 004 — Initial program counter 2 008 Fault Access error 3 00C Fault Address error 4 010 Fault Illegal instruction 5 014 Fault Divide by zero 6–7 018–01C — 8 020 Fault Privilege violation 9 024 Next Trace 10 028 Fault Unimplemented line-a opcode 11 02C Fault Unimplemented line-f opcode 12 030 Next Non-PC breakpoint debug interrupt 13 034 Next PC breakpoint debug interrupt 14 038 Fault Format error 15 03C Next Uninitialized interrupt 16–23 040–05C — 24 060 Next Spurious interrupt 25–31 064–07C Next Level 1–7 autovectored interrupts 32–47 080–0BC Next Trap #0–15 instructions Reserved Reserved MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 3-37 Table 3-21. Exception Vector Assignments (Continued) Vector Numbers Vector Offset (Hex) 1 Stacked Program Counter1 Assignment 48 0C0 Fault Floating-point branch on unordered condition 49 0C4 NextFP or Fault Floating-point inexact result 50 0C8 NextFP Floating-point divide-by-zero 51 0CC NextFP or Fault Floating-point underflow 52 0D0 NextFP or Fault Floating-point operand error 53 0D4 NextFP or Fault Floating-point overflow 54 0D8 NextFP or Fault Floating-point input not-a-number (NAN) 55 0DC NextFP or Fault Floating-point input denormalized number 56–60 0E0–0F0 — 61 0F4 Fault 62–63 0F8–0FC — 64–255 100–3FC Next Reserved Unsupported instruction Reserved User-defined interrupts ‘Fault’ refers to the PC of the faulting instruction. ‘Next’ refers to the PC of the instruction immediately after the faulting instruction. NextFP’ refers to the PC of the next floating-point instruction. ColdFire processors inhibit sampling for interrupts during the first instruction of all exception handlers. This allows any handler to effectively disable interrupts, if necessary, by raising the interrupt mask level in the SR. 3.8.1 Exception Stack Frame Definition The first longword of the exception stack frame, Figure 3-15, holds the 16-bit format/vector word (F/V) and 16-bit status register. The second holds the 32-bit program counter address of the faulted or interrupted instruction. 31 A7→ + 0x04 28 FORMAT 27 26 25 FS[3–2] 18 VEC 17 16 15 FS[1–0] 0 STATUS REGISTER PROGRAM COUNTER [31:0] Figure 3-15. Exception Stack Frame Table 3-22 describes F/V fields. FS encodings added to support the CF4e MMU are noted. MCF548x Reference Manual, Rev. 5 3-38 Freescale Semiconductor Exception Processing Overview Table 3-22. Format/Vector Word 1 Bits Name Description 31–28 FORMAT Format field. Written with a value of {4,5,6,7} by the processor indicating a 2-longword frame format. FORMAT records any longword stack pointer misalignment when the exception occurred. 27–26 FS[3:2] 25–18 VEC 17–16 FS[1:0] A7 at Exception Bits 1–0 A7 at First Instruction of Handler Format 00 Original A7–8 0100 01 Original A7–9 0101 10 Original A7–10 0110 11 Original A7–11 0111 Fault status. Defined for access and address errors and for interrupted debug service routines. 0000 Not an access or address error nor an interrupted debug service routine 0001 Reserved 0010 Interrupt during a debug service routine for faults other than access errors. 1 [ 0011 Reserved 0100 Error (for example, protection fault) on instruction fetch 0101 TLB miss on opword of instruction fetch (New in CF4e) 0110 TLB miss on extension word of instruction fetch (New in CF4e) 0111 IFP access error while executing in emulator mode (New in CF4e) 1000 Error on data write 1001 Error on attempted write to write-protected space 1010 TLB miss on data write (New in CF4e) 1011 Reserved 1100 Error on data read 1101 Attempted read, read-modify-write of protected space (New in CF4e) 1110 TLB miss on data read, or read-modify-write (New in CF4e) 1111 OEP access error while executing in emulator mode (New in CF4e) Vector number. Defines the exception type. It is calculated by the processor for internal faults and is supplied by the peripheral for interrupts. See Table 3-21. See bits 27–26. This generally refers to taking an I/O interrupt during a debug service routine but also applies to other fault types. If an access error occurs during a debug service routine, FS is set to 0111 if it is due to an instruction fetch or to 1111 for a data access. This applies only to access errors with the MMU present. If an access error occurs without an MMU, FS is set to 0010. 3.8.2 Processor Exceptions Table 3-23 describes CF4e exceptions. Note that if a ColdFire processor encounters any fault while processing another fault, it immediately halts execution with a catastrophic fault-on-fault condition. A reset is required to force the processor to exit this halted state. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 3-39 Table 3-23. Processor Exceptions Type Description Access error If the MMU is disabled, access errors are reported only in conjunction with an attempted store to write-protected memory. Thus, access errors associated with instruction fetch or operand read accesses are not possible. The Version 4 processor, unlike the Version 2 and 3 processors, updates the condition code register if a write-protect error occurs during a CLR or MOV3Q operation to memory. accesses that fault (that is, terminated with a transfer error acknowledge) generate an access error exception. MMU TLB misses and access violations use the same fault. If the MMU is enabled, all TLB misses and protection violations generate an access error exception. To determine if a fault is due to a TLB miss or another type of access error, new FS encodings (described in Table 3-22) signal TLB misses on the following: • Instruction fetch • Instruction extension fetch • Data read • Data write Address error An address error is caused by an attempted execution transferring control to an odd instruction address (that is, if bit 0 of the target address is set), an attempted use of a word-sized index register (Xi.w) or by an attempted execution of an instruction with a full-format indexed addressing mode. If an address error occurs on a JSR instruction, the Version 4 processor first pushes the return address onto the stack and then calculates the target address. On Version 2 and 3 processors, the target address is calculated then the return address is pushed on stack. If an address error occurs on an RTS instruction, the Version 4 processor preserves the original return PC and writes the exception stack frame above this value. On Version 2 and 3 processors, the faulting return PC is overwritten by the address error stack frame. Illegal instruction The scope of illegal instruction detection is implementation-specific across the generations of ColdFire cores. For the CF4e core, the complete 16-bit opcode is decoded and this exception is generated if execution of an unsupported instruction is attempted. Additionally, attempting to execute an illegal line A or line F opcode generates unique exception types: vectors 10 and 11, respectively. ColdFire processors do not provide illegal instruction detection on extension words of any instruction, including MOVEC. Attempting to execute an instruction with an illegal extension word causes undefined results. Divide-by-zero Privilege violation Attempting to divide by zero causes an exception (vector 5, offset = 0x014). Caused by attempted execution of a supervisor mode instruction while in user mode. The ColdFire Programmer’s Reference Manual lists supervisor- and user-mode instructions. Trace exception Trace mode, which allows instruction-by-instruction tracing, is enabled by setting SR[T]. If SR[T] is set, instruction completion (for all but the STOP instruction) signals a trace exception.The STOP instruction has the following effects: 1 The instruction before the STOP executes and then generates a trace exception. In the exception stack frame, the PC points to the STOP opcode. 2 When the trace handler is exited, the STOP instruction is executed, loading the SR with the immediate operand from the instruction. 3 The processor then generates a trace exception. The PC in the exception stack frame points to the instruction after STOP, and the SR reflects the value loaded in the previous step. If the processor is not in trace mode and executes a STOP instruction where the immediate operand sets SR[T], hardware loads the SR and generates a trace exception. The PC in the exception stack frame points to the instruction after STOP, and the SR reflects the value loaded in step 2. Note that because ColdFire processors do not support hardware stacking of multiple exceptions, it is the responsibility of the operating system to check for trace mode after processing other exception types. For example, when a TRAP instruction executes in trace mode, the processor initiates the TRAP exception and passes control to the corresponding handler. If the system requires a trace exception, the TRAP exception handler must check for this condition (SR[15] in the exception stack frame set) and pass control to the trace handler before returning from the original exception. MCF548x Reference Manual, Rev. 5 3-40 Freescale Semiconductor Exception Processing Overview Table 3-23. Processor Exceptions (Continued) Type Description Unimplemented A line-a opcode results when bits 15–12 of the opword are 1010. This exception is generated by the line-a opcode attempted execution of an undefined line-a opcode. Unimplemented A line-f opcode results when bits 15–12 of the opword are 1111. This exception is generated under the line-f opcode following conditions: • When attempting to execute an undefined line-f opcode. • When attempting to execute an FPU instruction when the FPU has been disabled in the CACR. Debug interrupt The debug interrupt exception is caused by a hardware breakpoint register trigger. Rather than generating an IACK cycle, the processor internally calculates the vector number (12 or 13, depending on the type of breakpoint trigger). Additionally, SR[M,I] are unaffected by the interrupt. Separate exception vectors are provided for PC breakpoints and for address/data breakpoints. In the case of a two-level trigger, the last breakpoint determines the vector. The two unique entries occur when a PC breakpoint generates the 0x034 vector. In case of a two-level trigger, the last breakpoint event determines the vector. See Chapter 8, “Debug Support,” for more information. Format error When an RTE instruction executes, the processor first examines the 4-bit format field to validate the frame type. For a ColdFire processor, attempted execution of an RTE where the format is not equal to {4, 5, 6, 7} generates a format error. The exception stack frame for the format error is created without disturbing the original exception frame and the stacked PC points to RTE. The selection of the format value provides limited debug support for porting code from M68000 applications. On M68000 Family processors, the SR was at the top of the stack. Bit 30 of the longword addressed by the system stack pointer is typically zero. Attempting an RTE using this old format generates a format error on a ColdFire processor. If the format field defines a valid type, the processor does the following: 1 Reloads the SR operand. 2 Fetches the second longword operand. 3 Adjusts the stack pointer by adding the format value to the auto-incremented address after the first longword fetch. 4 Transfers control to the instruction address defined by the second longword operand in the stack frame. When the processor executes a FRESTORE instruction, if the restored FPU state frame contains a non-supported value, execution is aborted and a format error exception is generated. Trap Executing a TRAP instruction always forces an exception and is useful for implementing system calls. The trap instruction may be used to change from user to supervisor mode. Interrupt exception Please refer to Chapter 13, “Interrupt Controller.” Reset exception Asserting the reset input signal (RSTI) causes a reset exception, which has the highest exception priority and provides for system initialization and recovery from catastrophic failure. When assertion of RSTI is recognized, current processing is aborted and cannot be recovered. The reset exception places the processor in supervisor mode by setting SR[S] and disables tracing by clearing SR[T]. It clears SR[M] and sets SR[I] to the highest level (0b111, priority level 7). Next, VBR is cleared. Configuration registers controlling operation of all processor-local memories are invalidated, disabling the memories. Note: Implementation-specific supervisor registers are also affected at reset. After RSTI is negated, the processor waits 16 cycles before beginning the reset exception process. During this time, certain events are sampled, including the assertion of the debug breakpoint signal. If the processor is not halted, it initiates the reset exception by performing two longword read bus cycles. The longword at address 0 is loaded into the stack pointer and the longword at address 4 is loaded into the PC. After the initial instruction is fetched from memory, program execution begins at the address in the PC. If an access error or address error occurs before the first instruction executes, the processor enters a fault-on-fault halted state. Unsupported instruction exception If the CF4e attempts to execute a valid instruction but the required optional hardware module is not present in the OEP, a non-supported instruction exception is generated (vector 0x61). Control is then passed to an exception handler that can then process the opcode as required by the system. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 3-41 3.9 Precise Faults To support a demand-paged virtual memory environment, all memory references require precise, recoverable faults. The ColdFire instruction restart mechanism ensures that a faulted instruction restarts from the beginning of execution; that is, no internal state information is saved when an exception occurs and none is restored when the handler ends. Given the PC address defined in the exception stack frame, the processor reestablishes program execution by transferring control to the given location as part of the RTE (return from exception) instruction. The instruction restart recovery model requires program-visible register changes made during execution to be undone if that instruction subsequently faults. The Version 4 (and later) OEP structure naturally supports this concept for most instructions; program-visible registers are updated only in the final OEP stage when fault collection is complete. If any type of exception occurs, pending register updates are discarded. For V4 cores and later, most single-cycle instructions already support precise faults and instruction restart. Some complex instructions do not. Consider the following memory-to-memory move: mov.l (Ay)+,(Ax)+ # copy 4 bytes from source to destination On a Version 4 processor, this instruction takes one cycle to read the source operand (Ay) and one to write the data into Ax. Both the source and destination address pointers are updated as part of execution. Table 3-24 lists the operations performed in execute stage (EX). Table 3-24. OEP EX Cycle Operations EX Cycle Operations 1 Read source operand from memory @ (Ay), update Ay, new Ay = old Ay + 4 2 Write operand into destination memory @ (Ax), update Ax, new Ax = old Ax + 4, update CCR A fault detected with the destination memory write is reported during the second cycle. At this point, operations performed in the first cycle are complete, so if the destination write takes any type of access error, Ay is updated. After the access error handler executes and the faulting instruction restarts, the processor’s operation is incorrect because the source address register has an incorrect (post-incremented) value. To recover the original state of the programming model for all instructions, the CF4e CPU adds the needed hardware to support full register recovery. This hardware allows program-visible registers to be restored to their original state for multi-cycle instructions so that the instruction restart mechanism is supported. Memory-to-memory moves and move multiple loads are representative of the complex instructions needing the special recovery support. The other major pipeline change affects the IFP. The IFP and OEP are decoupled by a FIFO instruction buffer. In the V4 IFP, each buffer entry includes 48 bits of instruction data fetched from memory and 64 bits of early decode and branch prediction information. This datapath is expanded slightly to include IFP fault status information. Thus, every IFP access can be tagged in case an instruction fetch terminates with an error acknowledge. MCF548x Reference Manual, Rev. 5 3-42 Freescale Semiconductor Precise Faults NOTE For access errors signaled on instruction prefetches, an access error exception is generated only if instruction execution is attempted. If an instruction fetch access error exception is generated and the FS field indicates the fault occurred on an extension word, it may be necessary for the exception PC to be rounded-up to the next page address to determine the faulting instruction fetch address. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 3-43 MCF548x Reference Manual, Rev. 5 3-44 Freescale Semiconductor Chapter 4 Enhanced Multiply-Accumulate Unit (EMAC) This chapter describes the functionality, microarchitecture, and performance of the enhanced multiply-accumulate (EMAC) unit in the ColdFire family of processors. 4.1 Introduction The MAC design provides a set of DSP operations which can be used to improve the performance of embedded code while supporting the integer multiply instructions of the baseline ColdFire architecture. The MAC provides functionality in three related areas: • Signed and unsigned integer multiplies • Multiply-accumulate operations supporting signed and unsigned integer operands, as well as signed, fixed-point, fractional operands • Miscellaneous register operations The ColdFire family supports two MAC implementations with different performance levels and capabilities. The original MAC uses a three-stage execution pipeline optimized for 16-bit operands and featuring a 16 × 16 multiply array with a single 32-bit accumulator. The EMAC features a four-stage pipeline optimized for 32-bit operands, with a fully pipelined 32 × 32 multiply array and four 48-bit accumulators. The first ColdFire MAC supported signed and unsigned integer operands and was optimized for 16 × 16 operations, such as those found in a variety of applications, including servo control and image compression. As ColdFire-based systems proliferated, the desire for more precision on input operands increased. The result was an improved ColdFire MAC with user-programmable control to optionally enable use of fractional input operands. EMAC improvements target three primary areas: • Improved performance of 32 × 32 multiply operations. • Addition of three more accumulators to minimize EMAC pipeline stalls caused by exchanges between the accumulator and the pipeline’s general-purpose registers. • A 48-bit accumulation data path to allow the use of a 40-bit product plus the addition of 8 extension bits to increase the dynamic number range when implementing signal processing algorithms. The three areas of functionality are addressed in detail in following sections. The logic required to support this functionality is contained in a MAC module, as shown in Figure 4-1. Operand Y Operand X X Shift 0,1,-1 +/- Accumulator(s) Figure 4-1. Multiply-Accumulate Functionality Diagram MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 4-1 4.1.1 MAC Overview The MAC is an extension of the basic multiplier found in most microprocessors. It is typically implemented in hardware within an architecture and supports rapid execution of signal processing algorithms in fewer cycles than comparable non-MAC architectures. For example, small digital filters can tolerate some variance in an algorithm’s execution time, but larger, more complicated algorithms such as orthogonal transforms may have more demanding speed requirements beyond the scope of any processor architecture, and may require full DSP implementation. To strike a balance between speed, size, and functionality, the ColdFire MAC is optimized for a small set of operations that involve multiplication and cumulative additions. Specifically, the multiplier array is optimized for single-cycle pipelined operations with a possible accumulation after product generation. This functionality is common in many signal processing applications. The ColdFire core architecture also has been modified to allow an operand to be fetched in parallel with a multiply, increasing overall performance for certain DSP operations. Consider a typical filtering operation where the filter is defined,11 as in Figure 4-2. N–1 y(i) = N–1 ∑ a ( k )y ( i – k ) + ∑ b ( k )x ( i – k ) k=1 k=0 Figure 4-2. Infinite Impulse Response (IIR) Filter Here, the output y(i) is determined by past output values and past input values. This is the general form of an infinite impulse response (IIR) filter. A finite impulse response (FIR) filter can be obtained by setting coefficients a(k) to zero. In either case, the operations involved in computing such a filter are multiplies and product summing. To show this point, reduce the above equation to a simple, four-tap FIR filter, shown in Figure 4-3, in which the accumulated sum is a sum of past data values and coefficients. 3 y(i) = ∑ b ( k )x ( i – k ) = b ( 0 )x ( i ) + b ( 1 )x ( i – 1 ) + b ( 2 )x ( i – 2 ) + b ( 3 )x ( i – 3 ) k=0 Figure 4-3. Four-Tap FIR Filter 4.1.2 General Operation The MAC speeds execution of ColdFire integer multiply instructions (MULS and MULU) and provides additional functionality for multiply-accumulate operations. By executing MULS and MULU in the MAC, execution times are minimized and deterministic compared to the 2-bit/cycle algorithm with early termination that the OEP normally uses if no MAC hardware is present. The added MAC instructions to the ColdFire ISA provide for the multiplication of two numbers, followed by the addition or subtraction of the product to or from the value in an accumulator. Optionally, the product may be shifted left or right by 1 bit before addition or subtraction. Hardware support for saturation arithmetic can be enabled to minimize software overhead when dealing with potential overflow conditions. Multiply-accumulate operations support 16- or 32-bit input operands of the following formats: MCF548x Reference Manual, Rev. 5 4-2 Freescale Semiconductor Introduction • • • Signed integers Unsigned integers Signed, fixed-point, fractional numbers The EMAC is optimized for single-cycle, pipelined 32 × 32 multiplications. For word- and longword-sized integer input operands, the low-order 40 bits of the product are formed and used with the destination accumulator. For fractional operands, the entire 64-bit product is calculated and either truncated or rounded to the most-significant 40-bit result using the round-to-nearest (even) method before it is combined with the destination accumulator. For all operations, the resulting 40-bit product is extended to a 48-bit value (using sign-extension for signed integer and fractional operands, zero-fill for unsigned integer operands) before being combined with the 48-bit destination accumulator. Figure 4-4 and Figure 4-5 show relative alignment of input operands, the full 64-bit product, the resulting 40-bit product used for accumulation, and 48-bit accumulator formats. X Product Extended Product OperandY 32 OperandX 32 40 23 8 “0” 40 + Accumulator 8 Extension Byte Upper [7:0] 8 40 Accumulator [31:0] Extension Byte Lower [7:0] Figure 4-4. Fractional Alignment MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 4-3 X Product Extended Product OperandY 32 OperandX 32 8 32 8 8 32 8 8 32 24 + Accumulator Extension Byte Upper [7:0] Accumulator [31:0] Extension Byte Lower [7:0] Figure 4-5. Signed and Unsigned Integer Alignment Thus, the 48-bit accumulator definition is a function of the EMAC operating mode. Given that each 48-bit accumulator is the concatenation of 16-bit accumulator extension register (ACCextn) contents and 32-bit ACCn contents, the specific definitions are as follows: if MACSR[6:5] == 00/* signed integer mode */ Complete Accumulator[47:0] = {ACCextn[15:0], ACCn[31:0]} if MACSR[6:5] == -1/* signed fractional mode */ Complete Accumulator [47:0] = {ACCextn[15:8], ACCn[31:0], ACCextn[7:0]} if MACSR[6:5] == 10/* unsigned integer mode */ Complete Accumulator[47:0] = {ACCextn[15:0], ACCn[31:0]} The four accumulators are represented as an array, ACCn, where n selects the register. Although the multiplier array is implemented in a four-stage pipeline, all arithmetic MAC instructions have an effective issue rate of 1 cycle, regardless of input operand size or type. All arithmetic operations use register-based input operands, and summed values are stored internally in an accumulator. Thus, an additional move instruction is needed to store data in a general-purpose register. One new feature found in EMAC instructions is the ability to choose the upper or lower word of a register as a 16-bit input operand. This is useful in filtering operations if one data register is loaded with the input data and another is loaded with the coefficient. Two 16-bit multiply accumulates can be performed without fetching additional operands between instructions by alternating the word choice during the calculations. The EMAC has four accumulator registers versus the MAC’s one accumulator. The additional registers improve the performance of some algorithms by minimizing pipeline stalls needed to store an accumulator value back to general-purpose registers. Many algorithms require multiple calculations on a given data set. By applying different accumulators to these calculations, it is often possible to store one accumulator without any stalls while performing operations involving a different destination accumulator. MCF548x Reference Manual, Rev. 5 4-4 Freescale Semiconductor Memory Map/Register Definition The need to move large amounts of data presents an obstacle to obtaining high throughput rates in DSP engines. New and existing ColdFire instructions can accommodate these requirements. A MOVEM instruction can move large blocks of data efficiently by generating line-sized burst transfers. The ability to simultaneously load an operand from memory into a register and execute a MAC instruction makes some DSP operations such as filtering and convolution more manageable. The programming model includes a 16-bit mask register (MASK), which can optionally be used to generate an operand address during MAC + MOVE instructions. The application of this register with auto-increment addressing mode supports efficient implementation of circular data queues for memory operands. The additional MAC status register (MACSR) contains a 4-bit operational mode field and condition flags. Operational mode bits control whether operands are signed or unsigned and whether they are treated as integers or fractions. These bits also control the overflow/saturation mode and the way in which rounding is performed. Negative, zero, and multiple overflow condition flags are also provided. 4.2 Memory Map/Register Definition The EMAC provides the following program-visible registers: • Four 32-bit accumulators (ACCn = ACC0, ACC1, ACC2, and ACC3) • Eight 8-bit accumulator extensions (two per accumulator), packaged as two 32-bit values for load and store operations (ACCext01 and ACCext23) • One 16-bit mask register (MASK) • One 32-bit MAC status register (MACSR) including four indicator bits signaling product or accumulation overflow (one for each accumulator: PAV0–PAV3) These registers are shown in Figure 4-6. 31 0 MACSR ACC0 ACC1 ACC2 ACC3 ACCext01 ACCext23 MASK MAC status register MAC accumulator 0 MAC accumulator 1 MAC accumulator 2 MAC accumulator 3 Extensions for ACC0 and ACC1 Extensions for ACC2 and ACC3 MAC mask register Figure 4-6. EMAC Register Set 4.2.1 MAC Status Register (MACSR) MACSR functionality is organized as follows: • MACSR[11–8] contains one product/accumulation overflow flag per accumulator. • MACSR[7–4] defines the operating configuration of the MAC unit. • MACSR[3–0] contains indicator flags from the last MAC instruction execution. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 4-5 R 31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 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 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 0 0 0 OMC S/U F/I R/T N Z V EV 0 0 0 0 0 0 0 0 0 0 0 0 W Reset R PAVx W Reset 0 0 0 0 Reg Addr Figure 4-7. MAC Status Register (MACSR) Table 4-1 describes MACSR fields. Table 4-1. MACSR Field Descriptions Bits Name Description 31–12 — 11–8 PAVx Product/accumulation overflow flags. Contains four flags, one per accumulator, that indicate if past MAC or MSAC instructions generated an overflow during product calculation or the 48-bit accumulation. When a MAC or MSAC instruction is executed, the PAVx flag associated with the destination accumulator is used to form the general overflow flag, MACSR[V]. Once set, each flag remains set until V is cleared by a MOV.L , MACSR instruction or the accumulator is loaded directly. 7 OMC Operational mode field: Overflow/saturation mode. Used to enable or disable saturation mode on overflow. If set, the accumulator is set to the appropriate constant on any operation which overflows the accumulator. Once saturated, the accumulator remains unaffected by any other MAC or MSAC instructions until either the overflow bit is cleared or the accumulator is directly loaded. 6 S/U Operational mode field: Signed/unsigned operations. In integer mode: S/U determines whether operations performed are signed or unsigned. It also determines the accumulator value during saturation, if enabled. 0 Signed numbers. On overflow, if OMC is enabled, an accumulator saturates to the most positive (0x7FFF_FFFF) or the most negative (0x8000_0000) number, depending on both the instruction and the value of the product that overflowed. 1 Unsigned numbers. On overflow, if OMC is enabled, an accumulator saturates to the smallest value (0x0000_0000) or the largest value (0xFFFF_FFFF), depending on the instruction. In fractional mode: S/U controls rounding while storing an accumulator to a general-purpose register. 0 Move accumulator without rounding to a 16-bit value. Accumulator is moved to a general-purpose register as a 32-bit value. 1 The accumulator is rounded to a 16-bit value using the round-to-nearest (even) method when it is moved to a general-purpose register. See Section 4.2.1.1.1, “Rounding.” The resulting 16-bit value is stored in the lower word of the destination register. The upper word is zero-filled. The accumulator value is not affected by this rounding procedure. Reserved, should be cleared. MCF548x Reference Manual, Rev. 5 4-6 Freescale Semiconductor Memory Map/Register Definition Table 4-1. MACSR Field Descriptions (Continued) Bits Name Description 5 F/I Operational mode field: Fractional/integer mode Determines whether input operands are treated as fractions or integers. 0 Integers can be represented in either signed or unsigned notation, depending on the value of S/U. 1 Fractions are represented in signed, fixed-point, two’s complement notation. Values range from -1 to 1- 2-15 for 16-bit fractions and -1 to 1 - 2-31 for 32-bit fractions. See Section 4.3.2, “Data Representation." 4 R/T Operational mode field: Round/truncate mode. Controls the rounding procedure for MOV.L ACCx,Rx, or MSAC.L instructions when operating in fractional mode. 0 Truncate. The product’s lsbs are dropped before it is combined with the accumulator. Additionally, when a store accumulator instruction is executed (MOV.L ACCx,Rx), the 8 lsbs of the 48-bit accumulator logic are simply truncated. 1 Round-to-nearest (even). The 64-bit product of two 32-bit, fractional operands is rounded to the nearest 40-bit value. If the low-order 24 bits equal 0x80_0000, the upper 40 bits are rounded to the nearest even (lsb = 0) value.See Section 4.2.1.1.1, “Rounding.” Additionally, when a store accumulator instruction is executed (MOV.L ACCx,Rx), the lsbs of the 48-bit accumulator logic are used to round the resulting 16- or 32-bit value. If MACSR[S/U] = 0 and MACSR[R/T] = 1, the low-order 8 bits are used to round the resulting 32-bit fraction. If MACSR[S/U] = 1, the low-order 24 bits are used to round the resulting 16-bit fraction. 3 N Negative flag. Set if the msb of the result is set, otherwise cleared. N is affected only by MAC, MSAC, and load operations; it is not affected by MULS and MULU instructions. 2 Z Zero flag. Set if the result equals zero, otherwise cleared. This bit is affected only by MAC, MSAC, and load operations; it is not affected by MULS and MULU instructions. 1 V Overflow flag. Set if an arithmetic overflow occurs on a MAC or MSAC instruction indicating that the result cannot be represented in the limited width of the EMAC. V is set only if a product overflow occurs or the accumulation overflows the 48-bit structure. V is evaluated on each MAC or MSAC operation and uses the appropriate PAVx flag in the next-state V evaluation. 0 EV Extension overflow flag. Signals that the last MAC or MSAC instruction overflowed the 32 lsbs in integer mode or the 40 lsbs in fractional mode of the destination accumulator. However, the result is still accurately represented in the combined 48-bit accumulator structure. Although an overflow has occurred, the correct result, sign, and magnitude are contained in the 48-bit accumulator. Subsequent MAC or MSAC operations may return the accumulator to a valid 32/40-bit result. Table 4-2 summarizes the interaction of the MACSR[S/U,F/I,R/T] control bits. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 4-7 Table 4-2. Summary of S/U, F/I, and R/T Control Bits 4.2.1.1 S/U F/I R/T Operational Modes 0 0 x Signed, integer 0 1 0 Signed, fractional Truncate on MAC.L and MSAC.L No round on accumulator stores 0 1 1 Signed, fractional Round on MAC.L and MSAC.L Round-to-32-bits on accumulator stores 1 0 x Unsigned, integer 1 1 0 Signed, fractional Truncate on MAC.L and MSAC.L Round-to-16-bits on accumulator stores 1 1 1 Signed, fractional Round on MAC.L and MSAC.L Round-to-16-bits on accumulator stores Fractional Operation Mode This section describes behavior when the fractional mode is used (MACSR[F/I] is set). 4.2.1.1.1 Rounding When the processor is in fractional mode, there are two operations during which rounding can occur. • Execution of a store accumulator instruction (MOV.L ACCx,Rx). The lsbs of the 48-bit accumulator logic are used to round the resulting 16- or 32-bit value. If MACSR[S/U] is cleared, the low-order 8 bits are used to round the resulting 32-bit fraction. If MACSR[S/U] is set, the low-order 24 bits are used to round the resulting 16-bit fraction. • Execution of a MAC (or MSAC) instruction with 32-bit operands. If MACSR[R/T] is zero, multiplying two 32-bit numbers creates a 64-bit product that is truncated to the upper 40 bits; otherwise, it is rounded using round-to-nearest (even) method. To understand the round-to-nearest-even method, consider the following example involving the rounding of a 32-bit number, R0, to a 16-bit number. Using this method, the 32-bit number is rounded to the closest 16-bit number possible. Let the high-order 16 bits of R0 be named R0.U and the low-order 16 bits be R0.L. • If R0.L is less than 0x8000, the result is truncated to the value of R0.U. • If R0.L is greater than 0x8000, the upper word is incremented (rounded up). • If R0.L is 0x8000, R0 is half-way between two 16-bit numbers. In this case, rounding is based on the lsb of R0.U, so the result is always even (lsb = 0). — If the lsb of R0.U = 1 and R0.L = 0x8000, the number is rounded up. — If the lsb of R0.U = 0 and R0.L =0x8000, the number is rounded down. This method minimizes rounding bias and creates as statistically correct an answer as possible. The rounding algorithm is summarized in the following pseudocode: if R0.L < 0x8000 then Result = R0.U else if R0.L > 0x8000 MCF548x Reference Manual, Rev. 5 4-8 Freescale Semiconductor Memory Map/Register Definition then Result = R0.U + 1 else if lsb of R0.U = 0 /* R0.L = 0x8000 */ then Result = R0.U else Result = R0.U + 1 The round-to-nearest-even technique is also known as convergent rounding. 4.2.1.1.2 Saving and Restoring the EMAC Programming Model The presence of rounding logic in the output datapath of the EMAC requires that special care be taken during the EMAC’s save/restore process. In particular, any result rounding modes must be disabled during the save/restore process so the exact bit-wise contents of the EMAC registers are accessed. Consider the following memory structure containing the EMAC programming model: struct macState { int int int int int int int int acc0; acc1; acc2; acc3; accext01; accext02; mask; macsr; } macState; The following assembly language routine shows the proper sequence for a correct EMAC state save. This code assumes all Dn and An registers are available for use and the memory location of the state save is defined by A7. EMAC_state_save: move.l macsr,d7 clr.l d0 move.l d0,macsr move.l acc0,d0 move.l acc1,d1 move.l acc2,d2 move.l acc3,d3 move.l accext01,d4 move.l accext23,d5 move.l mask,d6 movem.l #0x00ff,(a7) ; ; ; ; save the macsr zero the register to ... disable rounding in the macsr save the accumulators ; save the accumulator extensions ; save the address mask ; move the state to memory The following code performs the EMAC state restore: EMAC_state_restore: movem.l move.l move.l move.l move.l move.l move.l move.l (a7),#0x00ff #0,macsr d0,acc0 d1,acc1 d2,acc2 d3,acc3 d4,accext01 d5,accext23 ; restore the state from memory ; disable rounding in the macsr ; restore the accumulators ; restore the accumulator extensions MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 4-9 move.l move.l d6,mask d7,macsr ; restore the address mask ; restore the macsr By executing this type of sequence, the exact state of the EMAC programming model can be correctly saved and restored. 4.2.1.1.3 MULS/MULU MULS and MULU are unaffected by fractional mode operation; operands are still assumed to be integers. 4.2.1.1.4 Scale Factor in MAC or MSAC Instructions The scale factor is ignored while the MAC is in fractional mode. 4.2.2 Mask Register (MASK) The 32-bit MASK implements the low-order 16 bits to minimize the alignment complications involved with loading and storing only 16 bits. When the MASK is loaded, the low-order 16 bits of the source operand are actually loaded into the register. When it is stored, the upper 16 bits are all forced to ones. This register performs a simple AND with the operand address for MAC instructions. That is, the processor calculates the normal operand address and, if enabled, that address is then ANDed with {0xFFFF, MASK[15:0]} to form the final address. Therefore, with certain MASK bits cleared, the operand address can be constrained to a certain memory region. This is used primarily to implement circular queues in conjunction with the (An)+ addressing mode. This feature minimizes the addressing support required for filtering, convolution, or any routine that implements a data array as a circular queue. For MAC + MOVE operations, the MASK contents can optionally be included in all memory effective address calculations. The syntax is as follows: MAC.sz Ry,RxSF,<ea>y&,Rw The & operator enables the use of MASK and causes bit 5 of the extension word to be set. The exact algorithm for the use of MASK is as follows: if extension word, bit [5] = 1, the MASK bit, then if <ea> = (An) oa = An & {0xFFFF, MASK} if <ea> = (An)+ oa = An An = (An + 4) & {0xFFFF, MASK} if <ea> =-(An) oa = (An - 4) & {0xFFFF, MASK} An = (An - 4) & {0xFFFF, MASK} if <ea> = (d16,An) oa = (An + se_d16) & {0xFFFF0x, MASK} Here, oa is the calculated operand address and se_d16 is a sign-extended 16-bit displacement. For auto-addressing modes of post-increment and pre-decrement, the calculation of the updated An value is also shown. Use of the post-increment addressing mode, {(An)+} with the MASK is suggested for circular queue implementations. MCF548x Reference Manual, Rev. 5 4-10 Freescale Semiconductor EMAC Instruction Set Summary 4.3 EMAC Instruction Set Summary Table 4-3 summarizes EMAC unit instructions. Table 4-3. EMAC Instruction Summary Command Mnemonic Description Multiply Signed MULS <ea>y,Dx Multiplies two signed operands yielding a signed result Multiply Unsigned MULU <ea>y,Dx Multiplies two unsigned operands yielding an unsigned result Multiply Accumulate MAC Ry,RxSF,ACCx MSAC Ry,RxSF,ACCx Multiplies two operands and adds/subtracts the product to/from an accumulator Multiply Accumulate with Load MAC Ry,Rx,<ea>y,Rw,ACCx Multiplies two operands and combines the product to an MSAC Ry,Rx,<ea>y,Rw,ACCx accumulator while loading a register with the memory operand Load Accumulator MOV.L {Ry,#imm},ACCx Loads an accumulator with a 32-bit operand Store Accumulator MOV.L ACCx,Rx Writes the contents of an accumulator to a CPU register Copy Accumulator MOV.L ACCy,ACCx Copies a 48-bit accumulator Load MACSR MOV.L {Ry,#imm},MACSR Writes a value to MACSR Store MACSR MOV.L MACSR,Rx Write the contents of MACSR to a CPU register Store MACSR to CCR MOV.L MACSR,CCR Write the contents of MACSR to the CCR Load MAC Mask Reg MOV.L {Ry,#imm},MASK Writes a value to the MASK register Store MAC Mask Reg MOV.L MASK,Rx Writes the contents of the MASK to a CPU register Load AccExtensions01 MOV.L {Ry,#imm},ACCext01 Loads the accumulator 0,1 extension bytes with a 32-bit operand Load AccExtensions23 MOV.L {Ry,#imm},ACCext23 Loads the accumulator 2,3 extension bytes with a 32-bit operand Store AccExtensions01 MOV.L ACCext01,Rx Writes the contents of accumulator 0,1 extension bytes into a CPU register Store AccExtensions23 MOV.L ACCext23,Rx Writes the contents of accumulator 2,3 extension bytes into a CPU register 4.3.1 EMAC Instruction Execution Timing The instruction execution times for the EMAC can be found in Section 3.7, “Instruction Execution Timing.” The ColdFire family supports two multiply-accumulate implementations that provide different levels of performance and capability for differing silicon costs. The EMAC features a four-stage execution pipeline, optimized for 32-bit operands with a fully-pipelined 32 × 32 multiply array and four 48-bit accumulators. The EMAC execution pipeline overlaps the AGEX stage of the OEP; that is, the first stage of the EMAC pipeline is the last stage of the basic OEP. EMAC units are designed for sustained, fully-pipelined operation on accumulator load, copy, and multiply-accumulate instructions. However, instructions that store contents of the multiply-accumulate programming model can generate OEP stalls that expose the EMAC execution pipeline depth, as in the following: mac.w Ry, Rx, Acc0 move.l Acc0, Rz MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 4-11 The mov.l instruction that stores the accumulator to an integer register (Rz) stalls until the program-visible copy of the accumulator is available. Figure 4-8 shows EMAC timing. Three-cycle regBusy stall mac DSOC mov mov AGEX mac mov EMAC EX1 mac mov mac EMAC EX2 mac EMAC EX3 mac EMAC EX4 Accumulator 0 new old Figure 4-8. EMAC-Specific OEP Sequence Stall In Figure 4-8, the OEP stalls the store-accumulator instruction for 3 cycles: the depth of the EMAC pipeline minus 1. The minus 1 factor is needed because the OEP and EMAC pipelines overlap by a cycle, the AGEX stage. As the store-accumulator instruction reaches the AGEX stage where the operation is performed, the just-updated accumulator 0 value is available. As with change or use stalls between accumulators and general-purpose registers, introducing intervening instructions that do not reference the busy register can reduce or eliminate sequence-related store-MAC instruction stalls. In fact, a major benefit of the EMAC is the addition of three accumulators to minimize stalls caused by exchanges between the accumulator(s) and the general-purpose registers. 4.3.2 Data Representation MACSR[S/U,F/I] selects one of the following three modes, where each mode defines a unique operand type: • Two’s complement signed integer: In this format, an N-bit operand value lies in the range -2(N-1) < operand < 2(N-1) - 1. The binary point is right of the lsb. • Unsigned integer: In this format, an N-bit operand value lies in the range 0 < operand < 2N - 1. The binary point is right of the lsb. • Two’s complement, signed fractional: In an N-bit number, the first bit is the sign bit. The remaining bits signify the first N-1 bits after the binary point. Given an N-bit number, aN-1aN-2aN-3... a2a1a0, its value is given by the equation in Figure 4-9. N–2 value = – ( 1 ⋅ a N – 1 ) + ∑ 2 (i + 1 – N) ⋅ ai i=0 Figure 4-9. Two’s Complement, Signed Fractional Equation MCF548x Reference Manual, Rev. 5 4-12 Freescale Semiconductor EMAC Instruction Set Summary This format can represent numbers in the range -1 < operand < 1 - 2(N-1). For words and longwords, the largest negative number that can be represented is -1, whose internal representation is 0x8000 and 0x8000_0000, respectively. The largest positive word is 0x7FFF or (1 - 2-15); the most positive longword is 0x7FFF_FFFF or (1 - 2-31). 4.3.3 EMAC Opcodes EMAC opcodes are described in the ColdFire Programmer’s Reference Manual. Note the following: • Unless otherwise noted, the value of MACSR[N,Z] is based on the result of the final operation that involves the product and the accumulator. • The overflow (V) flag is handled differently. It is set if the complete product cannot be represented as a 40-bit value (this applies to 32 × 32 integer operations only) or if the combination of the product with an accumulator cannot be represented in the given number of bits. The EMAC design includes an additional product/accumulation overflow bit for each accumulator that are treated as sticky indicators and are used to calculate the V bit on each MAC or MSAC instruction. See Section 4.2.1, “MAC Status Register (MACSR).” • For the MAC design, the assembler syntax of the MAC (multiply and add to accumulator) and MSAC (multiply and subtract from accumulator) instructions does not include a reference to the single accumulator. For the EMAC, it is expected that assemblers support this syntax and that no explicit reference to an accumulator is interpreted as a reference to ACC0. These assemblers would also support syntaxes where the destination accumulator is explicitly defined. • The optional 1-bit shift of the product is specified using the notation {<< | >>} SF, where <<1 indicates a left shift and >>1 indicates a right shift. The shift is performed before the product is added to or subtracted from the accumulator. Without this operator, the product is not shifted. If the EMAC is in fractional mode (MACSR[F/I] is set), SF is ignored and no shift is performed. Because a product can overflow, the following guidelines are implemented: — For unsigned word and longword operations, a zero is shifted into the product on right shifts. — For signed, word operations, the sign bit is shifted into the product on right shifts unless the product is zero. For signed, longword operations, the sign bit is shifted into the product unless an overflow occurs or the product is zero, in which case a zero is shifted in. — For all left shifts, a zero is inserted into the lsb position. The following pseudocode explains basic MAC or MSAC instruction functionality. This example is presented as a case statement covering the three basic operating modes with signed integers, unsigned integers, and signed fractionals. Throughout this example, a comma-separated list in curly brackets, {}, indicates a concatenation operation. switch (MACSR[6:5]) /* MACSR[S/U, F/I] */ { case 0: /* signed integers */ if (MACSR.OMC == 0 || MACSR.PAVx == 0) then { MACSR.PAVx = 0 /* select the input operands */ if (sz == word) then {if (U/Ly == 1) then operandY[31:0] = {sign-extended Ry[31], Ry[31:16]} else operandY[31:0] = {sign-extended Ry[15], Ry[15:0]} if (U/Lx == 1) then operandX[31:0] = {sign-extended Rx[31], Rx[31:16]} else operandX[31:0] = {sign-extended Rx[15], Rx[15:0]} MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 4-13 } else {operandY[31:0] = Ry[31:0] operandX[31:0] = Rx[31:0] } /* perform the multiply */ product[63:0] = operandY[31:0] * operandX[31:0] /* check for product overflow */ if ((product[63:39] != 0x0000_00_0) && (product[63:39] != 0xfff f_ff_1)) then { /* product overflow */ MACSR.PAVx = 1 MACSR.V = 1 if (inst == MSAC && MACSR.OMC == 1) then if (product[63] == 1) then result[47:0] = 0x0000_7fff_ffff else result[47:0] = 0xffff_8000_0000 else if (MACSR.OMC == 1) then /* overflowed MAC, saturationMode enabled */ if (product[63] == 1) then result[47:0] = 0xffff_8000_0000 else result[47:0] = 0x0000_7fff_ffff } /* sign-extend to 48 bits before performing any scaling */ product[47:40] = {8{product[39]}} /* sign-extend */ /* scale product before combining with accumulator */ switch (SF) /* 2-bit scale factor */ { case 0: /* no scaling specified */ break; case 1: /* SF = “<< 1” */ product[40:0] = {product[39:0], 0} break; case 2: /* reserved encoding */ break; case 3: /* SF = “>> 1” */ product[39:0] = {product[39], product[39:1]} break; } if (MACSR.PAVx == 0) then {if (inst == MSAC) then result[47:0] = ACCx[47:0] - product[47:0] else result[47:0] = ACCx[47:0] + product[47:0] } /* check for accumulation overflow */ if (accumulationOverflow == 1) then {MACSR.PAVx = 1 MCF548x Reference Manual, Rev. 5 4-14 Freescale Semiconductor EMAC Instruction Set Summary MACSR.V = 1 if (MACSR.OMC == 1) then /* accumulation overflow, saturationMode enabled */ if (result[47] == 1) then result[47:0] = 0x0000_7fff_ffff else result[47:0] = 0xffff_8000_0000 } /* transfer the result to the accumulator */ ACCx[47:0] = result[47:0] } MACSR.V = MACSR.PAVx MACSR.N = ACCx[47] if (ACCx[47:0] == 0x0000_0000_0000) then MACSR.Z = 1 else MACSR.Z = 0 if ((ACCx[47:31] == 0x0000_0) || (ACCx[47:31] == 0xffff_1)) then MACSR.EV = 0 else MACSR.EV = 1 break; case 1,3: /* signed fractionals */ if (MACSR.OMC == 0 || MACSR.PAVx == 0) then { MACSR.PAVx = 0 if (sz == word) then {if (U/Ly == 1) then operandY[31:0] = {Ry[31:16], 0x0000} else operandY[31:0] = {Ry[15:0], 0x0000} if (U/Lx == 1) then operandX[31:0] = {Rx[31:16], 0x0000} else operandX[31:0] = {Rx[15:0], 0x0000} } else {operandY[31:0] = Ry[31:0] operandX[31:0] = Rx[31:0] } /* perform the multiply */ product[63:0] = (operandY[31:0] * operandX[31:0]) << 1 /* check for product rounding */ if (MACSR.R/T == 1) then { /* perform convergent rounding */ if (product[23:0] > 0x80_0000) then product[63:24] = product[63:24] + 1 else if ((product[23:0] == 0x80_0000) && (product[24] == 1)) then product[63:24] = product[63:24] + 1 } /* sign-extend to 48 bits and combine with accumulator */ /* check for the -1 * -1 overflow case */ if ((operandY[31:0] == 0x8000_0000) && (operandX[31:0] == 0x8000_0000)) then product[71:64] = 0x00 /* zero-fill */ else product[71:64] = {8{product[63]}} /* sign-extend */ if (inst == MSAC) then result[47:0] = ACCx[47:0] - product[71:24] else result[47:0] = ACCx[47:0] + product[71:24] MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 4-15 /* check for accumulation overflow */ if (accumulationOverflow == 1) then {MACSR.PAVx = 1 MACSR.V = 1 if (MACSR.OMC == 1) then /* accumulation overflow, saturationMode enabled */ if (result[47] == 1) then result[47:0] = 0x007f_ffff_ff00 else result[47:0] = 0xff80_0000_0000 } /* transfer the result to the accumulator */ ACCx[47:0] = result[47:0] } MACSR.V = MACSR.PAVx MACSR.N = ACCx[47] if (ACCx[47:0] == 0x0000_0000_0000) then MACSR.Z = 1 else MACSR.Z = 0 if ((ACCx[47:39] == 0x00_0) || (ACCx[47:39] == 0xff_1)) then MACSR.EV = 0 else MACSR.EV = 1 break; case 2: /* unsigned integers */ if (MACSR.OMC == 0 || MACSR.PAVx == 0) then { MACSR.PAVx = 0 /* select the input operands */ if (sz == word) then {if (U/Ly == 1) then operandY[31:0] = {0x0000, else operandY[31:0] = {0x0000, if (U/Lx == 1) then operandX[31:0] = {0x0000, else operandX[31:0] = {0x0000, } else {operandY[31:0] = Ry[31:0] operandX[31:0] = Rx[31:0] } Ry[31:16]} Ry[15:0]} Rx[31:16]} Rx[15:0]} /* perform the multiply */ product[63:0] = operandY[31:0] * operandX[31:0] /* check for product overflow */ if (product[63:40] != 0x0000_00) then { /* product overflow */ MACSR.PAVx = 1 MACSR.V = 1 if (inst == MSAC && MACSR.OMC == 1) then result[47:0] = 0x0000_0000_0000 else if (MACSR.OMC == 1) then /* overflowed MAC, saturationMode enabled */ MCF548x Reference Manual, Rev. 5 4-16 Freescale Semiconductor EMAC Instruction Set Summary result[47:0] = 0xffff_ffff_ffff } /* zero-fill to 48 bits before performing any scaling */ product[47:40] = 0 /* zero-fill upper byte */ /* scale product before combining with accumulator */ switch (SF) /* 2-bit scale factor */ { case 0: /* no scaling specified */ break; case 1: /* SF = “<< 1” */ product[40:0] = {product[39:0], 0} break; case 2: /* reserved encoding */ break; case 3: /* SF = “>> 1” */ product[39:0] = {0, product[39:1]} break; } /* combine with accumulator */ if (MACSR.PAVx == 0) then {if (inst == MSAC) then result[47:0] = ACCx[47:0] - product[47:0] else result[47:0] = ACCx[47:0] + product[47:0] } /* check for accumulation overflow */ if (accumulationOverflow == 1) then {MACSR.PAVx = 1 MACSR.V = 1 if (inst == MSAC && MACSR.OMC == 1) then result[47:0] = 0x0000_0000_0000 else if (MACSR.OMC == 1) then /* overflowed MAC, saturationMode enabled */ result[47:0] = 0xffff_ffff_ffff } /* transfer the result to the accumulator */ ACCx[47:0] = result[47:0] } MACSR.V = MACSR.PAVx MACSR.N = ACCx[47] if (ACCx[47:0] == 0x0000_0000_0000) then MACSR.Z = 1 else MACSR.Z = 0 if (ACCx[47:32] == 0x0000) then MACSR.EV = 0 else MACSR.EV = 1 break; } MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 4-17 MCF548x Reference Manual, Rev. 5 4-18 Freescale Semiconductor Chapter 5 Memory Management Unit (MMU) This chapter describes the ColdFire virtual memory management unit (MMU), which provides virtual-to-physical address translation and memory access control. The MMU consists of memory-mapped control, status, and fault registers that provide access to translation-lookaside buffers (TLBs). Software can control address translation and access attributes of a virtual address by configuring MMU control registers and loading TLBs. With software support, the MMU provides demand-paged, virtual addressing. 5.1 Features The MMU has the following features: • MMU memory-mapped control, status, and fault registers — Support a flexible, software-defined virtual environment — Provide control and maintenance of TLBs — Provide fault status and recovery information functions • Separate, 32-entry, fully associative instruction and data TLBs (Harvard TLBs) — Resides in the controller — Operates in parallel with the memories — Suffers no performance penalty on TLB hits — Supports 1-, 4-, and 8-Kbyte and 1-Mbyte page sizes concurrently — Contains register-based TLB entries • Core extensions: — User stack pointer — All access error exceptions are precise and recoverable • Harvard TLB provides 97% of baseline performance on large embedded applications using equivalent V4 without MMU support as a baseline. 5.2 Virtual Memory Management Architecture The ColdFire memory management architecture provides a demand-paged, virtual-address environment with hardware address translation acceleration. It supports supervisor/user, read, write, and execute permission checking on a per-memory request basis. The architecture defines the MMU TLB, associated control logic, TLB hit/miss logic, address translation based on the TLB contents, and access faults due to TLB misses and access violations. It intentionally leaves some virtual environment details undefined to maximize the software-defined flexibility. These include the exact structure of the memory-resident pointer descriptor/page descriptor tables, the base registers for these tables, the exact information stored in the tables, the methodology (if any) for maintenance of access, and written information on a per-page basis. 5.2.1 MMU Architecture Features To add optional virtual addressing support, demand-page support, permission checking, and hardware address translation acceleration to the ColdFire architecture, the MMU architecture features the following: • Addresses from the core to the MMU are treated as physical or virtual addresses. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 5-1 • • • • 5.2.2 The address access control logic, address attribute logic, memories, and controller function as in previous ColdFire versions with the addition of the MMU. The MMU, its TLB, and associated control reside in the logic. The MMU appears as a memory-mapped device in the space. Information for access error fault processing is stored in the MMU. A precise fault (transfer error acknowledge) signals the core on translation (TLB miss) and access faults. The core supports an instruction restart model for this fault class. Note that this structure uses the existing ColdFire access error fault vector and needs no new ColdFire exception stack frames. The following additions are made to the memory access control to better support the fault processing and memory maintenance necessary for this virtual addressing environment. These additions improve memory performance and functionality for physical and virtual address environments: — New supervisor-protect bits to the access control registers (ACRs) and the cache control register (CACR) — Improved addressing of the ACRs MMU Architecture Location Figure 5-1 shows the placement of the MMU/TLB hardware. It follows a traditional model in which it is closely coupled to the processor local-memory controllers. MCF548x Reference Manual, Rev. 5 5-2 Freescale Semiconductor Virtual Memory Management Architecture Instruction Fetch Pipeline J IAG Branch Cache KC1 IC1 KC2 IC2 Branch Accel. Instruction Memory Physical KC1 IED IB Memory Management Unit (MMU) Operand Execution Pipeline DS Physical KC1 DS J OAG Data Memory KC1 OC1 KC2 OC2 EX M Bus K2M EMAC Misalignment Module FPU DA BDM DSCLK DSI DSDO DDATA PSTDDATA PSTCLK Figure 5-1. CF4e Processor Core Block with MMU 5.2.3 MMU Architecture Implementation This section describes ColdFire design additions and changes for the MMU architecture. It includes precise faults, MMU access, virtual mode, virtual memory references, instruction and data cache addresses, supervisor/user stack pointers, access error stack frame additions, expanded control register space, ACR address improvements, supervisor protection, and debugging in a virtual environment. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 5-3 5.2.3.1 Precise Faults The MMU architecture performs virtual-to-physical address translation and permission checking in the core. To support demand-paging, the core design provides a precise, recoverable fault for all references. 5.2.3.2 MMU Access The MMU TLB control registers are memory-mapped. The TLB entries are read and written indirectly through the MMU control registers. The memory space for these resources is defined by a new supervisor program model register, the MMU base address register (MMUBAR). This register defines a supervisor-mode, data-only space. It has the highest priority for the data address mode determination. 5.2.3.3 Virtual Mode Every instruction and data reference is either a virtual or physical address mode access. All addresses for special mode (interrupt acknowledges, emulator mode operations, etc.) accesses are physical. All addresses are physical if the MMU is not enabled. If the MMU is present and enabled, the address mode for normal accesses is determined by the MMUBAR, RAMBARs, and ACRs in the priority order listed. Addresses that hit in the MMUBAR, RAMBARs, and ACRs are treated as physical references. These addresses are not translated and their address attributes are sourced from the highest priority mapping register they hit. If an address hits none of these mapping registers, it is a virtual address and is sent to the MMU. If the MMU is enabled, the default CACR information is not used. 5.2.3.4 Virtual Memory References The ColdFire MMU architecture references the MMU for all virtual mode accesses to the . MMU, SRAM and ACR memory spaces are treated as physical address spaces and all permissions that apply to these spaces are contained in the respective mapping register. The virtual mode access either hits or misses in the TLB of the MMU. A TLB miss generates an access fault in the processor, allowing software to either load the appropriate translation into the TLB and restart the faulting instruction or abort the process. Each TLB hit checks permissions based on the access control information in the referenced TLB entry. 5.2.3.5 Instruction and Data Cache Addresses For a given page size, virtual address bits that reference within a page are called the in-page address. All bits above this are the virtual page number. Likewise, the physical address has a physical page number and in-page address bits. Virtual and physical in-page address bits are the same; the MMU translates the virtual page number to the physical page number. Instruction and data caches are accessed with the untranslated address. The translated address is used for cache allocation. That is, caches are virtual-address accessed and physical-address tagged. If instruction and data cache addresses are not larger than the in-page address for the smallest active MMU page, the cache is considered physically accessed; if they are larger, the cache can have aliasing problems between virtual and cache addresses. Software handles these problems by forcing the virtual address to be equal to the physical address for those bits addressing the cache, but above the in-page address of the smallest active page size. The number of these bits depends on cache and page sizes. Caches are addressed with the virtual address, because the cache uses synchronous memory elements, and an access starts at the rising-clock edge of the first pipeline stage. The MMU provides a physical address midway through this cycle. If the cache set address has fewer bits than the in-page address, the cache is considered physically addressed because these bits are the same in the virtual and physical addresses. If the cache set address has MCF548x Reference Manual, Rev. 5 5-4 Freescale Semiconductor Virtual Memory Management Architecture more bits than the in-page address, one or more of the low-order virtual page number bits are used to address the cache. The MMU translates these bits; the resulting low-order physical page number bits are used to determine cache hits. Address aliasing problems occur when two virtual addresses access one physical page. This is generally allowed and, if the page is cacheable, one coherent copy of the page image is mapped in the cache at any time. If multiple virtual addresses pointing to the same physical address differ only in the low-order virtual page number bits, conflicting copies can be allocated. For an 8-Kbyte, 4-way, set-associative cache with a 16-byte line size, the cache set address uses address bits 10–4. If virtual addresses 0x0_1000 and 0x0_1400 are mapped to physical address 0x0_1000, using virtual address 0x0_1000 loads cache set 0x00; using virtual address 0x0_1400 loads cache set 0x40. This puts two copies of the same physical address in the cache making this memory space not coherent. To avoid this problem, software must force low-order virtual page number bits to be equal to low-order physical address bits for all bits used to address the cache set. 5.2.3.6 Supervisor/User Stack Pointers To isolate supervisor and user modes, CF4e implements two A7 register stack pointers, one for supervisor mode (SSP) and one for user mode (USP). Two former M68000 family privileged instructions to load and store the user stack pointer are restored in the instruction set architecture. 5.2.3.7 Access Error Stack Frame accesses that fault (that is, terminate with a transfer error acknowledge) generate an access error exception. MMU TLB misses and access violations use the same fault. To quickly determine if a fault was due to a TLB miss or another type of access error, new fault status field (FS) encodings in the exception stack frame signal TLB misses on the following: • Instruction fetch • Instruction extension fetch • Data read • Data write See Section 5.4.3, “Access Error Stack Frame Additions,” for more information. 5.2.3.8 Expanded Control Register Space The MMU base address register (MMUBAR) is added for ColdFire virtual mode. Like other control registers, it can be accessed from the debug module or written using the privileged MOVEC instruction. See Section 5.5.3.1, “MMU Base Address Register (MMUBAR).” 5.2.3.9 Changes to ACRs and CACR New ACR and CACR bits, Table 5-1, improve address granularity and supervisor mode protection. These improvements are not necessary to implement the ColdFire MMU, but they improve memory functionality for physical and virtual address environments. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 5-5 Table 5-1. New ACR and CACR Bits Bits Name Description ACRn[10] AMM Address mask mode. Determines access to the associated address space. 0 The ACR hit function is the same as previous versions, allowing control of a 16-Mbyte or greater memory region. 1 The upper 8 bits of the address and ACR are compared without a mask function; bits 23–20 of the address and ACR are compared masked by ACR[19–16], allowing control of a 1- to 16-Mbyte region. Reset value is 0. ACRn[3] SP Supervisor protect. Determines access to the associated address space. 0 Supervisor and user access allowed. 1 Only supervisor access allowed. Attempted user access causes an access error exception. Reset value is 0. CACR[23] DDSP Default data supervisor protect. Determines access to the associated data space. 0 Supervisor and user access allowed. 1 Only supervisor access allowed. Attempted user access causes an access error exception. Reset value is 0. CACR[7] DISP Default instruction supervisor protect. Determines access to the associated instruction space. 0 Supervisor and user access allowed. 1 Only supervisor access allowed. Attempted user access causes access error exception Reset value is 0. 5.2.3.10 ACR Address Improvements ACRs provide a 16-Mbyte address window. For a given request address, if the ACR is valid and the request mode matches the mode specified in the supervisor mode field, ACRn[S], hit determination is specified as follows: ACRx_Hit = 0; if ((address[31:24] & ~ACRn[23:16]) == (ACRn[31:24] & ~ACRn[23:16])) ACRx_Hit = 1; With this hit function, ACRs can assign address attributes for user or supervisor requests to memory spaces of at least 16 Mbytes (through the address mask). With the MMU definition, the ACR hit function is improved by the address mask mode bit (ACRn[AMM]), which supports finer address granularity. See Table 5-1. The revised hit determination becomes the following: ACRx_Hit = 0; if (ACRn[10] == 1) if ((address[31–24] == ACRn[31–24])) && ((address[23–20] & ~ACRn[19–16]) == (ACRn[23–20] & ~ACRn[19–16]))) ACRx_Hit = 1; else if (address[31–24] & ~ACRn[23–16]) == (ACRn[31–24] & ~ACRn[23–16])) ACRx_Hit = 1; MCF548x Reference Manual, Rev. 5 5-6 Freescale Semiconductor Debugging in a Virtual Environment 5.2.3.11 Supervisor Protection Each instruction or data reference is either a supervisor or user access. The CPU’s status register supervisor bit (SR[S]) determines the operating mode. New ACR and CACR bits protect supervisor space. See Table 5-1. 5.3 Debugging in a Virtual Environment To support debugging in a virtual environment, numerous enhancements are implemented in the ColdFire debug architecture. These enhancements are collectively called Debug revision D and primarily relate to the addition of an 8-bit address space identifier (ASID) to yield a 40-bit virtual address. This expansion affects two major debug functions: • The ASID is optionally included in the hardware breakpoint registers specification. For example, the four PC breakpoint registers are expanded by 8 bits each, so that a specific ASID value can be part of the breakpoint instruction address. Likewise, data address/data breakpoint registers are expanded to include an ASID value. The new control registers define whether and how the ASID is included in the breakpoint comparison trigger logic. • The debug module implements the concept of ownership trace in which an ASID value can be optionally displayed as part of real-time trace. When enabled, real-time trace displays instruction addresses on any change-of-flow instruction that is not absolute or PC-relative. For Debug revision D architecture, the address display is expanded to optionally include ASID contents, thus providing the complete instruction virtual address on these instructions. Additionally, when a Sync_PC serial BDM command is loaded from the external development system, the processor displays the complete virtual instruction address, including the 8-bit ASID value. The MMU control registers are accessible through serial BDM commands. See Chapter 8, “Debug Support.” 5.4 Virtual Memory Architecture Processor Support To support the MMU, enhancements have been made to the exception model, the stack pointers, and the access error stack frame. 5.4.1 Precise Faults To support demand-paging, all memory references require precise, recoverable faults. The ColdFire instruction restart mechanism ensures that a faulted instruction restarts from the beginning of execution; that is, no internal state information is saved when an exception occurs and none is restored when the handler ends. Given the PC address defined in the exception stack frame, the processor reestablishes program execution by transferring control to the given location as part of the RTE (return from exception) instruction. For a detailed description, see Section 3.9, “Precise Faults.” 5.4.2 Supervisor/User Stack Pointers To provide the required isolation between these operating modes as dictated by a virtual memory management scheme, a user stack pointer (A7–USP) is added. The appropriate stack pointer register (SSP, USP) is accessed as a function of the processor’s operating mode. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 5-7 In addition, the following two privileged M68000 family instructions to load/store the USP are added to the ColdFire instruction set architecture: mov.l mov.l Ay,USP USP,Ax # move to USP: opcode = 0x4E6{0-7} # move from USP: opcode = 0x4E6{8–F} The address register number is encoded in the three low-order bits of the opcode. These instructions are described in detail in Section 5.7, “MMU Instructions.” 5.4.3 Access Error Stack Frame Additions ColdFire exceptions generate a standard 2-longword stack frame, signaling the contents of the SR and PC at the time of the exception, the exception type, and a 4-bit fault status field (FS). The first longword contains the 16-bit format/vector word (F/V) and the 16-bit status register. The second contains the 32-bit program counter address of the faulted instruction. 31 A7 → 28 FORMAT 27 26 25 FS[3–2] 18 VEC[7–0] + 0x04 17 16 15 FS[1–0] 0 STATUS REGISTER PROGRAM COUNTER [31–0] Figure 5-2. Exception Stack Frame The FS field is used for access and address errors. To optimize TLB miss exception handling, new FS encodings (Table 5-2) allow quick error classification. Table 5-2. Fault Status Encodings FS[3:0] 0000 0001, 001x Definition Not an access or address error Reserved 0100 Error (for example, protection fault) on instruction fetch 0101 TLB miss on opword of instruction fetch (New in CF4e) 0110 TLB miss on extension word of instruction fetch (New in CF4e) 0111 IFP access error while executing in emulator mode (New in CF4e) 1000 Error on data write 1001 Attempted write of protected space 1010 TLB miss on data write (New in CF4e) 1011 Reserved 1100 Error on data read 1101 Attempted read, read-modify-write of protected space (New in CF4e) 1110 TLB miss on data read, or read-modify-write (New in CF4e) 1111 OEP access error while executing in emulator mode (New in CF4e) MCF548x Reference Manual, Rev. 5 5-8 Freescale Semiconductor MMU Definition 5.5 MMU Definition The ColdFire MMU provides a virtual address, demand-paged memory architecture. The MMU supports hardware address translation acceleration using software-managed TLBs. It enforces permission checking on a per-memory request basis, and has control, status, and fault registers for MMU operation. 5.5.1 Effective Address Attribute Determination The ColdFire core generates an effective memory address for all instruction fetches and data read and write memory accesses. The previous ColdFire memory access control model was based strictly on physical addresses. Every memory request address is a physical address that is analyzed by this memory access control logic and assigned address attributes, which include the following: • Cache mode • SRAM enable information • Write protect information • Write mode information These attributes control processing of the memory request. The address itself is not affected by memory access control logic. Instruction and data references base effective address attributes and access mode on the instruction type and the effective address. Accesses are of the following two types: • Special mode accesses, including interrupt acknowledges, reads/writes to program-visible control registers (such as CACR, ROMBARs, RAMBARs, and ACRs), cache control commands (CPUSHL and INTOUCH), and emulator mode operations. These accesses have the following attributes: — Non-cacheable — Precise — No write protection Unless the CPU space/IACK mask bit is set, interrupt acknowledge cycles and emulator mode operations are allowed to hit in RAMBARs and ROMBARs. All other operations are normal mode accesses. • Normal mode accesses. For these accesses, an effective cache mode, precision and write-protection are calculated for each request. For data, a normal mode access address is compared with the following priority, from highest to lowest: RAMBAR0, RAMBAR1, ROMBAR0, ROMBAR1, ACR0, and ACR1. If no match is found, default attributes in the CACR are used. The priority for instruction accesses is RAMBAR0, RAMBAR1, ROMBAR0, ROMBAR1, ACR2, and ACR3. Again, if no match is found, default CACR attributes are used. Only the test-and-set (TAS) instruction can generate a normal mode access with implied cache mode and precision. TAS is a special, byte-sized, read-modify-write instruction used in synchronization routines. A TAS data access that does not hit in the RAMBARs is non-cacheable and precise. TAS uses the normal effective write protection. The ColdFire MMU is an optional enhancement to the memory access control. If the MMU is present and enabled, it adds two factors for calculating effective address attributes: • MMUBAR defines a memory-mapped, privileged data-only space with the highest priority in effective address attribute calculation for the data (that is, the MMUBAR has priority over RAMBAR0). MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 5-9 • If virtual mode is enabled, any normal mode access that does not hit in the MMUBAR, RAMBARs, ROMBARs, or ACRs is considered a normal mode virtual address request and generates its access attributes from the MMU. For this case, the default CACR address attributes are not used. The MMU also uses TLB contents to perform virtual-to-physical address translation. 5.5.2 MMU Functionality The MMU provides virtual-to-physical address translation and memory access control. The MMU consists of memory-mapped, control, status, and fault registers, and a TLB that can be accessed through MMU registers. Supervisor software can access these resources through MMUBAR. Software can control address translation and access attributes of a virtual address by configuring MMU control registers and loading the MMU’s TLB, which functions as a cache, associating virtual addresses to corresponding physical addresses and providing access attributes. Each TLB entry maps a virtual page. Several page sizes are supported. Features such as clear-all and probe-for-hit help maintain TLBs. Fault-free, virtual address accesses that hit in the TLB incur no pipeline delay. Accesses that miss the TLB or hit the TLB but violate an access attribute generate an access error exception. On an access error, software can reference address and information registers in the MMU to retrieve data. Depending on the fault source, software can obtain and load a new TLB entry, modify the attributes of an existing entry, or abort the faulting process. 5.5.3 MMU Organization Access to the MMU memory-mapped region is controlled by MMUBAR, a 32-bit supervisor control register at 0x008 that is accessed using MOVEC or the serial BDM debug port. The ColdFire Programmers Reference Manual describes the MOVEC instruction. 5.5.3.1 MMU Base Address Register (MMUBAR) Figure 5-3 shows MMUBAR. The default reset state is an invalid MMUBAR, so that the MMU is disabled and the memory-mapped space is not visible. 31 30 29 28 27 26 25 R 24 23 22 21 20 19 18 17 16 BA W Reset R 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 V 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 W Reset Reg Addr CPU + 0x008 Figure 5-3. MMU Base Address Register (MMUBAR) Table 5-3 describes MMU base address register fields. MCF548x Reference Manual, Rev. 5 5-10 Freescale Semiconductor MMU Definition Table 5-3. MMUBAR Field Descriptions Bits Name Description 31–16 BA Base address. Defines the base address for the 64-Kbyte address space mapped to the MMU. 15–1 — Reserved, should be cleared. Writes are ignored and reads return zeros. 0 V Valid. Indicates when MMUMBAR contents are valid. BA is not used unless V is set. 0 MMUBAR contents are not valid. 1 MMUBAR contents are valid. 5.5.3.2 MMU Memory Map MMUBAR holds the base address for the 64-Kbyte MMU memory map, shown in Table 5-4. The MMU memory map area is not visible unless the MMUBAR is valid and must be referenced aligned. A large portion of the map is reserved for future use. Table 5-4. MMU Memory Map Offset from MMUBAR + 0x0000 MMU control register (MMUCR) + 0x0004 MMU operation register (MMUOR) + 0x0008 MMU status register (MMUSR) + 0x000C Reserved + 0x0010 MMU fault, test, or TLB address register (MMUAR) + 0x0014 MMU read/write TLB tag register (MMUTR) + 0x0018 MMU read/write TLB data register (MMUDR) + 0x001C–0xFFFC 1 Name Reserved1 May be used for implementation-specific information/control registers The address space ID (ASID) is located in a CPU space control register. The 8-bit ASID value located in the low order byte of a 32-bit supervisor control register, mapped into CPU space at address 0x003 and accessed using a MOVEC instruction. The ColdFire Family Programmer’s Reference Manual describes MOVEC. This 8-bit field is the current user ASID. The ASID is an extension to the virtual address. Address space 0x00 may be reserved for supervisor mode. See address space mode functionality in Section 5.5.3.3, “MMU Control Register (MMUCR).” The other 255 address spaces are used to tag user processes. The TLB entry ASID values are compared to this value for user mode unless the TLB entry is marked shared (MMUTR[SG] is set). The TLB entry ASID value may be compared to 0x00 for supervisor accesses. 5.5.3.3 MMU Control Register (MMUCR) MMUCR, Figure 5-4, has the address space mode and virtual mode enable bits. The user must force pipeline synchronization after writing to this register. Therefore, all writes to this register must be immediately followed by a NOP instruction. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 5-11 R 31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 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 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 ASM EN 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 W Reset R W Reset Reg Addr MMUBAR + 0x000 Figure 5-4. MMU Control Register (MMUCR) Table 5-5 describes MMUCR fields. Table 5-5. MMUCR Field Descriptions Bits Name 31–2 — 1 ASM 0 EN 5.5.3.4 Description Reserved, should be cleared. Writes are ignored and reads return zeros. Address space mode. Controls how the address space ID is used for TLB hits. 0 TLB entry ASID values are compared to the address space ID register value for user or supervisor mode unless the TLB entry is marked shared (MMUTR[SG] = 1). The address space ID register value is the effective address space for all requests, supervisor and user. 1 Address space 0x00 is reserved for supervisor mode and the effective address space is forced to 0x00 for all supervisor accesses. The other 255 address spaces are used to tag user processes. The TLB entry ASID values are compared to the address space ID register for user mode unless the TLB entry is marked shared (SG = 1). The TLB entry ASID value is always compared to 0x00 for supervisor accesses. This allows two levels of sharing. All users but not the supervisor share an entry if SG = 1and ASID ¦ 0. All users and the supervisor share an entry if SG = 1 and ASID = 0 Virtual mode enabled. Indicates when virtual mode is enabled. 0 Virtual mode is disabled. 1 Virtual mode is enabled. MMU Operation Register (MMUOR) Figure 5-5 shows the MMUOR. MCF548x Reference Manual, Rev. 5 5-12 Freescale Semiconductor MMU Definition 31 30 29 28 27 26 25 24 R 23 22 21 20 19 18 17 16 AA W Reset R 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 0 0 0 0 0 0 R/W ACC UAA 0 0 0 0 0 0 0 0 0 0 STLB CA CNL CAS ITLB ADR W Reset Reg Addr 0 0 0 0 0 0 MMUBAR + 0x004 Figure 5-5. MMU Operation Register (MMUOR) Table 5-6 describes MMUOR fields. Table 5-6. MMUOR Field Descriptions Bits Name Description 31–16 AA TLB allocation address. This read-only field is maintained by MMU hardware. Its range and format depend on the TLB implementation (specific TLB size in entries, associativity, and organization). The access TLB function can use AA to read or write the addressed TLB entry. The MMU loads AA on the following three events: • On DTLB access errors, it loads the address of the TLB entry that caused the error. • If UAA is set, it loads the address of the TLB entry chosen by the MMU for replacement. • If STLB is set, it uses the data in MMUAR to search the TLB and if the TLB hits, loads the address of the TLB entry that hits, or if the TLB misses, loads the TLB entry chosen by the MMU for replacement. The MMU never picks a locked entry for replacement, and TLB hits of locked entries do not update hardware replacement algorithm information. This is so access error handlers mapped with locked TLB entries do not influence the replacement algorithm. Further, TLB search operations do not update the hardware replacement algorithm information while TLB writes (loads) do update the hardware replacement algorithm information. The algorithm used to choose the allocation address depends on the TLB implementation (such as LRU, round-robin, pseudo-random). 15–9 — Reserved, should be cleared. Writes are ignored and reads return zeros. 8 STLB 7 CA 6 CNL Search TLB. STLB always reads as zero. 0 No operation 1 The MMU searches the TLB using data in MMUAR. This operation updates the probe TLB hit bit in the status register plus loads the AA field as described above. Clear all TLB entries. CA always reads as zero. 0 No operation 1 Clear all TLB entries and all hardware TLB replacement algorithm information. Clear all non-locked TLB entries. Setting CNL clears all TLB entries that do not have their locked bit set. CNL always reads as zero. 0 No operation 1 Clear all non-locked TLB entries. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 5-13 Table 5-6. MMUOR Field Descriptions (Continued) Bits Name 5 CAS Clear all non-locked TLB entries that match ASID. CAS is always reads as a zero. 0 No operation 1 Clear all non-locked TLB entries that match ASID register. 4 ITLB ITLB operation. Used by TLB search and access operations that use the TLB allocation address. 0 The MMU uses the DTLB to search or update the allocation address. 1 The MMU uses the ITLB for searches and updates of the allocation address. 3 ADR TLB address select. Indicates which address to use when accessing the TLB. 0 Use the TLB allocation address for the TLB address. 1 Use MMUAR for the TLB address. 2 R/W TLB access read/write select. Indicates whether to do a read or a write when accessing the TLB. 0 Write 1 Read 1 ACC MMU TLB access. This bit always reads as a zero. STLB is used for search operations. 0 No operation. ACC should be a zero to search the TLB. 1 The MMU reads or writes the TLB depending on R/W. For TLB reads, TLB tag and data results are loaded into MMUTR and MMUDR. For TLB writes, the contents of these registers are written to the TLB. The TLB is accessed using the TLB allocation address if ADR is zero or using MMUAR if ADR is set. 0 UAA Update allocation address. UAA always reads as a zero. 0 No operation 1 MMU updates the allocation address field with the MMU’s choice for the allocation address in the ITLB or DTLB depending on the ITLB instruction operation bit. 5.5.3.5 Description MMU Status Register (MMUSR) MMUSR, Figure 5-6, is updated on all data access faults and search TLB operations. R 31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 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 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 0 0 0 0 0 0 0 0 0 SPF RF WF 0 HIT 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 W Reset R W Reset Reg Addr MMUBAR + 0x008 Figure 5-6. MMU Status Register (MMUSR) Table 5-7 describes MMUSR fields. MCF548x Reference Manual, Rev. 5 5-14 Freescale Semiconductor MMU Definition Table 5-7. MMUSR Field Descriptions Bits Name 31–6 — 5 SPF Supervisor protect fault. Indicates if the last data fault was a user mode access that hit in a TLB entry that had its supervisor protect bit set. 0 Last data access fault did not have a supervisor protect fault. 1 Last data access fault had a supervisor protect fault. 4 RF Read access fault. Indicates if the last data fault was an data read access that hit in a TLB entry that did not have its read bit set. 0 Last data access fault did not have a read protect fault. 1 Last data access fault had a read protect fault. 3 WF Write access fault. Indicates if the last data fault was an data write access that hit in a TLB entry that did not have its write bit set. 0 Last data access fault did not have a write protect fault. 1 Last data access fault had a write protect fault. 2 — Reserved, should be cleared. Writes are ignored and reads return zeros. 1 HIT 0 — 5.5.3.6 Description Reserved, should be cleared. Writes are ignored and reads return zeros. Search TLB hit. Indicates if the last data fault or the last search TLB operation hit in the TLB. 0 Last data access fault or search TLB operation did not hit in the TLB. 1 Last data access fault or search TLB operation hit in the TLB. Reserved, should be cleared. Writes are ignored and reads return zeros. MMU Fault, Test, or TLB Address Register (MMUAR) The MMUAR format, Figure 5-7, depends on how the register is used. 31 30 29 28 27 26 25 24 R 23 22 21 20 19 18 17 16 FA W Reset 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 0 0 0 0 0 0 0 R FA W Reset 0 0 0 0 0 Reg Addr 0 0 0 MMUBAR + 0x010 Figure 5-7. MMU Fault, Test, or TLB Address Register (MMUAR) Table 5-8 describes MMUAR fields. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 5-15 Table 5-8. MMUAR Field Descriptions Bits Name Description 31–0 FA Form address. Written by the MMU with the virtual address on DTLB misses and access faults. For this case, all 32 bits are address bits. This register may be written with a virtual address and address attribute information for searching the TLB (MMUCR[STLB]). For this case, FA[31–1] are the virtual page number and FA[0] is the supervisor bit. The current ASID is used for the TLB search. MMUAR can also be written with a TLB address for use with the access TLB function (using MMUCR[ACC]). 5.5.3.7 MMU Read/Write Tag and Data Entry Registers (MMUTR and MMUDR) Each TLB entry consists of a 32-bit TLB tag entry and a 32-bit TLB data entry. TLB entries are referenced through MMUTR and MMUDR. For read TLB accesses, the contents of the TLB tag and data entries referenced by the allocation address or MMUAR are loaded in MMUTR and MMUDR. TLB write accesses place MMUTR and MMUDR contents into the TLB tag and data entries defined by the allocation address or MMUAR. MMUTR, Figure 5-8, contains the virtual address tag, the address space ID (ASID), a shared page indicator, and the valid bit. 31 30 29 28 27 26 25 24 R 23 22 21 20 19 18 17 16 VA W Reset 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 SG V 0 0 R VA ID W Reset 0 0 0 0 0 Reg Addr 0 0 0 0 0 0 0 0 0 MMUBAR + 0x014 Figure 5-8. MMU Read/Write TLB Tag Register (MMUTR) Table 5-9 describes MMUTR fields. Table 5-9. MMUTR Field Descriptions Bits Name Description 31–10 VA Virtual address. Defines the virtual address mapped by this entry. The number of bits used in the TLB hit determination depends on the page size field in the corresponding TLB data entry. 9–2 ID Address space ID (ASID). This extension to the virtual address marks this entry as part of 1 of 256 possible address spaces. Address space 0x00 can be reserved for supervisor mode. The other 255 address spaces are used to tag user processes. TLB entry ASID values are compared to the ASID register value for user mode unless the TLB entry is marked shared (SG = 1). The TLB entry ASID value may be compared to 0x00 for supervisor accesses or to the ASID. The description of MMUCR[ASM] in Table 5-5 gives details on supervisor mode and ASID compares. MCF548x Reference Manual, Rev. 5 5-16 Freescale Semiconductor MMU Definition Table 5-9. MMUTR Field Descriptions (Continued) Bits Name Description 1 SG Shared global. Indicates when the entry is shared among user address spaces. If an entry is shared, its ASID is not part of the TLB hit determination for user accesses. 0 This entry is not shared globally. 1 This entry is shared globally. Note that the ASID can be used to determine supervisor mode hits to allow two sharing levels. If SG and MMUCR[ASM] are set and the ASID is not zero, all users (but not the supervisor) share an entry. If SG and MMUCR[ASM] are set and the ASID is zero, all users and the supervisor share an entry. The description of ASM in Table 5-5 details supervisor mode and ASID compares. 0 V Valid. Indicates when the entry is valid. Only valid entries generate a TLB hit. 0 Entry is not valid. 1 Entry is valid. MMUDR, Figure 5-9, contains the physical address, page size, cache mode field, supervisor-protect bit, read, write, execute permission bits, and lock-entry bit. 31 30 29 28 27 26 25 24 R 23 22 21 20 19 18 17 16 PA W Reset 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 SP R W X LK 0 0 0 0 0 0 0 R PA SZ CM W Reset 0 0 0 0 0 0 Reg Addr 0 0 0 0 MMUBAR + 0x014 Figure 5-9. MMU Read/Write TLB Data Register (MMUDR) Table 5-10 describes MMUDR fields. Table 5-10. MMUDR Field Descriptions Bits Name Descriptions 31–10 PA Physical address. Defines the physical address which is mapped by this entry. The number of bits used to build the effective physical address if this TLB entry hits depends on the page size field. 9–8 SZ Page size. Page size for this entry: 00 01 10 11 1 Mbyte: VA[31–20] used for TLB hit 4 Kbytes VA[31–12] used for TLB hit 8 Kbytes VA[31–13] used for TLB hit 1 Kbyte VA[31–10] used for TLB hit MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 5-17 Table 5-10. MMUDR Field Descriptions (Continued) 5.5.4 Bits Name Descriptions 7–6 CM Cache mode. If a Harvard TLB implementation is used, CM0 is a don’t care for the ITLB. CM is ignored on writes and always reads as zero for the ITLB. Instruction cache modes: 1x Page is non-cacheable. 