manualshive.com logo in svg

Xilinx Virtex-II Pro PPC405, Руководство пользователя

Поделиться
Скачать
Xilinx Virtex-II Pro PPC405 Скачать руководство пользователя страница 548

856

www.xilinx.com

March 2002 Release

1-800-255-7778

Virtex-II Pro™ Platform FPGA Documentation

Appendix E:

PowerPC

®

 6xx/7xx Compatibility

R

•

The programmable-interval timer (PIT) interrupt. This interrupt is triggered by a 
time-out on the PIT registers. Its function replaces that of the decrementer interrupt in 
the PowerPC 6xx/7xx family.

•

The fixed-interval timer (FIT) interrupt. This interrupt is triggered by a pre-
determined bit transition in the time base. This feature is not supported by the 
PowerPC 6xx/7xx family.

•

The watchdog timer (WDT) interrupt. This critical interrupt is triggered by a pre-
determined bit transition in the time base. This feature is not supported by the 
PowerPC 6xx/7xx family.

•

The timer-control register (TCR). This register controls the PowerPC 40x timer 
resources. It is not supported by the PowerPC 6xx/7xx family.

•

The timer-status register (TSR). This register is used by the PowerPC 40x timer 
resources to report status. It is not supported by the PowerPC 6xx/7xx family.

Other Differences

Instructions

PowerPC 40x processors can support implementation-specific instructions that are not 
supported in PowerPC 6xx/7xx processors. For example, the multiply-accumulate (MAC) 
instructions are not supported by PowerPC 6xx/7xx processors. Refer to 

Table B-32, 

page 798

, for a list of implementation dependent PPC405 instructions. This table also 

shows which PPC405 instructions are not supported by the PowerPC architecture.

Endian Support

The default memory-access order for all PowerPC processors is big-endian. The PowerPC 
embedded-environment architecture defines a true little-endian memory-access capability 
that is implemented using the endian storage attribute (E). The PPC405 supports this 
capability. The PowerPC architecture supports a little-endian mode that is implemented by 
PowerPC 6xx/7xx processors. This mode is not supported by the PPC405.

Debug Resources

Debug resources are implementation dependent. In general, all PowerPC 40x processors 
support debug events on both instruction addresses and data addresses. Debug events are 
controlled using the DBCR0 and DBCR1 registers. Debug status is reported by the DBSR 
register. PowerPC 6xx/7xx processors support debug resources to varying degrees, but the 
capabilities are often less comprehensive than those supported by PowerPC 40x 
processors.

Power Management

The PowerPC 40x family implements power management using the MSR[WE] bit. Setting 
this bit places the processor in the wait state. Power management is disabled when an 
interrupt occurs.

The PowerPC 6xx/7xx family similarly implements power management using the 
MSR[POW] bit. PowerPC 7xx processors support four different power states, programmed 
using the HID0 register. Power management is disabled when an interrupt occurs.

Содержание Virtex-II Pro PPC405

Страница 1: ...R Volume 2 a PPC405 User Manual Virtex II Pro Platform FPGA Developer s Kit March 2002 Release...

Страница 2: ...CodePack All other trademarks are the property of their respective owners Xilinx does not assume any liability arising out of the application or use of any product described or shown herein nor does...

Страница 3: ...entation of the PowerPC embedded environment architecture This chapter also contains an overview of the features supported by the PPC405 Chapter 2 Operational Concepts introduces the processor operati...

Страница 4: ...PC 6xx 7xx Compatibility describes the programming model differences between the PPC405 and PowerPC 6xx and 7xx series processors Appendix F PowerPC Book E Compatibility describes the programming mode...

Страница 5: ...tions Continued Convention Definition Table P 2 Instruction Field Definitions Field Location Description AA 30 Absolute address bit branch instructions 0 The immediate field represents an address rela...

Страница 6: ...e address of the next instruction MB 21 25 Mask begin Used in rotate and mask instructions to specify the beginning bit of a mask ME 26 30 Mask end Used in rotate and mask instructions to specify the...

Страница 7: ...ield Extended opcodes in decimal appear in the instruction format diagrams presented with individual instructions The XO field name does not appear in instruction descriptions XO 22 30 Extended opcode...

Страница 8: ...he result of extending n on the left with sign bits if then else Conditional execution if condition then a else b where a and b represent one or more pseudocode statements Indenting indicates the rang...

Страница 9: ...se register number formed using the split TBRF field in a mftb instruction Table P 3 Pseudocode Conventions Continued Convention Definition Table P 4 Operator Precedence Operators Associativity REGIST...

Страница 10: ...register rn Specifies GPR n r15 for example SGR Storage guarded register SLER Storage little endian register SPRGn SPR general purpose register n SRRn Save restore register n SU0R Storage user defined...

Страница 11: ...ructions between the processor and PLB It is used when cache misses occur and when access to non cacheable memory occurs flush A cache or TLB operation that involves writing back a modified entry to m...

Страница 12: ...program or portion of a program and any data required for the program to run problem state Synonym for user mode real address Synonym for physical address scalar Individual data objects and instructi...

Страница 13: ...6 Published by Warthman Associates Palo Alto CA ISBN 0 9649654 0 2 Optimizing PowerPC Code Programming the PowerPC Chip in Assembly Language by Gary Kacmarcik ASIN 0201408392 PowerPC Programming Pocke...

Страница 14: ...322 www xilinx com March 2002 Release 1 800 255 7778 Virtex II Pro Platform FPGA Documentation Preface R...

Страница 15: ...Overview The PowerPC architecture is a 64 bit architecture with a 32 bit subset The material in this document only covers aspects of the 32 bit architecture implemented by the PPC405 In general the P...

Страница 16: ...chitecture are defined in Table 1 1 The PowerPC architecture requires that all PowerPC implementations adhere to the UISA offering compatibility among all PowerPC application programs However differen...

Страница 17: ...t one level of the architecture and defined more specifically at another For example the UISA defines conditions that can cause an alignment exception and the OEA specifies the exception itself Featur...

Страница 18: ...s do not include the special 64 bit extensions to the PowerPC UISA Also floating point support can be provided either in hardware or software by PowerPC embedded environment processors Figure 1 1 show...

Страница 19: ...programs to create or modify code to manage storage coherency and to optimize memory access performance It defines the cache and memory models the timekeeping resources from a user perspective and res...

Страница 20: ...uctions Exception model Dual level interrupt structure supporting various exception types Specification of interrupt priorities and masking Privileged SPRs for controlling and handling exceptions Inte...

Страница 21: ...ed software is in general not compatible but the differences are relatively minor Software developers who are concerned with cross compatibility of privileged software between the PPC405 and PowerPC B...

Страница 22: ...uring cacheline fills and flushes Support for on chip memory OCM that can provide memory access performance identical to a cache hit Flexible memory management Translation of the 4 GB logical address...

Страница 23: ...mmediate value in the instruction Register indirect with index A base address is stored in a register and a displacement from the base address is stored in a second register Register indirect The data...

Страница 24: ...d by that operand For example if the four bytes in a word operand occupy addresses 4 5 6 and 7 the word address is 4 The PPC405 supports both big endian an operand s most significant byte is at the lo...

Страница 25: ...igure 1 2 PPC405 Registers UG011_51_033101 MSR Machine State Register CCR0 Core Configuration Register SPR General Purpose Registers SPRG0 SPRG1 SPRG2 SPRG3 SPRG4 SPRG5 SPRG6 SPRG7 Memory Management R...

Страница 26: ...ructions can set CR0 and compare instructions can set any CR field Additional instructions are provided to perform logical operations and tests on CR fields and bits within the fields The CR can be ac...

Страница 27: ...register or count register The default prediction can be overridden by software at assembly or compile time This capability is described further in Branch Prediction page 370 The PPC405 has a single...

Страница 28: ...he PPC405 Memory Management Unit The PPC405 supports 4 GB of flat non segmented address space The memory management unit MMU provides address translation protection functions and storage attribute con...

Страница 29: ...cacheline fill minimizing execution stalls caused by instruction cache misses When the ICU is accessed four instructions are read from the appropriate cacheline and placed temporarily in a line buffer...

Страница 30: ...r appears to be stopped Real time trace mode which supports event triggering for real time tracing Debug events are supported that allow developers to manage the debug process Debug modes and debug ev...

Страница 31: ...al processor state to facilitate software debugging This capability complies with the IEEE 1149 1 specification for vendor specific extensions and is therefore compatible with standard JTAG hardware f...

Страница 32: ...340 www xilinx com March 2002 Release 1 800 255 7778 Virtex II Pro Platform FPGA Documentation Chapter 1 Introduction to the PPC405 R...

Страница 33: ...sor systems The weakly consistent memory model allows system bus operations to be reordered dynamically The goal of reordering bus operations is to reduce the effect of memory latency and improving ov...

Страница 34: ...e the operation complete execution in the context they were initiated This includes privilege level translation mode and memory protection All instructions following the operation complete execution i...

Страница 35: ...leged mode User mode The processor operating mode is controlled by the privilege level field in the machine state register MSR PR When MSR PR 0 the processor operates in privileged mode When MSR PR 1...

Страница 36: ...ion address NIA when it is used to fetch an instruction sequentially or as the result of a branch The input values and method used by the processor to calculate an EA depend on the instruction that is...

Страница 37: ...d to translate the addresses generated by programs into physical memory addresses Memory management also consists of the mechanisms used to characterize memory region behavior also referred to as stor...

Страница 38: ...essing Modes Programs can use 32 bit effective addresses to reference the 4 GB physical address space using one of two addressing modes Real mode Virtual mode Real mode and virtual mode are enabled an...

Страница 39: ...bit is numbered 31 A bit set to 1 has a numerical value associated with its position b relative to the least significant bit lsb This value is equal to 2 lsb b For example if bit 5 is set to 1 in a b...

Страница 40: ...3 0 31 0x03 0x02 0x01 0x00 0x04 LSB MSB Byte 0 Byte 1 Byte 2 Byte 3 Memory Content Memory Address Halfword Byte 0 Byte 1 0 31 0x03 0x02 0x01 0x00 0x04 LSB MSB Byte 0 Byte 1 Byte Byte 0 0 31 0x00 0x03...

Страница 41: ...equential byte address is assigned to the next lowest byte and so on The term little endian is used because the little end of the scalar when considered as a binary number comes first in memory The fo...

Страница 42: ...nment architecture and PowerPC Book E architecture Little endian mode defined by the PowerPC architecture is not implemented by the PPC405 Little endian mode does not support true little endian memory...

Страница 43: ...reordering for both types of memory accesses Little Endian Instruction Fetching Instructions are word four byte data types that are always aligned on word boundaries in memory Instructions stored in a...

Страница 44: ...ata item is a byte is not reversed when the big endian and little endian mappings are compared For example the character A is located at address 14 in both the big endian and little endian mappings In...