0x Page is cacheable. Data cache modes 00 Page is cacheable writethrough. 01 Page is cacheable copyback. 10 Page is non-cacheable precise. 11 Page is non-cacheable imprecise. 5 SP Supervisor protect. Controls user mode access to the page mapped by this entry. 0 Entry is not supervisor protected. 1 Entry is supervisor protected. An attempted user mode access that matches this entry generates an access error exception. 4 R Read access enable. Indicates if data read accesses to this entry are allowed. If a Harvard TLB implementation is used, this bit is a don’t care for the ITLB. This bit is ignored on writes and always reads as zero for the ITLB. 0 Do not allow data read accesses. Attempted data read accesses that match this entry generate an access error exception. 1 Allow data read accesses. 3 W Write access enable. Indicates if data write accesses are allowed to this entry. If separate ITLB and DTLBs) are used, W is a don’t care for the ITLB. W is ignored on writes and reads as zero for the ITLB. 0 Do not allow data write accesses. Attempted data write accesses that match this entry generate an access error exception. 1 Allow data write accesses. 2 X Execute access enable. Indicates if instruction fetches to this entry are allowed. If separate ITLB and DTLBs are is used, X is a don’t care for the DTLB. X is ignored on writes and reads as zero for the DTLB. 0 Do not allow instruction fetches. Attempted instruction fetches that match this entry cause an access error exception. 1 Allow instruction fetch accesses. 1 LK Lock entry bit. Indicates if this entry is included in the replacement algorithm. TLB hits of locked entries do not update replacement algorithm information. 0 Include this entry when determining the best entry for a TLB allocation. 1 Do not allow this entry to be selected by the replacement algorithm. 0 — Reserved, should be cleared. Writes are ignored and reads return zeros. MMU TLB Each TLB entry consists of two 32-bit fields. The first is the TLB tag entry, and the second is the TLB data entry. TLB size and organization are implementation dependent. TLB entries can be read and written through MMU registers. TLB contents are unaffected by reset. MCF548x Reference Manual, Rev. 5 5-18 Freescale Semiconductor MMU Definition 5.5.5 MMU Operation The processor sends instruction fetch requests and data read/write requests to the MMU in the instruction and operand address generation cycles (IAG and OAG). The controller and memories occupy the next two pipeline stages, instruction fetch cycles 1 and 2 (IC1 and IC2) and operand fetch cycles 1 and 2 (OC1 and OC2). For late writes, optional data pipeline stages are added to the controller as well as any writable memories. Table 5-11 shows the association between memory pipeline stages and the processor’s pipeline structures, shown in Figure 5-1. . Table 5-11. Version 4 Memory Pipelines Memory Pipeline Stage Instruction Fetch Pipeline Operand Execution Pipeline J stage IAG OAG KC1 stage IC1 OC1 KC2 stage IC2 OC2 Operand execute stage n/a EX Late-write stage n/a DA Version 4 use the same 2-cycle read pipeline developed for Version 3. Each has 32-bit address and 32-bit read data paths. Version 4 uses synchronous memory elements for all memory control units. To support this, certain control information and all address bits are sent on the at the end of the cycle before the initial bus access cycle (The data has an additional 32-bit write data path). For processor store operations, Version 4 ColdFire uses a late-write strategy, which can require 2 additional data cycles. This strategy yields the pipeline behavior described in Table 5-12. Table 5-12. Pipeline Cycles Cycle J Description Control and partial address broadcast (to start synchronous memories) KC1 Complete address and control broadcast plus MMU information. It is during this cycle that all memory element read operations are performed; that is, memory arrays are accessed. KC2 Select appropriate memory as source, return data to processor, handle cache misses or hold pipeline as needed. EX Optional write stage, pipeline address and control for store operations. DA Data available for stores from processor; memory element update occurs in the next cycle. The contains two independent memory unit access controllers and two independent controllers. Each instruction and data is analyzed to see which, if any, controller is referenced. This information, along with cache mode, store precision, and fault information, is sourced during KC1. The optional MMU is referenced concurrently with the memory unit access controllers. It has two independent control sections to simultaneously process an instruction and data request. Figure 5-1 shows how the MMU and memory unit access controllers fit in the pipeline. As the diagram shows, core address and attributes are used to access the mapping registers and the MMU. By the middle of the KC1 cycle, the memory address is available along with its corresponding access control. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 5-19 Figure 5-10 shows more details of the MMU structure. The TLB is accessed at the beginning of the KC1 pipeline stage so the resulting physical address can be sourced to the cache controllers to factor into the cache hit/miss determination. This is required because caches are virtually indexed but physically mapped. JADDR, J Control To memory controllers J TLB data entries TLB tag entries Memory unit access control (MMUBAR, RAMBARs, ROMBARs, ACRs, CACR priority hit logic) Comp To control for TLB miss logic TLB hit entry data KC1 Translated address MMU’s access control TLB Hit Untranslated address mapping register’s access control To control for TLB miss logic Mapping register hit or special mode access To memory controllers plus bus interface KADDR_KC1 KC1 cycle access control Figure 5-10. Address and Attributes Generation 5.6 MMU Implementation The MMU implements a 64-entry full-associative Harvard TLB architecture with 32-entry ITLB and DTLB. This section provides more details of this specific TLB implementation. This section details the operation and looks at the size, frequency, miss rate, and miss recovery time of this specific TLB implementation. 5.6.1 TLB Address Fields Because the TLB has a total of 64 entries (32 each for the ITLB and DTLB), a 6-bit address field is necessary. TLB addresses 0–31 reference the ITLB, and TLB addresses 32–63 reference the DTLB. In the MMUOR, bits 0 through 5 of the TLB allocation address (AA[5–0]) have this address format for CF4e. The remaining TLB allocation address bits (AA[15–6]) are ignored on updates and always read as zero. MCF548x Reference Manual, Rev. 5 5-20 Freescale Semiconductor MMU Implementation When MMUAR is used for a TLB address, bits FA[5–0] also have this address format for CF4e. The remaining form address bits (FA[31–6]) are ignored when this register is being used for a TLB address. 5.6.2 TLB Replacement Algorithm The instruction and data TLBs provide low-latency access to recently used instruction and operand translation information. CF4e ITLBs and DTLBs are 32-entry fully associative caches. The 32 ITLB entries are searched on each instruction reference; the 32 DTLB entries are searched on each operand reference. CF4e TLBs are software controlled. The TLB clear-all function clears valid bits on every TLB entry and resets the replacement logic. A new valid entry is loaded in the TLBs may be designated as locked and unavailable for allocation. TLB hits to locked entries do not update replacement algorithm information. When a new TLB entry needs to be allocated, the user can specify the exact TLB entry to be updated (through MMUOR[ADR] and MMUAR) or let TLB hardware pick the entry to update based on the replacement algorithm. A pseudo-least-recently used (PLRU) algorithm picks the entry to be replaced on a TLB miss. The algorithm works as follows: • If any element is empty (non-valid), use the lowest empty element as the allocate entry (that is, entry 0 before 1, 2, 3, and so on). • If all entries are valid, use the entry indicated by the PLRU as the allocate entry. The PLRU algorithm uses 31 most-recently used state bits per TLB to track the TLB hit history. Table 5-13 lists these state bits. Table 5-13. PLRU State Bits State Bits Meaning rdRecent31To16 A one indicates 31To16 is more recent than 15To00 rdRecent31To24 A one indicates 31To24 is more recent than 23To16 rdRecent15To08 A one indicates 15To08 is more recent than 07To00 rdRecent31To28 A one indicates 31To28 is more recent than 27To24 rdRecent23To20 A one indicates 23To20 is more recent than 19To16 rdRecent15To12 A one indicates 15To12 is more recent than 11To08 rdRecent07To04 A one indicates 07To04 is more recent than 03To00 rdRecent31To30 A one indicates 31To30 is more recent than 29To28 rdRecent27To26 A one indicates 27To26 is more recent than 25To24 rdRecent23To22 A one indicates 23To22 is more recent than 21To20 rdRecent19To18 A one indicates 19To18 is more recent than 17To16 rdRecent15To14 A one indicates 15To14 is more recent than 13To12 rdRecent11To10 A one indicates 11To10 is more recent than 09To08 rdRecent07To06 A one indicates 07To06 is more recent than 05To04 rdRecent03To02 A one indicates 03To02 is more recent than 01To00 rdRecent31 A one indicates 31 is more recent than 30 rdRecent29 A one indicates 29 is more recent than 28 MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 5-21 Table 5-13. PLRU State Bits (Continued) State Bits Meaning rdRecent27 A one indicates 27 is more recent than 26 rdRecent25 A one indicates 25 is more recent than 24 rdRecent23 A one indicates 23 is more recent than 22 rdRecent21 A one indicates 21 is more recent than 20 rdRecent19 A one indicates 19 is more recent than 18 rdRecent17 A one indicates 17 is more recent than 16 rdRecent15 A one indicates 15 is more recent than 14 rdRecent13 A one indicates 13 is more recent than 12 rdRecent11 A one indicates 11 is more recent than 10 rdRecent09 A one indicates 09 is more recent than 08 rdRecent07 A one indicates 07 is more recent than 06 rdRecent05 A one indicates 05 is more recent than 04 rdRecent03 A one indicates 03 is more recent than 02 rdRecent01 A one indicates 01 is more recent than 00 Binary state bits are updated on all TLB write (load) operations, as well as normal ITLB and DTLB hits of non-locked entries. Also, if all entries in a binary state are locked, than that state is always set. That is, if entries 15, 14, 13, and 12 were locked, LRU state bit rdRecent15To14 is forced to one. For a completely valid TLB, binary state information determines the LRU entry. The CF4e replacement algorithm is deterministic and, for the case of a full TLB (with no locked entries and always touching new pages), the replacement entry repeats every 32 TLB loads. 5.6.3 TLB Locked Entries Figure 5-11 is a ColdFire MMU Harvard TLB block diagram. For TLB miss faults, the instruction restart model completely reexecutes an instruction on returning from the exception handler. An instruction can touch two instruction pages (a 32- or 48-bit instruction can straddle two pages) or four data pages (a memory-to-memory word or longword move where misaligned source and destination operands straddle two pages). Therefore, one instruction may take two ITLB misses and allocate two ITLB pages before completion. Likewise, one instruction may require four DTLB misses and allocate four DTLB pages. Because of this, a pool of unlocked TLB entries must be available if virtual memory is used. The above examples show the fewest entries needed to guarantee an instruction can complete execution. For good MMU performance, more unlocked TLB entries should be available. MCF548x Reference Manual, Rev. 5 5-22 Freescale Semiconductor MMU Instructions Current address space ID (ASID) J Instruction or data address and attributes TLB Tag Entry 31 TLB Tag Entry 0 TLB Tag Entry 31 TLB Tag Entry 0 KC1 Compare Compare Instruction or data hit select To control for instruction or DTLB miss logic IC1 or OC1 translated address IC1 or OC1 access control Figure 5-11. Version 4 ColdFire MMU Harvard TLB 5.7 MMU Instructions The MOVE to USP and MOVE from USP instructions have been added for accessing the USP. Refer to the ColdFire Programmer’s Reference Manual for more information. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 5-23 MCF548x Reference Manual, Rev. 5 5-24 Freescale Semiconductor Chapter 6 Floating-Point Unit (FPU) 6.1 Introduction This chapter describes instructions implemented in the floating-point unit (FPU) designed for use with the ColdFire family of microprocessors. The FPU conforms to the American National Standards Institute (ANSI)/Institute of Electrical and Electronics Engineers (IEEE) Standard for Binary Floating-Point Arithmetic (ANSI/IEEE Standard 754). The hardware unit is optimized for real-time execution with exceptions disabled and default results provided for specific operations, operands, and number types. The FPU does not support all IEEE-754 number types and operations in hardware. Exceptions can be enabled to support these cases in software. 6.1.1 Overview The FPU operates on 64-bit, double-precision, floating-point data and supports single-precision and signed integer input operands. The FPU programming model is like that in the MC68060 microprocessor. The FPU is intended to accelerate the performance of certain classes of embedded applications, especially those requiring high-speed floating-point arithmetic computations. See Section 6.7.3, “Key Differences between ColdFire and M68000 FPU Programming Models.” The FPU appears as another execute engine at the bottom stages of the operand execution pipeline (OEP), using operands from a dual-ported register file. Setting bit 4 in the cache control register (CACR[DF]) disables the FPU. If CACR[DF] is cleared, all FPU instructions are issued and executed, otherwise the processor responds with an unimplemented line-F instruction exception (vector 11). Operating systems often assume user applications are integer-only (to minimize the time required by save context) by setting CACR[DF] at process initiation. If the application includes floating-point instructions, the attempted execution of the first FP instruction generates the unimplemented line-F exception, which signals the kernel that the FPU registers must be included in the context for the application. The application then continues execution with CACR[DF] cleared to enable FPU execution. 6.1.1.1 Notational Conventions Table 6-1 defines notational conventions used in this chapter. Table 6-1. Notational Conventions Symbol Description Single- and Double-Precision Operand Operations + Arithmetic addition or postincrement indicator − Arithmetic subtraction or predecrement indicator × Arithmetic multiplication ÷ Arithmetic division or conjunction symbol ∼ Invert, operand is logically complemented. An overbar, , is also used for this operation. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 6-1 Table 6-1. Notational Conventions (Continued) Symbol Description & Logical AND | Logical OR → <op> <operand>tested sign-extended Source operand is moved to destination operand Any double-operand operation Operand is compared to zero and the condition codes are set appropriately All bits of the upper portion are made equal to the high-order bit of the lower portion Other Operations If <condition> then <operations> else <operations> Test the condition. If true, the operations after then are performed. If the condition is false and the optional else clause is present, the operations after else are performed. If the condition is false and else is omitted, the instruction performs no operation. Refer to the Bcc instruction description as an example. Register Specifications An Address register n (example: A3 is address register 3) Ay, Ax Source and destination address registers, respectively Dn Data register n (example: D3 is data register 3) Dy,Dx Source and destination data registers, respectively FPCR Floating-point control register FPIAR Floating-point instruction address register FPn FPSR FPy,FPx Floating-point data register n (example: FP3 is FPU data register 3) Floating-point status register Source and destination floating-point data registers, respectively PC Program counter Rn Address or data register Rx Destination register Ry Source register Xi Index register Table 6-2 describes addressing modes and syntax for floating-point instructions. MCF548x Reference Manual, Rev. 5 6-2 Freescale Semiconductor Operand Data Formats and Types Table 6-2. Floating-Point Addressing Modes Addressing Modes Syntax Register direct Address register direct Address register direct 6.2 Dy Ay Register indirect Address register indirect Address register indirect with postincrement Address register indirect with predecrement Address register indirect with displacement (Ay) –(Ay) (d16,Ay) Program counter indirect with displacement (d16,PC) Operand Data Formats and Types The FPU supports signed byte, word, and longword integer formats, which are identical to those supported by the integer unit. The FPU also supports single- and double-precision binary floating-point formats that fully comply with the IEEE-754 standard. 6.2.1 Signed-Integer Data Formats The FPU supports 8-bit byte (B), 16-bit word (W), and 32-bit longword (L) integer data formats. 6.2.2 Floating-Point Data Formats Figure 6-1 shows the two binary floating-point data formats. 31 30 S 63 S 62 8-Bit Exponent Sign of Mantissa 51 11-Bit Exponent 0 22 52-Bit Fraction 23-Bit Fraction Single 0 Double Sign of Mantissa Figure 6-1. Floating-Point Data Formats Note that, throughout this chapter, a mantissa is defined as the concatenation of an integer bit, the binary point, and a fraction. A fraction is the term designating the bits to the right of the binary point in the mantissa. Mantissa (integer bit).(fraction) Figure 6-2. Mantissa The integer bit is implied to be set for normalized numbers and infinities, clear for zeros and denormalized numbers. For not-a-numbers (NANs), the integer bit is ignored. The exponent in both floating-point formats is an unsigned binary integer with an implied bias added to it. Subtracting the bias from exponent MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 6-3 yields a signed, two’s complement power of two. This represents the magnitude of a normalized floating-point number when multiplied by the mantissa. By definition, a normalized mantissa always takes values starting from 1.0 and going up to, but not including, 2.0; that is, [1.0...2.0). 6.2.3 Floating-Point Data Types Each floating-point data format supports five unique data types: normalized numbers, zeros, infinities, NANs, and denormalized numbers. The normalized data type, Figure 6-3, never uses the maximum or minimum exponent value for a given format. 6.2.3.1 Normalized Numbers Normalized numbers include all positive or negative numbers with exponents between the maximum and minimum values. For single- and double-precision normalized numbers, the implied integer bit is one and the exponent can be zero. Min < Exponent < Max Fraction = Any bit pattern Sign of Mantissa, 0 or 1 Figure 6-3. Normalized Number Format 6.2.3.2 Zeros Zeros can be positive or negative and represent real values, + 0.0 and – 0.0. See Figure 6-4. Exponent = 0 Fraction = 0 Sign of Mantissa, 0 or 1 Figure 6-4. Zero Format 6.2.3.3 Infinities Infinities can be positive or negative and represent real values that exceed the overflow threshold. A result’s exponent greater than or equal to the maximum exponent value indicates an overflow for a given data format and operation. This overflow description ignores the effects of rounding and the user-selectable rounding models. For single- and double-precision infinities, the fraction is a zero. See Figure 6-5. Exponent = Maximum Fraction = 0 Sign of Mantissa, 0 or 1 Figure 6-5. Infinity Format MCF548x Reference Manual, Rev. 5 6-4 Freescale Semiconductor Operand Data Formats and Types 6.2.3.4 Not-A-Number When created by the FPU, NANs represent the results of operations having no mathematical interpretation, such as infinity divided by infinity. Operations using a NAN operand as an input return a NAN result. User-created NANs can protect against uninitialized variables and arrays or can represent user-defined data types. See Figure 6-6. Exponent = Maximum Fraction = Any nonzero bit pattern Sign of Mantissa, 0 or 1 Figure 6-6. Not-a-Number Format If an input operand to an operation is a NAN, the result is an FPU-created default NAN. When the FPU creates a NAN, the NAN always contains the same bit pattern in the fraction: all fraction bits are ones and the sign bit is zero. When the user creates a NAN, any nonzero bit pattern can be stored in the fraction and the sign bit. 6.2.3.5 Denormalized Numbers Denormalized numbers represent real values near the underflow threshold. Denormalized numbers can be positive or negative. For denormalized numbers in single- and double-precision, the implied integer bit is a zero. See Figure 6-7. Exponent = 0 Fraction = Any nonzero bit pattern Sign of Mantissa, 0 or 1 Figure 6-7. Denormalized Number Format Traditionally, the detection of underflow causes floating-point number systems to perform a flush-to-zero. The IEEE-754 standard implements gradual underflow: the result mantissa is shifted right (denormalized) while the result exponent is incremented until reaching the minimum value. If all the mantissa bits of the result are shifted off to the right during this denormalization, the result becomes zero. Denormalized numbers are not supported directly in the hardware of this implementation but can be handled in software if needed (software for the input denorm exception could be written to handle denormalized input operands, and software for the underflow exception could create denormalized numbers). If the input denorm exception is disabled, all denormalized numbers are treated as zeros. Table 6-3 summarizes the data type specifications for byte, word, longword, single- and double-precision data formats. Table 6-3. Real Format Summary Parameter Data Format Single-Precision 3130 s 23 22 e Double-Precision 0 f 6362 s 52 51 e 0 f Field Size in Bits Sign (s) 1 1 MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 6-5 Table 6-3. Real Format Summary (Continued) Parameter Single-Precision Double-Precision Biased exponent (e) 8 11 Fraction (f) 23 52 Total 32 64 Interpretation of Sign Positive fraction s=0 s=0 Negative fraction s=1 s=1 Normalized Numbers Bias of biased exponent Range of biased exponent Range of fraction +127 (0x7F) +1023 (0x3FF) 0 < e < 255 (0xFF) 0 < e < 2047 (0x7FF) Zero or Nonzero Zero or Nonzero 1.f 1.f (–1)s × 2e–127 × 1.f (–1)s × 2e–1023 × 1.f Mantissa Relation to representation of real numbers Denormalized Numbers Biased exponent format minimum Bias of biased exponent Range of fraction 0 (0x00) 0 (0x000) +126 (0x7E) +1022 (0x3FE) Nonzero Nonzero 0.f 0.f (–1)s × 2–126 × 0.f (–1)s × 2–1022 × 0.f Mantissa Relation to representation of real numbers Signed Zeros Biased exponent format minimum 0 (0x00) 0 (0x00) Mantissa 0.f = 0.0 0.f = 0.0 Signed Infinities Biased exponent format maximum Mantissa 255 (0xFF) 2047 (0x7FF) 0.f = 0.0 0.f = 0.0 NANs Sign Don’t Care 0 or 1 Biased exponent format maximum 255 (0xFF) 2047 (0x7FF) Nonzero Nonzero xxxxx…xxxx 11111…1111 xxxxx…xxxx 11111…1111 Fraction Representation of Fraction Nonzero Bit Pattern Created by User Fraction When Created by FPU MCF548x Reference Manual, Rev. 5 6-6 Freescale Semiconductor Register Definition Table 6-3. Real Format Summary (Continued) Parameter Single-Precision Double-Precision Approximate Ranges Maximum Positive Normalized 3.4 × 1038 1.8 x 10308 Minimum Positive Normalized 1.2 × 10–38 2.2 x 10–308 Minimum Positive Denormalized 1.4 × 10–45 4.9 x 10–324 6.3 Register Definition The programmer’s model for the FPU consists of the following: • Eight 64-bit floating-point data registers (FP0–FP7) • One 32-bit floating-point control register (FPCR) • One 32-bit floating-point status register (FPSR) • One 32-bit floating-point instruction address register (FPIAR) Figure 6-8 shows the FPU programming model. 63 0 FP0 FP1 FP2 FP3 FP4 FP5 FP6 FP7 FPCR FPSR FPIAR Floating-point data registers Floating-point control register Floating-point status register Floating-point instruction address register Figure 6-8. Floating-Point Programmer’s Model 6.3.1 Floating-Point Data Registers (FP0–FP7) Floating-point data registers are analogous to the integer data registers for the 68K/ColdFire family. They always contain numbers in double-precision format, even though the operand may be a single-precision value used in a single-precision calculation. All external operands, regardless of the source data format, are converted to double-precision format before being used in any calculation or being stored in a floating-point data register. A reset or a null-restore operation sets FP0–FP7 to positive, nonsignaling NANs. 6.3.2 Floating-Point Control Register (FPCR) The FPCR, Figure 6-9, contains an exception enable byte (EE) and a mode control byte (MC). Each EE bit corresponds to a floating-point exception class. The user can separately enable traps for each class of floating-point exceptions. The MC bits control FPU operating modes. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 6-7 The user can read or write to FPCR using FMOVE or FRESTORE. A processor reset or a restore operation of the null state clears the FPCR. When this register is cleared, the FPU never generates exceptions. 31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 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 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 R W Reset Exception Enable Byte (EE) R BSUN INAN OPERR OVFL UNFL Mode Control Byte (MC) DZ INEX IDE 0 PREC 0 0 0 0 0 RND 0 0 0 0 0 0 0 0 W Reset 0 0 0 0 0 Reg Addr 0 0 CPU + 0x824 Figure 6-9. Floating-Point Control Register (FPCR) Table 6-4 describes FPCR fields. Table 6-4. FPCR Field Descriptions Bits Field Description 31–16 — 15 BSUN Branch set on unordered 14 INAN Input not-a-number 13 OPERR 12 OVFL Overflow 11 UNFL Underflow 10 DZ 9 INEX 8 IDE 7 — 6 PREC 5–4 RND 3–0 — Reserved, should be cleared. Operand error Divide by zero Inexact operation Input denormalized Reserved, should be cleared. Rounding precision 0 Double (D) 1 Single (S) Rounding mode 00 To nearest (RN) 01 To zero (RZ) 10 To minus infinity (RM) 11 To plus infinity (RP) Reserved, should be cleared. MCF548x Reference Manual, Rev. 5 6-8 Freescale Semiconductor Register Definition 6.3.3 Floating-Point Status Register (FPSR) The FPSR, Figure 6-10, contains a floating-point condition code byte (FPCC), a floating-point exception status byte (EXC), and a floating-point accrued exception byte (AEXC). The user can read or write all FPSR bits. Execution of most floating-point instructions modifies FPSR. FPSR is loaded using FMOVE or FRESTORE. A processor reset or a restore operation of the null state clears the FPSR. The floating-point condition code byte contains 4 condition code bits that are set after completion of all arithmetic instructions involving the floating-point data registers. The floating-point store operation, FMOVEM, and move system control register instructions do not affect the FPCC. The exception status byte contains a bit for each floating-point exception that might have occurred during the most recent arithmetic instruction or move operation. This byte is cleared at the start of all operations that generate floating-point exceptions (except FBcc only affects BSUN and that only for nonaware tests). Operations that do not generate floating-point exceptions do not clear this byte. An exception handler can use this byte to determine which floating-point exception or exceptions caused a trap. The equations below the table show the comparative relationship between the EXC byte and AEXC byte. The accrued exception byte contains 5 required bits for IEEE-754 exception-disabled operations. These exceptions are logical combinations of EXC bits. AEXC records all floating-point exceptions since AEXC was last cleared, either by writing to FPSR or as a result of reset or a restore operation of the null state. Many users disable traps for some or all floating-point exception classes. AEXC eliminates the need to poll EXC after each floating-point instruction. At the end of arithmetic operations, EXC bits are logically combined to form an AEXC value that is logically ORed into the existing AEXC byte (FBcc only updates IOP). This operation creates sticky floating-point exception bits in AEXC that the user can poll only at the end of a series of floating-point operations. A sticky bit is one that remains set until the user clears it. Setting or clearing AEXC bits neither causes nor prevents an exception. The equations below the table show relationships between EXC and AEXC. Comparing the current value of an AEXC bit with a combination of EXC bits derives a new value in the corresponding AEXC bit. These boolean equations apply to setting AEXC bits at the end of each operation affecting AEXC. 31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 Floating-Point Condition Code Byte (FPCC) R 0 0 0 0 N Z I NAN 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 W Reset Exception Status Byte (EXC) Floating-Point Accrued Exception Byte (AEXC) R BSUN INAN OPERR OVFL UNFL DZ INEX IDE IOP OVFL UNFL DZ INEX 0 0 0 0 0 0 0 0 W Reset 0 0 0 0 Reg Addr 0 0 0 0 0 0 0 CPU + 0x822 Figure 6-10. Floating-Point Status Register (FPSR) Table 6-5 describes FPSR fields. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 6-9 Table 6-5. FPSR Field Descriptions Bits Field Description 31–28 — Reserved, should be cleared. 27 N Negative 26 Z Zero 25 I Infinity 24 NAN 23–16 — 15 BSUN Branch/set on unordered 14 INAN Input not-a-number 13 OPERR 12 OVFL Overflow 11 UNFL Underflow 10 DZ Divide by zero 9 INEX Inexact result 8 IDE Input is denormalized 7 IOP Invalid operation 6 OVFL Overflow 5 UNFL Underflow 4 DZ Divide by zero 3 INEX Inexact result 2–0 — Not-a-number Reserved, should be cleared. Operand error Reserved, should be cleared. For AEXC[OVFL], AEXC[DZ], and AEXC[INEX], the next value is determined by ORing the current AEXC value with the EXC equivalent, as shown in the following: • Next AEXC[OVFL] = Current AEXC[OVFL] | EXC[OVFL] • Next AEXC[DZ] = Current AEXC[DZ] | EXC[DZ] • Next AEXC[INEX] = Current AEXC[INEX] | EXC[INEX] For AEXC[IOP] and AEXC[UNFL], the next value is calculated by ORing the current AEXC value with EXC bit combinations, as follows: • Next AEXC[IOP] = Current AEXC[IOP] | EXC[BSUN | INAN | OPERR] • Next AEXC[UNFL] = Current AEXC[UNFL] | EXC[UNFL & INEX] 6.3.4 Floating-Point Instruction Address Register (FPIAR) The ColdFire OEP can execute integer and floating-point instructions simultaneously. As a result, the PC value stacked by the processor in response to a floating-point exception trap may not point to the instruction that caused the exception. MCF548x Reference Manual, Rev. 5 6-10 Freescale Semiconductor Floating-Point Computational Accuracy For FPU instructions that can generate exception traps, the 32-bit FPIAR is loaded with the instruction PC address before the FPU begins execution. In case of an FPU exception, the trap handler can use the FPIAR contents to determine the instruction that generated the exception. FMOVE to/from FPCR, FPSR, or FPIAR and FMOVEM instructions cannot generate floating-point exceptions; therefore, they do not modify FPIAR. A reset or a null-restore operation clears FPIAR. 6.4 Floating-Point Computational Accuracy The FPU performs all floating-point internal operations in double-precision. It supports mixed-mode arithmetic by converting single-precision operands to double-precision values before performing the specified operation. The FPU converts all memory data formats to the double-precision data format and stores the value in a floating-point register or uses it as the source operand for an arithmetic operation. When moving a double-precision floating-point value from a floating-point data register, the FPU can convert the data depending on the destination, as follows: • Valid data formats for memory destination: B, W, L, S, or D • Valid data formats for integer data register destinations: B, W, L, or S Normally if the input operand is a denormalized number, the number must be normalized before an FPU instruction can be executed. A denormalized input operand is converted to zero if the input denorm exception (IDE) is disabled. If IDE is enabled, the floating-point engine traps to allow software action to be taken by the handler. 6.4.1 Intermediate Result All FPU calculations use an intermediate result. When the FPU performs any operation, the calculation is carried out using double-precision inputs, and the intermediate result is calculated as if to produce infinite precision. After the calculation is complete, any necessary rounding of the intermediate result for the selected precision is performed and the result is stored in the destination. Figure 6-11 shows the intermediate result format. The intermediate result’s exponent for some dyadic operations (for example, multiply and divide) can easily overflow or underflow the 11-bit exponent of the designated floating-point register. To simplify overflow and underflow detection, intermediate results in the FPU maintain a 12-bit two’s complement, integer exponent. Detection of an intermediate result overflow or underflow always converts the 12-bit exponent into a 11-bit biased exponent before being stored in a floating-point data register. The FPU internally maintains a 56-bit mantissa for rounding purposes. The mantissa is always rounded to 53 bits (or fewer, depending on the selected rounding precision) before it is stored in a floating-point data register. 56-Bit Intermediate Mantissa 12-Bit Exponent 52-Bit Fraction Integer lsb Guard Round Sticky Figure 6-11. Intermediate Result Format If the destination is a floating-point data register, the result is in double-precision format but may be rounded to single-precision, if required by the rounding precision, before being stored. If the single-precision mode is selected, the exponent value is in the correct range even if it is stored in MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 6-11 double-precision format. If the destination is a memory location or an integer data register, rounding precision is ignored. In this case, a number in the double-precision format is taken from the source floating-point data register, rounded to the destination format precision, and then written to memory or the integer data register. Depending on the selected rounding mode or destination data format, the location of the lsb of the mantissa and the locations of the guard, round, and sticky bits in the 56-bit intermediate result mantissa vary. Guard and round bits are calculated exactly. The sticky bit creates the illusion of an infinitely wide intermediate result. As the arrow in Figure 6-11 shows, the sticky bit is the logical OR of all bits to the right of the round bit in the infinitely precise result. During calculation, nonzero bits generated to the right of the round bit set the sticky bit. Because of the sticky bit, the rounded intermediate result for all required IEEE arithmetic operations in RN mode can err by no more than one half unit in the last place. 6.4.2 Rounding the Result The FPU supports the four rounding modes specified by the IEEE-754 standard: round-to-nearest (RN), round-toward-zero (RZ), round-toward-plus-infinity (RP), and round-toward-minus-infinity (RM). The RM and RP modes are often referred to as directed-rounding-modes and are useful in interval arithmetic. Rounding is accomplished through the intermediate result. Single-precision results are rounded to a 24-bit mantissa boundary; double-precision results are rounded to a 53-bit mantissa boundary. The current floating-point instruction can specify rounding precision, overriding the rounding precision specified in FPCR for the duration of the current instruction. For example, the rounding precision for FADD is determined by FPCR, while the rounding precision for FSADD is single-precision, independent of FPCR. Range control helps emulate devices that support only single-precision arithmetic by rounding the intermediate result’s mantissa to the specified precision and checking that the intermediate exponent is in the representable range of the selected rounding precision. If the intermediate result’s exponent exceeds the range, the appropriate underflow or overflow value is stored as the result in the double-precision format exponent. For example, if the data format and rounding mode is single-precision RM and the result of an arithmetic operation overflows the single-precision format, the maximum normalized single-precision value is stored as a double-precision number in the destination floating-point data register; that is, the unbiased 11-bit exponent is 0x0FF and the 52-bit fraction is 0xF_FFFF_E000_0000. If an infinity is the appropriate result for an underflow or overflow, the infinity value for the destination data format is stored as the result; that is, the exponent has the maximum value and the mantissa is zero. Figure 6-12 shows the algorithm for rounding an intermediate result to the selected rounding precision and destination data format. If the destination is a floating-point register, the rounding boundary is determined by either the selected rounding precision specified by FPCR[PREC] or by the instruction itself. For example, FSADD and FDADD specify single- and double-precision rounding regardless of FPCR[PREC]. If the destination is memory or an integer data register, the destination data format determines the rounding boundary. If the rounded result of an operation is inexact, INEX is set in FPSR[EXC]. MCF548x Reference Manual, Rev. 5 6-12 Freescale Semiconductor Floating-Point Computational Accuracy Entry Guard, Round and Sticky Bits = 0 INEX 1 Select Rounding Mode Check Intermediate Result RN Pos G and lsb = 1, R and S = 0 or G = 1, R or S = 1 RM Neg RP Pos G, R, or S = 1 N Y RZ Neg G, R, or S = 1 Y N Exact Result G,R, and S are chopped Add 1 to lsb Add 1 to lsb Overflow = 1 Shift mantissa right 1 bit, Add 1 to exponent Guard Round Sticky 0 0 0 Exit Exit Figure 6-12. Rounding Algorithm Flowchart The 3 additional bits beyond the double-precision format, the difference between the intermediate result’s 56-bit mantissa and the storing result’s 53-bit mantissa, allow the FPU to perform all calculations as though it were performing calculations using a compute engine with infinite bit precision. The result is always correct for the specified destination’s data format before rounding (unless an overflow or underflow error occurs). The specified rounding produces a number as close as possible to the infinitely precise intermediate value and still representable in the selected precision. The tie case in Table 6-6 shows how the 56-bit mantissa allows the FPU to meet the error bound of the IEEE specification. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 6-13 Table 6-6. Tie-Case Example Result Integer 52-Bit Fraction Guard Round Sticky Intermediate x xxx…x00 1 0 0 Rounded-to-Nearest x xxx…x00 0 0 0 The lsb of the rounded result does not increment even though the guard bit is set in the intermediate result. The IEEE-754 standard specifies this way of handling ties. If the destination data format is double-precision and there is a difference between the infinitely precise intermediate result and the round-to-nearest result, the relative difference is 2–53 (the value of the guard bit). This error is equal to half of the lsb’s value and is the worst case error that can be introduced with RN mode. Thus, the term one-half unit in the last place correctly identifies the error bound for this operation. This error specification is the relative error present in the result; the absolute error bound is equal to 2exponent x 2–53. Table 6-7 shows the error bound for other rounding modes. Table 6-7. Round Mode Error Bounds Result Integer 52-Bit Fraction Guard Round Sticky Intermediate x xxx…x00 1 1 1 Rounded-to-Zero x xxx…x00 0 0 0 The difference between the infinitely precise result and the rounded result is 2–53 + 2–54 + 2–55, which is slightly less than 2–52 (the value of the lsb). Thus, the error bound for this operation is not more than one unit in the last place. The FPU meets these error bounds for all arithmetic operations, providing accurate, repeatable results. 6.5 Floating-Point Post-Processing Most operations end with post-processing, for which the FPU provides two steps. First, FPSR[FPCC] bits are set or cleared at the end of each arithmetic or move operation to a single floating-point data register. FPCC bits are consistently set based on the result of the operation. Second, the FPU supports 32 conditional tests that allow floating-point conditional instructions to test floating-point conditions in the same way that integer conditional instructions test the integer condition code. The combination of consistently set FPCC bits and the simple programming of conditional instructions gives the processor a highly flexible, efficient way to change program flow based on floating-point results. When the summary for each instruction is read, it should be assumed that an instruction performs post processing, unless the summary specifically states otherwise. The following paragraphs describe post processing in detail. 6.5.1 Underflow, Round, and Overflow During calculation of an arithmetic result, the FPU has more precision and range than the 64-bit double-precision format. However, the final result is a double-precision value. In some cases, an intermediate result becomes either smaller or larger than can be represented in double-precision. Also, the operation can generate a larger exponent or more bits of precision than can be represented in the chosen rounding precision. For these reasons, every arithmetic instruction ends by checking for underflow, rounding the result and checking for overflow. At the completion of an arithmetic operation, the intermediate result is checked to see if it is too small to be represented as a normalized number in the selected precision. If so, the underflow (UNFL) bit is set in FPSR[EXC]. If no underflow occurs, the intermediate result is rounded according to the user-selected MCF548x Reference Manual, Rev. 5 6-14 Freescale Semiconductor Floating-Point Post-Processing rounding precision and mode. After rounding, the inexact bit (INEX) is set as described in Figure 6-12. Lastly, the magnitude of the result is checked to see if it exceeds the current rounding precision. If so, the overflow (OVFL) bit is set, and a correctly signed infinity or correctly signed largest normalized number is returned, depending on the rounding mode. NOTE INEX can also be set by OVFL, UNFL, and when denormalized numbers are encountered. 6.5.2 Conditional Testing Unlike operation-dependent integer condition codes, an instruction either always sets FPCC bits in the same way or does not change them at all. Therefore, instruction descriptions do not include FPCC settings. This section describes how FPCC bits are set. FPCC bits differ slightly from integer condition codes. An FPU operation’s final result sets or clears FPCC bits accordingly, independent of the operation itself. Integer condition code bits N and Z have this characteristic, but V and C are set differently for different instructions. Table 6-8 lists FPCC settings for each data type. Loading FPCC with another combination and executing a conditional instruction can produce an unexpected branch condition. Table 6-8. FPCC Encodings Data Type N Z I NAN + Normalized or Denormalized 0 0 0 0 – Normalized or Denormalized 1 0 0 0 +0 0 1 0 0 –0 1 1 0 0 + Infinity 0 0 1 0 – Infinity 1 0 1 0 + NAN 0 0 0 1 – NAN 1 0 0 1 The inclusion of the NAN data type in the IEEE floating-point number system requires each conditional test to include FPCC[NAN] in its boolean equation. Because it cannot be determined whether a NAN is bigger or smaller than an in-range number (since it is unordered), the compare instruction sets FPCC[NAN] when an unordered compare is attempted. All arithmetic instructions that result in a NAN also set the NAN bit. Conditional instructions interpret NAN being set as the unordered condition. The IEEE-754 standard defines the following four conditions: • Equal to (EQ) • Greater than (GT) • Less than (LT) • Unordered (UN) The standard requires only the generation of the condition codes as a result of a floating-point compare operation. The FPU can test for these conditions and 28 others at the end of any operation affecting condition codes. For floating-point conditional branch instructions, the processor logically combines the 4 bits of the FPCC condition codes to form 32 conditional tests, 16 of which cause an exception if an MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 6-15 unordered condition is present when the conditional test is attempted (IEEE nonaware tests). The other 16 do not cause an exception (IEEE-aware tests). The set of IEEE nonaware tests is best used in one of the following cases: • When porting a program from a system that does not support the IEEE standard to a conforming system • When generating high-level language code that does not support IEEE floating-point concepts (that is, the unordered condition). An unordered condition occurs when one or both of the operands in a floating-point compare operation is a NAN. The inclusion of the unordered condition in floating-point branches destroys the familiar trichotomy relationship (greater than, equal, less than) that exists for integers. For example, the opposite of floating-point branch greater than (FBGT) is not floating-point branch less than or equal (FBLE). Rather, the opposite condition is floating-point branch not greater than (FBNGT). If the result of the previous instruction was unordered, FBNGT is true, whereas both FBGT and FBLE would be false because unordered fails both of these tests (and sets BSUN). Compiler code generators should be particularly careful of the lack of trichotomy in the floating-point branches, because it is common for compilers to invert the sense of conditions. When using the IEEE nonaware tests, the user receives a BSUN exception if a branch is attempted and FPCC[NAN] is set, unless the branch is an FBEQ or an FBNE. If the BSUN exception is enabled in FPCR, the exception takes a BSUN trap. Therefore, the IEEE nonaware program is interrupted if an unexpected condition occurs. Users knowledgeable of the IEEE-754 standard should use IEEE-aware tests in programs that contain ordered and unordered conditions. Because the ordered or unordered attribute is explicitly included in the conditional test, EXC[BSUN] is not set when the unordered condition occurs. Table 6-9 summarizes conditional mnemonics, definitions, equations, predicates, and whether EXC[BSUN] is set for the 32 floating-point conditional tests. The equation column lists FPCC bit combinations for each test in the form of an equation. Condition codes with an overbar indicate cleared bits; all other bits are set. Table 6-9. Floating-Point Conditional Tests Mnemonic Definition Equation Predicate 1 EXC[BSUN] Set IEEE Nonaware Tests EQ Equal Z 000001 No NE Not equal Z 001110 No GT Greater than NAN | Z | N 010010 Yes Not greater than NAN | Z | N 011101 Yes Greater than or equal Z | (NAN | N) 010011 Yes Not greater than or equal NAN | (N & Z) 011100 Yes Less than N & (NAN | Z) 010100 Yes NLT Not less than NAN | (Z | N) 011011 Yes LE Less than or equal Z | (N & NAN) 010101 Yes Not less than or equal NAN | (N | Z) 011010 Yes Greater or less than NAN | Z 010110 Yes NGL Not greater or less than NAN | Z 011001 Yes GLE Greater, less or equal NAN 010111 Yes NGT GE NGE LT NLE GL MCF548x Reference Manual, Rev. 5 6-16 Freescale Semiconductor Floating-Point Exceptions Table 6-9. Floating-Point Conditional Tests (Continued) Mnemonic NGLE Definition Equation Not greater, less or equal Predicate 1 EXC[BSUN] Set 011000 Yes NAN IEEE-Aware Tests EQ Equal Z 000001 No NE Not equal Z 001110 No OGT Ordered greater than NAN | Z | N 000010 No ULE Unordered or less or equal NAN | Z | N 001101 No OGE Ordered greater than or equal Z | (NAN | N) 000011 No ULT Unordered or less than NAN | (N & Z) 001100 No OLT Ordered less than N & (NAN | Z) 000100 No UGE Unordered or greater or equal NAN | (Z | N) 001011 No OLE Ordered less than or equal Z | (N & NAN) 000101 No UGT Unordered or greater than NAN | (N | Z) 001010 No OGL Ordered greater or less than NAN | Z 000110 No UEQ Unordered or equal NAN | Z 001001 No OR Ordered NAN 000111 No UN Unordered NAN 001000 No Miscellaneous Tests 1 6.6 F False False 000000 No T True True 001111 No SF Signaling false False 010000 Yes ST Signaling true True 011111 Yes SEQ Signaling equal Z 010001 Yes SNE Signaling not equal Z 011110 Yes This column refers to the value in the instruction’s conditional predicate field that specifies this test. Floating-Point Exceptions This section describes floating-point exceptions and how they are handled. Table 6-10 lists the vector numbers related to floating-point exceptions. If the exception is taken pre-instruction, the PC contains the address of the next floating-point instruction (nextFP). If the exception is taken post-instruction, the PC contains the address of the faulting instruction (fault). MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 6-17 Table 6-10. Floating-Point Exception Vectors Vector Number Vector Offset Program Counter Assignment 48 0x0C0 Fault Floating-point branch/set on unordered condition 49 0x0C4 NextFP or Fault Floating-point inexact result 50 0x0C8 NextFP Floating-point divide-by-zero 51 0x0CC NextFP or Fault Floating-point underflow 52 0x0D0 NextFP or Fault Floating-point operand error 53 0x0D4 NextFP or Fault Floating-point overflow 54 0x0D8 NextFP or Fault Floating-point input NAN 55 0x0DC NextFP or Fault Floating-point input denormalized number In addition to these vectors, attempting to execute a FRESTORE instruction with a unsupported frame value generates a format error exception (vector 14). See the FRESTORE instruction in the ColdFire Programmer’s Reference Manual. Attempting to execute an FPU instruction with an undefined or unsupported value in the 6-bit effective address, the 3-bit source/destination specifier, or the 7-bit opmode generates a line-F emulator exception, vector 11. See Table 6-23. 6.6.1 Floating-Point Arithmetic Exceptions This section describes floating-point arithmetic exceptions; Table 6-11 lists these exceptions in order of priority: Table 6-11. Exception Priorities Priority Exception 1 Branch/set on unordered (BSUN) 2 Input Not-a-Number (INAN) 3 Input denormalized number (IDE) 4 Operand error (OPERR) 5 Overflow (OVFL) 6 Underflow (UNFL) 7 Divide-by-zero (DZ) 8 Inexact (INEX) Most floating-point exceptions are taken when the next floating-point arithmetic instruction is encountered (this is called a pre-instruction exception). Exceptions set during a floating-point store to memory or to an integer register are taken immediately (post-instruction exception). Note that FMOVE is considered an arithmetic instruction because the result is rounded. Only FMOVE with any destination other than a floating-point register (sometimes called FMOVE OUT) can generate post-instruction exceptions. Post-instruction exceptions never write the destination. After a post-instruction exception, processing continues with the next instruction. MCF548x Reference Manual, Rev. 5 6-18 Freescale Semiconductor Floating-Point Exceptions A floating-point arithmetic exception becomes pending when the result of a floating-point instruction sets an FPSR[EXC] bit and the corresponding FPCR[ENABLE] bit is set. A user write to the FPSR or FPCR that causes the setting of an exception bit in FPSR[EXC] along with its corresponding exception enabled in FPCR, leaves the FPU in an exception-pending state. The corresponding exception is taken at the start of the next arithmetic instruction as a pre-instruction exception. Executing a single instruction can generate multiple exceptions. When multiple exceptions occur with exceptions enabled for more than one exception class, the highest priority exception is reported and taken. It is up to the exception handler to check for multiple exceptions. The following multiple exceptions are possible: • Operand error (OPERR) and inexact result (INEX) • Overflow (OVFL) and inexact result (INEX) • Underflow (UNFL) and inexact result (INEX) • Divide-by-zero (DZ) and inexact result (INEX) • Input denormalized number (IDE) and inexact result (INEX) • Input not-a-number (INAN) and input denormalized number (IDE) In general, all exceptions behave similarly. If the exception is disabled when the exception condition exists, no exception is taken, a default result is written to the destination (except for BSUN exception, which has no destination), and execution proceeds normally. If an enabled exception occurs, the same default result above is written for pre-instruction exceptions but no result is written for post-instruction exceptions. An exception handler is expected to execute FSAVE as its first floating-point instruction. This also clears FPCR, which keeps exceptions from occurring during the handler. Because the destination is overwritten for floating-point register destinations, the original floating-point destination register value is available for the handler on the FSAVE state frame. The address of the instruction that caused the exception is available in the FPIAR. When the handler is done, it should clear the appropriate FPSR exception bit on the FSAVE state frame, then execute FRESTORE. If the exception status bit is not cleared on the state frame, the same exception occurs again. Alternatively, instead of executing FSAVE, an exception handler could simply clear appropriate FPSR exception bits, optionally alter FPCR, and then return from the exception. Note that exceptions are never taken on FMOVE to or from the status and control registers and FMOVEM to or from the floating-point data registers. At the completion of the exception handler, the RTE instruction must be executed to return to normal instruction flow. 6.6.1.1 Branch/Set on Unordered (BSUN) A BSUN results from performing an IEEE nonaware conditional test associated with the FBcc instruction when an unordered condition is present. Any pending floating-point exception is first handled by a pre-instruction exception, after which the conditional instruction restarts. The conditional predicate is evaluated and checked for a BSUN exception before executing the conditional instruction. A BSUN exception occurs if the conditional predicate is an IEEE non-aware branch and FPCC[NAN] is set. When this condition is detected, FPSR[BSUN] is set. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 6-19 Table 6-12. BSUN Exception Enabled/Disabled Results Condition BSUN Description Exception disabled 0 The floating-point condition is evaluated as if it were the equivalent IEEE-aware conditional predicate. No exceptions are taken. Exception Enabled 1 The processor takes a floating-point pre-instruction exception. The BSUN exception is unique in that the exception is taken before the conditional predicate is evaluated. If the user BSUN exception handler fails to update the PC to the instruction after the excepting instruction when returning, the exception executes again. Any of the following actions prevent taking the exception again: • Clearing FPSR[NAN] • Disabling FPCR[BSUN] • Incrementing the stored PC in the stack bypasses the conditional instruction. This applies to situations where fall-through is desired. Note that to accurately calculate the PC increment requires knowledge of the size of the bypassed conditional instruction. 6.6.1.2 Input Not-A-Number (INAN) The INAN exception is a mechanism for handling a user-defined, non-IEEE data type. If either input operand is a NAN, FPSR[INAN] is set. By enabling this exception, the user can override the default action taken for NAN operands. Because FMOVEM, FMOVE FPCR, and FSAVE instructions do not modify status bits, they cannot generate exceptions. Therefore, these instructions are useful for manipulating INANs. See Table 6-13. Table 6-13. INAN Exception Enabled/Disabled Results Condition INAN Description Exception disabled 0 If the destination data format is single- or double-precision, a NAN is generated with a mantissa of all ones and a sign of zero transferred to the destination. If the destination data format is B, W, or L, a constant of all ones is written to the destination. Exception enabled 1 The result written to the destination is the same as the exception disabled case unless the exception occurs on a FMOVE OUT, in which case the destination is unaffected. 6.6.1.3 Input Denormalized Number (IDE) The input denorm bit, FPCR[IDE], provides software support for denormalized operands. When the IDE exception is disabled, the operand is treated as zero, FPSR[INEX] is set, and the operation proceeds. When the IDE exception is enabled and an operand is denormalized, an IDE exception is taken, but FPSR[INEX] is not set to allow the handler to set it appropriately. See Table 6-14. Note that the FPU never generates denormalized numbers. If necessary, software can create them in the underflow exception handler. Table 6-14. IDE Exception Enabled/Disabled Results Condition IDE Description Exception disabled 0 Any denormalized operand is treated as zero, FPSR[INEX] is set, and the operation proceeds. Exception enabled 1 The result written to the destination is the same as the exception disabled case unless the exception occurs on a FMOVE OUT, in which case the destination is unaffected. FPSR[INEX] is not set to allow the handler to set it appropriately. MCF548x Reference Manual, Rev. 5 6-20 Freescale Semiconductor Floating-Point Exceptions 6.6.1.4 Operand Error (OPERR) The operand error exception encompasses problems arising in a variety of operations, including errors too infrequent or trivial to merit a specific exception condition. Basically, an operand error occurs when an operation has no mathematical interpretation for the given operands. Table 6-15 lists possible operand errors. When one occurs, FPSR[OPERR] is set. Table 6-15. Possible Operand Errors Instruction Condition Causing Operand Error FADD [(+∞) + (-∞)] or [(-∞) + (+∞)] FDIV (0 ÷ 0) or (∞ ÷ ∞) FMOVE OUT (to B, W, or L) Integer overflow, source is NAN or ±∞ FMUL One operand is 0 and the other is ±∞ FSQRT Source is < 0 or -∞ FSUB [(+∞) - (+∞)] or [(-∞) - (-∞)] Table 6-16 describes results when the exception is enabled and disabled. Table 6-16. OPERR Exception Enabled/Disabled Results Condition OPERR Description Exception disabled 0 When the destination is a floating-point data register, the result is a double-precision NAN, with its mantissa set to all ones and the sign set to zero (positive). For a FMOVE OUT instruction with the format S or D, an OPERR exception is impossible. With the format B, W, or L, an OPERR exception is possible only on a conversion to integer overflow, or if the source is either an infinity or a NAN. On integer overflow and infinity source cases, the largest positive or negative integer that can fit in the specified destination size (B, W, or L) is stored. In the NAN source case, a constant of all ones is written to the destination. Exception enabled 1 The result written to the destination is the same as for the exception disabled case unless the exception occurred on a FMOVE OUT, in which case the destination is unaffected. If desired, the user OPERR handler can overwrite the default result. 6.6.1.5 Overflow (OVFL) An overflow exception is detected for arithmetic operations in which the destination is a floating-point data register or memory when the intermediate result’s exponent is greater than or equal to the maximum exponent value of the selected rounding precision. Overflow occurs only when the destination is S- or D-precision format; overflows for other formats are handled as operand errors. At the end of any operation that could potentially overflow, the intermediate result is checked for underflow, rounded, and then checked for overflow before it is stored to the destination. If overflow occurs, FPSR[OVFL,INEX] are set. Even if the intermediate result is small enough to be represented as a double-precision number, an overflow can occur if the magnitude of the intermediate result exceeds the range of the selected rounding precision format. See Table 6-17. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 6-21 Table 6-17. OVFL Exception Enabled/Disabled Results Condition OVFL Description Exception disabled 0 The values stored in the destination based on the rounding mode defined in FPCR[MODE]. RN Infinity, with the sign of the intermediate result. RZ Largest magnitude number, with the sign of the intermediate result. RM For positive overflow, largest positive normalized number For negative overflow, -∞. RP For positive overflow, +∞ For negative overflow, largest negative normalized number. Exception enabled 1 The result written to the destination is the same as for the exception disabled case unless the exception occurred on a FMOVE OUT, in which case the destination is unaffected. If desired, the user OVFL handler can overwrite the default result. 6.6.1.6 Underflow (UNFL) An underflow exception occurs when the intermediate result of an arithmetic instruction is too small to be represented as a normalized number in a floating-point register or memory using the selected rounding precision; that is, when the intermediate result exponent is less than or equal to the minimum exponent value of the selected rounding precision. Underflow can only occur when the destination format is single or double precision. When the destination is byte, word, or longword, the conversion underflows to zero without causing an underflow or an operand error. At the end of any operation that could underflow, the intermediate result is checked for underflow, rounded, and checked for overflow before it is stored in the destination. FPSR[UNFL] is set if underflow occurs. If the underflow exception is disabled, FPSR[INEX] is also set. Even if the intermediate result is large enough to be represented as a double-precision number, an underflow can occur if the magnitude of the intermediate result is too small to be represented in the selected rounding precision. Table 6-18 shows results when the exception is enabled or disabled. Table 6-18. UNFL Exception Enabled/Disabled Results Condition UNFL Description Exception disabled 0 The stored result is defined below. The UNFL exception also sets FPSR[INEX] if the UNFL exception is disabled. RN Zero, with the sign of the intermediate result. RZ Zero, with the sign of the intermediate result. RM For positive underflow, + 0 For negative underflow, smallest negative normalized number. RP For positive underflow, smallest positive normalized number For negative underflow, - 0 Exception enabled 1 The result written to the destination is the same as for the exception disabled case unless the exception occurs on a FMOVE OUT, in which case the destination is unaffected. If desired, the user UNFL handler can overwrite the default result. The UNFL exception does not set FPSR[INEX] if the UNFL exception is enabled so the exception handler can set FPSR[INEX] based on results it generates. 6.6.1.7 Divide-by-Zero (DZ) Attempting to use a zero divisor for a divide instruction causes a divide-by-zero exception. When a divide-by-zero is detected, FPSR[DZ] is set. Table 6-19 shows results when the exception is enabled or disabled. MCF548x Reference Manual, Rev. 5 6-22 Freescale Semiconductor Floating-Point Exceptions Table 6-19. DZ Exception Enabled/Disabled Results Condition DZ Exception disabled 0 The destination floating-point data register is written with infinity with the sign set to the exclusive OR of the signs of the input operands. Exception enabled 1 The destination floating-point data register is written as in the exception is disabled case. 6.6.1.8 Description Inexact Result (INEX) An INEX exception condition exists when the infinitely precise mantissa of a floating-point intermediate result has more significant bits than can be represented exactly in the selected rounding precision or in the destination format. If this condition occurs, FPSR[INEX] is set and the infinitely-precise result is rounded according to Table 6-20. Table 6-20. Inexact Rounding Mode Values Mode Result RN The representable value nearest the infinitely-precise intermediate value is the result. If the two nearest representable values are equally near, the one whose lsb is 0 (even) is the result. This is sometimes called round-to-nearest-even. RZ The result is the value closest to and no greater in magnitude than the infinitely-precise intermediate result. This is sometimes called chop-mode, because the effect is to clear bits to the right of the rounding point. RM The result is the value closest to and no greater than the infinitely-precise intermediate result (possibly -×). RP The result is the value closest to and no less than the infinitely-precise intermediate result (possibly +×). FPSR[INEX] is also set for any of the following conditions: • If an input operand is a denormalized number and the IDE exception is disabled • An overflowed result • An underflowed result with the underflow exception disabled Table 6-18 shows results when the exception is enabled or disabled. Table 6-21. INEX Exception Enabled/Disabled Results Condition INEX Description Exception disabled 0 The result is rounded and then written to the destination. Exception enabled 1 The result written to the destination is the same as for the exception disabled case unless the exception occurred on a FMOVE OUT, in which case the destination is unaffected. If desired, the user INEX handler can overwrite the default result. 6.6.2 Floating-Point State Frames Floating-point arithmetic exception handlers should have FSAVE as the first floating-point instruction; otherwise, encountering another floating-point arithmetic instruction will cause the exception to be reported again. After FSAVE executes, the handler should use FMOVEM to access floating-point data registers, because it cannot generate further exceptions or change the FPSR. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 6-23 Note that if no intervention is needed, instead of FSAVE, the handler can simply clear the appropriate FPCR and FPSR bits and then return from the exception. Because the FPCR and FPSR are written in the FSAVE frame, a context switch needs only execute FSAVE and FMOVEM for data registers. The new process needs to load data registers by using a FMOVEM/FRESTORE sequence before it can continue. FSAVE operations always write a 4-longword floating-point state frame that holds a 64-bit exception operand. Figure 6-13 shows FSAVE frame contents. 31 24 23 19 18 16 15 0 Format word Frame Format Control Register (FPCR) 0000_0 Vector Exception operand upper 32 bits Exception operand lower 32 bits Status register (FPSR) Figure 6-13. Floating-Point State Frame Contents Table 6-22 describes format word fields. Table 6-22. Format Word Field Descriptions Bits Name 31–24 Frame format 23–19 — 18–16 Vector Description Defines the format of the frame. 0x00 Null Frame (NULL) 0x05 Idle Frame (IDLE) 0xE5 Exception Frame (EXCP) Zeros Exception vector 000 BSUN 001 INEX 010 DZ 011 UNFL 100 OPERR 101 OVFL 110 INAN 111 IDE When FSAVE executes, the floating-point frame reflects the FPU state at the time of the FSAVE. Internally, the FPU can be in the NULL, IDLE, or EXCP states. Upon reset, the FPU is in NULL state, in which all floating-point registers contain NANs and the FPCR, FPSR, and FPIAR contain zeros. The FPU remains in NULL state until execution of an implemented floating-point instruction (except FSAVE). At this point, the FPU transitions from NULL to an IDLE state. A FRESTORE of NULL returns the FPU to NULL state. EXCP state is entered as a result of a floating-point exception or an unsupported data type exception. The vector field identifies exception types associated with the EXCP state. This field and the exception vector taken are determined directly from the exception control (FPCR) and status (FPSR) bits. An FSAVE instruction always clears FPCR after saving its state. Thus, after an FSAVE, a handler does not generate further floating-point exceptions unless the handler re-enables the exceptions. FRESTORE returns FPCR and FPSR to their previous state before entering the handler, as stored in the state frame. A handler could alter the state frame to restore the FPU (using FRESTORE) into a different state than that saved by using FSAVE. MCF548x Reference Manual, Rev. 5 6-24 Freescale Semiconductor Instructions Normally, an exception handler executes FSAVE, processes the exception, clears the exception bit in the FSAVE state frame status word, and executes FRESTORE. If appropriate exception bits set in the status word are not cleared, the same exception is taken again. If multiple exception bits are set in the status word, each should be processed, cleared, and restored by their respective handlers. In this way, all exceptions are processed in priority order. If it is not necessary to handle multiple exceptions, the exception model can be simplified (after any processing) by the handler manually loading FPCR and FPSR and then discarding the state frame before executing an RTE. Given that state frames are four longwords, it may be quicker to discard the state frame by incrementing the address pointer (often the system stack pointer, A7) by 16. The exception operand, contained in longwords two and three of the FSAVE frame, is always the value of the destination operand before the operation which caused the exception commenced. Thus, for dyadic register-to-register operations, the exception operand contains the value of the destination register before it was overwritten by the operation which caused the exception. This operand can be retrieved by an exception handler that needs both original operands in order to process the exception. 6.7 Instructions This section includes an instruction set summary, execution times, and differences between ColdFire and M68000 FPU programming models. For detailed instruction descriptions, see the ColdFire Programmer’s Reference Manual. 6.7.1 Floating-Point Instruction Overview ColdFire instructions are 16-, 32-, or 48-bits long. The general definition of a floating-point operation and effective addressing mode require 32 bits; some addressing modes require another 16-bit extension word. Table 6-23 shows the minimum size instruction formats. The first word is the opword; the second is extension word 1. Table 6-23. Floating-Point Instruction Formats Mnemonic Instruction Code FABS 1 1 1 1 0 0 1 0 0 0 ea mode ea reg 0 r/m 0 src spec dest reg opmode FADD 1 1 1 1 0 0 1 0 0 0 ea mode ea reg 0 r/m 0 src spec dest reg opmode FBcc 1 1 1 1 0 0 1 0 1 s z cond predicate 16b displacement or MS Word of 32b LS Word of 32b Displacement FCMP 1 1 1 1 0 0 1 0 0 0 ea mode ea reg 0 r/m 0 src spec dest reg 0 1 1 1 0 0 0 FDIV 1 1 1 1 0 0 1 0 0 0 ea mode ea reg 0 r/m 0 src spec dest reg FINT 1 1 1 1 0 0 1 0 0 0 ea mode ea reg 0 r/m 0 src spec dest reg 0 0 0 0 0 0 1 FINTRZ 1 1 1 1 0 0 1 0 0 0 ea mode ea reg 0 r/m 0 src spec dest reg 0 0 0 0 0 1 1 opmode MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 6-25 Table 6-23. Floating-Point Instruction Formats (Continued) Mnemonic FMOVE Instruction Code 1 1 1 1 0 0 1 0 0 0 ea mode ea reg 0 r/m 0 src spec dest reg 1 1 1 1 0 0 1 0 0 0 ea mode ea reg 0 1 1 dest fmt 1 1 1 1 0 0 1 0 0 0 ea mode ea reg 1 0 d r FMOVEM 1 1 1 1 0 0 1 0 0 0 ea mode ea reg 1 1 d 1 0 r FMUL 1 1 1 1 0 0 1 0 0 0 ea mode ea reg 0 r/m 0 src spec dest reg opmode FNEG 1 1 1 1 0 0 1 0 0 0 ea mode ea reg 0 r/m 0 src spec dest reg opmode FNOP 1 1 1 1 0 0 1 0 1 0 0 0 0 0 0 0 0 FRESTOR E 1 1 1 1 0 0 1 1 0 1 ea mode ea reg FSAVE 1 1 1 1 0 0 1 1 0 0 ea mode ea reg FSQRT 1 1 1 1 0 0 1 0 0 0 ea mode ea reg 0 r/m 0 src spec dest reg opmode FSUB 1 1 1 1 0 0 1 0 0 0 ea mode ea reg 0 r/m 0 src spec dest reg opmode FTST 1 1 1 1 0 0 1 0 0 0 ea mode ea reg 0 r/m 0 src spec dest reg 0 1 1 1 0 1 0 0 reg sel 0 0 0 src reg 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 opmode register list 0 0 0 0 0 0 0 0 Table 6-24 defines the terminology used in Table 6-23. Table 6-24. Instruction Format Terminology Term Definition Instructions Instructions appear in memory as sequential, 16-bit values, and are read in the above table left to right. An instruction can have from 1 to 3 16-bit words. A shaded block indicates this word is never used and is not present. EA MODE EA REG Defines the effective address for an operand located external to the FPU. For most FPU instructions, this field defines the location of an external source operand; for FP store operations, it specifies the destination location. R/M If R/M = 0, an FPU data register is one source operand, otherwise the source operand is specified by the EA {MODE, REG} fields. SRC SPEC Defines the format (byte, word, longword, single-, or double-precision) of an external operand. DEST REG Specifies the destination FPU data register. COND PREDICATE Defines the condition to be evaluated (EQ, NE, and so on) during the execution of the FPU conditional branch instruction. MCF548x Reference Manual, Rev. 5 6-26 Freescale Semiconductor Instructions Table 6-24. Instruction Format Terminology (Continued) Term Definition OPMODE Defines the exact operation to be performed by the FPU. SZ Defines the length of the PC-relative displacement for the FPU conditional branch instruction. If SZ = 0, the displacement is 16 bits, otherwise a 32-bit displacement is used. dr Specifies direction of the MOVE transfer. As a 0, it moves from memory to the FP; as 1, it moves from the FP to memory. REGISTER LIST Defines FPU data registers to be moved during the execution of the FMOVEM instruction. REG SEL 6.7.2 Indicates the FPU control register to be moved during execution of an FMOVE control register instruction. Floating-Point Instruction Execution Timing Table 6-25 shows the ColdFire execution times for the floating-point instructions in terms of processor core clock cycles. Each timing entry is presented as C(r/w). • C = The number of processor clock cycles including all applicable operand reads and writes plus all internal core cycles required to complete instruction execution • r = The number of operand reads • w = The number of operand writes NOTE Timing assumptions are the same as those for the ColdFire ISA. See the ColdFire Microprocessor Family Programmer’s Reference Manual. Table 6-25. Floating-Point Instruction Execution Times1, 2, 3 Effective Address <ea> Opcode Format FPn Dn (An) (An)+ -(An) (d16,An) (d16,PC) FABS <ea>y,FPx 1(0/0) 1(0/0) 1(1/0) 1(1/0) 1(1/0) 1(1/0) 1(1/0) FADD <ea>y,FPx 4(0/0) 4(0/0) 4(1/0) 4(1/0) 4(1/0) 4(1/0) 4(1/0) FBcc <label> — — — — — — 2(0/0) if correct, 9(0/0) if incorrect FCMP <ea>y,FPx 4(0/0) 4(0/0) 4(1/0) 4(1/0) 4(1/0) 4(1/0) 4(1/0) FDIV <ea>y,FPx 23(0/0) 23(0/0) 23(1/0) 23(1/0) 23(1/0) 23(1/0) 23(1/0) FINT <ea>y,FPx 4(0/0) 4(0/0) 4(1/0) 4(1/0) 4(1/0) 4(1/0) 4(1/0) FINTRZ <ea>y,FPx 4(0/0) 4(0/0) 4(1/0) 4(1/0) 4(1/0) 4(1/0) 4(1/0) FMOVE <ea>y,FPx 1(0/0) 1(0/0) 1(1/0) 1(1/0) 1(1/0) 1(1/0) 1(1/0) FPy,<ea>x — 2(0/1) 2(0/1) 2(0/1) 2(0/1) 2(0/1) — <ea>y,FP*R — 6(0/0) 6(1/0) 6(1/0) 6(1/0) 6(1/0) 6(1/0) FP*R,<ea>x — 1(0/0) 1(0/1) 1(0/1) 1(0/1) 1(0/1) — MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 6-27 Table 6-25. Floating-Point Instruction Execution Times1, 2, 3 (Continued) Effective Address <ea> Opcode Format FPn Dn (An) (An)+ -(An) (d16,An) (d16,PC) <ea>y,#list — — 2n(2n/0) — — 2n(2n/0) 2n(2n/0) #list,<ea>x — — 1+2n(0/2n) — — 1+2n(0/2n) — FMUL <ea>y,FPx 4(0/0) 4(0/0) 4(1/0) 4(1/0) 4(1/0) 4(1/0) 4(1/0) FNEG <ea>y,FPx 1(0/0) 1(0/0) 1(1/0) 1(1/0) 1(1/0) 1(1/0) 1(1/0) — — — — — — 2(0/0) FMOVEM 4 FNOP FRESTORE <ea>y — — 6(4/0) — — 6(4/0) 6(4/0) FSAVE <ea>x — — 7(0/4) — — 7(0/4) — FSQRT <ea>y,FPx 56(0/0) 56(0/0) 56(1/0) 56(1/0) 56(1/0) 56(1/0) 56(1/0) FSUB <ea>y,FPx 4(0/0) 4(0/0) 4(1/0) 4(1/0) 4(1/0) 4(1/0) 4(1/0) FTST <ea>y,FPx 1(0/0) 1(0/0) 1(1/0) 1(1/0) 1(1/0) 1(1/0) 1(1/0) 1 Add 1(1/0) for an external read operand of double-precision format for all instructions except FMOVEM, and 1(0/1) for FMOVE FPy,<ea>x when the destination is double-precision. 2 If the external operand is an integer format (byte, word, longword), there is a 4 cycle conversion time which must be added to the basic execution time. 3 If any exceptions are enabled, the execution time for FMOVE FPy,<ea>x increases by one cycle. If the BSUN exception is enabled, the execution time for FBcc increases by one cycle. 4 For FMOVEM, n refers to the number of registers being moved. The ColdFire architecture supports concurrent execution of integer and floating-point instructions. The latencies in this table define the execution time needed by the FPU. After a multi-cycle FPU instruction is issued, subsequent integer instructions can execute concurrently with the FPU execution. For this sequence, the floating-point instruction occupies only one OEP cycle. 6.7.3 Key Differences between ColdFire and M68000 FPU Programming Models This section is intended for compiler developers and developers porting assembly language routines from the M68000 family to ColdFire. It highlights major differences between the ColdFire FPU instruction set architecture (ISA) and the equivalent M68000 family ISA, using the MC68060 as the reference. The internal FPU datapath width is the most obvious difference. ColdFire uses 64-bit double-precision and the M68000 family uses 80-bit extended precision. Other differences pertain to supported addressing modes, both across all FPU instructions as well as specific opcodes. Table 6-26 lists key differences. Because all ColdFire implementations support instruction sizes of 48 bits or less, M68000 operations requiring larger instruction lengths cannot be supported. . Table 6-26. Key Programming Model Differences Feature Internal datapath width Support for fpGEN d8(An,Xi),FPx M68000 ColdFire 80 bits 64 bits Yes No MCF548x Reference Manual, Rev. 5 6-28 Freescale Semiconductor Instructions Table 6-26. Key Programming Model Differences (Continued) Feature M68000 ColdFire Support for fpGEN xxx.{w,l},FPx Yes No Support for fpGEN d8(PC,Xi),FPx Yes No Support for fpGEN #xxx,FPx Yes No Support for fmovem (Ay)+,#list Yes No Support for fmovem #list,-(Ax) Yes No Support for fmovem FP Control Registers Yes No Some differences affect function activation and return. M68000 subroutines typically began with FMOVEM #list,-(a7) to save registers on the system stack, with each register occupying three longwords. In ColdFire, each register occupies two longwords and the stack pointer must be adjusted before the FMOVEM instruction. A similar sequence generally occurs at the end of the function, preparing to return control to the calling routine. The examples in Table 6-27, Table 6-28, and Table 6-29 show a M68000 operation and the equivalent ColdFire sequence. Table 6-27. M68000/ColdFire Operation Sequence 11 M68000 ColdFire Equivalent fmovem.x #list,-(a7) lea -8*n(a7),a7;allocate stack space fmovem.d #list,(a7) ;save FPU registers fmovem.x (a7)+,#list fmovem.d (a7),#list ;restore FPU registers lea 8*n(a7),a7 ;deallocate stack space 1 n is the number of FP registers to be saved/restored. If the subroutine includes LINK and UNLK instructions, the stack space needed for FPU register storage can be factored into these operations and LEA instructions are not required. The M68000 FPU supports loads and stores of multiple control registers (FPCR, FPSR, and FPIAR) with one instruction. For ColdFire, only one can be moved at a time. For instructions that require an unsupported addressing mode, the operand address can be formed with a LEA instruction immediately before the FPU operation. See Table 6-28. Table 6-28. M68000/ColdFire Operation Sequence 2 M68000 ColdFire Equivalent fadd.s label,fp2 lea label,a0;form pointer to data fadd.s (a0),fp2 fmul.d (d8,a1,d7),fp5 lea (d8,a1,d7),a0;form pointer to data fmul.d (a0),fp5 fcmp.l (d8,pc,d2),fp3 lea (d8,pc,d2),a0;form pointer to data fcmp.l (a0),fp3 The M68000 FPU allows floating-point instructions to directly specify immediate values; the ColdFire FPU does not support these types of immediate constants. It is recommended that floating-point immediate MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 6-29 values be moved into a table of constants that can be referenced using PC-relative addressing or as an offset from another address pointer. See Table 6-29. Table 6-29. M68000/ColdFire Operation Sequence 3 M68000 ColdFire Equivalent fadd.l #imm1,fp3 fadd.l (imm1_label,pc),fp3 fsub.s #imm2,fp4 fsub.s (imm2_label,pc),fp3 fdiv.d #imm3,fp5 fdiv.d (imm3_label,pc),fp3 align 4 imm1_label: long imm1 ;integer longword imm2_label: long imm2 ;single-precision imm3_label: long imm3_upper, imm3_lower ;double-precision Finally, ColdFire and the M68000 differ in how exceptions are made pending. In the ColdFire exception model, asserting both an FPSR exception indicator bit and the corresponding FPCR enable bit makes an exception pending. Thus, a pending exception state can be created by loading FPSR and/or FPCR. On the M68000, this type of pending exception is not possible. Analysis of compiled floating-point applications indicates these differences account for most of the changes between M68000-compatible text and the equivalent ColdFire program. MCF548x Reference Manual, Rev. 5 6-30 Freescale Semiconductor Chapter 7 Local Memory This chapter describes the MCF548x implementation of the ColdFire Version 4e local memory specification. It consists of two major sections. • Section 7.2, “SRAM Overview,” describes the MCF548x core’s local static RAM (SRAM) implementation. It covers general operations, configuration, and initialization. It also provides information and examples showing how to minimize power consumption when using the SRAM. • Section 7.7, “Cache Overview,” describes the MCF548x cache implementation, including organization, configuration, and coherency. It describes cache operations and how the cache interfaces with other memory structures. 7.1 Interactions between Local Memory Modules Depending on configuration information, instruction fetches and data read accesses may be sent simultaneously to the SRAM and cache controllers. This approach is required because all three controllers are memory-mapped devices, and the hit/miss determination is made concurrently with the read data access. Power dissipation can be minimized by configuring the RAMBARs to mask unused address spaces whenever possible. If the access address is mapped into the region defined by the SRAM (and this region is not masked), the SRAM provides the data back to the processor, and the cache data is discarded. Accesses from the SRAM module are never cached. The complete definition of the processor’s local bus priority scheme for read references is as follows: if (SRAM “hits”) SRAM supplies data to the processor else if (data cache “hits”) data cache supplies data to the processor else system memory reference to access data For data write references, the memory mapping into the local memories is resolved before the appropriate destination memory is accessed. Accordingly, only the targeted local memory is accessed for data write transfers. NOTE The two SRAMs discussed in this chapter is on the processor local bus. There is a third 32-Kbyte SRAM on the MCF548x device. See Chapter 16, “32-Kbyte System SRAM,” for more information. 7.2 SRAM Overview The two 4-Kbyte, on-chip SRAM modules provide the core with pipelined, single-cycle access to memory. Memory can be independently mapped to any 0-modulo-4K location in the 4-Gbyte address space and configured to respond to either instruction or data accesses. The following summarizes features of the MCF548x SRAM implementation: • Two 4-Kbyte SRAMs, organized as 1024 x 32 bits • Single-cycle throughput. When the pipeline is full, one access can occur per clock cycle. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 7-1 • • • • 7.3 Physical location on the processor’s high-speed local bus with a user-programmed connection to the internal instruction or data bus Memory location programmable on any 0-modulo-4K address boundary Byte, word, and longword address capabilities The RAM base address registers (RAMBAR0 and RAMBAR1) define the logical base address, attributes, and access types for the two SRAM modules. SRAM Operation Each SRAM module provides a general-purpose memory block that the ColdFire processor can access with single-cycle throughput. The location of the memory block can be specified to any 0-module-4K address boundary in the 4-Gbyte address space by RAMBARn[BA], described in Section 7.4.1, “SRAM Base Address Registers (RAMBAR0/RAMBAR1).” The memory is ideal for storing critical code or data structures or for use as the system stack. Because the SRAM module connects physically to the processor’s high-speed local bus, it can service processor-initiated accesses or memory-referencing debug module commands. The Version 4e ColdFire processor core implements a Harvard memory architecture. Each SRAM module may be logically connected to either the processor’s internal instruction or data bus. This logical connection is controlled by a configuration bit in the RAM base address registers (RAMBAR0 and RAMBAR1). If an instruction fetch is mapped into the region defined by the SRAM, the SRAM sources the data to the processor and any cache data is discarded. Likewise, if a data access is mapped into the region defined by the SRAM, the SRAM services the access and the cache is not affected. Accesses from SRAM modules are never cached, and debug-initiated references are treated as data accesses. Note also that the SRAMs cannot be accessed by the on-chip DMAs. The on-chip system configuration allows concurrent core and DMA execution, where the CPU can reference code or data from the internal SRAMs or caches while performing a DMA transfer. Accesses are attempted in the following order: 1. SRAM 2. Cache (if space is defined as cacheable) 3. System SRAM, MBAR space, or external access 7.4 SRAM Register Definition The SRAM programming model consists of RAMBAR0 and RAMBAR1. 7.4.1 SRAM Base Address Registers (RAMBAR0/RAMBAR1) The SRAM modules are configured through the RAMBARs, shown in Figure 7-1. Each RAMBAR holds the base address of the SRAM. The MOVEC instruction provides write-only access to this register from the processor. Each RAMBAR can be read or written from the debug module in a similar manner. All undefined RAMBAR bits are reserved. These bits are ignored during writes to the RAMBAR and return zeros when read from the debug module. The valid bits, RAMBARn[V], are cleared at reset, disabling the SRAM modules. All other bits are unaffected. NOTE RAMBARn is read/write by the debug module. MCF548x Reference Manual, Rev. 5 7-2 Freescale Semiconductor SRAM Register Definition 31 30 29 28 27 26 25 24 R 23 22 21 20 19 18 17 16 BA W Reset 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 0 0 WP D/I 0 C/I SC SD UC UD V 0 0 0 0 0 0 0 0 0 0 0 0 R BA W Reset 0 0 0 Reg Addr 0 CPU space + 0xC04 (RAMBAR0), 0xC05 (RAMBAR1) Figure 7-1. SRAM Base Address Registers (RAMBARn) RAMBARn fields are described in detail in Table 7-1. Table 7-1. RAMBARn Field Description Bits Name Description 31–12 BA Base address. Defines the SRAM module’s word-aligned base address. Each SRAM module occupies a 4-Kbyte space defined by the contents of BA. SRAM may reside on any 4-Kbyte boundary in the 4 Gbyte address space. 11–9 — Reserved. Should be cleared. 8 WP Write protect. Controls read/write properties of the SRAM. 0 Allows read and write accesses to the SRAM module 1 Allows only read accesses to the SRAM module. Any attempted write reference generates an access error exception to the ColdFire processor core. 7 D/I Data/instruction bus. Indicates whether SRAM is connected to the internal data or instruction bus. 0 Data bus 1 Instruction bus 6 — Reserved, should be cleared. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 7-3 Table 7-1. RAMBARn Field Description (Continued) Bits Name Description 5 C/I 4 SC 3 SD 2 UC 1 UD Address space masks (ASn). These fields allow certain types of accesses to be masked, or inhibited from accessing the SRAM module. These bits are useful for power management as described in Section 7.6, “Power Management.” In particular, C/I is typically set. The address space mask bits are follows: C/I = CPU space/interrupt acknowledge cycle mask. Note that C/I must be set if BA = 0. SC = Supervisor code address space mask SD = Supervisor data address space mask UC = User code address space mask UD = User data address space mask For each ASn bit: 0 An access to the SRAM module can occur for this address space 1 Disable this address space from the SRAM module. If a reference using this address space is made, it is inhibited from accessing the SRAM module and is processed like any other non-SRAM reference. 0 V Valid. Enables/disables the SRAM module. V is cleared at reset. 0 RAMBAR contents are not valid. 1 RAMBAR contents are valid. The mapping of a given access into the SRAM uses the following algorithm to determine if the access hits in the memory: if (RAMBAR[0] = 1) if (((access = instructionFetch) & (RAMBAR[7] = 1)) | ((access = dataReference) & (RAMBAR[7] = 0))) if (requested address[31:10] = RAMBAR[31:10]) if (requested address[31:n] = RAMBAR[31:n] if (ASn of the requested type = 0) Access is mapped to the SRAM module if (access = read) Read the SRAM and return the data if (access = write) if (RAMBAR[8] = 0) Write the data into the SRAM else Signal a write-protect access error ASn refers to the five address space mask bits: C/I, SC, SD, UC, and UD. 7.5 SRAM Initialization After a hardware reset, the contents of each SRAM module are undefined. The valid bits, RAMBARn[V], are cleared, disabling the SRAM modules. If the SRAM requires initialization with instructions or data, the following steps should be performed: 1. Load RAMBARn with bit 7 = 0, mapping the SRAM module to the desired location. Clearing RAMBARn[7] logically connects the SRAM module to the processor’s data bus. 2. Read the source data and write it to the SRAM. Various instructions support this function, including memory-to-memory move instructions and the move multiple instruction (MOVEM). MOVEM is optimized to generate line-sized burst fetches on line-aligned addresses, so it generally provides maximum performance. MCF548x Reference Manual, Rev. 5 7-4 Freescale Semiconductor SRAM Initialization 3. After the data is loaded into the SRAM, it may be appropriate to revise the RAMBAR attribute bits, including the write-protect and address-space mask fields. If the SRAM contains instructions, RAMBAR[D/I] must be set to logically connect the memory to the processor’s internal instruction bus. Remember that the SRAM cannot be accessed by the on-chip DMAs. The on-chip system configuration allows concurrent core and DMA execution, where the core can execute code out of internal SRAM or cache during DMA access. The ColdFire processor or an external emulator using the debug module can perform these initialization functions. 