Страница 45: ...e aligned Alignment and Endian Storage Control The endian storage control attribute E does not affect how the processor handles operand alignment Data alignment is handled identically for accesses to...

Страница 46: ...lignment Interrupt 0x0600 page 510 for more information Instruction Conventions Instruction Forms Opcode tables and instruction listings often contain information regarding the instruction form This i...

Страница 47: ...lass Defined instructions contain all the instructions defined by the PowerPC architecture Defined instructions are guaranteed to be supported by all implementations of the PowerPC architecture The on...

Страница 48: ...served instruction as described in the following section An attempt to execute an illegal instruction causes an illegal instruction error program exception With the exception of an instruction consist...

Страница 49: ...ntations The last three columns indicate which PPC405 instructions are part of the allocated instruction class Support of these instructions by PowerPC Book E processors is implementation dependent De...

Страница 50: ...Class The reserved instruction class consists of all instruction primary opcodes and associated extended opcodes if applicable that do not belong to either the defined class or the allocated class Pre...

Страница 51: ...are typically creates a context execution environment that protects itself and other applications from the effects of an errant application program The remaining chapters in this book generally descri...

Страница 52: ...A Special Purpose Registers page 770 for a complete list of all SPRs user and privileged supported by the PPC405 Simplified instruction mnemonics are available for the mtspr and mfspr instructions fo...

Страница 53: ...pecific bits in the CR The integer arithmetic instructions can update CR0 to reflect their result The integer compare instructions can update a specific CR field to reflect their result Conditional br...

Страница 54: ...bit is set when the result is positive and not zero otherwise it is cleared 2 EQ Zero 0 Result is not equal to zero 1 Result is equal to zero This bit is set when the result is zero otherwise it is cl...

Страница 55: ...Table 3 3 Fixed Point Exception Register XER Bit Definitions Bit Name Function Description 0 SO Summary overflow 0 No overflow occurred 1 Overflow occurred SO is set to 1 whenever an instruction excep...

Страница 56: ...unt that is decremented by a conditional branch instruction with an appropriately coded BO opcode field The value in the CTR wraps to 0xFFFF_FFFF if the value in the register is 0 prior to the decreme...

Страница 57: ...G7 263 0x107 These registers can be read using the mfspr instruction In privileged mode system software accesses these registers using different SPR numbers see page 432 Time Base Registers The time b...

Страница 58: ...on storage interrupt handler to be invoked Alignment Exception An attempt to access memory with an invalid effective address alignment for the specific instruction causes the alignment interrupt handl...

Страница 59: ...R or fields within the branch instruction Optionally a branch return address can be automatically loaded into the LR by setting the LK instruction opcode bit to 1 This option is useful for specifying...

Страница 60: ...n Branch Unconditional Table 3 6 lists the PowerPC unconditional branch instructions These branches specify a 26 bit signed displacement to the branch target address by appending the 24 bit LI instruc...

Страница 61: ...anch Unconditional Instructions Mnemonic Name Operation Operand Syntax b Branch Branch to relative address tgt_addr ba Branch Absolute Branch to absolute address bl Branch and Link Branch to relative...

Страница 62: ...rchitecture provides software with the ability to override reverse the dynamic prediction using a static prediction hint encoded in the instruction opcode This can be useful when it is known at compil...

Страница 63: ...ed by the processor Setting the y bit to 1 reverses the above behavior For branch always encoding BO 0 BO 2 branch prediction cannot be reversed no y bit is recognized The sign of the displacement ope...

Страница 64: ...mpute the effective address EA of the next instruction using the following addressing modes Branch to relative conditional and unconditional Branch to absolute conditional and unconditional Branch to...

Страница 65: ...30 to 0 The link register update option is enabled by setting the LK instruction field bit 31 to 1 This option causes the effective address of the instruction following the branch instruction to be lo...

Страница 66: ...ons that use branch conditional to absolute addressing generate the next instruction address by appending 0b00 to the immediate displacement operand BD and sign extending the result Branches using thi...

Страница 67: ...anch conditional to count register instruction generates the next instruction address by reading the contents of the CTR and clearing the two low order bits to zero The link register update option is...

Страница 68: ...vice routine Executing the sc instruction causes a system call exception to occur See System Call Interrupt 0x0C00 page 514 for more information on the operation of this instruction Table 3 10 Conditi...

Страница 69: ...tions See Trap Instructions page 832 for more information The TO operand field in the system trap instructions specifies the test conditions performed on the remaining two operands Multiple test condi...

Страница 70: ...of three addressing modes register indirect with immediate index register indirect with index or register indirect These addressing modes are described in the following sections For some instructions...

Страница 71: ...of these two registers are added to generate the effective address If the rA instruction field is 0 specifying r0 a value of zero rather than the contents of r0 is added to rB The option to specify rA...

Страница 72: ...o to be generated The option to specify rA or 0 is shown in the instruction descriptions as rA 0 Figure 3 19 shows how an effective address is generated when using register indirect addressing Figure...

Страница 73: ...PC405 the above invalid instruction forms produce a boundedly undefined result In other PowerPC implementations those forms can cause a program exception Load Byte and Zero Table 3 14 lists the PowerP...

Страница 74: ...oad Halfword and Zero Register indirect with immediate index EA rA 0 d rD d rA lhzu Load Halfword and Zero with Update Register indirect with immediate index EA rA d rA EA rA 0 rA rD lhzx Load Halfwor...

Страница 75: ...ng Mode Operand Syntax lha Load Halfword Algebraic Register indirect with immediate index EA rA 0 d rD d rA lhau Load Halfword Algebraic with Update Register indirect with immediate index EA rA d rA E...

Страница 76: ...C implementations that form can cause a program exception Store Byte Table 3 18 lists the PowerPC store byte instructions These instructions store the lower eight bits of rS into the specified byte lo...

Страница 77: ...a in little endian order Likewise when used in a system operating with little endian byte order these instructions have the effect of loading Table 3 19 Store Halfword Instructions Mnemonic Name Addre...

Страница 78: ...n Table 3 21 Load and Store with Byte Reverse Instructions Mnemonic Name Addressing Mode Operand Syntax lhbrx Load Halfword Byte Reverse Indexed Register indirect with index EA rA 0 rB rD rA rB lwbrx...

Страница 79: ...PowerPC load and store string instructions and their addressing modes See the individual instruction listings in Chapter 11 Instruction Set for more information on their operation and restrictions on...

Страница 80: ...NB 0 then n 32 The indexed forms take the byte count from XER 25 31 Unlike the immediate forms if XER 25 31 0 then n 0 For the lswx instruction the contents of rD are undefined if n 0 The n bytes are...

Страница 81: ...signed integers The following types of integer instructions are supported by the PowerPC architecture Arithmetic Instructions Logical Instructions Compare Instructions Rotate Instructions Shift Instru...

Страница 82: ...and in some cases between GPRs and signed immediate values Integer Addition Instructions Table 3 24 shows the PowerPC integer addition instructions The instructions in this table are grouped by the t...

Страница 83: ...A 0 SIMM 0x0000 addis Add Immediate Shifted XER and CR0 are not updated rD rA SIMM Add Extended Instructions rD is loaded with the sum rA rB XER CA adde Add Extended XER CA is updated to reflect the r...

Страница 84: ...is represented by the register pair r3 r4 where r3 contains the most significant 32 bits of i and r4 contains the least significant 32 bits The 64 bit integer j is similarly represented by the registe...

Страница 85: ...rB subfe Subtract from Extended and Record XER CA and CR0 are updated to reflect the re sult subfeo Subtract from Extended with Overflow Enabled XER CA OV SO are updated to reflect the result subfeo...

Страница 86: ...bled and Record XER OV SO and CR0 are updated to reflect the result Table 3 27 Multiply Instructions Mnemonic Name Operation Operand Syntax Multiply Low Word Instructions rD is loaded with the low 32...

Страница 87: ...tination register rather than a source register rS is used to specify one of the source registers AND and NAND Instructions Table 3 29 shows the PowerPC AND and NAND instructions For each type of inst...

Страница 88: ...i AND Immediate and Record CR0 is updated to reflect the result rA rS UIMM AND Immediate Shifted Instructions rA is loaded with the logical result rS AND UIMM 0x0000 andis AND Immediate Shifted and Re...

Страница 89: ...OR UIMM 0x0000 oris OR Immediate Shifted CR0 is not updated rA rS UIMM OR with Complement Instructions rA is loaded with the logical result rS OR rB orc OR with Complement CR0 is not updated rA rS rB...

Страница 90: ...ion on the CR fields The second operand specifies the operand length This is referred to the L bit in the compare instruction encoding When using the compare instructions on 32 bit PowerPC implementat...

Страница 91: ...ion has the record Rc bit set to 1 in the instruction encoding CR0 CR 0 3 is updated to reflect the result of the operation A set Rc bit is indicated by the suffix in the instruction mnemonic Rotate i...

Страница 92: ...r is summarized as if MB ME then mask MB ME 1 s mask all remaining bits 0 s else mask MB 31 ones mask 0 ME ones mask all remaining bits 0 s Figure 3 23 shows the generated mask for both cases Rotate L...

Страница 93: ...left rotated 16 bits and the result is written to rA after ANDing with the mask This has the effect of extracting byte 0 from rS rS 0 7 and placing it in byte 2 of rA rA 16 23 Rotate Left then Mask In...

Страница 94: ...ll other bit positions The example shows the original contents of the destination register rA and the source register rS rS is rotated 16 bits and the result is inserted into rA after ANDing with the...

Страница 95: ...instructions In the operand syntax for shift instructions the rA operand specifies the destination register rather than a source register rS is used to specify the source register Simplified mnemonics...

Страница 96: ...tion Figure 3 26 Logical Shift Examples UG011_18_033101 Shift by 7 bits rS 0 31 1000_0111_0110_0101_0100_0011_0010_0001 rA 0 31 0000_0001_0000_1110_1100_1010_1000_0110 0 31 0000_0001_0000_1110_1100_10...

Страница 97: ...instructions negative multiply accumulate instructions and multiply halfword instructions Modulo and Saturating Arithmetic The multiply accumulate and negative multiply accumulate instructions produc...

Страница 98: ...it intermediate result Generally rD is loaded with the lower 32 bits of the 33 bit intermediate result However if the instruction performs saturating arithmetic and the intermediate result overflows r...

Страница 99: ...bits of this result are stored in rD Otherwise the nearest representable value is stored in rD macchwsu Multiply Accumulate Cross Halfword to Word Saturate Unsigned XER and CR0 are not updated rD rA...

Страница 100: ...dicates the multiply accumulate operation performed The column also shows on an instruction by instruction basis how the XER and CR registers are updated if at all Figure 3 28 Multiply Accumulate Cros...