7.5.1 SRAM Initialization Code The code segment below initializes the SRAM using RAMBAR0. The code sets the base address of the SRAM at 0x2000_0000 before it initializes the SRAM to zeros. RAMBASE RAMVALID move.l movec.l EQU 0x20000000 EQU 0x00000035 #RAMBASE+RAMVALID,D0 D0, RAMBAR0 ;set this variable to 0x20000000 ;load RAMBASE + valid bit into D0 ;load RAMBAR0 and enable SRAM The following loop initializes the entire SRAM to zero: lea.l move.l RAMBASE,A0 #1024,D0 ;load pointer to SRAM ;load loop counter into D0 (A0)+ #1,D0 SRAM_INIT_LOOP ;clear 4 bytes of SRAM ;decrement loop counter ;exit if done; else continue looping SRAM_INIT_LOOP: clr.l subq.l bne.b The following function copies the number of bytesToMove from the source (*src) to the processor’s local SRAM at an offset relative to the SRAM base address defined by destinationOffset. The bytesToMove must be a multiple of 16. For best performance, source and destination SRAM addresses should be line-aligned (0-modulo-16). ; copyToCpuRam (*src, destinationOffset, bytesToMove) RAMBASE RAMFLAGS EQU EQU lea.l movem.l ; ; ; ; ; ; 0x20000000 0x00000035 ;SRAM base address ;RAMBAR valid + mask bits -12(a7),a7;allocate temporary space #0x1c,(a7);store D2/D3/D4 registers stack arguments and locations +0 saved d2 +4 saved d3 +8 saved d4 +12 returnPc +16 pointer to source operand MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 7-5 ; +20 ; +24 loop: 7.6 destinationOffset bytesToMove move.l movec.l RAMBASE+RAMFLAGS,a0 ;define RAMBAR0 contents a0,rambar0;load it move.l 16(a7),a0;load argument defining *src lea.l add.l RAMBASE,a1;memory pointer to SRAM base 20(a7),a1;include destinationOffset move.l asr.l 24(a7),d4;load byte count #4,d4 ;divide by 16 to convert to loop count .align movem.l movem.l lea.l lea.l subq.l bne.b 4 ;force loop on 0-mod-4 address (a0),#0xf;read 16 bytes from source #0xf,(a1);store into SRAM destination 16(a0),a0;increment source pointer 16(a1),a1;increment destination pointer #1,d4 ;decrement loop counter loop ;if done, then exit, else continue movem.l lea.l rts (a7),#0x1c;restore d2/d3/d4 registers 12(a7),a7;deallocate temporary space Power Management Because processor memory references may be simultaneously sent to an SRAM module and cache, power can be minimized by configuring RAMBAR address space masks as precisely as possible. For example, if an SRAM is mapped to the internal instruction bus and contains instruction data, setting the ASn mask bits associated with operand references can decrease power dissipation. Similarly, if the SRAM contains data, setting ASn bits associated with instruction fetches minimizes power. Table 7-2 shows typical RAMBAR configurations. . Table 7-2. Examples of Typical RAMBAR Settings Data Contained in SRAM 7.7 RAMBAR[5–0] Code only 0x2B Data only 0x35 Both code and data 0x21 Cache Overview This section describes the MCF548x cache implementation, including organization, configuration, and coherency. It describes cache operations and how the cache interacts with other memory structures. The MCF548x implements a special branch instruction cache for accelerating branches, enabled by a bit in the cache access control register (CACR[BEC]). The branch cache is described in Section 3.2.1.1.1, “Branch Acceleration.” MCF548x Reference Manual, Rev. 5 7-6 Freescale Semiconductor Cache Organization The MCF548x processor’s Harvard memory structure includes a 32-Kbyte data cache and a 32-Kbyte instruction cache. Both are nonblocking and 4-way set-associative with a 16-byte line. The cache improves system performance by providing single-cycle access to the instruction and data pipelines. This decouples processor performance from system memory performance, increasing bus availability for on-chip DMA or external devices. Figure 7-2 shows the organization and integration of the data cache. Cache Control External Bus Control Control Logic Control Data Array ColdFire Processor Core System Integration Unit (SIU) Directory Array Data Data Address Address/ Data Data Path Address Address Path Figure 7-2. Data Cache Organization Both caches implement line-fill buffers to optimize line-sized burst accesses. The data cache supports operation of copyback, write-through, or cache-inhibited modes. A four-entry, 32-bit buffer supports cache line-push operations, and can be configured to defer write buffering in write-through or cache-inhibited modes. The cache lock feature can be used to guarantee deterministic response for critical code or data areas. A nonblocking cache services read hits or write hits from the processor while a fill (caused by a cache allocation) is in progress. As Figure 7-2 shows, accesses use a single bus connected to the cache. All addresses from the processor to the cache are physical addresses. A cache hit occurs when an address matches a cache entry. For a read, the cache supplies data to the processor. For a write, which is permitted only to the data cache, the processor updates the cache. If an access does not match a cache entry (misses the cache) or if a write access must be written through to memory, the cache performs a bus cycle on the internal bus and correspondingly on the external bus by way of the system integration unit (SIU). The cache module does not implement bus snooping; cache coherency with other possible bus masters must be maintained in software. 7.8 Cache Organization A four-way set associative cache is organized as four ways (levels). There are 512 sets in the 32-Kbyte data cache with each line containing 16 bytes (4 longwords). The 32-Kbyte instruction cache has 512 sets. Entire cache lines are loaded from memory by burst-mode accesses that cache 4 longwords of data or instructions. All 4 longwords must be loaded for the cache line to be valid. Figure 7-3 shows data cache organization as well as terminology used. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 7-7 Way 0 Way 1 Way 2 Way 3 • • • • • • Line • • • • • • Set 0 Set 1 Set 510 Set 511 Cache Line Format TAG V M Longword 0 Longword 1 Longword 2 Longword 3 Where: TAG—21-bit address tag V—Valid bit for line M—Modified bit for line (data cache only) Figure 7-3. Data Cache Organization and Line Format A set is a group of four lines (one from each level, or way), corresponding to the same index into the cache array. 7.8.1 Cache Line States: Invalid, Valid-Unmodified, and Valid-Modified As shown in Table 7-3, a data cache line can be invalid, valid-unmodified (often called exclusive), or valid-modified. An instruction cache line can be valid or invalid. Table 7-3. Valid and Modified Bit Settings V M Description 0 x Invalid. Invalid lines are ignored during lookups. 1 0 Valid, unmodified. Cache line has valid data that matches system memory. 1 1 Valid, modified. Cache line contains most recent data, data at system memory location is stale. A valid line can be explicitly invalidated by executing a CPUSHL instruction. 7.8.2 The Cache at Start-Up As Figure 7-4 (A) shows, after power-up, cache contents are undefined; V and M may be set on some lines even though the cache may not contain the appropriate data for start up. Because reset and power-up do not invalidate cache lines automatically, the cache should be cleared explicitly by setting CACR[DCINVA,ICINVA] before the cache is enabled (B). After the entire cache is flushed, cacheable entries are loaded first in way 0. If way 0 is occupied, the cacheable entry is loaded into the same set in way 1, as shown in Figure 7-4 (D). This process is described in detail in Section 7.9, “Cache Operation.” MCF548x Reference Manual, Rev. 5 7-8 Freescale Semiconductor Cache Organization Invalid (V = 0) Valid, not modified (V = 1, M = 0) Valid, modified (V = 1, M = 1) A: Cache population at start-up Way 0Way 1Way 2Way 3 B: Cache after invalidation, C: Cache after loads in before it is enabled Way 0 Way 0Way 1Way 2Way 3 Way 0Way 1Way 2Way 3 At reset, cache contents are indeterminate; V and M may be set. The cache should be cleared explicitly by setting CACR[DCINVA] before the cache is enabled. Setting CACR[DCINVA] invalidates the entire cache. D: First load in Way 1 Way 0Way 1Way 2Way 3 Set 0 Set 511 Initial cacheable accesses to memory-fill positions in way 0. A line is loaded in way 1 only if that set is full in way 0. Figure 7-4. Data Cache—A: at Reset, B: after Invalidation, C and D: Loading Pattern MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 7-9 7.9 Cache Operation Figure 7-5 shows the general flow of a caching operation using the 32-Kbyte data cache as an example. The discussion in this chapter assumes a data cache. Instruction cache operations are similar except that there is no support for writing to the cache; therefore, such notions of modified cache lines and write allocation do not apply. Address 31 13 12 Tag Data/Tag Reference 4 3 0 Way 3 Way 2 Way 1 Way 0 Index Set 0 Set Select A[12:4] Set 1 • • • Set 511 TAG STATUS LW0 LW1 LW2 LW3 • • • TAG • • • • • • • • • • • • • • • STATUS LW0 LW1 LW2 LW3 Data Address A[31:13] MUX 3 2 1 Comparator 0 Line Select Hit 3 Hit 2 Hit 1 Hit 0 Logical OR Hit Figure 7-5. Data Caching Operation The following steps determine if a data cache line is allocated for a given address: 1. The cache set index, A[12:4], selects one cache set. 2. A[31:13] and the cache set index are used as a tag reference or are used to update the cache line tag field. Note that A[31:13] can specify 19 possible address lines that can be mapped to one of the four ways. 3. The four tags from the selected cache set are compared with the tag reference. A cache hit occurs if a tag matches the tag reference and the V bit is set, indicating that the cache line contains valid data. If a cacheable write access hits in a valid cache line, the write can occur to the cache line without having to load it from memory. If the memory space is copyback, the updated cache line is marked modified (M = 1), because the new data has made the data in memory out of date. If the memory location is write-through, the write is passed on to system memory and the M bit is never used. Note that the tag does not have TT or TM bits. To allocate a cache entry, the cache set index selects one of the cache’s 512 sets. The cache control logic looks for an invalid cache line to use for the new entry. If none is available, the cache controller uses a MCF548x Reference Manual, Rev. 5 7-10 Freescale Semiconductor Cache Operation pseudo-round-robin replacement algorithm to choose the line to be deallocated and replaced. First the cache controller looks for an invalid line, with way 0 the highest priority. If all lines have valid data, a 2-bit replacement counter is used to choose the way. After a line is allocated, the pointer increments to point to the next way. Cache lines from ways 0 and 1 can be protected from deallocation by enabling half-cache locking. If CACR[DHLCK,IHLCK] = 1, the replacement pointer is restricted to way 2 or 3. As part of deallocation, a valid, unmodified cache line is invalidated. It is consistent with system memory, so memory does not need to be updated. To deallocate a modified cache line, data is placed in a push buffer (for an external cache line push) before being invalidated. After invalidation, the new entry can replace it. The old cache line may be written after the new line is read. When a cache line is selected to host a new cache entry, the following three things happen: 1. The new address tag bits A[31:13] are written to the tag. 2. The cache line is updated with the new memory data. 3. The cache line status changes to a valid state (V = 1). Read cycles that miss in the cache allocate normally as previously described. Write cycles that miss in the cache do not allocate on a cacheable write-through region, but do allocate for addresses in a cacheable copyback region. A copyback byte, word, longword, or line write miss causes the following: 1. The cache initiates a line fill or flush. 2. Space is allocated for a new line. 3. V and M are both set to indicate valid and modified. 4. Data is written in the allocated space. No write to memory occurs. Note the following: • Read hits cannot change the status bits and no deallocation or replacement occurs; the data or instructions are read from the cache. • If the cache hits on a write access, data is written to the appropriate portion of the accessed cache line. Write hits in cacheable, write-through regions generate an external write cycle and the cache line is marked valid, but is never marked modified. Write hits in cacheable copyback regions do not perform an external write cycle; the cache line is marked valid and modified (V = 1 and M = 1). • Misaligned accesses are broken into at least two cache accesses. • Validity is provided only on a line basis. Unless a whole line is loaded on a cache miss, the cache controller does not validate data in the cache line. Write accesses designated as cache-inhibited by the CACR or ACR bypass the cache and perform a corresponding external write. Normally, cache-inhibited reads bypass the cache and are performed on the external bus. The exception to this normal operation occurs when all of the following conditions are true during a cache-inhibited read: • The cache-inhibited fill buffer bit, CACR[DNFB], is set. • The access is an instruction read. • The access is normal (that is, transfer type (TT) equals 0). In this case, an entire line is fetched and stored in the fill buffer. It remains valid there, and the cache can service additional read accesses from this buffer until either another fill or a cache-invalidate-all operation occurs. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 7-11 Valid cache entries that match during cache-inhibited address accesses are neither pushed nor invalidated. Such a scenario suggests that the associated cache mode for this address space was changed. To avoid this, it is generally recommended to use the CPUSHL instruction to push or invalidate the cache entry or set CACR[DCINVA] to invalidate the data cache before switching cache modes. 7.9.1 Caching Modes For every memory reference generated by the processor or debug module, a set of effective attributes is determined based on the address and the ACRs. Caching modes determine how the cache handles an access. A data access can be cacheable in either write-through or copyback mode; it can be cache-inhibited in precise or imprecise modes. For normal accesses, the ACRn[CM] bit corresponding to the address of the access specifies the caching modes. If an address does not match an ACR, the default caching mode is defined by CACR[DDCM,IDCM]. The specific algorithm is as follows: if (address == ACR0-address including mask) effective attributes = ACR0 attributes else if (address == ACR1-address including mask) effective attributes = ACR1 attributes else effective attributes = CACR default attributes Addresses matching an ACR can also be write-protected using ACR[W]. Addresses that do not match either ACR can be write-protected using CACR[DW]. Reset disables the cache and clears all CACR bits. As shown in Figure 7-4, reset does not automatically invalidate cache entries; they must be invalidated through software. The ACRs allow the defaults selected in the CACR to be overridden. In addition, some instructions (for example, CPUSHL) and processor core operations perform accesses that have an implicit caching mode associated with them. The following sections discuss the different caching accesses and their associated cache modes. 7.9.1.1 Cacheable Accesses If ACRn[CM] or the default field of the CACR indicates write-through or copyback, the access is cacheable. A read access to a write-through or copyback region is read from the cache if matching data is found. Otherwise, the data is read from memory and the cache is updated. When a line is being read from memory for either a write-through or copyback read miss, the longword within the line that contains the core-requested data is loaded first and the requested data is given immediately to the processor, without waiting for the three remaining longwords to reach the cache. The following sections describe write-through and copyback modes in detail. Note that some of this information applies to data caches only. 7.9.1.1.1 Write-Through Mode (Data Cache Only) Write accesses to regions specified as write-through are always passed on to the external bus, although the cycle can be buffered, depending on the state of CACR[DESB]. Writes in write-through mode are handled with a no-write-allocate policy—that is, writes that miss in the cache are written to the external bus but do not cause the corresponding line in memory to be loaded into the cache. Write accesses that hit always write through to memory and update matching cache lines. The cache supplies data to data-read accesses that hit in the cache; read misses cause a new cache line to be loaded into the cache. MCF548x Reference Manual, Rev. 5 7-12 Freescale Semiconductor Cache Operation 7.9.1.1.2 Copyback Mode (Data Cache Only) Copyback regions are typically used for local data structures or stacks to minimize external bus use and reduce write-access latency. Write accesses to regions specified as copyback that hit in the cache update the cache line and set the corresponding M bit without an external bus access. The cache should be flushed using the CPUSHL instruction before invalidating the cache in copyback mode using the CINV bit. Modified cache data is written to memory only if the line is replaced because of a miss or a CPUSHL instruction pushes the line. If a byte, word, longword, or line write access misses in the cache, the required cache line is read from memory, thereby updating the cache. When a miss selects a modified cache line for replacement, the modified cache data moves to the push buffer. The replacement line is read into the cache and the push buffer contents are then written to memory. 7.9.1.2 Cache-Inhibited Accesses Memory regions can be designated as cache-inhibited, which is useful for memory containing targets such as I/O devices and shared data structures in multiprocessing systems. It is also important to not cache the MCF548x memory-mapped registers. If the corresponding ACRn[CM] or CACR[DDCM] indicates cache-inhibited, precise or imprecise, the access is cache-inhibited. The caching operation is identical for both cache-inhibited modes, which differ only regarding recovery from an external bus error. In determining whether a memory location is cacheable or cache-inhibited, the CPU checks memory-control registers in the following order: 1. RAMBARs 2. ACR0 and ACR2 3. ACR1 and ACR3 4. If an access does not hit in the RAMBARs or the ACRs, the default is provided for all accesses in CACR. Cache-inhibited write accesses bypass the cache, and a corresponding external write is performed. Cache-inhibited reads bypass the cache and are performed on the external bus, except when all of the following conditions are true: • The cache-inhibited fill-buffer bit, CACR[DNFB], is set. • The access is an instruction read. • The access is normal (that is, TT = 0). In this case, a fetched line is stored in the fill buffer and remains valid there; the cache can service additional read accesses from this buffer until another fill occurs or a cache-invalidate-all operation occurs. If ACRn[CM] indicates cache-inhibited mode, precise or imprecise, the controller bypasses the cache and performs an external transfer. If a line in the cache matches the address and the mode is cache-inhibited, the cache does not automatically push the line if it is modified, nor does it invalidate the line if it is valid. Before switching cache mode, execute a CPUSHL instruction or set CACR[DCINVA,ICINVA] to invalidate the entire cache. If ACRn[CM] indicates precise mode, the sequence of read and write accesses to the region is guaranteed to match the instruction sequence. In imprecise mode, the processor core allows read accesses that hit in the cache to occur before completion of a pending write from a previous instruction. Writes are not deferred past data-read accesses that miss the cache (that is, that must be read from the bus). Precise operation forces data-read accesses for an instruction to occur only once by preventing the instruction from being interrupted after data is fetched. Otherwise, if the processor is not in precise mode, MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 7-13 an exception aborts the instruction and the data may be accessed again when the instruction is restarted. These guarantees apply only when ACRn[CM] indicates precise mode and aligned accesses. CPU space-register accesses using the MOVEC instruction are treated as cache-inhibited and precise. 7.9.2 Cache Protocol The following sections describe the cache protocol for processor accesses and assumes that the data is cacheable (that is, write-through or copyback). Note that the discussion of write operations applies to the data cache only. 7.9.2.1 Read Miss A processor read that misses in the cache requests the cache controller to generate a bus transaction. This bus transaction reads the needed line from memory and supplies the required data to the processor core. The line is placed in the cache in the valid state. 7.9.2.2 Write Miss (Data Cache Only) The cache controller handles processor writes that miss in the data cache differently for write-through and copyback regions. Write misses to copyback regions cause the cache line to be read from system memory, as shown in Figure 7-6. 1. Writing character X to 0x0B generates a write miss. Data cannot be written to an invalid line. Cache Line 0x0C 0x08 MCF548x 0x04 0x00 V=0 M=0 X 2. The cache line (characters A–P) is updated from system memory, and the line is marked valid. 0x0C 0x08 0x04 0x00 V=1 ABCD EFGH IJKL MNOP M = 0 System Memory 3. After the cache line is filled, the write that initiated the write miss (the character X) completes to 0x0B. MCF548x 0x0C 0x08 0x04 0x00 V=1 ABCD EXGH IJKL MNOP M =1 Figure 7-6. Write-Miss in Copyback Mode The new cache line is then updated with write data and the M bit is set for the line, leaving it in modified state. Write misses to write-through regions write directly to memory without loading the corresponding cache line into the cache. MCF548x Reference Manual, Rev. 5 7-14 Freescale Semiconductor Cache Operation 7.9.2.3 Read Hit On a read hit, the cache provides the data to the processor core and the cache line state remains unchanged. If the cache mode changes for a specific region of address space, lines in the cache corresponding to that region that contain modified data are not pushed out to memory when a read hit occurs within that line. First execute a CPUSHL instruction or set CACR[DCINVA,ICINVA] before switching the cache mode. 7.9.2.4 Write Hit (Data Cache Only) The cache controller handles processor writes that hit in the data cache differently for write-through and copyback regions. For write hits to a write-through region, portions of cache lines corresponding to the size of the access are updated with the data. The data is also written to external memory. The cache line state is unchanged. For copyback accesses, the cache controller updates the cache line and sets the M bit for the line. An external write is not performed and the cache line state changes to (or remains in) the modified state. 7.9.3 Cache Coherency (Data Cache Only) The MCF548x provides limited cache coherency support in multiple-master environments. Both write-through and copyback memory update techniques are supported to maintain coherency between the cache and memory. The cache does not support snooping (that is, cache coherency is not supported while external or DMA masters are using the bus). Therefore, on-chip DMAs of the MCF548x cannot access local memory and do not maintain coherency with the data cache. 7.9.4 Memory Accesses for Cache Maintenance The cache controller performs all maintenance activities that supply data from the cache to the core, including requests to the SIU for reading new cache lines and writing modified lines to memory. The following sections describe memory accesses resulting from cache fill and push operations. Chapter 17, “FlexBus,” describes required bus cycles in detail. 7.9.4.1 Cache Filling When a new cache line is required, a line read is requested from the SIU, which generates a burst-read transfer by indicating a line access with the size signals, SIZ[1:0]. The responding device supplies 4 consecutive longwords of data. Burst operations can be inhibited or enabled through the burst read/write enable bits (BSTR/BSTW) in the chip-select control registers (CSCR0–CSCR7). SIU line accesses implicitly request burst-mode operations from memory. For more information regarding external bus burst-mode accesses, see Chapter 17, “FlexBus.” The first cycle of a cache-line read loads the longword entry corresponding to the requested address. Subsequent transfers load the remaining longword entries. A burst operation is aborted by an a write-protection fault, which is the only possible access error. Exception processing proceeds immediately. Note that unlike Version 2 and Version 3 access errors, the program counter stored on the exception stack frame points to the faulting instruction. See Section 3.8.2, “Processor Exceptions.” MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 7-15 7.9.4.2 Cache Pushes Cache pushes occur for line replacement and as required for the execution of the CPUSHL instruction. To reduce the requested data’s latency in the new line, the modified line being replaced is temporarily placed in the push buffer while the new line is fetched from memory. After the bus transfer for the new line completes, the modified cache line is written back to memory and the push buffer is invalidated. 7.9.4.2.1 Push and Store Buffers The 16-byte push buffer reduces latency for requested new data on a cache miss by holding a displaced modified data cache line while the new data is read from memory. If a cache miss displaces a modified line, a miss read reference is immediately generated. While waiting for the response, the current contents of the cache location load into the push buffer. When the burst-read bus transaction completes, the cache controller can generate the appropriate line-write bus transaction to write the push buffer contents into memory. In imprecise mode, the FIFO store buffer can defer pending writes to maximize performance. The store buffer can support as many as four entries (16 bytes maximum) for this purpose. Data writes destined for the store buffer cannot stall the core. The store buffer effectively provides a measure of decoupling between the pipeline’s ability to generate writes (one per cycle maximum) and the external bus’s ability to retire those writes. In imprecise mode, writes stall only if the store buffer is full and a write operation is on the internal bus. The internal write cycle is held, stalling the data execution pipeline. If the store buffer is not used (that is, store buffer disabled or cache-inhibited precise mode), external bus cycles are generated directly for each pipeline write operation. The instruction is held in the pipeline until external bus transfer termination is received. Therefore, each write is stalled for 5 cycles, making the minimum write time equal to 6 cycles when the store buffer is not used. See Section 3.2.1.2, “Operand Execution Pipeline (OEP).” The data store buffer enable bit, CACR[DESB], controls the enabling of the data store buffer. This bit can be set and cleared by the MOVEC instruction. DESB is zero at reset and all writes are performed in order (precise mode). ACRn[CM] or CACR[DDCM] generates the mode used when DESB is set. Cacheable write-through and cache-inhibited imprecise modes use the store buffer. The store buffer can queue data as much as 4 bytes wide per entry. Each entry matches the corresponding bus cycle it generates; therefore, a misaligned longword write to a write-through region creates two entries if the address is to an odd-word boundary. It creates three entries if it is to an odd-byte boundary—one per bus cycle. 7.9.4.2.2 Push and Store Buffer Bus Operation As soon as the push or store buffer has valid data, the internal bus controller uses the next available external bus cycle to generate the appropriate write cycles. In the event that another cache fill is required (for example, cache miss to process) during the continued instruction execution by the processor pipeline, the pipeline stalls until the push and store buffers are empty, then generate the required external bus transaction. Supervisor instructions, the NOP instruction, and exception processing synchronize the processor core and guarantee the push and store buffers are empty before proceeding. Note that the NOP instruction should be used only to synchronize the pipeline. The preferred no-operation function is the TPF instruction. See the ColdFire Programmer’s Reference Manual for more information on the TPF instruction. MCF548x Reference Manual, Rev. 5 7-16 Freescale Semiconductor Cache Operation 7.9.5 Cache Locking Ways 0 and 1 of the data cache can be locked by setting CACR[DHLCK]; likewise, ways 0 and 1 of the instruction cache can be locked by setting CACR[IHLCK]. If a cache is locked, cache lines in ways 0 and 1 are not subject to being deallocated by normal cache operations. As Figure 7-7 (B and C) shows, the algorithm for updating the cache and for identifying cache lines to be deallocated is otherwise unchanged. If ways 2 and 3 are entirely invalid, cacheable accesses are first allocated in way 2. Way 3 is not used until the location in way 2 is occupied. Ways 0 and 1 are still updated on write hits (D in Figure 7-7) and may be pushed or cleared only explicitly by using specific cache push/invalidate instructions. However, new cache lines cannot be allocated in ways 0 and 1. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 7-17 Invalid (V = 0) Valid, not modified (V = 1, M = 0) Valid, modified (V = 1, M = 1) A: Ways 0 and 1 are filled. Ways 2 and 3 are invalid. B: CACR[DHLCK] is set, locking ways 0 and 1. C: When a set in Way 2 is D: Write hits to ways 0 occupied, the set in way 3 and 1 update cache is used for a cacheable lines. access. Way 0Way 1Way 2Way 3 Way 0Way 1Way 2Way 3 Way 0Way 1Way 2Way 3 Way 0Way 1Way 2Way 3 After CACR[DHLCK] is set, subsequent cache accesses go to ways 2 and 3. While the cache is locked and after a position in ways is full, the set in Way 3 is updated. While the cache is locked, ways 0 and 1 can be updated by write hits. In this example, memory is configured as copyback, so updated cache lines are marked modified. Set 0 Set 511 After reset, the cache is invalidated, ways 0 and 1 are then written with data that should not be deallocated. Ways 0 and 1 can be filled systematically by using the INTOUCH instruction. Figure 7-7. Data Cache Locking MCF548x Reference Manual, Rev. 5 7-18 Freescale Semiconductor Cache Register Definition 7.10 Cache Register Definition This section describes the MCF548x implementation of the Version 4e cache registers. 7.10.1 Cache Control Register (CACR) The CACR in Figure 7-8 contains bits for configuring the cache. It can be written by the MOVEC register instruction and can be read or written from the debug facility. A hardware reset clears CACR, which disables the cache; however, reset does not affect the tags, state information, or data in the cache. NOTE CACR is read/write by the debug module. 31 30 29 28 27 26 R DEC DW DESB DDPI DHLCK 25 DDCM 24 23 DCINVA DDSP 22 21 20 0 0 0 19 18 BEC BCINVA 17 16 0 0 W Reset 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 ICINVA IDSP 0 0 0 0 0 0 0 0 0 0 0 0 0 R IEC 0 DNFB IDPI IHLCK IDCM EUSP DF W Reset 0 0 Reg Addr 0 0 0 0 0 0x002 Figure 7-8. Cache Control Register (CACR) Table 7-4 describes CACR fields. Note that some implementations may include fields not defined here; consult the part-specific documentation. Table 7-4. CACR Field Descriptions Bits Name Description 31 DEC Enable data cache. 0 Cache disabled. The data cache is not operational, but data and tags are preserved. 1 Cache enabled. 30 DW Data default write-protect. For normal operations that do not hit in the RAMBARs or ACRs, this field defines write-protection. See Section 7.9.1, “Caching Modes.” 0 Not write protected. 1 Write protected. Write operations cause an access error exception. 29 DESB Enable data store buffer. Affects the precision of transfers. 0 Imprecise-mode, write-through or cache-inhibited writes bypass the store buffer and generate bus cycles directly. Section 7.9.4.2.1, “Push and Store Buffers,” describes the associated performance penalty. 1 The four-entry FIFO store buffer is enabled; to maximize performance, this buffer defers pending imprecise-mode, write-through or cache-inhibited writes. Precise-mode, cache-inhibited accesses always bypass the store buffer. Precise and imprecise modes are described in Section 7.9.1.2, “Cache-Inhibited Accesses.” MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 7-19 Table 7-4. CACR Field Descriptions (Continued) Bits Name Description 28 DDPI Disable CPUSHL invalidation. 0 Normal operation. A CPUSHL instruction causes the selected line to be pushed if modified, then invalidated. 1 No clear operation. A CPUSHL instruction causes the selected line to be pushed if modified, then left valid. 27 DHLCK Half-data cache lock mode 0 Normal operation. The cache allocates the lowest invalid way. If all ways are valid, the cache allocates the way pointed at by the counter and then increments this counter. 1 Half-cache operation. The cache allocates to the lower invalid way of levels 2 and 3; if both are valid, the cache allocates to Way 2 if the high-order bit of the round-robin counter is zero; otherwise, it allocates Way 3 and increments the round-robin counter. This locks the contents of ways 0 and 1. Ways 0 and 1 are still updated on write hits and may be pushed or cleared by specific cache push/invalidate instructions. 26–25 DDCM Default data cache mode. For normal operations that do not hit in the RAMBARs, ROMBARs, or ACRs, this field defines the effective cache mode. 00 Cacheable write-through imprecise 01 Cacheable copyback 10 Cache-inhibited precise 11 Cache-inhibited imprecise Precise and imprecise accesses are described in Section 7.9.1.2, “Cache-Inhibited Accesses.” 24 DCINVA Data cache invalidate all. Writing a 1 to this bit initiates entire cache invalidation. Once invalidation is complete, this bit automatically returns to 0; it is not necessary to clear it explicitly. Note the caches are not cleared on power-up or normal reset, as shown in Figure 7-4. 0 No invalidation is performed. 1 Initiate invalidation of the entire data cache. The cache controller sequentially clears V and M bits in all sets. Subsequent data accesses stall until the invalidation is finished, at which point, this bit is automatically cleared. In copyback mode, the cache should be flushed using a CPUSHL instruction before setting this bit. 23 DDSP Data default supervisor-protect. For normal operations that do not hit in the RAMBAR, ROMBAR, or ACRs, this field defines supervisor-protection 0 Not supervisor protected 1 Supervisor protected. User operations cause a fault 22–20 — 19 BEC 18 BCINVA 17–16 — 15 IEC 14 — Reserved, should be cleared. Enable branch cache. 0 Branch cache disabled. This may be useful if code is unlikely to be reused. 1 Branch cache enabled. Branch cache invalidate. Invalidation occurs when this bit is written as a 1. Note that branch caches are not cleared on power-up or normal reset. 0 No invalidation is performed. 1 Initiate an invalidation of the entire branch cache. Reserved, should be cleared. Enable instruction cache 0 Instruction cache disabled. All instructions and tags in the cache are preserved. 1 Instruction cache enabled. Reserved, should be cleared. MCF548x Reference Manual, Rev. 5 7-20 Freescale Semiconductor Cache Register Definition Table 7-4. CACR Field Descriptions (Continued) Bits Name Description 13 DNFB Default cache-inhibited fill buffer 0 Fill buffer does not store cache-inhibited instruction accesses (16 or 32 bits). 1 Fill buffer can store cache-inhibited accesses. The buffer is used only for normal (TT = 0) instruction reads of a cache-inhibited region. Instructions are loaded into the buffer by a burst access (line fill). They stay in the buffer until they are displaced; subsequent accesses may not appear on the external bus. Setting DNFB can cause a coherency problem for self-modifying code. If a cache-inhibited access uses the buffer while DNFB = 1, instructions remain valid in the buffer until a cache-invalidate-all instruction, another cache-inhibited burst, or a miss that initiates a fill. A write to the line in the fill goes to the external bus without updating or invalidating the buffer. Subsequent reads of that written data are serviced by the fill buffer and receive stale information. Note: Freescale discourages the use of self-modifying code. 12 IDPI 11 IHLCK Instruction cache half-lock. 0 Normal operation. The cache allocates to the lowest invalid way; if all ways are valid, the cache allocates to the way pointed at by the round-robin counter and then increments this counter. 1 Half cache operation. The cache allocates to the lowest invalid way of ways 2 and 3; if both of these ways are valid, the cache allocates to way 2 if the high-order bit of the round-robin counter is zero; otherwise, it allocates way 3 and then increments the round-robin counter. This locks the contents of ways 0 and 1. Ways 0 and 1 are still updated on write hits and may be pushed or cleared by specific cache push/invalidate instructions. 10 IDCM Instruction default cache mode. For normal operations that do not hit in the RAMBARs or ACRs, this field defines the effective cache mode. 0 Cacheable 1 Cache-inhibited 9 — 8 ICINVA Instruction cache invalidate. Invalidation occurs when this bit is written as a 1. Note the caches are not cleared on power-up or normal reset. 0 No invalidation is performed. 1 Initiate invalidation of instruction cache. The cache controller sequentially clears all V bits. Subsequent local memory bus accesses stall until invalidation completes, at which point ICINVA is cleared automatically without software intervention. For copyback mode, use CPUSHL before setting ICINVA. 7 IDSP Default instruction supervisor protection bit. For normal operations that do not hit in the RAMBAR, ROMBAR, or ACRs, this field defines supervisor-protection. 0 Not supervisor protected 1 Supervisor protected. User operations cause a fault 6 — 5 EUSP 4 DF Disable FPU. Determines whether the FPU is enabled. See Section 6.1.1, “Overview.” 0 FPU enabled. 1 FPU disabled 3–0 — Reserved, should be cleared. Instruction CPUSHL invalidate disable. 0 Normal operation. A CPUSHL instruction causes the selected line to be invalidated. 1 No clear operation. A CPUSHL instruction causes the selected line to be left valid. Reserved, should be cleared. Reserved, should be cleared. Enable USP. Enables the use of the user stack pointer. 0 USP disabled. Core uses a single stack pointer. 1 USP enabled. Core uses separate supervisor and user stack pointers. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 7-21 7.10.2 Access Control Registers (ACR0–ACR3) The ACRs, Figure 7-9, assign control attributes, such as cache mode and write protection, to specified memory regions. ACR0 and ACR1 control data attributes; ACR2 and ACR3 control instruction attributes. Registers are accessed with the MOVEC instruction with the Rc encodings in Figure 7-9. For overlapping data regions, ACR0 takes priority; ACR2 takes priority for overlapping instruction regions. Data transfers to and from these registers are longword transfers. NOTE The MBAR region should be mapped as cache-inhibited through an ACR or the CACR. NOTE ACR0–ACR3 is read/write by the debug module. 31 30 29 28 R 27 26 25 24 23 22 21 BA 20 19 18 17 16 ADMSK W Reset R 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 0 AMM 0 0 0 0 SP W1 0 0 0 0 0 0 0 0 0 0 0 0 0 E S CM W Reset 0 0 Reg Addr 0 0 0 ACR0: 0x004; ACR1: 0x005; ACR2: 0x006; ACR3: 0x007 1 Reserved in ACR2 and ACR3. Figure 7-9. Access Control Register Format (ACRn) Table 7-5 describes ACRn fields. I Table 7-5. ACRn Field Descriptions Bits Name Description 31–24 BA Base address. Compared with address bits A[31:24]. Eligible addresses that match are assigned the access control attributes of this register. 23–16 ADMSK Address mask. Setting a mask bit causes the corresponding address base bit to be ignored. The low-order mask bits can be set to define contiguous regions larger than 16 Mbytes. The mask can define multiple noncontiguous regions of memory. 15 E Enable. Enables or disables the other ACRn bits. 0 Access control attributes disabled 1 Access control attributes enabled MCF548x Reference Manual, Rev. 5 7-22 Freescale Semiconductor Cache Management Table 7-5. ACRn Field Descriptions (Continued) Bits Name Description 14–13 S Supervisor mode. Specifies whether only user or supervisor accesses are allowed in this address range or if the type of access is a don’t care. 00 Match addresses only in user mode 01 Match addresses only in supervisor mode 1x Execute cache matching on all accesses 12–11 — Reserved, should be cleared. 10 AMM 9–7 — Reserved; should be cleared. 6–5 CM Cache mode. Selects the cache mode and access precision. Precise and imprecise modes are described in Section 7.9.1.2, “Cache-Inhibited Accesses.” 00 Cacheable, write-through 01 Cacheable, copyback 10 Cache-inhibited, precise 11 Cache-inhibited, imprecise 4 — Reserved, should be cleared. 3 SP Supervisor protect. 0 Indicates supervisor and user mode access allowed, reset value is 0 1 Indicates only supervisor access is allowed to this address space and attempted user mode accesses generate an access error exception 2 W ACR0/ACR1 only. Write protect. Selects the write privilege of the memory region. ACR2[W] and ACR3[W] are reserved. 0 Read and write accesses permitted 1 Write accesses not permitted 1–0 — Reserved, should be cleared. 7.11 Address mask mode. 0 The ACR hit function allows control of a 16 Mbytes or greater memory region. 1 The upper 8 bits of the address and ACR are compared without a mask function. Address bits [23:20] of the address and ACR are compared using ACR[19:16] as a mask, allowing control of a 1–16 Mbyte memory region. Cache Management The cache can be enabled and configured by using a MOVEC instruction to access CACR. A hardware reset clears CACR, disabling the cache and removing all configuration information; however, reset does not affect the tags, state information, and data in the cache. Set CACR[DCINVA,ICINVA] to invalidate the caches before enabling them. The privileged CPUSHL instruction supports cache management by selectively pushing and invalidating cache lines. The address register used with CPUSHL directly addresses the cache’s directory array. The CPUSHL instruction flushes a cache line. The value of CACR[DDPI,IDPI] determines whether CPUSHL invalidates a cache line after it is pushed. To push the entire cache, implement a software loop to index through all sets and through each of the four lines within each set (a total of 512 lines for the data cache and 1024 lines for the instruction cache). The state of CACR[DEC,IEC] does not affect the operation of CPUSHL or CACR[DCINVA,ICINVA]. Disabling a cache by setting CACR[IEC] or CACR[DEC] makes the cache nonoperational without affecting tags, state information, or contents. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 7-23 The contents of An used with CPUSHL specify cache row and line indexes. This differs from the 68K family where a physical address is specified. Figure 7-11 shows the An format for the data cache. The contents of An used with CPUSHL specify cache row and line indexes. Figure 7-10 shows the An format for the data cache. 31 13 12 0 4 3 0 Set Index Way Index Figure 7-10. An Format (Data Cache) Figure 7-11 shows the An format for the instruction cache. 31 13 12 0 4 3 0 Set Index Way Index Figure 7-11. An Format (Instruction Cache) The following code example flushes the entire data cache: _cache_disable: nop move.w jsr clr.l movec movec move.l movec rts #0x2700,SR _cache_flush d0 d0,ACR0 d0,ACR1 #0x01000000,d0 d0,CACR ;mask off IRQs ;flush the cache completely ;ACR0 off ;ACR1 off ;Invalidate and disable cache _cache_flush: nop moveq.l moveq.l move.l #0,d0 #0,d1 d0,a0 ;synchronize—flush store buffer ;initialize way counter ;initialize set counter ;initialize cpushl pointer cpushl add.l addq.l cmpi.l bne dc,(a0) #0x0010,a0 #1,d1 #511,d1 setloop ;push cache line a0 ;increment set index by 1 ;increment set counter ;are sets for this way done? moveq.l addq.l move.l cmpi.l bne rts #0,d1 #1,d0 d0,a0 #4,d0 setloop ;set counter to zero again ;increment to next way ;set = 0, way = d0 ;flushed all the ways? setloop: The following CACR loads assume the instruction cache has been invalidated, the default instruction cache mode is cacheable, and the default data cache mode is copyback. MCF548x Reference Manual, Rev. 5 7-24 Freescale Semiconductor Cache Management dataCacheLoadAndLock: move.l movec #0xa3080800,d0; enable and invalidate data cache ... d0,cacr ; ... in the CACR The following code preloads half of the data cache (16 Kbytes). It assumes a contiguous block of data is to be mapped into the data cache, starting at a 0-modulo-16K address. move.l lea dataCacheLoop: tst.b lea subq.l bne.b #1024,d0 ;256 16-byte lines in 16K space data_,a0 ; load pointer defining data area (a0) ;touch location + load into data cache 16(a0),a0;increment address to next line #1,d0 ;decrement loop counter dataCacheLoop;if done, then exit, else continue ; A 16K region has been loaded into ways 0 and 1 of the 32K data cache. lock it! move.l movec rts #0xaa088000,d0;set the data cache lock bit ... d0,cacr ; ... in the CACR align 16 The following CACR loads assume the data cache has been invalidated, the default instruction cache mode is cacheable and the default operand cache mode is copyback. Note that this function must be mapped into a cache inhibited or SRAM space, or these text lines will be prefetched into the instruction cache, possibly displacing some of the 8-Kbyte space being explicitly fetched. instructionCacheLoadAndLock: move.l movec #0xa2088100,d0;enable and invalidate the instruction d0,cacr ;cache in the CACR The following code segments preload half of the instruction cache (8 Kbytes). It assumes a contiguous block of data is to be mapped, starting at a 0-modulo-8K address move.l #512,d0 ;512 16-byte lines in 8K space lea code_,a0 ;load pointer defining code area instCacheLoop: intouch (a0) ;touch location + load into instruction cache ; Note in the assembler we use, there is no INTOUCH opcode. The following ; is used to produce the required binary representation cpushl #nc,(a0) ;touch location + load into ;instruction cache lea 16(a0),a0;increment address to next line subq.l #1,d0 ;decrement loop counter bne.b instCacheLoop;if done, then exit, else continue ; A 8K region was loaded into levels 0 and 1 of the 16-Kbyte instruction cache. ; lock it! move.l movec rts #0xa2088800,d0;set the instruction cache lock bit d0,cacr ;in the CACR MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 7-25 7.12 Cache Operation Summary This section gives operational details for the cache and presents instruction and data cache-line state diagrams. 7.12.1 Instruction Cache State Transitions Because the instruction cache does not support writes, it supports fewer operations than the data cache. As Figure 7-12 shows, an instruction cache line can be in one of two states, valid or invalid. Modified state is not supported. Transitions are labeled with a capital letter indicating the previous state and a number indicating the specific case listed in Table 7-6. These numbers correspond to the equivalent operations on data caches, described in Section 7.12.2, “Data Cache State Transitions.” II5—ICINVA II6—CPUSHL & IDPI II7—CPUSHL & IDPI IV1—CPU read miss IV2—CPU read hit IV7—CPUSHL & IDPI II1—CPU read miss Invalid V=0 Valid V=1 IV5—ICINVA IV6—CPUSHL & IDPI Figure 7-12. Instruction Cache Line State Diagram Table 7-6 describes the instruction cache state transitions shown in Figure 7-12. Table 7-6. Instruction Cache Line State Transitions Current State Access Invalid (V = 0) Valid (V = 1) Read miss II1 Read line from memory and update cache; IV1 Read new line from memory and update cache; supply data to processor; supply data to processor; stay in valid state. go to valid state. Read hit II2 Not possible IV2 Supply data to processor; stay in valid state. Write miss II3 Not possible IV3 Not possible Write hit II4 Not possible IV4 Not possible Cache invalidate II5 No action; stay in invalid state. IV5 No action; go to invalid state. Cache push II6, No action; II7 stay in invalid state. IV6 No action; go to invalid state. IV7 No action; stay in valid state. MCF548x Reference Manual, Rev. 5 7-26 Freescale Semiconductor Cache Operation Summary 7.12.2 Data Cache State Transitions Using the V and M bits, the data cache supports a line-based protocol allowing individual cache lines to be invalid, valid, or modified. To maintain memory coherency, the data cache supports both write-through and copyback modes, specified by the corresponding ACR[CM], or CACR[DDCM] if no ACR matches. Read or write misses to copyback regions cause the cache controller to read a cache line from memory into the cache. If available, tag and data from memory update an invalid line in the selected set. The line state then changes from invalid to valid by setting the V bit. If all lines in the row are already valid or modified, the pseudo-round-robin replacement algorithm selects one of the four lines and replaces the tag and data. Before replacement, modified lines are temporarily buffered and later copied back to memory after the new line has been read from memory. Figure 7-13 shows the three possible data cache line states and possible processor-initiated transitions for memory configured as copyback. Transitions are labeled with a capital letter indicating the previous state and a number indicating the specific case; see Table 7-7. CI5—DCINVA CI6—CPUSHL & DDPI CI7—CPUSHL & DDPI CV1—CPU read miss CV2—CPU read hit CV7—CPUSHL & DDPI CI1—CPU read miss Invalid V=0 Valid V=1 M=0 CV5—DCINVA CV6—CPUSHL & DDPI CI3—CPU write miss CD1—CPU read miss CD7—CPUSHL CD5—DCINVA & DDPI CV3—CPU write miss CD6—CPUSHL & DDPI CV4—CPU write hit Modified V=1 M=1 CD2—CPU read hit CD3—CPU write miss CD4—CPU write hit Figure 7-13. Data Cache Line State Diagram—Copyback Mode Figure 7-14 shows the two possible states for a cache line in write-through mode. WV1—CPU read miss WV2—CPU read hit WV3—CPU write miss WV4—CPU write hit WV7—CPUSHL & DDPI WI3—CPU write miss WI5—DCINVA WI6—CPUSHL & DDPI WI1—CPU read miss Invalid V=0 Valid V=1 WV5—DCINVA WV6—CPUSHL & DDPI Figure 7-14. Data Cache Line State Diagram—Write-Through Mode Table 7-7 describes data cache line transitions and the accesses that cause them. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 7-27 Table 7-7. Data Cache Line State Transitions Current State Access Invalid (V = 0) Valid (V = 1, M = 0) Modified (V = 1, M = 1) Read miss (C,W)I1 Read line from memory and update cache; supply data to processor; go to valid state. (C,W)V1 Read new line from memory and update cache; supply data to processor; stay in valid state. CD1 Push modified line to buffer; read new line from memory and update cache; supply data to processor; write push buffer contents to memory; go to valid state. Read hit (C,W)I2 Not possible. (C,W)V2 Supply data to processor; stay in valid state. CD2 Supply data to processor; stay in modified state. Write miss (copyback) CI3 Read line from memory and update cache; write data to cache; go to modified state. CV3 Read new line from memory and update cache; write data to cache; go to modified state. CD3 Push modified line to buffer; read new line from memory and update cache; write push buffer contents to memory; stay in modified state. Write miss (writethrough) WI3 Write data to memory; WV3 stay in invalid state. Write data to memory; stay in valid state. WD3 Write data to memory; stay in modified state. Cache mode changed for the region corresponding to this line. To avoid this state, execute a CPUSHL instruction or set CACR[DCINVA,ICINVA] before switching modes. Write hit (copyback) CI4 Not possible. CV4 Write data to cache; go to modified state. CD4 Write data to cache; stay in modified state. Write hit (writethrough) WI4 Not possible. WV4 Write data to memory and to cache; stay in valid state. WD4 Write data to memory and to cache; go to valid state. Cache mode changed for the region corresponding to this line. To avoid this state, execute a CPUSHL instruction or set CACR[DCINVA,ICINVA] before switching modes. Cache (C,W)I5 No action; invalidate stay in invalid state. (C,W)V5 No action; go to invalid state. CD5 No action (modified data lost); go to invalid state. Cache push (C,W)V6 No action; go to invalid state. CD6 Push modified line to memory; go to invalid state. (C,W)V7 No action; stay in valid state. CD7 Push modified line to memory; go to valid state. (C,W)I6 No action; (C,W)I7 stay in invalid state. The following tables present the same information as Table 7-7, organized by the current state of the cache line. In Table 7-8 the current state is invalid. MCF548x Reference Manual, Rev. 5 7-28 Freescale Semiconductor Cache Operation Summary Table 7-8. Data Cache Line State Transitions (Current State Invalid) Access Response Read miss (C,W)I1 Read line from memory and update cache; supply data to processor; go to valid state. Read hit (C,W)I2 Not possible Write miss (copyback) CI3 Read line from memory and update cache; write data to cache; go to modified state. Write miss (write-through) WI3 Write data to memory; stay in invalid state. Write hit (copyback) CI4 Not possible Write hit (write-through) WI4 Not possible Cache invalidate (C,W)I5 No action; stay in invalid state. Cache push (C,W)I6 No action; stay in invalid state. Cache push (C,W)I7 No action; stay in invalid state. In Table 7-9 the current state is valid. Table 7-9. Data Cache Line State Transitions (Current State Valid) Access Response Read miss (C,W)V1 Read new line from memory and update cache; supply data to processor; stay in valid state. Read hit (C,W)V2 Supply data to processor; stay in valid state. Write miss (copyback) CV3 Read new line from memory and update cache; write data to cache; go to modified state. Write miss (write-through) WV3 Write data to memory; stay in valid state. Write hit (copyback) CV4 Write data to cache; go to modified state. Write hit (write-through) WV4 Write data to memory and to cache; stay in valid state. Cache invalidate (C,W)V5 No action; go to invalid state. Cache push (C,W)V6 No action; go to invalid state. Cache push (C,W)V7 No action; stay in valid state. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 7-29 In Table 7-10 the current state is modified. Table 7-10. Data Cache Line State Transitions (Current State Modified) Access Response Read miss CD1 Push modified line to buffer; read new line from memory and update cache; supply data to processor; write push buffer contents to memory; go to valid state. Read hit CD2 Supply data to processor; stay in modified state. Write miss (copyback) CD3 Push modified line to buffer; read new line from memory and update cache; write push buffer contents to memory; stay in modified state. Write miss (write-through) WD3 Write data to memory; stay in modified state. Cache mode changed for the region corresponding to this line. To avoid this state, execute a CPUSHL instruction or set CACR[DCINVA,ICINVA] before switching modes. Write hit (copyback) CD4 Write data to cache; stay in modified state. Write hit (write-through) WD4 Write data to memory and to cache; go to valid state. Cache mode changed for the region corresponding to this line. To avoid this state, execute a CPUSHL instruction or set CACR[DCINVA,ICINVA] before switching modes. Cache invalidate CD5 No action (modified data lost); go to invalid state. Cache push CD6 Push modified line to memory; go to invalid state. Cache push CD7 Push modified line to memory; go to valid state. 7.13 Cache Initialization Code The following example sets up the cache for FLASH or ROM space only. move.l #0xA70C8100,D0 movec D0, CACR move.l #0xFF00C000,D0 movec D0,ACR0 //enable cache, invalidate it, //default mode is cache-inhibited imprecise //cache FLASH space, enable, //ignore FC2, cacheable, writethrough MCF548x Reference Manual, Rev. 5 7-30 Freescale Semiconductor Chapter 8 Debug Support 8.1 Introduction This chapter describes the Revision D enhanced hardware debug support in the ColdFire Version 4. This revision of the ColdFire debug architecture encompasses earlier revisions. An expanded set of debug functionality is defined as Revision B (or Rev. B). The further enhanced debug architecture implemented in the Version 4 ColdFire is known as Revision C (or Rev. C). The addition of the memory management unit (MMU) in the Version 4e ColdFire requires corresponding enhancements to the ColdFire debug functionality, resulting in Revision D. 8.1.1 Overview The debug module interface is shown in Figure 8-1. High-speed local bus ColdFire CPU Core Debug Module Trace Port Control PSTDDATA[7:0] BKPT PSTCLK Communication Port DSCLK, DSI, DSO Figure 8-1. Processor/Debug Module Interface Debug support is divided into three areas: • Real-time trace support: The ability to determine the dynamic execution path through an application is fundamental for debugging. The ColdFire solution implements an 8-bit parallel output bus that reports processor execution status and data to an external BDM emulator system. See Section 8.3, “Real-Time Trace Support.” • Background debug mode (BDM): Provides low-level debugging in the ColdFire processor complex. In BDM, the processor complex is halted and a variety of commands can be sent to the processor to access memory and registers. The external BDM emulator uses a three-pin, serial, full-duplex channel. See Section 8.5, “Background Debug Mode (BDM),” and Section 8.4, “Memory Map/Register Definition.” • Real-time debug support: BDM requires the processor to be halted, which many real-time embedded applications cannot do. Debug interrupts let real-time systems execute a unique service routine that can quickly save key register and variable contents and return the system to normal operation without halting. External development systems can access saved data, because the hardware supports concurrent operation of the processor and BDM-initiated commands. In addition, the option is provided to allow interrupts to occur. See Section 8.6, “Real-Time Debug Support.” The Version 2 ColdFire core implemented the original debug architecture, now called Revision A. Based on feedback from customers and third-party developers, enhancements have been added to succeeding MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 8-1 generations of ColdFire cores. For Revision A, CSR[HRL] is 0. See Section 8.4.2, “Configuration/Status Register (CSR).” The Version 3 core implements Revision B of the debug architecture, offering more flexibility for configuring the hardware breakpoint trigger registers and removing the restrictions involving concurrent BDM processing while hardware breakpoint registers are active. For Revision B, CSR[HRL] is 1. Revision C of the debug architecture more than doubles the on-chip breakpoint registers and provides an ability to interrupt debug service routines. For Revision C, CSR[HRL] is 2. Differences between Revision B and C are summarized as follows: • Debug Revision B has separate PST[3:0] and DDATA[3:0] signals. • Debug Revision C adds breakpoint registers and supports normal interrupt request service during debug. It combines debug signals into PSTDDATA[7:0]. The addition of the memory management unit (MMU) to the baseline architecture requires corresponding enhancements to the ColdFire debug functionality, resulting in Revision D. For Revision D, the revision level bit, CSR[HRL], is 3. With software support, the MMU can provide a demand-paged, virtual address environment. To support debugging in this virtual environment, the debug enhancements are primarily related to the expansion of the virtual address to include the 8-bit address space identifier (ASID). Conceptually, the virtual address is expanded to a 40-bit value: the 8-bit ASID plus the 32-bit address. The expansion of the virtual address affects two major debug functions: • The ASID is optionally included in the specification of the hardware breakpoint registers. As an example, the four PC breakpoint registers are each expanded by 8 bits, so that a specific ASID value may be programmed as part of the breakpoint instruction address. Likewise, each operand address/data breakpoint register is expanded to include an ASID value. Finally, new control registers define if and how the ASID is to be included in the breakpoint comparison trigger logic. • The debug module implements the concept of ownership trace in which the ASID value may be optionally displayed as part of the real-time trace functionality. When enabled, real-time trace displays instruction addresses on every change-of-flow instruction that is not absolute or PC-relative. For Rev. D, this instruction address display optionally includes the contents of the ASID, thus providing the complete instruction virtual address on these instructions. Additionally when a Sync_PC serial BDM command is loaded from the external development system, the processor optionally displays the complete virtual instruction address, including the 8-bit ASID value. In addition to these ASID-related changes, the new MMU control registers are accessible by using serial BDM commands. The same BDM access capabilities are also provided for the EMAC and FPU programming models. Finally, a new serial BDM command is implemented (FORCE_TA) to assist debugging when a software error generates an incorrect memory address that hangs the external bus. The new BDM command attempts to break this condition by forcing a bus termination. 8.2 Signal Descriptions Table 8-1 describes debug module signals. All ColdFire debug signals are unidirectional and related to a rising edge of the processor core’s clock signal. The standard 26-pin debug connector is shown in Section 8.9, “Freescale-Recommended BDM Pinout.” MCF548x Reference Manual, Rev. 5 8-2 Freescale Semiconductor Signal Descriptions Table 8-1. Debug Module Signals Signal Description DSCLK Development Serial Clock-Internally synchronized input. (The logic level on DSCLK is validated if it has the same value on two consecutive rising bus clock edges.) Clocks the serial communication port to the debug module during packet transfers. Maximum frequency is PSTCLK/5. At the synchronized rising edge of DSCLK, the data input on DSI is sampled and DSO changes state. DSI Development Serial Input -Internally synchronized input that provides data input for the serial communication port to the debug module, once the DSCLK has been seen as high (logic 1). DSO Development Serial Output -Provides serial output communication for debug module responses. DSO is registered internally. The output is delayed from the validation of DSCLK high. BKPT Breakpoint - Input used to request a manual breakpoint. Assertion of BKPT puts the processor into a halted state after the current instruction completes. Halt status is reflected on processor status/debug data signals (PSTDDATA[7:0]) as the value 0xF. If CSR[BKD] is set (disabling normal BKPT functionality), asserting BKPT generates a debug interrupt exception in the processor. PSTCLK Processor Status Clock - Half-speed version of the processor clock. Its rising edge appears in the center of the two-processor-cycle window of valid PSTDDATA output. See Figure 8-2. PSTCLK indicates when the development system should sample PSTDDATA values. If real-time trace is not used, setting CSR[PCD] keeps PSTCLK and PSTDDATA outputs from toggling without disabling triggers. Non-quiescent operation can be reenabled by clearing CSR[PCD], although the external development systems must resynchronize with the PSTDDATA output. PSTCLK starts clocking only when the first non-zero PST value (0xC, 0xD, or 0xF) occurs during system reset exception processing. Table 8-4 describes PST values. PSTDDATA[7:0] Processor Status/Debug Data - These outputs, which change on the negative edge of PSTCLK, indicate both processor status and captured address and data values and are discussed more thoroughly in Section 8.2.1, “Processor Status/Debug Data (PSTDDATA[7:0]).” Figure 8-2 shows PSTCLK timing with respect to PSTDDATA. PSTCLK STDDATA Figure 8-2. PSTCLK Timing 8.2.1 Processor Status/Debug Data (PSTDDATA[7:0]) Processor status data outputs are used to indicate both processor status and captured address and data values. They operate at half the processor’s frequency. Given that real-time trace information appears as a sequence of 4-bit data values, there are no alignment restrictions; that is, the processor status (PST) values and operands may appear on either nibble of PSTDDATA[7:0]. The upper nibble (PSTDDATA[7:4]) is the more significant and yields values first. CSR controls capturing of data values to be presented on PSTDDATA. Executing the WDDATA instruction captures data that is displayed on PSTDDATA too. These signals are updated each processor cycle and display two values at a time for two processor clock cycles. Table 8-2 shows the PSTDDATA MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 8-3 output for the processor’s sequential execution of single-cycle instructions (A, B, C, D...). Cycle counts are shown relative to processor frequency. These outputs indicate the current processor pipeline status and are not related to the current bus transfer. Table 8-2. PSTDDATA: Sequential Execution of Single-Cycle Instructions Cycles PSTDDATA[7:0] T+0, T+1 {PST for A, PST for B} T+2, T+3 {PST for C, PST for D} T+4, T+5 {PST for E, PST for F} The signal timing for the example in Table 8-2 is shown in Figure 8-3. T+0 T+1 T+2 T+3 T+4 T+5 T+6 Processor Clock PSTCLK {A, B} PSTDDATA {C, D} {E, F} Figure 8-3. PSTDDATA: Single-Cycle Instruction Timing Table 8-3 shows the case where a PSTDDATA module captures a memory operand on a simple load instruction: mov.l <mem>,Rx. Table 8-3. PSTDDATA: Data Operand Captured Cycle T PSTDDATA[7:0] {PST for mov.l, PST marker for captured operand) = {0x1, 0xB} T+1 {0x1, 0xB} T+2 {Operand[3:0], Operand[7:4]} T+3 {Operand[3:0], Operand[7:4]} T+4 {Operand[11:8], Operand[15:12]} T+5 {Operand[11:8], Operand[15:12]} T+6 {Operand[19:16], Operand[23:20]} T+7 {Operand[19:16], Operand[23:20]} T+8 {Operand[27:24], Operand[31:28]} T+9 {Operand[27:24], Operand[31:28]} T+10 (PST for next instruction) T+11 (PST for next instruction,...) MCF548x Reference Manual, Rev. 5 8-4 Freescale Semiconductor Real-Time Trace Support NOTE A PST marker and its data display are sent contiguously. Except for this transmission, the IDLE status (0x0) can appear anytime. Again, given that real-time trace information appears as a sequence of 4-bit values, there are no alignment restrictions. That is, PST values and operands may appear on either nibble of PSTDDATA. 8.3 Real-Time Trace Support Real-time trace, which defines the dynamic execution path, is a fundamental debug function. The ColdFire solution is to include a parallel output port providing encoded processor status and data to an external development system. This 8-bit port is partitioned into two consecutive 4-bit nibbles. Each nibble can either transmit information concerning the processor’s execution status (PST) or debug data (DDATA). The processor status may not be related to the current bus transfer, due to the decoupling FIFOs. External development systems can use PSTDDATA outputs with an external image of the program to completely track the dynamic execution path. This tracking is complicated by any change in flow, especially when branch target address calculation is based on the contents of a program-visible register (variant addressing). PSTDDATA outputs can be configured to display the target address of such instructions in sequential nibble increments across multiple processor clock cycles, as described in Section 8.3.1, “Begin Execution of Taken Branch (PST = 0x5).” Four 32-bit storage elements form a FIFO buffer connecting the processor’s high-speed local bus to the external development system through PSTDDATA[7:0]. The buffer captures branch target addresses and certain data values for eventual display on the PSTDDATA port, two nibbles at a time starting with the least significant bit (lsb). Execution speed is affected only when three storage elements contain valid data to be dumped to the PSTDDATA port. This occurs only when two values are captured simultaneously in a read-modify-write operation. The core stalls until two FIFO entries are available. Table 8-4 shows the encoding of these signals. Table 8-4. Processor Status Encoding PST[3:0] Definition Hex Binary 0x0 0000 Continue execution. Many instructions execute in one processor cycle. If an instruction requires more clock cycles, subsequent clock cycles are indicated by driving PSTDDATA outputs with this encoding. 0x1 0001 Begin execution of one instruction. For most instructions, this encoding signals the first clock cycle of an instruction’s execution. Certain change-of-flow opcodes, plus the PULSE and WDDATA instructions, generate different encodings. 0x2 0010 Begin execution of two instructions. For superscalar instruction dispatches, this encoding signals the first clock cycle of the simultaneous instructions’ execution. 0x3 0011 Entry into user-mode. Signaled after execution of the instruction that caused the ColdFire processor to enter user mode. If the display of the ASID is enabled (CSR[3] = 1), the following occurs: • The 8-bit ASID follows the instruction address; that is, the PSTDDATA sequence is {0x3, 0x5, marker, instruction address, 0x8, ASID}, where 0x8 is the ASID data marker. • Whenever the current ASID is loaded by the privileged MOVEC instruction, the ASID is displayed on PSTDDATA. The resulting PSTDDATA sequence for the MOVEC instruction is then {0x1, 0x8, ASID}, where the 0x8 is the data marker for the ASID. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 8-5 Table 8-4. Processor Status Encoding (Continued) PST[3:0] Definition Hex Binary 0x4 0100 Begin execution of PULSE and WDDATA instructions. PULSE defines logic analyzer triggers for debug or performance analysis. WDDATA lets the core write any operand (byte, word, or longword) directly to the PSTDDATA port, independent of debug module configuration. When WDDATA is executed, a value of 0x4 is signaled, followed by the appropriate marker, and then the data transfer on the PSTDDATA port. Transfer length depends on the WDDATA operand size. 0x5 0101 Begin execution of taken branch or SYNC_PC command. For some opcodes, a branch target address may be displayed on PSTDDATA depending on the CSR settings. CSR also controls the number of address bytes displayed, indicated by the PST marker value preceding the DDATA nibble that begins the data output. See Section 8.3.1, “Begin Execution of Taken Branch (PST = 0x5).” Also indicates that the SYNC_PC command has been issued. 0x6 0110 Begin execution of instruction plus a taken branch. The processor completes execution of a taken conditional branch instruction and simultaneously starts executing the target instruction. This is achieved through branch folding. 0x7 0111 Begin execution of return from exception (RTE) instruction. 0x8–0xB 1000–1011 Indicates the number of bytes to be displayed on the DDATA port on subsequent clock cycles. The value is driven onto the PSTDDATA port one cycle before the data is displayed. 0x8 Begin 1-byte transfer on PSTDDATA. 0x9 Begin 2-byte transfer on PSTDDATA. 0xA Begin 3-byte transfer on PSTDDATA. 0xB Begin 4-byte transfer on PSTDDATA. 0xC 1100 Normal exception processing. Exceptions that enter emulation mode (debug interrupt or optionally trace) generate a different encoding, as described below. Because the 0xC encoding defines a multiple-cycle mode, PSTDDATA outputs are driven with 0xC until exception processing completes. 0xD 1101 Emulator mode exception processing. Displayed during emulation mode (debug interrupt or optionally trace). Because this encoding defines a multiple-cycle mode, PSTDDATA outputs are driven with 0xD until exception processing completes. 0xE 1110 A breakpoint state change causes this encoding to assert for one cycle only followed by the trigger status value. If the processor stops waiting for an interrupt, the encoding is asserted for multiple cycles. See Section 8.3.2, “Processor Stopped or Breakpoint State Change (PST = 0xE).” 0xF 1111 Processor is halted. Because this encoding defines a multiple-cycle mode, the PSTDDATA outputs display 0xF until the processor is restarted or reset. (see Section 8.5.1, “CPU Halt”) 8.3.1 Begin Execution of Taken Branch (PST = 0x5) PST is 0x5 when a taken branch is executed. For some opcodes, a branch target address may be displayed on PSTDDATA depending on the CSR settings. CSR also controls the number of address bytes displayed, which is indicated by the PST marker value immediately preceding the PSTDDATA nibble that begins the data output. Multiple byte DDATA values are displayed in least-to-most-significant order. The processor captures only those target addresses associated with taken branches which use a variant addressing mode, that is, RTE and RTS instructions, JMP and JSR instructions using address register indirect or indexed addressing modes, and all exception vectors. MCF548x Reference Manual, Rev. 5 8-6 Freescale Semiconductor Real-Time Trace Support The simplest example of a branch instruction using a variant address is the compiled code for a C language case statement. Typically, the evaluation of this statement uses the variable of an expression as an index into a table of offsets, where each offset points to a unique case within the structure. For such change-of-flow operations, the V4 microarchitecture uses the debug pins to output the following sequence of information on two successive processor clock cycles: 1. Use PSTDDATA (0x5) to identify that a taken branch is executed. 2. Optionally signal the target address to be displayed sequentially on the PSTDDATA pins. Encodings 0x9–0xB identify the number of bytes displayed. 3. The new target address is optionally available on subsequent cycles using the PSTDDATA port. The number of bytes of the target address displayed on this port is configurable (2, 3, or 4 bytes, where the encoding is 0x9, 0xA, and 0xB, respectively). Another example of a variant branch instruction would be a JMP (A0) instruction. Figure 8-4 shows when the PSTDDATA outputs that indicate when a JMP (A0) executed, assuming the CSR was programmed to display the lower 2 bytes of an address. Processor Clock PSTCLK PSTDDATA 0x59 A0[3–0,7–4] A0[11–8,15–12] Figure 8-4. Example JMP Instruction Output on PSTDDATA PSTDDATA is driven two nibbles at a time with a 0x59; 0x5 indicates a taken branch and the marker value 0x9 indicates a 2-byte address. Thus, the subsequent 4 nibbles display the lower 2 bytes of address register A0 in least-to-most-significant nibble order. The PSTDDATA output after the JMP instruction continues with the next instruction. 8.3.2 Processor Stopped or Breakpoint State Change (PST = 0xE) The 0xE encoding is generated either as a one- or multiple-cycle issue as follows: • When the core is stopped by a STOP instruction, this encoding appears in multiple-cycle format. The ColdFire processor remains stopped until an interrupt occurs; thus, PSTDDATA outputs display 0xE until stopped mode is exited. • When a breakpoint status change is to be output on PSTDDATA, 0xE is displayed for one cycle, followed immediately with the 4-bit value of the current trigger status, where the trigger status is left justified rather than in the CSR[BSTAT] description. Section 8.4.2, “Configuration/Status Register (CSR),” shows that status is right justified. That is, the displayed trigger status on PSTDDATA after a single 0xE is as follows: — 0x0 = no breakpoints enabled — 0x2 = waiting for level-1 breakpoint — 0x4 = level-1 breakpoint triggered — 0xA = waiting for level-2 breakpoint — 0xC = level-2 breakpoint triggered Thus, 0xE can indicate multiple events, based on the next value, as Table 8-5 shows. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 8-7 Table 8-5. 0xE Status Posting PSTDDATA Stream Includes 8.3.3 Result {0xE, 0x2} Breakpoint state changed to waiting for level-1 trigger {0xE, 0x4} Breakpoint state changed to level-1 breakpoint triggered {0xE, 0xA} Breakpoint state changed to waiting for level-2 trigger {0xE, 0xC} Breakpoint state changed to level-2 breakpoint triggered {0xE, 0xE} Stopped mode. Processor Halted (PST = 0xF) PST is 0xF when the processor is halted (see Section 8.5.1, “CPU Halt”). Because this encoding defines a multiple-cycle mode, the PSTDDATA outputs display 0xF until the processor is restarted or reset. Therefore, PSTDDATA[7:0] continuously are 0xFF. NOTE HALT can be distinguished from a data output 0xFF by counting 0xFF occurrences on PSTDDATA. Because data always follows a marker (0x8, 0x9, 0xA, or 0xB), the longest occurrence in PSTDDATA of 0xFF in a data output is four. Two scenarios exist for data 0xFFFF_FFFF: • The B marker occurs on the most-significant nibble of PSTDDATA with the data of 0xFF following: PSTDDATA[7:0] 0xBF 0xFF 0xFF 0xFF 0xFX (X indicates that the next PST value is guaranteed to not be 0xF.) • The B marker occurs on the least-significant nibble of PSTDDATA with the data of 0xFF following: PSTDDATA[7:0] 0xYB 0xFF 0xFF 0xFF 0xFF 0xXY (X indicates the PST value is guaranteed not to be 0xF, and Y signifies a PSTDDATA value that doesn’t affect the 0xFF count.) NOTE As the result of the above, a count of at least nine or more sequential single 0xF values or five or more sequential 0xFF values indicates the HALT condition. MCF548x Reference Manual, Rev. 5 8-8 Freescale Semiconductor Memory Map/Register Definition 8.4 Memory Map/Register Definition In addition to the existing BDM commands that provide access to the processor’s registers and the memory subsystem, the debug module contains 19 registers to support the required functionality. These registers are also accessible from the processor’s supervisor programming model by executing the WDEBUG instruction (write only). Thus, the breakpoint hardware in the debug module can be read or written by the external development system using the debug serial interface or written by the operating system running on the processor core. Software is responsible for guaranteeing that accesses to these resources are serialized and logically consistent. Hardware provides a locking mechanism in the CSR to allow the external development system to disable any attempted writes by the processor to the breakpoint registers (setting CSR[IPW]). BDM commands must not be issued if the WDEBUG instruction is used to access debug module registers or the resulting behavior is undefined. These registers, shown in Figure 8-5, are treated as 32-bit quantities, regardless of the number of implemented bits. 31 31 31 31 15 7 15 15 0 AATR Address attribute trigger register ABLR ABHR Address low breakpoint register Address high breakpoint register AATR1 Address 1 attribute trigger register 0 7 15 0 0 ABLR1 Address low breakpoint 1 register ABHR1 Address high breakpoint 1 register 31 31 31 31 15 15 15 15 7 0 BAAR BDM address attributes register CSR Configuration/status register DBR DBMR Data breakpoint register Data breakpoint mask register 0 0 0 Data breakpoint 1 register DBR1 DBMR1 Data breakpoint mask 1 register 31 31 31 15 15 15 0 PBR PBR1 PBR2 PBR3 PBMR PC breakpoint register PC breakpoint 1 register PC breakpoint 2 register PC breakpoint 3 register PC breakpoint mask register TDR Trigger definition register XTDR Extended trigger definition register 0 0 Note: Each debug register is accessed as a 32-bit register; shaded fields above are not used (don’t care). All debug control registers are writable from the external development system or the CPU via the WDEBUG instruction. CSR is write-only from the programming model. It can be read from and written to through the BDM port. CSR is accessible in supervisor mode as debug control register 0x00 using the WDEBUG instruction and Figure 8-5. Debug Programming Model MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 8-9 The registers in Table 8-7 are accessed through the BDM port by BDM commands, WDMREG and RDMREG, described in Section 8.5.3.3, “Command Set Descriptions.” These commands contain a 5-bit field, DRc, that specifies the register, as shown in Table 8-6. Table 8-6. BDM/Breakpoint Registers DRc[4–0] 0x00 0x01–0x05 Register Name Configuration/status register1 Reserved Initial State Section/ Page CSR 0x0020_0000 8.4.2/8-11 — — — 0x04 PC breakpoint ASID control PBAC — 8.4.3/8-14 0x05 BDM address attribute register BAAR 0x0000_0005 8.4.4/8-15 0x06 Address attribute trigger register AATR 0x0000_0005 8.4.5/8-16 0x07 Trigger definition register TDR 0x0000_0000 8.4.6/8-17 0x08 Program counter breakpoint register PBR — 8.4.7/8-20 0x09 Program counter breakpoint mask register PBMR — 8.4.7/8-20 — — — 0x0A–0x0B Reserved 0x0C Address breakpoint high register ABHR — 8.4.8/8-21 0x0D Address breakpoint low register ABLR — 8.4.8/8-21 0x0E Data breakpoint register DBR — 8.4.9/8-22 0x0F Data breakpoint mask register DBMR — 8.4.9/8-22 — — — PBASID — 8.4.10/8-24 — — — 0x10–0x153 Reserved 1 Abbreviation 0x14 PC breakpoint ASID register 0x15 Reserved 0x16 Address attribute trigger register 1 AATR1 0x0000_0005 8.4.5/8-16 0x17 Extended trigger definition register XTDR 0x0000_0000 8.4.11/8-25 0x18 Program counter breakpoint 1 register PBR1 0x0000_0000 8.4.7/8-20 0x19 Reserved — — — 0x1A Program counter breakpoint register 2 PBR2 0x0000_0000 8.4.7/8-20 0x1B Program counter breakpoint register 3 PBR3 0x0000_0000 8.4.7/8-20 0x1C Address high breakpoint register 1 ABHR1 — 8.4.8/8-21 0x1D Address low breakpoint register 1 ABLR1 — 8.4.8/8-21 0x1E Data breakpoint register 1 DBR1 — 8.4.9/8-22 0x1F Data breakpoint mask register 1 DBMR1 — 8.4.9/8-22 CSR is write-only from the programming model. It can be read or written through the BDM port using the RDMREG and WDMREG commands. These registers are also accessible from the processor’s supervisor programming model through the execution of the WDEBUG instruction. Thus, the external development system and the operating system running on the processor core can access the breakpoint hardware. It is the responsibility of the software MCF548x Reference Manual, Rev. 5 8-10 Freescale Semiconductor Memory Map/Register Definition to guarantee that all accesses to these resources are serialized and logically consistent. The hardware provides a locking mechanism in the CSR to allow the external development system to disable any attempted writes by the processor to the breakpoint registers (setting IPW = 1). BDM commands must not be issued if the ColdFire processor is accessing debug module registers with the WDEBUG instruction or the resulting behavior is undefined. The ColdFire debug architecture supports a number of hardware breakpoint registers, that can be configured into single- or double-level triggers based on the PC or operand address ranges with an optional inclusion of specific data values. With the addition of the MMU capabilities, the breakpoint specifications must be expanded to optionally include the address space identifier (ASID) in these user-programmable virtual address triggers. The core includes four PC breakpoint triggers and two sets of operand address breakpoint triggers, each with two independent address registers (to allow specification of a range) and a data breakpoint with masking capabilities. Core breakpoint triggers are accessible through the serial BDM interface or written through the supervisor programming model using the WDEBUG instruction. Two ASID-related registers (PBAC and PBASID) are added for the PC breakpoint qualification, and two existing registers (AATR and AATR1) are expanded for the address breakpoint qualification. 8.4.1 Revision A Shared Debug Resources In the Revision A implementation of the debug module, certain hardware structures are shared between BDM and breakpoint functionality, as shown in Table 8-7. Table 8-7. Rev. A Shared BDM/Breakpoint Hardware Register BDM Function Breakpoint Function AATR Bus attributes for all memory commands Attributes for address breakpoint ABHR Address for all memory commands Address for address breakpoint DBR Data for all BDM write commands Data for data breakpoint Thus, loading a register to perform a specific function that shares hardware resources is destructive to the shared function. For example, a BDM command to access memory overwrites an address breakpoint in ABHR. A BDM write command overwrites the data breakpoint in DBR. Revision B added hardware registers to eliminate these shared functions. The BAAR is used to specify bus attributes for BDM memory commands and has the same format as the LSB of the AATR. Note that the registers containing the BDM memory address and the BDM data are not program visible. 8.4.2 Configuration/Status Register (CSR) The configuration/status register (CSR) defines the debug configuration for the processor and memory subsystem and contains status information from the breakpoint logic. CSR is write-only from the programming model. CSR is accessible in supervisor mode as debug control register 0x00 using the WDEBUG instruction and through the BDM port using the RDMREG and WDMREG commands. It can be read from and written to through the BDM port. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 8-11 31 R 30 29 28 BSTAT 27 26 25 24 FOF TRG HALT 23 22 BKPT 21 20 HRL 19 0 18 17 16 BKD0 PCD0 IPW0 W Reset 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 NPL 0 0 0 0 0 0 0 0 0 0 R MAP TRC EMU DDC UHE BTB SSM OTE W Reset 0 0 0 0 0 0 0 Reg Addr 0 0 0 CPU + 0x00 Figure 8-6. Configuration/Status Register (CSR) Table 8-8 describes CSR fields. Table 8-8. CSR Field Descriptions Bits Name Description 31–28 BSTAT Breakpoint status. Provides read-only status information concerning hardware breakpoints. Also output on PSTDDATA when it is not displaying PST or other processor data. BSTAT is cleared by a TDR or XTDR write or by a CSR read when either a level-2 breakpoint is triggered or a level-1 breakpoint is triggered and the level-2 breakpoint is disabled. 0000 No breakpoints enabled 0001 Waiting for level-1 breakpoint 0010 Level-1 breakpoint triggered 0101 Waiting for level-2 breakpoint 0110 Level-2 breakpoint triggered 27 FOF Fault-on-fault. If FOF is set, a catastrophic halt occurred and forced entry into BDM. 26 TRG Hardware breakpoint trigger. If TRG is set, a hardware breakpoint halted the processor core and forced entry into BDM. Reset, and the debug GO command clear TRG. 25 HALT Processor halt. If HALT is set, the processor executed a HALT and forced entry into BDM. Reset, and the debug GO command clear HALT. 24 BKPT Breakpoint assert. If BKPT is set, BKPT is asserted, forcing the processor into BDM. Reset, and the debug GO command clear BKPT. 23–20 HRL Hardware revision level. Indicates the level of debug module functionality. An emulator could use this information to identify the level of functionality supported. 0000 Initial debug functionality (Revision A) 0001 Revision B 0010 Revision C 0011 Revision D 19 — Reserved, should be cleared. MCF548x Reference Manual, Rev. 5 8-12 Freescale Semiconductor Memory Map/Register Definition Table 8-8. CSR Field Descriptions (Continued) Bits Name Description 18 BKD Breakpoint disable. Used to disable the normal BKPT input functionality and to allow the assertion of BKPT to generate a debug interrupt. 0 Normal operation 1 BKPT is edge-sensitive: a high-to-low edge on BKPT signals a debug interrupt to the processor. The processor makes this interrupt request pending until the next sample point, when the exception is initiated. In the ColdFire architecture, the interrupt sample point occurs once per instruction. There is no support for nesting debug interrupts. 17 PCD PSTCLK disable. Setting PCD disables generation of PSTCLK and PSTDDATA outputs and forces them to remain quiescent. 16 IPW Inhibit processor writes. Setting IPW inhibits processor-initiated writes to the debug module’s programming model registers. IPW can be modified only by commands from the external development system. 15 MAP Force processor references in emulator mode. 0 All emulator-mode references are mapped into supervisor code and data spaces. 1 The processor maps all references while in emulator mode to a special address space, TT = 10, TM = 101 or 110. The internal SRAM and caches are disabled. 14 TRC Force emulation mode on trace exception. If TRC = 1, the processor enters emulator mode when a trace exception occurs. If TRC=0, the processor enters supervisor mode. 13 EMU Force emulation mode. If EMU = 1, the processor begins executing in emulator mode. See Section 8.6.1.1, “Emulator Mode.” 12–11 DDC Debug data control. Controls operand data capture for PSTDDATA, which displays the number of bytes defined by the operand reference size before the actual data; byte displays 8 bits, word displays 16 bits, and long displays 32 bits (one nibble at a time across multiple clock cycles). See Table 8-4. 00 No operand data is displayed. 01 Capture all write data. 10 Capture all read data. 11 Capture all read and write data. 10 UHE User halt enable. Selects the CPU privilege level required to execute the HALT instruction. 0 HALT is a supervisor-only instruction. 1 HALT is a supervisor/user instruction. 9–8 BTB Branch target bytes. Defines the number of bytes of branch target address PSTDDATA displays. 00 0 bytes 01 Lower 2 bytes of the target address 10 Lower 3 bytes of the target address 11 Entire 4-byte target address See Section 8.3.1, “Begin Execution of Taken Branch (PST = 0x5).” 7 — Reserved, should be cleared. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 8-13 Table 8-8. CSR Field Descriptions (Continued) Bits Name Description 6 NPL Non-pipelined mode. Determines whether the core operates in pipelined or mode. 0 Pipelined mode 1 Non-pipelined mode. The processor effectively executes one instruction at a time with no overlap. This adds at least 5 cycles to the execution time of each instruction. Superscalar instruction dispatch is disabled when operating in this mode. Given an average execution latency of 1.6, throughput in non-pipeline mode would be 6.6, approximately 25% or less of pipelined performance. Regardless of the NPL state, a triggered PC breakpoint is always reported before the triggering instruction executes. In normal pipeline operation, the occurrence of an address or data breakpoint trigger is imprecise. In non-pipeline mode, triggers are always reported before the next instruction begins execution and trigger reporting can be considered precise. An address or data breakpoint should always occur before the next instruction begins execution. Therefore, the occurrence of the address/data breakpoints should be guaranteed. 5 — 4 SSM Single-step mode. Setting SSM puts the processor in single-step mode. 0 Normal mode. 1 Single-step mode. The processor halts after execution of each instruction. While halted, any BDM command can be executed. On receipt of the GO command, the processor executes the next instruction and halts again. This process continues until SSM is cleared. 3 OTE Ownership-trace enable. 1 The display of the ASID on the PSTDDATA outputs by entering in user mode, by loading the ASID by a MOVEC, or by executing a BDM SYNC_PC command. 