Страница 101: ...ts of this result are stored in rD Otherwise the nearest representable value is stored in rD machhwsu Multiply Accumulate High Halfword to Word Saturate Unsigned XER and CR0 are not updated rD rA rB m...

Страница 102: ...egister rD producing a 33 bit intermediate result Generally rD is loaded with the lower 32 bits of the 33 bit intermediate result However if the instruction performs saturating arithmetic and the inte...

Страница 103: ...Otherwise the nearest representable value is stored in rD maclhws Multiply Accumulate Low Halfword to Word Saturate Signed XER and CR0 are not updated rD rA rB maclhws Multiply Accumulate Low Halfword...

Страница 104: ...o Word Modulo Unsigned XER and CR0 are not updated rD rA rB maclhwu Multiply Accumulate Low Halfword to Word Modulo Unsigned and Record CR0 is updated to reflect the result maclhwuo Multiply Accumulat...

Страница 105: ...This product is negated and added to the value in the destination register rD producing a 33 bit intermediate result this is the same as subtracting the product from rD Generally rD is loaded with th...

Страница 106: ...the result nmacchwo Negative Multiply Accumulate Cross Halfword to Word Modulo Signed with Overflow Enabled and Record XER OV SO and CR0 are updated to reflect the result Negative Multiply Accumulate...

Страница 107: ...tination register rD producing a 33 bit intermediate result this is the same as subtracting the product from rD Generally rD is loaded with the lower 32 bits of the 33 bit intermediate result However...

Страница 108: ...h Overflow Enabled XER OV SO are updated to reflect the result nmachhwo Negative Multiply Accumulate High Halfword to Word Modulo Signed with Overflow Enabled and Record XER OV SO and CR0 are updated...

Страница 109: ...result this is the same as subtracting the product from rD Generally rD is loaded with the lower 32 bits of the 33 bit intermediate result However if the instruction performs saturating arithmetic an...

Страница 110: ...the result nmaclhwo Negative Multiply Accumulate Low Halfword to Word Modulo Signed with Overflow Enabled and Record XER OV SO and CR0 are updated to reflect the result Negative Multiply Accumulate L...

Страница 111: ...depending on the instruction For each type of instruction shown in Table 3 46 the Operation column indicates the multiply operation performed The column also shows on an instruction by instruction bas...

Страница 112: ...f at all The XER register is not updated by these instructions Multiply Cross Halfword to Word Unsigned Instructions rD is loaded with the unsigned product rA 16 31 rB 0 15 mulchwu Multiply Cross Half...

Страница 113: ...updated if at all The XER register is not updated by these instructions Multiply High Halfword to Word Unsigned Instructions rD is loaded with the unsigned product rA 0 15 rB 0 15 mulhhwu Multiply Hi...

Страница 114: ...egal instruction and call the appropriate library routines to emulate the floating point instruction using integer instructions This method is not preferred due to the overhead associated with executi...

Страница 115: ...4 bit CR fields with the most significant CRM bit corresponding to CR0 and the least significant CRM bit corresponding to CR7 When mtcrf is executed a CR field is loaded with the corresponding bits i...

Страница 116: ...tion For more information see Split Field Notation page 571 Synchronizing Instructions Table 3 51 lists the PowerPC synchronization instructions The types of synchronization defined by the PowerPC arc...

Страница 117: ...only when necessary Synchronization Effects of PowerPC Instructions Additional PowerPC instructions can cause synchronizing operations to occur All instructions that result in some form of synchroniza...

Страница 118: ...If the reservation exists when the store is executed the store is performed and CR0 EQ is set to 1 If the reservation does not exist when the store is executed the target memory location is not modifi...

Страница 119: ...supported on all PowerPC implementations It is good programming practice to always specify identical addresses for lwarx and stwcx pairs The PPC405 can maintain only one reservation at a time The add...

Страница 120: ...428 www xilinx com March 2002 Release 1 800 255 7778 Virtex II Pro Platform FPGA Documentation Chapter 3 User Programming Model R...

Страница 121: ...r 8 Timer Resources describes the time base and timer registers Chapter 9 Debugging describes the resources available in the PPC405 for debugging software and hardware Privileged Registers Figure 4 1...

Страница 122: ...1 SPR 0x111 SPRG2 SPR 0x112 SPRG3 SPR 0x113 SPRG4 SPR 0x114 SPRG5 SPR 0x115 SPRG6 SPR 0x116 SPRG7 SPR 0x117 PVR Processor Version Register SPR 0x11F Memory Management Registers PID SPR 0x3B1 ZPR SPR 0...

Страница 123: ...ation on these instructions The MSR is also modified during execution of the system call instruction sc return from interrupt instructions rfi and rfci and by the exception mechanism during a control...

Страница 124: ...0200 page 504 for more information 20 FE0 Floating Point Exception Mode 0 Unsupported This bit is unsupported and ignored by the PPC405 Software should clear this bit to 0 21 DWE Debug Wait Enable 0 D...

Страница 125: ...ectively The bit definitions are shown in Table 4 2 The PVR is a privileged read only SPR with an address of 287 0x11F It is read using the mfspr instruction Write access is not supported 0 31 General...

Страница 126: ...system call interrupt handler determines which system service routine to call and whether the calling application has permission to call that service If permission is granted the system call interrupt...

Страница 127: ...tion page 342 for more information Special Purpose Register Instructions The special purpose register instructions shown in Table 4 6 are used to read and write the special purpose registers SPRs usin...

Страница 128: ...se of the processor wait state Wait state is a low power operating mode that can be used to conserve processor energy when the processor is not busy Wait state is entered when software sets the wait s...

Страница 129: ...sually through a memory controller The PowerPC architecture does not define the type organization implementation or existence of internal or external caches The cache structure of other PowerPC proces...

Страница 130: ...pipelining and prefetching for cache misses and non cacheable requests Buffering of up to eight non cacheable instructions in the fill buffer Support for non cacheable hits into the fill buffer Flash...

Страница 131: ...t have a dirty bit The 512 total lines of 32 bytes each yields a 16 KB cache size Data is selected from the data cache using fields within the data address Likewise an instruction is selected from the...

Страница 132: ...two tags are present in a congruence class The tag field in the data address is compared to both tags in the congruence class A hit occurs when the data address tag field is equal to one of the two ta...

Страница 133: ...the ICU examines the instruction cache for a hit When a hit occurs the cacheline is read from the instruction cache and loaded into the line buffer Individual instructions are sent from the line buffe...

Страница 134: ...onfiguration Register page 459 describes additional software controls that can be used to manage instruction prefetching from cacheable and non cacheable memory Instruction Cache Hint Instruction The...

Страница 135: ...resulting in as many as eight synonyms 4 KB pages one bit EA19 is used in indexing and tag comparison resulting in two possible synonyms The following two options are available for preventing cache sy...

Страница 136: ...a cacheline flush to occur prior to loading the cacheline from the fill buffer A cacheline flush updates system memory with the modified data from the cache All 32 bytes in a cacheline are written seq...

Страница 137: ...contiguous addresses Loads from non cacheable memory and those that do not allocate cachelines as described above can be programmed to generate eight word PLB requests or to generate only the number...

Страница 138: ...if the first load cannot be completed immediately If a subsequent DCU request of any kind is made it is not accepted until the previous loads are completed by the DCU A store to non cacheable memory...

Страница 139: ...ine corresponding to the specified address Data cache block zero dcbz This instruction allocates a cacheline corresponding to the specified address and clears the cacheline contents to zero It can be...

Страница 140: ...No other access is guaranteed to be atomic particularly the following Load and store operations using unaligned operands Accesses resulting from execution of the lmw stmw lswi lswx stswi or stswx ins...

Страница 141: ...ve accesses are inappropriate For example an attempt to fetch instructions from addresses that do not contain instructions can cause a program to fail Speculatively reading data from a memory mapped I...

Страница 142: ...ents speculative accesses from occurring outside the appropriate regions The system call and interrupt return instructions sc rfi and rfci are not recognized by the processor as breaks in program flow...

Страница 143: ...controls the caching policy of a memory region When the W attribute is cleared to 0 the memory region has a write back caching policy Writes that hit the cache update the cacheline but they do not upd...

Страница 144: ...ion condition can be enabled using the U0 exception enable bit U0XE in the CCR0 register see Core Configuration Register page 459 When CCR0 U0XE 1 a store to memory locations with U0 1 cause a data st...

Страница 145: ...bits in the DCWR are cleared to 0 This establishes a write back caching policy for all real mode memory The DCWR is a privileged SPR with an address of 954 0x3BA and can be read and written using the...

Страница 146: ...d and written using the mfspr and mtspr instructions Instruction Cache Cacheability Register ICCR The instruction cache cacheability register ICCR specifies real mode instruction memory cacheability t...

Страница 147: ...2 containing compressed instructions In those implementations memory regions with U0 1 contain compressed instructions and memory regions with U0 0 contain uncompressed instructions System software c...

Страница 148: ...on name the term cache block often appears A cache block is synonymous with a cacheline Table 5 3 summarizes which cache control instructions are privileged and which instructions can be executed in u...

Страница 149: ...g EA rA 0 rB rA rB icbt Instruction Cache Block Touch If the instruction specified by the effective address EA is cacheable and is not currently cached by the instruction cache the cacheline containin...

Страница 150: ...EA is cached by the data cache the cacheline containing that byte is invalidated If the cacheline is modified dirty those modifications are lost EA is calculated using register indirect with index add...

Страница 151: ...on Cache debug features Figure 5 13 shows the format of the CCR0 The fields in CCR0 are defined as shown in Table 5 6 dcbz Data Cache Block Clear to Zero An effective address EA is calculated using re...

Страница 152: ...e data cache unit onto the processor local bus PLB Bit 0 is controlled by the processor and cannot be controlled by software See PLB Request Priority page 461 for more information 10 11 IPP ICU PLB Pr...

Страница 153: ...available GPRs 22 NCRS Non Cacheable Request Size 0 Request size is four words 1 Request size is eight words Specifies the number of instructions requested from non cacheable memory when an instructio...

Страница 154: ...ht instructions are touched into a single cacheline This function example contains seven instructions If more than eight instructions are required additional lines must be touched into the cache icbt...

Страница 155: ...present in the cache are not updated For example when a DMA controller reads and writes cacheable system memory coherency can be lost because The processor does not automatically supply the DMA contr...

Страница 156: ...ndicates a replacement value in the cache unrelated to the program Next assume an external device updates the words at system memory addresses 0x100C 0x1024 while at the same time a cacheline reload f...

Страница 157: ...mory regions It is recommended that the alignment and size of shared memory regions be a multiple of the cacheline size By configuring all shared memory regions to start on a cacheline boundary and sp...