3–0 — 8.4.3 Reserved, should be cleared. Reserved, should be cleared. PC Breakpoint ASID Control Register (PBAC) The PBAC configures the breakpoint qualification for each PC breakpoint register (PBR, PBR1, PBR2, and PBR3). Four bits are dedicated for each breakpoint register and specify how the ASID is used in PC breakpoint qualification. R 31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 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 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 W Reset R PBR3AC PBR2AC PBR1AC PBRAC W Reset Reg Addr 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 CPU + 0x0A Figure 8-7. PC Breakpoint ASID Control Register (PBAC) PBR3AC, PBR2AC, PBR1AC, and PBRAC apply to PBR3, PBR2, PBR1, and PBR, respectively, and are functionally identical. They enable or disable ASID, supervisor mode, and user mode breakpoint MCF548x Reference Manual, Rev. 5 8-14 Freescale Semiconductor Memory Map/Register Definition qualification. Reset clears these fields, disabling qualifications and defaulting to the Revision C debug module functionality. Table 8-9. PBAC Field Descriptions Bits Name 31-16 — 15–12 PBR3AC 11–8 PBR2AC 7–4 PBR1AC 3–0 PBRAC 8.4.4 Description Reserved, should be cleared. PBRn ASID control. Corresponds to the ASID control associated with PBRn. Determines whether the ASID is included in the PC breakpoint comparison and whether the operating mode (supervisor or user) is included in the comparison logic. x00x No ASID qualification; no mode qualification x010 No ASID qualification; user mode qualification enabled x011 No ASID qualification; supervisor mode qualification enabled x10x ASID qualification enabled; no mode qualification x110 ASID qualification enabled; user mode qualification enabled x111 ASID qualification enabled; supervisor mode qualification enabled BDM Address Attribute Register (BAAR) The BAAR defines the address space for memory-referencing BDM commands. To maintain compatibility with Revision A, BAAR is loaded with any data written to the LSB of AATR. See Figure 8-8. The reset value of 0x5 sets supervisor data as the default address space. BAAR is write only. BAAR[R,SZ] are loaded directly from the BDM command. BAAR[TT,TM] can be programmed as debug control register 0x05 from the external development system. For compatibility with Rev. A, BAAR is loaded each time AATR is written. R 31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 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 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 0 0 0 0 0 0 0 R 0 0 0 0 0 0 0 0 0 W Reset R SZ TT TM W Reset Reg Addr 0 0 0 0 1 0 1 CPU + 0x05 Figure 8-8. BDM Address Attribute Register (BAAR) Table 8-10 describes BAAR fields. Table 8-10. BAAR Field Descriptions Bits Name Description 31-8 — Reserved 7 R Read/write 0 Write 1 Read MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 8-15 Table 8-10. BAAR Field Descriptions Bits Name 6–5 SZ Size 00 Longword 01 Byte 10 Word 11 Reserved 4–3 TT Transfer type. See the TT definition in Table 8-11. 2–0 TM Transfer modifier. See the TM definition in Table 8-11. 8.4.5 Description Address Attribute Trigger Registers (AATR, AATR1) The AATR and AATR1, Figure 8-9, define address attributes and a mask to be matched in the trigger. The register value is compared with address attribute signals from the processor’s local high-speed bus, as defined by the setting of the trigger definition register (TDR) for AATR and the extended trigger definition register (XTDR) for AATR1. This register is expanded to include an optional ASID specification and a control bit that enables the use of the ASID field. 31 R 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 0 0 0 0 0 0 0 ASIDCTRL1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 ATTRASID W Reset R RM SZM TTM TMM R SZ TT TM W Reset 0 0 Reg Addr 0 0 0 0 0 0 0 0 0 0 0 1 0 1 CPU + 0x06 (AATR), 0x16( AATR1) 1 Write only. AATR and AATR1 are accessible in supervisor mode as debug control register 0x06 and 0x16 respectively using the WDEBUG instruction and through the BDM port using the WDMREG command. Figure 8-9. Address Attribute Trigger Registers (AATR, AATR1) Table 8-11 describes AATR and AATR1 fields. Table 8-11. AATR and AATR1 Field Descriptions Bits Name 31–25 — 24 Description Reserved, should be cleared. ASIDCTRL ABLR/ABHR/ATTR address breakpoint ASID enable. Corresponds to the ASID control enable for the address breakpoint defined in ABLR, ABHR, and ATTR. 0 Disable ASID qualifier (reset default) 1 Enable ASID qualifier MCF548x Reference Manual, Rev. 5 8-16 Freescale Semiconductor Memory Map/Register Definition Table 8-11. AATR and AATR1 Field Descriptions (Continued) Bits 23–16 Name Description ATTRASID ABLR/ABHR/ATTR ASID. Corresponds to the ASID to be included in the address breakpoint specified by ABLR, ABHR, and ATTR. 15 RM Read/write mask. Setting RM masks R in address comparisons. 14–13 SZM Size mask. Setting an SZM bit masks the corresponding SZ bit in address comparisons. 12–11 TTM Transfer type mask. Setting a TTM bit masks the corresponding TT bit in address comparisons. 10–8 TMM Transfer modifier mask. Setting a TMM bit masks the corresponding TM bit in address comparisons. 7 R Read/write. R is compared with the R/W signal of the processor’s local bus. 6–5 SZ Size. Compared to the processor’s local bus size signals. 00 Longword 01 Byte 10 Word 11 Reserved 4–3 TT Transfer type. Compared with the local bus transfer type signals. 00 Normal processor access 01 Reserved 10 Emulator mode access 11 Acknowledge/CPU space access These bits also define the TT encoding for BDM memory commands. In this case, the 01 encoding indicates an external or DMA access (for backward compatibility). These bits affect the TM bits. 2–0 TM Transfer modifier. Compared with the local bus transfer modifier signals, which give supplemental information for each transfer type. TT = 00 (normal mode): 000 Data and instruction cache line push 001 User data access 010 User code access 011 Instruction cache invalidate 100 Data cache push/Instruction cache invalidate 101 Supervisor data access 110 Supervisor code access 111 INTOUCH instruction access TT = 10 (emulator mode): 0xx–100 Reserved 101 Emulator mode data access 110 Emulator mode code access 111 Reserved TT = 11 (acknowledge/CPU space transfers): 000 CPU space access 001–111 Interrupt acknowledge levels 1–7 These bits also define the TM encoding for BDM memory commands (for backward compatibility). 8.4.6 Trigger Definition Register (TDR) The TDR, shown in Table 8-10, configures the operation of the hardware breakpoint logic that corresponds with the ABHR/ABLR/AATR, PBR/PBR1/PBR2/PBR3/PBMR, and DBR/DBMR registers within the debug module. In conjunction with the XTDR and its associated debug registers, TDR controls the actions MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 8-17 taken under the defined conditions. Breakpoint logic may be configured as one- or two-level triggers. TDR[31–16] or XTDR[31–16] define second-level triggers, and bits 15–0 define first-level triggers. TDR is accessible in supervisor mode as debug control register 0x07 using the WDEBUG instruction and through the BDM port using the WDMREG command. NOTE The debug module has no hardware interlocks, so to prevent spurious breakpoint triggers while the breakpoint registers are being loaded, disable TDR and XTDR (by clearing TDR[29,13] and XTDR[29,13]) before defining triggers. A write to TDR clears the CSR trigger status bits, CSR[BSTAT]. When cleared, the data enable bits (EDxx) for both the second level and first level triggers disable data breakpoints. When set, these bits enable the corresponding data breakpoint condition based on the size and placement on the processor’s local data bus. The address breakpoint for each trigger is enabled by setting the address enable bits (EAx); clearing all three bits disables the corresponding breakpoint. 31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 DI 2 EAI 2 EAR 2 EAL 2 EPC 2 PCI 2 Second Level Triggers R TRC EBL 2 W Reset EDLW EDWL EDWU EDLL EDLM EDUM EDUU 2 2 2 2 2 2 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 DI 1 EAI 1 EAR 1 EAL 1 EPC 1 PCI 1 0 0 0 0 0 0 First Level Triggers R 0 0 EBL 1 0 0 0 W Reset EDLW EDWL EDWU EDLL EDLM EDUM EDUU 1 1 1 1 1 1 1 0 0 Reg Addr 0 0 0 0 0 CPU + 0x07 Figure 8-10. Trigger Definition Register (TDR) Table 8-12 describes TDR fields. Table 8-12. TDR Field Descriptions Bits Name Description 31–30 TRC Trigger response control. Determines how the processor responds to a completed trigger condition. The trigger response is always displayed on PSTDDATA. 00 Display on PSTDDATA only 01 Processor halt 10 Debug interrupt 11 Reserved 29 EBL2 Enable breakpoint. Global enable for the breakpoint trigger. Setting TDR[EBL] or XTDR[EBL] enables a breakpoint trigger. If both TDL[EBL] and XTDL[EBL] are cleared, all breakpoints are disabled. MCF548x Reference Manual, Rev. 5 8-18 Freescale Semiconductor Memory Map/Register Definition Table 8-12. TDR Field Descriptions (Continued) Bits Name Description 28 EDLW2 Data enable bit: Data longword. Entire processor’s local data bus. 27 EDWL2 Data enable bit: Lower data word. 26 EDWU2 Data enable bit: Upper data word. 25 EDLL2 Data enable bit: Lower lower data byte. Low-order byte of the low-order word. 24 EDLM2 Data enable bit: Lower middle data byte. High-order byte of the low-order word. 23 EDUM2 Data enable bit: Upper middle data byte. Low-order byte of the high-order word. 22 EDUU2 Data enable bit: Upper upper data byte. High-order byte of the high-order word. 21 DI2 Data breakpoint invert. Provides a way to invert the logical sense of all the data breakpoint comparators. This can develop a trigger based on the occurrence of a data value other than the DBR contents. 20 EAI2 Address enable bit: Enable address breakpoint inverted. Breakpoint is based outside the range between ABLR and ABHR. Trigger if address > ABHR or if address < ABLR. 19 EAR2 Address enable bit: Enable address breakpoint range. The breakpoint is based on the inclusive range defined by ABLR and ABHR. Trigger if address Š ABHR or if address ð ABLR. 18 EAL2 Address enable bit: Enable address breakpoint low. The breakpoint is based on the address in the ABLR. Trigger address = ABLR 17 EPC2 Enable PC breakpoint. If set, this bit enables the PC breakpoint for the second level trigger. 16 PCI2 Breakpoint invert. If set, this bit allows execution outside a given region as defined by PBR/PBR1/PBR2/PBR3 and PBMR to enable a trigger. If cleared, the PC breakpoint is defined within the region defined by PBR/PBR1/PBR2/PBR3 and PBMR. 15–14 — 13 EBL1 12 EDLW1 Data enable bit: Data longword. Entire processor’s local data bus. 11 EDWL1 Data enable bit: Lower data word. 10 EDWU1 Data enable bit: Upper data word. 9 EDLL1 Data enable bit: Lower lower data byte. Low-order byte of the low-order word. 8 EDLM1 Data enable bit: Lower middle data byte. High-order byte of the low-order word. 7 EDUM1 Data enable bit: Upper middle data byte. Low-order byte of the high-order word. 6 EDUU1 Data enable bit: Upper upper data byte. High-order byte of the high-order word. 5 DI1 Data breakpoint invert. Provides a way to invert the logical sense of all the data breakpoint comparators. This can develop a trigger based on the occurrence of a data value other than the DBR contents. 4 EAI1 Address enable bit: Enable address breakpoint inverted. Breakpoint is based outside the range between ABLR and ABHR. Trigger if address > ABHR or if address < ABLR. 3 EAR1 Address enable bit: Enable address breakpoint range. The breakpoint is based on the inclusive range defined by ABLR and ABHR. Trigger if address Š ABHR or if address ð ABLR. Reserved, should be cleared. Enable breakpoint. Global enable for the breakpoint trigger. Setting TDR[EBL] or XTDR[EBL] enables a breakpoint trigger. If both TDL[EBL] and XTDL[EBL] are cleared, all breakpoints are disabled. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 8-19 Table 8-12. TDR Field Descriptions (Continued) Bits Name Description 2 EAL1 Address enable bit: Enable address breakpoint low. The breakpoint is based on the address in the ABLR. Trigger address = ABLR 1 EPC1 Enable PC breakpoint. If set, this bit enables the PC breakpoint for the first level trigger. 0 PCI1 Breakpoint invert. If set, this bit allows execution outside a given region as defined by PBR/PBR1/PBR2/PBR3 and PBMR to enable a trigger. If cleared, the PC breakpoint is defined within the region defined by PBR/PBR1/PBR2/PBR3 and PBMR. 8.4.7 Program Counter Breakpoint and Mask Registers (PBRn, PBMR) Each PC breakpoint register (PBR, PBR1, PBR2, PBR3) defines an instruction address for use as part of the trigger. These registers’ contents are compared with the processor’s program counter register when the appropriate valid bit is set, and TDR or XTDR are configured appropriately. PBR bits are masked by setting corresponding PBMR bits. Results are compared with the processor’s program counter register, as defined in TDR or XTDR. PBR1–PBR3 are not masked. Figure 8-11 shows the PC breakpoint register. PC breakpoint registers are accessible in supervisor mode using the WDEBUG instruction and through the BDM port using the RDMREG and WDMREG commands using values shown in Section 8.5.3.3, “Command Set Descriptions.” 31 30 29 28 27 26 25 R 24 23 22 21 20 19 18 17 16 CNTRAD W Reset 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 R CNTRAD 0 V1 W Reset 0 Reg Addr 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 CPU + 0x08 (PBR); 0x18 (PBR1); 0x1A (PBR2); 0x1B (PBR3) 1 PBR0 does not have a valid bit. PBR0 is read as 0 and should be cleared. Figure 8-11. Program Counter Breakpoint Registers (PBR, PBR1, PBR2, PBR3) Table 8-13 describes PBR, PBR1, PBR2, and PBR3 fields. MCF548x Reference Manual, Rev. 5 8-20 Freescale Semiconductor Memory Map/Register Definition Table 8-13. PBR, PBR1, PBR2, PBR3 Field Descriptions Bits Name Description 31–1 CNTRAD PC breakpoint address. The 31-bit address to be compared with the PC as a breakpoint trigger. 0 V Valid. 0 Breakpoint registers are not compared with the processor’s program counter register 1 Breakpoint registers are compared with the processor’s program counter register when the appropriate valid bit is set and TDR or XTDR are configured appropriately. Note: This bit is not implemented on PBR0; it is implemented on PBR[1:3]. Figure 8-12 shows PBMR. PBMR is accessible in supervisor mode as debug control register 0x09 using the WDEBUG instruction and via the BDM port using the WDMREG command. 31 30 29 28 27 26 25 R 24 23 22 21 20 19 18 17 16 CNTRMSK W Reset 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 0 0 0 0 0 0 R CNTRMSK W Reset 0 0 0 0 0 Reg Addr 0 0 0 0 CPU + 0x09 Figure 8-12. Program Counter Breakpoint Mask Register (PBMR) Table 8-14 describes PBMR fields. Table 8-14. PBMR Field Descriptions Bits Name 31–0 8.4.8 Description CNTRMSK PC breakpoint mask. 0 This PBMR bit causes the corresponding PBR bit to be compared to the appropriate program counter register bit. 1 This PBMR bit causes the corresponding PBR bit to be ignored. Address Breakpoint Registers (ABLR/ABLR1, ABHR/ABHR1) The ABLR, ABLR1, ABHR, and ABHR1, shown in Figure 8-13, define regions in the processor’s data address space that can be used as part of the trigger. These register values are compared with the address for each transfer on the processor’s high-speed local bus. The trigger definition register (TDR) identifies the trigger as one of three cases: • Identically the value in ABLR • Inside the range bound by ABLR and ABHR inclusive • Outside that same range XTDR determines the same for ABLR1 and ABHR1. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 8-21 31 30 29 28 27 26 25 24 R 23 22 21 20 19 18 17 16 AD W1 Reset 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 0 0 0 0 1 0 1 R AD W1 Reset 0 0 0 Reg Addr 0 0 0 0 0 CPU + 0x0D (ABLR); 0x1D (ABLR1); 0x0C (ABHR); 0x1C (ABHR1) 1 ABHR and ABHR1 are accessible in supervisor mode as debug control registers 0x0C and 0x1C, using the WDEBUG instruction and via the BDM port using the RDMREG and WDMREG commands. Figure 8-13. Address Breakpoint Registers (ABLR, ABHR, ABLR1, ABHR1) Table 8-15 describes ABLR and ABLR1 fields. Table 8-15. ABLR and ABLR1 Field Description Bits Name Description 31–0 AD Low address. Holds the 32-bit address marking the lower bound of the address breakpoint range. Breakpoints for specific addresses are programmed into ABLR or ABLR1. Table 8-16 describes ABHR and ABHR1 fields. Table 8-16. ABHR and ABHR1 Field Description 8.4.9 Bits Name Description 31–0 AD High address. Holds the 32-bit address marking the upper bound of the address breakpoint range. Data Breakpoint and Mask Registers (DBR/DBR1, DBMR/DBMR1) The data breakpoint registers (DBR/DBR1, Figure 8-14), specify data patterns used as part of the trigger into debug mode. DBRn bits are masked by setting corresponding DBMR bits, as defined in TDR. DBR and DBR1 are accessible in supervisor mode as debug control register 0x0E and 0x1E, using the WDEBUG instruction and through the BDM port using the RDMREG and WDMREG commands. MCF548x Reference Manual, Rev. 5 8-22 Freescale Semiconductor Memory Map/Register Definition 31 30 29 28 27 26 25 R 24 23 22 21 20 19 18 17 16 DATA (DBR/DBR1) W Reset 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 0 0 0 0 0 0 R DATA (DBR/DBR1) W Reset 0 0 0 0 0 0 Reg Addr 0 0 0 CPU + 0x0E (DBR), 0x1E (DBR1) Figure 8-14. Data Breakpoint Registers (DBR/DBR1) Table 8-17 describes DBRn fields. Table 8-17. DBRn Field Descriptions Bits Name 31–0 DATA Description Data breakpoint value. Contains the value to be compared with the data value from the processor’s local bus as a breakpoint trigger. DBMR and DBMR1 are accessible in supervisor mode as debug control register 0x0F and 0x1F, using the WDEBUG instruction and via the BDM port using the WDMREG command. 31 30 29 28 27 26 R 25 24 23 22 21 20 19 18 17 16 MSK (DBMR/DBMR1) W Reset 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 0 0 0 0 0 R MSK (DBMR/DBMR1) W Reset 0 0 0 0 0 Reg Addr 0 0 0 0 0 CPU + 0x0F (DBMR), 0x1F (DBMR1) Figure 8-15. Data Breakpoint Mask Registers (DBMR/DBMR1) Table 8-18 describes DBMRn fields. Table 8-18. DBMRn Field Descriptions Bits Name Description 31–0 MSK Data breakpoint mask. The 32-bit mask for the data breakpoint trigger. Clearing a DBRn bit allows the corresponding DBRn bit to be compared to the appropriate bit of the processor’s local data bus. Setting a DBMRn bit causes that bit to be ignored. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 8-23 DBRs support both aligned and misaligned references. Table 8-19 shows relationships between processor address, access size, and location within the 32-bit data bus. Table 8-19. Access Size and Operand Data Location 8.4.10 A[1:0] Access Size Operand Location 00 Byte D[31:24] 01 Byte D[23:16] 10 Byte D[15:8] 11 Byte D[7:0] 0x Word D[31:16] 1x Word D[15:0] xx Longword D[31:0] PC Breakpoint ASID Register (PBASID) Each PC breakpoint register (PBR, PBR1, PBR2, or PBR3) specifies an instruction address that can be used to trigger a breakpoint. To support debugging in a virtual environment, an ASID can optionally be associated with the instruction address in the PC breakpoint registers. The optional specification of an ASID value is made using PBASID and its exact inclusion within the breakpoint specification defined by the PBAC. 31 30 29 R 28 27 26 25 24 23 22 21 PBR3ASID 20 19 18 17 16 PBR2ASID W Reset 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 0 0 R PBR1ASID PBRASID W Reset 0 0 0 Reg Addr 0 0 0 0 0 0 0 0 0 0 CPU + 0x14 Figure 8-16. PC Breakpoint ASID Register (PBASID) PBASID contains one 8-bit ASID values for each PC breakpoint register, as described in Table 8-20, which allows each PC breakpoint register to be associated with a unique virtual address and process. Table 8-20. PBASID Field Descriptions Bits Name Description 31–24 PBA3SID PBR3ASID. Corresponds to the ASID associated with PBR3. 23–16 PBA2SID PBR2ASID Corresponds to the ASID associated with PBR2. MCF548x Reference Manual, Rev. 5 8-24 Freescale Semiconductor Memory Map/Register Definition Table 8-20. PBASID Field Descriptions (Continued) Bits Name 15–8 PBA1SID PBR1ASID. Corresponds to the ASID associated with PBR1. 7–0 PBASID PBRASID. Corresponds to the ASID associated with PBR. 8.4.11 Description Extended Trigger Definition Register (XTDR) The XTDR configures the operation of the hardware breakpoint logic that corresponds with the ABHR1/ABLR1/AATR1 and DBR1/DBMR1 registers within the debug module and, in conjunction with the TDR and its associated debug registers, controls the actions taken under the defined conditions. The breakpoint logic may be configured as a one- or two-level trigger, where TDR[31–16] or XTDR[31–16] define the second-level trigger and bits 15–0 define the first-level trigger. The XTDR is accessible in supervisor mode as debug control register 0x17 using the WDEBUG instruction and via the BDM port using the WDMREG command. NOTE The debug module has no hardware interlocks, so to prevent spurious breakpoint triggers while the breakpoint registers are being loaded, disable TDR and XTDR (by clearing TDR[29,13] and XTDR[29,13]) before defining triggers. A write to the XTDR clears the trigger status bits, CSR[BSTAT]. When cleared, the data enable bits (EDxx) for both the second level and first level triggers disable data breakpoints. When set, these bits enable the corresponding data breakpoint condition based on the size and placement on the processor’s local data bus. The address breakpoint for each trigger is enabled by setting the address enable bits (EAx); clearing all three bits disables the corresponding breakpoint. Section 8.4.11.1, “Resulting Set of Possible Trigger Combinations,” describes how to handle multiple breakpoint conditions. 31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 DI 2 EAI 2 EAR 2 EAL 2 0 0 Second Level Triggers R 0 0 W Reset EBL 2 EDLW EDWL EDWU EDLL EDLM EDUM EDUU 2 2 2 2 2 2 2 — 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 DI 1 EAI 1 EAR 1 EAL 1 0 0 0 0 0 0 0 First Level Triggers R 0 W Reset 0 — 0 EBL 1 0 Reg Addr EDLW EDWL EDWU EDLL EDLM EDUM EDUU 1 1 1 1 1 1 1 0 0 0 0 0 0 0 — 0 CPU + 0x17 Figure 8-17. Extended Trigger Definition Register (XTDR) MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 8-25 Table 8-21 describes XTDR fields. Table 8-21. XTDR Field Descriptions Bits Name Description 31–30 — 29 EBL2 28 EDLW2 Data enable bit: Data longword. Entire processor’s local data bus. 27 EDWL2 Data enable bit: Lower data word. 26 EDWU2 Data enable bit: Upper data word. 25 EDLL2 Data enable bit: Lower lower data byte. Low-order byte of the low-order word. 24 EDLM2 Data enable bit: Lower middle data byte. High-order byte of the low-order word. 23 EDUM2 Data enable bit: Upper middle data byte. Low-order byte of the high-order word. 22 EDUU2 Data enable bit: Upper upper data byte. High-order byte of the high-order word. 21 DI2 Data breakpoint invert. Provides a way to invert the logical sense of all the data breakpoint comparators. This can develop a trigger based on the occurrence of a data value other than the DBR1 contents. 20 EAI2 Address enable bit: Enable address breakpoint inverted. Breakpoint is based outside the range between ABLR1 and ABHR1. Trigger if address > ABHR or if address < ABLR. 19 EAR2 Address enable bit: Enable address breakpoint range. The breakpoint is based on the inclusive range defined by ABLR1 and ABHR1. Trigger if address Š ABHR or if address ð ABLR. 18 EAL2 Address enable bit: Enable address breakpoint low. The breakpoint is based on the address in the ABLR1. Trigger address = ABLR 17–14 — 13 EBL1 12 EDLW1 Data enable bit: Data longword. Entire processor’s local data bus. 11 EDWL1 Data enable bit: Lower data word. 10 EDWU1 Data enable bit: Upper data word. 9 EDLL1 Data enable bit: Lower lower data byte. Low-order byte of the low-order word. 8 EDLM1 Data enable bit: Lower middle data byte. High-order byte of the low-order word. 7 EDUM1 Data enable bit: Upper middle data byte. Low-order byte of the high-order word. 6 EDUU1 Data enable bit: Upper upper data byte. High-order byte of the high-order word. 5 DI1 Data breakpoint invert. Provides a way to invert the logical sense of all the data breakpoint comparators. This can develop a trigger based on the occurrence of a data value other than the DBR contents. 4 EAI1 Address enable bit: Enable address breakpoint inverted. Breakpoint is based outside the range between ABLR1 and ABHR1. Trigger if address > ABHR or if address < ABLR. Reserved, should be cleared. Enable breakpoint level. If set, EBL2 is the global enable for the breakpoint trigger; that is, if TDR[EBL2] or XTDR[EBL2] is set, a breakpoint trigger is enabled. Clearing both disables all breakpoints. Reserved, should be cleared. Enable breakpoint level. If set, EBL1 is the global enable for the breakpoint trigger; that is, if TDR[EBL1] or XTDR[EBL1] is set, a breakpoint trigger is enabled. Clearing both disables all breakpoints. MCF548x Reference Manual, Rev. 5 8-26 Freescale Semiconductor Memory Map/Register Definition Table 8-21. XTDR Field Descriptions (Continued) Bits Name 3 EAR1 Address enable bit: Enable address breakpoint range. The breakpoint is based on the inclusive range defined by ABLR1 and ABHR1. Trigger if address Š ABHR or if address ð ABLR. 2 EAL1 Address enable bit: Enable address breakpoint low. The breakpoint is based on the address in the ABLR1. Trigger address = ABLR 1–0 — 8.4.11.1 Description Reserved, should be cleared. Resulting Set of Possible Trigger Combinations The resulting set of possible breakpoint trigger combinations consist of the following options where || denotes logical OR, && denotes logical AND, and {} denotes an optional additional trigger term: One-level triggers of the form: if if if (PC_breakpoint) (PC_breakpoint|| Address_breakpoint{&& Data_breakpoint}) (PC_breakpoint|| Address_breakpoint{&& Data_breakpoint} || Address1_breakpoint{&& Data1_breakpoint}) if if (Address_breakpoint {&& Data_breakpoint}) ((Address_breakpoint {&& Data_breakpoint}) || (Address1_breakpoint{&& Data1_breakpoint})) if (Address1_breakpoint {&& Data1_breakpoint}) Two-level triggers of the form: if (PC_breakpoint) then if (Address_breakpoint{&& Data_breakpoint}) if if (PC_breakpoint) then if || (Address_breakpoint{&& Data_breakpoint} Address1_breakpoint{&& Data1_breakpoint}) (PC_breakpoint) then if (Address1_breakpoint{&& Data1_breakpoint}) if (Address_breakpoint {&& Data_breakpoint}) then if (Address1_breakpoint{&& Data1_breakpoint}) if (Address1_breakpoint {&& Data1_breakpoint}) then if (Address_breakpoint{&& Data_breakpoint}) if (Address_breakpoint {&& Data_breakpoint}) then if (PC_breakpoint) if (Address1_breakpoint {&& Data1_breakpoint}) then if (PC_breakpoint) if (Address_breakpoint {&& Data_breakpoint}) MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 8-27 then if || if (PC_breakpoint Address1_breakpoint{&& Data1_breakpoint}) (Address1_breakpoint {&& Data1_breakpoint}) then if (PC_breakpoint || Address_breakpoint{&& Data_breakpoint}) In this example, PC_breakpoint is the logical summation of the PBR/PBMR, PBR1, PBR2, and PBR3 breakpoint registers; Address_breakpoint is a function of ABHR, ABLR, and AATR; Data_breakpoint is a function of DBR and DBMR; Address1_breakpoint is a function of ABHR1, ABLR1, and AATR1; and Data1_breakpoint is a function of DBR1 and DBMR1. In all cases, the data breakpoints can be included with an address breakpoint to further qualify a trigger event as an option. 8.5 Background Debug Mode (BDM) The ColdFire Family implements a low-level system debugger in the microprocessor hardware. Communication with the development system is handled through a dedicated, high-speed serial command interface. The ColdFire architecture implements the BDM controller in a dedicated hardware module. Although some BDM operations, such as CPU register accesses, require the CPU to be halted, all other BDM commands, such as memory accesses, can be executed while the processor is running. BDM is useful for the following reasons: • In-circuit emulation is not needed, so physical and electrical characteristics of the system are not affected. • BDM is always available for debugging the system and provides a communication link for upgrading firmware in existing systems. • Provides high-speed cache downloading (500 Kbytes/sec), especially useful for flash programming • Provides absolute control of the processor, and thus the system. This feature allows quick hardware debugging with the same tool set used for firmware development. 8.5.1 CPU Halt Although most BDM operations can occur in parallel with CPU operations, unrestricted BDM operation requires the CPU to be halted. The sources that can cause the CPU to halt are listed below, in order of priority: 1. A catastrophic fault-on-fault condition automatically halts the processor. 2. A hardware breakpoint can be configured to generate a pending halt condition similar to the assertion of BKPT. This type of halt is always first made pending in the processor. Next, the processor samples for pending halt and interrupt conditions once per instruction. When a pending condition is asserted, the processor halts execution at the next sample point. See Section 8.6.1, “Theory of Operation.” 3. The execution of a HALT instruction immediately suspends execution. Attempting to execute HALT in user mode while CSR[UHE] = 0 generates a privilege violation exception. If CSR[UHE] = 1, HALT can be executed in user mode. After HALT executes, the processor can be restarted by serial shifting a GO command into the debug module. Execution continues at the instruction after HALT. MCF548x Reference Manual, Rev. 5 8-28 Freescale Semiconductor Background Debug Mode (BDM) 4. The assertion of the BKPT input is treated as a pseudo-interrupt; that is, asserting BKPT creates a pending halt, which is postponed until the processor core samples for halts/interrupts. The processor samples for these conditions once during the execution of each instruction; if a pending halt is detected then, the processor suspends execution and enters the halted state. The assertion of BKPT should be considered in the following two special cases: • After the system reset signal is negated, the processor waits for 16 processor clock cycles before beginning reset exception processing. If the BKPT input is asserted within eight cycles after RSTI is negated, the processor enters the halt state, signaling halt status (0xF) on the PSTDDATA outputs. While the processor is in this state, all resources accessible through the debug module can be referenced. This is the only chance to force the processor into emulation mode through CSR[EMU]. After system initialization, the processor’s response to the GO command depends on the set of BDM commands performed while it is halted for a breakpoint. Specifically, if the PC register was loaded, the GO command causes the processor to exit halted state and pass control to the instruction address in the PC, bypassing normal reset exception processing. If the PC was not loaded, the GO command causes the processor to exit halted state and continue reset exception processing. • The ColdFire architecture also handles a special case of BKPT being asserted while the processor is stopped by execution of the STOP instruction. For this case, the processor exits the stopped mode and enters the halted state. At this point, all BDM commands may be exercised. When restarted, the processor continues by executing the next sequential instruction, that is, the instruction following the STOP opcode. CSR[27–24] indicates the halt source, showing the highest priority source for multiple halt conditions. Debug module Revisions A and B clear CSR[27–24] upon a read of the CSR, but Revision C and D (in V4) do not. The debug GO command clears CSR[26–24]. HALT can be recognized by counting 0xFF occurrences on PSTDDATA. The count is necessary to determine between a possible data output value of 0xFF and the HALT condition. Because data always follows a marker (0x8, 0x9, 0xA, or 0xB), PSTDDATA can display no more than four data 0xFFs. Two such scenarios exist: • A B marker occurs on the left nibble of PSTDDATA with the data of 0xFF following: PSTDDATA[7:0] 0xBF 0xFF 0xFF 0xFF 0xFX (X indicates that the next PST value is guaranteed to not be 0xF) • A B marker occurs on the right nibble of PSTDDATA with the data of 0xFF following: PSTDDATA[7:0] 0xYB 0xFF 0xFF 0xFF 0xFF 0xXY (X indicates that the PST value is guaranteed to not be 0xF, and Y indicates a PSTDDATA value that doesn’t affect the 0xFF count). Thus, a count of either nine or more sequential single 0xF values or five or more sequential 0xFF values signifies the HALT condition. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 8-29 8.5.2 BDM Serial Interface When the CPU is halted and PSTDDATA reflects the halt status, the development system can send unrestricted commands to the debug module. The debug module implements a synchronous protocol using two inputs (DSCLK and DSI) and one output (DSO), where DSO is specified as a delay relative to the rising edge of the processor clock. See Table 8-1. The development system serves as the serial communication channel master and must generate DSCLK. The serial channel operates at a frequency from DC to 1/5 of the PSTCLK frequency. The channel uses full-duplex mode, where data is sent and received simultaneously by both master and slave devices. The transmission consists of 17-bit packets composed of a status/control bit and a 16-bit data word. As shown in Figure 8-18, all state transitions are enabled on a rising edge of the PSTCLK clock when DSCLK is high; that is, DSI is sampled and DSO is driven. C0 C1 C2 C3 C4 PSTCLK DSCLK DSI BDM State Machine DSO Current Current State Past Next Next State Current Figure 8-18. Maximum BDM Serial Interface Timing DSCLK and DSI are synchronized inputs. DSCLK acts as a pseudo clock enable and is sampled, along with DSI, on the rising edge of PSTCLK. DSO is delayed from the DSCLK-enabled PSTCLK rising edge (registered after a BDM state machine state change). All events in the debug module’s serial state machine are based on the PSTCLK rising edge. DSCLK must also be sampled low (on a positive edge of PSTCLK) between each bit exchange. The msb is sent first. Because DSO changes state based on an internally recognized rising edge of DSCLK, DSO cannot be used to indicate the start of a serial transfer. The development system must count clock cycles in a given transfer. C0–C4 are described as follows: • C0: Set the state of the DSI bit. • C1: First synchronization cycle for DSI (DSCLK is high). • C2: Second synchronization cycle for DSI (DSCLK is high). • C3: BDM state machine changes state depending upon DSI and whether the entire input data transfer has been transmitted. • C4: DSO changes to next value. NOTE A not-ready response can be ignored except during a memory-referencing cycle. Otherwise, the debug module can accept a new serial transfer after 32 processor clock periods. 8.5.2.1 Receive Packet Format The basic receive packet, Figure 8-19, consists of 16 data bits and 1 status bit MCF548x Reference Manual, Rev. 5 8-30 Freescale Semiconductor Background Debug Mode (BDM) . 16 15 0 S Data Field [15:0] Figure 8-19. Receive BDM Packet Table 8-22 describes receive BDM packet fields. Table 8-22. Receive BDM Packet Field Description Bits Name Description 16 S Status. Indicates the status of CPU-generated messages listed below. The not-ready response can be ignored unless a memory-referencing cycle is in progress. Otherwise, the debug module can accept a new serial transfer after 32 processor clock periods. S 0 0 1 1 1 15–0 Data 8.5.2.2 DataMessage xxxx Valid data transfer 0xFFFFStatus OK 0x0000Not ready with response; come again 0x0001Error: Terminated bus cycle; data invalid 0xFFFFIllegal command Data. Contains the message to be sent from the debug module to the development system. The response message is always a single word, with the data field encoded as shown above. Transmit Packet Format The basic transmit packet, Figure 8-20, consists of 16 data bits and 1 control bit. 16 15 0 C D[15:0] Figure 8-20. Transmit BDM Packet Table 8-23 describes transmit BDM packet fields. Table 8-23. Transmit BDM Packet Field Description 8.5.3 Bits Name 16 C 15–0 Data Description Control. This bit is reserved. Command and data transfers initiated by the development system should clear C. Contains the data to be sent from the development system to the debug module. BDM Command Set Table 8-24 summarizes the BDM command set. Subsequent paragraphs contain detailed descriptions of each command. Issuing a BDM command when the processor is accessing debug module registers using the WDEBUG instruction causes undefined behavior. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 8-31 Table 8-24. BDM Command Summary Description CPU State1 Command (Hex) Command Mnemonic Read A/D register rareg/ rdreg Read the selected address or data register and return the results through the serial interface. Halted 8.5.3.3.1 0x218 {A/D, Reg[2:0]} Write A/D register wareg/ wdreg Write the data operand to the specified address or data register. Halted 8.5.3.3.2 0x208 {A/D, Reg[2:0]} Read memory location read Read the data at the memory location specified by the longword address. Steal 8.5.3.3.3 0x1900—byte 0x1940—word 0x1980—lword Write memory location write Write the operand data to the memory location specified by the longword address. Steal 8.5.3.3.4 0x1800—byte 0x1840—word 0x1880—lword Dump memory block dump Used with READ to dump large blocks of memory. An initial READ is executed to set up the starting address of the block and to retrieve the first result. A DUMP command retrieves subsequent operands. Steal 8.5.3.3.5 0x1D00—byte 0x1D40—word 0x1D80—lword Fill memory block fill Used with WRITE to fill large blocks of memory. An initial WRITE is executed to set up the starting address of the block and to supply the first operand. A FILL command writes subsequent operands. Steal 8.5.3.3.6 0x1C00—byte 0x1C40—word 0x1C80—lword Resume execution go The pipeline is flushed and refilled before resuming instruction execution at the current PC. Halted 8.5.3.3.7 0x0C00 No operation nop Perform no operation; may be used as a null command. Parallel 8.5.3.3.8 0x0000 Output the current PC sync_pc Capture the current PC and display it on the PSTDDATA output pins. Parallel 8.5.3.3.9 0x0001 Read control register rcreg Read the system control register. Halted 8.5.3.3.11 0x2980 Write control register wcreg Write the operand data to the system control register. Halted 8.5.3.3.15 0x2880 Read debug module register rdmreg Read the debug module register. Parallel 8.5.3.3.16 0x2D {0x42 DRc[4:0]} Write debug module register wdmreg Write the operand data to the debug module register. Parallel 8.5.3.3.17 0x2C {0x42 DRc[4:0]} Section 1 General command effect and/or requirements on CPU operation: - Halted. The CPU must be halted to perform this command. - Steal. Command generates bus cycles that can be interleaved with bus accesses. - Parallel. Command is executed in parallel with CPU activity. 2 0x4 is a three-bit field. Unassigned command opcodes are reserved by Freescale. All unused command formats within any revision level perform a NOP and return the illegal command response. MCF548x Reference Manual, Rev. 5 8-32 Freescale Semiconductor Background Debug Mode (BDM) 8.5.3.1 ColdFire BDM Command Format All ColdFire Family BDM commands include a 16-bit operation word followed by an optional set of one or more extension words, as shown in Figure 8-21. 15 10 Operation 9 8 0 R/W 7 6 5 4 3 Op Size 0 0 A/D 2 0 Register Extension Word(s) Figure 8-21. BDM Command Format Table 8-25 describes BDM fields. Table 8-25. BDM Field Descriptions Bit Name 15–10 Operation 9 — 8 R/W 7–6 Operand Size 5–4 — 3 A/D 2–0 Register 8.5.3.1.1 Description Specifies the command. These values are listed in Table 8-24. Reserved Direction of operand transfer. 0 Data is written to the CPU or to memory from the development system. 1 The transfer is from the CPU to the development system. Operand data size for sized operations. Addresses are expressed as 32-bit absolute values. Note that a command performing a byte-sized memory read leaves the upper 8 bits of the response data undefined. Referenced data is returned in the lower 8 bits of the response. Operand SizeBit Values 00 Byte8 bits 01 Word16 bits 10 Longword32 bits 11 Reserved— Reserved Address/data. Determines whether the register field specifies a data or address register. 0 Indicates a data register. 1 Indicates an address register. Contains the register number in commands that operate on processor registers. Extension Words as Required Some commands require extension words for addresses or immediate data. Addresses require two extension words because only absolute long addressing is permitted. Longword accesses are forcibly longword-aligned and word accesses are forcibly word-aligned. Immediate data can be 1 or 2 words long. Byte and word data each requires one extension word and longword data requires two extension words. Operands and addresses are transferred most-significant word first. In the following descriptions of the BDM command set, the optional set of extension words is defined as address, data, or operand data. 8.5.3.2 Command Sequence Diagrams The command sequence diagram in Figure 8-22 shows serial bus traffic for commands. Each bubble represents a 17-bit bus transfer. The top half of each bubble indicates the data the development system MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 8-33 sends to the debug module; the bottom half indicates the debug module’s response to the previous development system commands. Command and result transactions overlap to minimize latency. Commands transmitted to the debug module Command code transmitted during this cycle High-order 16 bits of memory address Low-order 16 bits of memory address Non-serial-related activity READ (LONG) ??? MS ADDR ’NOT READY’ LS ADDR ’NOT READY’ XXX ’ILLEGAL’ NEXT CMD ’NOT READY’ READ MEMORY LOCATION Sequence taken if operation has not completed XXX ’NOT READY’ Next Command Code XXX MS RESULT NEXT CMD LS RESULT XXX BERR NEXT CMD ’NOT READY’ Data used from this transfer Sequence taken if illegal command is received by debug module Results from previous command Responses from the debug module Sequence taken if bus error occurs on memory access High- and low-order 16 bits of result Figure 8-22. Command Sequence Diagram The sequence is as follows: • In cycle 1, the development system command is issued (READ in this example). The debug module responds with either the low-order results of the previous command or a command complete status of the previous command, if no results are required. • In cycle 2, the development system supplies the high-order 16 address bits. The debug module returns a not-ready response unless the received command is decoded as unimplemented, which is indicated by the illegal command encoding. If this occurs, the development system should retransmit the command. NOTE A not-ready response can be ignored except during a memory-referencing cycle. Otherwise, the debug module can accept a new serial transfer after 32 processor clock periods. • • • In cycle 3, the development system supplies the low-order 16 address bits. The debug module always returns a not-ready response. At the completion of cycle 3, the debug module initiates a memory read operation. Any serial transfers that begin during a memory access return a not-ready response. Results are returned in the two serial transfer cycles after the memory access completes. For any command performing a byte-sized memory read operation, the upper 8 bits of the response data are undefined and the referenced data is returned in the lower 8 bits. The next command’s opcode is sent to the debug module during the final transfer. If a memory or register access is terminated with a bus error, the error status (S = 1, DATA = 0x0001) is returned instead of result data. MCF548x Reference Manual, Rev. 5 8-34 Freescale Semiconductor Background Debug Mode (BDM) 8.5.3.3 Command Set Descriptions The following sections describe the commands summarized in Table 8-24. NOTE The BDM status bit (S) is 0 for normally completed commands. S = 1 for illegal commands, not-ready responses, and transfers with bus-errors. Section 8.5.2, “BDM Serial Interface,” describes the receive packet format. Freescale reserves unassigned command opcodes for future expansion. Unused command formats in any revision level perform a NOP and return an illegal command response. 8.5.3.3.1 Read A/D Register (RAREG/RDREG) Read the selected address or data register and return the 32-bit result. A bus error response is returned if the CPU core is not halted. Command/Result Formats: 15 Command 12 11 0x2 8 7 0x1 4 0x8 Result 3 A/D 2 0 Register D[31:16] D[15:0] Figure 8-23. RAREG/RDREG Command Format Command Sequence: RAREG/RDREG ??? XXX MS RESULT NEXT CMD LS RESULT XXX BERR NEXT CMD ’NOT READY’ Figure 8-24. RAREG/RDREG Command Sequence Operand Data: Result Data: 8.5.3.3.2 None The contents of the selected register are returned as a longword value, most-significant word first. Write A/D Register (WAREG/WDREG) The operand longword data is written to the specified address or data register. A write alters all 32 register bits. A bus error response is returned if the CPU core is not halted. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 8-35 Command Format: 15 12 11 0x2 8 7 0x0 4 0x8 3 2 A/D 0 Register D[31:16] D[15:0] Figure 8-25. WAREG/WDREG Command Format Command Sequence WAREG/WDREG ??? MS DATA ’NOT READY’ LS DATA ’NOT READY’ XXX BERR NEXT CMD ’NOT READY’ NEXT CMD ’CMD COMPLETE’ Figure 8-26. WAREG/WDREG Command Sequence Operand Data Result Data 8.5.3.3.3 Longword data is written into the specified address or data register. The data is supplied most-significant word first. Command complete status is indicated by returning 0xFFFF (with S cleared) when the register write is complete. Read Memory Location (READ) Read data at the longword address. Address space is defined by BAAR[TT,TM]. Hardware forces low-order address bits to zeros for word and longword accesses to ensure that word addresses are word-aligned and longword addresses are longword-aligned. Command/Result Formats: MCF548x Reference Manual, Rev. 5 8-36 Freescale Semiconductor Background Debug Mode (BDM) 15 12 Byte 11 8 0x1 7 0x9 4 3 0x0 Command 0 0x0 A[31:16] A[15:0] Result Word Command X X X X X X 0x1 X X 0x9 D[7:0] 0x4 0x0 0x8 0x0 A[31:16] A[15:0] Result Longword Command D[15:0] 0x1 0x9 A[31:16] A[15:0] Result D[31:16] D[15:0] Figure 8-27. READ Command/Result Formats Command Sequence: READ (B/W) ??? MS ADDR ’NOT READY’ LS ADDR ’NOT READY’ READ MEMORY LOCATION XXX ’NOT READY’ NEXT CMD RESULT XXX BERR READ (LONG) ??? MS ADDR ’NOT READY’ LS ADDR ’NOT READY’ READ MEMORY LOCATION NEXT CMD ’NOT READY’ XXX ’NOT READY’ XXX MS RESULT NEXT CMD LS RESULT XXX BERR NEXT CMD ’NOT READY’ Figure 8-28. READ Command Sequence Operand Data The only operand is the longword address of the requested location. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 8-37 Result Data 8.5.3.3.4 Word results return 16 bits of data; longword results return 32. Bytes are returned in the LSB of a word result, the upper byte is undefined. 0x0001 (S = 1) is returned if a bus error occurs. Write Memory Location (WRITE) Write data to the memory location specified by the longword address. The address space is defined by BAAR[TT,TM]. Hardware forces low-order address bits to zeros for word and longword accesses to ensure that word addresses are word-aligned and longword addresses are longword-aligned. Command Formats: 15 12 Byte 11 8 0x1 7 0x8 4 0x0 3 1 0x0 A[31:16] A[15:0] X Word X X 0x1 X X X X X 0x8 D[7:0] 0x4 0x0 0x8 0x0 A[31:16] A[15:0] D[15:0] Longword 0x1 0x8 A[31:16] A[15:0] D[31:16] D[15:0] Figure 8-29. WRITE Command Format MCF548x Reference Manual, Rev. 5 8-38 Freescale Semiconductor Background Debug Mode (BDM) Command Sequence: WRITE (B/W) ??? MS ADDR ’NOT READY’ LS ADDR ’NOT READY’ DATA ’NOT READY’ WRITE MEMORY LOCATION XXX ’NOT READY’ NEXT CMD ’CMD COMPLETE’ XXX BERR NEXT CMD ’NOT READY’ WRITE (LONG) ??? MS ADDR ’NOT READY’ LS ADDR ’NOT READY’ MS DATA ’NOT READY’ LS DATA ’NOT READY’ WRITE MEMORY LOCATION XXX ’NOT READY’ NEXT CMD ’CMD COMPLETE’ XXX BERR NEXT CMD ’NOT READY’ Figure 8-30. WRITE Command Sequence Operand Data Result Data 8.5.3.3.5 This two-operand instruction requires a longword absolute address that specifies a location to which the data operand is to be written. Byte data is sent as a 16-bit word, justified in the LSB; 16- and 32-bit operands are sent as 16 and 32 bits, respectively. Command complete status is indicated by returning 0xFFFF (with S cleared) when the register write is complete. A value of 0x0001 (with S set) is returned if a bus error occurs. Dump Memory Block (DUMP) DUMP is used with the READ command to access large blocks of memory. An initial READ is executed to set up the starting address of the block and to retrieve the first result. If an initial READ is not executed before the first DUMP, an illegal command response is returned. The DUMP command retrieves subsequent operands. The initial address is incremented by the operand size (1, 2, or 4) and saved in a temporary register. Subsequent DUMP commands use this address, perform the memory read, increment it by the current operand size, and store the updated address in the temporary register. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 8-39 NOTE DUMP does not check for a valid address; it is a valid command only when preceded by NOP, READ, or another DUMP command. Otherwise, an illegal command response is returned. NOP can be used for intercommand padding without corrupting the address pointer. The size field is examined each time a dynamically altered. DUMP command is processed, allowing the operand size to be Command/Result Formats: 15 Byte Command Result Word 12 11 8 0x1 X Command X 7 0xD X X X 0x1 X 4 3 0x0 X Result 0x0 X 0xD 0 D[7:0] 0x4 0x0 0x8 0x0 D[15:0] Longword Command 0x1 0xD Result D[31:16] D[15:0] Figure 8-31. DUMP Command/Result Formats Command Sequence: READ MEMORY LOCATION DUMP (B/W) ??? XXX ’NOT READY’ NEXT CMD RESULT XXX ’ILLEGAL’ NEXT CMD ’NOT READY’ READ MEMORY LOCATION DUMP (LONG) ??? XXX ’ILLEGAL’ XXX BERR NEXT CMD ’NOT READY’ XXX ’NOT READY’ NEXT CMD ’NOT READY’ NEXT CMD MS RESULT NEXT CMD LS RESULT XXX BERR NEXT CMD ’NOT READY’ Figure 8-32. DUMP Command Sequence Operand Data: None MCF548x Reference Manual, Rev. 5 8-40 Freescale Semiconductor Background Debug Mode (BDM) Result Data: 8.5.3.3.6 Requested data is returned as either a word or longword. Byte data is returned in the least-significant byte of a word result. Word results return 16 bits of significant data; longword results return 32 bits. A value of 0x0001 (with S set) is returned if a bus error occurs. Fill Memory Block (FILL) A FILL command is used with the WRITE command to access large blocks of memory. An initial WRITE is executed to set up the starting address of the block and to supply the first operand. The FILL command writes subsequent operands. The initial address is incremented by the operand size (1, 2, or 4) and saved in a temporary register after the memory write. Subsequent FILL commands use this address, perform the write, increment it by the current operand size, and store the updated address in the temporary register. If an initial returned. WRITE is not executed preceding the first FILL command, the illegal command response is NOTE The FILL command does not check for a valid address: FILL is a valid command only when preceded by another FILL, a NOP, or a WRITE command. Otherwise, an illegal command response is returned. The NOP command can be used for intercommand padding without corrupting the address pointer. The size field is examined each time a FILL command is processed, allowing the operand size to be altered dynamically. Command Formats: 15 12 Byte 11 8 0x1 X Word X 0xC X 0x1 7 X X X 4 3 0x0 X X 0xC 0 0x0 D[7:0] 0x4 0x0 0x8 0x0 D[15:0] Longword 0x1 0xC D[31:16] D[15:0] Figure 8-33. FILL Command Format MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 8-41 Command Sequence: FILL (LONG) ??? MS DATA ’NOT READY’ LS DATA ’NOT READY’ XXX ’ILLEGAL’ NEXT CMD ’NOT READY’ WRITE MEMORY LOCATION XXX ’NOT READY’ NEXT CMD ’CMD COMPLETE’ XXX BERR FILL (B/W) ??? DATA ’NOT READY’ WRITE MEMORY LOCATION XXX ’ILLEGAL’ NEXT CMD ’NOT READY’ NEXT CMD ’NOT READY’ XXX ’NOT READY’ NEXT CMD ’CMD COMPLETE’ NEXT CMD ’NOT READY’ XXX BERR Figure 8-34. FILL Command Sequence Operand Data: Result Data: 8.5.3.3.7 A single operand is data to be written to the memory location. Byte data is sent as a 16-bit word, justified in the least-significant byte; 16- and 32-bit operands are sent as 16 and 32 bits, respectively. Command complete status (0xFFFF) is returned when the register write is complete. A value of 0x0001 (with S set) is returned if a bus error occurs. Resume Execution (GO) The pipeline is flushed and refilled before normal instruction execution resumes. Prefetching begins at the current address in the PC and at the current privilege level. If any register (such as the PC or SR) is altered by a BDM command while the processor is halted, the updated value is used when prefetching resumes. If a GO command is issued and the CPU is not halted, the command is ignored. 15 12 0x0 11 8 0xC 7 4 0x0 3 0 0x0 Figure 8-35. GO Command Format Command Sequence: GO ??? NEXT CMD ’CMD COMPLETE’ Figure 8-36. GO Command Sequence Operand Data: None MCF548x Reference Manual, Rev. 5 8-42 Freescale Semiconductor Background Debug Mode (BDM) Result Data: 8.5.3.3.8 NOP The command-complete response (0xFFFF) is returned during the next shift operation. No Operation (NOP) performs no operation and may be used as a null command where required. Command Formats: 15 12 0x0 11 8 0x0 7 4 0x0 3 0 0x0 Figure 8-37. NOP Command Format Command Sequence: NOP ??? NEXT CMD ’CMD COMPLETE’ Figure 8-38. NOP Command Sequence Operand Data: Result Data: 8.5.3.3.9 None The command-complete response, 0xFFFF (with S cleared), is returned during the next shift operation. Synchronize PC to the PSTDDATA Lines (SYNC_PC) The SYNC_PC command captures the current PC and displays it on the PSTDDATA outputs. After the debug module receives the command, it sends a signal to the ColdFire processor that the current PC must be displayed. The processor then forces an instruction fetch at the next PC with the address being captured in the DDATA logic under control of CSR[BTB]. The specific sequence of PSTDDATA values is as follows: 1. Debug signals a SYNC_PC command is pending. 2. CPU completes the current instruction. 3. CPU forces an instruction fetch to the next PC, generates a PST = 0x5 value indicating a taken branch and signals the capture of DDATA. 4. The instruction address corresponding to the PC is captured. 5. The PST marker (0x9–0xB) is generated and displayed as defined by CSR[BTB] followed by the captured PC address. If the option to display ASID is enabled (CSR[3] = 1), the 8-bit ASID follows the address. That is, the PSTDDATA sequence is {0x5, Marker, Instruction Address, 0x8, ASID}, where the 0x8 is the marker for the ASID. The SYNC_PC command can be used to dynamically access the PC for performance monitoring. The execution of this command is considerably less obtrusive to the real-time operation of an application than a HALT-CPU/READ-PC/RESUME command sequence. Command Formats: MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 8-43 15 12 11 0x0 8 7 4 0x0 3 0x0 0 0x1 Figure 8-39. SYNC_PC Command Format Command Sequence: SYNC_PC NEXT CMD ??? “CMD COMPLETE” Figure 8-40. SYNC_PC Command Sequence Operand Data: Result Data: 8.5.3.3.10 None Command complete status (0xFFFF) is returned when the register write is complete. Force Transfer Acknowledge (FORCE_TA) DEBUG_D logic implements the new FORCE_TA serial BDM command to resolve a hung bus condition. In some system designs, references to certain unmapped memory addresses may cause the external bus to hang with no transfer acknowledge generated by any bus responders. The FORCE_TA forces generation of a transfer acknowledge signal, which can be logically summed into the normal acknowledge logic located in the system integration module (SIM) outside of the ColdFire core. There are two scenarios of interest, one caused by a processor access and the other caused by a BDM access. The following sequences identify the operations needed to break the hung bus condition: • Bus hang caused by processor or external or internal alternate master: — Assert the breakpoint input to force a processor core halt. — If the bus hang was caused by a processor access, send in FORCE_TA commands until the processor is halted, as signaled by PST = 0xF. Due to pipeline and store buffer depths, many memory accesses may be queued up behind the access causing the bus hang. Repeated FORCE_TA commands eventually allow processing of all these pending accesses. As soon as the processor is halted, the system reaches a quiescent, controllable state. — If the hang was caused by another master, such as a DMA channel, the processor can halt immediately. In this case as well, multiple assertions of the FORCE_TA command may be required to terminate the alternate master’s errant access. • Bus hang caused by BDM access: — It is assumed the processor is already halted at the time of the errant BDM access. To resolve the hung bus, it is necessary to process four or more FORCE_TA commands, because the BDM command may have initiated a cache line access that fetches 4 longwords, each needing a unique transfer acknowledge. Formats: 15 12 0x0 11 8 0x0 7 4 0x0 3 0 0x2 Figure 8-41. FORCE_TA Command Command Sequence: MCF548x Reference Manual, Rev. 5 8-44 Freescale Semiconductor Background Debug Mode (BDM) FORCE_TA NEXT CMD ??? “CMD COMPLETE” Figure 8-42. FORCE_TA Command Sequence Operand Data: Result Data: 8.5.3.3.11 None The command complete response, 0xFFFF (with the status bit cleared), is returned during the next shift operation. This response indicates the FORCE_TA command was processed correctly and does not necessarily reflect the status of any internal bus. Read Control Register (RCREG) Read the selected control register and return the 32-bit result. Accesses to the processor/memory control registers are always 32 bits wide, regardless of register width. The second and third words of the command form a 32-bit address, which the debug module uses to generate a special bus cycle to access the specified control register. The 12-bit Rc field is the same as that used by the MOVEC instruction. Command/Result Formats: 15 Command 12 11 8 7 4 3 0 0x2 0x9 0x8 0x0 0x0 0x0 0x0 0x0 0x0 Rc Result D[31:16] D[15:0] Figure 8-43. RCREG Command/Result Formats Command Sequence: RCREG ??? MS ADDR ’NOT READY’ MS ADDR ’NOT READY’ READ CONTROL REGISTER XXX ’NOT READY’ NEXT CMD MS RESULT NEXT CMD LS RESULT XXX BERR NEXT CMD ’NOT READY’ Figure 8-44. RCREG Command Sequence Operand Data: Result Data: The only operand is the 32-bit Rc control register select field. Control register contents are returned as a longword, most-significant word first. The implemented portion of registers smaller than 32 bits is guaranteed correct; other bits are undefined. Rc encoding: See Table 8-26. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 8-45 Table 8-26. ColdFire CPU Control Register Map Name CPU Space (Rc) Register Name Memory Management Control Registers CACR 0x002 Cache control register ASID 0x003 Address space identifier ACR0–ACR3 MMUBAR 0x004–0x007 Access control registers 0–3 0x008 MMU base address register Processor General-Purpose Registers D0–D7 0x(0,1)80–0x(0,1)87 Data registers 0–7 (0 = load, 1 = store) A0–A7 0x(0,1)88–0x(0,1)8F Address registers 0–7 (0 = load, 1 = store) A7 is user stack pointer Processor Miscellaneous Registers OTHER_A7 0x800 Other stack pointer VBR 0x801 Vector base register MACSR 0x804 MAC status register MASK 0x805 MAC address mask register ACC0–ACC3 0x806–0x80B MAC accumulators 0–3 ACCext01 0x807 MAC accumulator 0, 1 extension bytes ACCext23 0x808 MAC accumulator 2, 3 extension bytes SR 0x80E Status register PC 0x80F Program counter Processor Floating-Point Registers FPU0 0x810 32 msbs of floating-point data register 0 FPL0 0x811 32 lsbs of floating-point data register 0 FPU1 0x812 32 msbs of floating-point data register 1 FPL1 0x813 32 lsbs of floating-point data register 1 FPU2 0x814 32 msbs of floating-point data register 2 FPL2 0x815 32 lsbs of floating-point data register 2 FPU3 0x816 32 msbs of floating-point data register 3 FPL3 0x817 32 lsbs of floating-point data register 3 FPU4 0x818 32 msbs of floating-point data register 4 FPL4 0x819 32 lsbs of floating-point data register 4 FPU5 0x81A 32 msbs of floating-point data register 5 FPL5 0x81B 32 lsbs of floating-point data register 5 FPU6 0x81C 32 msbs of floating-point data register 6 MCF548x Reference Manual, Rev. 5 8-46 Freescale Semiconductor Background Debug Mode (BDM) Table 8-26. ColdFire CPU Control Register Map (Continued) Name CPU Space (Rc) Register Name FPL6 0x81D 32 lsbs of floating-point data register 6 FPU7 0x81E 32 msbs of floating-point data register 7 FPL7 0x81F 32 lsbs of floating-point data register 7 FPIAR 0x821 Floating-point instruction address register FPSR 0x822 Floating-point status register FPCR 0x824 Floating-point control register Local Memory and Module Control Registers RAMBAR0 0xC04 RAM base address register 0 RAMBAR1 0xC05 RAM base address register 1 MBAR 0xC0F Primary module base address register (not a core register) 8.5.3.3.12 BDM Accesses of the Stack Pointer Registers (A7: SSP and USP) The Version 4 ColdFire core supports two unique stack pointer (A7) registers: the supervisor stack pointer (SSP) and the user stack pointer (USP). The hardware implementation of these two programmable-visible 32-bit registers does not uniquely identify one as the SSP and the other as the USP. Rather, the hardware uses one 32-bit register as the currently-active A7; the other is named simply the OTHER_A7. Thus, the contents of the two hardware registers is a function of the operating mode of the processor: if SR[S] = 1 then else A7 = Supervisor Stack Pointer OTHER_A7 = User Stack Pointer A7 = User Stack Pointer OTHER_A7 = Supervisor Stack Pointer The BDM programming model supports reads and writes to A7 and OTHER_A7 directly. It is the responsibility of the external development system to determine the mapping of A7 and OTHER_A7 to the two program-visible definitions (supervisor and user stack pointers), based on the SR[S]. 8.5.3.3.13 BDM Accesses of the EMAC Registers The presence of rounding logic in the output datapath of the EMAC requires special care for BDM-initiated reads and writes of its programming model. In particular, any result rounding modes must be disabled during the read/write process so the exact bit-wise EMAC register contents are accessed. For example, a BDM read of an accumulator (ACCx) requires the following sequence: BdmReadACCx ( rcreg wcreg rcreg wcreg ) macsr; #0,macsr; ACCx; #saved_data,macsr; // // // // read current macsr contents & save disable all rounding modes read the desired accumulator restore the original macsr MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 8-47 Likewise, to write an accumulator register, the following BDM sequence is needed: BdmWriteACCx ( rcreg wcreg wcreg wcreg ) macsr; #0,macsr; #data,ACCx; #saved_data,macsr; // // // // read current macsr contents & save disable all rounding modes write the desired accumulator restore the original macsr Additionally, writes to the accumulator extension registers must be performed after the corresponding accumulators are updated because a write to any accumulator alters the corresponding extension register contents. For more information on saving and restoring the complete EMAC programming model, see the appropriate section of the EMAC chapter. 8.5.3.3.14 BDM Accesses of Floating-Point Data Registers (FPn) The ColdFire debug architecture allows BDM accesses of the entire programming model (including all FPU-related registers) of the processor core using RCREG and WCREG. However, certain hardware restrictions require the accesses related to the 64-bit FPn data registers be performed in a certain manner to guarantee correct operation. The serial BDM command structure supports 8-, 16- and 32-bit accesses, but there is no direct mechanism for accessing 64-bit data values. Rather than changing this well-established protocol and command set, BDM accesses of 64-bit data values are treated as two independent 32-bit references. In particular, 64-bit FPn data registers are treated as two separate values from the BDM perspective. Each FPn is partitioned into upper and lower longwords, FPUn and FPLn. Either longword can be read first. The processor treats the BDM read command as a pseudo-FMOVEM. Accordingly, all rounding modes and exception enables are ignored and the 32-bit contents of FPUn or FPLn are sent to the debug module for transmission over the serial communication channel. The FPU programming model is unchanged. To write to an FPU data register, FPUn must be written first and followed by a write to FPLn. The processor operates as follows: the BDM write to FPUn is performed, which loads the upper 32 bits of an internal double-precision operand register; the BDM write to FPLn loads the supplied operand into the lower 32 bits of the same internal register, and the entire 64-bit value is loaded into the selected FPn. Failure to execute this sequence of commands produces an undefined value in the FPUn. Note that any BDM write of an FPU register changes the internal state from NULL to IDLE. 8.5.3.3.15 Write Control Register (WCREG) The operand (longword) data is written to the specified control register. The write alters all 32 register bits. See the RCREG instruction description for the Rc encoding and for additional notes on writes to the A7 stack pointers and the EMAC and FPU programming models. Command/Result Formats: MCF548x Reference Manual, Rev. 5 8-48 Freescale Semiconductor Background Debug Mode (BDM) 15 12 Command 11 8 7 4 3 0 0x2 0x8 0x8 0x0 0x0 0x0 0x0 0x0 0x0 Rc Result D[31:16] D[15:0] Figure 8-45. WCREG Command/Result Formats Command Sequence: WCREG ??? MS ADDR ’NOT READY’ MS ADDR ’NOT READY’ MS DATA ’NOT READY’ LS DATA ’NOT READY’ WRITE CONTROL REGISTER XXX ’NOT READY’ NEXT CMD ’CMD COMPLETE’ XXX BERR NEXT CMD ’NOT READY’ Figure 8-46. WCREG Command Sequence Operand Data: Result Data: 8.5.3.3.16 This instruction requires two longword operands. The first selects the register to which the operand data is to be written; the second contains the data. Successful write operations return 0xFFFF. Bus errors on the write cycle are indicated by the setting of bit 16 in the status message and by a data pattern of 0x0001. Read Debug Module Register (RDMREG) Read the selected debug module register and return the 32-bit result. The only valid register selection for the RDMREG command is CSR (DRc = 0x00). Note that this read of the CSR clears the trigger status bits (CSR[BSTAT]) if either a level-2 breakpoint has been triggered or a level-1 breakpoint has been triggered and no level-2 breakpoint has been enabled. Command/Result Formats: MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 8-49 15 Command 12 11 0x2 8 7 5 0xD 4 0 100 Result DRc D[31:16] D[15:0] Figure 8-47. RDMREG BDM Command/Result Formats Table 8-27 shows the definition of DRc encoding. Table 8-27. Definition of DRc Encoding—Read DRc[4:0] Debug Register Definition Mnemonic Initial State Page 0x00 Configuration/Status CSR 0x0 p. 8-11 0x01–0x1F Reserved — — — Command Sequence: RDMREG ??? XXX MS RESULT NEXT CMD LS RESULT XXX ’ILLEGAL’ NEXT CMD ’NOT READY’ Figure 8-48. RDMREG Command Sequence Operand Data: Result Data: 8.5.3.3.17 None The contents of the selected debug register are returned as a longword value. The data is returned most-significant word first. Write Debug Module Register (WDMREG) The operand (longword) data is written to the specified debug module register. All 32 bits of the register are altered by the write. DSCLK must be inactive while the debug module register writes from the CPU accesses are performed using the WDEBUG instruction. Command Format: Figure 8-49. WDMREG BDM Command Format 15 12 0x2 11 8 7 0xC 5 100 4 0 DRc D[31:16] D[15:0] Table 8-6 shows the definition of the DRc write encoding. MCF548x Reference Manual, Rev. 5 8-50 Freescale Semiconductor Real-Time Debug Support Command Sequence: WDMREG ??? MS DATA ’NOT READY’ LS DATA ’NOT READY’ XXX ’ILLEGAL’ NEXT CMD ’NOT READY’ NEXT CMD ’CMD COMPLETE’ Figure 8-50. WDMREG Command Sequence Operand Data: Longword data is written into the specified debug register. The data is supplied most-significant word first. Command complete status (0xFFFF) is returned when register write is complete. Result Data: 8.6 Real-Time Debug Support The ColdFire Family provides support debugging real-time applications. For these types of embedded systems, the processor must continue to operate during debug. The foundation of this area of debug support is that while the processor cannot be halted to allow debugging, the system can generally tolerate the small intrusions of the BDM inserting instructions into the pipeline with minimal effect on real-time operation. The debug module provides three types of breakpoints: PC with mask, operand address range, and data with mask. These breakpoints can be configured into one- or two-level triggers with the exact trigger response also programmable. The debug module programming model can be written from either the external development system using the debug serial interface or from the processor’s supervisor programming model using the WDEBUG instruction. Only CSR is readable using the external development system. 8.6.1 Theory of Operation Breakpoint hardware can be configured through TDR[TCR] to respond to triggers by displaying PSTDDATA, initiating a processor halt, or generating a debug interrupt. As shown in Table 8-28, when a breakpoint is triggered, an indication (CSR[BSTAT]) is provided on the PSTDDATA output port of the DDATA information when it is not displaying captured processor status, operands, or branch addresses. See Section 8.3.2, “Processor Stopped or Breakpoint State Change (PST = 0xE).” Table 8-28. PSTDDATA Nibble/CSR[BSTAT] Breakpoint Response PSTDDATA Nibble/CSR[BSTAT] 1 1 Breakpoint Status 0000/0000 No breakpoints enabled 0010/0001 Waiting for level-1 breakpoint 0100/0010 Level-1 breakpoint triggered 1010/0101 Waiting for level-2 breakpoint 1100/0110 Level-2 breakpoint triggered Encodings not shown are reserved for future use. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 8-51 The breakpoint status is also posted in CSR. Note that CSR[BSTAT] is cleared by a CSR read when either a level-2 breakpoint is triggered or a level-1 breakpoint is triggered and a level-2 breakpoint is not enabled. Status is also cleared by writing to either TDR or XTDR to disable trigger options. BDM instructions use the appropriate registers to load and configure breakpoints. As the system operates, a breakpoint trigger generates the response defined in TDR. PC breakpoints are treated in a precise manner: exception recognition and processing are initiated before the excepting instruction is executed. All other breakpoint events are recognized on the processor’s local bus, but are made pending to the processor and sampled like other interrupt conditions. As a result, these interrupts are said to be imprecise. In systems that tolerate the processor being halted, a BDM-entry can be used. With TDR[TRC] = 01, a breakpoint trigger causes the core to halt (PST = 0xF). If the processor core cannot be halted, the debug interrupt can be used. With this configuration, TDR[TRC] = 10, the breakpoint trigger becomes a debug interrupt to the processor, which is treated higher than the nonmaskable level-7 interrupt request. As with all interrupts, it is made pending until the processor reaches a sample point, which occurs once per instruction. Again, the hardware forces the PC breakpoint to occur before the targeted instruction executes and is precise. This is possible because the PC breakpoint is enabled when interrupt sampling occurs. For address and data breakpoints, reporting is considered imprecise because several instructions may execute after the triggering address or data is detected. As soon as the debug interrupt is recognized, the processor aborts execution and initiates exception processing. This event is signaled externally by the assertion of a unique PST value (PST = 0xD) for multiple cycles. The core enters emulator mode when exception processing begins. After the standard 8-byte exception stack is created, the processor fetches a unique exception vector from the vector table. Table 8-29 describes the two unique entries that distinguish PC breakpoints from other trigger events. Table 8-29. Exception Vector Assignments Vector Number Vector Offset (Hex) Stacked Program Counter Assignment 12 0x030 Next Non-PC-breakpoint debug interrupt 13 0x034 Next PC-breakpoint debug interrupt (Refer to the ColdFire Programmer’s Reference Manual.) In the case of a two-level trigger, the last breakpoint event determines the exception vector; however, if the second-level trigger is PC || Address {&& Data} (as shown in the last condition in the code example in Section 8.4.11.1, “Resulting Set of Possible Trigger Combinations”), the vector taken is determined by the first condition that occurs after the first-level trigger: vector 13 if PC occurs first or vector 12 if Address {&& Data} occurs first. If both occur simultaneously, the non-PC-breakpoint debug interrupt is taken (vector number 12). Execution continues at the instruction address in the vector corresponding to the breakpoint triggered. The debug interrupt handler can use supervisor instructions to save the necessary context such as the state of all program-visible registers into a reserved memory area. During a debug interrupt service routine, all normal interrupt requests are evaluated and sampled once per instruction. If any exception occurs, the processor responds as follows: 1. It saves a copy of the current value of the emulator mode state bit and then exits emulator mode by clearing the actual state. MCF548x Reference Manual, Rev. 5 8-52 Freescale Semiconductor Real-Time Debug Support 2. Bit 1 of the fault status field (FS1) in the next exception stack frame is set to indicate the processor was in emulator mode when the interrupt occurred. This corresponds to bit 17 of the longword at the top of the system stack. See Section 3.8.1, “Exception Stack Frame Definition.” 3. It passes control to the appropriate exception handler. 4. It executes an RTE instruction when the exception handler finishes. During the processing of the RTE, FS1 is reloaded from the system stack. If this bit is set, the processor sets the emulator mode state and resumes execution of the original debug interrupt service routine. This is signaled externally by the generation of the PST value that originally identified the debug interrupt exception, that is, PST = 0xD. Fault status encodings are listed in Table 5-2. Implementation of this debug interrupt handling fully supports the servicing of a number of normal interrupt requests during a debug interrupt service routine. The emulator mode state bit is essentially changed to be a program-visible value, stored into memory during exception stack frame creation, and loaded from memory by the RTE instruction. When debug interrupt operations complete, the RTE instruction executes and the processor exits emulator mode. After the debug interrupt handler completes execution, the external development system can use BDM commands to read the reserved memory locations. In Revision A, if a hardware breakpoint such as a PC trigger is left unmodified by the debug interrupt service routine, another debug interrupt is generated after the completion of the RTE instruction. In Revisions B and C, the generation of another debug interrupt during the first instruction after the RTE exits emulator mode is inhibited. This behavior is consistent with the existing logic involving trace mode where the first instruction executes before another trace exception is generated. Thus, all hardware breakpoints are disabled until the first instruction after the RTE completes execution, regardless of the programmed trigger response. 8.6.1.1 Emulator Mode Emulator mode is used to facilitate nonintrusive emulator functionality. This mode can be entered in three different ways: • Setting CSR[EMU] forces the processor into emulator mode. EMU is examined only if RSTI is negated and the processor begins reset exception processing. It can be set while the processor is halted before reset exception processing begins. See Section 8.5.1, “CPU Halt.” • A debug interrupt always puts the processor in emulation mode when debug interrupt exception processing begins. • Setting CSR[TRC] forces the processor into emulation mode when trace exception processing begins. While operating in emulation mode, the processor exhibits the following properties: • Unmasked interrupt requests are serviced. The resulting interrupt exception stack frame has FS[1] set to indicate the interrupt occurred while in emulator mode. • If CSR[MAP] = 1, all caching of memory and the SRAM module are disabled. All memory accesses are forced into a specially mapped address space signaled by TT = 0x2, TM = 0x5 or 0x6. This includes stack frame writes and the vector fetch for the exception that forced entry into this mode. The RTE instruction exits emulation mode. The processor status output port provides a unique encoding for emulator mode entry (0xD) and exit (0x7). MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 8-53 8.6.2 Concurrent BDM and Processor Operation The debug module supports concurrent operation of both the processor and most BDM commands. BDM commands may be executed while the processor is running, except the following: • Read/write address and data registers • Read/write control registers For BDM commands that access memory, the debug module requests the processor’s local bus. The processor responds by stalling the instruction fetch pipeline and waiting for current bus activity to complete before freeing the local bus for the debug module to perform its access. After the debug module bus cycle, the processor reclaims the bus. NOTE Breakpoint registers must be carefully configured in a development system if the processor is executing. The debug module contains no hardware interlocks, so TDR and XTDR should be disabled while breakpoint registers are loaded, after which TDR and XTDR can be written to define the exact trigger. This prevents spurious breakpoint triggers. Because there are no hardware interlocks in the debug unit, no BDM operations are allowed while the CPU is writing the debug’s registers (DSCLK must be inactive). 8.7 Debug C Definition of PSTDDATA Outputs This section specifies the ColdFire processor and debug module’s generation of the PSTDDATA output on an instruction basis. In general, the PSTDDATA output for an instruction is defined as follows: PSTDDATA = 0x1, {[0x89B], operand} where the {...} definition is optional operand information defined by the setting of the CSR. The CSR provides capabilities to display operands based on reference type (read, write, or both). A PST value {0x8, 0x9, or 0xB} identifies the size and presence of valid data to follow on the PSTDDATA output {1, 2, or 4 bytes}. Additionally, for certain change-of-flow branch instructions, CSR[BTB] provides the capability to display the target instruction address on the PSTDDATA output {2, 3, or 4 bytes} using a PST value of {0x9, 0xA, or 0xB}. 8.7.1 User Instruction Set Table 8-30 shows the PSTDDATA specification for user-mode instructions. Rn represents any {Dn, An} register. In this definition, the ‘y’ suffix generally denotes the source and ‘x’ denotes the destination operand. For a given instruction, the optional operand data is displayed only for those effective addresses referencing memory. Table 8-30. PSTDDATA Specification for User-Mode Instructions Instruction Operand Syntax PSTDDATA add.l <ea>y,Dx PSTDDATA = 0x1,{0xB, source operand} add.l Dy,<ea>x PSTDDATA = 0x1,{0xB, source},{0xB, destination} adda.l <ea>y,Ax PSTDDATA = 0x1,{0xB, source operand} addi.l #<data>,Dx PSTDDATA = 0x1 MCF548x Reference Manual, Rev. 5 8-54 Freescale Semiconductor Debug C Definition of PSTDDATA Outputs Table 8-30. PSTDDATA Specification for User-Mode Instructions (Continued) Instruction Operand Syntax PSTDDATA addq.l #<data>,<ea>x PSTDDATA = 0x1,{0xB, source},{0xB, destination} addx.l Dy,Dx PSTDDATA = 0x1 and.l <ea>y,Dx PSTDDATA = 0x1,{0xB, source operand} and.l Dy,<ea>x PSTDDATA = 0x1,{0xB, source},{0xB, destination} andi.l #<data>,Dx PSTDDATA = 0x1 asl.l {Dy,#<data>},Dx PSTDDATA = 0x1 asr.l {Dy,#<data>},Dx PSTDDATA = 0x1 bcc.{b,w,l} if taken, then PSTDDATA = 0x5, else PSTDDATA = 0x1 bchg.{b,l} #<data>,<ea>x PSTDDATA = 0x1,{0x8, source},{0x8, destination} bchg.{b,l} Dy,<ea>x PSTDDATA = 0x1,{0x8, source},{0x8, destination} bclr.{b,l} #<data>,<ea>x PSTDDATA = 0x1,{0x8, source},{0x8, destination} bclr.{b,l} Dy,<ea>x PSTDDATA = 0x1,{0x8, source},{0x8, destination} bra.{b,w,l} PSTDDATA = 0x5 bset.{b,l} #<data>,<ea>x PSTDDATA = 0x1,{0x8, source},{0x8, destination} bset.{b,l} Dy,<ea>x PSTDDATA = 0x1,{0x8, source},{0x8, destination} bsr.{b,w,l} PSTDDATA = 0x5,{0xB, destination operand} btst.{b,l} #<data>,<ea>x PSTDDATA = 0x1,{0x8, source operand} btst.{b,l} Dy,<ea>x PSTDDATA = 0x1,{0x8, source operand} clr.b <ea>x PSTDDATA = 0x1,{0x8, destination operand} clr.l <ea>x PSTDDATA = 0x1,{0xB, destination operand} clr.w <ea>x PSTDDATA = 0x1,{0x9, destination operand} cmp.b <ea>y,Dx PSTDDATA = 0x1, {0x8, source operand} cmp.l <ea>y,Dx PSTDDATA = 0x1,{0xB, source operand} cmp.w <ea>y,Dx PSTDDATA = 0x1, {0x9, source operand} cmpa.l <ea>y,Ax PSTDDATA = 0x1,{0xB, source operand} cmpa.w <ea>y,Ax PSTDDATA = 0x1, {0x9, source operand} cmpi.b #<data>,Dx PSTDDATA = 0x1 cmpi.l #<data>,Dx PSTDDATA = 0x1 cmpi.w #<data>,Dx PSTDDATA = 0x1 divs.l <ea>y,Dx PSTDDATA = 0x1,{0xB, source operand} divs.w <ea>y,Dx PSTDDATA = 0x1,{0x9, source operand} divu.l <ea>y,Dx PSTDDATA = 0x1,{0xB, source operand} divu.w <ea>y,Dx PSTDDATA = 0x1,{0x9, source operand} MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 8-55 Table 8-30. PSTDDATA Specification for User-Mode Instructions (Continued) Instruction Operand Syntax PSTDDATA eor.l Dy,<ea>x PSTDDATA = 0x1,{0xB, source},{0xB, destination} eori.l #<data>,Dx PSTDDATA = 0x1 ext.l Dx PSTDDATA = 0x1 ext.w Dx PSTDDATA = 0x1 extb.l Dx PSTDDATA = 0x1 PSTDDATA = 0x11 illegal jmp <ea>y PSTDDATA = 0x5, {[0x9AB], target address} 2 jsr <ea>y PSTDDATA = 0x5, {[0x9AB], target address},{0xB , destination operand}2 lea.l <ea>y,Ax PSTDDATA = 0x1 link.w Ay,#<displacement> PSTDDATA = 0x1,{0xB, destination operand} lsl.l {Dy,#<data>},Dx PSTDDATA = 0x1 lsr.l {Dy,#<data>},Dx PSTDDATA = 0x1 mov3q.l #<data>,<ea>x PSTDDATA = 0x1, {0xB, destination operand} move.b <ea>y,<ea>x PSTDDATA = 0x1,{0x8, source},{0x8, destination} move.l <ea>y,<ea>x PSTDDATA = 0x1,{0xB, source},{0xB, destination} move.w <ea>y,<ea>x PSTDDATA = 0x1,{0x9, source},{0x9, destination} move.w CCR,Dx PSTDDATA = 0x1 move.w {Dy,#<data>},CCR PSTDDATA = 0x1 movea.l <ea>y,Ax PSTDDATA = 0x1,{0xB, source} movea.w <ea>y,Ax PSTDDATA = 0x1,{0x9, source} movem.l #list,<ea>x PSTDDATA = 0x1,{0xB, destination},... 3 movem.l <ea>y,#list PSTDDATA = 0x1,{0xB, source},... 3 moveq.l #<data>,Dx PSTDDATA = 0x1 muls.l <ea>y,Dx PSTDDATA = 0x1,{0xB, source operand} muls.w <ea>y,Dx PSTDDATA = 0x1,{0x9, source operand} mulu.l <ea>y,Dx PSTDDATA = 0x1,{0xB, source operand} mulu.w <ea>y,Dx PSTDDATA = 0x1,{0x9, source operand} mvs.b <ea>y,Dx PSTDDATA = 0x1, {0x8, source operand} mvs.w <ea>y,Dx PSTDDATA = 0x1, {0x9, source operand} mvz.b <ea>y,Dx PSTDDATA = 0x1, {0x8, source operand} mvz.w <ea>y,Dx PSTDDATA = 0x1, {0x9, source operand} neg.l Dx PSTDDATA = 0x1 negx.l Dx PSTDDATA = 0x1 MCF548x Reference Manual, Rev. 5 8-56 Freescale Semiconductor Debug C Definition of PSTDDATA Outputs Table 8-30. PSTDDATA Specification for User-Mode Instructions (Continued) Instruction Operand Syntax nop PSTDDATA PSTDDATA = 0x1 not.l Dx PSTDDATA = 0x1 or.l <ea>y,Dx PSTDDATA = 0x1,{0xB, source operand} or.l Dy,<ea>x PSTDDATA = 0x1,{0xB, source},{0xB, destination} ori.l #<data>,Dx PSTDDATA = 0x1 pea.l <ea>y PSTDDATA = 0x1,{0xB, destination operand} pulse PSTDDATA = 0x4 rems.l <ea>y,Dw:Dx PSTDDATA = 0x1,{0xB, source operand} remu.l <ea>y,Dw:Dx PSTDDATA = 0x1,{0xB, source operand} rts PSTDDATA = 0x1, PSTDDATA = 0x5, {[0x9AB], target address} sats.l Dx PSTDDATA = 0x1 scc.b Dx PSTDDATA = 0x1 sub.l <ea>y,Dx PSTDDATA = 0x1,{0xB, source operand} sub.l Dy,<ea>x PSTDDATA = 0x1,{0xB, source},{0xB, destination} suba.l <ea>y,Ax PSTDDATA = 0x1,{0xB, source operand} subi.l #<data>,Dx PSTDDATA = 0x1 subq.l #<data>,<ea>x PSTDDATA = 0x1,{0xB, source},{0xB, destination} subx.l Dy,Dx PSTDDATA = 0x1 swap.w Dx PSTDDATA = 0x1 tas.b <ea>x PSTDDATA = 0x1, {0x8, source}, {0x8, destination} tpf PST = 0x1 tpf.l #<data> PST = 0x1 tpf.w #<data> PST = 0x1 trap #<data> PSTDDATA = 0x11 tst.b <ea>x PSTDDATA = 0x1,{0x8, source operand} tst.l <ea>y PSTDDATA = 0x1,{0xB, source operand} tst.w <ea>y PSTDDATA = 0x1,{0x9, source operand} unlk Ax PSTDDATA = 0x1,{0xB, destination operand} wddata.b <ea>y PSTDDATA = 0x4, {0x8, source operand wddata.l <ea>y PSTDDATA = 0x4, {0xB, source operand wddata.w <ea>y PSTDDATA = 0x4, {0x9, source operand MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 8-57 1 During normal exception processing, the PSTDDATA output is driven to a 0xC indicating the exception processing state. The exception stack write operands, as well as the vector read and target address of the exception handler may also be displayed. Exception ProcessingPSTDDATA = 0xC,{0xB,destination},// stack frame {0xB,destination},// stack frame {0xB,source},// vector read PSTDDATA = 0x5,{[0x9AB],target}// handlerPC The PSTDDATA specification for the reset exception is shown below: Exception ProcessingPSTDDATA = 0xC, PSTDDATA = 0x5,{[0x9AB],target}// handlerPC The initial references at address 0 and 4 are never captured nor displayed since these accesses are treated as instruction fetches. For all types of exception processing, the PSTDDATA = 0xC value is driven at all times, unless the PSTDDATA output is needed for one of the optional marker values or for the taken branch indicator (0x5). 2 For JMP and JSR instructions, the optional target instruction address is displayed only for those effective address fields defining variant addressing modes. This includes the following <ea>x values: (An), (d16,An), (d8,An,Xi), (d8,PC,Xi). 3 For Move Multiple instructions (MOVEM), the processor automatically generates line-sized transfers if the operand address reaches a 0-modulo-16 boundary and there are four or more registers to be transferred. For these line-sized transfers, the operand data is never captured nor displayed, regardless of the CSR value. The automatic line-sized burst transfers are provided to maximize performance during these sequential memory access operations. Table 8-31 shows the PSTDDATA specification for multiply-accumulate instructions. Table 8-31. PSTDDATA Values for User-Mode Multiply-Accumulate Instructions Instruction Operand Syntax PSTDDATA mac.l Ry,Rx PSTDDATA = 0x1 mac.l Ry,Rx,<ea>y,Rw,ACCx PSTDDATA = 0x1,{0xB, source operand} mac.l Ry,Rx,ACCx PSTDDATA = 0x1 mac.l Ry,Rx,ea,Rw PSTDDATA = 0x1,{0xB, source operand} mac.w Ry,Rx PSTDDATA = 0x1 mac.w Ry,Rx,<ea>y,Rw,ACCx PSTDDATA = 0x1,{0xB, source operand} mac.w Ry,Rx,ACCx PSTDDATA = 0x1 mac.w Ry,Rx,ea,Rw PSTDDATA = 0x1,{0xB, source operand} move.l {Ry,#<data>},ACCext01 PSTDDATA = 0x1 move.l {Ry,#<data>},ACCext23 PSTDDATA = 0x1 move.l {Ry,#<data>},ACCx PSTDDATA = 0x1 move.l {Ry,#<data>},MACSR PSTDDATA = 0x1 move.l {Ry,#<data>},MASK PSTDDATA = 0x1 move.l ACCext01,Rx PSTDDATA = 0x1 move.l ACCext23,Rx PSTDDATA = 0x1 move.l ACCy,ACCx PSTDDATA = 0x1 MCF548x Reference Manual, Rev. 5 8-58 Freescale Semiconductor Debug C Definition of PSTDDATA Outputs Table 8-31. PSTDDATA Values for User-Mode Multiply-Accumulate Instructions (Continued) Instruction Operand Syntax PSTDDATA move.l ACCy,Rx PSTDDATA = 0x1 move.l MACSR,CCR PSTDDATA = 0x1 move.l MACSR,Rx PSTDDATA = 0x1 move.l MASK,Rx PSTDDATA = 0x1 msac.l Ry,Rx PSTDDATA = 0x1 msac.l Ry,Rx,<ea>y,Rw,ACCx PSTDDATA = 0x1,{0xB, source operand} msac.l Ry,Rx,ACCx PSTDDATA = 0x1 msac.l Ry,Rx,<ea>y,Rw PSTDDATA = 0x1,{0xB, source},{0xB, destination} msac.w Ry,Rx PSTDDATA = 0x1 msac.w Ry,Rx,<ea>y,Rw,ACCx PSTDDATA = 0x1,{0xB, source operand} msac.w Ry,Rx,ACCx PSTDDATA = 0x1 msac.w Ry,Rx,<ea>y,Rw PSTDDATA = 0x1,{0xB, source},{0xB, destination} Table 8-32 shows the PSTDDATA specification for floating-point instructions; note that <ea>y includes FPy, Dy, Ay, and <mem>y addressing modes. The optional operand capture and display applies only to the <mem>y addressing modes. Note also that the PSTDDATA values are the same for a given instruction, regardless of explicit rounding precision. Table 8-32. PSTDDATA Values for User-Mode Floating-Point Instructions Instruction 1 Operand Syntax PSTDDATA fabs.sz <ea>y,FPx PSTDDATA = 0x1, [89B], source} fadd.sz <ea>y,FPx PSTDDATA = 0x1, [89B], source} fbcc.{w,l} <label> if taken, then PSTDDATA = 5, else PSTDDATA = 0x1 fcmp.sz <ea>y,FPx PSTDDATA = 0x1, [89B], source} fdiv.sz <ea>y,FPx PSTDDATA = 0x1, [89B], source} fint.sz <ea>y,FPx PSTDDATA = 0x1, [89B], source} fintrz.sz <ea>y,FPx PSTDDATA = 0x1, [89B], source} fmove.sz <ea>y,FPx PSTDDATA = 0x1, [89B], source} fmove.sz FPy,<ea>x PSTDDATA = 0x1, [89B], destination} fmove.l <ea>y,FP*R PSTDDATA = 0x1, B, source} fmove.l FP*R,<ea>x PSTDDATA = 0x1, B, destination} fmovem <ea>y,#list PSTDDATA = 0x1 fmovem #list,<ea>x PSTDDATA = 0x1 fmul.sz <ea>y,FPx PSTDDATA = 0x1, [89B], source} MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 8-59 Table 8-32. PSTDDATA Values for User-Mode Floating-Point Instructions (Continued) Instruction 1 fneg.sz Operand Syntax PSTDDATA <ea>y,FPx PSTDDATA = 0x1, [89B], source} fnop PSTDDATA = 0x1 fsqrt.sz <ea>y,FPx PSTDDATA = 0x1, [89B], source} fsub.sz <ea>y,FPx PSTDDATA = 0x1, [89B], source} ftst.sz <ea>y PSTDDATA = 0x1, [89B], source} 1 The FP*R notation refers to the floating-point control registers: FPCR, FPSR, and FPIAR. Depending on the size of any external memory operand specified by the f<op>.fmt field, the data marker is defined as shown in Table 8-33. Table 8-33. Data Markers and FPU Operand Format Specifiers 8.7.2 Format Specifier Data Marker .b 8 .w 9 .l B .s B .d Never captured Supervisor Instruction Set The supervisor instruction set has complete access to the user mode instructions plus the opcodes shown below. The PSTDDATA specification for these opcodes is shown in Table 8-34. Table 8-34. PSTDDATA Specification for Supervisor-Mode Instructions Instruction Operand Syntax PSTDDATA cpushl dc,(Ax) ic,(Ax) bc,(Ax) PSTDDATA = 0x1 frestore <ea>y PSTDDATA = 0x1 fsave <ea>x PSTDDATA = 0x1 halt PSTDDATA = 0x1, PSTDDATA = 0xF intouch (Ay) PSTDDATA = 0x1 move.l Ay,USP PSTDDATA = 0x1 move.l USP,Ax PSTDDATA = 0x1 move.w SR,Dx PSTDDATA = 0x1 move.w {Dy,#<data>},SR PSTDDATA = 0x1, {0x3} MCF548x Reference Manual, Rev. 5 8-60 Freescale Semiconductor ColdFire Debug History Table 8-34. PSTDDATA Specification for Supervisor-Mode Instructions (Continued) Instruction movec.l Operand Syntax Ry,Rc rte PSTDDATA PSTDDATA = 0x1, {8, ASID} PSTDDATA = 0x7, {0xB, source operand}, {3},{0xB, source operand}, {DD}, PSTDDATA = 0x5, {[0x9AB], target address} stop #<data> PSTDDATA = 0x1, PSTDDATA = 0xE wdebug.l <ea>y PSTDDATA = 0x1, {0xB, source, 0xB, source} The move-to-SR and RTE instructions include an optional PSTDDATA = 0x3 value, indicating an entry into user mode. Additionally, if the execution of a RTE instruction returns the processor to emulator mode, a multiple-cycle status of 0xD is signaled. Similar to the exception processing mode, the stopped state (PSTDDATA = 0xE) and the halted state (PSTDDATA = 0xF) display this status throughout the entire time the ColdFire processor is in the given mode. 8.8 ColdFire Debug History This section describes the origins of the ColdFire debug systems. 8.8.1 ColdFire Debug Classic: The Original Definition The original design, Revision A, provided debug support in three separate areas: • Real-time trace • Background debug mode (BDM) • Real-time debug The real-time debug features may be accessed from the external BDM emulator or from the supervisor programming model of the processor. The hardware breakpoint registers include: a PC breakpoint + mask, two address registers for defining a specific address or a range of addresses, and a data breakpoint + mask. The original design supported breakpoints of the form: if PC_breakpoint is triggered then respond using user-defined configuration if Address_breakpoint {&& Data_breakpoint} is triggered then respond using user-defined configuration Two-level triggers of the form: if PC_breakpoint is triggered then if Address_breakpoint {&& Data_breakpoint} is triggered then respond using user-defined configuration if Address_breakpoint {&& Data_breakpoint} is triggered then if PC_breakpoint is triggered then respond using user-defined configuration MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 8-61 The data_breakpoint can be included as an optional part of an address breakpoint. The ColdFire debug architecture was created to provide this set of functionality without requiring the traditional connection to the external system bus. Rather, the functionality is provided using only a connection to a Freescale-defined 26-pin debug connector. By providing the required debug signals in customer-specific designs, standard third-party emulators can be used for debug of these designs. NOTE The baseline debug functionality is described in any of the ColdFire MCF52xx User’s Manuals, which are available as PDF files at: http://www.freescale.com/ColdFire/. As an example, see the debug section of the MCF5272 User’s Manual located under MCF5272 Product Information. 8.8.2 ColdFire Debug Revision B During development of the Version 3 ColdFire design, there were a number of enhancements to the original debug functionality requested by customers and third-party developers. These requests resulted in an expanded set of debug functionality named Revision B. The Rev. B enhancements are as follows: • Addition of a BDM SYNC_PC command to display the processor’s current PC • Creation of more flexible hardware breakpoint triggers, i.e., support for “OR” combinations • Removal of the restrictions involving concurrent hardware breakpoint use and BDM command activity • Redefinition of the processor status values for the RTS instruction • An external mechanism to generate a debug interrupt • A mechanism to inhibit debug interrupts after the RTE exit • A mechanism to identify the revision level of the debug module Rev. B enhancements provide backward compatibility with the original design. 8.8.3 ColdFire Debug Revision C Continuing discussions with customers and the developer community led to Revision C design enhancements primarily related to improvements in the real-time debug capabilities of the ColdFire architecture. The remainder of this section details these enhancements. 8.8.3.1 Debug Interrupts and Interrupt Requests (Emulator Mode) In Rev. A and Rev. B ColdFire debug implementations, the response to a user-defined breakpoint trigger can be configured to be one of three possibilities: • The breakpoint trigger can merely be displayed on the DDATA bus, with no internal reaction to the trigger. The trigger state information is displayed on DDATA in all situations. • The breakpoint trigger can force the processor to halt and allow BDM activities. • The breakpoint trigger can generate a special debug interrupt to allow real-time systems to quickly process the interrupt and return to normal system executing as rapidly as possible. The occurrence of the debug interrupt exception is treated as a special type of interrupt. It is considered to be higher in priority than all normal interrupt requests and has special processor status values to provide an external indication that this interrupt has occurred. MCF548x Reference Manual, Rev. 5 8-62 Freescale Semiconductor Freescale-Recommended BDM Pinout Additionally, the execution of the debug interrupt service routine is forced to be interrupt-inhibited by the processor hardware. While in this service routine, there is an optional capability to map all instruction and operand references into a separate address space, so that an emulator could define the routine dynamically. The current processor implementations actually include a program-invisible state bit that defines this emulator mode of operation. Also note, the interrupt mask level is not modified during the processing of a debug interrupt. Customers with real-time embedded systems have specifically asked for the ability to service normal interrupt requests while processing the debug interrupt service routine. In many systems of this type, motion-based servo interrupts must be considered as the highest priority interrupt request. To provide this functionality and be able to service any number of normal interrupt requests (including the possibility of nested interrupts), the processor state signaling emulator mode must be included as part of the exception stack frame. As part of the Rev. C functionality, the operation of the debug interrupt is modified in the following manner: 1. The occurrence of the breakpoint trigger, configured to generate a debug interrupt, is treated exactly as before. The debug interrupt is treated as a higher priority exception relative to the normal interrupt requests encoded on the interrupt priority input signals. 2. At the appropriate sample point, the ColdFire processor initiates debug interrupt exception processing. This event is signaled externally by the generation of a unique PST value (PST = 0xD) asserted for multiple cycles. The processor sets the emulator mode state bit as part of this exception processing. 3. While the processor in the debug interrupt service routine, all normal interrupt requests are evaluated and sampled once per instruction. While in this routine, if any type of exception occurs, the processor responds in the following manner: a) In response to the new exception, the processor saves a copy of the current value of the emulator mode state bit and then exits emulator mode by clearing the actual state. b) The new exception stack frame sets bit 1 of the fault status field, using the saved emulator mode bit, indicating execution while in emulator mode has been interrupted. This corresponds to bit 17 of the longword at the top of the system stack. c) Control is passed to the appropriate exception handler. d) When the exception handler is complete, a Return From Exception (RTE) instruction is executed. During the processing of the RTE, FS[1] is reloaded from the system stack. If this bit is asserted, the processor sets the emulator mode state and resumes execution of the original debug interrupt service routine. This is signaled externally by the generation of the PST value that originally identified the occurrence of a debug interrupt exception, that is, PST = 0xD. Implementation of this revised debug interrupt handling fully supports the servicing of any number of normal interrupt requests while in a debug interrupt service routine. The emulator mode state bit is essentially changed to be a program-visible value, stored into memory during exception stack frame creation and loaded from memory by the RTE instruction. 8.9 Freescale-Recommended BDM Pinout The ColdFire BDM connector, Figure 8-51, is a 26-pin Berg connector arranged 2 x 13. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 8-63 Developer reserved 1 1 2 BKPT GND 3 4 DSCLK GND 5 6 Developer reserved1 RESET 7 8 DSI VDD_IO 2 9 10 DSO GND 11 12 PSTDDATA7 PSTDDATA6 13 14 PSTDDATA5 PSTDDATA4 15 16 PSTDDATA3 PSTDDATA2 17 18 PSTDDATA1 PSTDDATA0 19 20 GND Freescale reserved 21 22 Freescale reserved GND 23 24 PSTCLK VDD_CPU 25 26 TA 1 Pins 2 reserved for BDM developer use. Supplied by target. Figure 8-51. Recommended BDM Connector MCF548x Reference Manual, Rev. 5 8-64 Freescale Semiconductor Part II System Integration Unit Part II describes the system integration unit, which provides overall control of the bus and serves as the interface between the ColdFire core processor complex and internal peripheral devices. It includes a general description of the SIU and individual chapters that describe components of the SIU, such as the interrupt controller, general purpose timers, slice timers, and GPIOs. Contents Part II contains the following chapters: • Chapter 9, “System Integration Unit (SIU),” describes the SIU programming model, bus arbitration, and system-protection functions for the MCF548x. • Chapter 10, “Internal Clocks and Bus Architecture,” describes the clocking and internal buses of the MCF548x and discusses the main functional blocks controlling the XL bus and the XL bus arbiter • Chapter 11, “General Purpose Timers (GPT),” describes the functionality of the four general purpose timers, GPT0–GPT3. • Chapter 12, “Slice Timers (SLT),” describes the two slice timers, shorter term periodic interrupts, used in the MCF548x. • Chapter 13, “Interrupt Controller,” describes operation of the interrupt controller portion of the SIU. It includes descriptions of the registers in the interrupt controller memory map and the interrupt priority scheme. • Chapter 14, “Edge Port Module (EPORT),” describes EPORT module functionality. • Chapter 15, “GPIO,” describes the operation and programming model of the parallel port pin assignment, direction-control, and data registers. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor i MCF548x Reference Manual, Rev. 5 ii Freescale Semiconductor Chapter 9 System Integration Unit (SIU) 9.1 Introduction The system integration unit (SIU) of the MCF548x family integrates several timer functions required by most embedded systems. The SIU contains the following components: • Slice timers • Watchdog timer • General purpose timers • General purpose I/O ports • Interrupt controller Two internal 32-bit slice timers are provided to create short cycle periodic interrupts, typically utilized for RTOS scheduling and alarm functionality. A watchdog timer is included that will reset the processor if not regularly serviced, catching software hang-ups. Up to four 32-bit general purpose timers are included, which are capable of input capture, output compare, and PWM functionality. Most peripheral I/O pins on the MCF548x family are muxed with GPIO, adding flexibility and usability to pins on the chip. The programmable interrupt controller multiplexes the external interrupts, general purpose timers, slice timers, and peripheral sources to the CF4e core. Refer to Chapter 13, “Interrupt Controller,” for information about the MCF548x interrupt controller. The SIU timers are discussed in the following chapters: • General purpose timers and watchdog timer (GPT0) are described in Chapter 11, “General Purpose Timers (GPT).” — The watchdog timer is further detailed in Section 10.3.2.3, “Watchdog Functions.” • Slice timers are detailed in Chapter 12, “Slice Timers (SLT).” • GPIO functionality is discussed in Chapter 15, “GPIO.” 9.2 Features The system integration unit has the following features: • Interrupt controller • Two 32-bit slice timers for periodic alarm and interrupt generation • Software watchdog timer with programmable secondary bus monitor • Up to four 32-bit general-purpose timers with capture, compare, and PWM capability • General-purpose I/O ports multiplexed with peripheral pins • System protection and reset status and control 9.3 Memory Map/Register Definition Table 9-1 shows the programming model for the SIU. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 9-1 Table 9-1. SIU Register Map Address (MBAR +) Name CPU+0xC0F Module Base Address Register 0x04 Byte0 SDRAM Drive Strength Register Byte3 Access R/W 1 SDRAMDS R/W SBCR R/W Reserved System Breakpoint Control Register 0x1–0x1C Reserved 0x20 SDRAM Chip Select 0 Configuration Register1 CS0CFG01 R/W 0x24 SDRAM Chip Select 1 Configuration Register1 CS1CFG11 R/W 0x28 SDRAM Chip Select 2 Configuration Register1 CS2CFG21 R/W 0x2C SDRAM Chip Select 3 Configuration Register1 CS3CFG31 R/W 0x30–0x34 0x38 RESERVED Sequential Access Control Register 0x3C–0x40 0x44 0x50 SECSACR R/W RSR R/W JTAGID R RESERVED Reset Status Register 0x48–0x4C 1 Byte2 MBAR 1 0x08–0x0C 0x10 Byte1 RESERVED JTAG Device Identification Number The SDRAM Drive Strength and Chip Select Configuration registers are discussed in Chapter 18, “SDRAM Controller (SDRAMC).” They are shown in this memory map for reference purposes. 9.3.1 Module Base Address Register (MBAR) The supervisor-level MBAR, Figure 9-1, specifies the base address and allowable access types for all internal peripherals. It is written with a MOVEC instruction using the CPU address 0xC0F (refer to the ColdFire Family Programmer’s Reference Manual). MBAR can be read or written through the debug modules as a read/write register, as described in Chapter 8, “Debug Support.” Only the debug module can read MBAR. The MBAR is initialized to 0x8000_0000 at reset; however, it can be relocated to a new base address. To access internal peripherals, write MBAR with the appropriate base address (BA) after system reset. All internal peripheral registers occupy a single relocatable memory block along 256-KByte boundaries. MBAR[BA] is compared to the upper 14 bits of the full 32-bit internal address to determine if an internal peripheral is being accessed. Any accesses in this range, whether to a valid peripheral address or not, will be made internally rather than using the external bus. NOTE The MBAR region must be mapped to non-cacheable space. MCF548x Reference Manual, Rev. 5 9-2 Freescale Semiconductor Memory Map/Register Definition 31 30 29 28 27 26 25 R 24 23 22 21 20 19 18 BA 17 16 0 0 W Reset R 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 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 0 0 W Reset Reg Addr CPU + 0xC0F Figure 9-1. Module Base Address Register (MBAR) 9.3.1.1 System Breakpoint Control Register (SBCR) The System Breakpoint Control Register allows for discrete control over functionality of the BKPT signal. The assertion of the BKPT signal can be programmed to halt the core, DMA, and DSPI or any combination. In addition, a halt condition in the DMA can be programmed to halt the CPU, or a halt in the CPU can halt the DMA. 31 R PIN2 CPU W Reset R 30 29 28 27 PIN2 CPU2 DMA2 PIN2 DMA DMA CPU DSPI 26 25 24 23 22 21 20 19 18 17 16 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 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 0 0 W Reset Reg Addr MBAR + 0x0010 Figure 9-2. System Breakpoint Control Register (SBCR) Table 9-2. SBCR Field Descriptions Bit Name Description 31 PIN2CPU Pin control of the ColdFire V4e breakpoint. This bit controls whether the BKPT pin can halt the ColdFire V4e. 0 The assertion of BKPT will not halt the ColdFire V4e core. 1 The assertion of BKPT will halt the ColdFire V4e core. 30 PIN2DMA Pin control of the multichannel DMA breakpoint. This bit controls whether the BKPT pin can halt the DMA. 0 The assertion of BKPT will not halt the DMA. 1 The assertion of BKPT will halt the DMA. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 9-3 Table 9-2. SBCR Field Descriptions (Continued) Bit Name Description 29 CPU2DMA ColdFire V4e control of the multichannel DMA breakpoint. This bit controls whether a ColdFire V4e halt condition causes the assertion of the DMA breakpoint. 0 A ColdFire V4e halt condition will not halt the DMA. 1 A ColdFire V4e halt condition will halt the DMA. 28 DMA2CPU DMA control of the ColdFire V4e breakpoint. This bit controls whether a DMA halt condition causes the assertion of the ColdFire V4e breakpoint. 0 A DMA halt condition will not halt the ColdFire V4e. 1 A DMA halt condition will halt the ColdFire V4e. 27 PIN2DSPI 26-0 — 9.3.1.2 Pin control of the DSPI breakpoint. This bit controls whether the BKPT pin can halt the DSPI. 0 The assertion of BKPT will not halt the DSPI. 1 The assertion of BKPT will halt the DSPI. Reserved, should be cleared. SEC Sequential Access Control Register (SECSACR) This register is used to control bus accesses to the SEC module. If a sequential accesses to the SEC are enabled, then data will be buffered to create a single 64-bit access to the SEC instead of splitting up the transfer into two longwords. This can help to improve overall SEC performance. R 31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 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 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 SEQEN 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 W Reset R W Reset Reg Addr MBAR + 0x38 Figure 9-3. SEC Sequential Access Control Register (SECSACR) Table 9-3. SECSACR Field Descriptions Bits Name 31–1 — 0 SEQEN Description Reserved SEC Sequential access enable. 0 SEC Sequential Access is disabled. 1 SEC Sequential Access is enabled. Note: Setting this bit is recommended when the SEC is in use. MCF548x Reference Manual, Rev. 5 9-4 Freescale Semiconductor Memory Map/Register Definition 9.3.1.3 Reset Status Register (RSR) RSR allows the software, particularly the reset exception service routine, to know what type of reset has been asserted. When a reset signal is asserted, the associated status bit is set, and it maintains its value until the software explicitly clears the bit. 31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 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 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 0 0 0 0 0 0 0 0 0 0 0 RST JTG 0 RST WD RST 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 R W Reset R W Reset Reg Addr MBAR + 0x44 Figure 9-4. Reset Status Register (RSR) Table 9-4. RSR Field Descriptions Bits Name 31–4 — 3 RSTJTG 2 — 1 RSTWD General purpose watchdog timer reset asserted. Cleared by writing 1 to this bit position or by external reset. 0 RST External reset (PLL Lock qualification) asserted. Cleared by writing a 1 to this bit position. 9.3.1.4 Description Reserved, should be cleared. JTAG reset asserted. Cleared by writing 1 to this bit position or by external reset. Reserved, should be cleared. JTAG Device Identification Number (JTAGID) 31 30 29 28 27 26 25 R 24 23 22 21 20 19 18 17 16 6 5 4 3 2 1 0 JTAGID W Reset See Table 9-5 15 14 R 13 12 11 10 9 8 7 JTAGID W Reset See Table 9-5 Reg Addr MBAR + 0x50 Figure 9-5. JTAG Device ID Register (JTAGID) MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 9-5 Table 9-5. JTAGID Field Descriptions Bits Name Description 31–0 JTAGID The JTAG Identification Number Register is a read only register which contains the JTAG ID number for the MCF548x. Its value is hard coded and cannot be modified. Values for the MCF548x are the following: MCF5485 0x0800c01d MCF5484 0x0800d01d MCF5483 0x0800e01d MCF5482 0x0800f01d MCF5481 0x0801001d MCF5480 0x0801101d MCF548x Reference Manual, Rev. 5 9-6 Freescale Semiconductor Chapter 10 Internal Clocks and Bus Architecture 10.1 Introduction This chapter describes the clocking and internal buses of the MCF548x and discusses the main functional blocks controlling the XL bus and the XL bus arbiter. 10.1.1 Block Diagram Figure 10-1 shows a top-level block diagram of the MCF548x products. ColdFire V4e Core FPU, MMU EMAC 32K D-cache 32K I-cache PLL DDR SDRAM Interface FlexBus Interface XL Bus Arbiter Memory Controller FlexBus Controller Cryptography Accelerator*** XL Bus Read/Write Write DMA DMA Bus Read 32K System SRAM GP Timers x 4 PCI 2.2 Controller Multichannel DMA Master Bus Interface and FIFOs FlexCAN x2 PCI Interface & FIFOs CommBus DSPI I2C PSC x 4 FEC0 FEC1** Perpheral Communications I/O Interface & Ports USB 2.0 DEVICE* Communications I/O Subsystem Slice Timers x 2 PCI I/O Interface and Ports Watchdog Timer R/W Master/Slave Interface Crypto Interrupt Controller Slave Perpheral I/O Interface & Ports System Integration Unit XL Bus USB 2.0 PHY* *Available in MCF5485, MCF5484, MCF5483, and MCF5482 devices. **Available in MCF5485, MCF5484, MCF5481, and MCF5480 devices. ***Available in MCF5485, MCF5483, and MCF5481 devices. Figure 10-1. MCF548x Internal Bus Architecture MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 10-1 10.1.2 Clocking Overview The MCF548x requires a clock generated externally to be input to the CLKIN signal. The MCF548x uses this clock as the reference clock for the internal PLL. The internal PLL then generates the clocks needed by the CPU core and integrated peripherals. The external PCI and FlexBus signals are always clocked at the same frequency as the CLKIN signal. A programmable clock multiplier (determined by the AD[12:8] signals at reset) is used to determine the XL bus frequency. All integrated peripherals and the 32KB system SRAM are clocked at the same frequency as the XLB. The ColdFire V4e core complex (core, MMU, FPU, SRAMs, etc.) is always clocked at twice the XLB frequency. Table 10-1 shows the supported PLL encodings and the corresponding clock frequency ranges. Table 10-1. MCF548x Divide Ratio Encodings 1 AD[12:8]1 Clock Ratio CLKIN–PCI and FlexBus Frequency Range (MHz) Internal XLB, SDRAM bus, and PSTCLK Frequency Range (MHz) Core Frequency Range (MHz) 00011 1:2 41.67–50.0 83.33–100 166.66–200 00101 1:2 25.0–41.67 50.0–83.33 100.0–166.66 01111 1:4 25.0 100 200 All other values of AD[12:8] are reserved. Figure 10-2 correlates CLKIN, internal bus, and core clock frequencies for the 2x–4x multipliers. CLKIN Internal Clock Core Clock 2x 25.0 2x 50.0 50.0 100.0 100.0 4x 2x 25.0 25 200.0 100.0 50 70 CLKIN (MHz) 30 50 70 90 110 200.0 130 60 80 100 120 Internal Clock (MHz) 140 160 180 200 220 240 260 Core Clock (MHz) Figure 10-2. CLKIN, Internal Bus, and Core Clock Ratios 10.1.3 Internal Bus Overview There are three main internal buses in the MCF548x—the extended local bus (XL bus), the internal peripheral bus (slave bus), and the communication subsystem bus (CommBus). See Figure 10-1. • XL bus — Interface between the ColdFire core, memory controller, communication subsystem, FlexBus controller, and PCI controller. • Internal peripheral bus (slave bus) — The control/data interface from the core to the communication subsystem or peripheral programming registers and FIFOs. The base address of this memory-mapped bus will be stored in the internal peripheral bus base address register (MBAR). • CommBus — The data transfer interface between the multichannel DMA and each peripheral function. MCF548x Reference Manual, Rev. 5 10-2 Freescale Semiconductor Introduction 10.1.4 XL Bus Features Features of the XL bus and its integration modules include the following: • 32-bit physical address • 64-bit data bus width • Split-transaction bus; address and data tenures occur independently. • One-level address pipeline; supports up to two complete address tenures before the first data tenure completes. • Strict, in-order, address and data tenures are enforced. • Address and data bus “parking” may be used to remove arbitration phase from the address and data tenures—most recent master, programmed master, or no parking methods supported. • Access can occur in single (1-8 bytes) beat, or four-beat (32 bytes) burst transfers. • Eight-level arbitration priority that is hardware selectable for each master with a least recently used (LRU) protocol for masters of equal priority. Priority may change dynamically based on specific system requirements. • Fully static, multiplexed bus architecture. 10.1.5 Internal Bus Transaction Summaries The XL bus can be mastered by the ColdFire core, multichannel DMA controller, and the PCI controller (external PCI master). Any of these masters can access all resources available to the XL bus. Bus masters can access any on-chip or off-chip resources via the XL bus. The sequence is as follows: • Bus masters gains mastership of the XL bus from the XL bus arbiter. • The bus master’s address is asserted during the address tenure. XL bus slave devices (SDRAM, PCI, etc.) decode the address. If the address falls within a slave’s space, it returns an address acknowledge. • The bus master initiates the data tenure and transfers the data to the appropriate slave device. 10.1.6 XL Bus Interface Operations This section describes how the XLB interface operates. 10.1.6.1 Basic Transfer Protocol An XLB interface memory transaction is illustrated in Figure 10-3. It shows that the transaction consists of distinct address and data tenures, each having three phases: arbitration, transfer, and termination. The separation of these operations allows address pipelines and split transactions to be implemented. Split-bus transaction capability allows one master to have mastership of the address bus, while another master has mastership of the data bus. Pipelines allows the address tenure of a bus transaction to begin before the data tenure of the previous transaction finishes. The data transfer phase can either be one beat or four, depending on whether or not the transaction is a burst. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 10-3 Address Tenure Arbitration Transfer Termination Data Tenure Arbitration Transfer Termination Figure 10-3. Address and Data Tenures The following outlines the basic functions of each of the phases: • Address tenure: — Arbitration: During arbitration, address bus arbitration signals are used to gain mastership of the address bus. — Transfer: After mastership is obtained, the address bus master transfers the address and transfer attributes on the address bus. Address signals and transfer attribute signals control the address transfer. — Termination: After the address transfer, the system signals that the address tenure is complete or that it must be repeated. • Data tenure: — Arbitration: To begin a data tenure, the master arbitrates for data bus mastership. — Transfer: After mastership is obtained, the data bus master samples the data bus for read operations or drives the data bus for write operations. — Termination: Data termination signals are required after each data beat in a data transfer. In a single-beat transaction, data termination signals also indicate the end of the tenure; in burst accesses, data termination signals apply to individual beats and indicate the end of the tenure only after the final data beat. 10.1.6.2 Address Pipelines The XLB protocol provides independent address and data bus capability to support pipeline and split-bus transaction system organizations. The XLB arbiter allows for one level of pipeline. This feature can be enabled and disabled in the Arbiter Configuration Register (XARB_CFG). While this feature does not improve latency, it can significantly improve bus/memory throughput, so it should be considered for systems that expect to stress bus throughput capacity. The XLB arbiter effects pipelines by regulating address bus grant, data bus grants, and address acknowledge signals. For example, a one-level pipeline is enabled by asserting the address acknowledge signal to the current address bus master, as well as granting the address bus to the next requesting master before the current data bus tenure completes. MCF548x Reference Manual, Rev. 5 10-4 Freescale Semiconductor PLL 10.2 PLL 10.2.1 PLL Memory Map/Register Descriptions Table 10-2. System PLL Memory Map 10.2.2 MBAR Offset Name 0x300 System PLL Control Register Byte0 Byte1 Byte2 Byte3 SPCR Access R/W System PLL Control Register (SPCR) The system PLL control register (SPCR) defines the clock enables used to control clocks to a set of peripherals. Unused peripherals can have their clock stopped, reducing power consumption. In addition, the SPCR contains a read-only bit for the system PLL lock status. At reset, the clock enables are set, enabling all system PLL gated output clocks. R 31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 PLLK 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 COR EN CRY ENB 0 PSC EN 0 FB EN PCI EN MEM EN 0 1 1 1 1 1 1 1 1 W Reset R W Reset CRY CAN1 ENA EN 1 1 Reg Addr USB FEC1 FEC0 DMA CAN0 EN EN EN EN EN 1 1 1 1 1 MBAR + 0x300 Figure 10-4. System PLL Control Register (SPCR) Table 10-3. SPCR Field Descriptions Bits Name Description 31 PLLK 30-15 — 14 COREN Core & Communications Sub-System Clock Enable - Controls clocks for the CF4 Core, System SRAM, CommBus Arbiter, I2C, Comm Timers, and External DMA modules 13 CRYENB Crypto Clock Enable B - Controls the fast clock to the SEC 12 CRYENA Crypto Clock Enable A - Controls the slow clock to the SEC 11 CAN1EN CAN1 Clock Enable 10 — System PLL Lock Status - Read-only lock status of the system PLL. 1 PLL has obtained frequency lock 0 PLL has not locked Reserved, should be cleared. Reserved, should be cleared. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 10-5 Table 10-3. SPCR Field Descriptions (Continued) Bits Name 9 PSCEN 8 — 7 USBEN USB Clock Enable 6 FEC1EN FEC1 Clock Enable 5 FEC0EN FEC0 Clock Enable 4 DMAEN Multi-channel DMA Clock Enable 3 CAN0EN CAN0 Clock Enable 2 FBEN FlexBus Clock Enable 1 PCIEN PCI Bus Clock Enable 0 MEMEN 10.3 Description PSC Clock Enable - Controls clock for all PSC modules. Reserved, should be cleared. Memory Clock Enable - Controls clocks of the SDRAM controller module XL Bus Arbiter The XL bus arbiter handles bus arbitration between XL bus masters. 10.3.1 Features The arbiter features are as follows: • Eight priority levels • Priority levels may be changed dynamically by XL bus masters • XL bus arbitration support for eight masters • Least recently used (LRU) priority scheme for masters of equal priority • Multiple masters at each priority level supported • One level of address pipelines is enforced by the arbiter • Bus grant parking modes: — No parking — Park on last master — Park on programmed master • Watchdog timers for various XL bus time-out conditions 10.3.2 10.3.2.1 Arbiter Functional Description Prioritization The prioritization function will indicate that a master is requesting the bus and indicate which master has priority. Priority is determined first by using the hardcoded master priority or the master n priority bits in the arbiter master priority register (XARB_PRIEN), depending on the arbiter master priority enable bit for each master. Secondly, masters at the same level of priority will be further sorted by a least recently used MCF548x Reference Manual, Rev. 5 10-6 Freescale Semiconductor XL Bus Arbiter algorithm (LRU). Once a requesting master is identified as having priority and is granted the bus, that master will be continue to be granted the bus if: 1. It is requesting the bus. The request must occur immediately after the required 1 clock de-assertion after a qualified bus grant. and 2. It is the highest priority device. and 3. There is no address retry. Multiple masters at level 0 will only be able to perform one tenure before the bus is passed to the next master at level 0 using the LRU algorithm. The priority level of each master may be changed while the arbiter is running. This allows dynamic changes in priority such as an aging scheme. The arbiter recognizes changes after one clock. It is possible to control priority by enabling the master priority enable bits for a master (XARB_PRIEN). This causes the priority to be determined from the master n priority bits in the arbiter master priority register (XARB_PRI). Once again a system dependent dynamic scheme may be employed. 10.3.2.2 10.3.2.2.1 Bus Grant Mechanism Bus Grant The bus grant mechanism generates the address bus grant signals to the masters using the signals from the prioritization function. It will also generate required indicators of state to the prioritization and watchdog functions. The bus grant mechanism will enforce the one level address pipeline. The critical condition is that before a third address tenure is granted, the first tenure (address and, if needed, data) must be completed. The arbiter will assert a bus grant to a master when there are masters requesting, or if parking is enabled and the one level pipeline condition is met. 10.3.2.2.2 Parking Modes The bus grant mechanism will support the no parking, park on programmed master, and park on last master bus parking modes. • When in no parking mode, the arbiter will not assert a bus grant when there are no masters asserting a bus request. • In park on programmed master mode, the arbiter will assert a bus grant to the master indicated in the select parked master field (ACFG[SP]) when no masters are asserting a bus request and the one level pipeline will not be violated. • In park on last master mode, the arbiter will assert a bus grant to the last master granted the bus when no masters are asserting a bus request and the one level pipeline will not be violated. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 10-7 10.3.2.3 Watchdog Functions 10.3.2.3.1 Timer Functions There are three watchdog timers: address tenure time out, data tenure time out, and bus activity time out. Each has a programmable timer count and can be disabled. A timer time-out will set a status bit and trigger an interrupt if that interrupt is enabled. • The address tenure watchdog is a 32-bit timer. If an acknowledge is not detected by the programmed number of clocks after bus grant is accepted, the address watchdog timer will expire and the arbiter will issue an acknowledge. The related data tenure will be terminated with a transfer error acknowledge. The arbiter will set the Address Tenure Time-out Status bit in the arbiter status register and issue an interrupt if that interrupt is enabled. The upper 28 bits of address tenure time-out are programmed via the address tenure time-out register. The lower 4 bits are always 0xF. • The data tenure watchdog is a 32-bit timer. If a data tenure is not terminated, the data watchdog timer will expire and the arbiter will issue a transfer error acknowledge. The arbiter will set the Data Tenure Time-out Status bit in the arbiter status register and issue an interrupt if that interrupt is enabled. Address Time-out (32 bits) = {address tenure time-out register (28bits), 0xF} Data Time-out (32 bits) = {data tenure time-out register (28 bits), 0xF} • The bus activity watchdog is a 32-bit timer. If no bus activity is detected by the programmed number of clocks, the bus activity watchdog timer will expire and the arbiter will set the Bus Activity Time-out Status bit in the arbiter status register and issue an interrupt if that interrupt is enabled. NOTE Enabling the data time-out will also enable the address time-out. It is recommended that the data watchdog timer should always be programmed to a value that is larger than the address watchdog timer. This prevents the XL bus arbiter from generating a transfer error acknowledge due to expiration of the data watchdog timer while the address tenure has not completed. 10.3.3 XLB Arbiter Register Descriptions The XLB Arbiter registers and their locations are defined in Table 10-4. Table 10-4. XL Bus Arbiter Memory Map MBAR Offset Name 0x240 Arbiter Configuration Register XARB_CFG R/W 0x244 Arbiter Version Register XARB_VER R 0x248 Arbiter Status Register XARB_SR R/W 0x24C Arbiter Interrupt Mask Register XARB_IMR R/W 0x250 Arbiter Address Capture XARB_ADRCAP R/W 0x254 Arbiter Signal Capture XARB_SIGCAP R/W Byte0 Byte1 Byte2 Byte3 Access MCF548x Reference Manual, Rev. 5 10-8 Freescale Semiconductor XL Bus Arbiter Table 10-4. XL Bus Arbiter Memory Map (Continued) MBAR Offset Name 0x258 Arbiter Address Timeout XARB_ADRTO R/W 0x25C Arbiter Data Timeout XARB_DATTO R/W 0x260 Arbiter Bus Timeout XARB_BUSTO R/W 0x264 Arbiter Master Priority Enable XARB_PRIEN R/W 0x268 Arbiter Master Priority XARB_PRI R/W 10.3.3.1 Byte0 Byte1 Byte2 Byte3 Access Arbiter Configuration Register (XARB_CFG) The arbiter configuration register is used to enable watchdog functions and arbiter protocol functions. 31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 0 0 0 0 0 BA DT AT 0 0 0 0 0 0 0 0 1 1 0 R PLDIS W Reset R SP 0 PM W Reset 0 0 Reg Addr 0 0 0 0 MBAR + 0x0240 Figure 10-5. Arbiter Configuration Register (XARB_CFG) Table 10-5. XARB_CFG Bit Descriptions Bit Name Description 31 PLDIS 30–11 — Reserved, should be cleared. 10–8 SP Select Parked Master. These bits set the master that is used in Park on Programmed Master mode. 000 Master 0 001 Master 1 ... 111 Master 7). 7 — Reserved, should be cleared. 6–5 PM[1:0] Pipeline Disable. This bit is used to control the pipeline functionality 0 Enable pipeline 1 Disable pipeline Parking Mode. Parking modes are detailed in Section 10.3.2.2.2, “Parking Modes.” 00 No parking (default) 01 Reserved 10 Park on most recently used master 11 Park on programmed master as specified by the Select Parked Master bits 21:23 above. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 10-9 Table 10-5. XARB_CFG Bit Descriptions (Continued) Bit Name 4 — Reserved, should be cleared. 3 BA Bus Activity Time-out Enable. If enabled, the arbiter will set the Bus Activity Time-out Status bit (XARB_SR[BA]) when the Bus Activity Time-out is reached. Bus Activity Time-out is derived from the arbiter bus activity time out count register. 0 Disable bus activity time-out 1 Enable bus activity time-out 2 DT Data Tenure Time-out Enable. If enabled, the arbiter will transfer error acknowledge when the Data Tenure Time-out is reached. Data Tenure Time-out is derived from the arbiter data tenure time out count register. Also, the arbiter will set the Data Tenure Time-out Status bit (Arbiter Status Register Bit 30). Setting this bit will also enable the Address Tenure Time-out. This is required to ensure that a data time-out will not occur before an address acknowledge. 0 Disable data tenure time-out 1 Enable data tenure time-out 1 AT Address Tenure Time-out Enable. If enabled, the arbiter will AACK and TEA (if required) when the Address Tenure Time-out is reached. Address Tenure Time-out is derived from the Arbiter Address Tenure Time Out Count register. Also, the arbiter will set the Address Tenure Time-out Status bit (Arbiter Status Register Bit 31). Address Tenure Time-out is also enabled by the DT bit above. 0 Disable address tenure time-out 1 Enable address tenure time-out 0 — Reserved, should be cleared. 10.3.3.2 31 Description Arbiter Version Register (XARB_VER) 30 29 28 27 26 25 24 R 23 22 21 20 19 18 17 16 VER W Reset 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 0 0 0 0 0 0 1 R VER W Reset 0 0 Reg Addr 0 0 0 0 0 0 MBAR + 0x0244 Figure 10-6. Arbiter Version Register (XARB_VER) Table 10-6. VER Field Descriptions Bit Name 31–0 VER Description Hardware Version ID. The current version number is 0x0001. MCF548x Reference Manual, Rev. 5 10-10 Freescale Semiconductor XL Bus Arbiter 10.3.3.3 Arbiter Status Register (XARB_SR) The arbiter status register indicates the state of watchdog functions. When a monitored condition occurs, the respective bit is set to 1. The bit will stay set until the bit is cleared by writing a 1 into that bit. Even if the causal condition is removed, the bit will remain set until cleared. R 31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 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 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 0 0 0 0 0 0 SEA MM TTA TTR ECW TTM BA DT AT 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 W Reset R W Reset Reg Addr MBAR + 0x0248 Figure 10-7. Arbiter Status Register (XARB_SR) Table 10-7. XARB_SR Field Descriptions Bits Name 31–9 — 8 SEA Slave Error Acknowledge. This bit is set when an error is detected by any slave devices during the transfer. 7 MM Multiple Masters at priority 0. If more than 1 master is recognized at priority 0, this bit is set. Once this occurs this bit will remain set until cleared. This bit is intended to help in tuning dynamic priority algorithm development. 6 TTA TT Address Only. The arbiter automatically AACKs for address only TT codes. This bit is set when this occurs. 5 TTR TT Reserved. The arbiter automatically AACKs for reserved TT codes. This bit is set when this occurs. 4 ECW External Control Word Read/Write. External Control Word Read/Write operations are not supported on the XL bus. If either occur, the arbiter AACKs and TEAs and sets this bit. 3 TTM TBST/TSIZ mismatch. Set when an illegal/reserved TBST and TSIZ[0:2] combination occurs. These combinations are TBST asserted and TSIZ[0:2] = 000, 001, 011, or 1xx (x is 0 or 1). 2 BA Bus Activity Tenure Time-out. Set when the bus activity time-out counter expires. 1 DT Data Tenure Time-out. Set when the data tenure time-out counter expires. 0 AT Address Tenure Time-out. Set when the address tenure time-out counter expires. 10.3.3.4 Description Reserved, should be cleared. Arbiter Interrupt Mask Register (XARB_IMR) The arbiter interrupt mask register is used to enable a status bit to cause an interrupt. If the interrupt mask and corresponding status bits are set in the arbiter status register and arbiter interrupt mask register, the arbiter will assert the interrupt signal. Normally, an interrupt service routine would read the status register MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 10-11 to determine the state of the arbiter. It is possible that multiple conditions exist that would cause an interrupt. Disabling an interrupt by writing a 0 to a bit in this register will not clear the status bit in the arbiter status register. R 31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 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 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 0 0 0 0 0 0 BAE DTE ATE 0 0 0 0 0 0 0 0 0 0 W Reset R SEAE MME TTAE TTRE ECWE TTME W Reset Reg Addr 0 0 0 0 0 0 MBAR + 0x024C Figure 10-8. Arbiter Interrupt Mask Register (XARB_IMR) Table 10-8. XARB_IMR Field Descriptions Bits Name Description 31–9 — 8 SEAE Slave Error Acknowledge interrupt enable. 0 The corresponding interrupt source is masked. 1 The corresponding interrupt source is enabled. 7 MME Multiple Masters at priority 0 interrupt enable. 0 The corresponding interrupt source is masked. 1 The corresponding interrupt source is enabled. 6 TTAE TT Address Only interrupt enable. 0 The corresponding interrupt source is masked. 1 The corresponding interrupt source is enabled. 5 TTRE TT Reserved interrupt enable. 0 The corresponding interrupt source is masked. 1 The corresponding interrupt source is enabled. 4 ECWE External Control Word Read/Write interrupt enable. 0 The corresponding interrupt source is masked. 1 The corresponding interrupt source is enabled. 3 TTME TBST/TSIZ mismatch interrupt enable. 0 The corresponding interrupt source is masked. 1 The corresponding interrupt source is enabled. 2 BAE Bus Activity Tenure Time-out interrupt enable. 0 The corresponding interrupt source is masked. 1 The corresponding interrupt source is enabled. Reserved, should be cleared. MCF548x Reference Manual, Rev. 5 10-12 Freescale Semiconductor XL Bus Arbiter Table 10-8. XARB_IMR Field Descriptions (Continued) Bits Name 1 DTE Data Tenure Time-out interrupt enable. 0 The corresponding interrupt source is masked. 1 The corresponding interrupt source is enabled. 0 ATE Address Tenure Time-out interrupt enable. 0 The corresponding interrupt source is masked. 1 The corresponding interrupt source is enabled. 10.3.3.5 Description Arbiter Address Capture Register (XARB_ADRCAP) The arbiter address capture register will capture the address for a tenure that has an address time-out, data time-out, or there is a transfer error acknowledge from another source. This value is held until unlocked by writing any value to the arbiter address capture register or arbiter bus signal capture register. This value is also unlocked by writing a 1 to either XARB_SR[DT] or XARB_SR[AT]. Unlocking the register does not clear its contents. 31 30 29 28 27 26 25 R 24 23 22 21 20 19 18 17 16 ADRCAP W Reset 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 0 0 0 0 0 0 R ADRCAP W Reset 0 0 0 Reg Addr 0 0 0 0 0 0 MBAR + 0x0250 Figure 10-9. Arbiter Address Capture Register (XARB_ADRCAP) Table 10-9. XARB_ADRCAP Field Descriptions Bits Name Description 31–0 ADRCAP Address that is captured when a bus error occurs. This happens on an address time-out, data time-out, or any transfer error acknowledge. 10.3.3.6 Arbiter Bus Signal Capture Register (XARB_SIGCAP) Important bus signals are captured when a bus error occurs. This happens on an address time-out, data time-out, or any transfer error acknowledge. The arbiter bus signal capture register will capture TT, TBST, and TSIZ for a tenure that has an address time-out or data time-out, or there is a transfer error acknowledge from another source. These values are held until unlocked by writing any value to the arbiter address capture register (XARB_ADRCAP) or arbiter bus signal capture register (XARB_SIGCAP). These values are also unlocked by writing a 1 to either XARB_SR[DT] or XARB_SR[AT]. Unlocking the register does not clear its contents. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 10-13 R 31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 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 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 0 0 0 0 0 — TBST 0 0 0 0 0 0 0 0 0 0 W Reset R TSIZ[0:2] TT[0:4] W Reset 0 Reg Addr 0 0 0 0 0 MBAR + 0x0254 Figure 10-10. Arbiter Bus Signal Capture Register (XARB_SIGCAP) Table 10-10. XARB_SIGCAP Field Descriptions Bits Name Description 31–10 — Reserved, should be cleared. 9–7 TSIZ[0:2] TSIZ[0:2] encodings. 001 1 byte 010 2 bytes 011 3 bytes 100 4 bytes 101 5 bytes 110 6 bytes 111 7 bytes 000 8 bytes 010 32 bytes (when TBST=0) 6 — 5 TBST 4–0 TT Reserved, should be cleared TBST. 1 Non-burst 0 Burst TT[0:4] encodings. 01010 Read 00010 Write-with-flush 00110 Write-with-kill MCF548x Reference Manual, Rev. 5 10-14 Freescale Semiconductor XL Bus Arbiter 10.3.3.7 R Arbiter Address Tenure Time Out Register (XARB_ADRTO) 31 30 29 28 27 26 25 24 23 22 0 0 0 0 0 0 0 0 1 1 1 1 1 1 15 14 13 12 11 10 9 8 7 21 20 19 18 17 16 1 1 1 1 1 1 6 5 4 3 2 1 0 1 1 1 1 1 1 1 ADRTO W Reset R ADRTO W Reset 1 1 1 1 1 1 1 Reg Addr 1 1 MBAR + 0x0258 Figure 10-11. Arbiter Address Tenure Time Out Register (XARB_ADRTO) Table 10-11. XARB_ADRTO Field Descriptions Bits Name 31–28 — 27–0 ADRTO 10.3.3.8 R Description Reserved, should be cleared. Upper 28-bits of the Address time-out counter value. This field is prepended to 0xF to generate the full 32-bit time-out counter value. Arbiter Data Tenure Time Out Register (XARB_DATTO) 31 30 29 28 27 26 25 24 23 22 0 0 0 0 0 0 0 0 1 1 1 1 1 1 15 14 13 12 11 10 9 8 7 21 20 19 18 17 16 1 1 1 1 1 1 6 5 4 3 2 1 0 1 1 1 1 1 1 1 DATTO W Reset R DATTO W Reset 1 1 1 Reg Addr 1 1 1 1 1 1 MBAR + 0x025C Figure 10-12. Arbiter Data Tenure Time Out Register (XARB_DATTO) MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 10-15 Table 10-12. XARB_DATTO Field Descriptions Bits Name 31–28 — 27–0 DATTO 10.3.3.9 Description Reserved, should be cleared. Upper 28-bits fo the Data time-out counter value. This field is prepended to 0xF to generate the full 32-bit time-out counter value. Arbiter Bus Activity Time Out Register (XARB_BUSTO) 31 30 29 28 27 26 25 24 R 23 22 21 20 19 18 17 16 BUSTO W Reset 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 1 1 1 1 1 1 1 R BUSTO W Reset 1 1 1 Reg Addr 1 1 1 1 1 1 MBAR + 0x0260 Figure 10-13. Arbiter Bus Activity Time Out Register (XARB_BUSTO) Table 10-13. XARB_BUSTO Field Descriptions Bits Name 31–0 BUSTO Description Bus activity time-out counter value in XLB clocks. 10.3.3.10 Arbiter Master Priority Enable Register (XARB_PRIEN) The arbiter master priority enable register determines whether the arbiter uses the hardwired or software programmable priority for a master. The default is enabled for all masters. Both methods may be used at the same time for different masters. This register may be written at any time. The change will become effective 1 clock after the register is written. MCF548x Reference Manual, Rev. 5 10-16 Freescale Semiconductor XL Bus Arbiter R 31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 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 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 0 0 0 0 0 0 0 — — — — M3 M2 — M0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 W Reset R W Reset Reg Addr MBAR + 0x0264 Figure 10-14. Arbiter Master Priority Enable Register (XARB_PRIEN) Table 10-14. XARB_PRIEN Field Descriptions Bits Name Description 31–4 — Reserved, should be cleared. 3 M3 Master 3 Priority Register Enable 2 M2 Master 2 Priority Register Enable 1 — Reserved, should be cleared. 0 M0 Master 0 Priority Register Enable When enabled, the software programmable value in the arbiter master priority register (XARB_PRI) is used as the priority for the master. When disabled, the master’s priority is determined as follows: Table 10-15. Hardcoded Master Priority Master Priority Description M7–M4 — Unused M3 7 PCI Target Interface M2 7 Multichannel DMA M1 — Unused M0 7 ColdFire core 10.3.3.11 Arbiter Master Priority Register (XARB_PRI) The master n priority bits of the arbiter master priority register are used to set the priority of each master if the corresponding arbiter master priority enable register bit is enabled. This XARB_PRI register, in conjunction with the arbiter master priority enable (XARB_PRIEN) register, allows master priorities to be set, ignoring the hardcoded priority. This register may be written at anytime. The change will become effective 1 clock after the register is written. Valid values are from 0 to 7, with 0 being the highest priority. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 10-17 31 R 30 0 29 28 Reserved 27 26 0 25 24 Reserved 23 22 0 21 20 Reserved 19 18 0 17 16 Reserved W Reset R 0 1 1 1 0 1 1 1 0 1 1 1 0 1 1 1 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 M3 Priority 0 M2 Priority 0 Reserved 0 M0 Priority W Reset 0 1 1 1 0 1 Reg Addr 1 1 0 1 1 1 0 1 1 1 MBAR + 0x0268 Figure 10-15. Arbiter Master Priority Register (XARB_PRI) Table 10-16. XARB_PRI Field Descriptions Bits Name 31–15 — 14–12 M3P 11 — 10–8 M2P 7–3 — 2–0 M0P Description Reserved, should be cleared. Master 3 Priority Reserved, should be cleared. Master 2 Priority Reserved, should be cleared. Master 0 Priority MCF548x Reference Manual, Rev. 5 10-18 Freescale Semiconductor Chapter 11 General Purpose Timers (GPT) 11.1 Introduction This chapter describes the operation of the MCF548x general purpose timers. 11.1.1 Overview The MCF548x has four general-purpose timers (GPT[0:3]) that are configurable for the following functions: • Input capture • Output capture • Pulse width modulation (PWM) output • Simple GPIO • Internal CPU timer • Watchdog timer (on GPT0 only) Timer modules run off the internal peripheral bus clock. Each timer is associated to a single I/O signal. Each timer has a 16-bit prescaler and 16-bit counter, thus achieving a 32-bit range (but only 16-bit resolution). 11.1.2 Modes of Operation The following gives a brief description of the available GPT modes: 1. Input Capture—When enabled in this mode, the counters run until the specified capture event occurs (rise, fall, or pulse) on TIN[3:0]. At the capture event, the counter value is latched in the status register. When this occurs, a CPU interrupt is generated. 2. Output Capture—When enabled in this mode, the counters run until they reach the programmed terminal count value. At this point, the specified output event is generated (toggle, pulse high, or pulse low) on TOUT[3:0]. When this occurs, a CPU interrupt is generated. 3. PWM (pulse width modulation)—In this mode the user can program period and width values to create an adjustable, repeating output waveform on TOUT[3:0]. A CPU interrupt can be generated at the beginning of each PWM period, at which time a new width value can be loaded. The new width value, which represents “ON time,” is automatically applied at the beginning of the next period. This mode is suitable for PWM audio encoding. 4. Simple GPIO—In this mode TOUT[3:0] and TIN[3:0] operate as a GPIO. Either TOUT[3:0] or TIN[3:0] are specified, according to the programmable GPIO field. GPIO mode is mutually exclusive of modes 1 through 3 (listed above). In GPIO mode, modes 5 through 6 (listed below) remain available. 5. CPU Timer—The I/O signal is not used in this mode. Once enabled, the counters run until they reach a programmed terminal count. When this occurs, an interrupt can be generated to the CPU. This timer mode can be used simultaneously with the simple GPIO mode. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 11-1 6. Watchdog Timer—This is a special CPU timer mode, available only on GPT0. The user must enable the watchdog timer mode, which is not active upon reset. The terminal count value is programmable. If the counter is allowed to expire, a full reset occurs. To prevent the watchdog timer from expiring, software must periodically write 0xA5 to the GMS0[OCPW] field. This causes the counter to reset. 11.2 External Signals The GPT signals are the following: • TIN[3:0]—External timer input • TOUT[3:0]—External timer output 11.3 Memory Map/Register Definition Each GPT uses four 32-bit registers. These registers are located at MBAR + the GPT offset 0x800. Table 11-1 summarizes the GPT control registers. Table 11-1. General Purpose Timer Memory Map Address (MBAR +) Name 0x800 GPT Enable and Mode Select Register 0 GMS0 R/W 0x804 GPT Counter Input Register 0 GCIR0 R/W 0x808 GPT PWM Configuration Register 0 GPWM0 R/W 0x80C GPT Status Register 0 GSR0 R/W 0x810 GPT Enable and Mode Select Register 1 GMS1 R/W 0x814 GPT Counter Input Register 1 GCIR1 R/W 0x818 GPT PWM Configuration Register 1 GPWM1 R/W 0x81C GPT Status Register 1 GSR1 R/W 0x820 GPT Enable and Mode Select Register 2 GMS2 R/W 0x824 GPT Counter Input Register 2 GCIR2 R/W 0x828 GPT PWM Configuration Register 2 GPWM2 R/W 0x82C GPT Status Register 2 GSR2 R/W 0x830 GPT Enable and Mode Select Register 3 GMS3 R/W 0x834 GPT Counter Input Register 3 GCIR3 R/W 0x838 GPT PWM Configuration Register 3 GPWM3 R/W 0x83C GPT Status Register 3 GSR3 R/W Byte 0 Byte 1 Byte 2 Byte 3 Access MCF548x Reference Manual, Rev. 5 11-2 Freescale Semiconductor Memory Map/Register Definition 11.3.1 GPT Enable and Mode Select Register (GMSn) 31 30 29 R 28 27 26 25 24 OCPW 23 22 0 0 21 20 19 18 0 0 OCT 17 16 ICT W Reset 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 0 CE 0 SC OD IEN 0 0 0 0 0 0 0 0 0 0 0 R WDE N W Reset 0 Reg Addr GPIO 0 0 0 0 TMS 0 0 0 MBAR + 0x800 (GMS0), 0x810 (GMS1), 0x820 (GMS2), 0x830 (GSM3) Figure 11-1. GPT Enable and Mode Select Register (GMSn) Table 11-2. GMSn Field Descriptions Bits Name Description 31–24 OCPW Output capture pulse width. Applies to OC pulse types only. This field specifies the number of clocks (non-prescaled) to create a short output pulse at each output event. This pulse is generated at the end of the output capture period and overlays the next OC period (rather than adding to the period). This field is alternately used as the watchdog reset field if watchdog timer mode is enabled. 23–22 — 21–20 OCT 19–18 — 17–16 ICT Reserved, should be cleared. Output capture type. Describes action to occur at each output capture event, as follows: 00 Special case, output is immediately forced low without respect to each output capture event. 01 Output pulses highs, initial value is low (OCPW field applies). 10 Output pulses low, initial value is high (OCPW field applies). 11 Output toggles. GPIO modalities can be used to achieve an initial output state prior to enabling OC mode. It is important to move directly from GPIO output mode to OC mode and not to pass through the TMS=000 state. To prevent the internal timer mode from engaging during the GPIO state, CE bit should be cleared during the configuration steps. GPIO initialization is needed when presetting the I/O to 1 in conjunction with a simple toggle OCT setting. Reserved, should be cleared. Input capture type. Describes the input transition type required to trigger an input capture event, as follows: 00 Any input transition causes an IC event. 01 IC event occurs at input rising edge. 10 IC event occurs at input falling edge. 11 IC event occurs at any input pulse (i.e., at the second input edge). MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 11-3 Table 11-2. GMSn Field Descriptions (Continued) Bits Name Description 15 WDEN Watchdog enable. Enables watchdog operation. A timer expiration causes an internal MCF548x reset. Watchdog operation requires the TMS field be set for internal timer mode and the CE bit to be set. In this mode the OCPW byte field operates as a watchdog reset field. Writing A5 to the OCPW field resets the watchdog timer, preventing it from expiring. As long as the timer is properly configured, the watchdog operation continues. This bit (and functionality) is implemented only for GPT0. 0 Watchdog not enabled 1 Watchdog enabled 14–13 — Reserved, should be cleared. 12 CE Counter enable. Enables or resets the internal counter during internal timer modes only. CE must be set to enable these modes. If cleared, counter is held in reset. 0 Timer counter held in reset 1 Timer counter enabled This bit is secondary to the timer mode select bits (TMS). If TMS is1XX, internal timer modes are enabled. CE can then enable or reset the internal counter without changing the TMS field. GPIO operation is also available in this mode. 11 — Reserved, should be cleared. 10 SC Stop/continuous mode. 0 Stops the operation 1 Continues the operation The SC bit applies to multiple modes, as follows: IC mode (input capture mode) Stop operation—At each IC event, counter is reset. Continuous operation—counter is not reset at each IC event. Effect is to create status count values that are cumulative between capture events. If the special pulse mode capture type is specified, the SC bit is not used, operation fixed as if it were stop. OC mode (output capture mode) Stop operation—Counter resets and stops at the first output capture event. Software needs to pass through TMS=000 state to restart timer. Continuous operation—counter resets and continues at each OC event. The effect to is create back-to-back periodic OC events. PWM mode (pulse width modulation mode) The SC bit is not used; operation is always continuous. CPU Timer mode Stop operation—On counter expiration, timer waits until status bit is cleared by passing through TMS=000 state before beginning a new cycle. Continuous operation—On counter expiration, timer resets and immediately begin a new cycle. The effect is to generate fixed periodic timeouts. WatchDog Timer and GPIO modes The SC bit is not used. 9 OD Open drain. 0 Normal I/O 1 Open Drain emulation—affects all modes that drive the I/O pin (GPIO, OC, and PWM). Any output “1” is converted to a tri-state at the I/O pin. MCF548x Reference Manual, Rev. 5 11-4 Freescale Semiconductor Memory Map/Register Definition Table 11-2. GMSn Field Descriptions (Continued) Bits Name Description 8 IEN Interrupt enable. Enables interrupt generation to the CPU for all modes (IC, OC, PWM, and Internal Timer). IEN is not required for watchdog expiration to create a reset. 0 Interrupt disabled 1 Interrupt enabled 7–6 — 5–4 GPIO 3 — 2–0 TMS 11.3.2 Reserved, should be cleared. GPIO mode type. Simple GPIO functionality that can be used simultaneously with the internal timer mode. It is not compatible with IC, OC, or PWM modes, because these modes dictate the usage of the I/O signals. 0X Timer enabled as simple GPIO input on TINn 10 Timer enabled as simple GPIO output, TOUTn=0 11 Timer enabled as simple GPIO output, TOUTn=1 (tri-state if OD=1) While in GPIO modes, internal timer mode is also available. To prevent undesired timer expiration, keep the CE bit cleared. Reserved, should be cleared. Timer mode select (and module enable). 000 Timer module not enabled. All timer operation is completely disabled. Control and status registers are still accessible. This mode should be entered when the timer is to be re-configured,. 001 Timer enabled for input capture. 010 Timer enabled for output capture. 011 Timer enabled for PWM. 1XX Timer enabled for simple GPIO. Internal timer modes available. CE bit controls timer counter. GPT Counter Input Register (GCIRn) 31 30 29 28 27 26 25 24 R 23 22 21 20 19 18 17 16 PRE W Reset 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 0 0 0 0 0 0 0 R CNT W Reset 0 0 Reg Addr 0 0 0 0 0 0 MBAR + 0x804 (GCIR0), 0x814 (GCIR1), 0x824 (GCIR2), 0x834 (GCIR3) Figure 11-2. GPT Counter Input Register (GCIRn) MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 11-5 Table 11-3. GCIRn Field Descriptions Bits Name Description 31–16 PRE Prescaler. Prescale amount applied to internal counter (in clocks). Note that in addition to other enable bits and field settings, the PRE field must be written as non-zero to enable counter operation for all modes except the simple GPIO mode. A prescale of 0x0001 means one clock per count increment. 15–0 CNT Count value. Sets number of prescaled counts applied to reference events, as follows: IC—Field has no effect, internal counter starts at 0. OC—Number of prescaled counts counted before creating output event. PWM—Number of prescaled counts defining the PWM output period. Internal Timer—Number of prescaled counts counted before timer (or watchdog) expires. Reading this register only returns the programmed value, intermediate values of the internal counter are not available to software. 11.3.3 GPT PWM Configuration Register (GPWMn) 31 30 29 28 27 26 25 R 24 23 22 21 20 19 18 17 16 WIDTH W Reset R 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 0 0 0 0 0 0 PWM OP 0 0 0 0 0 0 0 LOAD 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 W Reset Reg Addr MBAR + 0x808 (GPWM0), 0x818 (GPWM1), 0x828 (GPWM2), 0x838 (GPWM3) Figure 11-3. GPT PWM Configuration Register (GPWMn) Table 11-4. GPWMn Field Descriptions Bits Name 31–16 WIDTH 15–9 — 8 PWMOP Description PWM width. Used in PWM mode only. Defines ON time for output in prescaled counts. Similar to count value, which defines the period. ON time overlays the period time. If WIDTH = 0, output is always OFF. If WIDTH exceeds count value, output is always ON. ON and OFF polarity is set by the PWMOP bit. Reserved. Should be cleared. PWM output polarity. Defines PWM output polarity for OFF time. Opposite state is ON time. PWM cycles begin with ON time. 0 PWM output is low during OFF time 1 PWM output is high during OFF time MCF548x Reference Manual, Rev. 5 11-6 Freescale Semiconductor Memory Map/Register Definition Table 11-4. GPWMn Field Descriptions (Continued) Bits Name 7–1 — 0 LOAD 11.3.4 Description Reserved. Should be cleared. Bit forces immediate period update. Bit auto clears itself. A new period begins immediately with the current count and width settings. If LOAD = 0, new count or width settings are not updated until end of current period. Prescale setting is not part of this process. Changing prescale value while PWM is active causes unpredictable results for the period in which it was changed. The same is true for PWMOP bit. GPT Status Register (GSRn) 31 30 29 28 27 26 25 R 24 23 22 21 20 19 18 17 16 CAPTURE W Reset R 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 0 0 PIN 0 0 0 0 0 OVF W Reset 0 0 0 Reg Addr 0 0 0 0 0 0 0 0 0 TEXP PWMP COMP CAPT w1c w1c w1c w1c 0 0 0 0 MBAR + 0x80C (GSR0), 0x81C (GSR1), 0x82C (GSR2), 0x83C (GSR3) Figure 11-4. GPT Status Register (GSRn) Table 11-5. GSRn Field Descriptions Bits 31–16 Name Description CAPTURE Read of internal counter, latch at reference event. This is pertinent only in IC mode, in which case it represents the count value at the time the input event occurred. Capture status does not shadow the internal counter while an event is pending, it is updated only at the time the input event occurs. If ICT is set to 11, which is Pulse Capture Mode, the Capture value records the width of the pulse. Also, the SC bit is irrelevant in Pulse Capture Mode, operation is as if SC were 0. 15 — 14–12 OVF 11–9 — 8 PIN 7–4 — 3 TEXP Reserved. Should be cleared. Overflow counter. Represents how many times internal counter has rolled over. This is pertinent only during IC mode and would represent an extremely long period of time between input events. However, if SC = 1 (indicating cumulative reporting of input events), this field could come into play. This field is cleared by any “sticky bit” status write in the TEXP, PWMP, COMP, or CAPT bit fields. Reserved GPIO input value. This bit reflects the registered state of the TINn pin (all modes). The clock registers the state of the input. Valid, even if timer is not enabled. Reserved. Should be cleared. Timer expired in internal timer mode. Cleared by writing 1 to this bit position. Also cleared if TMS is 000 (i.e., timer not enabled). MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 11-7 Table 11-5. GSRn Field Descriptions (Continued) Bits Name 2 PWMP PWM end of period occurred. Cleared by writing 1 to this bit position. Also cleared if TMS is 000 (i.e., timer not enabled). 1 COMP OC reference event occurred. Cleared by writing 1 to this bit position. Also cleared if TMS is 000 (i.e., timer not enabled). 0 CAPT IC reference event occurred. Cleared by writing 1 to this bit position. Also cleared if TMS is 000 (i.e., timer not enabled). 11.4 11.4.1 Description Functional Description Timer Configuration Method Use the following method to configure each timer: 1. Determine the mode select field (GMSn[TMS]) value for the desired operation. 2. Program any other registers associated with this mode. 3. Program interrupt enable as desired. 4. Enable the timer by writing the mode select value into the TMS field. 11.4.2 Programming Notes Programmers should observe the following notes: 1. Intermediate values of the timer internal counters are not readable by software. 2. In PWM mode, an interrupt occurs at the beginning of a pulse. An interrupt service routine prepares the new pulse width of the next pulse while the current pulse is running. 3. The stop/continuous mode bit (GMSn[SC] ) operates differently for different modes. In general, this bit controls whether the timer halts at the end of a current mode, or resets and continues with a repetition of the mode. See Table 11-2 for precise operation. 4. The GMSn[TMS] field operates somewhat as a global enable. If it is zero, then all timer modes are disabled and internal counters are reset. See Table 11-2 for more detail. 5. There is a counter enable bit (GMSn[CE]) that operates somewhat independently of the TMS field. This bit controls the counter for CPU timer or watchdog timer modes only. See Table 11-2 to understand the operation of these bits across the various modes. MCF548x Reference Manual, Rev. 5 11-8 Freescale Semiconductor Chapter 12 Slice Timers (SLT) 12.1 Introduction This chapter explains the operation of the MCF548x slice timers. 12.1.1 Overview Two slice timers are included to provide shorter term periodic interrupts—SLT0 and SLT1. Each timer consists of a 32-bit counter with no prescale. The counters count down from a prescribed value and expire/interrupt when they reach zero. They can be configured to automatically preset to the prescribed value and resume counting or wait until the status/interrupt is serviced before beginning a new cycle. The current count value can be read without disturbing the count operation. Each SLT has a status bit to indicate the timer has expired. If enabled, a CPU interrupt is generated at count expiration. Each timer has a separate interrupt. Clearing the status and/or interrupt is accomplished by writing 1 to the status bit, or disabling the timer entirely with the timer enable (SCR[TEN]) bit. Software should write a terminal count value of greater than 255. 12.2 Memory Map/Register Definition There are two slice timers. Each one uses four 32-bit registers. These registers are located at an offset from MBAR of 0x900. Table 12-1 summarizes the SLT control registers. Table 12-1. Slice Timer Memory Map Address (MBAR +) Name 0x900 SLT Terminal Count Register 0 STCNT0 R/W 0x904 SLT Control Register 0 SCR0 R/W 0x908 SLT Count Value Register 0 SCNT0 R 0x90C SLT Status Register 0 SSR0 R/W 0x910 SLT Terminal Count Register 1 STCNT1 R/W 0x914 SLT Control Register 1 SCR1 R/W 0x918 SLT Count Value Register 1 SCNT1 R 0x91C SLT Status Register 1 SSR1 R/W Byte 0 Byte 1 Byte 2 Byte 3 Access MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 12-1 12.2.1 SLT Terminal Count Register (STCNTn) 31 30 29 28 27 26 25 24 R 23 22 21 20 19 18 17 16 TC W Reset 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 0 0 0 0 0 0 0 R TC W Reset 0 0 0 0 0 Reg Addr 0 0 0 MBAR + 0x900 (STCNT0), + 0x910 (STCNT1) Figure 12-1. SLT Terminal Count Register (STCNTn) Table 12-2. STCNTn Field Descriptions Bits Name Description 31–0 TC Terminal count. GPIO output bit set. The user programs this register to set the terminal count value to be used by the SLT. This register can be updated even if the timer is running; the new value takes effect immediately. The new value also clears any existing interrupt. Note: Software should not write a value less than 255 to the timer. 12.2.2 R SLT Control Register (SCRn) 31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 0 0 0 0 0 RUN IEN TEN 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 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 0 0 W Reset R W Reset Reg Addr MBAR + 0x904 (SCR0), + 0x914 (SCR1) Figure 12-2. SLT Control Register (SCRn) MCF548x Reference Manual, Rev. 5 12-2 Freescale Semiconductor Memory Map/Register Definition Table 12-3. SCRn Field Descriptions Bits Name 31–27 — 26 RUN Run or wait mode 0 Timer counter expires, but then waits until the timer is cleared (either by writing 1 to the status bit or by disabling and re-enabling the timer), before resuming operation. 1 Timer is enabled, and runs continuously. When the timer counter expires the terminal count value immediately is reloaded and resumes counting down. 25 IEN Interrupt enable. A CPU interrupt is generated only if this bit is set. 0 Interrupt is not generated 1 Interrupt is generated This bit does not affect operation of the timer counter or status bit registers. 24 TEN Timer enable 0 Timer is reset, then remains idle 1 Normal timer operation 23–0 — 12.2.3 Description Reserved, should be cleared. Reserved, should be cleared. SLT Timer Count Register (SCNTn) 31 30 29 28 27 26 25 24 R 23 22 21 20 19 18 17 16 CNT W Reset 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 0 0 0 0 0 0 0 R CNT W Reset 0 0 0 Reg Addr 0 0 0 0 0 MBAR + 0x908 (SCNT0), + 0x918 (SCNT1) Figure 12-3. SLT Count Register (SCNTn) Table 12-4. SCNTn Field Descriptions Bits Name Description 31–0 CNT Timer count. GPIO output bit set. Provides the current state of the timer counter. This register does not change while a read is in progress, but the actual timer counter continues unaffected. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 12-3 12.2.4 R SLT Status Register (SSRn) 31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 0 0 0 0 0 0 BE ST 0 0 0 0 0 0 0 0 w1c w1c W Reset R 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 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 0 0 W Reset Reg Addr MBAR + 0x90C (SSR0), + 0x91C (SSR1) Figure 12-4. SLT Status Register (SSRn) Table 12-5. SSRn Field Descriptions Bits Name Description 31–26 — Reserved, should be cleared 25 BE Bus Error Status. Provides information on attempted write to read-only register. The bit is cleared by writing 1 to its bit position. 24 ST SLT timeout. This status bit is set whenever the timer has expired. The bit is cleared by writing 1 to its bit position. If interrupts are enabled, clearing this status bit also clears the interrupt. 23–0 — Reserved, should be cleared. MCF548x Reference Manual, Rev. 5 12-4 Freescale Semiconductor Chapter 13 Interrupt Controller 13.1 Introduction This section details the functionality for the MCF548x interrupt controller. The general features of the interrupt controller include: • 63 interrupt sources, organized as: — 56 fully-programmable interrupt sources — 7 fixed-level interrupt sources • Each of the 63 sources has a unique interrupt control register (ICRn) to define the software-assigned levels and priorities within the level • Unique vector number for each interrupt source • Ability to mask any individual interrupt source, plus global mask-all capability • Support for both hardware and software interrupt acknowledge cycles • “Wake-up” signal from stop mode The 56 fully-programmable and seven fixed-level interrupt sources for each of the two interrupt controllers on the MCF548x handle the complete set of interrupt sources from all of the modules on the device. This section describes how the interrupt sources are mapped to the interrupt controller logic, and how interrupts are serviced. 13.1.1 68K/ColdFire Interrupt Architecture Overview Before continuing with the specifics of the MCF548x interrupt controller, a brief review of the interrupt architecture of the 68K/ColdFire family is appropriate. The interrupt architecture of ColdFire is exactly the same as the M68000 family, where there is a 3-bit encoded interrupt priority level sent from the interrupt controller to the core, providing 7 levels of interrupt requests. Level 7 represents the highest priority interrupt level, while level 1 is the lowest priority. The processor samples for active interrupt requests once per instruction by comparing the encoded priority level against a 3-bit interrupt mask value (I) contained in bits 10:8 of the machine’s status register (SR). If the priority level is greater than the SR[I] field at the sample point, the processor suspends normal instruction execution and initiates interrupt exception processing. Level 7 interrupts are treated as non-maskable and edge-sensitive within the processor, while levels 1-6 are treated as level-sensitive and may be masked depending on the value of the SR[I] field. For correct operation, the ColdFire requires that, once asserted, the interrupt source remain asserted until explicitly disabled by the interrupt service routine. During the interrupt exception processing, the CPU enters supervisor mode, disables trace mode, and then fetches an 8-bit vector from the interrupt controller. This byte-sized operand fetch is known as the interrupt acknowledge (IACK) cycle, with the ColdFire implementation using a special encoding of the transfer type and transfer modifier attributes to distinguish this data fetch from a “normal” memory access. The fetched data provides an index into the exception vector table that contains 256 addresses, each pointing to the beginning of a specific exception service routine. In particular, vectors 64–255 of the exception vector table are reserved for user interrupt service routines. The first 64 exception vectors are reserved for the processor to handle reset, error conditions (access, address), arithmetic faults, system calls, etc. Once the interrupt vector number has been retrieved, the processor continues by creating a stack frame in memory. For ColdFire, all exception stack frames are 2 longwords in length and contain 32 bits of vector MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 13-1 and status register data, along with the 32-bit program counter value of the instruction that was interrupted (see Section 3.8.1, “Exception Stack Frame Definition,” for more information on the stack frame format). After the exception stack frame is stored in memory, the processor accesses the 32-bit pointer from the exception vector table using the vector number as the offset, and then jumps to that address to begin execution of the service routine. After the status register is stored in the exception stack frame, the SR[I] mask field is set to the level of the interrupt being acknowledged, effectively masking that level and all lower values while in the service routine. For many peripheral devices, the processing of the IACK cycle directly negates the interrupt request, while other devices require that request to be explicitly negated during the processing of the service routine. For the MCF548x, the processing of the interrupt acknowledge cycle is fundamentally different than previous 68K/ColdFire cores. In the new approach, all IACK cycles are directly handled by the interrupt controller, so the requesting peripheral device is not accessed during the IACK. As a result, the interrupt request must be explicitly cleared in the peripheral during the interrupt service routine. For more information, see Section 13.1.1.1.3, “Interrupt Vector Determination.” Unlike the M68000 family, all ColdFire processors guarantee that the first instruction of the service routine is executed before sampling for interrupts is resumed. By making this initial instruction a load of the SR, interrupts can be safely disabled, if required. During the execution of the service routine, the appropriate actions must be performed on the peripheral to negate the interrupt request. For more information on exception processing, see the ColdFire Programmer’s Reference Manual at http://www.freescale.com/coldfire 13.1.1.1 Interrupt Controller Theory of Operation To support the interrupt architecture of the 68K/ColdFire programming model, the combined 63 interrupt sources are organized as 7 levels, with each level supporting up to nine prioritized requests. Consider the interrupt priority structure shown in Table 13-1, which orders the interrupt levels/priorities from highest to lowest. Table 13-1. Interrupt Priority Scheme Interrupt Level ICR[IL] Priority ICR[IP] Supported Interrupt Sources 7 6 #8–63 5 4 7 — (Mid-point) #7 (IRQ7) 3 2 #8–63 1 0 MCF548x Reference Manual, Rev. 5 13-2 Freescale Semiconductor Introduction Table 13-1. Interrupt Priority Scheme (Continued) Interrupt Level ICR[IL] 6 5 4 3 2 1 Priority ICR[IP] Supported Interrupt Sources 7–4 #8–63 — (Mid-point) #6 (IRQ6) 3–0 #8–63 7–4 #8–63 — (Mid-point) #5 (IRQ5) 3–0 #8–63 7–4 #8–63 — (Mid-point) #4 (IRQ4) 3–0 #8–63 7–4 #8–63 — (Mid-point) #3 (IRQ3) 3–0 #8–63 7–4 #8–63 — (Mid-point) #2 (IRQ2) 3–0 #8–63 7–4 #8–63 — (Mid-point) #1 (IRQ1) 3–0 #8–63 The level and priority is fully programmable for all sources except interrupt sources 1–7. Interrupt source 1–7 (the external interrupts) are fixed at the corresponding level’s midpoint priority. Thus, a maximum of eight fully-programmable interrupt sources are mapped into a single interrupt level. The fixed interrupt source is hardwired to the given level and represents the mid-point of the priority within the level. For the fully-programmable interrupt sources, the 3-bit level and the 3-bit priority within the level are defined in the 8-bit interrupt control register (ICRn). The operation of the interrupt controller can be broadly partitioned into three activities: • Recognition • Prioritization • Vector determination during IACK 13.1.1.1.1 Interrupt Recognition The interrupt controller continuously examines the request sources and the interrupt mask register to determine if there are active requests. This is the recognition phase. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 13-3 13.1.1.1.2 Interrupt Prioritization As an active request is detected, it is translated into the programmed interrupt level, and the resulting 7-bit decoded priority level (IRQ[7:1]) is driven out of the interrupt controller. 13.1.1.1.3 Interrupt Vector Determination Once the core has sampled for pending interrupts and begun interrupt exception processing, it generates an interrupt acknowledge cycle (IACK). The IACK transfer is treated as a memory-mapped byte read by the processor and routed to the interrupt controller. Next, the interrupt controller extracts the level being acknowledged from address bits[4:2], determines the highest priority interrupt request active for that level, and returns the 8-bit interrupt vector for that request to complete the cycle. The 8-bit interrupt vector is formed using the following algorithm: vector_number = 64 + interrupt source number Recall vector numbers 0—63 are reserved for the ColdFire processor and its internal exceptions. Thus, the mapping of bit positions to vector numbers that apply are the following: if interrupt source 1 is active and acknowledged, then vector_number = 65 if interrupt source 2 is active and acknowledged, then vector_number = 66 if interrupt source 8 is active and acknowledged, then vector_number = 72 if interrupt source 9 is active and acknowledged, then vector_number = 73 ... ... if interrupt source 63 is active and acknowledged, then vector_number = 127 The net effect is a fixed mapping between the bit position within the source to the actual interrupt vector number. If there is no active interrupt source for the given level, a special “spurious interrupt” vector (vector_number = 24) is returned, and it is the responsibility of the service routine to handle this error situation. Note this protocol implies the interrupting peripheral is not accessed during the acknowledge cycle since the interrupt controller completely services the acknowledge. This means the interrupt source must be explicitly cleared in the interrupt service routine. This design provides unique vector capability for all interrupt requests, regardless of the “complexity” of the peripheral device. Vector number 64 is unused. 13.2 Memory Map/Register Descriptions The register programming model for the interrupt controllers is memory-mapped to a 256-byte space. In the following discussion, there are a number of program-visible registers greater than 32 bits in size. For these control fields, the physical register is partitioned into two 32-bit values: a register “High” (the upper longword) and a register “Low” (the lower longword). The nomenclature <reg_name>H and <reg_name>L is used to reference these values. The registers and their locations are defined in Table 13-2. MCF548x Reference Manual, Rev. 5 13-4 Freescale Semiconductor Memory Map/Register Descriptions Table 13-2. Interrupt Controller Memory Map Address Offset Name Byte0 Byte1 Byte2 Byte3 Access 0x700 Interrupt Pending Register High [63:32] IPRH R 0x704 Interrupt Pending Register Low [31:0] IPRL R 0x708 Interrupt Mask Register High [63:32] IMRH R/W 0x70c Interrupt Mask Register Low [31:0] IMRL R/W 0x710 Interrupt Force Register High [63:32] INTFRCH R/W 0x714 Interrupt Force Register Low [31:0] INTFRCL R 0x718 Interrupt Request Level Register and Interrupt Acknowledge Level and Priority Register 0x71C– 0x73C — 0x740 Interrupt Control Registers IRLR[7:1] IACKLPR Reserved Reserved R — Reserved ICR01 ICR02 ICR03 R 0x744 ICR04 ICR05 ICR06 ICR07 R 0x748 ICR08 ICR09 ICR10 ICR11 R/W 0x74c ICR12 ICR13 ICR14 ICR15 R/W 0x750 ICR16 ICR17 ICR18 ICR19 R/W 0x754 ICR20 ICR21 ICR22 ICR23 R/W 0x758 ICR24 ICR25 ICR26 ICR27 R/W 0x75C ICR28 ICR29 ICR30 ICR31 R/W 0x760 ICR32 ICR33 ICR34 ICR35 R/W 0x764 ICR36 ICR37 ICR38 ICR39 R/W 0x768 ICR40 ICR41 ICR42 ICR43 R/W 0x76C ICR44 ICR45 ICR46 ICR47 R/W 0x770 ICR48 ICR49 ICR50 ICR51 R/W 0x774 ICR52 ICR53 ICR54 ICR55 R/W 0x778 ICR56 ICR57 ICR58 ICR59 R/W 0x77C ICR60 ICR61 ICR62 ICR63 R/W 0x780-0x7D C — Reserved 0x7E0 Software IACK Register SWIACK Reserved — R MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 13-5 Table 13-2. Interrupt Controller Memory Map (Continued) Address Offset Name Byte0 0x7E4 Level N IACK Registers L1IACK Reserved R 0x7E8 L2IACK Reserved R 0x7EC L3IACK Reserved R 0x7F0 L4IACK Reserved R 0x7F4 L5IACK Reserved R 0x7F8 L6IACK Reserved R 0x7FC L7IACK Reserved R 13.2.1 Byte1 Byte2 Byte3 Access Register Descriptions 13.2.1.1 Interrupt Pending Registers (IPRH, IPRL) The IPRH and IPRL registers, Figure 13-1 and Figure 13-2, are each 32 bits in size and provide a bit map for each interrupt request to indicate if there is an active request for the given source (1 = active request, 0 = no request). The state of the interrupt mask register does not affect the IPR. The IPR is cleared by reset. The IPR is a read-only register, so any attempted write to this register is ignored. Bit 0 is not implemented and reads as a zero. 31 30 29 28 27 26 25 R 24 23 22 21 20 19 18 17 16 INT[63:48] W Reset 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 0 0 0 0 0 0 R INT[47:32] W Reset 0 0 Reg Addr 0 0 0 0 0 0 0 MBAR + 0x700 Figure 13-1. Interrupt Pending Register High (IPRH) Table 13-3. IPRH Field Descriptions Bits Name 31–0 INT[63:32] Description Interrupt pending. Each bit corresponds to an interrupt source. The corresponding IMRH bit determines whether an interrupt condition can generate an interrupt. At every system clock, the IPRH samples the signal generated by the interrupting source. The corresponding IPRH bit reflects the state of the interrupt signal even if the corresponding IMRH bit is set. 0 The corresponding interrupt source does not have an interrupt pending 1 The corresponding interrupt source has an interrupt pending MCF548x Reference Manual, Rev. 5 13-6 Freescale Semiconductor Memory Map/Register Descriptions 31 30 29 28 27 26 25 R 24 23 22 21 20 19 18 17 16 INT[31:16] W Reset 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 R INT[15:1] 0 W Reset 0 0 0 0 0 0 0 Reg Addr 0 0 0 0 0 0 0 0 0 MBAR + 0x704 Figure 13-2. Interrupt Pending Register Low (IPRL) Table 13-4. IPRL Field Descriptions Bits Name Description 31–1 INT[31:1] Interrupt Pending. Each bit corresponds to an interrupt source. The corresponding IMRL bit determines whether an interrupt condition can generate an interrupt. At every system clock, the IPRL samples the signal generated by the interrupting source. The corresponding IPRL bit reflects the state of the interrupt signal even if the corresponding IMRL bit is set. 0 The corresponding interrupt source does not have an interrupt pending 1 The corresponding interrupt source has an interrupt pending 0 — 13.2.1.2 Reserved, should be cleared. Interrupt Mask Register (IMRH, IMRL) The IMRH and IMRL registers are each 32 bits in size and provide a bit map for each interrupt to allow the request to be disabled (1 = disable the request, 0 = enable the request). The IMR is set to all ones by reset, disabling all interrupt requests. The IMR can be read and written. A write that sets bit 0 of the IMR forces the other 63 bits to be set, disabling all interrupt sources and providing a global mask-all capability. NOTE If an interrupt source is masked in the interrupt controller mask register (IMR) or a module’s interrupt mask register while the interrupt mask in the status register (SR[I]) is set to a value lower than the interrupt’s level, a spurious interrupt may occur. This situation occurs because by the time the status register acknowledges the interrupt, it has been masked and the CPU cannot determine the interrupt source. To avoid this situation for interrupt sources with levels 1–6, first write a higher level interrupt mask to the status register before setting the mask in the IMR or the module’s interrupt mask register. After the mask is set, return the interrupt mask in the status register to its previous value. Since level 7 interrupts cannot be disabled in the status register prior to masking, use of the IMR or module interrupt mask registers to disable level 7 interrupts is not recommended. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 13-7 31 30 29 28 27 26 25 R 24 23 22 21 20 19 18 17 16 INT_MASK[63:48] W Reset 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 1 1 1 1 1 1 1 R INT_MASK[47:32] W Reset 1 1 1 1 1 1 1 Reg Addr 1 1 MBAR + 0x708 Figure 13-3. Interrupt Mask Register High (IMRH) Table 13-5. IMRH Field Descriptions Bits 31–0 31 Name Description INT_MASK Interrupt mask. Each bit corresponds to an interrupt source. The corresponding IMRH bit determines whether an interrupt condition can generate an interrupt. The corresponding IPRH bit reflects the state of the interrupt signal even if the corresponding IMRH bit is set. 0 The corresponding interrupt source is not masked 1 The corresponding interrupt source is masked 30 29 28 27 26 25 R 24 23 22 21 20 19 18 17 16 INT_MASK[31:16] W Reset 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 R INT_MASK[15:1] MASK ALL W Reset Reg Addr 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 MBAR + 0x70C Figure 13-4. Interrupt Mask Register Low (IMRL) MCF548x Reference Manual, Rev. 5 13-8 Freescale Semiconductor Memory Map/Register Descriptions Table 13-6. IMRL Field Descriptions Bits Name 31–1 INT_MASK Interrupt mask. Each bit corresponds to an interrupt source. The corresponding IMRL bit determines whether an interrupt condition can generate an interrupt. The corresponding IPRL bit reflects the state of the interrupt signal even if the corresponding IMRL bit is set. 0 The corresponding interrupt source is not masked 1 The corresponding interrupt source is masked 0 13.2.1.3 Description MASKALL Mask all interrupts. Setting this bit will force the other 63 bits of the IMRH and IMRL to ones, disabling all interrupt sources, and providing a global mask-all capability. Interrupt Force Registers (INTFRCH, INTFRCL) The INTFRCH and INTFRCL registers are each 32 bits in size and provide a mechanism to allow software generation of interrupts for each possible source for functional or debug purposes. The system design may reserve one or more sources to allow software to self-schedule interrupts by forcing one or more of these bits in the appropriate INTFRC register (1 = force request, 0 = negate request). The assertion of an interrupt request via the INTFRC register is not affected by the interrupt mask register. The INTFRC register is cleared by reset. 31 30 29 28 27 26 25 R 24 23 22 21 20 19 18 17 16 INTFRC[63:48] W Reset 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 0 0 0 0 0 0 R INTFRC[47:32] W Reset 0 0 Reg Addr 0 0 0 0 0 0 0 MBAR + 0x710 Figure 13-5. Interrupt Force Register High (INTFRCH) Table 13-7. INTFRCH Field Descriptions Bits Name Description 31–0 INTFRC Interrupt force. Allows software generation of interrupts for each possible source for functional or debug purposes. 0 No interrupt forced on corresponding interrupt source 1 Force an interrupt on the corresponding source MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 13-9 31 30 29 28 27 26 25 24 R 23 22 21 20 19 18 17 16 INTFRC[31:16] W Reset 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 R INTFRC[16:1] — W Reset 0 0 0 0 0 0 0 0 Reg Addr 0 0 0 0 0 0 0 0 MBAR + 0x714 Figure 13-6. Interrupt Force Register Low (INTFRCL) . Table 13-8. INTFRCL Field Descriptions Bits Name Description 31–1 INTFRC Interrupt force. Allows software generation of interrupts for each possible source for functional or debug purposes. 0 No interrupt forced on corresponding interrupt source 1 Force an interrupt on the corresponding source 0 — 13.2.1.4 Reserved, should be cleared. Interrupt Request Level Register (IRLR) This 7-bit register is updated each machine cycle and represents the current interrupt requests for each interrupt level, where bit 7 corresponds to level 7, bit 6 to level 6, etc. 7 6 5 R 4 3 2 1 0 IRQ 0 W Reset 0 0 Reg Addr 0 0 0 0 0 0 MBAR + 0x718 Figure 13-7. Interrupt Request Level Register (IRLR) Table 13-9. IRQn Field Descriptions Bits Name 7–1 IRQ 0 — Description Interrupt requests. Represents the prioritized active interrupts for each level. 