Страница 158: ...smaller than the data cache The following code sequence is an example of how shared memory can be flushed from the data cache r1 start of shared memory region r2 end of shared memory region loop dcbf...

Страница 159: ...tes executable code from one cacheable memory location to another requires the same coherency treatment as self modifying code Although instructions are not changed they are treated as data by the pro...

Страница 160: ...its EA22 26 A way is selected from the congruence class using the cache way select field CWS in the CCR0 register CCR0 CWS 0 selects way A and CCR0 CWS 1 selects way B The cacheline information in the...

Страница 161: ...ngruence class and way is loaded into the destination GPR rD Figure 5 15 shows the format of the cache information loaded into rD The information fields loaded in rD are defined as shown in Table 5 8...

Страница 162: ...470 www xilinx com March 2002 Release 1 800 255 7778 Virtex II Pro Platform FPGA Documentation Chapter 5 Memory System Management R...

Страница 163: ...rovides greater control over memory protection Blocks of memory as small as 1 KB can be individually protected from unauthorized access Protection and relocation enable system software to support mult...

Страница 164: ...ess protection is limited the U0 storage attribute can be used for write protection Implementation of a real mode memory manager is more straightforward than a virtual mode memory manager Real mode is...

Страница 165: ...LB When translating a virtual address the MMU examines the page translation entries for a matching VPN PID and EPN Rather than examining all entries in the table only entries contained in the processo...

Страница 166: ...ticular embedded system Both the page table size and access time can be optimized Independent page sizes can be used for application modules device drivers system service routines and data Independent...

Страница 167: ...lation into a physical page determine the protection characteristics of the page and specify the storage attributes associated with the page The PPC405 TLB is physically implemented as three separate...

Страница 168: ...TLBHI also referred to as the tag entry and TLBLO also referred to as the data entry The fields within a TLB entry are categorized as follows Virtual page identification These fields identify the pag...

Страница 169: ...sed by all processes Physical Page Identification Fields The physical page identification portion of a TLB entry contains the following field RPN Physical page number or real page number TLBLO bits 0...

Страница 170: ...formation M Memory coherent TLBLO bit 30 Setting and clearing this bit does not affect memory accesses in the PPC405 In implementations that support multi processing this bit can be used to improve th...

Страница 171: ...D value of 0x00 A PID value of 0x00 does not identify a process that can access any page When PID 0x00 a page translation hit only occurs when TID 0x00 It is possible for software to load the TLB with...

Страница 172: ...n real mode all address translation and memory protection checks performed by the MMU are disabled After system software initializes the UTLB with page translation entries management of the PPC405 UTL...

Страница 173: ...age Exception When instruction address translation is enabled MSR IR 1 an instruction storage exception occurs when access to a page is not permitted for any of the following reasons From user mode Th...

Страница 174: ...fetched from a page instead causing an instruction storage interrupt ISI to occur The ISI does not occur when the instruction is fetched but instead occurs when the instruction is executed This preven...

Страница 175: ...LO WR only Because this is a privileged instruction access cannot be denied by zone protection dcbz Affected by TLBLO WR and in user mode only ZPR Zn 00 Other cache control instructions can invalidate...

Страница 176: ...cause an exception and are instead treated as a no operation dcba Affected by TLBLO WR and in user mode only ZPR Zn 00 dcbt Affected by ZPR Zn 00 in user mode only This instruction is treated as a lo...

Страница 177: ...dcbz Data storage interrupt Data storage interrupt dccci Data storage interrupt No violation privileged instruction dcread No violation treated as load No violation privileged instruction icbi No viol...

Страница 178: ...sically available Some of the software and data pages must be stored outside physical memory such as on a hard drive when they are not used Ideally the most frequently used pages stay in physical memo...

Страница 179: ...also record the information in a separate data structure associated with the page translation entry Page modification information is used to indicate whether an old page can be overwritten with a new...

Страница 180: ...he interrupt handler updates the UTLB and returns from the interrupt handler using rfi without enabling virtual mode no additional context synchronization is required However if virtual mode is enable...

Страница 181: ...ransfers that occur as a result of an exception An interrupt occurs when the processor suspends execution of a program after detecting an exception The processor saves the suspended program machine st...

Страница 182: ...instruction to have completed execution partially completed execution or not have begun execution No instructions following the excepting instruction are executed prior to transferring control to the...

Страница 183: ...should be used to reference memory Also one of the following two rules must be followed The memory operand must be aligned on the operand size boundary see Table 2 1 page 353 The accessed memory locat...

Страница 184: ...ct when transferring program control to an interrupt handler Alignment 0x0600 Noncritical Synchronous Precise Unaligned operand of dcread lwarx stwcx dcbz to non cacheable or write through memory Prog...

Страница 185: ...e return address Refer to the specific interrupt description in Interrupt Reference page 502 for information on the saved return address 2 Save the interrupted program state Figure 7 1 PPC405 Exceptio...

Страница 186: ...in real mode Certain interrupts are disabled depending on the exception 6 Synchronize the processor context All interrupts are context synchronizing The processor fetches and executes the first instru...

Страница 187: ...though critical and noncritical exceptions use different save restore register pairs simultaneous occurrences of these exceptions are also processed serially The PPC405 uses the interrupt priority sh...

Страница 188: ...instruction FPU unavailable Attempted execution of an implemented floating point instruction when MSR FP 0 Not implemented by the PPC405D5 APU unavailable Attempted execution of an implemented auxilia...

Страница 189: ...R CE bit is used to enable and disable the interrupt The machine check interrupt can be disabled but the exception is not persistent Machine check exception status is recorded in the machine check int...