0 There are no active interrupts at this level 1 There is an active interrupt at this level Reserved MCF548x Reference Manual, Rev. 5 13-10 Freescale Semiconductor Memory Map/Register Descriptions 13.2.1.5 Interrupt Acknowledge Level and Priority Register (IACKLPR) Each time an IACK is performed, the interrupt controller responds with the vector number of the highest priority source within the level being acknowledged. In addition to providing the vector number directly for the byte-sized IACK read, this 8-bit register is also loaded with information about the interrupt level and priority being acknowledged. This register provides the association between the acknowledged “physical” interrupt request number and the programmed interrupt level/priority. The contents of this read-only register are described in Figure 13-8 and Table 13-10. 7 R 6 5 — 4 3 2 LEVEL 1 0 0 0 PRI W Reset 0 0 0 Reg Addr 0 0 0 MBAR + 0x719 Figure 13-8. IACK Level and Priority Register (IACKLPR) Table 13-10. IACKLPR Field Descriptions Bits Name 7 — 6–4 LEVEL 3–0 PRI 13.2.1.6 Description Reserved Interrupt level. Represents the interrupt level currently being acknowledged. Interrupt Priority. Represents the priority within the interrupt level of the interrupt currently being acknowledged. 0 Priority 0 1 Priority 1 2 Priority 2 3 Priority 3 4 Priority 4 5 Priority 5 6 Priority 6 7 Priority 7 8 Mid-Point Priority associated with the fixed level interrupts only Interrupt Control Registers 1–63 (ICRn) Each ICRn specifies the interrupt level (1–7) and the priority within the level (0–7). All ICRn registers can be read, but only ICR8 to ICR63 can be written. It is software’s responsibility to program the ICRn registers with unique and non-overlapping level and priority definitions. Failure to program the ICRn registers in this matter can result in undefined behavior. If a specific interrupt request is completely unused, the ICRn value can remain in its reset (and disabled) state. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 13-11 R 7 6 0 0 0 0 5 4 3 2 IL 1 0 IP W Reset 0 Reg Addr 0 0 0 0 0 See Table 13-2 for register offsets Figure 13-9. Interrupt Control Registers 1–63 (ICRn) Table 13-11. ICRn Field Descriptions Bits Name 7–6 — Reserved, should be cleared. 5–3 IL Interrupt level. Indicates the interrupt level assigned to each interrupt input. 2–0 IP Interrupt priority. Indicates the interrupt priority for internal modules within the interrupt-level assignment. 000b represents the lowest priority and 111b represents the highest. For the fixed level interrupt sources, the priority is fixed at the midpoint for the level, and the IP field will always read as 000b. 13.2.1.6.1 Description Interrupt Sources Table 13-12 lists the interrupt sources for each interrupt request line Table 13-12. Interrupt Source Assignments Sourc e Module Flag 1 EPORT EPF1 Edge port flag 1 Write ‘1’ to EPFR[EPF1] 2 EPF2 Edge port flag 2 Write ‘1’ to EPFR[EPF2] 3 EPF3 Edge port flag 3 Write ‘1’ to EPFR[EPF3] 4 EPF4 Edge port flag 4 Write ‘1’ to EPFR[EPF4] 5 EPF5 Edge port flag 5 Write ‘1’ to EPFR[EPF5] 6 EPF6 Edge port flag 6 Write ‘1’ to EPFR[EPF6] 7 EPF7 Edge port flag 7 Write ‘1’ to EPFR[EPF7] 8–14 Source Description Flag Clearing Mechanism Not used MCF548x Reference Manual, Rev. 5 13-12 Freescale Semiconductor Memory Map/Register Descriptions Table 13-12. Interrupt Source Assignments (Continued) Sourc e Module Flag 15 USB 2.0 EP0ISR Endpoint 0 interrupt Write ‘1’ to appropriate bit in EP0ISR 16 EP1ISR Endpoint 1 interrupt Write ‘1’ to appropriate bit in EP1ISR 17 EP2ISR Endpoint 2 interrupt Write ‘1’ to appropriate bit in EP2ISR 18 EP3ISR Endpoint 3 interrupt Write ‘1’ to appropriate bit in EP3ISR 19 EP4ISR Endpoint 4 interrupt Write ‘1’ to appropriate bit in EP4ISR 20 EP5ISR Endpoint 5 interrupt Write ‘1’ to appropriate bit in EP5ISR 21 EP6ISR Endpoint 6 interrupt Write ‘1’ to appropriate bit in EP6ISR 22 USBISR USB 2.0 general interrupt Write ‘1’ to appropriate bit in USBISR 23 Source Description USBAISR USB 2.0 core interrupt 24 Write ‘0’ to appropriate bit in USBAISR OR of all USB interrupts Clear appropriate USB interrupt(s) RFOF | TFUF DSPI overflow or underflow Write ‘1’ to DSR[RFDF] and/or DSR[TFUF] 26 RFOF Receive FIFO overflow interrupt Write ‘1’ to DSR[RFOF] 27 RFDF Receive FIFO drain interrupt Write ‘1’ to DSR[RFDF] or DMA acknowledge 28 TFUF Transmit FIFO underflow interrupt Write ‘1’ to DSR[TFUF] 29 TCF Transfer complete interrupt Write ‘1’ to DSR[TCF] 30 TFFF Transfer FIFO fill interrupt Write ‘1’ to DSR[TFFF] or DMA acknowledge 31 EOQF End of queue interrupt Write ‘1’ to DSR[EOQF] 25 — Flag Clearing Mechanism DSPI 32 PSC3 — PSC3 interrupt Cleared when service complete 33 PSC2 — PSC2 interrupt Cleared when service complete 34 PSC1 — PSC1 interrupt Cleared when service complete 35 PSC0 — PSC0 interrupt Cleared when service complete 36 CommTim TC Combined interrupts from comm timers Write ‘1’ to CTCRn[I] 37 SEC — SEC interrupt Service interrupt and write ‘1’ to SICR 38 FEC1 — FEC1 interrupt Write appropriate interrupt condition bit = 1 39 FEC0 — FEC0 interrupt Write appropriate interrupt condition bit = 1 40 I2C — I2C interrupt Write IIF = 0 41 PCIARB — PCI arbiter interrupt Write ‘1’ to PASR[EXTMBK] or PASR[ITLMBK] 42 CBPCI — Comm bus PCI interrupt Clear FIFO alarm condition 43 XLBPCI — XLB PCI interrupt Write ‘1’ to appropriate PCIISR bit(s) 44–46 47 Not used XLBARB — XLBARB to CPU interrupt Write ‘1’ to appropriate ARB_SR bit(s) MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 13-13 Table 13-12. Interrupt Source Assignments (Continued) Sourc e Module Flag 48 DMA — 49 CAN0 Source Description Flag Clearing Mechanism Multichannel DMA interrupt Write ‘1’ to DIPR[TASKn] ERROR FlexCAN error interrupt Read error bits in ESR or write ERR_INT = 0 50 BUSOFF FlexCAN bus off interrupt Write BOFF_INT = 0 51 MBOR Message buffer ORed interrupt Write BUFnI = 1 after reading BUFnI = 1 52 53 Not used Slice Timer SLT1 Slice timer 1 interrupt Write ST = 1 SLT0 Slice timer 0 interrupt Write ST = 1 CAN1 ERROR FlexCAN error interrupt Read error bits in ESR or write ERR_INT = 0 56 BUSOFF FlexCAN bus off interrupt Write BOFF_INT = 0 57 MBOR Message buffer ORed interrupt Write BUFnI = 1 after reading BUFnI = 1 54 55 58 59 Not used GPT3 GPT3 interrupt Write ‘1’ to appropriate GSR bit 60 GPT2 GPT2 interrupt Write ‘1’ to appropriate GSR bit 61 GPT1 GPT1 interrupt Write ‘1’ to appropriate GSR bit 62 GPT0 GPT0 interrupt Write ‘1’ to appropriate GSR bit 63 13.2.1.7 GPTs Not used Software and Level n IACK Registers (SWIACKR, L1IACK–L7IACK) The eight IACK registers can be explicitly addressed via the CPU, or implicitly addressed via a processor-generated interrupt acknowledge cycle during exception processing. In either case, the interrupt controller’s actions are very similar. First, consider an IACK cycle to a specific level: that is, a level-n IACK. When this type of IACK arrives in the interrupt controller, the controller examines all the currently-active level-n interrupt requests, determines the highest priority within the level, and then responds with the unique vector number corresponding to that specific interrupt source. The vector number is supplied as the data for the byte-sized IACK read cycle. In addition to providing the vector number, the interrupt controller also loads the level and priority number for the level into the IACKLPR, where it may be retrieved later. This interrupt controller design also supports the concept of a software IACK. A software IACK is a useful concept that allows an interrupt service routine to determine if there are other pending interrupts, so that the overhead associated with interrupt exception processing (including machine state save/restore functions) can be minimized. In general, the software IACK is performed near the end of an interrupt service routine, and if there are additional active interrupt sources, the current interrupt service routine (ISR) passes control to the appropriate service routine, but without taking another interrupt exception. When the interrupt controller receives a software IACK read, it returns the vector number associated with the highest level, highest priority unmasked interrupt source for that interrupt controller. The IACKLPR is also loaded as the software IACK is performed. If there are no active sources, the interrupt controller returns an all-zero vector as the operand. For this situation, the IACKLPR is also cleared. MCF548x Reference Manual, Rev. 5 13-14 Freescale Semiconductor Memory Map/Register Descriptions In addition to the software IACK registers within each interrupt controller, there are global software IACK registers. A read from the global SWIACK will return the vector number for the highest level and priority unmasked interrupt source from all interrupt controllers. A read from one of the LnIACK registers will return the vector for the highest priority unmasked interrupt within a level for all interrupt controllers. 7 6 5 4 R 3 2 1 0 0 0 0 0 VECTOR W Reset 0 0 Reg Addr 0 0 See Table 13-2 for register offsets Figure 13-10. Software and Level n IACK Registers (SWIACKR, L1IACK–L7IACK) Table 13-13. SWIACK and L1IACK–L7IACK Field Descriptions Bits Name Description 7–0 VECTOR Vector number. A read from the SWIACK register returns the vector number associated with the highest level, highest priority unmasked interrupt source. A read from one of the LnACK registers returns the highest priority unmasked interrupt source within the level. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 13-15 MCF548x Reference Manual, Rev. 5 13-16 Freescale Semiconductor Chapter 14 Edge Port Module (EPORT) 14.1 Introduction The edge port module (EPORT) has seven external interrupt pins, IRQ[7:1]. Each pin can be configured individually as a level-sensitive interrupt pin, an edge-detecting interrupt pin (rising edge, falling edge, or both), or a general-purpose input/output (I/O) pin. See Figure 14-1. Stop Mode EPPAR[2n, 2n + 1] Edge Detect Logic EPFRn D0 Internal Bus Q D0 D1 Q D1 To Interrupt Controller EPPDRn Synchronizer Rising Edge of System Clock EPIERn EPDRn IRQn PIN EPDDRn Figure 14-1. EPORT Block Diagram 14.2 Interrupt/General-Purpose I/O Pin Descriptions All interrupt pins default to general-purpose input pins at reset. The pin value is synchronized to the rising edge of the internal clock when read from the EPORT pin data register (EPPDR). The values used in the edge/level detect logic are also synchronized to the rising edge of the internal clock. These pins use Schmitt-triggered input buffers which have built-in hysteresis that decrease the probability of generating false edge-triggered interrupts for slow rising and falling input signals. When a pin is configured as an output, it is driven to a state whose level is determined by the corresponding bit in the EPORT data register (EPDR). All bits in the EPDR are high at reset. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 14-1 NOTE The GPIO functionality of the external interrupt pins is controlled by the EPORT module. However, some external interrupt signals are muxed with other functions. In this case, the pin’s IRQ functionality must be enabled in the GPIO module’s pin assignment register in order to use the pin’s GPIO function via the EPORT registers. For more information, refer to Chapter 15, “GPIO.” 14.3 Memory Map/Register Definition This subsection describes the memory map and register structure. 14.3.1 Memory Map Refer to Table 14-1 for a description of the EPORT memory map. The EPORT has an MBAR offset for base address of 0xF00. Table 14-1. Edge Port Module Memory Map MBAR Offset Name 0xF00 EPORT pin assignment register 0xF04 EPORT data direction register EPORT interrupt enable register EPDDR 0xF08 EPORT data register EPORT pin data register EPDR 0xF0C EPORT flag register EPFR Byte0 Byte1 Byte2 Byte3 Access1 —2 S EPIER —2 S/U EPPDR —2 EPPAR —2 1 S = CPU supervisor mode access only. S/U = CPU supervisor or user mode access. User mode accesses to supervisor only addresses have no effect and result in a cycle termination transfer error. 2 Writing to reserved address locations has no effect, and reading returns 0s. 14.3.2 Register Descriptions The EPORT programming model consists of these registers: • The EPORT pin assignment register (EPPAR) controls the function of each pin individually. • The EPORT data direction register (EPDDR) controls the direction of each pin individually. • The EPORT interrupt enable register (EPIER) enables interrupt requests for each pin individually. • The EPORT data register (EPDR) holds the data to be driven to the pins. • The EPORT pin data register (EPPDR) reflects the current state of the pins. • The EPORT flag register (EPFR) individually latches EPORT edge events. MCF548x Reference Manual, Rev. 5 14-2 Freescale Semiconductor Memory Map/Register Definition 14.3.2.1 15 R EPORT Pin Assignment Register (EPPAR) 14 13 EPPA7 12 11 EPPA6 10 9 EPPA5 8 EPPA4 7 6 5 EPPA3 4 3 EPPA2 2 EPPA1 1 0 0 0 0 0 W Reset 0 0 0 0 0 0 0 0 Reg Addr 0 0 0 0 0 0 MBAR + 0xF00 Figure 14-2. EPORT Pin Assignment Register (EPPAR) Table 14-2. EPPAR Field Descriptions Bits Name Description 15–2 EPPAn EPORT pin assignment select fields. The read/write EPPAn fields configure EPORT pins for level detection and rising and/or falling edge detection. Pins configured as level-sensitive are inverted so that a logic 0 on the external pin represents a valid interrupt request. Level-sensitive interrupt inputs are not latched. To guarantee that a level-sensitive interrupt request is acknowledged, the interrupt source must keep the signal asserted until acknowledged by software. Level sensitivity must be selected to bring the device out of stop mode with an IRQn interrupt. Pins configured as edge-triggered are latched and need not remain asserted for interrupt generation. A pin configured for edge detection can trigger an interrupt regardless of its configuration as input or output. Interrupt requests generated in the EPORT module can be masked by the interrupt controller module. EPPAR functionality is independent of the selected pin direction. Reset clears the EPPAn fields. 00 Pin IRQn level-sensitive 01 Pin IRQn rising edge triggered 10 Pin IRQn falling edge triggered 11 Pin IRQn both falling edge and rising edge triggered 1–0 — 14.3.2.2 Reserved, should be cleared. EPORT Data Direction Register (EPDDR) R 7 6 5 4 3 2 1 0 EPDD7 EPDD6 EPDD5 EPDD4 EPDD3 EPDD2 EPDD1 0 0 0 0 0 0 0 0 0 W Reset Reg Addr MBAR + 0xF04 Figure 14-3. EPORT Data Direction Register (EPDDR) MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 14-3 Table 14-3. EPDDR Field Descriptions Bits Name Description 7–1 EPDDn Setting any bit in the EPDDR configures the corresponding pin as an output. Clearing any bit in EPDDR configures the corresponding pin as an input. Pin direction is independent of the level/edge detection configuration. Reset clears EPDD7–EPDD1. To use an EPORT pin as an external interrupt request source, its corresponding bit in EPDDR must be clear. Software can generate interrupt requests by programming the EPORT data register when the EPDDR selects output. 0 Corresponding EPORT pin configured as input 1 Corresponding EPORT pin configured as output 0 — 14.3.2.3 Reserved, should be cleared. Edge Port Interrupt Enable Register (EPIER) R 7 6 5 4 3 2 1 0 EPIE7 EPIE6 EPIE5 EPIE4 EPIE3 EPIE2 EPIE1 0 0 0 0 0 0 0 0 0 W Reset Reg Addr MBAR + 0xF05 Figure 14-4. EPORT Port Interrupt Enable Register (EPIER) Table 14-4. EPIER Field Descriptions Bits Name 7–1 EPIEn 0 — 14.3.2.4 Description Edge port interrupt enable bits enable EPORT interrupt requests. If a bit in EPIER is set, EPORT generates an interrupt request when: • The corresponding bit in the EPORT flag register (EPFR) is set or later becomes set. • The corresponding pin level is low and the pin is configured for level-sensitive operation. Clearing a bit in EPIER negates any interrupt request from the corresponding EPORT pin. Reset clears EPIE7–EPIE1. 0 Interrupt requests from corresponding EPORT pin disabled 1 Interrupt requests from corresponding EPORT pin enabled Reserved, should be cleared. Edge Port Data Register (EPDR) R 7 6 5 4 3 2 1 0 EPD7 EPD6 EPD5 EPD4 EPD3 EPD2 EPD1 0 1 1 1 1 1 1 1 1 W Reset Reg Addr MBAR + 0xF08 Figure 14-5. EPORT Port Data Register (EPDR) MCF548x Reference Manual, Rev. 5 14-4 Freescale Semiconductor Memory Map/Register Definition Table 14-5. EPDR Field Descriptions Bits Name Description 7–1 EPDx Edge port data bits. Data written to EPDR is stored in an internal register; if any pin of the port is configured as an output, the bit stored for that pin is driven onto the pin. Reading EDPR returns the data stored in the register. Reset sets EPD7-EPD1. 0 — 14.3.2.5 Reserved, should be cleared. Edge Port Pin Data Register (EPPDR) R 7 6 5 4 3 2 1 0 EPPD7 EPPD6 EPPD5 EPPD4 EPPD3 EPPD2 EPPD1 0 W Reset Current pin state Reg Addr 0 MBAR + 0xF09 Figure 14-6. EPORT Port Pin Data Register (EPPDR) Table 14-6. EPPDR Field Descriptions Bits Name Description 7–1 EPPDx Edge port pin data bits. The read-only EPPDR reflects the current state of the EPORT pins IRQ7–IRQ1. Writing to EPPDR has no effect, and the write cycle terminates normally. Reset does not affect EPPDR. 0 — 14.3.2.6 Reserved, should be cleared. Edge Port Flag Register (EPFR) R 7 6 5 4 3 2 1 0 EPF7 EPF6 EPF5 EPF4 EPF3 EPF2 EPF1 0 0 0 0 0 0 0 0 0 W Reset Reg Addr MBAR + 0xF0C Figure 14-7. EPORT Port Flag Register (EPFR) MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 14-5 Table 14-7. EPFR Field Descriptions Bits Name 7–1 EPFn 0 — Description Edge port flag bits. When an EPORT pin is configured for edge triggering, its corresponding read/write bit in EPFR indicates that the selected edge has been detected. Reset clears EPF7–EPF1. Bits in this register are set when the selected edge is detected on the corresponding pin. A bit remains set until cleared by writing a 1 to it. Writing 0 has no effect. If a pin is configured as level-sensitive (EPPARn = 00), pin transitions do not affect this register. 0 Selected edge for IRQx pin has not been detected. 1 Selected edge for IRQx pin has been detected. Reserved, should be cleared. MCF548x Reference Manual, Rev. 5 14-6 Freescale Semiconductor Chapter 15 GPIO 15.1 Introduction Many of the MCF548x pins whose primary function is to serve as the external interface to off-chip resources may also be used for general-purpose digital I/O (GPIO) access and for one or two secondary functions. When used for GPIO purposes, the port x pins (PXXX) indicate which port is being accessed. In some cases, the pin function is set by the operating mode, and the alternate pin functions are not supported. The MCF548x GPIO signals are grouped into 8-bit ports; however, some ports do not use all eight bits. Each GPIO port has registers that configure, monitor, and control the port signals. Figure 15-1 is a block diagram of the MCF548x GPIO module. NOTE The actual signals and functions available vary for different members of the MCF548x family. See Chapter 2, “Signal Descriptions,” for more details. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 15-1 PORT FBCTL PORT FBCS PORT DMA PORT FEC0H PORT FEC0L PORT FEC1H PORT FEC1L PORT FECI2C BWE[3:0] / PFBCTL[7:4] OE / PFBCTL3 R/W / PFBCTL2 TA / PFBCTL1 ALE / PFBCTL0 PORT PCIBG PCIBG[4:0] / PPCIBG[4:0] PORT PCIBR PCIBR[4:0] / PPCIBR[4:0] FBCS[5:1] / PFBCS[5:1] DACK[1:0] / PDMA[3:2] DREQ[1:0] / PDMA[1:0] PORT PSC3PSC2 FEC0TXCLK / PFEC0H7 FEC0TXEN / PFEC0H6 FEC0TXD0 / PFEC0H5 FEC0COL / PFEC0H4 FEC0RXCLK / PFEC0H3 FEC0RXDV / PFEC0H2 FEC0RXD0 / PFEC0H1 FEC0CRS / PFEC0H0 PORT PSC1PSC0 FEC0TXD3 / PFEC0L7 FEC0TXD2 / PFEC0L6 FEC0TXD1 / PFEC0L5 FEC0TXER / PFEC0L4 FEC0RXD3 / PFEC0L3 FEC0RXD2 / PFEC0L2 FEC0RXD1 / PFEC0L1 FEC0RXER / PFEC0L0 FEC1TXCLK / PFEC1H7 FEC1TXEN / PFEC1H6 FEC1TXD0 / PFEC1H5 FEC1COL / PFEC1H4 FEC1RXCLK / PFEC1H3 FEC1RXDV / PFEC1H2 FEC1RXD0 / PFEC1H1 FEC1CRS / PFEC1H0 PORT DSPI PORT IRQ1 FEC1TXD3 / PFEC1L7 FEC1TXD2 / PFEC1L6 FEC1TXD1 / PFEC1L5 FEC1TXER / PFEC1L4 FEC1RXD3 / PFEC1L3 FEC1RXD2 / PFEC1L2 FEC1RXD1 / PFEC1L1 FEC1RXER / PFEC1L0 PORT TIM2 PSC3CTS / PPSC3PSC27 PSC3RTS / PPSC3PSC26 PSC3RXD / PPSC3PSC25 PSC3TXD / PPSC3PSC24 PSC2CTS / PPSC3PSC23 PSC2RTS / PPSC3PSC22 PSC2RXD / PPSC3PSC21 PSC2TXD / PPSC3PSC20 PSC1CTS / PPSC1PSC07 PSC1RTS / PPSC1PSC06 PSC1RXD / PPSC1PSC05 PSC1TXD / PPSC1PSC04 PSC0CTS / PPSC1PSC03 PSC0RTS / PPSC1PSC02 PSC0RXD / PPSC1PSC01 PSC0TXD / PPSC1PSC00 DSPIPCS5 / PCSS / PDSPI6 DSPIPCS3 / PDSPI5 DSPIPCS2 / PDSPI4 DSPIPCS0 / SS / PDSPI3 DSPISCK / PDSPI2 DSPISIN / PDSPI1 DSPISOUT / PDSPI1 IRQ[7:1] / PIRQ[5:1] TOUT[3:0] / PTIM[7:4] TIN[3:0] / PTIM[3:0] FEC0EMDIO / PFECI2C3 FEC0EMDC / PFECI2C2 SCL / PFECI2C1 SDA / PFECI2C0 1 The 2 The port IRQ GPIO functionality is provided through the EPORT module. port TIM GPIO funtionality is provided through the GPT module. Figure 15-1. MCF548x GPIO Module Block Diagram 15.1.1 Overview The MCF548x GPIO module controls the configuration and use for the following external GPIO ports (register types in parentheses): • ColdFire bus (FlexBus) accesses (FBCTL, FBCS) MCF548x Reference Manual, Rev. 5 15-2 Freescale Semiconductor External Pin Description • • • • • • External DMA request and acknowledge (DMA) PCI bus access (PCIGNT, PCIREQ) Ethernet data and control (FEC0H, FEC0L, FEC1H, FEC1L, FECI2C) I2C serial control (FECI2C) DMA serial peripheral interface (DSPI) Programmable serial control (PSC1PSC0 and PSC3PSC2) 15.1.2 Features The MCF548x GPIO module includes these distinctive features: • Control of primary function use of the supported GPIO ports indicated in Section 15.1.1, “Overview” • General purpose I/O support for all ports: — Registers for storing output pin data — Registers for controlling pin data direction — Registers for reading current pin state — Registers for setting and clearing output pin data registers 15.2 External Pin Description The MCF548x GPIO module controls the functionality of several external pins. These pins are listed in Table 15-1. Table 15-1. MCF548x GPIO Module External Pins Primary Function (Pin Name)1 GPIO Alternate Function 1 Alternate Function 2 Description Flexbus Control BWE[3:2] PFBCTL[7:6] BE / BWE[3:2] TSIZ[1:0] Byte write strobes for external data transfer / Port FBCTL[7:4] / Byte enables for external data transfer / FlexBus transfer size BWE[1:0] PFBCTL[5:4] BE / BWE[1:0] FBADDR[1:0] Byte write strobes for external data transfer / Port FBCTL[7:4] / Byte enables for external data transfer / FlexBus address[1:0] OE PFBCTL3 — — Output enable for external reads / Port FBCTL3 R/W PFBCTL2 TBST — Read/write indication for external data transfer / Port FBCTL2 / FlexBus transfer burst TA PFBCTL1 — — Transfer acknowledge for external data transfer / Port FBCTL1 ALE PFBCTL0 TBST — Address latch enable indication for external data transfer / Port FBCTL0 / FlexBus transfer burst Flexbus Chip Selects FBCS[5:1] PFBCS[5:1] — — Flexbus chip selects 5 – 1 / Port FBCS[5:4] DMA Controller DACK1 PDMA3 TOUT1 — DMA acknowledge 1 / Port DMA3 / GP timer output 1 MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 15-3 Table 15-1. MCF548x GPIO Module External Pins (Continued) Primary Function (Pin Name)1 GPIO Alternate Function 1 Alternate Function 2 DACK0 PDMA2 TOUT0 — DREQ1 PDMA1 TIN1 IRQ1 DREQ0 PDMA0 TIN0 — Description DMA acknowledge 0 / Port DMA2 / GP timer output 0 DMA request 1 / Port DMA1 / GP timer input 1 / Interrupt 1 DMA request 0 / Port DMA0 / GP timer input 0 Fast Ethernet Controller 0 FEC0TXCLK PFEC0H7 — — Ethernet Controller 0 transmit clock / Port FEC0H7 FEC0TXEN PFEC0H6 — — Ethernet Controller 0 transmit enable / PFEC0H6 FEC0TXD0 PFEC0H5 — — Ethernet Controller 0 transmit data 0 / Port FEC0H5 FEC0COL PFEC0H4 — — Ethernet Controller 0 collision / Port FEC0H4 FEC0RXCLK PFEC0H3 — — Ethernet Controller 0 receive clock / Port FEC0H3 FEC0RXDV PFEC0H2 — — Ethernet Controller 0 receive data valid / Port FEC0H2 FEC0RXD0 PFEC0H1 — — Ethernet Controller 0 receive data 0 / Port FEC0H1 FEC0CRS PFEC0H0 — — Ethernet Controller 0 carrier receive sense / Port FEC0H0 FEC0TXD[3:1] PFEC0L[7:5] — — Ethernet Controller 0 transmit data / Port FEC0L[7:5] FEC0TXER PFEC0L4 — — Ethernet Controller 0 transmit error / Port FEC0L4 FEC0RXD[3:1] PFEC0L[3:1] — — Ethernet Controller 0 receive data [3:1] / Port FEC0L[3:1] FEC0RXER PFEC0L0 — — Ethernet Controller 0 receive error / Port FEC0L0 FEC0MDIO PFECI2C3 — — Ethernet Controller 0 management data control / Port FECI2C3 FEC0MDC PFECI2C2 — — Ethernet Controller 0 management data clock / Port FECI2C2 Fast Ethernet Controller 1 FEC1TXCLK PFEC1H7 — — Ethernet Controller 1 transmit clock / Port FEC1H7 FEC1TXEN PFEC1H6 — — Ethernet Controller 1 transmit enable / Port FEC1H6 FEC1TXD0 PFEC1H5 — — Ethernet Controller 1 transmit data 0 / Port FEC1H5 FEC1COL PFEC1H4 — — Ethernet Controller 1 collision / Port FEC1H4 FEC1RXCLK PFEC1H3 — — Ethernet Controller 1 receive clock / Port FEC1H3 FEC1RXDV PFEC1H2 — — Ethernet Controller 1 receive data valid / Port FEC1H2 FEC1RXD0 PFEC1H1 — — Ethernet Controller 1 receive data 0 / Port FEC1H1 FEC1CRS PFEC1H0 — — Ethernet Controller 1 carrier receive sense / Port FEC1H0 FEC1TXD[3:1] PFEC1L[7:5] — — Ethernet Controller 1 transmit data / Port FEC1L[7:5] FEC1TXER PFEC1L4 — — Ethernet Controller 1 transmit error / Port FEC1L4 FEC1RXD[3:1] PFEC1L[3:1] — — Ethernet Controller 1 receive data [3:1] / Port FEC1L[3:1] FEC1RXER PFEC1L0 — — Ethernet Controller 1 receive error / Port FEC1L0 FEC1MDIO — SDA CANRX0 Ethernet Controller 1 management data control / I2C serial data / FlexCAN 0 receive data MCF548x Reference Manual, Rev. 5 15-4 Freescale Semiconductor External Pin Description Table 15-1. MCF548x GPIO Module External Pins (Continued) Primary Function (Pin Name)1 GPIO Alternate Function 1 Alternate Function 2 Description FEC1MDC — SCL CANTX0 Ethernet Controller 1 management data clock / I2C serial clock / FlexCAN 0 transmit data I2C Serial Control SDA PFECI2C1 — — I2C serial data / Port FECI2C1 SCL PFECI2C0 — — I2C serial clock / Port FECI2C0 External Interrupts IRQ6 2 PIRQ6 CANRX1 — Interrupt 6 / Port IRQ6 / FlexCAN 1 receive data IRQ5 PIRQ52 CANRX1 — Interrupt 5 / Port IRQ5 / FlexCAN 1 receive data DMA Serial Peripheral Interface DSPICS5/PCSS PDSPI6 — DSPICS3 PDSPI5 TOUT3 DSPICS2 PDSPI4 TOUT2 DSPICS0/SS PDSPI3 PSC3RTS PSC3FSYNC DSPISCK PDSPI2 PSC3CTS PSC3BCLK DSPISIN PDSPI1 PSC3RXD — DSPI serial data input / Port DSPI1 / PSC3 receive data DSPISOUT PDSPI0 PSC3TXD — DSPI serial data output / Port DSPI 0 / PSC3 transmit data — DSPI synchronous peripheral chip select 3 / Port DSPI6 CANTX1 DSPI synchronous peripheral chip select 3 / Port DSPI5 / GP timer out 3 / FlexCAN 1 transmit data CANTX1 DSPI synchronous peripheral chip select 3 / Port DSPI4 / GP timer out 2 / FlexCAN 1 transmit data DSPI synchronous peripheral chip select 3 / Port DSPI3 / PSC3 request-to-send / PSC3 frame sync DSPI serial clock / Port DSPI2 / PSC3 clear-to-send / PSC3 modem clock Programmable Serial Control Module 3 PSC3CTS PPSC3PSC27 PSC3BCLK — PSC3 clear-to-send indication / Port PSC3PSC27 / PSC3 modem clock PSC3RTS PPSC3PSC26 PSC3FSYNC — PSC3 request-to-send indication / Port PSC3PSC26 / PSC3 frame sync PSC3RXD PPSC3PSC25 — — PSC3 receive data / Port PSC3PSC25 PSC3TXD PPSC3PSC24 — — PSC3 transmit data / Port PSC3PSC24 Programmable Serial Control Module 2 PSC2CTS PPSC3PSC23 PSC2BCLK CANRX0 PSC2 clear-to-send indication / Port PSC3PSC23 / PSC2 modem clock PSC2RTS PPSC3PSC22 PSC2FSYNC CANTX0 PSC2 request-to-send indication / Port PSC3PSC22 / PSC2 frame sync PSC2RXD PPSC3PSC21 — — PSC2 receive data / Port PSC3PSC21 PSC2TXD PPSC3PSC20 — — PSC2 transmit data / Port PSC3PSC20 Programmable Serial Control Module 1 PSC1CTS PPSC1PSC07 PSC1BCLK — PSC1 clear-to-send indication / Port PSC1PSC07 / PSC1 modem clock MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 15-5 Table 15-1. MCF548x GPIO Module External Pins (Continued) Primary Function (Pin Name)1 GPIO Alternate Function 1 Alternate Function 2 PSC1RTS PPSC1PSC06 PSC1FSYNC — PSC1 request-to-send indication / Port PSC1PSC06 / PSC1 frame sync PSC1RXD PPSC1PSC05 — — PSC1 receive data / Port PSC1PSC05 PSC1TXD PPSC1PSC04 — — PSC1 transmit data / Port PSC1PSC04 Description Programmable Serial Control Module 0 PSC0CTS PPSC1PSC03 PSC0BCLK — PSC0 clear-to-send indication / Port PSC1PSC03 / PSC0 modem clock PSC0RTS PPSC1PSC02 PSC0FSYNC — PSC0 request-to-send indication / Port PSC1PSC02 / PSC0 frame sync PSC0RXD PPSC1PSC01 — — PSC0 receive data / Port PSC1PSC01 PSC0TXD PPSC1PSC00 — — PSC0 transmit data / Port PSC1PSC00 Peripheral Component Interface PCIBG4 PPCIGNT4 TBST — PCI bus grant 4 / Port PCIGNT4 / Flexbus transfer burst PCIBG3 PPCIGNT3 TOUT3 — PCI bus grant 3 / Port PCIGNT3 / GP timer out 3 PCIBG2 PPCIGNT2 TOUT2 — PCI bus grant 2 / Port PCIGNT2 / GP timer out 2 PCIBG1 PPCIGNT1 TOUT1 — PCI bus grant 1 / Port PCIGNT1 / GP timer out 1 PCIBG0 PPCIGNT0 TOUT0 — PCI bus grant 0 / Port PCIGNT0 / GP timer out 0 PCIBR4 PPCIREQ4 IRQ4 — PCI bus request 4 / Port PCIREQ4 / Interrupt 4 PCIBR3 PPCIREQ3 TIN2 — PCI bus request 3 / Port PCIREQ3 / GP timer in 3 PCIBR2 PPCIREQ2 TIN2 — PCI bus request 2 / Port PCIREQ2 / GP timer in 2 PCIBR1 PPCIREQ1 TIN1 — PCI bus request 1 / Port PCIREQ1 / GP timer in 1 PCIBR0 PPCIREQ0 TIN0 — PCI bus request 0 / Port PCIREQ0 / GP timer in 0 General Purpose Timer 2 TIN3 PTIM3 IRQ3 CANRX1 TOUT3 PTIM72 CANTX1 — 2 TIN2 PTIM2 IRQ2 CANRX1 TOUT2 PTIM62 CANTX1 — GP timer in 3 / Port TIM7 / Interrupt 3 / FlexCAN 1 receive data GP timer out 3 / Port TIM6 / FlexCAN 1 transmit data GP timer in 2 / Port TIM5 / Interrupt 1 / FlexCAN 2 receive data GP timer out 2 / Port TIM4 / FlexCAN 1 transmit data 1 The primary functionality of a pin is not necessarily the default function of the pin after reset. Most pins that have muxed GPIO functionality will default to GPIO inputs. See the reset value of the associated pin assignment register. See Section 15.3.2.5, “Port x Pin Assignment Registers (PAR_x)”) for more information on default pin functionality. 2 GPIO is supported, but the GPIO functionality is controlled by the timer or EPORT module instead of the GPIO module. Signals are listed because there are pin assignment registers in the GPIO module for controlling the signal functions. Refer to the signals chapter of the MCF548x chip specification for more detailed descriptions of these signals. The function of most of the pins (primary function, GPIO, etc.) is determined by the GPIO module pin assignment registers. MCF548x Reference Manual, Rev. 5 15-6 Freescale Semiconductor Memory Map/Register Definition It should be noted from Table 15-1 that there are several cases where a function is mapped to more than one pin. While it is possible to enable the function on more than one pin simultaneously, this type of programming should be avoided for input functions to prevent unexpected behavior. All multiple-pin functions are listed in Table 15-2. Table 15-2. MCF548x Multiple-Pin Functions Function Direction Associated Pins GP timer in 3 (TIN3) I TIN3, PCIBR3 GP timer in 2 (TIN2) I TIN2, PCIBR2 GP timer in 1 (TIN1) I TIN1, PCIBR1, DREQ1 GP timer in 0 (TIN0) I TIN0, PCIBR0, DMA_REQ0 GP timer out 3 (TOUT3) O TOUT3, PCIBG3, DSPI_PSC3 GP timer out 2 (T2OUT) O TOUT2, PCIBG2, DSPI_PSC2 GP timer out 1 (T1OUT) O TOUT1, PCIBG1, DACK1 GP timer out 0 (T0OUT) O TOUT0, PCIBG0, DACK0 FlexCAN 0 transmit data (CANTX0) O PSC2RTS, FEC1MDC FlexCAN 0 receive data (CANRX0) I PSC2CTS, FEC1MDIO FlexCAN 1 transmit data (CANTX1) O T3OUT, T2OUT, DSPI_PCS3, DSPI_PCS2 FlexCAN 1 receive data (CANRX1) I T3IN, T2IN, IRQ6, IRQ5 2C I serial data (SDA) I/O SDA, FEC1MDC I2 C serial clock (SCL) I/O SDA, FEC1MDIO PSC3 request-to-send (PSC3RTS) O PSC3RTS, DSPIPCS0/SS PSC3 clear-to-send (PSC3CTS) I PSC3CTS, DSPISCK PSC3 modem clock (PSC3BCLK) I PSC3CTS, DSPISCK PSC3 frame sync (PSC3FSYNC) I PSC3CTS, DSPIPCS0/SS PSC3 uart receive data (PSC3RXD) I PSC3RXD, DSPISIN PSC3 uart transmit data (PSC3TXD) O PSC3TXD, DSPISOUT 15.3 15.3.1 Memory Map/Register Definition Register Overview Table 15-3 summarizes all the registers in the MCF548x GPIO module address space. Table 15-3. MCF548x GPIO Module Memory Map MBAR Offset 31–24 23–16 15–8 7–0 Access1 Port Output Data Registers 0xA00 PODR_FBCTL PODR_FBCS PODR_DMA Reserved3 S/U 0xA04 PODR_FEC0H PODR_FEC0L PODR_FEC1H PODR_FEC1L S/U MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 15-7 Table 15-3. MCF548x GPIO Module Memory Map (Continued) MBAR Offset 31–24 23–16 15–8 7–0 Access1 0xA08 PODR_FECI2C PODR_PCIBG PODR_PCIBR Reserved3 S/U 0xA0C PODR_PSC3PSC2 PODR_PSC1PSC0 PODR_DSPI Reserved3 S/U Port Data Direction Registers 0xA10 PDDR_FBCTL PDDR_FBCS PDDR_DMA Reserved2 S/U 0xA14 PDDR_FEC0H PDDR_FEC0L PDDR_FEC1H PDDR_FEC1L S/U 0xA18 PDDR_FECI2C PDDR_PCIBG PDDR_PCIBR Reserved3 S/U 3 S/U 0xA1C PDDR_PSC3PSC2 PDDR_PSC1PSC0 PDDR_DSPI Reserved Port Pin Data/Set Data Registers 0xA20 PPDSDR_FBCTL PPDSDR_FBCS PPDSDR_DMA Reserved3 S/U 0xA24 PPDSDR_FEC0H PPDSDR_FEC0L PPDSDR_FEC1H PPDSDR_FEC1L S/U S/U 0xA28 PPDSDR_FECI2C PPDSDR_PCIBG PPDSDR_PCIBR Reserved3 0xA2C PPDSDR_PSC3PSC2 PPDSDR_PSC1PSC0 PPDSDR_DSPI Reserved3 S/U Port Clear Output Data Registers 0xA30 PCLRR_FBCTL PCLRR_FBCS PCLRR_DMA Reserved3 S/U 0xA34 PCLRR_FEC0H PCLRR_FEC0L PCLRR_FEC1H PCLRR_FEC1L S/U 0xA38 PCLRR_FECI2C PCLRR_PCIBG PCLRR_PCIBR Reserved3 S/U PCLRR_DSPI Reserved3 S/U PAR_DMA S/U 0xA3C PCLRR_PSC3PSC2 PCLRR_PSC1PSC0 Pin Assignment Registers 0xA40 PAR_FBCTL 0xA44 PAR_FECI2CIRQ Reserved3 S/U 0xA48 PAR_PCIBG PAR_PCIBR S/U 0xA4C 0xA50 PAR_PSC3 PAR_FBCS PAR_PSC2 PAR_DSPI 2 PAR_PSC0 S/U PAR_TIMER 3 S/U Reserved Reserved3 0xA54– 0xA7F 1 PAR_PSC1 S/U = supervisor or user mode access. Reads to reserved locations return 0s. Writes have no effect. 15.3.2 15.3.2.1 Register Descriptions Port x Output Data Registers (PODR_x) The PODR registers store the data to be driven on the corresponding port x pins when the pins are configured for general purpose output. MCF548x Reference Manual, Rev. 5 15-8 Freescale Semiconductor Memory Map/Register Definition Most PODR_x registers have full 8-bit implementations, as shown in Figure 15-2. The remaining PODR_x registers use fewer than eight bits. These registers are shown in Figure 15-3, Figure 15-4, Figure 15-5, and Figure 15-6. The PODR_x registers are read/write. At reset, all implemented bits in the PODR_x registers are set. Unimplemented bits always remain cleared. Reading a PODR_x register returns the current values in the register, not the port x pin values. To set bits in a PODR_x register, write 1s to the PODR_x bits, or write 1s to the corresponding bits in the PORTP_x/SET_x register. To clear bits in a PODR_x register, write 0s to the PODR_x bits, or write 0s to the corresponding bits in the PCLRR_x register. 15.3.2.1.1 8-Bit PODR_x Registers The 8-bit PODR_x registers include the following: • PODR_FBCTL • PODR_FEC0H • PODR_FEC0L • PODR_FEC1H • PODR_FEC1L • PODR_PSC3PSC2 • PODR_PSC1PSC0 Figure 15-2 displays the 8-bit PODR_x registers. R 7 6 5 4 3 2 1 0 PODRx7 PODRx6 PODRx5 PODRx4 PODRx3 PODRx2 PODRx1 PODRx0 1 1 1 1 1 1 1 1 W Reset Reg MBAR + 0xA00 (PODR_FBCTL), 0xA04 (PODR_FEC0H), 0xA05 (PODR_FEC0L), 0xA06 (PODR_FEC1H), Addr 0xA07 (PODR_FEC1L), 0xA0C (PODR_PSC3PSC2), 0xA0D (PODR_PSC1PSC0) Figure 15-2. 8-Bit Port Output Data Registers (PODR_x) Table 15-4. 8-Bit PODR_x Field Descriptions Bits Name 7–0 PODRxn 15.3.2.1.2 Description PODRx Output Data Bits 0 Drive 0 when PORT x pin is general purpose output 1 Drive 1 when PORT x pin is general purpose output 7-Bit PODR_x Register The 7-bit PODR_DSPI register is the output data register for the PDSPIn port. Figure 15-3 displays the 7-bit PODR_x register. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 15-9 7 R 0 6 5 4 3 2 1 0 PODRDSPI6 PODRDSPI5 PODRDSPI4 PODRDSPI3 PODRDSPI2 PODRDSPI1 PODRDSPI0 W Reset 0 1 1 Reg Addr 1 1 1 1 1 MBAR + 0xA0E (PODR_DSPI) Figure 15-3. 7-Bit PODR_DSPI Register (PODR_x) Table 15-5. 7-Bit PODR_DSPI Field Descriptions Bits Name 7 — 6–0 PODRDSPIn 15.3.2.1.3 Description Reserved, should be cleared PODR DSPI output data bits 0 Drive 0 when PDSPIn pin is general purpose output 1 Drive 1 when PDSPIn pin is general purpose output 5-Bit PODR_x Registers The 5-bit PODR_x registers are the output data registers for PPCIBGn (PODR_PCIBG) and PPCIBRn (PODR_PCIBR). Figure 15-4 displays the 5-bit PODR_x registers. R 7 6 5 4 3 2 1 0 0 0 0 PODRx4 PODRx3 PODRx2 PODRx1 PODRx0 0 0 0 1 1 1 1 1 W Reset Reg Addr MBAR + 0xA09 (PODR_PCIBG), 0xA0A (PODR_PCIBR) Figure 15-4. 5-Bit PODR_PCIBG and PODR_PCIBR Registers Table 15-6. 5-Bit PODR_PCIBG and PODR_PCIBR Field Descriptions Bits Name 7–5 — 4–0 PODRxn 15.3.2.1.4 Description Reserved, should be cleared PODR_PCIBG and PODR_PCIBR output data bits 0 Drive 0 when PPCIBGn or PPCIBRn pin is general purpose output 1 Drive 1 when PPCIBGn or PPCIBRn pin is general purpose output 4-Bit PODR_x Registers The 4-bit PODR_x registers are the output data registers for PDMAn (PODR_DMA) and PFECI2Cn (PODR_FECI2C). Figure 15-3 displays the 4-bit PODR_x registers. MCF548x Reference Manual, Rev. 5 15-10 Freescale Semiconductor Memory Map/Register Definition R 7 6 5 4 3 2 1 0 0 0 0 0 PODRx3 PODRx2 PODRx1 PODRx0 0 0 0 0 1 1 1 1 W Reset Reg Addr MBAR + 0xA02 (PORT_DMA), 0xA08 (PORT_FECI2C) Figure 15-5. 4-Bit PODR_DMA and PODR_FECI2C Registers Table 15-7. 4-Bit PODR_DMA and PODR_FECI2C Field Descriptions Bits Name 7–4 — 3–0 PODRxn 15.3.2.1.5 Description Reserved, should be cleared PORT_DMA and PORT_FECI2C output data bits 0 Drive 0 when PDMAn or PFECI2Cn pin is general purpose output 1 Drive 1 when PDMAn or PFECI2Cn pin is general purpose output FBCS Register (PODR_FBCS) The 5-bit PODR_FBCS register is the output data register for PFBCSn (PODR_FBCS). Figure 15-6 displays the 5-bit PODR_FBCS register. R 7 6 5 4 3 2 1 0 0 0 PODRFB5 PODRFB4 PODRFB3 PODRFB2 PODRFB1 0 0 0 0 0 0 0 0 0 W Reset Reg Addr MBAR + 0xA01 (PODR_FBCS) Figure 15-6. 5-Bit PODR_FBCS Register Table 15-8. 5-Bit PODR_FBCS Field Descriptions Bits Name 7–6 — 5–1 0 15.3.2.2 Description Reserved, should be cleared PODRFBn PORT_FBCS output data 0 Drive 0 when PFBCSn pin is general purpose output 1 Drive 1 when PFBCSn pin is general purpose output — Reserved, should be cleared Port x Data Direction Registers (PDDR_x) The PDDR registers control the direction of the port x pin drivers when the pins are configured for general purpose I/O. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 15-11 Most PDDR_x registers have a full 8-bit implementation, as shown in Figure 15-7. The remaining PDDR_x registers use fewer than eight bits. Their bit definitions are shown in Figure 15-8, Figure 15-9, Figure 15-10, and Figure 15-11. The PDDR_x registers are read/write. At reset, all bits in the PDDR_x registers are cleared. Setting any bit in a PDDR_x register configures the corresponding port x pin as an output. Clearing any bit in a PDDR_x register configures the corresponding pin as an input. 15.3.2.2.1 8-Bit PDDR_x Registers The 8-bit PDDR_x registers include the following: • PDDR_FBCTL • PDDR_FEC0H • PDDR_FEC0L • PDDR_FEC1H • PDDR_FEC1L • PDDR_PSC3PSC2 • PDDR_PSC1PSC0 Figure 15-7 displays the 8-bit PDDR_x registers. R 7 6 5 4 3 2 1 0 DDx7 DDx6 DDx5 DDx4 DDx3 DDx2 DDx1 DDx0 0 0 0 0 0 0 0 0 W Reset Reg MBAR + 0xA10 (PDDR_FBCTL), 0xA14 (PDDR_FEC0H), 0xA15 (PDDR_FEC0L), 0xA16 (PDDR_FEC1H), 0xA17 Addr (PDDR_FEC1L), 0xA1C (PDDR_PSC3PSC2), 0xA1D (PDDR_PSC1PSC0) Figure 15-7. 8-Bit Port Data Direction Registers Table 15-9. 8-Bit PDDR_x Field Descriptions Bits Name 7–0 DDxn 15.3.2.2.2 Description PDDR_x Data Direction Bits 0 PORT x pin is configured as input 1 PORT x pin is configured as output 7-Bit PDDR_x Register The 7-bit PDDR_DSPI register sets the data direction for the PDSPIn port. Figure 15-8 displays the 7-bit PDDR_DSPI register. MCF548x Reference Manual, Rev. 5 15-12 Freescale Semiconductor Memory Map/Register Definition R 7 6 5 4 3 2 1 0 0 DDDSP 6 DDDSP5 DDDSP4 DDDSPI3 DDDSPI2 DDDSP1 DDDSP0 0 0 0 0 0 0 0 0 W Reset Reg Addr MBAR + 0xA1E (PDDR_DSPI) Figure 15-8. 7-Bit PDDR_DSPI Data Direction Register Table 15-10. 7-Bit PDDR_DSPI Field Descriptions Bits Name 7 — 6–0 DDDSPn 15.3.2.2.3 Description Reserved, should be cleared PODR_DSPI data direction 0 PDSPIn pin configured as input 1 PDSPIn pin configured as output 5-Bit PDDR_x Registers The 5-bit PDDR_x registers are the data direction registers for PPCIBGn (PDDR_PCIBG) and PPCIBRn (PDDR_PCIBR). Figure 15-9 displays the 5-bit PDDR_x registers. R 7 6 5 4 3 2 1 0 0 0 0 DDx4 DDx3 DDx2 DDx1 DDx0 0 0 0 0 0 0 0 0 W Reset Reg Addr MBAR + 0xA11 (PDDR_FBCS), 0xA19 (PDDR_PCIBG), 0xA1A (PDDR_PCIBR) Figure 15-9. 5-Bit PDDR_PCIBG and PDDR_PCIBR Registers Table 15-11. 5-Bit PDDR_PCIBG and PDDR_PCIBR Field Descriptions Bits Name 7–5 — 4–0 DDxn 15.3.2.2.4 Description Reserved, should be cleared PDDR_PCIBG and PDDR_PCIBR Data Direction 0 PPCIBGn or PPCIBRn pin is configured as input 1 PPCIBGn or PPCIBRn pin is configured as output 4-Bit PDDR_x Registers The 4-bit PDDR_x registers are for data direction of PDMAn (PDDR_DMA) and (PDDR_FECI2C). Figure 15-10 displays the 4-bit PDDR_x registers. PFECI2Cn MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 15-13 R 7 6 5 4 3 2 1 0 0 0 0 0 DDx3 DDx2 DDx1 DDx0 0 0 0 0 0 0 0 0 W Reset Reg Addr MBAR + 0xA12 (PDDR_DMA), 0xA18 (PDDR_FECI2C) Figure 15-10. 4-Bit PDDR_DMA and PDDR_FECI2C Registers Table 15-12. 4-Bit PDDR_DMA and PDDR_FECI2C Field Descriptions Bits Name 7–4 — 3–0 DDxn 15.3.2.2.5 Description Reserved, should be cleared PDDR_DMA and PDDR_FECI2C Data Direction 0 PDMAn or PFECI2Cn pin is configured as input 1 PDMAn or PFECI2Cn pin is configured as output FBCS Register (PDDR_FBCS) The 5-bit PDDR_FBCS register is for data direction of PFBCSn. Figure 15-11 displays the 5-bit PDDR_FBCS register. R 7 6 5 4 3 2 1 0 0 0 DDFB5 DDFB4 DDFB3 DDFB2 DDFB1 0 0 0 0 0 0 0 0 0 W Reset Reg Addr MBAR + 0xA11 (PDDR_FBCS) Figure 15-11. 5-Bit PDDR_FBCS Register Table 15-13. 5-Bit PDDR_FBCS Field Descriptions Bits Name 7–6 — 5–1 DDFBn 0 — 15.3.2.3 Description Reserved, should be cleared PDDR_FBCS data direction 0 PFBCSn pin is configured as input 1 PFBCSn pin is configured as output Reserved, should be cleared Port x Pin Data/Set Data Registers (PPDSDR_x) The PPDSDR registers reflect the current pin states and control the setting of output pins when the pin is configured for general purpose I/O. MCF548x Reference Manual, Rev. 5 15-14 Freescale Semiconductor Memory Map/Register Definition Most PPDSDR_x registers have a full 8-bit implementation, as shown in Figure 15-12. The remaining PPDSDR_x registers use fewer than eight bits. Their bit definitions are shown in Figure 15-13, Figure 15-14, Figure 15-15, and Figure 15-16. The PPDSDR_x registers are read/write. At reset, the bits in the PPDSDR_x registers are set to the current pin states. Reading a PPDSDR_x register returns the current state of the port x pins. Writing 1s to a PPDSDR_x register sets the corresponding bits in the PODR_x register. Writing 0s has no effect. 15.3.2.3.1 8-Bit PPDSDR_x Registers The 8-bit PPDSDR_x registers include the following: • PPDSDR_FBCTL • PPDSDR_FEC0H • PPDSDR_FEC0L • PPDSDR_FEC1H • PPDSDR_FEC1L • PPDSDR_PSC3PSC2 • PPDSDR_PSC1PSC0 • PPDSDR_PSC3PSC2 Figure 15-12 displays the 8-bit PPDSDR_x registers. 7 6 5 4 3 2 1 0 R PPDx7 PPDx6 PPDx5 PPDx4 PPDx3 PPDx2 PPDx1 PPDx0 W PSDx7 PSDx6 PSDx5 PSDx4 PSDx3 PSDx2 PSDx1 PSDx0 P1 P1 P1 P1 P1 P1 P1 P1 Reset Reg Addr MBAR + 0xA20 (PPDSDR_FBCTL), 0xA24 (PPDSDR_FEC0H), 0xA25 (PPDSDR_FEC0L), 0xA26 (PPDSDR_FEC1H), 0xA27 (PPDSDR_FEC1L), 0xA2C (PPDSDR_PSC3PSC2), 0xA2D (PPDSDR_PSC1PSC0) 1 P = the current pin state. The exception is that PPDSDR_FBCTL is always reset to 0. Figure 15-12. 8-Bit Port Pin Data / Set Data Registers Table 15-14. 8-Bit PPDSDR_x Field Descriptions Bits Name 7–0 PPDxn Port pin data. This is read-only. 0 Port x pin state is low 1 Port x pin state is high PSDxn Port set data. 0 No effect 1 Corresponding PODR_x bit is set 15.3.2.3.2 Description 7-Bit PPDSDR_x Register The 7-bit PPDSDR_x register is for pin data and set data for PDSPIn. Figure 15-13 displays the 7-bit PPDSDR_DSPI register. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 15-15 R 7 6 5 4 3 2 1 0 0 PPDx6 PPDx5 PPDx4 PPDx3 PPDx2 PPDx1 PPDx0 PSDx6 PSDx5 PSDx4 PSDx3 PSDx2 PSDx1 PSDx0 W Reset 0 P 1 1 1 P Reg Addr 1 P P 1 1 P P P1 MBAR + 0xA2E (PPDSDR_DSPI) 1 P = the current pin state. Figure 15-13. 7-Bit Port Pin Data / Set Data Registers Table 15-15. 7-Bit PPDSDR_DSPI Field Descriptions Bits Name 7 — 6–0 PPDxn PPDSDR_DSPI pin data. This is Read-only. 0 PDSPIn pin state is low 1 PDSPIn pin state is high PSDxn PPDSDR_DSPI set data. 0 No effect 1 Corresponding PODR_DSPI bit is set 15.3.2.3.3 Description Reserved, should be cleared. 5-Bit PPDSDR_x Registers The 5-bit PPDSDR_x registers are the pin data and set data registers for PPCIBGn (PPDSDR_PCIBG) and PPCIBRn (PPDSDR_PCIBR). Figure 15-14 displays the 5-bit PPDSDR_x registers. R 7 6 5 4 3 2 1 0 0 0 0 PPDx4 PPDx3 PPDx2 PPDx1 PPDx0 PSDx4 PSDx3 PSDx2 PSDx1 PSDx0 P1 P1 P1 P1 P1 W Reset 0 0 Reg Addr 0 MBAR + 0xA29 (PPDSDR_PCIBG) and 0xA2A (PPDSDR_PCIBR) 1 P = the current pin state. Figure 15-14. 5-Bit PPDSDR_PCIBG and PPDSDR_PCIBR Registers Table 15-16. 5-Bit PPDSDR_PCIBG and PPDSDR_PCIBR Field Descriptions Bits Name 7–5 — Description Reserved, should be cleared. MCF548x Reference Manual, Rev. 5 15-16 Freescale Semiconductor Memory Map/Register Definition Table 15-16. 5-Bit PPDSDR_PCIBG and PPDSDR_PCIBR Field Descriptions (Continued) Bits Name 4–0 PPDxn PPDSDR_PCIBG and PPDSDR_PCIBR pin data. This is Read-only. 0 PPCIBGn or PPCIBRn pin state is low 1 PPCIBGn or PPCIBRn pin state is high PSDxn PPDSDR_PCIBG and PPDSDR_PCIBR set data. 0 No effect 1 Corresponding PODR_PCIBGn or PODR_PCIBRn bit is set 15.3.2.3.4 Description 4-Bit PPDSDR_x Registers The 4-bit PPDSDR_x registers are the pin data and set data registers for PDMAn (PPDSDR_DMA) and PFECI2Cn (PPDSDR_FECI2C). Figure 15-15 displays the 4-bit PPDSDR_DMA and PPDSDR_FECI2C registers. R 7 6 5 4 3 2 1 0 0 0 0 0 PPDx3 PPDx2 PPDx1 PPDx0 PSDx3 PSDx2 PSDx1 PSDx0 P1 P1 P1 P1 W Reset 0 0 Reg Addr 0 0 MBAR + 0xA22 (PPDSDR_DMA) and 0xA28 (PPDSDR_FECI2C) 1 P = the current pin state. Figure 15-15. 4-Bit PPDSDR_DMA and PPDSDR_FECI2C Registers Table 15-17. 4-Bit PPDSDR_DMA and PPDSDR_FECI2C Field Descriptions Bits Name 7–4 — 4–0 PPDxn PPDSDR_DMA and PPDSDR_FECI2C pin data. This is Read-only. 0 PDMAn or PFECI2Cn pin state is low 1 PDMAn or PFECI2Cn pin state is high PSDXn PPDSDR_DMA and PPDSDR_FECI2C set data. 0 No effect 1 Corresponding PODR_DMA or PODR_FECI2C bit is set 15.3.2.3.5 Description Reserved, should be cleared. FBCS Register (PPDSDR_FBCS) The 5-bit PPDSDR_FBCS register is for pin data and set data for PFBCSn. Figure 15-16 displays the 5-bit PPDSDR_FBCS register. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 15-17 R 7 6 5 4 3 2 1 0 0 0 PPDx5 PPDx4 PPDx3 PPDx2 PPDx1 0 PSDx5 PSDx4 PSDx3 PSDx2 PSDx1 W Reset 0 1 0 P Reg Addr 1 P 1 P 1 P P1 0 MBAR + 0xA21 (PDDSDR_FBCS) 1 P = the current pin state. Figure 15-16. 5-Bit PDDSDR_FBCS Register Table 15-18. 5-Bit PDDSDR_FBCS Field Descriptions Bits Name 7–6 — 5–1 PPDxn PDDSDR_FBCS pin data. This is Read-only. 0 PFBCSn pin state is low 1 PFBCSn pin state is high PSDxn PDDSDR_FBCS set data. 0 No effect 1 Corresponding PODR_FBCS bit is set 0 15.3.2.4 — Description Reserved, should be cleared. Reserved, should be cleared. Port x Clear Output Data Registers (PCLRR_x) Writing 0s to a PCLRR_x register clears the corresponding bits in the PODR_x register. Writing 1s has no effect. Reading the PCLRR_x register returns 0s. Most PCLRR_x registers have a full 8-bit implementation, as shown in Figure 15-17. The remaining PCLRR_x registers use fewer than eight bits. Their bit definitions are shown in Figure 15-18, Figure 15-19, Figure 15-20, and Figure 15-21. The PCLRR_x registers are read/write. The 8-bit PCLRR_x registers include the following: • PCLRR_FBCTL • PCLRR_FEC0H • PCLRR_FEC0L • PCLRR_FEC1H • PCLRR_FEC1L • PCLRR_PSC3PSC2 • PCLRR_PSC1PSC0 Figure 15-17 displays the 8-bit PCLRR_x registers. MCF548x Reference Manual, Rev. 5 15-18 Freescale Semiconductor Memory Map/Register Definition 7 6 5 4 3 2 1 0 R 0 0 0 0 0 0 0 0 W CLRx7 CLRx6 CLRx5 CLRx4 CLRx3 CLRx2 CLRx1 CLRx0 0 0 0 0 0 0 0 0 Reset Reg MBAR + 0xA30 (PCLRR_FBCTL), 0xA34 (PCLRR_FEC0H), 0xA35 (PCLRR_FEC0L), 0xA36 (PCLRR_FEC1H), Addr 0xA37 (PCLRR_FEC1L), 0xA3C (PCLRR_PSC3PSC2), 0xA3D (PCLRR_PSC1PSC0) Figure 15-17. 8-Bit Port Clear Output Data Registers Table 15-19. 8-Bit PCLRR_x Field Descriptions Bits Name 7–0 CLRxn 15.3.2.4.1 Description Clear output data registers 0 Corresponding PODR_x bit is cleared 1 No effect 7-Bit PCLRR_x Register The 7-bit PCLRR_DSPI register is the clear output data register for PDSPIn. Figure 15-18 displays the 7-bit PCLRR_DSPI register. R 7 6 5 4 3 2 1 0 0 0 0 0 0 0 0 0 CLRDSP6 CLRDSP5 CLRDSP4 CLRDSP3 CLRDSP2 CLRDSP1 CLRDSP0 0 0 0 0 0 0 0 W Reset 0 Reg Addr MBAR + 0xA3E (PCLRR_DSPI) Figure 15-18. 7-Bit Port Clear Output Data DSPI Register Table 15-20. 7-Bit PCLRR_DSPI Field Descriptions Bits Name 7 — 6–0 CLRDSPn 15.3.2.4.2 Description Reserved, should be cleared PCLRR_DSPI Clear output data register 0 Corresponding PODR_DSPI bit is cleared 1 No effect 5-Bit PCLRR_x Registers The 5-bit PCLRR_x registers are the pin data and set data registers for PPCIBGn (PCLRR_PCIBG) and PPCIBRn (PCLRR_PCIBR). Figure 15-19 displays the 5-bit PCLRR_x registers. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 15-19 R 7 6 5 4 3 2 1 0 0 0 0 0 0 0 0 0 PCLRRx4 PCLRRx3 PCLRRx2 PCLRRx1 PCLRRx0 0 0 0 0 0 W Reset 0 0 Reg Addr 0 MBAR + 0xA39 (PCLRR_PCIBG) and 0xA3A (PCLRR_PCIBR) Figure 15-19. 5-Bit PCIBG and PCIBR Clear Output Data Register Table 15-21. 5-Bit PCLRR_PCIBG and PCLRR_PCIBR Field Descriptions Bits Name 7–5 — 4–0 PCLRRxn 15.3.2.4.3 Description Reserved, should be cleared PCLRR_PCIBG and PCLRR_PCIBR clear output data registers 0 Corresponding PODR_PCIGNT or PODR_PCIBR bit is cleared 1 No effect 4-Bit PCLRR_x Registers The 4-bit PCLRR_x registers are the clear output data registers for PDMAn (PCLRR_DMA) and PFECI2Cn (PCLRR_FECI2C). Figure 15-20 displays the 4-bit PCLRR_x registers. R 7 6 5 4 3 2 1 0 0 0 0 0 0 0 0 0 PCLRRx3 PCLRRx2 PCLRRx1 PCLRRx0 0 0 0 0 W Reset 0 0 Reg Addr 0 0 MBAR + 0xA32 (PCLRR_DMA) and 0xA38 (PCLRR_FECI2C) Figure 15-20. 4-Bit DMA and FECI2C Clear Output Data Registers Table 15-22. 4-Bit PCLRR_DMA and PCLRR_FECI2C Field Descriptions Bits Name 7–4 — 3–0 PCLRRxn 15.3.2.4.4 Description Reserved, should be cleared PCLRR_DMA and PCLRR_FECI2C clear output data registers 0 Corresponding PODR_DMA or PODR_FECI2C bit is cleared 1 No effect 5-Bit PCLRR_FBCS Registers The 5-bit PCLRR_FBCS register is the clear output data register for PFBCSn. Figure 15-21 displays the 5-bit PCLRR_FBCS register. MCF548x Reference Manual, Rev. 5 15-20 Freescale Semiconductor Memory Map/Register Definition R 7 6 5 4 3 2 1 0 0 0 0 0 0 0 0 0 CLRFB5 CLRFB4 CLRFB3 CLRFB2 CLRFB1 0 0 0 0 0 W Reset 0 0 Reg Addr 0 MBAR + 0xA31 (PCLRR_FBCS) Figure 15-21. 5-Bit FlexBus Clear Output Data Register Table 15-23. 5-Bit PCLRR_FBCS Field Descriptions Bits Name 7–6 — 5–1 CLRFBn 0 — 15.3.2.5 Description Reserved, should be cleared PCLRR_FBCS clear output data register 0 Corresponding PODR_FBCS bit is cleared 1 No effect Reserved, should be cleared Port x Pin Assignment Registers (PAR_x) The PAR_x registers select the signal function that will be driven on the physical pin. 15.3.2.5.1 FlexBus Control Pin Assignment Register (PAR_FBCTL) The FlexBus control pin assignment (PAR_FBCTL) register controls the function of the FlexBus control signal pins. The PAR_FBCTL register is read/write. R 15 14 13 12 11 10 9 8 7 6 0 PAR_ BWE3 0 PAR_ BWE2 0 PAR_ BWE1 0 PAR_ BWE0 0 PAR_ OE 0 1 0 1 0 1 0 1 0 1 W Reset Reg Addr 5 4 PAR_RWB 1 1 3 2 0 PAR_ TA 0 1 1 0 PAR_ALE 1 1 MBAR + 0xA40 (PAR_FBCTL) Figure 15-22. FlexBus Control Pin Assignment Register (PAR_FBCTL) Table 15-24. PAR_FBCTL Field Descriptions Bits Name 15 — 14 13 Description Reserved, should be cleared. PAR_BWE3 The PAR_BWE bit configures the BE3/BWE3 pin for its primary function or general purpose I/O. 0 BE3/BWE3 pin configured for general purpose I/O (PFBCTL7) 1 BE3/BWE3 pin configured for FlexBus BE3/BWE3 or TSIZ1 function. The function chosen depends on the reset configuration. — Reserved, should be cleared. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 15-21 Table 15-24. PAR_FBCTL Field Descriptions (Continued) Bits Name 12 Description PAR_BWE2 The PAR_BWE bit configures the BE2/BWE2 pin for its primary function or general purpose I/O. 0 BE2/BWE2 pin configured for general purpose I/O (PFBCTL6) 1 BE2/BWE2 pin configured for FlexBus BE2/BWE2 or TSIZ0 function. The function chosen depends on the reset configuration. 11 — Reserved, should be cleared. Writes have no effect and terminate without transfer error exception PAR_BWE1 The PAR_BWE bit configures the BE1/BWE1 pin for its primary function or general purpose I/O. 0 BE1/BWE1 pin configured for general purpose I/O (PFBCTL5) 1 BE1/BWE1 pin configured for FlexBus BE1/BWE1 or FBADDR1 function. The function chosen depends on the reset configuration. 10 9 — Reserved, should be cleared. PAR_BWE0 The PAR_BWE bit configures the BE0/BWE0 pin for its primary function or general purpose I/O. 0 BE0/BWE0 pin configured for general purpose I/O (PFBCTL4) 1 BE0/BWE0 pin configured for FlexBus BE0/BWE0 or FBADDR0 function. The function chosen depends on the reset configuration. 8 7 — 6 PAR_OE 5–4 Reserved, should be cleared. The PAR_OE bit configures the OE pin for its primary function or general purpose I/O. 0 OE pin configured for general purpose I/O (PFBCTL3) 1 OE pin configured for Flexbus OE function. PAR_RWB The PAR_RWB bit configures the R/W pin for its primary function or general purpose I/O 0x R/W pin configured for general purpose I/O (PFBCTL2) 10R/W pin configured for Flexbus TBST function 11R/W pin configured for Flexbus R/W function 3 — 2 PAR_TA 1–0 PAR_ALE 15.3.2.6 Reserved, should be cleared. The PAR_TA bit configures the TA pin for its primary function or general purpose I/O 0 TA pin configured for general purpose I/O (PFBCTL1) 1 TA pin configured for Flexbus TA function The PAR_ALE bit configures the ALE pin for one of its primary functions or general purpose I/O. 0X ALE pin configured for general purpose I/O (PFBCTL0) 10 ALE pin configured for Flexbus TBST function 11 ALE pin configured for Flexbus ALE function FlexBus Chip Select Pin Assignment Register (PAR_FBCS) The PAR_FBCS register controls the function of the FlexBus chip select signal pins. The PAR_FBCS register is read/write. R 7 6 5 4 3 2 1 0 0 0 PAR_CS5 PAR_CS4 PAR_CS3 PAR_CS2 PAR_CS1 0 0 0 1 1 1 1 1 0 W Reset Reg Addr MBAR + 0xA42 (PAR_FBCS) Figure 15-23. Flexbus Chip Select Pin Assignment Register (PAR_FBCS) MCF548x Reference Manual, Rev. 5 15-22 Freescale Semiconductor Memory Map/Register Definition Table 15-25. PAR_FBCS Field Descriptions Bits Name 7–6 — 5–1 PAR_CSn 0 — 15.3.2.7 Description Reserved, should be cleared. The PAR_CSn bit configures the FBCSn pin for its primary function or general purpose I/O. 0 FBCSn pin configured for general purpose I/O (PFBCS[5:1]) 1 FBCSn pin configured for FlexBus FBCSn function Reserved, should be cleared. DMA Pin Assignment Register (PAR_DMA) The PAR_DMA register controls the function of the four MCF548x DMA pins. The PAR_DMA register is read/write 7 R 6 PAR_DACK1 5 4 PAR_DACK0 3 2 PAR_DREQ1 1 0 PAR_DREQ0 W Reset 0 Reg Addr 0 0 0 0 0 0 0 MBAR + 0xA43 (PAR_DMA) Figure 15-24. DMA Pin Assignment Register (PAR_DMA) Table 15-26. PAR_DMA Field Descriptions Bits Name Description 7–6 PAR_DACK1 The PAR_DACK1 field configures the DACK1 pin for its primary functions or general purpose I/O. 0X DACK1 pin configured for general purpose I/O (PDMA3) 10 DACK1 pin configured for GP Timer TOUT1 function 11 DACK1 pin configured for DACK1 function 5–4 PAR_DACK0 The PAR_DACK0 field configures the DACK0 pin for its primary functions or general purpose I/O. 0X DACK0 pin configured for general purpose I/O (PDMA2) 10 DACK0 pin configured for GP Timer TOUT0 function 11 DACK0 pin configured for DACK0 function 3–2 PAR_DREQ1 The PAR_DREQ1 field configures the DREQ1 pin for its primary functions or general purpose I/O. 00 = DREQ1 pin configured for general purpose I/O (PDMA1) 01 = DREQ1 pin configured for IRQ1 function 10 = DREQ1 pin configured for GP Timer TIN1 function 11 = DREQ1 pin configured for DREQ1 function 1–0 PAR_DREQ0 The PAR_DREQ0 field configures the DREQ0 pin for its primary functions or general purpose I/O. 0X = DREQ0 pin configured for general purpose I/O (PDMA0) 10 = DREQ0 pin configured for GP Timer TIN0 function 11 = DREQ0 pin configured for DREQ0 function 15.3.2.8 FEC/I2C/IRQ Pin Assignment Register (PAR_FECI2CIRQ) The PAR_FECI2CIRQ register controls the functions of the FEC0, FEC1, I2C, and IRQ pins. The PAR_FECI2CIRQ register is read/write MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 15-23 15 14 13 12 11 10 9 8 7 6 R PAR_ PAR_ PAR_ PAR_ PAR_ PAR_ PAR_E1MDIO PAR_E1MDC E07 E0MII E0MDIO E0MDC E17 E1MII W Reset 0 0 0 Reg Addr 0 0 0 1 1 1 1 5 4 0 0 0 0 3 2 1 PAR_ PAR_ PAR_ PAR_ SDA SCL IRQ6 IRQ5 0 0 1 MBAR + 0xA44 (PAR_FECI2CIRQ) Figure 15-25. FEC/I2C/IRQ Pin Assignment Register (PAR_FECI2CIRQ) Table 15-27. PAR_FEC/I2C/IRQ Field Descriptions Bits Name Description 15 PAR_E07 FEC0 7-wire mode pin assignment. Configures all the FEC0 7-wire mode pins (port FEC0H pins, except for E0CRS) for their primary functions or general purpose I/O. 0 All FEC1 7-wire mode pins configured for GPIO (PFEC0H[7:1]) 1 All FEC1 7-wire mode pins configured for their primary functions 14 PAR_E0MII FEC1 MII mode-only pin assignment. Configures all the FEC0 MII mode-only pins (port FEC0L pins, plus FEC0_CRS) for their primary functions or general purpose I/O. 0 All FEC0 MII mode-only pins configured for GPIO (PFEC0H0 and PFEC0L[7:0] 1 All FEC0 MII mode-only pins configured for their primary functions 13 PAR_E0MDIO FEC0 MDIO pin assignment. Configures the E0MDIO pin for its primary function or general purpose I/O. 0 E0MDIO pin configured for GPIO (PFECI2C3) 1 E0MDIO pin configured for E0MDIO function 12 PAR_E0MDC FEC0 MDC pin assignment. Configures the E0MDC pin for its primary function or general purpose I/O. 0 E0MDC pin configured for GPIO (PFECI2C2) 1 E0MDC pin configured for E0MDC function 11 PAR_E17 FEC1 7-wire mode pin assignment. Configures all the FEC1 7-wire mode pins (port FEC1H pins, except for E1CRS) for their primary functions or general purpose I/O. 0 All FEC1 7-wire mode pins configured for GPIO (PFEC1H[7:1]) 1 All FEC1 7-wire mode pins configured for their primary functions 10 PAR_E1MII FEC1 MII mode-only pin assignment. Configures all the FEC1 MII mode-only pins (port FEC1L pins, plus E1CRS) for their primary functions or general purpose I/O. 0 All FEC1 MII mode-only pins configured for GPIO (PFEC1H0 and PFEC1L[7:0]) 1 All FEC1 MII mode-only pins configured for their primary functions 9–8 PAR_ E1MDIO FEC1 MDIO pin assignment. Configures the E1MDIO pin for one of its primary functions. There is no GPIO capability on this pin. 0X E1MDIO pin configured for FlexCAN CANRX0 10 E1MDIO pin configured for I2C SDA function 11 E1MDIO pin configured for FEC1 E1MDIO function 7–6 PAR_ E1MDC FEC1 MDC pin assignment. Configures the E1MDC pin for one of its primary functions. There is no GPIO capability on this pin. 0X E1MDC pin configured for FlexCAN CANTX0 10 E1MDC pin configured for I2C SCL function 11 E1MDC pin configured for FEC1 E1MDC function 5–4 — Reserved, should be cleared. MCF548x Reference Manual, Rev. 5 15-24 0 Freescale Semiconductor 1 Memory Map/Register Definition Table 15-27. PAR_FEC/I2C/IRQ Field Descriptions (Continued) Bits Name 3 PAR_SDA SDA Pin Assignment. Configures the SDA pin for its primary function or general purpose I/O. 0 SDA pin configured for general purpose input (PFECI2C1) 1 SDA pin configured for SDA function 2 PAR_SCL SCL Pin Assignment. Configures the SCL pin for its primary function or general purpose I/O. 0 SCL pin configured for GPIO (PFECI2C0) 1 SCL pin configured for SCL function 1 PAR_ IRQ6 IRQ6 Pin Assignment. Configures the IRQ6 pin for one of its primary functions. 0 IRQ6 pin configured for FlexCAN CANRX1 1 IRQ6 pin configured for IRQ6 function Note that GPIO is obtained on the IRQ6 pin by (1) writing a 1 to PAR_IRQ6 and (2) disabling the IRQ6 function in the EPORT module. 0 PAR_ IRQ5 IRQ5 Pin Assignment. Configures the IRQ5 pin for one of its primary functions. 0 IRQ5 pin configured for FlexCAN CANRX1 1 IRQ5 pin configured for IRQ5 function Note that GPIO is obtained on the IRQ5 pin by (1) writing a 1 to PAR_IRQ5 and (2) disabling the IRQ5 function in the EPORT module. 15.3.2.9 Description PCI Grant Pin Assignment Register (PAR_PCIBG) The PAR_PCIBG register controls the functions of the PCI grant pins. The PAR_PCIBG register is read/write. R 15 14 13 12 11 10 9 0 0 0 0 0 0 PAR_ PCIBG4 PAR_ PCIBG3 PAR_ PCIBG2 PAR_ PCIBG1 PAR_ PCIBG0 0 0 0 0 0 0 0 0 0 0 0 W Reset Reg Addr 8 0 7 6 0 5 4 0 3 2 0 1 0 0 MBAR + 0xA48 (PAR_PCIBG) Figure 15-26. PCI Grant Pin Assignment Register (PAR_PCIBG) Table 15-28. PAR_PCIBG Field Descriptions Bits Name Description 15–10 — 9–8 PAR_ PCIBG4 PCIBG4 pin assignment. Configures the PCIBG4 pin for one of its primary functions or GPIO. 0X PCIBG4 pin configured for general purpose I/O (PPCIGNT4) 10 PCIBG4 pin configured for FlexBus TBST function 11 PCIBG4 pin configured for PCIBG4 function 7–6 PAR_ PCIBG3 PCIBG3 pin assignment. Configures the PCIBG3 pin for one of its primary functions or GPIO. 0X PCIBG3 pin configured for general purpose I/O (PPCIGNT3) 10 PCIBG3 pin configured for GP timer TOUT3 function 11 PCIBG3 pin configured for PCIBG3 function Reserved, should be cleared. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 15-25 Table 15-28. PAR_PCIBG Field Descriptions (Continued) Bits Name Description 5–4 PAR_ PCIBG2 PCIBG2 pin assignment. Configures the PCIBG2 pin for one of its primary functions or GPIO. 0X PCIBG2 pin configured for general purpose I/O (PPCIGNT2) 10 PCIBG2 pin configured for GP timer TOUT2 function 11 PCIBG2 pin configured for PCIBG2 function 3–2 PAR_ PCIBG1 PCIBG1 pin assignment. Configures the PCIBG1 pin for one of its primary functions or GPIO. 0X PCIBG1 pin configured for general purpose I/O (PPCIGNT1) 10 PCIBG1 pin configured for GP timer TOUT1 function 11 PCIBG1 pin configured for PCIBG1 function 1–0 PAR_ PCIBG0 PCIBG0 pin assignment. Configures the PCIBG0 pin for one of its primary functions or GPIO. 0X PCIBG0 pin configured for general purpose I/O (PPCIGNT0) 10 PCIBG0 pin configured for GP timer TOUT0 function 11 PCIBG0 pin configured for PCIBG0 function 15.3.2.10 PCI Request Pin Assignment Register (PAR_PCIBR) The PAR_PCIBR controls the functions of the PCI request pins. The PAR_PCIBR is read/write. R 15 14 13 12 11 10 0 0 0 0 0 0 0 0 0 0 0 0 9 8 7 6 5 4 3 2 1 0 PAR_PCIBR4 PAR_PCIBR3 PAR_PCIBR2 PAR_PCIBR1 PAR_PCIBR0 W Reset Reg Addr 0 0 0 0 0 0 0 0 0 0 MBAR + 0xA4A (PAR_PCIBR) Figure 15-27. PCI Request Pin Assignment Register (PAR_PCIBR) Table 15-29. PAR_PCIBR Field Descriptions Bits Name 15–10 — Description Reserved, should be cleared. Writes have no effect and terminate without transfer error exception 9–8 PAR_PCIBR4 PCIBR4 Pin Assignment. Configures the PCIBR4 pin for one of its primary functions or GPIO. 0X PCIBR4 pin configured for general purpose I/O (PPCIREQ4) 10 PCIBR4 pin configured for IRQ4 function 11 PCIBR4 pin configured for PCIBR4 function 7–6 PAR_PCIBR3 PCIBR3 Pin Assignment. Configures the PCIBR3 pin for one of its primary functions or GPIO. 0X PCIBR3 pin configured for general purpose I/O (PPCIREQ3) 10 PCIBR3 pin configured for GP timer TIN3 function 11 PCIBR3 pin configured for PCIBR3 function 5–4 PAR_PCIBR2 PCIBR2 Pin Assignment. Configures the PCIBR2 pin for one of its primary functions or GPIO. 0X PCIBR2 pin configured for general purpose I/O (PPCIREQ2) 10 PCIBR2 pin configured for GP timer TIN2 function 11 PCIBR2 pin configured for PCIBR2 function MCF548x Reference Manual, Rev. 5 15-26 Freescale Semiconductor Memory Map/Register Definition Table 15-29. PAR_PCIBR Field Descriptions (Continued) Bits Description Name 3–2 PAR_PCIBR1 PCIBR1 Pin Assignment. Configures the PCIBR1 pin for one of its primary functions or GPIO. 0X PCIBR1 pin configured for general purpose I/O (PPCIREQ1) 10 PCIBR1 pin configured for GP timer TIN1 function 11 PCIBR1 pin configured for PCIBR1 function 1–0 PAR_PCIBR0 PCIBR0 Pin Assignment. Configures the PCIBR0 pin for one of its primary functions or GPIO. 0X PCIBR0 pin configured for general purpose I/O (PPCIREQ0) 10 PCIBR0 pin configured for GP timer TIN0 function 11 PCIBR0 pin configured for PCIBR0 function 15.3.2.11 PSC3 Pin Assignment Register (PAR_PSC3) The PAR_PSC3 register controls the functions of the PSC3 pins. The PAR_PSC3 register is read/write. 7 6 R 5 PAR_CTS3 4 PAR_RTS3 3 2 1 0 PAR_RXD3 PAR_TXD3 0 0 0 0 0 0 W Reset 0 Reg Addr 0 0 0 MBAR + 0xA4C (PAR_PSC3) Figure 15-28. PSC3 Pin Assignment Register (PAR_PCS3) Table 15-30. PAR_PSC3 Descriptions Bits Name Description 7–6 PAR_CTS3 PSC3CTS pin assignment. Configures the PSC3CTS pin for one of its primary functions or general purpose I/O. 0X PSC3CTS pin configured for general purpose I/O (PPSC3PSC27) 10 PSC3CTS pin configured for PSC3BCLK function 11 PSC3CTS pin configured for PSC3CTS function 5–4 PAR_RTS3 PSC3RTS pin assignment. Configures the PSC3RTS pin for one of its primary functions or general purpose I/O. 0X PSC3RTS pin configured for general purpose I/O (PPSC3PSC26) 10 PSC3RTS pin configured for PSC3FSYNC function 11 PSC3RTS pin configured for PSC3RTS function 3 PAR_RXD3 PSC3RXD pin assignment. Configures the PSC3RXD pin for its primary function or general purpose I/O. 0 PSC3RXD pin configured for general purpose I/O (PPSC3PSC25) 1 PSC3RXD pin configured for PSC3RXD function 2 PAR_TXD3 PSC3TXD pin assignment. Configures the PSC3TXD pin for its primary function or general purpose I/O. 0 PSC3TXD pin configured for general purpose I/O (PPSC3PSC24) 1 PSC3TXD pin configured for PSC3TXD function 1–0 — Reserved, should be cleared. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 15-27 15.3.2.12 PSC2 Pin Assignment Register (PAR_PSC2) The PAR_PSC2 register controls the functions of the PSC2 pins. The PAR_PSC2 register is read/write. 7 R 6 5 PAR_CTS2 4 PAR_RTS2 3 2 1 0 PAR_RXD2 PAR_TXD2 0 0 0 0 0 0 W Reset 0 0 Reg Addr 0 0 MBAR + 0xA4D (PAR_PSC2) Figure 15-29. PSC2 Pin Assignment Register (PAR_PSC2) Table 15-31. PAR_PSC2 Descriptions Bits Name Description 7–6 PAR_CTS2 PSC2CTS pin assignment. Configures the PSC2CTS pin for one of its primary functions or general purpose I/O. 00 PSC2CTS pin configured for general purpose I/O (PPSC3PSC23) 01 PSC2CTS pin configured for FlexCAN CANRX0 10 PSC2CTS pin configured for PSC2BCLK function 11 PSC2CTS pin configured for PSC2CTS function 5–4 PAR_RTS2 PSC2RTS pin assignment. Configures the PSC2RTS pin for one of its primary functions or general purpose I/O. 00 PSC2RTS pin configured for general purpose I/O (PPSC3PSC22) 01 PSC2RTS pin configured for FlexCAN CANTX0 10 PSC2RTS pin configured for PSC2FSYNC function 11 PSC2RTS pin configured for PSC2RTS function 3 PAR_RXD2 PSC2RXD pin assignment. Configures the PSC2RXD pin for its primary function or general purpose I/O. 0 PSC2RXD pin configured for general purpose I/O (PPSC3PSC21) 1 PSC2RXD pin configured for PSC2RXD function 2 PAR_TXD2 PSC2TXD pin assignment. Configures the PSC2TXD pin for its primary function or general purpose I/O. 0 PSC2TXD pin configured for general purpose I/O (PPSC3PSC20) 1 PSC2TXD pin configured for PSC2TXD function 1–0 — Reserved, should be cleared. 15.3.2.13 PSC1 Pin Assignment Register (PAR_PSC1) The PAR_PSC1 register controls the functions of the PSC1 pins. The PAR_PSC1 register is read/write. MCF548x Reference Manual, Rev. 5 15-28 Freescale Semiconductor Memory Map/Register Definition 7 R 6 5 PAR_CTS1 4 PAR_RTS1 3 2 PAR_RXD1 PAR_TXD1 1 0 0 0 0 0 W Reset 0 0 0 Reg Addr 0 0 0 MBAR + 0xA4E (PAR_PSC1) Figure 15-30. PSC1 Pin Assignment Register (PAR_PSC1) Table 15-32. PAR_PCS1 Descriptions Bits Name Description 7–6 PAR_CTS1 PSC1CTS pin assignment. Configures the PSC1CTS pin for one of its primary functions or general purpose I/O. 0X PSC1CTS pin configured for general purpose I/O (PPSC1PSC07) 10 PSC1CTS pin configured for PSC1BCLK function 11 PSC1CTS pin configured for PSC1CTS function 5–4 PAR_RTS1 PSC1RTS pin assignment. Configures the PSC1RTS pin for one of its primary functions or general purpose I/O. 0X PSC1RTS pin configured for general purpose I/O (PPSC1PSC06) 10 PSC1RTS pin configured for PSC1FSYNC function 11 PSC1RTS pin configured for PSC1RTS function 3 PAR_RXD1 PSC1RXD Pin Assignment. Configures the PSC1RXD pin for its primary function or general purpose I/O. 0 PSC1RXD pin configured for general purpose I/O (PPSC1PSC05) 1 PSC1RXD pin configured for PSC1RXD function 2 PAR_TXD1 PSC1TXD Pin Assignment. Configures the PSC1TXD pin for its primary function or general purpose I/O. 0 PSC1TXD pin configured for general purpose I/O (PPSC1PSC04) 1 PSC1TXD pin configured for PSC1TXD function 1–0 — Reserved, should be cleared. 15.3.2.14 PSC0 Pin Assignment Register (PAR_PSC0) The PAR_PSC0 register controls the functions of the PSC0 pins. The PAR_PSC0 register is read/write. 7 R 6 5 PAR_CTS0 4 PAR_RTS0 3 2 1 0 PAR_RXD0 PAR_TXD0 0 0 0 0 0 0 W Reset 0 Reg Addr 0 0 0 MBAR + 0xA4F (PAR_PSC0) Figure 15-31. PSC0 Pin Assignment Register (PAR_PSC0) MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 15-29 Table 15-33. PAR_PCS0 Descriptions Bits Name Description 7–6 PAR_CTS0 PSC0CTS pin assignment. Configures the PSC0CTS pin for one of its primary functions or general purpose I/O. 0X PSC0CTS pin configured for general purpose I/O (PPSC1PSC03) 10 PSC0CTS pin configured for PSC0BCLK function 11 PSC0CTS pin configured for PSC0CTS function 5–4 PAR_RTS0 PSC0RTS pin assignment. Configures the PSC0RTS pin for one of its primary functions or general purpose I/O. 0X PSC0RTS pin configured for general purpose I/O (PPSC1PSC02) 10 PSC0RTS pin configured for PSC0FSYNC function 11 PSC0RTS pin configured for PSC0RTS function 3 PAR_RXD0 PSC0RXD Pin Assignment. Configures the PSC0RXD pin for its primary function or general purpose I/O. 0 PSC0RXD pin configured for general purpose I/O (PPSC1PSC01) 1 PSC0RXD pin configured for PSC0RXD function 2 PAR_TXD0 PSC0TXD Pin Assignment. Configures the PSC0TXD pin for its primary function or general purpose I/O. 0 PSC0TXD pin configured for general purpose I/O (PPSC1PSC00) 1 PSC0TXD pin configured for PSC0TXD function 1–0 — Reserved, should be cleared. 15.3.2.15 DSPI Pin Assignment Register (PAR_DSPI) The PAR_DSPI register controls the functions of MCF548x DSPI pins. The PAR_DSPI register is read/write. R 15 14 13 12 0 0 0 PAR_ CS5 0 0 0 0 W Reset Reg Addr 11 10 9 8 7 6 PAR_CS3 PAR_CS2 PAR_CS0 0 0 0 0 0 5 4 PAR_SCK 0 0 0 3 2 PAR_SIN 0 0 1 0 PAR_SOUT 0 0 MBAR + 0xA50 (PAR_DSPI) Figure 15-32. DSPI Pin Assignment Register (PAR_DSPI) Table 15-34. PAR_DSPI Descriptions Bits Name 15–13 — 12 PAR_CS5 Description Reserved, should be cleared. DSPICS5/PCSS pin assignment. Configures the DSPICS5/PCSS pin for its primary function or general purpose I/O. 0 DSPICS5/PCSS pin configured for general purpose I/O (PDSPI6) 1 DSPICS5/PCSS pin configured for DSPICS5/PCSS function MCF548x Reference Manual, Rev. 5 15-30 Freescale Semiconductor Memory Map/Register Definition Table 15-34. PAR_DSPI Descriptions (Continued) Bits Name Description 11–10 PAR_CS3 DSPICS3 pin assignment. Configures the DSPICS3 pin for its primary function or general purpose I/O. 00 DSPICS3 pin configured for general purpose I/O (PDSPI5) 01 DSPICS3 pin configured for FlexCAN CANTX1 10 DSPICS3 pin configured for GP timer TOUT3 function 11 DSPICS3 pin configured for DSPICS3 function 9–8 PAR_CS2 DSPICS2 pin assignment. Configures the DSPICS2 pin for its primary function or general purpose I/O. 00 DSPICS2 pin configured for general purpose I/O (PDSPI4) 01 DSPICS2 pin configured for FlexCAN CANTX1 10 DSPICS2 pin configured for GP timer TOUT2 function 11 DSPICS2 pin configured for DSPICS2 function 7–6 PAR_CS0 DSPICS0/SS pin assignment. Configures the DSPICS0/SS pin for its primary function or general purpose I/O. 00 DSPICS0/SS pin configured for general purpose I/O (PDSPI3) 01 DSPICS0/SS pin configured for PSC3FSYNC data 10 DSPICS0/SS pin configured for PSC3RTS function 11 DSPICS0/SS pin configured for DSPICS0/SS function 5–4 PAR_SCK DSPISCK pin assignment. Configures the DSPISCK pin for its primary function or general purpose I/O. 00 DSPISCK pin configured for general purpose I/O (PDSPI2) 01 DSPISCK pin configured for PSC3BCLK data 10 DSPISCK pin configured for PSC3CTS function 11 DSPISCK pin configured for DSPISCK function 3–2 PAR_SIN DSPISIN pin assignment. Configures the DSPISIN pin for its primary function or general purpose I/O. 0X DSPISIN pin configured for general purpose I/O (PDSPI1) 10 DSPISIN pin configured for PSC3RXD function 11 DSPISIN pin configured for DSPISIN function 1–0 PAR_SOUT DSPISOUT pin assignment. Configures the DSPISOUT pin for its primary function or general purpose I/O. 0X DSPISOUT pin configured for general purpose I/O (PDSPI0) 10 DSPISOUT pin configured for PSC3TXD function 11 DSPISOUT pin configured for DSPISOUT function 15.3.2.16 General Purpose Timer Pin Assignment Register (PAR_TIMER) The PAR_TIMER register controls the functions of MCF548x general purpose timer pins. The PAR_TIMER register is read/write. R 7 6 0 0 0 0 5 4 PAR_TIN3 3 2 PAR_TOUT3 1 PAR_TIN2 0 PAR_TOUT2 W Reset Reg Addr 1 1 1 1 1 1 MBAR + 0xA52 (PAR_TIMER) Figure 15-33. General Purpose Timer Pin Assignment Register (PAR_TIMER) MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 15-31 Table 15-35. PAR_TIMER Descriptions Bits Name 7–6 — 5–4 PAR_TIN3 3 2–1 0 Description Reserved, should be cleared. TIN3 pin assignment. Configures the TIN3 pin for its primary function 0X TIN3 pin configured for FlexCAN CANRX1 10 TIN3 pin configured for IRQ3 function 11 TIN3 pin configured for GP timer TIN3 function or general purpose input Note: General purpose input is obtained on the TIN3 pin by (1) writing 3 to the PAR_TIN3 field and (2) disabling the timer function in the general purpose timer module. General purpose output is not possible on the TIN3 pin. PAR_TOUT TOUT3 pin assignment. Configures the TOUT3 pin for its primary function 3 0 TOUT3 pin configured for FlexCAN CANTX1 1 TOUT3 pin configured for GP timer TOUT3 function or general purpose output Note: General purpose output is obtained on the TOUT3 pin by (1) writing 1 to the PAR_TOUT3 field and (2) disabling the timer function in the general purpose timer module. General purpose input is not possible on the TOUT3 pin. PAR_TIN2 TIN2 pin assignment. Configures the TIN2 pin for its primary function 0X TIN2 pin configured for FlexCAN CANRX1 10 TIN2 pin configured for IRQ2 function 11 TIN2 pin configured for GP timer TIN2 function or general purpose input Note: General purpose input is obtained on the TIN2 pin by (1) writing 3 to the PAR_TIN2 field and (2) disabling the timer function in the general purpose timer module. General purpose output is not possible on the TIN2 pin. PAR_TOUT TOUT2 pin assignment. Configures the TOUT2 pin for its primary function 2 0 TOUT2 pin configured for FlexCAN CANTX1 1 TOUT2 pin configured for GP timer TOUT2 function or general purpose output Note: General purpose output is obtained on the TOUT2 pin by (1) writing 1 to the PAR_TOUT2 field and (2) disabling the timer function in the general purpose timer module. General purpose input is not possible on the TOUT2 pin. NOTE Explicit pin function assignment capability for the TIN1, TOUT1, TIN0, and TOUT0 pins is not needed in the GPIO module since these pins only have the primary timer functions and general purpose I/O. Switching between the primary timer functions and GPIO is handled by the general purpose timer module. 15.4 15.4.1 Functional Description Overview Initial pin function is determined during reset configuration. See Chapter 2, “Signal Descriptions,” for more details. Most pins are configured as general purpose I/O by default. The notable exceptions to this are FlexBus control pins. These pins are configured for their primary functions after reset. The pin assignment registers allow the user to select among various primary functions and general purpose I/O after reset. Every general purpose I/O pin is individually configurable as an input or an output via a data direction register (PDDR_x). Every GPIO port has an output data register (PODR_x) and a pin data register MCF548x Reference Manual, Rev. 5 15-32 Freescale Semiconductor Functional Description (PPDSDR_x) to monitor and control the state of its pins. Data written to a PODR_x register is stored and then driven to the corresponding port x pins configured as outputs. Reading a PODR_x register returns the current state of the register regardless of the state of the corresponding pins. Reading a PPDSDR_x register returns the current state of the corresponding pins when configured as general purpose I/O, regardless of whether the pins are inputs or outputs. Every GPIO port has a PPDSDR_x register and a clear register (PCLRR_x) for setting or clearing individual bits in the PODR_x register. The MCF548x GPIO module does not generate interrupt requests. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 15-33 MCF548x Reference Manual, Rev. 5 15-34 Freescale Semiconductor Part III On-Chip Integration Part III describes on-chip integration for the MCF548x device. It includes descriptions of the system SRAM, SDRAM controller, PCI, FlexBus interface, FlexCAN, SEC cryptography accelerator, and JTAG. Contents Part III contains the following chapters: • Chapter 16, “32-Kbyte System SRAM,” describes the MCF548x on-chip system SRAM implementation. It covers general operations, configuration, and initialization. • Chapter 17, “FlexBus,” describes data transfer operations, error conditions, and reset operations. It describes transfers initiated by the MCF548x and by an external master, and includes detailed timing diagrams showing the interaction of signals in supported bus operations. • Chapter 18, “SDRAM Controller (SDRAMC),” describes configuration and operation of the synchronous DRAM controller component of the SIU. It includes a description of signals involved in DRAM operations, including chip select signals and their address, mask, and control registers. • Chapter 19, “PCI Bus Controller,” details the operation of the PCI bus controller for the MCF548x. • Chapter 20, “PCI Bus Arbiter Module,” describes the MCF548x PCI bus arbiter module, including timing for request and grant handshaking, the arbitration process, and the register in the PCI bus arbiter programing model. • Chapter 21, “FlexCAN,” describes the MCF548x implementation of the controller area network (CAN) protocol. This chapter describes FlexCAN module operation and provides a programming model. • Chapter 22, “Integrated Security Engine (SEC),” provides an overview of the MCF548x security encryption controller. • Chapter 23, “IEEE 1149.1 Test Access Port (JTAG),” describes configuration and operation of the MCF548x JTAG test implementation. It describes the use of JTAG instructions and provides information on how to disable JTAG functionality. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor i MCF548x Reference Manual, Rev. 5 ii Freescale Semiconductor Chapter 16 32-Kbyte System SRAM 16.1 Introduction This chapter explains the operation of the MCF548x 32-Kbyte system SRAM. 16.1.1 Block Diagram The system SRAM is organized as four 8-Kbyte banks, each organized as 2048 × 32-bits. The four banks occupy a contiguous block of memory but can be optionally interleaved on long-word boundaries. When configured for interleaved access, each bank contains the data for long word address modulo {bank #} (e.g. bank 2 contains data for all long word address modulo 2 locations). Figure 16-1 shows the SRAM organization in both linear and interleaved modes. Byte Address Byte Address Long Word 0 Long Word 1 Long Word 2 . . . 0x1_0000 0x1_0004 0x1_0008 . . . 0x1_1FFC 0x1_2000 0x1_2004 0x1_2008 . . . 0x1_3FFC 0x1_4000 0x1_4004 0x1_4008 . . . 0x1_5FFC 0x1_6000 0x1_6004 0x1_6008 . . . 0x1_7FFC Long Word 2047 Long Word 2048 Long Word 2049 Long Word 2050 . . . Long Word 4095 Long Word 4096 Long Word 4097 Long Word 4098 . . . Long Word 6143 Long Word 6144 Long Word 6145 Long Word 6146 . . . Long Word 8191 Linear Organization Long Word 0 Long Word 4 Long Word 8 . . . 0x1_0000 Bank 0 Bank 1 Bank 2 Bank 3 0x1_0010 0x1_0020 . . . 0x1_7FF0 0x1_0004 Long Word 8188 Long Word 1 Long Word 5 Long Word 9 . . . 0x1_0014 0x1_0024 . . . 0x1_7FF4 0x1_0008 Long Word 8189 Long Word 2 Long Word 6 Long Word 10 . . . 0x1_0018 0x1_0028 . . . 0x1_7FF8 0x1_000C Long Word 8190 Long Word 3 Long Word 7 Long Word 11 . . . 0x1_001C 0x1_002C . . . 0x1_7FFC Bank 0 Bank 1 Bank 2 Bank 3 Long Word 8191 Interleaved Organization Figure 16-1. SRAM Organization MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 16-1 The system SRAM contents always reside at MBAR + 0x0001 0000; therefore, it can be relocated by changing the MBAR contents. 16.1.2 Features The 32-Kbyte system SRAM is intended primarily as a fast scratch memory and data buffer for DMA and SEC processing, and as memory accessed through the shared bus by all system masters. The module features are the following: • Four 8-Kbyte banks, each organized as 2048 × 32-bits • Dedicated 32-bit data bus per bank • Optionally interleaved along long-word boundaries under software control • Single cycle access when accessed by the DMA • Byte, word, and longword addressing capabilities • Independent arbitration mechanism per bank 16.1.3 Overview This module provides a general-purpose memory block that can be accessed by the masters in the system (ColdFire core, SEC, DMA, and PCI) via the shared internal system bus. The SRAM is also accessed directly (without going through the system bus) by the SEC and DMA. This allows a mechanism for the sharing of parameter data among the various masters as well as a dedicated fast scratch memory and data buffer for DMA and SEC processing tasks. In order to maximize concurrent utilization, the system SRAM is organized as four banks. Each master is allocated a maximum transfer count and must give up access to the bank when its transfer count has been depleted. In this fashion, each master is given the opportunity to access each bank to prevent starvation of any given master. The transfer counts are configurable under software control for each master and each bank, so it can be optimized to maximize the SRAM utilization for specific tasks. Optionally, a master can be set to “own” a bank, whereby all other masters can access the bank only when the “own” master is not making accesses to the bank. 16.2 Memory Map/Register Definition Table 16-1 shows the memory map of the system SRAM module. For more information about a particular register, refer to the description of the register in the following sections. Table 16-1. System SRAM Memory Map Address (MBAR + ) Name Byte 0 Byte 1 Byte 2 Byte 3 Access 0x1_0000– 0x1_7FFC SRAM Contents R/W 0x1_FFC0 System SRAM Configuration Register SSCR R/W 0x1_FFC4 Transfer Count Configuration Register TCCR R/W 0x1_FFC8 Transfer Count Configuration Register - DMA Read Channel TCCRDR R/W MCF548x Reference Manual, Rev. 5 16-2 Freescale Semiconductor Memory Map/Register Definition Table 16-1. System SRAM Memory Map (Continued) Address (MBAR + ) Name Byte 0 Byte 1 Byte 2 Byte 3 Access 0x1_0000– 0x1_7FFC SRAM Contents 0x1_FFCC Transfer Count Configuration Register - DMA Write Channel TCCRDW R/W 0x1_FFD0 Transfer Count Configuration Register - SEC TCCRSEC R/W 16.2.1 R/W System SRAM Configuration Register (SSCR) This register is used to define the base address of the system SRAM and whether to interleave the banks. R 31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 INLV 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 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 0 0 W Reset R W Reset Reg Addr MBAR + 0x1_FFC0 Figure 16-2. System SRAM Configuration Register (SSCR) Each field is described in Table 16-2. Table 16-2. SSCR Register Field Descriptions Bits Name 31–17 — 16 INLV 15–0 — Description Reserved, should be cleared. Interleave enable. Controls whether the banks are interleaved along longword boundaries or linear. 0 The four SRAM banks are not interleaved (linear). 1 The four SRAM banks are interleaved. SRAM bank # contains data for long word address modulo {bank #} Reserved. Should be cleared. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 16-3 16.2.2 Transfer Count Configuration Register (TCCR) This register is used to configure the allocated maximum transfer count for each bank for the following masters: the ColdFire core, DMA, SEC, or PCI. This occurs as they access memory through the shared system bus. The DMA and the SEC can access the system SRAM either via the system bus or via their dedicated ports. Refer to sections 16.2.3 through 16.2.5. 31 30 29 28 0 0 0 0 0 0 0 0 1 1 1 15 14 13 12 11 10 9 0 0 0 0 0 0 0 0 R 27 26 25 24 23 22 21 20 0 0 0 0 1 0 0 0 0 1 1 1 1 8 7 6 5 4 3 2 1 0 0 0 0 0 0 0 0 0 BANK3_TC 19 18 17 16 BANK2_TC W Reset R BANK1_TC BANK0_TC W Reset 1 1 Reg Addr 1 1 1 1 1 1 MBAR + 0x1_FFC4 Figure 16-3. Transfer Count Configuration Register (TCCR) Each field is described in the Table 16-3. Table 16-3. TCCR Register Field Descriptions Bits Name 31–28 — 27–24 23–20 19–16 15–12 11–8 7–4 3–0 Description Reserved, should be cleared. BANK3_TC Bank three transfer count. This field indicates the maximum transfer count for bank 3. The master can make at most 4 * {field value} 32-bit transfers to/from bank 3 before it must wait for other masters to complete their transfers. If this field is programmed to “0” the master can “own” bank 3 for arbitrarily long transfers. — Reserved, should be cleared. BANK2_TC Bank two transfer count. This field indicates the maximum transfer count for bank 2. The master can make at most 4 * {field value} 32-bit transfers to/from bank 2 before it must wait for other masters to complete their transfers. If this field is programmed to “0” the master can “own” bank 2 for arbitrarily long transfers. — Reserved. Should be cleared. BANK1_TC Bank one transfer count. This field indicates the maximum transfer count for bank 1. The master can make at most 4 * {field value} 32-bit transfers to/from bank 1 before it must wait for other masters to complete their transfers. If this field is programmed to “0” the master can “own” bank 1 for arbitrarily long transfers. — Reserved. Should be cleared. BANK0_TC Bank zero transfer count. This field indicates the maximum transfer count for bank 0. The master can make at most 4 * {field value} 32-bit transfers to/from bank 0 before it must wait for other masters to complete their transfers. If this field is programmed to “0” the master can “own” bank 0 for arbitrarily long transfers. MCF548x Reference Manual, Rev. 5 16-4 Freescale Semiconductor Memory Map/Register Definition 16.2.3 Transfer Count Configuration Register—DMA Read Channel (TCCRDR) This register is used to configure the allocated maximum transfer count for each bank for the DMA read channel as it accesses SRAM directly, without going through the system bus. R 31 30 29 28 27 26 25 0 0 0 0 0 0 0 0 1 1 1 15 14 13 12 11 10 9 0 0 0 0 0 0 0 0 24 23 22 21 20 0 0 0 0 1 0 0 0 0 1 1 1 1 8 7 6 5 4 3 2 1 0 0 0 0 0 0 0 0 0 BANK3_TC 19 18 17 16 BANK2_TC W Reset R BANK1_TC BANK0_TC W Reset 1 1 Reg Addr 1 1 1 1 1 1 MBAR + 0x1_FFC8 Figure 16-4. Transfer Count Configuration Register—DMA Read Channel (TCCRDR) Each field is described in the table below. Table 16-4. TCCRDR Register Field Descriptions Bits Name 31–28 — 27–24 23–20 19–16 15–12 11–8 7–4 3–0 Description Reserved, should be cleared. BANK3_TC Bank three transfer count. This field indicates the maximum transfer count for bank 3. The DMA read channel can make at most 4 * {field value} 32-bit transfers from bank 3 before it must wait for other masters to complete their transfers. If this field is programmed to “0” the DMA read channel can “own” bank 3 for arbitrarily long transfers. — Reserved, should be cleared. BANK2_TC Bank two transfer count. This field indicates the maximum transfer count for bank 2. The DMA read channel can make at most 4 * {field value} 32-bit transfers from bank 2 before it must wait for other masters to complete their transfers. If this field is programmed to “0” the DMA read channel can “own” bank 2 for arbitrarily long transfers. — Reserved, should be cleared. BANK1_TC Bank one transfer count. This field indicates the maximum transfer count for bank 1. The DMA read channel can make at most 4 * {field value} 32-bit transfers from bank 1 before it must wait for other masters to complete their transfers. If this field is programmed to “0” the DMA read channel can “own” bank 1 for arbitrarily long transfers. — Reserved, should be cleared. BANK0_TC Bank zero transfer count. This field indicates the maximum transfer count for bank 0. The DMA read channel can make at most 4 * {field value} 32-bit transfers from bank 0 before it must wait for other masters to complete their transfers. If this field is programmed to “0” the DMA read channel can “own” bank 0 for arbitrarily long transfers. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 16-5 16.2.4 Transfer Count Configuration Register—DMA Write Channel (TCCRDW) This register is used to configure the allocated maximum transfer count for each bank of the DMA write channel as it accesses SRAM directly, without going through the system bus. R 31 30 29 28 27 26 25 0 0 0 0 0 0 0 0 1 1 1 15 14 13 12 11 10 9 0 0 0 0 0 0 0 0 24 23 22 21 20 0 0 0 0 1 0 0 0 0 1 1 1 1 8 7 6 5 4 3 2 1 0 0 0 0 0 0 0 0 0 BANK3_TC 19 18 17 16 BANK2_TC W Reset R BANK1_TC BANK0_TC W Reset 1 1 Reg Addr 1 1 1 1 1 1 MBAR + 0x1_FFCC Figure 16-5. Transfer Count Configuration Register—DMA Write Channel (TCCRDW) Each field is described in the table below. Table 16-5. TCCRDW Register Field Descriptions Bits Name 31–28 — 27–24 23–20 19–16 15–12 11–8 7–4 3–0 Description Reserved, should be cleared. BANK3_TC Bank three transfer count. This field indicates the maximum transfer count for bank 3. The DMA write channel can make at most 4 * {field value} 32-bit transfers to bank 3 before it must wait for other masters to complete their transfers. If this field is programmed to “0” the DMA write channel can “own” bank 3 for arbitrarily long transfers. — Reserved, should be cleared. BANK2_TC Bank two transfer count. This field indicates the maximum transfer count for bank 2. The DMA write channel can make at most 4 * {field value} 32-bit transfers to bank 2 before it must wait for other masters to complete their transfers. If this field is programmed to “0” the DMA write channel can “own” bank 2 for arbitrarily long transfers. — Reserved, should be cleared. BANK1_TC Bank one transfer count. This field indicates the maximum transfer count for bank 1. The DMA write channel can make at most 4 * {field value} 32-bit transfers to bank 1 before it must wait for other masters to complete their transfers. If this field is programmed to “0” the DMA write channel can “own” bank 1 for arbitrarily long transfers. — Reserved, should be cleared. BANK0_TC Bank zero transfer count. This field indicates the maximum transfer count for bank 0. The DMA write channel can make at most 4 * {field value} 32-bit transfers to bank 0 before it must wait for other masters to complete their transfers. If this field is programmed to “0” the DMA write channel can “own” bank 0 for arbitrarily long transfers. MCF548x Reference Manual, Rev. 5 16-6 Freescale Semiconductor Memory Map/Register Definition 16.2.5 Transfer Count Configuration Register—SEC (TCCRSEC) This register is used to configure the allocated maximum transfer count for each bank for the SEC as it accesses SRAM directly, without going through the system bus. R 31 30 29 28 27 26 25 0 0 0 0 0 0 0 0 1 1 1 15 14 13 12 11 10 9 0 0 0 0 0 0 0 0 24 23 22 21 20 0 0 0 0 1 0 0 0 0 1 1 1 1 8 7 6 5 4 3 2 1 0 0 0 0 0 0 0 0 0 BANK3_TC 19 18 17 16 BANK2_TC W Reset R BANK1_TC BANK0_TC W Reset 1 1 Reg Addr 1 1 1 1 1 1 MBAR + 0x1_FFD0 Figure 16-6. Transfer Count Configuration Register—SEC (TCCRSEC)) Each field is described in the table below. Table 16-6. TCCRSEC Register Field Descriptions Bits Name 31–28 — 27–24 23–20 19–16 15–12 11–8 7–4 3–0 Description Reserved, should be cleared. BANK3_TC Bank three transfer count. This field indicates the maximum transfer count for bank 3. The SEC can make at most 4 * {field value} 32-bit transfers to/from bank 3 before it must wait for other masters to complete their transfers. If this field is programmed to “0” the SEC can “own” bank 3 for arbitrarily long transfers. — Reserved, should be cleared. BANK2_TC Bank two transfer count. This field indicates the maximum transfer count for bank 2. The SEC can make at most 4 * {field value} 32-bit transfers to/from bank 2 before it must wait for other masters to complete their transfers. If this field is programmed to “0” the SEC can “own” bank 2 for arbitrarily long transfers. — Reserved, should be cleared. BANK1_TC Bank one transfer count. This field indicates the maximum transfer count for bank 1. The SEC can make at most 4 * {field value} 32-bit transfers to/from bank 1 before it must wait for other masters to complete their transfers. If this field is programmed to “0” the SEC can “own” bank 1 for arbitrarily long transfers. — Reserved, should be cleared. BANK0_TC Bank zero transfer count. This field indicates the maximum transfer count for bank 0. The SEC can make at most 4 * {field value} 32-bit transfers to/from bank 0 before it must wait for other masters to complete their transfers. If this field is programmed to “0” the SEC can “own” bank 0 for arbitrarily long transfers. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 16-7 16.3 Functional Description The system SRAM decodes the addresses for all four banks to determine which master is trying to access which bank. The system SRAM module provides a bus arbitration mechanism for granting access of each bank to each master. All masters simply request a data transfer and the SRAM grants a specified cycle count to the appropriate master. The arbitration is overlapped with the address phase of SRAM transfers and therefore imposes no performance penalty or overhead. The current master pointer for each bank is determined as shown in Figure 16-7. The current master pointer transitions to another master when the current master’s maximum transfer count is exceeded, or the current master is idle and another master requests access to the bank. Otherwise, the current master pointer remains unchanged. RE S (Ba ET nk 0 ) ET RES 1) k (Ban Master DMA-R SEC DMA-W SET RE ) nk 3 (Ba RES E (Ban T k 2) Figure 16-7. SRAM Arbitration MCF548x Reference Manual, Rev. 5 16-8 Freescale Semiconductor Chapter 17 FlexBus 17.1 Introduction This chapter describes data transfer operations, error conditions, and reset operations. It describes transfers initiated by the MCF548x and includes detailed timing diagrams showing the interaction of signals in supported bus operations. NOTE Unless otherwise noted, in this chapter the term ‘clock’ refers to the CLKIN used for the bus. 17.1.1 Overview A multi-function external bus interface called the FlexBus interface controller is provided on the MCF548x with basic functionality of interfacing to slave-only devices up to a maximum bus frequency of 50 MHz. It can be directly connected to asynchronous or synchronous devices such as external boot ROMs, flash memories, gate-array logic, or other simple target (slave) devices with little or no additional circuitry. For asynchronous devices a simple chip-select based interface can be used. The FlexBus interface has six general purpose chip-selects (FBCS[5:0]). Chip-select FBCS0 can be dedicated to boot ROM access and can be programmed to be byte (8 bits), word (16 bits), or longword (32 bits) wide. Control signal timing is compatible with common ROM / flash memories. 17.1.2 Features The following list summarizes the key FlexBus features: • Six independent, user-programmable chip-select signals (FBCS[5:0]) that can interface with SRAM, PROM, EPROM, EEPROM, Flash, and other peripherals • 8-, 16-, and 32-bit port sizes with configuration for multiplexed or non-multiplexed address and data buses • Byte, word, and longword, and line sized transfers • Programmable burst and burst-inhibited transfers selectable for each chip select and transfer direction • Programmable address setup time with respect to the assertion of chip select • Programmable address hold time with respect to the negation of chip select and transfer direction 17.1.3 Modes of Operation The FlexBus interface is a configurable multiplexed bus that is set to one of four modes: • Multiplexed 32-bit address and 32-bit data • Multiplexed 32-bit address and 16-bit data (non-multiplexed 16-bit address and 16-bit data) • Multiplexed 32-bit address and 8-bit data (non-multiplexed 24-bit address and 8-bit data) • Non-multiplexed 32-bit address with 32-bit data MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 17-1 17.2 Byte Lanes Figure 17-1 shows the byte lanes that external memory should be connected to and the sequential transfers if a longword is transferred for three port sizes. For example, an 8-bit memory should be connected to AD[31:24] (BE/BWE0). A longword transfer takes four transfers on AD[31:24], starting with the MSB and going to the LSB. Byte Select Processor External Data Bus BE/BWE0 BE/BWE1 BE/BWE2 BE/BWE3 AD[31:24] AD[23:16] AD[15:8] AD[7:0] 32-Bit Port Memory Byte 0 Byte 1 16-Bit Port Memory Byte 0 Byte 1 Byte 2 Byte 3 8-Bit Port Memory Byte 2 Byte 3 Driven with address values Byte 0 Byte 1 Byte 2 Driven with address values Byte 3 Figure 17-1. Connections for External Memory Port Sizes 17.3 Address Latch Because the FlexBus uses a multiplexed address and data bus, external logic might be needed in some cases to capture the address phase as shown in Figure 17-2. MCF548x Reference Manual, Rev. 5 17-2 Freescale Semiconductor External Signals External Device / Peripheral DATA[31:Y] AD[31:0] FlexBus Address ADDR[X:0] Latch Logic ALE Interface ALE Controller R/W R/W TSIZ[1:0] SIZ[1:0] TBST BURST BE/BWE[3:0] BE/BWE[3:0] OE OE TA TA FBCSx CS Figure 17-2. Multiplexed FlexBus Implementation 17.4 External Signals This section describes the external signals that are involved in data transfer operations. Table 17-1 summarizes the MCF548x FlexBus signals. Table 17-1. FlexBus Signal Summary Signal Name Direction Description Reset State FBCS[5:0] O General purpose chip-selects Hi-Z AD[31:0] I/O Address / Data bus Hi-Z ALE O Address Latch Enable Hi-Z BE/BWE[3:0] O Byte Selects Hi-Z OE O Output Enable Hi-Z R/W O Read/Write. 1 = Read, 0 = Write Hi-Z TBST O Burst Transfer indicator Hi-Z TSIZ[1:0] O Transfer Size Hi-Z TA I Transfer Acknowledge — MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 17-3 17.4.1 Chip-Select (FBCS[5:0]) The chip-select signal indicates which device is being selected. A particular chip-select asserts when the transfer address is within the device’s address space as defined in the base and mask address registers, see Section 17.5.2, “Chip-Select Registers.” 17.4.2 Address/Data Bus (AD[31:0]) The AD[31:0] bus carries address and data. The full 32-bit address is always driven on the first clock of a bus cycle (address phase). The number of byte lanes used to carry the data during the data phase is determined by the port size associated with the matching chip select. In non-multiplexed mode, it is divided into sub-buses: address (output) and data (input/output). In multiplexed mode, it carries the address during the address phase and the data during the data phase. Note that in multiplexed mode and during the data phase, the address continues driving on the lower byte lanes if these lanes are not used to carry the data. 17.4.3 Address Latch Enable (ALE) The assertion of ALE indicates that the MCF548x has begun a bus transaction and that the address and attributes are valid. ALE is asserted for one bus clock cycle. In multiplexed bus mode, ALE is used externally as an address latch enable to capture the address phase of the bus transfer, as shown in Figure 17-2. 17.4.4 Read/Write (R/W) MCF548x drives the R/W signal to indicate the direction of the current bus operation. It is driven high during read bus cycles and driven low during write bus cycles. 17.4.5 Transfer Burst (TBST) Transfer Burst indicates that a burst transfer is in progress as driven by the MCF548x. A burst transfer can be 2 to 16 beats depending on TSIZ[1:0] and the port size. NOTE When burst (TBST = 0) and transfer size is 16 bytes (TSIZ = 2’b11) and the address is misaligned within the 16-byte boundary, the external device must be able to wrap around the address. 17.4.6 Transfer Size (TSIZ[1:0]) For memory accesses, these signals, along with TBST, indicate the data transfer size of the current bus operation. The FlexBus interface supports byte, word, and longword operand transfers and allows accesses to 8-, 16-, and 32-bit data ports. For misaligned transfers, TSIZ[1:0] indicate the size of each transfer. For example, if a longword access through a 32-bit port device occurs at a misaligned offset of 0x1, a byte is transferred first (TSIZ[1:0] = 01), a word is next transferred at offset 0x2 (TSIZ[1:0] = 10), then the final byte is transferred at offset 0x4 (TSIZ[1:0] = 01). MCF548x Reference Manual, Rev. 5 17-4 Freescale Semiconductor External Signals For aligned transfers larger than the port size, TSIZ[1:0] behaves as follows: • If bursting is used, TSIZ[1:0] is driven to the size of transfer. • If bursting is inhibited, TSIZ[1:0] first shows the size of the entire transfer and then shows the port size. Table 17-2. Data Transfer Size TSIZ[1:0] Transfer Size 00 4 bytes (longword) 01 1 byte 10 2 bytes (word) 11 16 bytes (line) For burst-inhibited transfers, TSIZ[1:0] changes with each ALE assertion to reflect the next transfer size. For transfers to port sizes smaller than the transfer size, TSIZ[1:0] indicates the size of the entire transfer on the first access and the size of the current port transfer on subsequent transfers. For example, for a longword write to an 8-bit port, TSIZ[1:0] = 2’b00 for the first transaction and 2’b01 for the next three transactions. If bursting is used and in the case of longword write to an 8-bit port, TSIZ[1:0] is driven to 2’b00 for the entire transfer. 17.4.7 Byte Selects (BE/BWE[3:0]) The byte strobe (BE/BWE[3:0]) outputs indicate that data is to be latched or driven onto a byte of the data when driven low as shown in Table 17-1. BE/BWEn signals are asserted only to the memory bytes used during a read or write access. 17.4.8 Output Enable (OE) The output enable signal (OE) is sent to the interfacing memory and/or peripheral to enable a read transfer. OE is asserted only when a chip select matches the current address decode. 17.4.9 Transfer Acknowledge (TA) This signal indicates that the external data transfer is complete. During a read cycle, when the processor recognizes TA, it latches the data and then terminates the bus cycle. During a write cycle, when the processor recognizes TA, the bus cycle is terminated. If auto-acknowledge is disabled, the external device drives TA to terminate the bus transfer; if auto-acknowledge is enabled, the TA is generated internally after a specified wait states or the external device may assert external TA before the wait-state countdown, terminating the cycle early. The MCF548x negates FBCSn a cycle after the last TA asserts. During read cycles, the peripheral must continue to drive data until TA is recognized. For write cycles, the processor continues to drive data one clock after FBCSn is negated. The number of wait states is determined either by internally programmed auto acknowledgement or by the external TA input. If the external TA is used, the peripheral has total control on the number of wait states. NOTE External devices should only assert TA while the FBCSn signal to the external device is asserted. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 17-5 17.5 Chip-Select Operation Each chip-select has a dedicated set of the following registers for configuration and control: • Chip-select address registers (CSARn) control the base address space of the chip-select. See Section 17.5.2.1, “Chip-Select Address Registers (CSAR0–CSAR5).” • Chip-select mask registers (CSMRn) provide 16-bit address masking and access control. See Section 17.5.2.2, “Chip-Select Mask Registers (CSMR0–CSMR5).” • Chip-select control registers (CSCRn) provide port size and burst capability indication, wait-state generation, address setup and hold times, and automatic acknowledge generation features. See Section 17.5.2.3, “Chip-Select Control Registers (CSCR0–CSCR5).” FBCS0 is a global chip-select after reset and provides re-locatable boot ROM capability. 17.5.1 General Chip-Select Operation When a bus cycle is initiated, the MCF548x first compares its address with the base address and mask configurations programmed for chip-selects 0–5 (configured in CSCR0–CSCR5). If the driven address matches a programmed chip-select, the appropriate chip-select is asserted fulfilling the requirements as programmed in the respective configuration register. 17.5.1.1 8-, 16-, and 32-Bit Port Sizing Static bus sizing is programmable through the port size bits, CSCR[PS]. See Section 17.5.2.3, “Chip-Select Control Registers (CSCR0–CSCR5).” Note that the MCF548x always drives 32-bit address on the AD bus in the first cycle regardless of the external device’s address size. The external device must connect its address lines to the appropriate AD bits starting from AD0 and upward. It must also connect its data lines to the AD bus starting from the AD31 and downward. No bit ordering is required when connecting address and data lines to the AD bus. For example, a 16-bit address/16-bit data device would connect its addr[15:0] to AD[15:0] and data[15:0] to AD[31:16]. See Figure 17-6 for graphical connection. 17.5.1.2 Global Chip-Select Operation FBCS0, the global (boot) chip-select, allows address decoding for boot ROM before system initialization. Its operation differs from other external chip-select outputs after system reset. After system reset, FBCS0 is asserted for every external access. No other chip-select can be used until the valid bit, CSMR0[V], is set, at which point FBCS0 functions as configured. After this, FBCS[5:1] can be used as well. At reset, the port size, and automatic acknowledge functions of the global chip-select are determined by the logic levels on the AD[2:0] signals. Table 17-3, Table 17-4, and Table 17-5 list the various reset encodings for the configuration signals. Table 17-3. AD4/FB_CONFIG Selection of Non-Multiplexed 32-bit Address/32-bit Data Mode AD4 1 FlexBus Operating Mode 0 AD[31:0] used for data. PCIAD[31:0] used for address1 1 PCIAD[31:0] used for PCI bus. AD[31:0] used for both address and data. If the non-multiplexed 32-bit address/32-bit data mode is selected the PCI bus cannot be used. MCF548x Reference Manual, Rev. 5 17-6 Freescale Semiconductor Chip-Select Operation Table 17-4. AD[2]/AA Automatic Acknowledge of Boot FBCS0 AD[2]/AA Boot FBCS0 AA Configuration at Reset 0 Disabled 1 Enabled with 63 wait states Table 17-5. AD[1:0]/PS[1:0], Port Size of Boot FBCS0 17.5.2 AD[1:0]/PS[1:0] Boot FBCS0 Port Size at Reset 00 32-bit port 01 8-bit port 1x 16-bit port Chip-Select Registers The following tables describe in detail the registers and bit meanings for configuring chip-select operation. The chip-select controller register map is accessed relative to the memory base address register (MBAR). Table 17-6 shows the chip-select register memory map. Reading unused or reserved locations terminates normally and returns zeros. Table 17-6. Chip-Select Registers Register Offset [31:24] [23:16] [15:8] [7:0] ResetValue Access 1 0x0500 Chip-select address register—bank 0 (CSAR0) 0x0000_0000 R/W 0x0504 Chip-select mask register—bank 0 (CSMR0) 0x0000_0000 R/W 0x0508 Chip-select control register—bank 0 (CSCR0) BSTW = 0 BSTR = 0 PS = AD[1:0] AA = AD[2] WS = 111111 WRAH = 11 RDAH = 11 ASET = 11 SWSEN = 0 SWS = 000000 R/W 0x050C Chip-select address register—bank 1 (CSAR1) 0x0000_0000 R/W 0x0510 Chip-select mask register—bank 1 (CSMR1) 0x0000_0000 R/W 0x054 Chip-select control register—bank 1 (CSCR1) 0x0000_0000 R/W 0x0518 Chip-select address register—bank 2 (CSAR2) 0x0000_0000 R/W 0x051C Chip-select mask register—bank 2 (CSMR2) 0x0000_0000 R/W 0x0520 Chip-select control register—bank 2 (CSCR2) 0x0000_0000 R/W 0x0524 Chip-select address register—bank 3 (CSAR3) 0x0000_0000 R/W 0x0528 Chip-select mask register—bank 3 (CSMR3) 0x0000_0000 R/W MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 17-7 Table 17-6. Chip-Select Registers (Continued) Register Offset [31:24] [23:16] [15:8] [7:0] ResetValue Access 1 0x052C Chip-select control register—bank 3 (CSCR3) 0x0000_0000 R/W 0x0530 Chip-select address register—bank 4 (CSAR4) 0x0000_0000 R/W 0x0534 Chip-select mask register—bank 4 (CSMR4) 0x0000_0000 R/W 0x0538 Chip-select control register—bank 4 (CSCR4) 0x0000_0000 R/W 0x053C Chip-select address register—bank 5 (CSAR5) 0x0000_0000 R/W 0x0540 Chip-select mask register—bank 5 (CSMR5) 0x0000_0000 R/W 0x0544 Chip-select control register—bank 5 (CSCR5) 0x0000_0000 R/W 1 The access column indicates whether the corresponding register allows both read/write functionality (R/W), read-only functionality (R), or write-only functionality (W). A read access to a write-only register returns zeros. A write access to a read-only register has no effect. 2 Addresses not assigned to a register and undefined register bits are reserved for expansion. Write accesses to these reserved address spaces and reserved register bits have no effect. 17.5.2.1 Chip-Select Address Registers (CSAR0–CSAR5) CSARn, Figure 17-3, specify the chip-select base addresses. 31 30 29 28 27 26 25 24 R 23 22 21 20 19 18 17 16 BA W Reset 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 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 0 0 R W Reset Reg Addr MBAR + 0x500 (CSAR0); 0x50C (CSAR1); 0x518 (CSAR2); 0x524 (CSAR3); 0x530 (CSAR4); 0x53C (CSAR5) Figure 17-3. Chip-Select Address Registers (CSARn) Table 17-7. CSARn Field Descriptions Bits Name Description 31–16 BA Base address. Defines the base address for memory dedicated to chip-select FBCSn. BA is compared to bits 31–16 on the internal address bus to determine if chip-select memory is being accessed. 15–0 — Reserved, should be cleared MCF548x Reference Manual, Rev. 5 17-8 Freescale Semiconductor Chip-Select Operation 17.5.2.2 Chip-Select Mask Registers (CSMR0–CSMR5) CSMRn, Figure 17-4, are used to specify the address mask and allowable access types for the respective chip-selects. 31 30 29 28 27 26 25 24 R 23 22 21 20 19 18 17 16 BAM W Reset R 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 0 0 0 0 0 0 WP 0 0 0 0 0 0 0 V 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 W Reset Reg Addr MBAR + 0x504 (CSMR0); 0x510 (CSMR1); 0x51C (CSMR2); 0x528 (CSMR3); 0x534 (CSMR4); 0xr540 (CSMR5) Figure 17-4. Chip-Select Mask Registers (CSMRn) Table 17-8 describes CSMR fields. Table 17-8. CSMRn Field Descriptions Bits Name Description 31–16 BAM 15–9 — 8 WP 7–1 — Reserved, should be cleared 0 V Valid bit. Indicates whether the corresponding CSAR, CSMR, and CSCR contents are valid. Programmed chip-selects do not assert until V bit is set (except for FBCS0, which acts as the global chip-select). Reset clears each CSMRn[V]. At reset, no chip-select other than FBCS0 can be used until the CSMR0[V] is set. At which point FBCS[5:0] functions as configured. 0 chip-select invalid 1 chip-select valid Base address mask. Defines the chip-select block size by masking address bits. Setting a BAM bit causes the corresponding CSAR bit to be a “don’t care” in the decode. 0 Corresponding address bit is used in chip-select decode. 1 Corresponding address bit is a don’t care in chip-select decode. The block size for FBCSn is 2n; n = (number of bits set in respective CSMR[BAM]) + 16. For example, if CSAR0 = 0x0000 and CSMR0[BAM] = 0x0008, FBCS0 would address two discontinuous 64-Kbyte memory blocks: one from 0x0000–0xFFFF and one from 0x8_0000–0x8_FFFF. Likewise, for FBCS0 to access 32 Mbytes of address space starting at location 0x0, FBCS1 must begin at the next byte after FBCS0 for a 16-Mbyte address space. Then CSAR0 = 0x0000, CSMR0[BAM] = 0x01FF, CSAR1 = 0x0200, and CSMR1[BAM] = 0x00FF. Reserved, should be cleared Write protect. Controls write accesses to the address range in the corresponding CSAR. Attempting to write to the range of addresses for which CSARn[WP] = 1 results in the appropriate chip-select not being selected. No exception occurs. 0 Both read and write accesses are allowed 1 Only read accesses are allowed MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 17-9 17.5.2.3 Chip-Select Control Registers (CSCR0–CSCR5) Each CSCRn, Figure 17-5, controls the auto acknowledge, address setup and hold times, port size, burst capability, and activation of each chip-select. Note that to support the global chip-select, FBCS0, the CSCR0 reset values differ from the other CSCRs. FBCS0 allows address decoding for boot ROM before system initialization. 31 30 29 R 28 27 26 SWS 25 24 23 22 0 0 SWS EN — W 21 20 19 ASET 18 RDAH 17 16 WRAH Reset: CSCR0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 Reset: CSCRs 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 AA PS BEM BSTR BSTW 0 0 0 AD[1:0] AD3 0 0 0 0 0 0 0 0 0 0 0 R WS W Reset: CSCR0 1 1 1 1 1 1 0 AD2 Reset: CSCRs 0 0 0 0 0 0 0 0 Reg Addr 0 0 MBAR + 0x508 (CSCR0); 0x514 (CSCR1); 0x520 (CSCR2); 0x52C (CSCR3); 0x538 (CSCR4); 0x544 (CSCR5) Figure 17-5. Chip-Select Control Registers (CSCRn) Table 17-9 describes CSCRn fields. Table 17-9. CSCRn Field Descriptions Bits Name Description 31–26 SWS Secondary wait states. The number of wait states inserted before an internal transfer acknowledge is generated for burst transfer except for the first termination, which is controlled by the wait state count. The secondary wait state is only used if the secondary wait state enable is set, otherwise the wait state value is used for all burst transfers. 25–24 — 23 SWSEN 22 — 21–20 ASET Reserved, should be cleared Secondary wait state enable. If set (SWSEN = 1), then the secondary wait state value is used to insert wait states before an internal transfer acknowledge is generated for burst transfer secondary terminations. If cleared (SWSEN = 0), then the wait state value is used to insert wait states before an internal transfer acknowledge is generated for all transfers. Reserved, should be cleared Address setup. This field controls the asserting of chip-select with respect to assertion of a valid address and attributes. Note that the address and attributes are considered valid at the same time ALE asserts. 00 Assert chip-select on rising clock edge after address is asserted. (Default FBCSn) 01 Assert chip-select on second rising clock edge after address is asserted. 10 Assert chip-select on third rising clock edge after address is asserted. 11 Assert chip-select on fourth rising clock edge after address is asserted.(Reset FBCS0) MCF548x Reference Manual, Rev. 5 17-10 Freescale Semiconductor Chip-Select Operation Table 17-9. CSCRn Field Descriptions (Continued) Bits Name Description 19–18 RDAH Read Address Hold or (Deselect). This field controls the address and attribute hold time after the termination during a read cycle that hits in the chip-select address space. The hold time only applies at the end of a transfer. Therefore, a burst transfer only has a hold time added after the last bus cycle. RDAH = 00; Hold address and attributes one cycle after FBCSn negates on reads. (Default FBCSn) 01 Hold address and attributes two cycles after FBCSn negates on reads. 10 Hold address and attributes three cycles after FBCSn negates on reads. 11 Hold address and attributes four cycles after FBCSn negates on reads. (Reset FBCS0) 17–16 WRAH Write Address Hold or (Deselect). This field controls the address, data and attribute hold time after the termination of a write cycle that hits in the chip-select address space.The hold time only applies at the end of a transfer. Therefore, a burst transfer only has a hold time added after the last bus cycle. WRAH = 00; Hold address and attributes one cycle after FBCSn negates on writes. (Default FBCSn) 01 Hold address and attributes two cycles after FBCSn negates on writes. 10 Hold address and attributes three cycles after FBCSn negates on writes. 11 Hold address and attributes four cycles after FBCSn negates on writes. (Reset FBCS0) 15–10 WS Wait states. The number of wait states inserted after FBCSn asserts and before an internal transfer acknowledge is generated (WS = 0 inserts zero wait states, WS = 0x3F inserts 63 wait states). If AA = 0, TA must be asserted by the external system regardless of the number of wait states generated. In that case, the external transfer acknowledge ends the cycle. An external TA supersedes the generation of an internal TA. 9 — Reserved, should be cleared. 8 AA Auto-acknowledge enable. Determines the assertion of the internal transfer acknowledge for accesses specified by the chip-select address. 0 No internal TA is asserted. Cycle is terminated externally. 1 Internal TA is asserted as specified by WS. Note that if AA = 1 for a corresponding FBCSn and the external system asserts an external TA before the wait-state countdown asserts the internal TA, the cycle is terminated. Burst cycles increment the address bus between each internal termination. 7–6 PS Port size. Specifies the width of the data port associated with each chip-select. It determines where data is driven during write cycles and where data is sampled during read cycles. 00 32-bit port size. Valid data sampled and driven on D[31:0] 01 8-bit port size. Valid data sampled and driven on D[31:24] 1x 16-bit port size. Valid data sampled and driven on D[31:16] 5 BEM Byte enable mode. Specifies the byte enable operation. Certain SRAMs have byte enables that must be asserted during reads as well as writes. BEM can be set in the relevant CSCR to provide the appropriate mode of byte enable support in support of these SRAMs. 0 Neither BE or BWE is asserted for reads. BWE is generated for data write only. 1 BE is asserted for reads; BWE is asserted for writes. 4 BSTR Burst read enable. Specifies whether burst reads are used for memory associated with each FBCSn. 0 Data exceeding the specified port size is broken into individual, port-sized non-burst reads. For example, a longword read from an 8-bit port is broken into four 8-bit reads. 1 Enables data burst reads larger than the specified port size, including longword reads from 8and 16-bit ports and word reads from 8-bit ports. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 17-11 Table 17-9. CSCRn Field Descriptions (Continued) Bits Name Description 3 BSTW Burst write enable. Specifies whether burst writes are used for memory associated with each FBCSn. 0 Break data larger than the specified port size into individual port-sized, non-burst writes. For example, a longword write to an 8-bit port takes four byte writes. 1 Enables burst write of data larger than the specified port size, including longword writes to 8 and 16-bit ports and word writes to 8-bit ports. 2–0 — 17.6 17.6.1 Reserved, should be cleared. Functional Description Data Transfer Operation Data transfers between the MCF548x and other devices involve the following signals: • Address/data bus (AD[31:0]) • Control signals (ALE and TA) • FBCSn • OE • BE/BWE[3:0] • Attribute signals (R/W, TBST, TSIZ[1:0]) The address and write data (AD[31:0]), R/W, ALE, FBCSn, and all attribute signals change on the rising edge of the clock. Read data is registered in the MCF548x on the rising edge of the clock. The MCF548x FlexBus supports byte, word, and longword operand transfers and allows accesses to 8-, 16-, and 32-bit data ports.Transfer parameters such as address setup and hold, port size, the number of wait states for the external device being accessed, automatic internal transfer termination enable or disable, and burst enable or disable are programmed in the chip-select control registers (CSCRs), Section 17.5.2.3, “Chip-Select Control Registers (CSCR0–CSCR5).” 17.6.2 Data Byte Alignment and Physical Connections The MCF548x aligns data transfers in FlexBus byte lanes, the number of lanes depending on the width of the data port. Figure 17-6 shows the byte lanes that external memory should be connected to and the sequential transfers if a longword is transferred for three port sizes. For example, an 8-bit memory should be connected to the single lane AD[31:24]. A longword transfer through this 8-bit port takes four transfers on AD[31:24], starting with the MSB and going to the LSB. A longword transfer through a 32-bit port requires one transfer on each of the four byte lanes of the FlexBus. MCF548x Reference Manual, Rev. 5 17-12 Freescale Semiconductor Functional Description Byte Select BE/BWE0 BE/BWE1 BE/BWE2 BE/BWE3 Processor External Data Bus AD[31:24] AD[23:16] AD[15:8] AD[7:0] 32-Bit Port Memory Byte 0 Byte 1 16-Bit Port Memory Byte 0 Byte 1 Byte 2 Byte 3 8-Bit Port Memory Byte 2 Byte 3 Driven with address values Byte 0 Byte 1 Driven with address values Byte 2 Byte 3 Figure 17-6. Connections for External Memory Port Sizes 17.6.3 Address/Data Bus Multiplexing The MCF548x FlexBus uses a 32-bit wide multiplexed address and data bus (AD[31:0]). The full 32-bit address will always be driven on the first clock of a bus cycle. During the data phase, which AD[31:0] lines are used for data is determined by the programmed port size for the corresponding chip select. The MCF548x continues to drive the address on any AD[31:0] lines that are not used for data. Table 17-10 lists the supported combinations of address and data bus widths. Table 17-10. FlexBus Operating Modes Port Size Address Signals During Address Phase Data Signals During Data Phase Address Signals During Data Phase 32-bit1 AD[31:0] AD[31:0] -- 16-bit AD[31:0] AD[31:16] AD[15:0] 8-bit AD[31:0] AD[31:24] AD[23:0] 1 17.6.4 The 32-bit Address/32-bit Data non-multiplexed mode uses the PCI address/data bus to provide a second 32-bit bus for the address. PCI cannot be used if this mode is selected. Bus Cycle Execution As shown in Figure 17-9 and Figure 17-11, basic bus operations occur in four clocks, as follows: 1. At the first clock edge, the address, attributes, and ALE are driven. 2. FBCSn is asserted at the second rising clock edge to indicate which device has been selected and by that time the address and attributes are valid and stable. ALE is negated at this edge. For a write transfer, data is driven on the bus at this clock edge and continues to be driven until one clock cycle after FBCSn negates. For a read transfer, data is also returned at this cycle. External slave asserts TA at this clock edge. 3. Read data and TA are sampled on the third clock edge. TA can be negated after this edge and read data can then be tristated. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 17-13 4. FBCSn is negated at the fourth rising clock edge. This last clock of the bus cycle uses what would be an idle clock between cycles to provide hold time for address, attributes, and write data. 17.6.4.1 Data Transfer Cycle States The data transfer operation in the MCF548x is controlled by an on-chip state machine. The state transition diagram for basic read and write cycles is shown in Figure 17-7. Next Cycle S0 Wait States S3 S1 S2 Figure 17-7. Data Transfer State Transition Diagram Table 17-11 describes the states as they appear in subsequent timing diagrams. Table 17-11. Bus Cycle States State Cycle Description S0 All The read or write cycle is initiated. On the rising clock edge, the MCF548x places a valid address on AD[31:0], asserts ALE, and drives R/W high for a read and low for a write, if these signals are not already in the appropriate state. S1 All ALE is negated on the rising edge of CLK, and FBCSn is asserted. Data is driven on AD[31:Y] for writes, and AD[31:Y] is three-stated for reads. Address continues to be driven on AD[X:0] pins that are unused for data. If TA is recognized asserted, then the cycle moves on to S2. If TA is not asserted either internally or externally, then the S1 state continues to repeat. S2 Read Data is made available by the external device before the rising edge of CLK with TA asserted. The the MCF548x will latch data on this rising clock edge. All For internal termination, both the FBCSn and internal TA will be negated. For external termination, the external device should negate TA, and FBCSn select is negated after the rising edge of CLK at the end of S2. Read S3 All The external device can stop driving data after the rising edge of CLK at the beginning of S2. However, data can be driven until the end of S3 or any additional address hold cycles. Address, data, and R/W go invalid off the rising edge of CLK at the end of S3, terminating the read or write cycle. MCF548x Reference Manual, Rev. 5 17-14 Freescale Semiconductor Functional Description 17.6.5 17.6.5.1 FlexBus Timing Examples Basic Read Bus Cycle During a read cycle, the MCF548x receives data from memory or from a peripheral device. Figure 17-8 is a read cycle flowchart. NOTE Throughout this chapter AD[X:0] is used to indicate an address bus that can be 32-, 24-, or 16-bits in width. AD[31:Y] is a data bus that can be 32-, 16-, or 8-bits wide. MCF548X System 1. Set R/W to read. 2. Place address on AD[31:0]. 3. Assert ALE. 1. Decode address. 1. Negate ALE. 2. Assert FBCSn. 1. CS unit asserts internal TA (auto acknowledge/internal termination). 2. Sample TA low and latch data. 1. Select the appropriate slave device. 2. Drive data on AD[31:Y]. 3. Assert TA (external termination). 1. Start next cycle. 1. Negate TA (external termination). Figure 17-8. Read Cycle Flowchart The read cycle timing diagram is shown in Figure 17-9. NOTE In the following timing diagrams, the dotted lines indicate TA, OE, and FBCSn timing when internal termination is used (CSCR[AA] = 1). The external and internal TA assert at the same time; however, TA is not driven externally for internally terminated bus cycles. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 17-15 S0 S1 S2 S3 CLK ADDR[X:0] AD[X:0] A[31:Y] AD[31:Y] DATA R/W ALE TSIZ[1:0] TSIZ[1:0] FBCSn, BE/BWEn OE TA Figure 17-9. Basic Read Bus Cycle 17.6.5.2 Basic Write Bus Cycle During a write cycle, the MCF548x sends data to memory or to a peripheral device. The write cycle flowchart is shown in Figure 17-10. NOTE Throughout this chapter AD[X:0] is used to indicate an address bus that can be 32-, 24-, or 16-bits in width. AD[31:Y] is a data bus that can be 32-, 16-, or 8-bits wide. MCF548X System 1. Set R/W to write. 2. Place address on AD[31:0]. 3. Assert ALE. 1. Decode address. 1. Negate ALE. 2. Assert FBCSn. 1. CS unit asserts internal TA (auto acknowledge/internal termination). 2. Sample TA low. 1. Select the appropriate slave device. 2. Drive data on AD[31:Y]. 3. Assert TA (external termination). 1. Start next cycle. 1. Negate TA (external termination). Figure 17-10. Write Cycle Flowchart MCF548x Reference Manual, Rev. 5 17-16 Freescale Semiconductor Functional Description The write cycle timing diagram is shown in Figure 17-11. S0 S1 S2 S3 CLK ADDR[X:0] AD[X:0] AD[31:Y] A[31:Y] DATA R/W ALE TSIZ[1:0] TSIZ[1:0] FBCSn, BE/BWEn OE TA Figure 17-11. Basic Write Bus Cycle 17.6.5.3 Bus Cycle Multiplexing This section shows timing diagrams for various port size scenarios. Figure 17-12 illustrates the basic word read transfer to a 16-bit device with no wait states. The address is driven on the full AD[31:0] bus in the first clock. The MCF548x tristates AD[31:16] on the second clock and continues to drive address on AD[15:0] throughout the bus cycle. The external device returns the read data on AD[31:16] and may tristate the data line or continue to drive the data one clock after TA is sampled asserted. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 17-17 S0 S1 S2 S3 CLK AD[31:24] A[31:24] D[15:8] AD[23:16] A[23:16] D[7:0] AD[15:8] ADDR[15:8] AD[7:0] ADDR[7:0] R/W ALE 10 TSIZ[1:0] FBCSn, BE/BWEn OE TA Figure 17-12. Single Word Read Transfer with Muxed 32-A / 16-D or Non-Muxed 16-A / 16-D Figure 17-13 shows the similar configuration for a write transfer. The data is driven from the second clock on AD[31:16]. S0 S1 S2 S3 CLK AD[31:24] A[31:24] DATA[15:8] AD[23:16] A[23:16] DATA[7:0] AD[15:8] ADDR[15:8] AD[7:0] ADDR[7:0] R/W ALE TSIZ[1:0] 10 FBCSn, BE/BWEn OE TA Figure 17-13. Single Word Write Transfer with Muxed 32-A / 16-D or Non-Muxed 16-A / 16-D MCF548x Reference Manual, Rev. 5 17-18 Freescale Semiconductor Functional Description Figure 17-14 illustrates the basic byte read transfer to an 8-bit device with no wait states. The address is driven on the full AD[31:0] bus in the first clock. The MCF548x tristates AD[31:24] on the second clock and continues to drive address on AD[23:0] throughout the bus cycle. The external device returns the read data on AD[31:24], and may tristate the data line or continue to drive the data one clock after TA is sampled asserted. S0 S1 S2 S3 CLK AD[31:24] A[31:24] D[7:0] AD[23:16] ADDR[23:16] AD[15:8] ADDR[15:8] AD[7:0] ADDR[7:0] R/W ALE TSIZ[1:0] 01 FBCSn, BE/BWEn OE TA Figure 17-14. Single Byte Read Transfer with Muxed 32-A / 8-D or Non-Muxed 24-A / 8-D Figure 17-15 shows the similar configuration for a write transfer. The data is driven from the second clock on AD[31:24]. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 17-19 S0 S1 S2 S3 CLK AD[31:24] A[31:24] DATA[7:0] AD[23:16] ADDR[23:16] AD[15:8] ADDR[15:8] AD[7:0] ADDR[7:0] R/W ALE 01 TSIZ[1:0] FBCSn, BE/BWEn OE TA Figure 17-15. Single Byte Write Transfer with Muxed 32-A / 8-D or Non-Muxed 24-A / 8-D Figure 17-16 depicts a longword read through a 32-bit device. Notice that when the device port size is 32 bits, the only mode the bus supports is multiplexing address and data lines. S0 S1 S2 S3 CLK AD[31:24] A[31:24] D[31:24] AD[23:16] A[23:16] D[23:16] AD[15:8] A[15:8] D[15:8] AD[7:0] A[7:0] D[7:0] R/W ALE TSIZ[1:0] 00 FBCSn, BE/BWEn OE TA Figure 17-16. Longword Read Transfer with Muxed 32-A / 32-D MCF548x Reference Manual, Rev. 5 17-20 Freescale Semiconductor Functional Description Figure 17-17 illustrates the longword write to a 32-bit device. S0 S1 S2 S3 CLK AD[31:24] A[31:24] DATA[31:24] AD[23:16] A[23:16] DATA[23:16] AD[15:8] A[15:8] DATA[15:8] AD[7:0] A[7:0] DATA[7:0] R/W ALE TSIZ[1:0] 00 FBCSn, BE/BWEn OE TA Figure 17-17. Longword Write Transfer with Muxed 32-A / 32-D 17.6.5.4 Timing Variations The MCF548x has several features that can be used to change the timing characteristics of a basic read or write bus cycle to provide additional address setup, address hold, and time for a device to provide or latch data. 17.6.5.4.1 Wait States Wait states can be inserted before each beat of a transfer by programming the CSCRn registers. Wait states can be used to give the peripheral or memory more time to return read data or sample write data. Figure 17-18 and Figure 17-19 show the basic read and write bus cycles (also shown in Figure 17-9 and Figure 17-11). This is the default case with no wait states. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 17-21 S0 S1 S2 S3 CLK ADDR[X:0] AD[X:0] AD[31:Y] A[31:Y] DATA R/W ALE TSIZ[1:0] TSIZ[1:0] FBCSn, BE/BWEn OE TA Figure 17-18. Basic Read Bus Cycle (No Wait States) S0 S1 S2 S3 CLK ADDR[X:0] AD[X:0] AD[31:Y] A[31:Y] DATA R/W ALE TSIZ[1:0] TSIZ[1:0] FBCSn, BE/BWEn OE TA Figure 17-19. Basic Write Bus Cycle (No Wait States) If wait states are used, then the S1 state will repeat continuously until either the internal TA is asserted by the chip select auto-acknowledge unit or the external TA is recognized as asserted. Figure 17-20 and Figure 17-21 show a read and write cycle with one wait state. MCF548x Reference Manual, Rev. 5 17-22 Freescale Semiconductor Functional Description S0 S1 WS S2 S3 CLK ADDR[X:0] AD[X:0] AD[31:Y] A[31:Y] DATA R/W ALE TSIZ[1:0] TSIZ[1:0] FBCSn, BE/BWEn OE TA Figure 17-20. Read Bus Cycle (One Wait State) S0 S1 WS S2 S3 CLK ADDR[X:0] AD[X:0] AD[31:Y] A[31:Y] DATA R/W ALE TSIZ[1:0] TSIZ[1:0] FBCSn, BE/BWEn OE TA Figure 17-21. Write Bus Cycle (One Wait State) 17.6.5.4.2 Address Setup and Hold The timing of the assertion and negation of the chip selects, byte selects, and output enable can be programmed on a chip select basis. Each chip select can be programmed to assert one to four clocks after address latch enable (ALE) is asserted. Figure 17-22 and Figure 17-23 show read and write bus cycles with two clocks of address setup. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 17-23 S0 AS S1 S2 S3 CLK ADDR[X:0] AD[X:0] A[31:Y] AD[31:Y] DATA R/W ALE TSIZ[1:0] TSIZ[1:0] FBCSn, BE/BWEn OE TA Figure 17-22. Read Bus Cycle with Two Clock Address Setup (No Wait States) S0 AS S1 S2 S3 CLK ADDR[X:0] AD[X:0] AD[31:Y] A[31:Y] DATA R/W ALE TSIZ[1:0] TSIZ[1:0] FBCSn, BE/BWEn OE TA Figure 17-23. Write Bus Cycle with Two Clock Address Setup (No Wait States) In addition to address setup, there is also a programmable address hold option for each chip select. Address and attributes can be held one to four clocks after chip select, byte selects, and output enable negate. Figure 17-24 and Figure 17-25 show read and write bus cycles with two clocks of address hold. MCF548x Reference Manual, Rev. 5 17-24 Freescale Semiconductor Functional Description S0 S1 S2 S3 AH CLK ADDR[X:0] AD[X:0] AD[31:Y] A[31:Y] DATA R/W ALE TSIZ[1:0] TSIZ[1:0] FBCSn, BE/BWEn OE TA Figure 17-24. Read Cycle with Two Clock Address Hold (No Wait States) S0 S1 S2 S3 AH CLK ADDR[X:0] AD[X:0] AD[31:Y] A[31:Y] DATA R/W ALE TSIZ[1:0] TSIZ[1:0] FBCSn, BE/BWEn OE TA Figure 17-25. Write Cycle with Two Clock Address Hold (No Wait States) Figure 17-26 shows a bus cycle that uses address setup, wait states, and address hold. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 17-25 S0 AS S1 WS S2 S3 AH CLK ADDR[X:0] AD[X:0] A[31:Y] AD[31:Y] DATA R/W ALE TSIZ[1:0] TSIZ[1:0] FBCSn, BE/BWEn OE TA Figure 17-26. Write Cycle with Two Clock Address Setup and Two Clock Hold (One Wait State) 17.6.6 Burst Cycles The MCF548x can be programmed to initiate burst cycles if its transfer size exceeds the size of the port it is transferring to. The initiation of a burst cycle is encoded on the size pins. For burst transfers to smaller port sizes, TSIZ[1:0] indicate the size of the entire transfer. For example, with bursting enabled, a word transfer to an 8-bit port would take a 2-byte burst cycle, for which TSIZ[1:0] = 10 throughout. A longword transfer to an 8-bit port would take a 4-byte burst cycle, for which TSIZ[1:0] = 00 throughout. With bursting disabled, any transfer is larger than port size is broken into multiple individual transfers. With bursting enabled, an access is larger than port size would result a burst cycle of multiple beats. Table 17-12 shows the result of such transfer translations. Table 17-12. Transfer Size and Port Size Translation Port Size PS[1:0] Transfer Size TSIZ[1:0] Burst-inhibited: number of transfers Burst enabled: number of beats 01 (8-bit) 10 (word) 2 00 (longword) 4 11 (line) 16 00 (longword) 2 11 (line) 8 11 (line) 4 1- (16-bit) 00 (32-bit) The MCF548x bus can support 2-1-1-1 burst cycles and optimize DMA transfers. A user can add wait states by delaying termination of the cycle. If internal termination is used, different wait state counters can be used for the first access and the following beats. MCF548x Reference Manual, Rev. 5 17-26 Freescale Semiconductor Functional Description NOTE Line-sized transfers requested by the core or cache are broken up into four individual longword transfers, but the DMA can request line-sized transfers when the read line or combine write flags are set. See Section 24.4.9, “Line Buffers,” for more information. CSCRs are used to enable bursting for reads, writes, or both. Memory spaces can be declared burst-inhibited for reads and writes by clearing the appropriate CSCRn[BSTR,BSTW]. Figure 17-27 shows a longword read through an 8-bit device programmed for burst enable. The transfer results in a 4-beat burst and the data is driven on AD[31:24]. Notice that the transfer size is driven at longword (2’b00) throughout the bus cycle. S0 S1 S2 S2 S2 S2 S3 CLK ADDR[23:0] AD[23:0] AD[31:24] A[31:24] DATA DATA DATA DATA R/W ALE TSIZ[1:0] 00 FBCSn, BE/BWEn TBST OE TA Figure 17-27. Longword Read Burst from 8-Bit Port 2-1-1-1 (No Wait States) Figure 17-28 shows a longword write through an 8-bit device programmed for burst enable. The transfer results in a 4-beat burst and the data is driven on AD[31:24]. Notice that the transfer size is driven at longword (2’b00) throughout the bus cycle. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 17-27 S0 S1 S2 S2 S2 S2 S3 CLK ADDR[23:0] AD[23:0] AD[31:24] A[31:24] DATA DATA DATA DATA R/W ALE 00 TSIZ[1:0] FBCSn, BE/BWEn TBST OE TA Figure 17-28. Longword Write Burst to 8-Bit Port 2-1-1-1 (No Wait States) Figure 17-29 shows a longword read through an 8-bit device with burst inhibited. The transfer results in four individual transfers. Notice that the transfer size is driven at longword (2’b00) during the first transfer and at byte (2’b01) during the next three transfers. S0 S2 S1 S0 S1 S2 S0 S1 S2 S0 S2 S1 S3 CLK ADDR[23:0] AD[23:0] AD[31:24] A[31:24] DATA A[31:24] DATA A[31:24] DATA A[31:24] DATA DATA R/W ALE TSIZ[1:0] 00 01 FBCSn, BE/BWEn TBST OE TA Figure 17-29. Longword Read Burst-Inhibited from 8-Bit Port (No Wait States) MCF548x Reference Manual, Rev. 5 17-28 Freescale Semiconductor Functional Description Figure 17-30 shows a longword write through an 8-bit device with burst inhibited. The transfer results in four individual transfers. Notice that the transfer size is driven at longword (2’b00) during the first transfer and at byte (2’b01) during the next three transfers. S1 S0 S2 S0 S1 S0 S2 S1 S2 S0 S2 S1 S3 CLK ADDR[23:0] AD[23:0] AD[31:24] A[31:24] DATA A[31:24] DATA A[31:24] DATA A[31:24] DATA R/W ALE 00 TSIZ[1:0] 01 FBCSn, BE/BWEn TBST OE TA Figure 17-30. Longword Write Burst-Inhibited to 8-Bit Port (No Wait States) Figure 17-31 illustrates another read burst transfer, but in this case a wait state is added between individual beats. S0 S1 WS S2 WS/SWS S2 WS/SWS S2 WS/SWS S2 S3 CLK ADDR[23:0] AD[23:0] AD[31:24] A[31:24] DATA DATA DATA DATA R/W ALE TSIZ[1:0] 00 FBCSn, BE/BWEn TBST OE TA Figure 17-31. Longword Read Burst from 8-Bit Port 3-2-2-2 (One Wait State) MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 17-29 Figure 17-31 illustrates a write burst transfer with one wait state. S0 S1 WS S2 WS/SWS S2 WS/SWS S2 WS/SWS S2 S3 CLK ADDR[23:0] AD[23:0] A[31:24] AD[31:24] DATA DATA DATA DATA R/W ALE 00 TSIZ[1:0] FBCSn, BE/BWEn TBST OE TA Figure 17-32. Longword Write Burst to 8-Bit Port 3-2-2-2 (One Wait State) If address setup and hold are used, only the first and last beat of the burst cycle will be affected as shown in Figure 17-33. S0 AS S1 S2 S2 S2 S2 S3 AH CLK ADDR[23:0] AD[23:0] AD[31:24] A[31:24] DATA DATA DATA DATA R/W ALE TSIZ[1:0] 11 FBCSn, BE/BWEn OE TBST TA Figure 17-33. Longword Read Burst from 8-Bit Port 3-1-1-1 (Address Setup and Hold) Figure 17-34 shows a write cycle with one clock of address setup and address hold. MCF548x Reference Manual, Rev. 5 17-30 Freescale Semiconductor Functional Description S0 AS S1 S2 S2 S2 S2 S3 AH CLK AD[23:0] AD[31:24] ADDR[23:0] A[31:24] DATA DATA DATA DATA R/W ALE 11 TSIZ[1:0] FBCSn, BE/BWEn OE TBST TA Figure 17-34. Longword Write Burst to 8-Bit Port 3-1-1-1 (Address Setup and Hold) 17.6.7 Misaligned Operands Because operands, unlike opcodes, can reside at any byte boundary, they are allowed to be misaligned. A byte operand is properly aligned at any address, a word operand is misaligned at an odd address, and a longword is misaligned at an address not a multiple of four. Although the MCF548x enforces no alignment restrictions for data operands (including program counter (PC) relative data addressing), additional bus cycles are required for misaligned operands. Instruction words and extension words (opcodes) must reside on word boundaries. Attempting to prefetch a misaligned instruction word causes an address error exception. The MCF548x converts misaligned, cache-inhibited operand accesses to multiple aligned accesses. Figure 17-35 shows the transfer of a longword operand from a byte address to a 32-bit port. First a byte is transferred at an offset of 0x1. The slave device supplies the byte and acknowledges the data transfer. When the MCF548x starts the second cycle, a word is transferred with a byte offset of 0x2. The next two bytes are transferred in this cycle. In the third cycle, byte 3 is transferred. The byte offset is now 0x0, the port supplies the final byte, and the operation is complete. 31 24 23 16 15 8 7 0 A[2:0] Transfer 1 –– Byte 0 –– –– 001 Transfer 2 –– –– Byte 1 Byte 2 010 Transfer 3 Byte 3 –– –– –– 100 Figure 17-35. Example of a Misaligned Longword Transfer (32-Bit Port) If an operand is cacheable and is misaligned across a cache-line boundary, both lines are loaded into the cache. The example in Figure 17-36 differs from the one in Figure 17-35 because the operand is word-sized and the transfer takes only two bus cycles. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 17-31 31 24 23 16 15 8 7 0 A[2:0] Transfer 1 –– –– –– Byte 0 001 Transfer 2 Byte 0 –– –– — 100 Figure 17-36. Example of a Misaligned Word Transfer (32-Bit Port) 17.6.8 Bus Errors The MCF548x has no bus monitor. If the auto-acknowledge feature is not enabled for the address that generates the error, the bus cycle can be terminated by asserting TA or by using the software watchdog timer. If it is required that the MCF548x handle a bus error differently, an interrupt handler can be invoked by asserting an interrupt to the core along with TA when the bus error occurs. MCF548x Reference Manual, Rev. 5 17-32 Freescale Semiconductor Chapter 18 SDRAM Controller (SDRAMC) 18.1 Introduction This chapter describes configuration and operation of the synchronous DRAM (SDRAM) controller. It begins with a general overview and includes a description of signals involved in SDRAM operations. The remainder of the chapter describes the programming model and signal timing, as well as the command set required for synchronous DRAM operations. It also includes examples that the designer can follow to better understand how to configure the SDRAM controller for synchronous operations. 18.2 18.2.1 Overview Features The MCF548x SDRAM controller contains the following features: • Supports a glueless interface to SDR and DDR SDRAMs • 32-bit fixed memory port width • 64-bit data bus interface to internal XLB 64-bit bus • 32 bytes critical word first burst transfer • Up to 13 row address lines, up to 12 column address lines, 2 bits of bank address, and a maximum of four chip selects. The maximum row bits plus column bits can be less than or equal to 24. • Supports up to 1 Gbyte of memory—13+11 or 12+12 bit RA+CA, 2 bit BA, four chip selects • Minimum memory configuration of 8 Mbyte—11 bit row address (RA), 8 bit column address (CA), 2 bit bank address (BA) and one chip select • Supports page mode to maximize the data rate • Supports sleep mode and self-refresh mode • Error detect and parity check are not supported 18.2.2 Terminology The following terminology is used in this chapter: • SDRAM block: Any group of DRAM memories selected by one of the MCF548x SDCS[3:0] signals. Thus, the MCF548x can support up to four independent memory blocks. The base address of each block is programmed in the DRAM address and control registers (DACR0 and DACR1). • SDRAM bank: An internal partition in an SDRAM device. For example, a 64-Mbit SDRAM component might be configured as four 512K x 32 banks. Banks are selected through the SD_BA[1:0] signals. • SDRAM: These are RAMs that operate like asynchronous DRAMs but with a synchronous clock, a pipelined, multiple-bank architecture, and a faster speed. • Single data rate (SDR) SDRAM: This is SDRAM that drives/latches data and command information on the rising edge of the clock. • Double data rate (DDR) SDRAM: This is SDRAM that latches command information on the rising edge of the clock, but data is driven/latched on both the rising and falling edges of the clock rather than on just the rising edge. This doubles data throughput rate without an increase in frequency. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 18-1 18.2.3 Block Diagram Column addr[29:4] Address Input MUX Bank Row Column Address Pipeline Latches Bank Row Address Output MUX SDADDR[12:0] SDBA[1:0] Select SDCS[3:0] RAS addr[1:3] SDRAM Controller State Machine CAS SDWE SDDQS SDCLK[1:0] SDCLK[1:0] SDCKE SDDM tsiz[1:0], tbst datain[63:0] dataout[63:0] Write Data Buffer SDDATA[31:0] Read Data Buffer SDDATA[31:0] Figure 18-1. SDRAM Controller Block Diagram 18.3 18.3.1 External Signal Description SDRAM Data Bus (SDDATA[31:0]) SDDATA[31:0] is the bidirectional, non-multiplexed data bus used for SDRAM accesses. Data is sampled by the MCF548x on the rising edge of SDCLK when in SDR mode, and on both the rising and falling edge of SDCLK when in DDR mode. 18.3.2 SDRAM Address Bus (SDADDR[12:0]) The SDADDR[12:0] signals are the 13-bit, uni-directional address bus used for multiplexed row and column addresses during SDRAM bus cycles. The address multiplexing supports up to 256 Mbytes of SDRAM per chip select. 18.3.3 SDRAM Bank Addresses (SDBA[1:0]) Each SDRAM module has four internal row banks. The SDBA[1:0] signals are used to select the row bank. It is also used to select the SDRAM internal mode register during power-up initialization. MCF548x Reference Manual, Rev. 5 18-2 Freescale Semiconductor External Signal Description 18.3.4 SDRAM Row Address Strobe (RAS) This output is the SDRAM synchronous row address strobe. 18.3.5 SDRAM Column Address Strobe (CAS) This output is the SDRAM synchronous column address strobe. 18.3.6 SDRAM Chip Selects (SDCS[3:0]) These signals interface to the chip select lines of the SDRAMs within a memory block. Thus, there is one SDCS line for each memory block (the MCF548x supports up to four SDRAM memory blocks). 18.3.7 SDRAM Write Data Byte Mask (SDDM[3:0]) These output signals are sampled by the SDRAM on both edges of SDDQS to determine which byte lanes of the SDRAM data bus should be latched during a write cycle. In DDR mode, these bits are ignored during read operations. 18.3.8 SDRAM Data Strobe (SDDQS[3:0]) These bidirectional signals indicate when valid data is on the SDRAM data bus. Table 18-1 shows the correspondence between SDDATA byte lanes and the SDDQS and SDDM signals. Table 18-1. SDDQS and SDDM to Byte Lane Mapping 18.3.9 Byte Lane SDDQS SDDM SDDATA[31:24] (MSB) SDDQS3 SDDM3 SDDATA[23:16] SDDQS2 SDDM2 SDDATA[15:8] SDDQS1 SDDM1 SDDATA[7:0] (LSB) SDDQS0 SDDM0 SDRAM Clock (SDCLK[1:0]) This is the output clock for SDRAM accesses. 18.3.10 Inverted SDRAM Clock (SDCLK[1:0]) This is the inverted version of the SDRAM clock. It is used with SDCLK to provide the differential clocks for DDR SDRAM. 18.3.11 SDRAM Write Enable (SDWE) The SDRAM write enable (SDWE) is asserted to signify that a DRAM write cycle is underway. A read cycle is indicated by the negation of SDWE. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 18-3 18.3.12 SDRAM Clock Enable (SDCKE) This output is the SDRAM clock enable. SDCKE negates to put the SDRAM into low-power, self-refresh mode. 18.3.13 SDR SDRAM Data Strobe (SDRDQS) This is connected to SDDQS inputs. It is used in SDR mode only. 18.3.14 SDRAM Memory Supply (SDVDD) These pins supply positive power to the SDRAM module. SDVDD should be connected to +2.5V for DDR operation and +3.3V for SDR. 18.3.15 SDRAM Reference Voltage (VREF) This is the input reference voltage for differential SSTL_2 inputs. It is used in both DDR and SDR modes. For DDR VREF should be connected to 1.25V, and for SDR VREF should be connected to 1.5V. 18.4 18.4.1 Interface Recommendations Supported Memory Configurations The SDRAM controller supports up to 13 row addresses and up to 12 column addresses. However, the maximum row and column addresses are not supported at the same time. The number of row and column addresses must be less than or equal to 24. In addition to row/column address lines, there are always two row bank address bits. Therefore, the greatest possible address space which can be accessed using a single chip select is (226) x 32 bits, or 256 Mbytes. Table 18-2 shows the address multiplexing used by the MCF548x for different configurations. When the SDRAM controller receives the internal module enable, it latches the internal bus address lines addr[27:2] and multiplexes them into row, column and row bank addresses. addr[9:2] are always used for CA[7:0], addr[11:10] are always used for BA[1:0], and addr[23:12] are always used for RA[11:0]. addr[27:24] can be used for additional row or column address bits, as needed. NOTE The SDRAMC only supports an external 32-bit data bus. It is not possible to connect a smaller device(s) to only part of the SDRAM’s data bus. For example, if 16-bit wide devices are used, then you must use two 16-bit devices connected as a 32-bit port. MCF548x Reference Manual, Rev. 5 18-4 Freescale Semiconductor Interface Recommendations Table 18-2. SDRAM Address Multiplexing Device Configur ation Row bit x Number Col bit x of Bank bit Devices 512K x 32 bit 11 x 8 x 2 4M x 16 bit 12 x 8 x 2 Total SDCR Block [MUX] Size Setting 27 26 25 24 1 8 MB 00 — — — — 2 16 MB 00 — — — — 00 — — — CA8 4 32 MB 01 — — — RA12 00 — — CA9 CA8 8 64 MB 01 — — CA8 RA12 1 16 MB 00 — — — — 00 — — — CA8 2 32 MB 01 — — — RA12 00 — — CA9 CA8 01 — — CA8 RA12 00 — CA11 CA9 CA8 01 — CA9 CA8 RA12 00 — — CA9 CA8 01 — — CA8 RA12 00 — CA11 CA9 CA8 01 — CA9 CA8 RA12 00 CA12 CA11 CA9 CA8 01 CA11 CA9 CA8 RA12 00 — CA11 CA9 CA8 01 — CA9 CA8 RA12 00 CA12 CA11 CA9 CA8 01 CA11 CA8 RA12 12 x 9 x 2 64 Mbits 8M x 8bit 13 x 8 x 2 12 x 10 x 2 16M x 4 bit 4M x 32 bit 13 x 9 x 2 12 x 8 x 2 12 x 9 x 2 8M x 16 bit 128 Mbits 13 x 8 x 2 12 x 10 x 2 16M x 8 bit 13 x 9 x 2 4 64 MB 8 128 MB 2 64 MB 4 128 MB 8 256 MB 2 128 MB 12 x 11 x 2 32M x 4 bit 13 x 10 x 2 12 x 10 x 2 16M x 16 bit 13 x 9 x 2 12 x 11 x 2 256 Mbits 32M x 8 bit 13 x 10 x 2 12 x 12 x 2 64M x 4 bit 13 x 11 x 2 12 x 11 x 2 32M x 16 bit 512 Mbits 13 x 10x 2 12 x 12 x 2 64M x 8bit 13 x 11 x 2 4 256 MB Internal Address CA9 23–12 11–10 9–2 RA11-0 BA1-0 CA7-0 RA11-0 BA1-0 CA7-0 RA11-0 BA1-0 CA7-0 RA11-0 BA1-0 CA7-0 All memory devices of a single chip select block must have the same configuration and row/col address width; however, this is not necessary between different blocks. If mixing different memory organizations in different blocks, the following guidelines will ensure that every block is fully contiguous. • If all devices’ row address width is 12 bits, the column address can be ≥ 8 bits. • If all devices’ row address width is 13 bits, the column address can be ≥ 8 bits. • If all devices’ column address width is 8 bits, the row address can be ≥ 11 bits. • x8 and x16 data width memory devices can be mixed (but not in the same space). • x32 data width memory devices cannot be mixed with any other width. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 18-5 18.4.2 SDRAM SDR Connections Figure 18-2 shows a block diagram of the connections between the MCF548x and SDR SDRAM components. SDR design requires special timing consideration for the SDDQS[3:0] signals. For reads from DDR SDRAMs, the memory will drive the DQS pins so that the data lines and DQS signals have concurrent edges. The MCF548x SDRAMC is designed to latch data 1/4 clock after the SDDQS[3:0] edge. For DDR SDRAM, this ensures that the latch time is in the middle of the data valid window. The SDRAMC also uses the SDDQS[3:0] signals to determine when read data can be latched for SDR SDRAM; however, SDR memories do not provide DQS outputs. Instead the SDRAMC provides an SDRDQS output that is routed back into the controller as SDDQS[3:0]. The SDRDQS signal should be routed such that the valid data from the SDRAM reaches the MCF548x at the same time or just before the SDRDQS reaches the SDDQS[3:0] inputs. When routing SDRDQS the outbound trace length should be matched to the SDCLK trace length. This will align SDRDQS to the SDCLK as if the memory had generated the DQS pulse. The inbound trace should be routed along the data path. This should synchronize the SDDQS so that the data is latched in the middle of the data valid window. MCF548X SDR SDRAM SDADDR[12:0] A[12:0] SDBA[1:0] BA[1:0] SDDATA[31:0] DQ[31:0] SDCSn CS RAS CAS SDWE RAS CAS WE SDCLK[1:0] SDCKE CLK CKE SDDM[3:0] DQM[3:0] SDRDQS SDDQS[3:0] Figure 18-2. MCF548x Connections to SDR SDRAM 18.4.3 SDRAM DDR Component Connections Figure 18-3 shows a block diagram of the connections between the MCF548x and DDR SDRAM components. MCF548x Reference Manual, Rev. 5 18-6 Freescale Semiconductor Interface Recommendations DDR SDRAM MCF548X SDADDR[12:0] A[12:0] SDBA[1:0] BA[1:0] SDDATA[31:0] DQ[31:0] SDCSn CS RAS CAS SDWE RAS CAS WE SD_CLK[1:0] SD_CLK[1:0] SD_CKE CLK CLK CKE SDDM[3:0] SDDQS[3:0] DM[3:0] DQS[3:0] Figure 18-3. MCF548x Connections to DDR SDRAM 18.4.4 SDRAM DDR DIMM Connections There is a JEDEC standard for a 100-pin DDR DIMM with a 32-bit wide data bus. This DIMM standard was designed specifically to support 32-bit processors. The MCF548x can support current DIMM configurations up to 512 Mbytes. Figure shows a block diagram of the connections between the MCF548x and DDR SDRAM DIMMs. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 18-7 DDR SDRAM MCF548X SDADDR[12:0] A[12:0] SDBA[1:0] BA[1:0] SDDATA[31:0] DQ[31:0] SDCS[1:0] S[1:0] RAS CAS SDWE RAS CAS WE SDCLK[1:0] SDCLK[1:0] SDCKE CLK[1:0] CLK[1:0] CKE SDDM[3:0] DM[3:0] SDDQS[3:0] DQS[3:0] SCL SCL SDA SDA SDVDD SA0 Figure 18-4. MCF548x Connections to 100-pin DDR SDRAM DIMM 18.4.5 DDR SDRAM Layout Considerations Due to the critical timing for DDR SDRAM, there are a number of considerations that should be taken into account during PCB layout: • Minimize overall trace lengths. • Each DQS, DM, and DQ group must have identical loading and similar routing to maintain timing integrity. • Control and clock signals are routed point-to-point. • Trace length for clock, address, and command signals should match. • Route DDR signals on layers adjacent to the ground plane. • Use a VREF plane under the SDRAM. • VREF is decoupled from both SDVDD and VSS. • To avoid crosstalk, keep address and command signals separate from data and data strobes. • Use different resistor packs for command/address and data/data strobes. • Use single series, single parallel termination (25 Ω series, 50 Ω parallel values are recommended, but standard resistor packs with similar values can be substituted). • Series termination should be between the MCF548x and memory, but closest to the processor. • The parallel termination at end of the signal line (close to the SDRAM). • 0.1 uF decoupling for every termination resistor pack. MCF548x Reference Manual, Rev. 5 18-8 Freescale Semiconductor SDRAM Overview 18.4.5.1 Termination Example Figure 18-5 shows the recommended termination circuitry for DDR SDRAM signals. VREF 50 Ω DDR SDRAM MCF548X 25 Ω Figure 18-5. MCF548x DDR SDRAM Termination Circuit 18.5 18.5.1 SDRAM Overview SDRAM Commands When an internal bus master accesses SDRAM address space, the memory controller generates the corresponding SDRAM command. Table 18-3 lists SDRAM commands supported by the memory controller. Table 18-3. SDRAM Commands Function Symbol CKE CS RAS CAS WE BA[1:0] AP/C MD Other A Command Inhibit INH H H X X X X X X No Operation NOP H L H H H X X X Row and Bank Active ACTV H L L H H V V V Read READ H L H L H V L V Write WRITE H L H L L V L V Precharge All Banks PALL H L L H L X H X Load Mode Register LMR H L L L L LL V V Load Extended Mode Register LEMR H L L L L LH V V CBR Auto Refresh REF H L L L H X X X MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 18-9 Table 18-3. SDRAM Commands (Continued) Function Symbol CKE CS RAS CAS WE BA[1:0] AP/C MD Other A Self-Refresh SREF H→L L L L H X X X Power-Down PDWN H→L H X X X X X X H = High L = Low V = Valid X = Don’t care Many commands require a delay before the next command may be issued; sometimes the delay depends on the type of the next command. These delay requirements are managed by the values programmed in the memory controller configuration registers (SDCFG1, SDCFG2). 18.5.1.1 Row and Bank Active Command (ACTV) The ACTV command is responsible for latching the row and bank address and activating the specified row in the memory array. Once the row is activated, it can be accessed using subsequent READ and WRITE commands. NOTE The SDRAMC will support one active row for each chip select block. See Section 18.6.1, “Page Management” for more information. 18.5.1.2 Read Command (READ) When the SDRAMC receives a read request, it first checks the row and bank of the new access. If the address falls within the active row of an active bank, it is a page hit, and the READ is issued as soon as possible (pending any delays required by previous commands). If the address is within the active row, but the needed bank is inactive, or if there is no active row, the memory controller will issue an ACTV followed by the READ command. If the address is not within the active row, the memory controller will issue a PALL command to close the active row. Then the SDRAMC issues ACTV to activate the necessary bank and row for the new access, followed finally by the READ to the SDRAM. The PALL and ACTV commands (if necessary) can sometimes be issued in parallel with an on-going data movement. All reads, whether burst or single, must be allowed to complete the entire burst length on the memory bus. With SDR memory, the data masks are negated throughout the entire read burst length. With DDR memory, the data masks are asserted throughout the entire read burst length; but DDR memory ignores the data masks during reads. 18.5.1.3 Write Command (WRITE) When the memory controller receives a write request, it first checks the row and bank of the new access. If the address falls within the active row of an active bank, it is a page hit, and the WRITE is issued as soon as possible (pending any delays required by previous commands). If the address is within the active row but the needed bank is inactive, or if there is no active row, the memory controller will issue an ACTV followed by the WRITE command. If the address is not within the active row, the memory controller will MCF548x Reference Manual, Rev. 5 18-10 Freescale Semiconductor SDRAM Overview issue a PALL command to close the active row. Then the SDRAMC issues ACTV to activate the necessary row and bank for the new access, followed finally by the WRITE command. The PALL and ACTV commands (if necessary) can sometimes be issued in parallel with an on-going data movement. With both SDR and DDR memory, a read command can be issued overlapping the masked beats at the end of a previous single write of the same SDCS; the read command aborts the remaining (unnecessary) write beats. This is not possible with SDR memory, because SDR memory cannot be read with the masks asserted. 18.5.1.4 Precharge All Banks Command (PALL) The precharge command puts SDRAM into an idle state. The SDRAM must be in this idle state before a REF, LMR, LEMR, or ACTV command to open a new row within a particular bank can be issued. The memory controller issues the PALL command only when necessary for one of the following conditions: • Access to a new row • Refresh interval elapsed • Software commanded precharge NOTE The SDRAMC does not support the precharge selected bank memory command. 18.5.1.5 Load Mode/Extended Mode Register Command (LMR, LEMR) All SDRAM devices contain mode registers that are used to configure the timing and burst mode for the SDRAM. These commands are used to access the mode registers that physically reside within the SDRAM devices. During the LMR or LEMR command the SDRAM will latch the address bus and load the value into the selected mode register. NOTE The LMR and LEMR commands are only used during SDRAM initialization. The following steps should be used to write the mode register and extended mode register: 1. Set the SDCR[MODE_EN] bit. 2. Write the SDMR[BA] bits to select the mode register. 3. Write the desired mode register value to the SDMR[ADDR]. Don’t overwrite the SDMR[BA] values. 4. Set the SDMR[CMD] bit. 5. For DDR, repeat from step 2 for the extended mode register. 6. Clear the SDCR[MODE_EN] bit. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 18-11 18.5.1.5.1 Mode Register Definition Figure 18-6 shows the mode register definition. Note that this is the SDRAM’s mode register not the SDRAMC’s mode/extended mode register (SDMR) defined in Section 18.7.3, “SDRAM Mode/Extended Mode Register (SDMR).” Field BA1 BA0 0 0 A11 A10 A9 A8 A7 A6 OP_MODE A5 A4 CASL A3 A2 BT A1 A0 BLEN Figure 18-6. Mode Register Table 18-4. Mode Register Field Descriptions Address Line Description BA[1:0] Bank Address. These must both be zero to select the mode register. A11–A7 Operating Mode. 00000 Normal Operation 00010 Reset DLL Other values should not be used. A6–A4 CAS latency. Delay in clocks from issuing a READ to valid data out. Check the SDRAM manufacturer’s spec as the CASL settings supported can vary from memory to memory. A3 Burst Type. 0 Sequential 1 Interleaved. This setting should not be used since the SDRAMC does not support interleaved bursts. A2–A0 Burst length. Determines the number of locations that are accessed for a single READ or WRITE. 000 One. This is only a valid setting for SDR. 001 Two 010 Four 011 Eight (This value should be used for the MCF548x SDRAMC) 100–110 Reserved 111 Full page. This setting should not be used since full page bursting is not supported by the SDRAMC. 18.5.1.5.2 Extended Mode Register Definition Figure 18-7 shows the extended mode register used by DDR SDRAMs. Note that this is the SDRAM’s extended mode register, not the SDRAMC’s mode/extended mode register (SDMR) defined in Section 18.7.3, “SDRAM Mode/Extended Mode Register (SDMR).” Field BA1 BA0 0 1 A11 A10 A9 A8 A7 A6 A5 OPTION A4 A3 A2 A1 A0 DLL Figure 18-7. Extended Mode Register MCF548x Reference Manual, Rev. 5 18-12 Freescale Semiconductor SDRAM Overview Table 18-5. Extended Mode Register Field Descriptions Address Line Description BA[1:0] Bank Address. 00 Does not select the extended mode register 01 Selects the extended mode register 1x Reserved A11–A1 Option. These bits are not defined by the DDR specification. Each DDR SDRAM manufacturer can use these bits to implement optional features. Check with SDRAM manufacturer to determine if any optional features have been implemented. For normal operation all bits should be cleared. A0 18.5.1.6 Delay locked loop. Controls enabling of the delay locked loop circuitry used for DDR timing. 0 Enabled 1 Disabled. Auto Refresh Command (REF) The memory controller issues auto refresh commands according to the SDCR[RC] value. Each time the programmed refresh interval elapses, the memory controller issues a PALL command followed by a REF command. If a memory access is in progress at the time the refresh interval elapses, the memory controller schedules the refresh after the transfer is finished; but the interval timer continues counting so that the average refresh rate is constant. After REF, the SDRAM is in an idle state and waits for an ACTV command. 18.5.1.7 Self-Refresh (SREF) and Power-Down (PDWN) Commands The memory controller issues either a PDWN or a SREF command if the SDCR[CKE] bit is cleared. If the SDCR[REF] bit is set when CKE is negated, the controller issues a SREF command; if the REF bit is cleared, the controller issues a PDWN command. The REF bit may be changed in the same register write that changes the CKE bit; the controller will act upon the new value of the REF bit. Just like a REF, the controller automatically issues a PALL command before the self-refresh command. The memory is reactivated from power-down or self-refresh mode by setting the CKE bit. If a normal refresh interval elapses while the memory is in self-refresh mode, a PALL and REF will be performed as soon as the memory is reactivated. If the memory is put into and brought out of self-refresh all within a single refresh interval, the next automatic refresh will occur on schedule. In self-refresh mode, the memory does not require an external clock. To restart periodic refresh when the memory is reactivated, the REF bit must be reasserted. This can be done before the memory is reactivated, or in the same control register write that sets CKE to exit self-refresh mode. 18.5.2 Power-Up Initialization SDRAMs have a prescribed initialization sequence. The following sections detail the memory initialization steps for both SDR and DDR SDRAM. The sequence might change slightly from device-to-device. Refer to the device datasheet as the most relevant reference. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 18-13 18.5.2.1 SDR Initialization SDR initialization requires the following steps: 1. After reset is deactivated, pause for the amount of time indicated in the SDRAM specification. Usually 100μs or 200μs. 2. Initialize the SDRAM drive strength (SDRAMDS) and SDRAM chip select configuration (CSnCFG) registers. 3. Program the SDRAM configuration registers (SDCFG1 and SDCFG2) with the correct delay and timing values. 4. Issue a PALL command. Initialize the SDRAM control register (SDCR) with SDCR[IPALL] set. The SDCR[MODE_EN, REF, and IREF] bits should all remain cleared for this step. 5. Refresh the SDRAM. The SDRAM spec should indicate a number of refresh cycles to be performed before issuing an LMR command. Write to the SDCR with the IREF bit set (SDCR[MODE_EN, REF, and IPALL] should be cleared). This will force a refresh of the SDRAM each time the IREF bit is set. Repeat this step until the specified number of refresh cycles have completed. 6. Set SDCR[REF] to enable automatic refreshing for the rest of the initialization and regular operation. SDCR[MODE_EN, REF, and IPALL] remain cleared. 7. Initialize the SDRAM’s mode register using the LMR command. See Section 18.5.1.5, “Load Mode/Extended Mode Register Command (LMR, LEMR)” for more instruction on issuing an LMR command. 18.5.2.2 DDR Initialization The steps for DDR initialization are similar to the SDR initialization sequence; however, there are some additional steps required for DDR: 1. After reset is deactivated, pause for the amount of time indicated in the SDRAM specification. Usually 100μs or 200μs. 2. Initialize the SDRAM drive strength (SDRAMDS) and SDRAM chip select configuration (CSnCFG) registers. 3. Program the SDRAM configuration registers (SDCFG1 and SDCFG2) with the correct delay and timing values. 4. Issue a PALL command. Initialize the SDRAM control register (SDCR) with SDCR[IPALL] set. The SDCR[REF, and IREF] bits should remain cleared for this step. 5. Initialize the SDRAM’s extended mode register to enable the DLL. See Section 18.5.1.5, “Load Mode/Extended Mode Register Command (LMR, LEMR)” for instructions on issuing an LEMR command. 6. Initialize the SDRAM’s mode register and reset the DLL using the LMR command. See Section 18.5.1.5, “Load Mode/Extended Mode Register Command (LMR, LEMR)” for more instruction on issuing an LMR command. During this step the OP_MODE field of the mode register should be set to “normal operation/reset DLL.” 7. Pause for the DLL lock time specified by the memory. MCF548x Reference Manual, Rev. 5 18-14 Freescale Semiconductor Functional Overview 8. Issue a second PALL command. Initialize the SDRAM control register (SDCR) with SDCR[IPALL] set. The SDCR[REF, and IREF] bits should remain cleared for this step. 9. Refresh the SDRAM. The SDRAM spec should indicate a number of refresh cycles to be performed before issuing an LMR command. Write to the SDCR with the IREF bit set (SDCR[MODE_EN, REF, and IPALL] should be cleared). This will force a refresh of the SDRAM each time the IREF bit is set. Repeat this step until the specified number of refresh cycles have been completed. 10. Initialize the SDRAM’s mode register using the LMR command. See Section 18.5.1.5, “Load Mode/Extended Mode Register Command (LMR, LEMR)” for more instruction on issuing an LMR command. During this step the OP_MODE field of the mode register should be set to “normal operation.” 11. Set SDCR[REF] to enable automatic refreshing, and clear SDCR[MODE_EN] to lock the SDMR. SDCR[MODE_EN, IREF, and IPALL] remain cleared. 18.6 18.6.1 Functional Overview Page Management SDRAM devices have four internal banks. A particular row and bank of memory must be activated to allow read and write accesses. The SDRAM controller supports paging mode to maximize the memory access throughput. During operation, the SDRAM controller maintains an open page address for each SDCS block. An open page is composed of the active rows in the internal banks. SDRAMs can have a different row address open in each bank, but the SDRAMC does not support this. The page size of a SDCS block is equal to the space size divided by the number of rows; but the page may not be contiguous in the XLB address space because the internal address bits used for memory column address [11:8] and column address [7:0] are not consecutive. Because the column address may be split across two portions of the XLB address, the contiguous page size is (number of banks) × (256 columns) × (number of bits). This gives a contiguous page size of 4 Kbytes. However, the total (possibly fragmented) page size is (number of banks) × (number of columns) × (number of bits). If a new access does not fall in the open page of a SDCS block, the open page must be closed (PALL) and the new page must be opened (ACTV), then the READ or WRITE command can proceed. An ACTV command only activates one bank of a page. If another read or write falls in an inactive bank of the open page, another ACTV is needed but no precharge is needed. If a read or write falls in any of the active banks of the open page, no PALL or ACTV is needed; the read or write command can be issued immediately. A page is kept open until one of the following conditions occurs: • an access outside the open page • a refresh cycle is started. All SDCS blocks are refreshed at the same time; the refresh closes all banks of every SDRAM block. 18.6.2 Transfer Size In the MCF548x, the internal data bus is 64 bits wide, while the SDRAM external interface bus is 32 bits wide. Therefore, each XLB data beat requires two memory data beats. The SDRAM controller manages the size translation (packing/unpacking) between 64- and 32-bit buses. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 18-15 The SDRAM controller supports all possible XLB transfer sizes. SDRAMs are “burst only” devices; unnecessary beats on the memory bus are masked (write) or discarded (read). The SDRAMC will perform line bursts (32 byte) for all SDRAM access. This requires two beats of 16 bytes on the XLB, or eight beats of 4 bytes (one longword) on the memory bus. The SDRAM controller transfers the critical longword first, followed by the next three sequential longwords. The burst size and transfer order must be programmed in the SDRAM mode registers during initialization (SDMR); the burst size also must be programmed in the memory controller (SDCFG2). In a write operation, the data masks, SDDM[3:0], are used to inhibit writing unused bytes of each beat. In a read operation, the excess read data is discarded. 18.7 Memory Map/Register Definition The SDRAM controller contains four programming registers. Table 18-6. SDRAMC Memory Map Address (MBAR +) Name Byte0 Byte1 Byte2 Byte3 Access SDRAM Chip Select and Drive Strength Registers 0x04 SDRAM Drive Strength Register SDRAMDS R/W 0x20 SDRAM Chip Select 0 Configuration CS0CFG R/W 0x24 SDRAM Chip Select 1 Configuration CS1CFG R/W 0x28 SDRAM Chip Select 2 Configuration CS2CFG R/W 0x2C SDRAM Chip Select 3 Configuration CS3CFG R/W SDRAMC Configuration Registers 0x0100 SDRAM Mode/Extended Mode Register SDMR R/W 0x0104 SDRAM Control Register SDCR R/W 0x0108 SDRAM Configuration Register 1 SDCFG1 R/W 0x010C SDRAM Configuration Register 2 SDCFG2 R/W MCF548x Reference Manual, Rev. 5 18-16 Freescale Semiconductor Memory Map/Register Definition 18.7.1 R SDRAM Drive Strength Register (SDRAMDS) 31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 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 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 0 0 0 0 0 0 0 0 0 0 0 W Reset R SB_E SB_C SB_A 1 1 SB_S SB_D W Reset Reg Addr 1 1 1 1 1 1 1 1 MBAR + 0x04 Figure 18-8. SDRAM Drive Strength Register (SDRAMDS) Table 18-7. SDRAMDS Field Descriptions Bits Name Description 31–10 — 9–8 SB_E Controls the drive strength of SDCKE. See Table 18-8 for encodings. 7–6 SB_C Controls the drive strength of SDRAM clocks. See Table 18-8 for encodings. 5–4 SB_A Controls the drive strength of SDCS[3:0], RAS, CAS, SDWE, SDADDR[12:0], and SDBA[1:0]. See Table 18-8 for encodings. 3–2 SB_S Controls the drive strength of SDRDQS. See Table 18-8 for encodings. 1–0 SB_D Controls the drive strength of SDDATA[31:0], SDDM[3:0], and SDQS[3:0]. See Table 18-8 for encodings. Reserved. Should be cleared Table 18-8. SDRAM Drive Strength Bit Encodings 1 SB_x[1:0] SD_VDD1 10 3.3 8mA; SSTL_3 Class I 01 3.3 16mA; SSTL_3 Class II 00 3.3 24mA; SSTL_3 10 2.5 7.6mA; SSTL_2 Class I 01 2.5 13mA 00 2.5 15mA; SSTL_2 Class II 11 X Drive No Drive;Hi-Z 3.3V is for SDR mode, 2.5V is for DDR mode MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 18-17 18.7.2 SDRAM Chip Select Configuration Registers (CSnCFG) 31 30 29 28 27 R 26 25 24 23 22 21 20 CSBA 19 18 17 16 0 0 0 0 W Reset 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 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 R CSSZ W Reset Reg Addr 0 0 0 MBAR + 0x20 (CS0CFG), 0x24 (CS1CFG), 0x28 (CS2CFG), 0x2C (CS3CFG) Figure 18-9. SRAM Chip Select Configuration Register (CSnCFG) Table 18-9. CFnCFG Field Descriptions Bits Name 31–20 CSBA 19–5 — 4–0 CSSZ Description Chip select base address. Reserved. Should be cleared. Chip select size. 00000 Disabled 00001–10010 Reserved 10011 1 Mbyte, compare A[31:20] 10100 2 Mbyte, compare A[31:21] 10101 4 Mbyte, compare A[31:22] 10110 8 Mbyte, compare A[31:23] 10111 16 Mbyte, compare A[31:24] 11000 32 Mbyte, compare A[31:25] 11001 64 Mbyte, compare A[31:26] 11010 128 Mbyte, compare A[31:27] 11011 256 Mbyte, compare A[31:28] 11100 512 Mbyte, compare A[31:29] 11101 1 Gbyte, compare A[31:30] 11110 2 Gbyte, compare A31 11111 4 Gbyte, ignore A[31:20] Any chip select can be enabled or disabled, independent of others. Any chip select can be allocated any size of address space from 1 Mbyte to 4 Gbyte, independent of others. Any chip select address space can begin at any size-aligned base address, independent of others. For contiguous memory with different sizes of mem banks, place largest bank at lowest address, then place smaller banks in descending size order at ascending base address. For example, assume CS0 = 16M, CS1 = empty, CS2 = 64M, CS3 = 64M, CS4 = 256M, CS5 = empty: CS0CFG = 98000017 = enable 16M @ 0x9800 0000-0x98FF FFFF CS1CFG = 00000000 = disable CS2CFG = 90000019 = 64M @ 0x9000 0000-0x93FF FFFF MCF548x Reference Manual, Rev. 5 18-18 Freescale Semiconductor Memory Map/Register Definition CS3CFG = 94000019 = 64M @ 0x9400 0000-0x97FF FFFF CS4CFG = 8000001b = 256M @ 0x8000 0000-0x8FFF FFFF CS5CFG = 00000000 = disable This gives 400 Mbyte total memory, at 0x8000 0000-0x98FF FFFF 18.7.3 SDRAM Mode/Extended Mode Register (SDMR) The SDMR, shown in Figure 18-10, is used to write to the mode and extended mode registers that physically reside within in the SDRAM chips. These registers must be programmed during SDRAM initialization. See Section 18.5.2, “Power-Up Initialization” for more information on the initialization sequence. 31 R 30 29 28 27 26 25 24 BNKAD 23 22 21 20 19 18 AD 17 16 0 CMD W Reset R Uninitialized 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 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 0 0 W Reset Reg Addr MBAR +0x0100 Figure 18-10. SDRAM Mode/Extended Mode Register (SDMR) Table 18-10. SDMR Field Descriptions Bits Name Description 31–30 BNKAD Bank address. Driven onto SDBA[1:0] along with a LMR/LEMR command. All SDRAM chip selects are asserted simultaneously. SDCR[CKE] must be set before attempting to generate an LMR/LEMR command. The SDBA[1:0] value is used to select between LMR and LEMR commands. 00 Load mode register command (LMR) 01 Load extended mode register command (LEMR) 10–11 Reserved 29–18 AD Address. Driven onto SDADDR[11:0] along with an LMR/LEMR command. The AD value is stored as the mode (or extended mode) register data. 17 — Reserved. Should be cleared. 16 CMD 15–0 — Command. 1 Generate an LMR/LEMR command 0 Do not generate any command Reserved. Should be cleared. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 18-19 18.7.4 SDRAM Control Register (SDCR) The SDCR, shown in Figure 18-11, controls SDRAMC operating modes including the refresh count and address line muxing. 31 30 R MODE CKE _EN W 29 28 27 26 DDR REF 0 0 25 24 MUX Reset 23 22 AP DRIV E 21 20 19 18 17 16 RCNT Uninitialized R 15 14 13 12 0 0 0 0 11 10 9 8 DQS_OE 7 6 5 4 3 2 1 0 0 0 0 BUFF 0 IREF IPALL 0 W Reset Uninitialized Reg Addr MBAR + 0x0104 Figure 18-11. SDRAM Control Register (SDCR) Table 18-11. SDCR Field Descriptions Bits 31 Name Description MODE_EN Mode enable. 0 Mode register locked, cannot be written 1 Mode register enabled, can be written 30 CKE Clock enable. 0 SDCKE is negated (low) 1 SDCKE is asserted (high) 29 DDR DDR mode select. 0 SDR mode 1 DDR mode 28 REF Refresh enable. 0 Automatic refresh disabled 1 Automatic refresh enabled 27–26 — Reserved. Should be cleared. 25–24 MUX 23 AP Muxing control. Selects routing of addr[7:4] as row or column address bits as shown in Table 18-2. Auto precharge control bit. 0 CA10 is the auto precharge control bit 1 Reserved MCF548x Reference Manual, Rev. 5 18-20 Freescale Semiconductor Memory Map/Register Definition Table 18-11. SDCR Field Descriptions (Continued) Bits Name Description 22 DRIVE Drive rule selection. 0 Tri-state except to write. SDDATA and SDDQS are only driven when necessary to perform a write. 1 Drive except to read. SDDATA and SDDQS are only tristated when necessary to perform a read. When not being driven for a write cycle, SDDATA hold the most recent value and SDDQS are driven low. This mode is intended for minimal applications only, to prevent floating signals and allow unterminated board traces. However, terminated wiring is always recommended over unterminated. 21–16 RCNT Refresh Count. Controls automatic refresh frequency. The number of bus clocks between refresh cycles is (RC + 1) x 64. RCNT = (tREFI/ (SDCLK x 64)) - 1, rounded down to the next integer value. 15–12 — 11–8 DQS_OE Reserved. Should be cleared. DQS output enable. Each DQS_OE bit is a master enable for the corresponding SDDQSn signal. 1 SDDQSn can drive as necessary, depending on commands and SDCR[DRIVE] setting. 0 SDDQSn can never drive. Use this value in SDR mode or in DDR mode with a “single DQS” memory. Some 32-bit DDR devices only have a single DQS pin. Enable one of the SDDQSn signals and disable the other three. Then short all 4 pins external to the part. 7–5 — 4 BUFF 3 — 2 IREF Initiate Refresh (REF) command. Used to force a software initiated Refresh command. 1 Generate a Refresh command. All SDCSn signals are asserted simultaneously. SDCR[CLK_EN] must be set before attempting to generate a software refresh command. 0 Do not generate a Refresh command. 1 IPALL Initiate Precharge All command. Used to force a software initiated PALL command. 1 Generate a PALL command. All SDCSn signals are asserted simultaneously. SDCR[CKE] must be set before attempting to generate a software PALL command. 0 Do not generate a PALL command. 0 — 18.7.5 Reserved. Should be cleared. Buffering mode. Selects between buffered and unbuffered memory timing. Buffered and unbuffered memory cannot be mixed. 1 System uses “buffered” memory modules. 0 System does not use “buffered” memory modules. Reserved. Should be cleared. Reserved. Should be cleared. SDRAM Configuration Register 1 (SDCFG1) The 32-bit read/write SDRAM configuration register 1 (SDCFG1) stores delay values necessary between specific SDRAM commands. During initialization, software loads values to the register according to the selected SDCLK frequency and SDRAM specifications. This register is reset only by a power-up reset signal. The read and write latency fields govern the relative timing of commands and data, and must be exact values. All other fields govern the relative timing from one command to another, they have minimum values but any larger value is also legal (but with decreased performance). MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 18-21 The minimum values of certain fields can be different for SDR and DDR SDRAM, even if the data sheet timing is the same, because: • In SDR mode, the memory controller counts the delay in SDCLK • In DDR mode, the memory controller counts the delay in SDCLK × 2 SDCLK—memory controller clock—is the speed of the SDRAM interface and is equal to the internal bus clock. SDCLK × 2—double frequency of SDCLK—DDR uses both edges of the bus-frequency clock (SDCLK) to read/write data 31 R 30 29 28 SRD2RW 27 26 0 25 24 23 SWT2RD 22 21 20 RDLAT 19 18 0 17 16 ACT2RW W Reset Uninitialized 15 R 0 14 13 12 PRE2ACT 11 10 9 REF2ACT 8 7 0 6 5 4 WTLAT 3 2 1 0 0 0 0 0 W Reset Uninitialized Reg Addr MBAR + 0x0108 Figure 18-12. SDRAM Configuration Register 1 (SDCFG1) Table 18-12. SDCFG1 Field Descriptions Bits Name 31–28 SRD2RW 27 — 26–24 SWT2RD 23–20 RDLAT Description Single Read to Read/Write/Precharge delay. Limiting case is usually Write to Precharge. DDR mode: SRD2RW = CASL + (BL/2) + 1 For DDR, suggested value = 0x7 SDR mode: SRD2RW = CASL + BL + 1 If CASL=2, suggested value = 0xB If CASL=3, suggested value = 0xC Reserved. should be cleared Single Write to Read/Write/Precharge delay. Limiting case is Write to Precharge. DDR mode: SWT2RD = tWR/SDCLK + 1, suggested value = 0x3 SDR mode: SWT2RD = tWR, suggested value = 0x2 Read CAS Latency. Read latency. Read command to read data available delay counter. DDR mode: If CASL = 2, write 0x6 If CASL = 2.5, write 0x7 SDR mode: If CASL = 2, write 0x2 If CASL = 3, write 0x3 Note: CASL=2.5 is not supported for SDR. MCF548x Reference Manual, Rev. 5 18-22 Freescale Semiconductor Memory Map/Register Definition Table 18-12. SDCFG1 Field Descriptions (Continued) Bits Name 19 — 18–16 ACT2RW Description Reserved. Should be cleared. Active to Read/Write delay. Active command to any following read or write delay counter. Suggested value = tRCD/SDCLK - 1 (Round up to nearest integer) EXAMPLE: If tRCD = 20ns and SDCLK = 99 MHz 20ns / 10.1 ns = 1.98; round to 2; write 0x1. Note: Count value is in SDCLK periods for both SDR and DDR mode. 15 — 14–12 PRE2ACT Reserved. Should be cleared. Precharge to Active delay. Precharge command to following Active command delay counter. Suggested value = tRP/SDCLK - 1 (Round up to nearest integer) EXAMPLE: If tRP = 20ns and SDCLK = 99MHz 20ns / 10.1ns = 1.98; round to 2; write 0x1. Note: Count value is in SDCLK periods for both SDR and DDR mode. 11–8 REF2ACT Refresh to Active delay. Refresh command to following Active or Refresh command delay counter. Suggested value = tRFC/SDCLK - 1 (Round up to nearest integer) EXAMPLE: If tRFC = 75ns and SDCLK = 99MHz 75ns / 10.1ns = 7.425; round to 8; write 0x7. Note: Count value is in SDCLK periods for both SDR and DDR mode. 7 — 6–4 WTLAT 3–0 — 18.7.6 Reserved. Should be cleared. Write latency. Write command to write data delay counter. For DDR, write 0x3 For SDR, write 0x0 Reserved. Should be cleared. SDRAM Configuration Register 2 (SDCFG2) The 32-bit read/write configuration register 2 stores delay values necessary between specific SDRAM commands. During initialization, software loads values to the register according to the SDRAM information obtained from the data sheet. This register is reset only by a power-up reset signal. The burst length (BL) field must be exact. All other fields govern the relative timing from one command to another, they have minimum values, but any larger value is also legal (but with decreased performance). All delays in this register are expressed in SDCLK. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 18-23 31 30 R 29 28 27 BRD2PRE 26 25 24 23 BWT2RW 22 21 20 19 18 17 16 BL BRD2WT W Reset R Uninitialized 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 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 0 0 W Reset Reg Addr MBAR + 0x010C Figure 18-13. SDRAM Configuration Register 2 (SDCFG2) Table 18-13. SDCFG2 Field Descriptions Bits 18.8 Name Description 31–28 BRD2PRE Burst Read to Read/Precharge delay. Limiting case is Read to Read. For DDR, suggested value = 0x4 (BurstLength/2) For SDR, suggested value = 0x8 (BurstLength) 27–24 BWT2RW Burst Write to Read/Write/Precharge delay. Limiting case is Write to Precharge. For DDR, suggested value = 0x6 (BurstLength/2 + tWR) For SDR, suggested value = 0x8 (BurstLength + tWR - 2 Clocks) 23–20 BRD2WT Burst Read to Write delay. For DDR, suggested value = 0x7 For SDR: If CASL = 2, suggested value = 0xB If CASL = 3, suggested value = 0xC 19–16 BL Burst Length. Write 0x7 (Burst Length - 1) 15–0 — Reserved. Should be cleared. SDRAM Example This example interfaces two 16M × 16-bit × 4 bank DDR SDRAM components to an MCF548x operating at a 120 MHz SDCLK frequency. Table 18-14 lists design specifications for this example. Table 18-14. SDRAM Example Specifications Parameter Specification 13 row and 9 column addresses Two bank-select lines to access four internal banks Allowable burst lengths 2, 4, or 8 CAS latency 2 Clock cycle time (tCK) 7.5ns (min) ACTV-to-read/write 15 ns (min) 18ns (max) delay (tRCD) MCF548x Reference Manual, Rev. 5 18-24 Freescale Semiconductor SDRAM Example Table 18-14. SDRAM Example Specifications (Continued) Parameter 18.8.1 Specification Write recovery timer (tWR) 15 ns Precharge command to ACTV command (tRP) 15 ns (min) 18ns (max) Auto refresh command period (tRFC) 72ns (min) 75ns (max) Average periodic refresh interval (tREFI) 7.8 μs SDRAM Signal Drive Strength Settings The SDRAMDS should be programmed as shown in Figure 18-14. The settings assume the normal drive strength for 2.5V drive, 7.6mA, is sufficient for the loading in the system. 31 30 29 28 27 26 25 24 Field 23 22 21 20 19 18 17 16 1 0 — Setting 0000_0000_0000_0000 (hex) 0 15 14 0 13 Field 12 11 10 0 9 — 8 SB_E Setting 7 6 0 5 SB_C 4 SB_A 3 2 SB_S SB_D 0000_0010_1010_1010 (hex) 0 2 A A Figure 18-14. SDRAM Example Drive Strength Settings (SDRAMDS) This configuration results in a value of SDRAMDS = 0x0000_02AA, as described in Table 18-15. Table 18-15. SDRAMDS Field Descriptions Bits Name Setting 31–10 — 0 Reserved. Should be cleared 9–8 SB_E 10 2.5V, 7.6mA SSTL_2 Class I drive 7–6 SB_C 10 2.5V, 7.6mA SSTL_2 Class I drive 5–4 SB_A 10 2.5V, 7.6mA SSTL_2 Class I drive 3–2 SB_S 10 2.5V, 7.6mA SSTL_2 Class I drive 1–0 SB_D 10 2.5V, 7.6mA SSTL_2 Class I drive 18.8.2 Description SDRAM Chip Select Settings For this example, the SDRAM will be connected to SDCS0 with a base address of 0x0. All other chip selects are unused and do not need to be initialized. The CS0CFG should be programmed as shown in Figure 18-15. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 18-25 31 30 29 28 27 26 Field 25 24 23 22 21 20 19 18 BA Setting 17 16 1 0 — 0000_0000_0000_0000 (hex) 0 15 14 0 13 12 11 Field 10 0 9 8 7 6 0 5 4 3 — 2 CSSZ Setting 0000_0000_0001_1001 (hex) 0 0 1 9 Figure 18-15. SDRAM Example Chip Select 0 Configuration Settings (CS0CFG) This configuration results in a value of SDRAMDS = 0x0000_0019, as described in Table 18-16. Table 18-16. CS0CFG Field Descriptions Bits Name Setting 31–20 BA 0 Base address is set to 0x0 19–5 — 0 Reserved. Should be cleared. 4–0 CSSZ 1101 18.8.3 Description Total size is 64 Mbytes. 2 x 256Mbit = 64Mbytes SDRAM Configuration 1 Register Settings The SDCFG1 register should be programmed as shown in Figure 18-16. 31 Field 30 29 28 27 SRD2RW 26 — 25 23 SWT2RD 22 21 20 RDLAT 19 18 — 17 16 ACT2RW 0111_0011_0110_0010 Setting (hex) 7 15 Field 24 — 14 3 13 12 11 PRE2ACT 10 9 8 7 6 REF2ACT Setting (hex) 6 2 5 4 3 2 1 WTLAT — 3 0 0 0010_1000_0011_0000 2 8 Figure 18-16. SDRAM Example Configuration Register 1 Settings (SDCFG1) This configuration results in a value of SDCFG1 = 0x7362_2830, as described in Table 18-17. Table 18-17. SDCFG1 Field Descriptions Bits Name Setting 31–28 SRD2RW 111 27 — 0 26–24 SWT2RD 011 Description SRD2RW = CASL + (burst length/2) + 1 = 2 + 4+ 1 = 7 Reserved. Should be cleared. SWT2RD = tWR/SDCLK + 1 = 15ns/8.3ns + 1 = 2.8 clocks, rounded up to 3 MCF548x Reference Manual, Rev. 5 18-26 Freescale Semiconductor SDRAM Example Table 18-17. SDCFG1 Field Descriptions (Continued) Bits Name Setting 23–20 RDLAT 0110 19 — 0 18–16 ACT2RW 010 15 — 0 14–12 PRE2ACT 010 PRE2ACT = tRP/SDCLK - 1 = 18ns/8.3ns - 1 = 2.16 - 1 = 1.16, rounded up to 2 11–8 REF2ACT 1000 REF2ACT = tRFC/SDCLK - 1 = 75ns/8.3ns - 1 = 9 - 1 = 8 7 — 0 6–4 WTLAT 011 0x3 is the recommended value for DDR 3–0 CSSZ 1101 Total size is 64 Mbytes. 2 x 256Mbit = 64Mbytes 18.8.4 Description 0x6 is the recommended value for DDR memory with a CASL of 2 Reserved. Should be cleared. ACT2RW = tRCD/SDCLK - 1 = 18ns/8.3ns - 1 = 2.16 - 1 = 1.16, rounded up to 2 Reserved. Should be cleared. Reserved. Should be cleared. SDRAM Configuration 2 Register Settings The SDCFG2 register should be programmed as shown in Figure 18-17. 31 Field 30 29 28 27 BRD2PRE 26 25 24 23 BWT2RW 22 21 20 19 18 BRD2WT 17 16 1 0 BL 0100_0110_0111_0111 Setting (hex) 4 15 14 6 13 12 11 10 7 9 Field 8 7 6 7 5 4 3 2 — Setting 0000_0000_0000_0000 (hex) 0 0 0 0 Figure 18-17. SDRAM Example Configuration Register 2 Settings (SDCFG2) This configuration results in a value of SDCFG2 = 0x4677_0000, as described in Table 18-18. Table 18-18. SDCFG2 Field Descriptions Bits Name Setting 31–28 BRD2PRE 0100 BRD2PRE = burst length/2 = 8/2 = 4 27–24 BWT2RW 0110 BWT2RW = burst length/2 + tWR = 8/2 + 2 = 4 + 2 = 6 23–20 BRD2WT 0111 0x7 is the recommended value for DDR 19–16 BL 0111 BL = burst length - 1 = 8 - 1 = 7 15–0 — 0 18.8.5 Description Reserved. Should be cleared. SDRAM Control Register Settings and PALL command The SDCR should be programmed as shown in Figure 18-18. Along with the base settings for the SDCR the MODE_EN and IPALL bits are set to issue a PALL command to the SDRAM and enable writing of the mode register. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 18-27 31 30 Field MODE CKE _EN 29 28 DDR REF 27 26 25 — 24 MUX Setting 23 22 21 AP DRIVE 20 19 18 17 16 RCNT 1110_0001_0000_1101 (hex) E 15 14 Field 1 13 12 — 11 10 0 9 8 DQS_OE Setting 7 6 D 5 — 4 3 BUFF — 2 1 IREF IPALL 0 — 0000_0000_0000_0010 (hex) 0 0 0 2 Figure 18-18. SDRAM Control Register Settings + MODE_EN and IPALL This configuration results in a value of SDCR = 0xE10D_0002, as described in Table 18-19. Table 18-19. SDCR + MODE_EN and IPALL Field Descriptions Bits Name Setting Description 31 MODE_EN 1 Mode register is writable. 30 CKE 1 SDCKE is asserted 29 DDR 1 DDR mode is enabled 28 REF 0 Automatic refresh is disabled 27–26 — 00 Reserved. Should be cleared. 25–24 MUX 01 01 is the MUX setting for a 13 x 9 x 4 memory. See Table 18-2. 23 AP 0 0 sets the auto precharge control bit to A10. 22 DRIVE 0 Data and DQS lines are only driven for a write cycle. 21–16 RCNT 001101 15–12 — 0000 Reserved. Should be cleared. 11–8 DQS_OE 0000 0x0 disables drive for all SDDQS pins for now. 7–5 — 000 Reserved. Should be cleared. 4 BUFF 0 0 indicates that a buffered memory module is not being used. 3 — 0 Reserved. Should be cleared. 2 IREF 0 Do not initiate a REF command. 1 IPALL 1 Initiate a PALL command. 0 — 0 Reserved. Should be cleared. RCNT = (tREFI/ (SDCLK x 64)) - 1 = (7800ns/(8.3ns x 64)) - 1 = 13.62, round down to 13 (0xD) MCF548x Reference Manual, Rev. 5 18-28 Freescale Semiconductor SDRAM Example 18.8.6 Set the Extended Mode Register The SDMR should be programmed as shown in Figure 18-19. This step enables the DDR memory’s DLL. 31 Field 30 29 28 27 26 25 BNKAD 24 23 22 21 20 19 OPTION Setting 18 17 16 DLL — CMD 1 0 0100_0000_0000_0001 (hex) 4 15 14 0 13 12 11 10 0 9 8 Field 7 6 1 5 4 3 2 — Setting 0000_0000_0000_0000 (hex) 0 0 0 0 Figure 18-19. SDRAM Mode/Extended Mode Register Settings (SDMR) This configuration results in a value of SDMR = 0x4001_0000, as described in Table 18-20. Table 18-20. SDMR Field Descriptions Bits Name Setting 31–30 BNKAD 01 01 selects the extended mode register. 29–18 OPTION 0 Optional operating modes for the DDR. 0 selects normal operation. 18 DLL 0 Enable the DLL. 17 — 0 Reserved. Should be cleared. 16 CMD 1 Initiate the LEMR command. 15–0 — 0 Reserved. Should be cleared. 18.8.7 Description Set the Mode Register and Reset DLL The SDMR should be programmed as shown in Figure 18-20. This step programs the mode register and resets the DLL. 31 Field 30 29 28 BNKAD 27 26 25 OP_MODE 23 22 21 CASL 20 BT 19 18 BLEN 17 16 — CMD 1 0 0000_0100_1000_1101 Setting (hex) 0 15 14 4 13 12 11 10 8 9 Field 8 7 6 D 5 4 3 2 — Setting (hex) 24 0000_0000_0000_0000 0 0 0 0 Figure 18-20. SDRAM Mode/Extended Mode Register Settings (SDMR) This configuration results in a value of SDMR = 0x048D_0000, as described in Table 18-21. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 18-29 Table 18-21. SDMR Field Descriptions Bits Name Setting 31–30 BNKAD 00 29–25 OP_MODE 0010 Selects normal operating mode and resets the DLL. 24–22 CASL 010 CAS latency of two clocks. 21 BT 0 20–18 BLEN 011 17 — 0 Reserved. Should be cleared. 16 CMD 1 Initiate the LMR command. 15–0 — 0 Reserved. Should be cleared. 18.8.8 Description 00 selects the mode register. Sequential burst type. Burst length of eight Issue a PALL command The SDCR should be programmed as shown in Figure 18-21. This will issue a second PALL command to the memory. The same SDCR value calculated in Section 18.8.5, “SDRAM Control Register Settings and PALL command” is used (0xE10D_0002). 31 30 Field MODE CKE _EN 29 28 27 DDR REF 26 25 — 24 MUX Setting 23 22 AP DRIV E 21 20 19 18 17 16 1 0 RCNT 1110_0001_0000_1101 (hex) E 15 Field 14 1 13 12 11 — 10 0 9 8 7 DQS_OE Setting 6 D 5 — 4 3 2 BUFF — IREF IPALL — 0000_0000_0000_0010 (hex) 0 0 0 2 Figure 18-21. SDRAM Control Register Settings + MODE_EN and IPALL This configuration results in a value of SDCR = 0xE10D_0002, as described in Table 18-22. Table 18-22. SDCR + MODE_EN and IPALL Field Descriptions Bits Name Setting Description 31 MODE_EN 1 Mode register is writable. 30 CKE 1 SDCKE is asserted 29 DDR 1 DDR mode is enabled 28 REF 0 Automatic refresh is disabled 27–26 — 00 Reserved. Should be cleared. 25–24 MUX 01 01 is the MUX setting for a 13 x 9 x 4 memory. See Table 18-2. 23 AP 0 0 sets the auto precharge control bit to A10. MCF548x Reference Manual, Rev. 5 18-30 Freescale Semiconductor SDRAM Example Table 18-22. SDCR + MODE_EN and IPALL Field Descriptions (Continued) Bits Name Setting 22 DRIVE 0 21–16 RCNT 001101 15–12 — 0000 Reserved. Should be cleared. 11–8 DQS_OE 0000 0x0 disables drive for all SDDQS pins for now. 7–5 — 000 Reserved. Should be cleared. 4 BUFF 0 0 indicates that a buffered memory module is not being used. 3 — 0 Reserved. Should be cleared. 2 IREF 0 Do not initiate a REF command. 1 IPALL 1 Initiate a PALL command. 0 — 0 Reserved. Should be cleared. 18.8.9 Description Data and DQS lines are only driven for a write cycle. RCNT = (tREFI/ (SDCLK x 64)) - 1 = (7800ns/(8.3ns x 64)) - 1 = 13.62, round down to 13 (0xD) Perform Two Refresh Cycles The SDCR should be programmed as shown in Figure 18-22. Along with the base settings for the SDCR the MODE_EN and IREF bits are set to issue an REF command to the SDRAM and enable writing of the mode register. The memory used in this example requires two refresh cycles, so this step is repeated twice. 31 30 Field MODE CKE _EN 29 28 DDR REF 27 26 25 — MUX Setting 23 22 21 AP DRIVE 20 19 18 17 16 1 0 RCNT 1110_0001_0000_1101 (hex) E 15 Field 24 1 14 13 12 — 11 10 0 9 8 7 DQS_OE Setting 6 D 5 — 4 3 BUFF — 2 IREF IPALL — 0000_0000_0000_0100 (hex) 0 0 0 4 Figure 18-22. SDRAM Control Register Settings + MODE_EN and IREF This configuration results in a value of SDCR = 0xE10D_0004, as described in Table 18-19. Table 18-23. SDCR + MODE_EN and IREF Field Descriptions Bits Name Setting Description 31 MODE_EN 1 Mode register is writable. 30 CKE 1 SDCKE is asserted 29 DDR 1 DDR mode is enabled 28 REF 0 Automatic refresh is disabled MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 18-31 Table 18-23. SDCR + MODE_EN and IREF Field Descriptions (Continued) Bits Name Setting Description 27–26 — 00 Reserved. Should be cleared. 25–24 MUX 01 01 is the MUX setting for a 13 x 9 x 4 memory. See Table 18-2. 23 AP 0 0 sets the auto precharge control bit to A10. 22 DRIVE 0 Data and DQS lines are only driven for a write cycle. 21–16 RCNT 001101 15–12 — 0000 Reserved. Should be cleared. 11–8 DQS_OE 0000 0x0 disables drive for all SDDQS pins for now. 7–5 — 000 Reserved. Should be cleared. 4 BUFF 0 0 indicates that a buffered memory module is not being used. 3 — 0 Reserved. Should be cleared. 2 IREF 1 Initiate a REF command. 1 IPALL 0 Do not initiate a PALL command. 0 — 0 Reserved. Should be cleared. RCNT = (tREFI/ (SDCLK x 64)) - 1 = (7800ns/(8.3ns x 64)) - 1 = 13.62, round down to 13 (0xD) 18.8.10 Clear the Reset DLL Bit in the Mode Register The SDMR should be programmed as shown in Figure 18-20. This step programs the mode register and enables normal operation of the DLL by clearing the “reset DLL” option. 31 Field 30 29 28 BNKAD 27 26 25 24 OP_MODE 23 22 21 CASL 20 BT 19 18 BLEN 17 16 — CMD 1 0 0000_0000_1000_1101 Setting (hex) 0 15 14 0 13 12 11 10 9 Field 8 7 6 D 5 4 3 2 — Setting (hex) 8 0000_0000_0000_0000 0 0 0 0 Figure 18-23. SDRAM Mode/Extended Mode Register Settings This configuration results in a value of SDMR = 0x008D_0000, as described in Table 18-21. Table 18-24. SDMR Field Descriptions Bits Name Setting Description 31–30 BNKAD 00 29–25 OP_MODE 0000 Selects normal operating mode. 24–22 CASL 010 CAS latency of two clocks. 00 selects the mode register. MCF548x Reference Manual, Rev. 5 18-32 Freescale Semiconductor SDRAM Example Table 18-24. SDMR Field Descriptions (Continued) Bits Name Setting Description 21 BT 0 Sequential burst type. 20–18 BLEN 011 Burst length of eight. 17 — 0 Reserved. Should be cleared. 16 CMD 1 Initiate the LMR command. 15–0 — 0 Reserved. Should be cleared. 18.8.11 Enable Automatic Refresh and Lock Mode Register The SDCR should be programmed as shown in Figure 18-24. Along with the base settings for the SDCR the REF bit is set to enable automatic refreshing of the memory. In addition, the MODE_EN bit is cleared to disable write to the SDMR. 31 30 Field MODE CKE _EN 29 28 27 DDR REF 26 25 — 24 MUX Setting 23 22 AP DRIV E 21 20 19 18 17 16 1 0 RCNT 0111_0001_0000_1101 (hex) 7 15 Field 14 1 13 12 11 — 10 0 9 8 7 DQS_OE Setting 6 D 5 — 4 3 BUFF — 2 IREF IPALL — 0000_1111_0000_0000 (hex) 0 F 0 0 Figure 18-24. SDRAM Control Register Settings + REF This configuration results in a value of SDCR = 0x710D_0F00, as described in Table 18-25. Table 18-25. SDCR + REF Field Descriptions Bits Name Setting Description 31 MODE_EN 0 Mode register is not writable. 30 CKE 1 SDCKE is asserted 29 DDR 1 DDR mode is enabled 28 REF 1 Automatic refresh is enabled. 27–26 — 00 Reserved. Should be cleared. 25–24 MUX 01 01 is the MUX setting for a 13 x 9 x 4 memory. See Table 18-2. 23 AP 0 0 sets the auto precharge control bit to A10. 22 DRIVE 0 Data and DQS lines are only driven for a write cycle. 21–16 RCNT 001101 RCNT = (tREFI/ (SDCLK x 64)) - 1 = (7800ns/(8.3ns x 64)) - 1 = 13.62, round down to 13 (0xD) MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 18-33 Table 18-25. SDCR + REF Field Descriptions (Continued) Bits Name Setting Description 15–12 — 0000 Reserved. Should be cleared. 11–8 DQS_OE 1111 0xF enables drive for all SDDQS pins. 7–5 — 000 Reserved. Should be cleared. 4 BUFF 0 0 indicates that a buffered memory module is not being used. 3 — 0 Reserved. Should be cleared. 2 IREF 0 Initiate a REF command. 1 IPALL 0 Do not initiate a PALL command. 0 — 0 Reserved. Should be cleared. 18.8.12 Initialization Code The following assembly code initializes the DDR SDRAM using the register values determined above. Basic Configuration and Initialization: move.l move.l move.l move.l move.l move.l move.l move.l #0x000002AA, d0, SDRAMDS #0x00000019, d0, CS0CFG #0x73622830, d0, SDCFG1 #0x46770000, d0, SDCFG2 d0//Initialize SDRAMDS d0//Initialize SDCS0 d0//Initialize SDCFG1 d0//Initialize SDCFG2 Precharge Sequence and enable write to SDMR: move.l move.l #0xE10D0002, d0//Initialize SDCR, send PALL, enable SDMR d0, SDCR Write Extended Mode Register: move.l move.l #0x40010000, d0//Write LEMR to enable DLL d0, SDMR Write Mode Register and Reset DLL: move.l move.l #0x048D0000, d0//Write LMR and reset DLL d0, SDMR Precharge Sequence: move.l move.l #0xE10D0002, d0//Send PALL d0, SDCR Refresh Sequence: move.l move.l move.l move.l #0xE10D0004, d0//Send first REF command d0, SDCR #0xE10D0004, d0//Send second REF command d0, SDCR Write Mode Register and Clear Reset DLL: MCF548x Reference Manual, Rev. 5 18-34 Freescale Semiconductor SDRAM Example move.l move.l #0x008D0000, d0//Write LMR and clear reset DLL d0, SDMR Enable Auto Refresh and Lock SDMR: move.l move.l #0x710D0F00, d0//Enable auto refresh and clear MODE_EN d0, SDCR MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 18-35 MCF548x Reference Manual, Rev. 5 18-36 Freescale Semiconductor Chapter 19 PCI Bus Controller 19.1 Introduction This chapter details the operation of the PCI bus controller for the MCF548x device. The PCI Bus Arbiter is detailed in Chapter 20, “PCI Bus Arbiter Module.” 19.1.1 Block Diagram PCI Arbiter External REQ/GNT Comm Bus Req/Gnt PCI Controller Block XL Bus Slave Bus (IP Bus) Configuration PCI Controller Configuration Interface Master Bus Target Target Interface Master Bus/ Comm Bus Initiator Initiator Interface External PCI Bus Figure 19-1. PCI Block Diagram 19.1.2 Overview The peripheral component interface (PCI) bus is a high-performance bus with multiplexed address and data lines. It is especially suitable for high data-rate applications. The PCI controller module supports a 32-bit PCI initiator (master) and target interface. As a target, access to the internal XL bus is supported. As an initiator, the PCI controller is coupled directly to the XL bus (as a slave) and available on the communication subsystem as a multichannel DMA peripheral. The MCF548x contains PCI central resource functions such as the PCI Arbiter (Chapter 20, “PCI Bus Arbiter Module”) and PCI reset control. The PCI bus clock must be provided by an external source. It must be phase aligned and either equal to 1, 1/2, or 1/4 the frequency of the system clock. 19.1.3 Features The following PCI features are supported in the MCF548x: • Supports system clock: PCI clock frequency ratios 1:1, 2:1, and 4:1 • Uses external CLKIN as clock reference MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 19-1 • • • • • • • • • • • • Compatible with PCI 2.2 specification PCI initiator and target operation Fully synchronous design 32-bit PCI address bus PCI 2.2 Type 0 configuration space header Supports the PCI 16/8 clock rule PCI master multichannel DMA or CPU access to PCI bus Ideal transfer rates up to 266 Mbytes/sec. (66 MHz clock, 128 byte buffer) PCI to system bus address translation Target response is medium DEVSEL generation Initiator latency time-outs Automatic retry of target disconnects 19.2 External Signal Description Table 19-1. PCI Module External Signals Name Type Function MCF548x Reset PCIAD[31:0] I/O PCI Address Data Bus Tristate PCICXBE[3:0] I/O PCI Command/Bytes Enables Tristate PCIDEVSEL I/O PCI Device Select Tristate PCIFRAME I/O PCI Frame Tristate PCIIDSEL I PCI Initialization Device Select Tristate PCIIRDY I/O PCI Initiator Ready Tristate PCIPAR I/O PCI Parity Tristate I PCI Clock Toggling PCIPERR I/O PCI Parity Error Tristate PCIRESET O PCI Reset 0 PCISERR I/O PCI System Error Tristate PCISTOP I/O PCI Stop Tristate PCITRDY I/O PCI Target Ready Tristate CLKIN For detailed description of the PCI bus signals, see the PCI Local Bus Specification, Revision 2.2. 19.2.1 Address/Data Bus (PCIAD[31:0]) The PCIAD[31:0] lines are a time multiplexed address data bus. The address is presented on the bus during the address phase while the data is presented on the bus during one or more data phases. 19.2.2 Command/Byte Enables (PCICXBE[3:0]) The PCICXBE[3:0] lines are time multiplexed. The PCI command is presented during the address phase and the byte enables are presented during the data phase. Byte enables are active low. MCF548x Reference Manual, Rev. 5 19-2 Freescale Semiconductor External Signal Description 19.2.3 Device Select (PCIDEVSEL) The PCIDEVSEL signal is asserted active low when the PCI controller decodes that it is the target of a PCI transaction from the address presented on the PCI bus during the address phase. 19.2.4 Frame (PCIFRAME) The PCIFRAME signal is asserted active low by a PCI initiator to indicate the beginning of a transaction. It is deasserted when the initiator is ready to complete the final data phase. 19.2.5 Initialization Device Select (PCIIDSEL) The PCIIDSEL signal is asserted active high during a PCI Type 0 Configuration Cycle to address the PCI Configuration header. 19.2.6 Initiator Ready (PCIIRDY) The PCIIRDY signal is asserted active low to indicate that the PCI initiator is ready to transfer data. During a write operation, assertion indicates that the master is driving valid data on the bus. During a read operation, assertion indicates that the master is ready to accept data. 19.2.7 Parity (PCIPAR) The PCIPAR signal indicates the parity on the PCIAD[31:0] and PCICXBE[3:0] lines. 19.2.8 PCI Clock (CLKIN) The CLKIN signal serves as a reference clock for generation of the internal PCI clock. For more information, see Section 19.4.7, “PCI Clock Scheme.” 19.2.9 Parity Error (PCIPERR) The PCIPERR signal is asserted active low when a data phase parity error is detected if enabled. 19.2.10 Reset (PCIRESET) The PCIRESET signal is asserted active low by the PCI controller to reset the PCI bus. This signal is asserted after MCF548x reset and must be negated to enable usage of the PCI bus. 19.2.11 System Error (PCISERR) The PCISERR signal, if enabled, is asserted active low when an address phase parity error is detected. 19.2.12 Stop (PCISTOP) The PCISTOP signal is asserted active low by the currently addressed target to indicate that it wishes to stop the current transaction. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 19-3 19.2.13 Target Ready (PCITRDY) The PCITRDY signal is asserted active low by the currently addressed target to indicate that it is ready to complete the current data phase. 19.3 Memory Map/Register Definition The MCF548x has several sets of registers that control and report status for the different interfaces to the PCI controller: PCI Type 0 configuration space registers, general status/control registers, and communication subsystem interface registers. All of these registers are accessible as offsets of MBAR. As an XL bus master, an external PCI bus master can access MBAR space for register updates. PCIRESET is controlled by a bit in the register space, PCIGSCR[PR], and must first be cleared before external PCI devices wake-up. In other words, an external PCI master cannot load configuration software across the PCI bus until this bit is cleared by software. Access to all internal registers is supported regardless of the value held in PCIGSCR[PR]. All registers are accessible at an offset of MBAR in the memory space. There are two module offsets for PCI configuration space. One is allocated to the communication subsystem interface registers and the other to all other PCI controller registers including the standard Type 0 PCI configuration space. Software reads from unimplemented registers return 0x00000000 and writes have no effect. Table 19-2. PCI Memory Map Address Name Size Description Access PCI Type 0 Configuration Registers MBAR + 0xB00 PCIIDR 32 PCI Device ID/Vendor ID R MBAR + 0xB04 PCISCR 32 PCI Status/Command R/W MBAR + 0xB08 PCICCRIR 32 PCI Class Code/Revision ID R MBAR + 0xB0C PCICR1 32 PCI Configuration 1 Register R/W MBAR + 0xB10 PCIBAR0 32 PCI Base Address Register 0 R/W MBAR + 0xB14 PCIBAR1 32 PCI Base Address Register 1 R/W MBAR + 0xB18–0xB24 — — Reserved — MBAR + 0xB28 PCICCPR 32 PCI Cardbus CIS Pointer R/W MBAR + 0xB2C PCISID 32 Subsystem ID/Subsystem Vendor ID R/W MBAR + 0xB30 PCIERBAR 32 PCI Expansion ROM R/W MBAR + 0xB34 PCICPR 32 PCI Capabilities Pointer R/W MBAR + 0xB38 — — Reserved — MBAR + 0xB3C PCICR2 32 PCI Configuration Register 2 R/W MBAR + 0xB40–0xB5C — — Reserved — 32 Global Status/Control Register R/W 32 Target Base Address Translation Register 0 R/W General Control/Status Registers MBAR + 0xB60 MBAR + 0xB64 PCIGSCR PCITBATR0 MCF548x Reference Manual, Rev. 5 19-4 Freescale Semiconductor Memory Map/Register Definition Table 19-2. PCI Memory Map (Continued) Address MBAR + 0xB68 MBAR + 0xB6C MBAR + 0xB70 MBAR + 0xB74 MBAR + 0xB78 Name PCITBATR1 PCITCR PCIIW0BTAR PCIIW1BTAR PCIIW2BTAR Size Description Access 32 Target Base Address Translation Register 1 R/W 32 Target Control Register R/W 32 Initiator Window 0 Base/Translation Address Register R/W 32 Initiator Window 1 Base/Translation Address Register R/W 32 Initiator Window 2 Base/Translation Address Register R/W MBAR + 0xB7C — — Reserved — MBAR + 0xB80 PCIIWCR 32 Initiator Window Configuration Register R/W MBAR + 0xB84 PCIICR 32 Initiator Control Register R/W MBAR + 0xB88 PCIISR 32 Initiator Status Register R/W MBAR + 0xB8C–0xBF4 — — Reserved — MBAR + 0xBF8 PCICAR 32 Configuration Address Register R/W MBAR + 0xBFC — — Reserved — CommBus FIFO Transmit Interface Registers1 MBAR + 0x8400 PCITPSR 32 Tx Packet Size Register R/W MBAR + 0x8404 PCITSAR 32 Tx Start Address Register R/W MBAR + 0x8408 PCITTCR 32 Tx Transaction Control Register R/W MBAR + 0x840C PCITER 32 Tx Enables Register R/W MBAR + 0x8410 PCITNAR 32 Tx Next Address Register R MBAR + 0x8414 PCITLWR 32 Tx Last Word Register R MBAR + 0x8418 PCITDCR 32 Tx Done Counts Register R MBAR + 0x841C PCITSR 32 Tx Status Register R/WC MBAR + 0x8420–0x843C — — Reserved — MBAR + 0x8440 PCITFDR 32 Tx FIFO Data Register R/W MBAR + 0x8444 PCITFSR 32 Tx FIFO Status Register R/WC MBAR + 0x8448 PCITFCR 32 Tx FIFO Control Register R/W MBAR + 0x844C PCITFAR 32 Tx FIFO Alarm Register R/W MBAR + 0x8450 PCITFRPR 32 Tx FIFO Read Pointer Register R/W MBAR + 0x8454 PCITFWPR 32 Tx FIFO Write Pointer Register R/W MBAR + 0x8458–0x847C — — Reserved — MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 19-5 Table 19-2. PCI Memory Map (Continued) Address Name Size Description Access CommBus FIFO Receive Interface Registers1 1 MBAR + 0x8480 PCIRPSR 32 Rx Packet Size Register R/W MBAR + 0x8484 PCIRSAR 32 Rx Start Address Register R/W MBAR + 0x8488 PCIRTCR 32 Rx Transaction Control Register R/W MBAR + 0x848C PCIRER 32 Rx Enables Register R/W MBAR + 0x8490 PCIRNAR 32 Rx Next Address Register R MBAR + 0x8494 — — Reserved — MBAR + 0x8498 PCIRDCR 32 Rx Done Counts Register R MBAR + 0x849C PCIRSR 32 Rx Status Register R/WC MBAR + 0x84A0–0x84BC — — Reserved — MBAR + 0x84C0 PCIRFDR 32 Rx FIFO Data Register R/W MBAR + 0x84C4 PCIRFSR 32 Rx FIFO Status Register R/WC MBAR + 0x84C8 PCIRFCR 32 Rx FIFO Control Register R/W MBAR + 0x84CC PCIRFAR 32 Rx FIFO Alarm Register R/W MBAR + 0x84D0 PCIRFRPR 32 Rx FIFO Read Pointer Register R/W MBAR + 0x84D4 PCIRFWPR 32 Rx FIFO Write Pointer Register R/W MBAR + 0x84D8–0x84FC — — Reserved — The PCI controller has separate control registers for transmit and receive operations via the communication subsystem DMA. See Section 19.3.3, “Communication Subsystem Interface Registers” for more information on these registers. 19.3.1 PCI Type 0 Configuration Registers The PCI controller supplies a type 0 PCI configuration space header. These registers are accessible as an offset from MBAR or through externally mastered PCI configuration cycles. PCI Dword Reserved space (0x10–0x3F) can be accessed only from external PCI configuration accesses. MCF548x Reference Manual, Rev. 5 19-6 Freescale Semiconductor Memory Map/Register Definition 19.3.1.1 Device ID/Vendor ID Register (PCIIDR)—PCI Dword Addr 0 31 30 29 28 27 26 25 R 24 23 22 21 20 19 18 17 16 Device ID W Reset 0 1 0 1 1 0 0 0 0 0 0 0 0 1 1 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 1 0 1 0 1 1 1 R Vendor ID W Reset 0 0 0 1 0 0 0 Reg Addr 0 0 MBAR + 0xB00 Figure 19-2. Device ID/Vendor ID Register (PCIIDR) Table 19-3. PCIIDR Field Descriptions Bits Name 31–16 Device ID This field is read-only and represents the PCI Device Id assigned to the MCF548x. Its value is: 0x5806. 15–0 Vendor ID This field is read-only and represents the PCI Vendor Id assigned to the MCF548x. Its value is: 0x1057. 19.3.1.2 Description PCI Status/Command Register (PCISCR)—PCI Dword Addr 1 31 R 30 29 28 27 SE MA TR TS rwc1 rwc1 rwc1 rwc1 0 0 0 0 0 0 1 15 14 13 12 11 10 0 0 0 0 0 0 0 0 0 0 PE 1 W rwc Reset R 26 25 24 23 22 21 20 19 18 17 16 DP FC R 66M C 0 0 0 0 0 1 0 1 0 0 0 0 0 9 8 7 6 5 4 3 2 1 0 0 F S ST PER V MW SP B M IO 0 0 0 0 0 0 0 0 0 0 0 DT rwc1 W Reset Reg Addr MBAR + 0xB04 1 Bits 31-27 and 24 are read-write-clear (rwc). —Hardware can set rwc bits, but cannot clear them. —Only PCI configuration cycles can clear rwc bits that are currently set by writing a 1 to the bit location. Writing a 1 to a rwc bit that is currently a 0 or writing a 0 to any rwc bit has no effect. Figure 19-3. PCI Status/Command Register (PCISCR) MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 19-7 Table 19-4. PCISCR Field Descriptions Bits Name Description 31 PE Parity error detected. This bit is set when a parity error is detected, even if the PCISCR[PER] is cleared. This bit is cleared by a PCI configuration cycle writing a ‘1’ to the bit. Writing ‘0’ has no effect. 30 SE System error signalled. This bit is set whenever the PCI controller generates a PCI system error on the PCISERR line. This bit is cleared by a PCI configuration cycle writing a ‘1’ to the bit. Writing ‘0’ has no effect. 29 MA Master abort received. This bit is set whenever the PCI controller is the PCI master and terminates a transaction (except for a special cycle) with a master-abort. This bit is cleared by a PCI configuration cycle writing a ‘1’ to the bit. Writing ‘0’ has no effect. 28 TR Target abort received. This bit is set whenever the PCI controller is the PCI master and a transaction is terminated by a target-abort from the currently addressed target. This bit is cleared by a PCI configuration cycle writing a ‘1’ to the bit. Writing ‘0’ has no effect. 27 TS Target abort signalled. This bit is set whenever the PCI controller is the target and it terminates a transaction with a target-abort. This bit is cleared by a PCI configuration cycle writing a ‘1’ to the bit. Writing ‘0’ has no effect. 26–25 DT DEVSEL timing. Fixed to ‘01’. These bits encode a medium DEVSEL timing. This defines the slowest DEVSEL timing as meduim timing when the PCI controller is the target (except configuration accesses). 24 DP Master data parity error. This bit applies only when the PCI controller is the master and is set only if the following conditions are met: • The PCI controller-as-master sets PERR itself during a read or the PCI controller-as-master detected it asserted by the target during a write • The PCISCR[PER] bit is set This bit is cleared by a PCI configuration cycle writing a ‘1’ to the bit. Writing ‘0’ has no effect. 23 FC Fast back-to-back capable. Fixed to 1. This read-only bit indicates that the PCI controller as target is capable of accepting fast back-to-back transactions with other targets. 22 R 21 66M 20 C Capabilities list. Fixed to 0. This bit indicates that the PCI controller does not implement the New Capabilities List Pointer Configuration Register in DWORD 13 of the configuration space. 19–10 — Reserved, should be cleared. 9 F Fast back-to-back transfer enable. This bit controls whether or not the PCI controller as master can do fast back-to-back transactions to different devices. Initialization software should set this bit if all targets are fast back-to-back capable. 0 Fast back-to-back transactions are only allowed to the same device 1 The master is allowed to generate fast back-to-back transactions to different devices. 8 S SERR enable. This bit is an enable bit for the PCISERR driver. 0 PCISERR driver disabled 1 PCISERR driver enabled Note: Address parity errors are reported only if this bit and bit 6 are set. 7 ST Reserved. Fixed to 0. Prior to the 2.2 PCI Spec, this was the UDF (user defined features) supported bit. 0 Does not support UDF 1 Supported user defined features 66 MHz capable. Fixed to 1. This bit indicates that the PCI controller is 66 MHz capable. Address and data stepping. Fixed to 0. This bit indicates that the PCI controller never uses address/data stepping. Initialization software should write a 0 to this bit location. MCF548x Reference Manual, Rev. 5 19-8 Freescale Semiconductor Memory Map/Register Definition Table 19-4. PCISCR Field Descriptions (Continued) Bits Name Description 6 PER Parity error response. This bit controls the device’s response to parity errors. 0 The device sets its Parity Error status bit (bit 31) in the event of a parity error, but does not assert PERR. 1 When a parity error is detected, the PCI controller asserts PERR 5 V VGA palette snoop enable. Fixed to 0. This bit indicates that the PCI controller is not VGA compatible. Initialization software should write a 0 to this bit location. 4 MW Memory write and invalidate enable. This bit is an enable for using the memory write and invalidate command. 0 Only memory write command can be used 1 PCI controller-as-master may generate the memory write and invalidate command. 3 SP Special cycle monitor or ignore. This bit is to determine whether or not to ignore PCI Special Cycles. Since PCI controller-as-target does not recognize messages delivered via the Special Cycle operation, a value of 1 should never be programmed to this register. This bit, however, is programmable (read/write from both the IP bus and PCI bus Configuration cycles). 2 B Bus master enable. This bit indicates whether or not the PCI controller has the ability to serve as a master on the PCI bus. A value of 1 indicates this ability is enabled. If the PCI controller is used as a master on the PCI bus (via the XL bus or comm bus), a 1 should be written to this bit during initialization. If the value of the register is 0, it will not inhibit mastered transactions. This bit is meant to be read by configuration software. 1 M Memory access control. This bit controls the PCI controller’s response to memory space accesses. 0 The PCI controller does not recognize memory accesses 1 The PCI controller recognizes memory accesses. 0 IO I/O access control. Fixed to 0. This bit is not implemented because there is no PCI controller I/O type space accessible from the PCI bus. The PCI base address registers are memory address ranges only. Initialization software should write a 0 to this bit location. 19.3.1.3 Revision ID/Class Code Register (PCICCRIR)—PCI Dword 3 31 30 29 28 27 26 25 R 24 23 22 21 20 19 18 17 16 Class Code W Reset 0 0 0 0 0 1 1 0 1 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 0 0 R Class Code Revision ID W Reset 0 0 Reg Addr 0 0 0 0 0 0 0 0 0 0 0 MBAR + 0xB08 Figure 19-4. Revision ID/Class Code Register (PCICCRIR) MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 19-9 Table 19-5. PCICCRIR Field Descriptions Bits Name Description 31–8 Class Code This field is read-only and represents the PCI Class Code assigned to processor. Its value is: 0x06 8000. (Other bridge device). 7–0 Revision ID This field is read-only and represents the PCI Revision ID for this version of the processor. Its value is: 0x00. 19.3.1.4 Configuration 1 Register (PCICR1)—PCI Dword 3 31 30 29 28 R 27 26 25 24 23 22 21 20 19 18 17 16 Header Type BIST W Reset 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 R Lat Timer [7:3] Lat Timer [2:0] Cache Line Size [7:4] Cache Line Size [3:0] 0 0 0 W Reset 0 0 0 Reg Addr 0 0 0 0 0 0 0 0 0 0 MBAR + 0xB0C Figure 19-5. Configuration 1 Register (PCICR1) Table 19-6. PCICR1 Field Descriptions Bits Name Description 31–24 BIST Built in self test. Fixed to 0x00. The PCI controller does not implement the Built-In Self Test register. Initialization software should write a 0x00 to this register location. 23–16 Header Type 15–11 Lat Timer 10-8 7–4 3–0 Header type. Fixed to 0x00. The PCI controller implements a Type 0 PCI configuration space Header. Initialization software should write a 0x00 to this register location. Latency timer [7:3]. This register contains the latency timer value, in PCI clocks, used when the PCI controller is the PCI master. The upper five bits are programmable. Latency timer must be programmed to a non-zero value before the PCI Controller will operate as master of the PCI bus. Latency timer [2:0] The lower three bits of the register are hardwired low Cache Line Cache line size[7:4] Specifies the cache line size in units of DWORDs. The higher four bits Size of the register are hardwired low Cache line size [3:0] Specifies the cache line size in units of DWORDs. MCF548x Reference Manual, Rev. 5 19-10 Freescale Semiconductor Memory Map/Register Definition 19.3.1.5 Base Address Register 0 (PCIBAR0)—PCI Dword 4 31 30 29 28 27 26 R 25 24 23 22 21 20 19 18 BAR 0 17 16 0 0 W Reset 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 0 0 0 0 0 0 0 0 0 0 0 PREF 0 0 0 0 0 0 0 0 0 0 0 0 0 R RANGE IO/M# W Reset Reg Addr 0 0 0 MBAR + 0xB10 Figure 19-6. Base Address Register 0 (PCIBAR0) Table 19-7. PCIBAR0 Field Descriptions Bits Name Description 31–18 BAR0 Base address register 0. PCI base address register 0 (256 Kbyte). Applies only when processor is target. These bits are programmable (read/write from both the IP bus and PCI bus Configuration cycles). 17–4 — 3 PREF Prefetchable access. Fixed to 0. This bit indicates that the memory space defined by BAR0 is not prefetchable. Configuration software should write a 0 to this bit location. 2–1 RANGE Fixed to 00. This register indicates that base address 0 is 32 bits wide and can be mapped anywhere in 32-bit address space. Configuration software should write 00 to these bit locations. 0 IO/M# Reserved, should be cleared. IO or memory space. Fixed to 0. This bit indicates that BAR0 is for memory space. Configuration software should write a 0 to this bit location. 0 Memory 1 I/O MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 19-11 19.3.1.6 Base Address Register 1 (PCIBAR1)—PCI Dword 5 31 R 30 BAR1 29 28 27 26 25 24 23 22 21 20 19 18 17 16 0 0 0 0 0 0 0 0 0 0 0 0 0 0 W Reset 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 0 0 0 0 0 0 0 0 0 0 0 PREF 0 0 0 0 0 0 0 0 0 0 0 0 1 R RANGE IO/M# W Reset Reg Addr 0 0 0 MBAR + 0xB14 Figure 19-7. Base Address Register 1 (PCIBAR1) Table 19-8. PCIBAR1 Field Descriptions Bits Name 31–30 BAR1 29–4 — 3 PREF Prefetchable access. Fixed to 1. This bit indicates that the memory space defined by BAR1 is prefetchable. Configuration software should write a 1 to this bit location. 2–1 RANGE Fixed to 00. This register indicates that base address 1 is 32 bits wide and can be mapped anywhere in 32-bit address space. Configuration software should write 00 to these bit locations. 0 IO/M# 19.3.1.7 Description Base address register 1. Processo PCI base address register 1 (1 Gbyte). Applies only when the processor is target. These bits are programmable (read/write from both the IP bus and PCI bus Configuration cycles). Reserved, should be cleared. IO or memory space. Fixed to 0. This bit indicates that BAR1 is for memory space. Configuration software should write a 0 to this bit location. 0 Memory 1 I/O CardBus CIS Pointer Register PCICCPR—PCI Dword A This optional register contains the pointer to the Card Information Structure (CIS) for the CardBus card. All 32 bits of the register are programmable by the slave bus. From the PCI bus, this register can only be read, not written. Its reset value is 0x0000 0000 and is accessible at address MBAR + 0xB28. 19.3.1.8 Subsystem ID/Subsystem Vendor ID Registers PCISID—PCI Dword B The Subsystem Vendor ID register contains the 16-bit manufacturer identification number of the add-in board or subsystem that contains this PCI device. The Subsystem ID register contains the 16-bit subsystem identification number of the add-in board or subsystem that contains this PCI device. A value of zero in these registers indicates there isn’t a Subsystem Vendor and Subsystem ID associated with the device. If used, software must write to these registers before any PCI bus master reads them. MCF548x Reference Manual, Rev. 5 19-12 Freescale Semiconductor Memory Map/Register Definition All 32 bits of the register are programmable by the slave bus. From the PCI bus, this register can only be read, not written. The reset value is 0x0000_0000 and is accessible at address MBAR + 0xB2C. 19.3.1.9 Expansion ROM Base Address PCIERBAR—PCI Dword C Not implemented. Fixed to 0x0000_0000 at address MBAR + 0xB30. 19.3.1.10 Capabilities Pointer (Cap_Ptr) PCICPR—PCI Dword D Not implemented. Fixed to 0x00 at address MBAR + 0xB34. 19.3.1.11 Configuration 2 Register (PCICR2)—PCI Dword F 31 30 29 R 28 27 26 25 24 23 22 21 Max_Lat 20 19 18 17 16 Min_Gnt W Reset 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 0 0 R Interrupt Pin Interrupt Line W Reset 0 0 Reg Addr 0 0 0 0 0 0 0 0 0 0 0 MBAR + 0xB3C Figure 19-8. Configuration 2 Register (PCICR2) Table 19-9. PCICR2 Field Descriptions Bits Name Description 31–24 Max_Lat Maximum latency. Specifies how often, in units of 1/4 microseconds, the PCI controller would like to have access to the PCI bus as master. A value of zero indicates the device has no stringent requirement in this area. The register is read/write to/from the slave bus, but read only from the PCI bus. 23–16 Min_Gnt Minimum grant. The value programmed to this register indicates how long the PCI controller as master would like to retain PCI bus ownership whenever it initiates a transaction. The register is programmable from the slave bus, but read only from the PCI bus. 15–8 Interrupt Pin Fixed to 0x00. Indicates that this device does not use an interrupt request pin. 7–0 Interrupt Line Fixed to 0x00. The Interrupt Line register stores a value that identifies which input on a PCI interrupt controller the function’s PCI interrupt request pin. Since no interrupt request pin is used, as specified in the Interrupt Pin register, this register has no function. 19.3.2 General Control/Status Registers The general control/status registers primarily address the configurability of the XL bus initiator and target interfaces, though some also address global options which affect the multichannel DMA interface. These MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 19-13 registers are accessed primarily internally as offsets of MBAR, but can also be accessed by an external PCI master if PCI base and target base address registers are configured to access the space. See Section 19.5.2, “Address Maps,” on configuring address windows. 19.3.2.1 R Global Status/Control Register (PCIGSCR) 31 30 29 28 27 0 0 PE SE 0 rwc1 rwc1 W Reset R 26 25 24 23 22 21 20 19 18 17 16 XLB2CLKIN 0 0 0 0 0 Reserved —2 0 0 0 0 0 Uninitialized 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 0 PEE SEE 0 0 0 0 0 0 0 0 0 0 0 PR 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 W Reset Reg Addr MBAR + 0xB60 1 Bits 29 and 28 are read-write-clear (rwc). —Hardware can set rwc bits, but cannot clear them. —Software can clear rwc bits that are currently set by writing a 1 to the bit location. Writing a 1 to a rwc bit that is currently a 0 or writing a 0 to any rwc bit has no effect. 2 The reset value of bits 26-24 and 18-16 is determined by the PLL multiplier. Figure 19-9. Global Status/Control Register (PCIGSCR) Table 19-10. PCIGSCR Field Descriptions Bits Name 31–30 — Reserved, should be cleared. 29 PE PERR detected. This bit is set when the PCI Parity Error line, PCIPERR, asserts (any device). A CPU interrupt will be generated if the PCIGSCR[PEE] bit is set. It is up to application software to clear this bit by writing ‘1’ to it. 28 SE SERR detected. This bit is set when a PCI System Error line, PCISERR, asserts (any device). A CPU interrupt will be generated if the PCIGSCR[SEE] bit is set. It is up to application software to clear this bit by writing ‘1’ to it. 27 — Reserved, should be cleared. 26–24 23–19 18–16 15–14 Description XLB2CLKIN This bit field stores the XL bus clock to external PCI clock (CLKIN)divide ratio. This field is read-only and the reset value is determined by the PLL multiplier (either 1, 2, or 4). Software can read these bits to determine a valid ratio. If the register contains a differential value that does not reflect the PLL settings, the PCI controller could malfunction. — Reserved, should be cleared. CLKINReser This field is reserved. ved — Reserved, should be cleared. MCF548x Reference Manual, Rev. 5 19-14 Freescale Semiconductor Memory Map/Register Definition Table 19-10. PCIGSCR Field Descriptions (Continued) Bits Name Description 13 PEE Parity error interrupt enable. This bit enables CPU Interrupt generation when the PCI Parity Error signal, PCIPERR, is sampled asserted. When enabled and PCIPERR asserts, software must clear the PE status bit to clear the interrupt condition. 12 SEE System error interrupt enable. This bit enables CPU Interrupt generation when a PCI system error is detected on the PCISERR line. When enabled and PCISERR asserts, software must clear the SE status bit to clear the interrupt condition. 11–1 — Reserved, should be cleared. 0 PR PCI reset. This bit controls the external PCIRESET. When this bit is cleared, the external PCIRESET deasserts. Setting this bit does not reset the internal PCI controller. The application software must not initiate PCI transactions while this bit is set. It is recommended that this bit be programmed last during initialization. The reset value of the bit is 1 (PCIRESET asserted). 19.3.2.2 Target Base Address Translation Register 0 (PCITBATR0) 31 30 29 28 27 R 26 25 24 23 22 21 20 19 18 Base Address Translation 0 17 16 0 0 W Reset R 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 EN 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 W Reset Reg Addr MBAR + 0xB64 Figure 19-10. Target Base Address Translation Register 0 (PCITBATR0) Table 19-11. PCITBATR0 Field Descriptions Bits 31–18 Name Description Base This base address register corresponds to a hit on the BAR0 in MCF548x PCI Type 0 Configuration Address space register from PCI space. When there is a hit on MCF548x PCI BAR0 (MCF548x as Target), Translation 0 the upper 14 bits of the address (256-Kbyte boundary) are written over by this register value to address some space in MCF548x. In normal operation, this value should be written during the initialization sequence only. 17–1 — 0 Enable 0 Reserved, should be cleared. This bit enables a transaction in BAR0 space. If this bit is zero and a hit on MCF548 PCIBAR0 occurs, the target interface gasket will abort the PCI transaction. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 19-15 19.3.2.3 Target Base Address Translation Register 1 (PCITBATR1) 31 30 R Base Address Translation 1 W Reset R 29 28 27 26 25 24 23 22 21 20 19 18 17 16 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 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 EN 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 W Reset Reg Addr MBAR + 0xB68 Figure 19-11. Target Base Address Translation Register 1 (PCITBATR1) Table 19-12. PCITBATR1 Field Descriptions Bits Name Description 31–30 Base Address Translation 1 This base address register corresponds to a hit on the BAR1 in MCF548 PCI Type 0 Configuration space register (PCI space). When there is a hit on MCF548 PCI BAR1 (MCF548 as Target), the upper 2 bits of the address (1-Gbyte boundary) are written over by this register value to address some 1-Gbyte space in MCF548. This register can be reprogrammed to move the window of MCF548 address space accessed during a hit in PCIBAR1. 29–1 — Reserved, should be cleared. 0 EN This bit enables a transaction in BAR1 space. If this bit is zero and a hit on MCF548 PCI BAR1 occurs, the target interface gasket will abort the PCI transaction. 19.3.2.4 R Target Control Register (PCITCR) 31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 0 0 0 0 0 0 0 LD 0 0 0 0 0 0 0 P 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 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 0 0 W Reset R W Reset Reg Addr MBAR + 0xB6C Figure 19-12. Target Control Register (PCITCR) MCF548x Reference Manual, Rev. 5 19-16 Freescale Semiconductor Memory Map/Register Definition Table 19-13. PCITCR Field Descriptions Bits Name 31–25 — Reserved, should be cleared. 24 LD Latency rule disable. This control bit applies only when MCF548 is Target. When set, it prevents the PCI Controller from automatically issuing a retry disconnect due to the PCI 16/8 clock rule. This bit should only be set when the XL<->PCI path is not in use. The only transactions that are retried on the XL bus by the PCI are reads. Writes are held on the XL bus until either all data is posted (PCI memory writes) and the XL bus data tenure is normally terminated or, in the case of I/O writes to PCI, access is granted to the PCI bus and the connected write completes. When the LD bit is set, there is never a timeout on the PCI bus because the PCI 16/8 clock rule is not obeyed. If there is inbound PCI traffic (PCI->MCF548) and an XL bus write is held open by the PCI Controller, the PCI traffic will not be granted access to XL bus. This is true for reads that have not been prefetched and when the inbound write buffer is full. Both buses hang. Normal operation relies on the LD bit being cleared. If used, the bit must be set before the 15th PCI clock for the first transfer and before the 7th clock for other transfers. 23–17 — Reserved, should be cleared. 16 P Prefetch reads. This bit controls fetching a line from memory in anticipation of a request from the external master. The target interface will continue to prefetch lines from memory as long as PCIFRAME is asserted and there is space to store the data in the target read buffer. Note: This bit only applies to PCI reads in the address range for BAR 1 (prefetchable memory). Note: Prefetching is performed in response to a PCI memory-read-multiple command even if this bit is cleared. 15–0 — Reserved, should be cleared. 19.3.2.5 31 Description Initiator Window 0 Base/Translation Address Register (PCIIW0BTAR) 30 R 29 28 27 26 25 24 23 22 Window 0 Base Address 21 20 19 18 17 16 Window 0 Address Mask W Reset 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 R Window 0 Translation Address W Reset 0 0 Reg Addr 0 0 0 0 0 0 MBAR + 0xB70 Figure 19-13. Initiator Window 0 Base/Translation Address Register (PCIIW0BTAR) MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 19-17 Table 19-14. PCIIW0BTAR Field Descriptions Bits Name Description 31–24 Window 0 Base Address One of three base address registers to determine an XL bus hit on PCI. At most, the upper byte of the address is decoded. The Window 0 Address Mask register determines what bits of this register to compare the XL bus address against to generate the hit. The smallest possible Window is a 16-Mbyte block. 23–16 Window 0 Address Mask The Window 0 Address Mask Register masks the corresponding XL bus base address bit of the base address for Window 0 (Window 0 Base Address) to instruct the address decode logic to ignore or “don’t care” the bit. If the base address mask bit is set, the associated base address bit of Window 0 is ignored when generating the PCI hit. Bit 16 masks bit 24, bit 17 masks bit 25, and so on. 0 Corresponding address bit is used in address decode. 1 Corresponding address bit is ignored in address decode. For XL bus accesses to Window 0 address range, this byte also determines which upper 8 bits of the XL bus address to pass on for presentation as a PCI address. Any address bit used to decode the XL bus address, indicated by a “0”, will be translated. This provides a way to overlay a PCI page address onto the XL bus address. A “1” in the Address Mask byte indicates that the XL bus address bit will be passed to PCI unaltered. 15–8 Window 0 For any translated bit (described above), the corresponding value here will be driven onto the PCI Translation address bus for the XL bus Window 0 address hit. Address The Window Translation operation can not be turned off. If a direct mapping from XL bus to PCI space is desired, program the same value to both the Window Base Address Register and Window Translation Address Register. 7–0 — 19.3.2.6 31 Reserved, should be cleared. Initiator Window 1 Base/Translation Address Register (PCIIW1BTAR) 30 R 29 28 27 26 25 24 23 22 Window 1 Base Address 21 20 19 18 17 16 Window 1 Address Mask W Reset 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 R Window 1 Translation Address W Reset Reg Addr 0 0 0 0 0 0 0 0 MBAR + 0xB74 Figure 19-14. Initiator Window 1 Base/Translation Address Register (PCIIW1BTAR) The field descriptions for this register are the same as for PCIIW0BTAR, except that they apply to Window 1. MCF548x Reference Manual, Rev. 5 19-18 Freescale Semiconductor Memory Map/Register Definition 19.3.2.7 31 Initiator Window 2 Base/Translation Address Register (PCIIW2BTAR) 30 R 29 28 27 26 25 24 23 22 Window 2 Base Address 21 20 19 18 17 16 Window 2 Address Mask W Reset 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 R Window 2 Translation Address W Reset 0 0 0 0 0 0 0 Reg Addr 0 MBAR + 0xB78 Figure 19-15. Initiator Window 2 Base/Translation Address Register (PCIIW2BTAR) The field descriptions for this register are the same as for PCIIW0BTAR, except that they apply to Window 2. 19.3.2.8 R Initiator Window Configuration Register (PCIIWCR) 31 30 29 28 27 26 25 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 0 0 0 0 0 0 0 0 24 23 22 21 20 0 0 0 0 0 0 0 0 0 0 0 0 0 8 7 6 5 4 3 2 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Window 0 Control 19 18 17 16 Window 1 Control W Reset R Window 2 Control W Reset Reg Addr 0 0 0 0 MBAR + 0xB80 Figure 19-16. Initiator Window Configuration Register (PCIIWCR) MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 19-19 Table 19-15. PCIIWCR Field Descriptions Bits Name 31–28 — 27–24 Description Reserved, should be cleared. Window 0 Bit[3]—IO/M#. Control[3:0] 0 Window is mapped to PCI memory. 1 Window is mapped to PCI I/O. Bit[2:1]—PCI read command (PRC). If bit[3] is programmed memory, “0”, then these bits are used to determine the type of PCI memory command to issue. See Table 19-57. If bit[3] is set to “1”, the value of these bits is meaningless. 00 PCI Memory Read. 01 PCI Memory Read Line. 10 PCI Memory Read Multiple. 11 Reserved. Bit[0]—Enable. This bit is set to indicate the address registers that control the XL bus initiator interface access to PCI initialized and will be used. The PCI Controller can begin to decode XL bus PCI accesses. 0 Do not decode XL bus PCI accesses to Window. 1 Registers initialized—decode accesses to Window. 23–20 19–16 — Window 1 Bit[3]—IO/M#. Control[3:0] Bit[2:1]—PRC. Bit[0]—Enable. 15–12 11–8 Reserved Reserved register. Write a zero to this register. Window 2 Bit[3]—IO/M#. Control[3:0] Bit[2:1]—PRC. Bit[0]—Enable. 7–0 — 19.3.2.9 R Reserved, should be cleared. Reserved, should be cleared. Initiator Control Register (PCIICR) 31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 0 0 0 0 0 REE IAE TAE 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 W Reset R Maximum Retries W Reset Reg Addr 1 1 1 1 1 MBAR + 0xB84 Figure 19-17. Initiator Control Register (PCIICR) MCF548x Reference Manual, Rev. 5 19-20 Freescale Semiconductor Memory Map/Register Definition Table 19-16. PCIICR Field Descriptions Bits Name Description 31–27 — 26 REE Retry error enable. This bit enables CPU Interrupt generation in the case of Retry Error termination of a transaction. It may be desirable to mask CPU interrupts, but in such a case, software should poll the status bits to prevent a possible lock-up condition. 25 IAE Initiator abort enable. This bit enables CPU Interrupt generation in the case of Initiator Abort termination of a transaction. It may be desirable to mask CPU interrupts, but in such a case, software should poll the status bits to prevent a possible lock-up condition. 24 TAE Target abort enable. This bit enables CPU Interrupt generation in the case of Target Abort termination of a transaction. It may be desirable to mask CPU interrupts, but in such a case, software should poll the status bits to prevent a possible lock-up condition. 23–8 — 7–0 Maximum Retries Reserved, should be cleared. Reserved, should be cleared. This bit field controls the maximum number of automatic PCI retries or master latency time-outs to permit per write transaction. The retry counter is reset at the beginning of each write transaction (i.e. it is not cumulative). Setting the Maximum Retries to 0x00 allows infinite automatic retry cycles and latency time-outs before the write transaction will abort and, if open, send back an error on XL bus. A slow or malfunctioning Target might issue infinite retry disconnects or hold the data tenure open indefinitely, and therefore, permanently tie up the PCI bus if no Target Abort occurs. The Maximum Retries register does not apply to reads because reads are always ARTRY’d on XL bus when retry-terminated by the PCI target. This is done to avoid livelock scenarios where the device we are requesting read data from needs to flush itself of posted writes going to MCF548 before it can return the read data. The incoming writes cannot be blocked in this case. 19.3.2.10 Initiator Status Register (PCIISR) R 31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 0 0 0 0 0 RE IA TA 0 0 0 0 0 0 0 0 rwc1 rwc1 rwc1 W Reset R 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 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 0 0 W Reset Reg Addr MBAR + 0xB88 1 Bits 26-24 are read-write-clear (rwc). —Hardware can set rwc bits, but cannot clear them. —Software can clear rwc bits that are currently set by writing a 1 to the bit location. Writing a 1 to a rwc bit that is currently a 0 or writing a 0 to any rwc bit has no effect. Figure 19-18. Initiator Status Register (PCIISR) MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 19-21 Table 19-17. PCIISR Field Descriptions Bits Name Description 31–27 — Reserved, should be cleared. 26 RE Retry error. This flag is set when the controller ARTRY’s a read on XL bus when retry-terminated by the PCI target or when the Max_Retries limit is reached for a single XL bus write transaction. A CPU interrupt will be generated if PCIICR[RE] bit is set. It is up to application software to clear this bit by writing ‘1’ to it. 25 IA Initiator abort. This flag bit is set if the PCI controller issues an Initiator Abort flag. This indicates that no Target responded by asserting DEVSEL within the time allowed for subtractive decoding. A CPU interrupt will be generated if the PCIICR[IAE] bit is set. It is up to application software to clear this bit by writing ‘1’ to it. 24 TA Target abort. This flag bit is set if the addressed PCI Target has signalled an Abort. A CPU interrupt will be generated if the PCIICR[TAE] bit is set. It is up to application software to query the Target’s status register and determine the source of the error. It is up to application software to clear this bit by writing ‘1’ to it. 23–0 — Reserved, should be cleared. 19.3.2.11 Configuration Address Register (PCICAR) R 31 30 29 28 27 26 25 24 23 22 21 20 E 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 19 18 17 16 0 0 0 0 3 2 1 0 0 0 0 0 Bus Number W Reset R Device Number Function Number DWORD W Reset 0 0 0 0 0 0 0 Reg Addr 0 0 0 0 0 0 0 MBAR + 0xBF8 Figure 19-19. Configuration Address Register (PCICAR) Table 19-18. PCICAR Field Descriptions Bits Name Description 31 E Enable. The enable flag that controls configuration space mapping. When enabled, subsequent access to initiator window space defined as I/O in the PCIIWCR is translated into a PCI configuration, special cycle, or interrupt acknowledge access using the configuration address register information (Section 19.4.4.2, “Configuration Mechanism”). When disabled, a read or write to the window is passed through to the PCI bus as an I/O transaction. 0 Disabled 1 Enabled 30–24 — Reserved, should be cleared. MCF548x Reference Manual, Rev. 5 19-22 Freescale Semiconductor Memory Map/Register Definition Table 19-18. PCICAR Field Descriptions (Continued) Bits Name Description 23–16 Bus Number This register field is an encoded value used to select the target bus of the configuration access. For target devices on the PCI bus connected to MCF548, this field should be set to 0x00. 15–11 Device Number This field is used to select a specific device on the target bus.Section 19.4.4.2, “Configuration Mechanism,” for more information. 10–8 Function Number This field is used to select a specific function in the requested device. Single-function devices should respond to function number ‘000’. 7–2 DWORD This field is used to select the Dword address offset in the configuration space of the target device. 1–0 — 19.3.3 Reserved, should be cleared. Communication Subsystem Interface Registers The communication subsystem/multichannel DMA interface has separate control registers for transmit and receive operations. 19.3.3.1 Comm Bus FIFO Transmit Interface PCI Tx is controlled by 14 32-bit registers. These registers are located at an offset from MBAR of 0x8400. Register addresses are relative to this offset. 19.3.3.1.1 31 Tx Packet Size Register (PCITPSR) 30 29 28 27 26 R 25 24 23 22 21 20 19 18 Packet_Size[15:2] 17 16 Packet_Size [1:0] W Reset R 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 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 0 0 W Reset Reg Addr MBAR + 0x8400 Figure 19-20. Tx Packet Size Register (PCITPSR) MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 19-23 Table 19-19. PCITPSR Field Descriptions Bits Name 31–18 Description Packet_Size Packet_Size [15:2]. The Packet_Size field indicates the number of bytes for the transmit controller to send over PCI. Only bits [15:2] are writable. Only 32-bit data transfers to the FIFO are allowed. Writing to this register also completes a Restart Sequence as long as the Master Enable bit, PCITER[ME], is high and Reset Controller bit, PCITER[RC], is low. 17–16 Packet_Size [1:0] The two low bits are hardwired low. 15–0 — 19.3.3.1.2 31 Reserved, should be cleared. Tx Start Address Register (PCITSAR) 30 29 28 27 26 25 R 24 23 22 21 20 19 18 17 16 Start_Add W Reset 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 0 0 0 0 0 0 R Start_Add W Reset 0 0 0 Reg Addr 0 0 0 0 0 0 MBAR + 0x8404 Figure 19-21. Tx Start Address Register (PCITSAR) Table 19-20. PCITSAR Field Descriptions Bits Name Description 31–0 Start_Add User writes the PCI address to be presented for the first DWORD of a PCI packet. The PCI Tx controller will track and calculate the necessary address for subsequent transactions. Addressing is assumed to be sequential from the start address unless the PCITTCR[DI] bit is set. This register will not increment as the PCI packet proceeds. MCF548x Reference Manual, Rev. 5 19-24 Freescale Semiconductor Memory Map/Register Definition 19.3.3.1.3 R Tx Transaction Control Register (PCITTCR) 31 30 29 28 27 26 25 24 23 22 21 18 17 16 0 0 0 0 0 0 0 0 0 1 1 1 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 0 0 0 0 0 0 0 W 0 0 0 DI 0 0 0 0 0 0 0 0 0 0 0 0 0 PCI_cmd 20 19 Max_Retries W Reset R Max_Beats W Reset 0 0 Reg Addr 0 MBAR + 0x8408 Figure 19-22. Tx Transaction Control Register (PCITTCR) Table 19-21. PCITTCR Field Descriptions Bits Name 31–28 — 27–24 PCI_cmd 23–16 Description Reserved, should be cleared. The user writes this field with the desired PCI command to present during the address phase of each PCI transaction. The default is Memory Write. This field is not checked for consistency and if written to an illegal value, unpredictable results will occur. If not using the default value, the user should write this register only once prior to any packet Restart. Max_Retries The user writes this field with the maximum number of retries to permit “per packet”. The retry counter is reset when the packet completes normally or is terminated by a master abort, target abort, or an abort due to exceeding the retry limit. A slow or malfunctioning Target might issue infinite disconnects and therefore permanently tie up the PCI bus. A finite (0x01 to 0xf) Max_Retries value will detect this condition and generate an interrupt. Setting Max_Retries to 0x00 will not generate an interrupt but will permit re-arbitration of the PCI bus between each disconnect. 15–11 — Reserved, should be cleared. 10–8 Max_Beats 7–5 — Reserved, should be cleared. 4 W Word transfer. The user writes this register to disable the two high byte enables of the PCI bus during write transactions initiated by this interface. The default setting is 0, enable all 4 byte enables. 3–1 — Reserved, should be cleared. 0 DI Disable address incrementing. The user writes this register to disable PCI address incrementing between transactions. The default setting is 0, increment address by 4 (4 byte data bus). The user writes this register with the desired number of PCI data beats to attempt on each PCI transaction. The default setting of 0 represents the maximum of eight beats per transaction. The transmit controller will wait until sufficient bytes are in the Transmit FIFO to support the indicated number of beats (NOTE: Each beat is four bytes). In the case that a packet is nearly complete and less than the Max_Beats number of bytes remain to complete the packet, the Transmit Controller will issue single-beat transactions automatically until the packet is finished. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 19-25 19.3.3.1.4 R Tx Enables Register (PCITER) 31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 RC RF 0 CM BE 0 0 ME 0 0 FEE SE RE TAE IAE NE 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 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 0 0 W Reset R W Reset Reg Addr MBAR + 0x840C Figure 19-23. Tx Enables Register (PCITER) Table 19-22. PCITER Field Descriptions Bits Name Description 31 RC Reset controller. User writes this bit high to put Transmit Controller in a reset state. Other register bits are not affected. This Reset is intended for recovery from an error condition or to reload the Start Address when Continuous mode is selected. This Reset bit does not prohibit register access but it must be negated in order to initiate a Restart sequence (i.e. writing the Packet_Size register). If it is used to reload a Start Address then the Start_Add register must be written prior to deasserting this Reset bit. 30 RF Reset FIFO. The FIFO will be reset and flushed of any existing data when set high. The Reset Controller bit and the Reset FIFO bit operate independently but clearly both must be low for normal operation. 29 — Reserved, should be cleared. 28 CM Continuous mode. User writes this bit high to activate Continuous mode. In Continuous mode the Start_Add value is ignored at each packet restart and the PCI address is auto-incremented from one packet to the next. Also, the Packets_Done status byte will become active, indicating how many packets have been transmitted since the last Reset Controller condition. If the Continuous bit is low, software is responsible for updating the Start_Add value at each packet Restart. 27 BE Bus error enable. User writes this bit high to enable bus error indications. Setting this bit allows the errors indicated by BE1, BE2, and BE3 in PCITSR to generate a bus error, which can result in a TEA on the XL bus. See Section 19.3.3.1.8, “Tx Status Register (PCITSR),” for bus error descriptions. Normally this bit will be low (negated) since illegal slave bus accesses are not destructive to register contents (although it may indicate broken software). This bit does not affect interrupt generation. 26–25 — Reserved, should be cleared. 24 ME Master enable. This is the Transmit Controller master enable signal. User must write it high to enable operation. It can be toggled low to permit out-of-order register updates prior to generating a Restart sequence (in which case transmission will begin when Master Enable is written back high), but it should not be used as such in Continuous mode because it can have the side effect of resetting the Packets_Done status counter. 23–22 — Reserved, should be cleared. MCF548x Reference Manual, Rev. 5 19-26 Freescale Semiconductor Memory Map/Register Definition Table 19-22. PCITER Field Descriptions (Continued) Bits Name Description 21 FEE FIFO error enable. User writes this bit high to enable CPU Interrupt generation in the case of FIFO error termination of a packet transmission. It may be desirable to mask CPU interrupts in the case that multichannel DMA is controlling operation, but in such a case software should poll the status bits to prevent a possible lock-up condition. 20 SE System error enable. User writes this bit high to enable CPU Interrupt generation in the case of system error termination of a packet transmission.. It may be desirable to mask CPU interrupts in the case that multichannel DMA is controlling operation, but in such a case software should be polling the status bits to prevent a possible lock-up condition. 19 RE Retry abort enable. User writes this bit high to enable CPU Interrupt generation in the case of retry abort termination of a packet transmission. It may be desirable to mask CPU interrupts in the case that multichannel DMA is controlling operation, but in such a case software should poll the status bits to prevent a possible lock-up condition. 18 TAE Target abort enable. User writes this bit high to enable CPU Interrupt generation in the case of target abort termination of a packet transmission. It may be desirable to mask CPU interrupts in the case that multichannel DMA is controlling operation, but in such a case software should poll the status bits to prevent a possible lock-up condition. 17 IAE Initiator abort enable. User writes this bit high to enable CPU Interrupt generation in the case of initiator abort termination of a packet transmission. It may be desirable to mask CPU interrupts in the case that multichannel DMA is controlling operation, but in such a case software should poll the status bits to prevent a possible lock-up condition. 16 NE Normal termination enable. User writes this bit high to enable CPU Interrupt generation at the conclusion of a normally terminated packet transmission. This may or may not be desirable depending on the nature of program control by multichannel DMA or the processor core. 15–0 — Reserved, should be cleared. 19.3.3.1.5 31 Tx Next Address Register (PCITNAR) 30 29 28 27 26 25 R 24 23 22 21 20 19 18 17 16 Next_Address W Reset 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 0 0 0 0 0 0 R Next_Address W Reset 0 0 Reg Addr 0 0 0 0 0 0 0 MBAR + 0x8410 Figure 19-24. Tx Next Address Register (PCITNAR) MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 19-27 Table 19-23. PCITNAR Field Descriptions Bits 31–0 19.3.3.1.6 31 Name Description Next_Address This status register contains the next (unwritten) PCI address and is updated at the successful completion of each PCI data beat. It represents a byte address and is updated with the user-written Start_Add value whenever the Start_Add is reloaded. It is intended to be accurate even in the case of abnormal terminations on the PCI bus. Tx Last Word Register (PCITLWR) 30 29 28 27 26 25 R 24 23 22 21 20 19 18 17 16 Last_Word W Reset 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 0 0 0 0 0 0 R Last_Word W Reset 0 0 0 0 0 0 0 Reg Addr 0 0 MBAR + 0x8414 Figure 19-25. Tx Last Word Register (PCITLWR) Table 19-24. PCITLWR Field Descriptions Bits 31–0 19.3.3.1.7 31 Name Description Last_Word This status register indicates the last 32-bit data fetched from the FIFO and is designed for the case in which an abnormal PCI termination has corrupted the integrity of the FIFO data (for that word). Tx Done Counts Register (PCITDCR) 30 29 28 27 26 25 R 24 23 22 21 20 19 18 17 16 Bytes_Done W Reset 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 0 0 0 0 0 0 R Packets_Done W Reset Reg Addr 0 0 0 0 0 0 0 0 0 MBAR + 0x8418 Figure 19-26. Tx Done Counts Register (PCITDCR) MCF548x Reference Manual, Rev. 5 19-28 Freescale Semiconductor Memory Map/Register Definition Table 19-25. PCITDCR Field Descriptions Bits Name Description 31–16 Bytes_Done This status register indicates the number of bytes transmitted since the start of a packet. It is updated at the end of each successful PCI data beat. For normally terminated packets the Bytes_Done value and the Packet_Size values will be equal. If Continuous Mode is active, the Bytes_Done value operates the same way. When the restart occurs for a continuous packet, however, Bytes_Done will read 0 and the Packets_Done field will increment. 15–0 Packets_Done This status register indicates the number of previous packets transmitted and is active only if continuous mode is in effect. The counter is reset if the following occurs: • Reset Controller bit, PCITER[RC], is asserted (normal way to restart continuous mode) • Master Enable bit, PCITER[ME], is negated during the current PCI data transmission and left negated until the NT status bit asserts The Master Enable bit, if negated as described, resets the Packets_Done status without disturbing continuous mode addressing.. At any point in time, the total number of Bytes transmitted can be calculated as: ( Packets_Done × Packet_Size ) + Bytes_Done assuming Packet_Size is the same for all restart sequences and the Packets_Done register has not been cleared. 19.3.3.1.8 R Tx Status Register (PCITSR) 31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 0 0 0 0 0 0 0 NT BE3 BE2 BE1 FE SE RE TA IA rwc1 rwc1 rwc1 rwc1 rwc1 rwc1 rwc1 rwc1 rwc1 W Reset R 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 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 0 0 W Reset Reg Addr MBAR + 0x841C 1 Bits 24-16 are read-write-clear (rwc). —Hardware can set rwc bits, but cannot clear them. —Software can clear rwc bits that are currently set by writing a 1 to the bit location. Writing a 1 to a rwc bit that is currently a 0 or writing a 0 to any rwc bit has no effect. Figure 19-27. Tx Status Register (PCITSR) Table 19-26. PCITSR Field Descriptions Bits Name Description 31–25 — Reserved, should be cleared. 24 NT Normal termination. This bit is set when any packet terminates normally. It is not set for abnormally terminated packets. An interrupt will be generated by this condition if the PCITER[NE] bit is set. This bit is cleared by writing ‘1’ to it. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 19-29 Table 19-26. PCITSR Field Descriptions (Continued) Bits Name Description 23 BE3 Bus error type 3. This bit is set whenever a slave bus transaction attempts to write to a Read-Only register. This flag bit is set regardless of the bus error enable bit (BE). If software is polling and wishes to disregard this error it must mask this bit out. No register bit corruption occurs for this (or any other) bus error case. This bit is cleared by writing ‘1’ to it. 22 BE2 Bus error type 2. This bit is set whenever a slave bus transaction attempts to write to a Reserved register (an entire 32-bit register, not just a Reserved bit or byte). This flag bit is set regardless of the bus error enable bit (BE). If software is polling and wishes to disregard this error it must mask this bit out. This bit is cleared by writing ‘1’ to it. 21 BE1 Bus error type 1. This bit is set whenever a slave bus transaction attempts to read a Reserved register (an entire 32-bit register, not just a Reserved bit or byte). This flag bit is set regardless of the bus error enable bit (BE). If software is polling and wishes to disregard this error it must mask this bit out. This bit is cleared by writing ‘1’ to it. 20 FE FIFO error. This bit is set whenever the Transmit FIFO asserts an unmasked error bit. An interrupt will be generated by this condition if the PCITER[FEE] bit is set. The source of the error must be determined by reading the FIFO status register PCITFSR. Also, the error condition must be cleared at the FIFO prior to clearing this Sticky bit or this flag will continue to assert. This bit is cleared by writing ‘1’ to it. 19 SE System error. This bit is set in response to the Transmit Controller entering an illegal state. System error indicates a malfunction of the block and should not occur in normal operation. An interrupt can be generated by this condition if the PCITER[SE] bit is set. In normal operation this should never occur. The only recovery is to assert the reset controller bit, PCITER[RC], and clear this flag by writing ‘1’ to it. 18 RE Retry error. This bit is set if Max_Retries is set to a finite value (0x01 to 0xff) and the PCI transaction has performed retries in excess of the setting. An interrupt will be generated by this condition if the PCITER[RE] bit is set. This retry counter is reset at the beginning of each packet, not at the beginning of each transaction.This bit is cleared by writing ‘1’ to it. 17 TA Target abort. This bit is set if the PCI controller has issued a Target Abort (which means the addressed PCI Target has signalled an Abort). An interrupt will be generated by this condition if the PCITER[TAE] bit is set. It is up to application software to query the Target’s status register and determine the source of the error. The coherency of the Transmit FIFO data and the Transmit Controller’s status registers (Next_Address, Bytes_Done, etc.) should remain valid. This bit is cleared by writing ‘1’ to it. 16 IA Initiator abort. This bit is set if the PCI controller issues an Initiator Abort. This indicates that no Target responded but further status information can be read from the PCI Configuration interface. An interrupt will be generated by this condition if the PCITER[IAE] bit is set. The coherency of the Transmit FIFO data and the Transmit Controller’s status registers (Next_Address, Bytes_Done, etc.) should remain valid.This bit is cleared by writing ‘1’ to it. 15–0 — Reserved, should be cleared. NOTE Registers MBAR + 0x8420 through MBAR + 0x843C are reserved for future use. Accesses to these registers will result in undefined behavior. MCF548x Reference Manual, Rev. 5 19-30 Freescale Semiconductor Memory Map/Register Definition 19.3.3.1.9 31 Tx FIFO Data Register (PCITFDR) 30 29 28 27 26 25 R 24 23 22 21 20 19 18 17 16 FIFO_Data_Word W Reset 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 0 0 0 0 0 0 R FIFO_Data_Word W Reset 0 0 0 0 0 0 0 Reg Addr 0 0 MBAR + 0x8440 Figure 19-28. Tx FIFO Data Register (PCITFDR) + Table 19-27. PCITFDR Field Descriptions Bits Name 31–0 Description FIFO_Data This is the data port to the FIFO. Reading from this location will “pop” data from the FIFO, writing _Word data will “push” data into the FIFO. During normal operation the multichannel DMA controller will be pushing data here. The PCI controller will pop data for transmission from a dedicated peripheral port, so the user program should not be reading here. Note: Only full 32-bit accesses are allowed. If all FIFO byte enables are not asserted when accessing this location, FIFO data will be corrupted. 19.3.3.1.10 Tx FIFO Status Register (PCITFSR) R 31 30 29 28 27 26 25 24 23 22 21 20 19 18 IP TXW 0 0 0 0 0 0 FAE RXW UF OF FR Full rwc1 rwc1 rwc1 rwc1 W rwc1 Reset R rwc1 17 16 Alarm Empt y 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 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 0 0 W Reset Reg Addr MBAR + 0x8444 1 Bits 31, 30 and 23-20 are read-write-clear (rwc). —Hardware can set rwc bits, but cannot clear them. —Software can clear rwc bits that are currently set by writing a 1 to the bit location. Writing a 1 to a rwc bit that is currently a 0 or writing a 0 to any rwc bit has no effect. Figure 19-29. Tx FIFO Status Register (PCITFSR) MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 19-31 Table 19-28. PCITFSR Field Descriptions Bits Name Description 31 IP Illegal Pointer. An address outside the FIFO controller’s memory range has been written to one of the user visible pointers. This bit will cause the FIFO error output to assert unless the IP_MASK bit in the FIFO Controller register is set. Resetting the FIFO will clear this condition and the bit is cleared by writing a one to it. 30 TXW Transmit Wait Condition. Since the Transmit Controller waits for enough data in the FIFO to satisfy each PCI transaction before the transfer initiates, this bit will not assert. 29–24 — 23 FAE Frame accept error. This module does not support data framing functionality, so this bit should be ignored. 22 RXW Receive wait condition. Since this FIFO is configured as a Transmit FIFO (i.e. the PCI controller only reads from this FIFO), this bit will not assert. 21 UF Underflow. This bit indicates that the read pointer has surpassed the write pointer. In other words the FIFO has been read beyond Empty. Resetting the FIFO will clear this condition and the bit is cleared by writing a one to it. 20 OF Overflow. This bit indicates that the write pointer has surpassed the read pointer. In other words the FIFO has been written beyond Full. Resetting the FIFO will clear this condition and the bit is cleared by writing a one to it. 19 FR Frame ready. The FIFO has a complete Frame of data ready for transmission. This module does not provide support for data framing functionality, so this bit should be ignored. 18 Full The FIFO is Full. This is not a sticky bit or error condition. The Full indication tracks with the state of the FIFO. 17 Alarm The FIFO is at or above the Alarm “watermark”, as set by the user according to the Alarm and Control registers settings. This is not a sticky bit or error indication. 16 Empty The FIFO is empty. This is not a sticky bit or error condition. 15–0 — Reserved, should be cleared. Reserved, should be cleared. 19.3.3.1.11 Tx FIFO Control Register (PCITFCR) R 31 30 29 28 27 26 25 24 0 0 WFR 0 0 0 0 0 0 0 0 0 1 0 0 1 0 0 15 14 13 12 11 10 9 8 7 6 5 4 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 GR 23 22 21 20 19 18 17 16 0 0 0 0 0 3 2 1 0 0 0 0 0 0 0 0 0 0 0 IP_ FAE_ RXW_ UF_ OF_ TXW_ MASK MASK MASK MASK MASK MASK W Reset R W Reset Reg Addr MBAR + 0x8448 Figure 19-30. Tx FIFO Control Register (PCITFCR) MCF548x Reference Manual, Rev. 5 19-32 Freescale Semiconductor Memory Map/Register Definition Table 19-29. PCITFCR Field Descriptions Bits Name Description 31–30 — 29 WFR 28-27 — 26–24 GR[2:0] Granularity. Control high “watermark” point at which FIFO negates Alarm condition (i.e., request for data). It represents the number of free bytes times 4. A granularity setting of zero should be avoided because it means the Alarm bit (and the Requestor signal) will not negate until the FIFO is completely full. The multichannel DMA module may perform up to 2 additional data writes after the negation of a Requestor due to its internal pipelining. 23 IP_MASK Illegal pointer mask. When this bit is set, the FIFO controller masks the Status register’s IP bit from generating an error. Reserved, should be cleared. Write frame. When this bit is set, the FIFO controller assumes next data transmitted is End of Frame (EOF). Note: This module does not support Framing. This bit should remain low. Reserved, should be cleared. 22 FAE_MASK Frame accept error mask. When this bit is set, the FIFO controller masks the Status Register’s FAE bit from generating an error. 21 RXW_MASK Receive wait condition mask. When this bit is set, the FIFO controller masks the Status Register’s RXW bit from generating an error. (To help with backward compatibility, this bit is asserted at reset.) 20 UF_MASK Underflow mask. When this bit is set, the FIFO controller masks the Status Register’s UF bit from generating an error. 19 OF_MASK Overflow mask. When this bit is set, the FIFO controller masks the Status Register’s OF bit from generating an error. 18 TXW_MASK Transmit wait condition mask. When this bit is set, the FIFO controller masks the Status Register’s TXW bit from generating an error. (To help with backward compatibility, this bit is asserted at reset.) 17–0 — Reserved, should be cleared. 19.3.3.1.12 Tx FIFO Alarm Register (PCITFAR) R 31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 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 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 0 0 0 0 0 0 0 1 1 1 W Reset R Alarm Alarm W Reset Reg Addr 0 0 0 0 0 0 0 1 1 MBAR + 0x844C Figure 19-31. Tx FIFO Alarm Register (PCITFAR) MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 19-33 Table 19-30. PCITFAR Field Descriptions Bits Name 31–12 — 11–7 Alarm Description Reserved, should be cleared. Bits 11-7 are hardwired low. 6–0 Bits 6-0 are programmable to control a 128-byte FIFO. User writes these bits to set low level “watermark”, which is the point where FIFO asserts request for multichannel DMA controller data filling. Value is in bytes. For example, with Alarm = 32 (0x20), an alarm condition occurs when the FIFO contains less than 32bytes. Once asserted, alarm does not negate until high level mark is reached, as specified by FIFO control register granularity (GR[2:0]) bits. Note: The Alarm setting should be programmed to a value greater than or equal to Max_Beats * 4 or else data transfer may stall. The Tx controller waits for enough data to form a burst of Max_Beats to be in the FIFO before it will transmit data. For a Max_Beats value of 0(8 beats), Alarm should be programmed to 32 or greater. 19.3.3.1.13 Tx FIFO Read Pointer Register (PCITFRPR) R 31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 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 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 W Reset R ReadPtr W Reset Reg Addr 0 0 0 0 MBAR + 0x8450 Figure 19-32. Tx FIFO Read Pointer Register (PCITFRPR) Table 19-31. PCITFRPR Field Descriptions Bits Name 31–7 — 6–0 ReadPtr Description Reserved, should be cleared. This value is maintained by FIFO hardware and is not normally written by the user. It can be adjusted in special cases, but this disrupts data flow integrity. The value represents the Read address presented to the FIFO RAM. MCF548x Reference Manual, Rev. 5 19-34 Freescale Semiconductor Memory Map/Register Definition 19.3.3.1.14 Tx FIFO Write Pointer Register (PCITFWPR) R 31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 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 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 W Reset R WritePtr W Reset Reg Addr 0 0 0 0 MBAR + 0x8454 Figure 19-33. Tx FIFO Write Pointer Register (PCITFWPR) Table 19-32. PCITFWPR Field Descriptions Bits Name 31–7 — 6–0 WritePtr Description Reserved, should be cleared. Value is maintained by FIFO hardware and is not normally written by user. It can be adjusted in special cases, but this disrupts data flow integrity. Value represents the Write address presented to the FIFO RAM. This marks the end of the PCI Comm Bus FIFO Transmit Interface description. 19.3.3.2 Comm Bus FIFO Receive Interface PCI Rx is controlled by 13 32-bit registers. These registers are located at an offset from MBAR. Register addresses are relative to this offset. MCF548x Reference Manual, Rev. 5 Freescale Semiconductor 19-35 19.3.3.2.1 31 Rx Packet Size Register (PCIRPSR) 30 29 28 27 26 R 25 24 23 22 21 20 19 18 Packet_Size[15:2] 17 16 Packet_Size [1:0] W Reset 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 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 0 0 R W Reset Reg Addr MBAR + 0x8480 Figure 19-34. Rx Packet Size Register (PCIRPSR) Table 19-33. PCIRPSR Field Descriptions Bits Name 31–18 Description Packet_Size Packet_Size [15:2]. The Packet_Size field indicates the number of bytes for the receive controller to read over PCI. Only bits [15:2] are writable. Only 32-bit data transfers to the FIFO are allowed. Writing to this register also completes a Restart Sequence as long as the Master Enable bit, PCIRER[ME], is high and Reset Controller bit, PCIRER[RC], is low. 17-16 Packet_Size [1:0] The two low bits are hardwired low. 15–0 — 19.3.3.2.2 31 Reserved, should be cleared. No Bus Error is generated. Rx Start Address Register (PCIRSAR) 30 29 28 27 26 25 R 24 23 22 21 20 19 18 17 16 Start_Add W Reset 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 0 0 0 0 0 0 R Start_Add W Reset Reg Addr 0 0 0 0 0 0 0 0 0 MBAR + 0x8484 Figure 19-35. Rx Start Address Register (PCIRSAR) MCF548x Reference Manual, Rev. 5 19-36 Freescale Semiconductor Memory Map/Register Definition Table 19-34. PCIRSAR Field Descriptions Bits Name Description 31–0 Start_Add The user writes this register with the desired starting address for the current packet. This is the address which will be first presented on the external PCI bus and then auto-incremented as necessary. Addressing is assumed to be sequential from the start address unless the PCIRTCR[DI] bit is set. This register will not increment as the PCI packet proceeds. 19.3.3.2.3 R Rx Transaction Control Register (PCIRTCR) 31 30 29 28 27 26 25 24 23 22 21 18 17 16 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0 0 0 0 0 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 0 0 FB 0 0 0 0 W 0 0 0 DI 0 0 0 0 0 0 0 0 0 0 0 0 0 PCI_cmd 20 19 Max_Retries W Reset R Max_Beats W Reset 0 0 Reg Addr 0 MBAR + 0x8488 Figure 19-36. Rx Transaction Control Register (PCIRTCR) Table 19-35. PCIRTCR Field Descriptions Bits Name 31–28 — 27–24 PCI_cmd 23–16 15–13 Description Reserved, should be cleared. The user writes this field with the desired PCI command to present during the address phase of each PCI transaction. The default is Memory Read Multiple. This field is not checked for consistency and if written to an illegal value, unpredictable results will occur. If not using the default value, the user should write this register only once prior to any packet Restart. Max_Retries The user writes this field with the maximum number of retries to permit “per packet”. The retry counter is reset when the packet completes normally or is terminated by a master abort, target abort, or an abort due to exceeding the retry limit. A slow or malfunctioning Target might issue infinite disconnects and therefore permanently tie up the PCI bus. A finite (0x01 to 0xf) Max_Retries value will detect this condition and generate an interrup