Страница 190: ...masked All Others No Change Critical input interrupts are enabled or disabled 15 All 0 Reserved 16 EE All 0 External interrupts are disabled masked 17 PR All 0 Processor is in privileged mode 18 FP A...

Страница 191: ...n address of 27 0x01B Both registers are read and written using the mfspr and mtspr instructions Save Restore Registers 2 and 3 The save restore registers 2 and 3 SRR2 and SRR3 are 32 bit registers us...

Страница 192: ...the EVPR are ignored by the processor The resulting 32 bit exception vector physical address is used by the interrupt mechanism to transfer control to the appropriate interrupt handler Figure 7 4 show...

Страница 193: ...ption occurred 6 PTR Program Trap Instruction 0 Did not occur 1 Occurred When set to 1 indicates a successful trap instruction compare occurred resulting in a trap instruction program exception 7 PEU...

Страница 194: ...the data access instruction that caused one of the following exceptions Alignment exception Data storage exception Data TLB miss exception Figure 7 6 shows the format of the DEAR register The DEAR is...

Страница 195: ...the critical interrupt input signal and forces a critical input interrupt to occur When MSR CE 0 the processor does not recognize the critical interrupt input signal and critical input interrupts cann...

Страница 196: ...has saved SRR2 and SRR3 Saving these registers avoids potential corruption of the interrupt handler should another machine check interrupt occur Instruction Machine Check Interrupt Instruction machine...

Страница 197: ...the machine check exception SRR3 Loaded with a copy of the MSR at the point the interrupt occurs ESR MCI 1 All remaining bits are cleared to 0 DEAR Not used MSR AP APE WE CE EE PR FP ME FE0 DWE DE FE1...

Страница 198: ...depends on the privilege mode In user mode any store or dcbz instruction can cause an exception for this reason No zone protection override can be specified the corresponding zone field value is not e...

Страница 199: ...dccci otherwise 0 DIZ 1 if the exception was caused by a zone protection violation otherwise 0 U0F 1 if the exception was caused by a U0 violation otherwise 0 MCI Unchanged All remaining bits are cle...

Страница 200: ...an only be specified when instruction virtual mode is enabled MSR IR 1 Non executable addresses have the write executable bit TLBLO EX in the corresponding TLB entry cleared to zero No zone protection...

Страница 201: ...cessor recognizes exceptions caused by asserting the noncritical interrupt input signal and forces an external interrupt to occur When MSR EE 0 the processor does not recognize the noncritical interru...

Страница 202: ...t is not aligned on a word boundary Executing an stwcx instruction with an operand that is not aligned on a word boundary From privileged mode MSR PR 0 executing a dcread instruction with an operand t...

Страница 203: ...e should either Replace the trap instruction with a non trapping instruction Modify the trap conditions to prevent a program interrupt Modify the address in SRR0 to point to the next sequential instru...

Страница 204: ...on otherwise 0 This bit is set if software attempts to execute a floating point instruction PPR 1 for attempted execution of a privileged instruction in user mode otherwise 0 PTR 1 for exceptions due...

Страница 205: ...of the PowerPC 405 processor Interrupt Classification Noncritical return using the rfi instruction Synchronous Precise Description FPU unavailable exceptions occur when a program attempts to execute a...

Страница 206: ...e sc instruction provides a means for a user level program to call a privileged system service routine It is assumed that the appropriate linkage information expected by the system call handler is ini...

Страница 207: ...of the PowerPC 405 processor Interrupt Classification Noncritical return using the rfi instruction Synchronous Precise Description APU unavailable exceptions occur when a program attempts to execute a...

Страница 208: ...to clear the PIT to 0 does not cause a PIT interrupt This exception is persistent To prevent repeated interrupts from occurring the interrupt handler must clear the exception status in TSR PIS before...

Страница 209: ...from occurring the interrupt handler must clear the exception status in TSR FIS before returning This interrupt is enabled only by setting both of the following The FIT interrupt enable bit in the tim...

Страница 210: ...og interrupt status bit in the timer status register TSR WIS is cleared to 0 A 0 to 1 transition occurs on the time base bit corresponding to the watchdog time period During the cycle following detect...

Страница 211: ...both A TAG field that matches the data effective address page number EA EPN A TID field that matches the current process ID PID Software cannot disable data TLB miss interrupts See TLB Access page 47...

Страница 212: ...ruction TLB miss exceptions are associated with the fetching of an instruction from memory However an instruction TLB miss interrupt occurs only if an attempt is made to execute the instruction as req...

Страница 213: ...set DBSR IDE This bit is set in addition to other debug status bits When debug interrupts are later enabled the set IDE bit causes a debug interrupt to occur immediately When exiting an interrupt hand...

Страница 214: ...effective address of the instruction following the instruction that caused the debug exception EDE Loaded with the 32 bit exception vector physical address of the exception that caused the debug inte...

Страница 215: ...er event interrupts A watchdog timer interrupt that provides the ability to set critical interrupts that can aid in recovery from system failures A programmable interval timer interrupt that provides...

Страница 216: ...me base The TBU and TBL registers are SPRs with user mode read access and privileged mode write access Reading the time base registers requires use of the move from time base register instruction This...

Страница 217: ...ftbu rz Read TBU again cmpw rz rx Check for TBU rollover by comparing old and new bne loop Read the time base again if a rollover occurred Following is a code example for writing the time base simplif...

Страница 218: ...presents the number of seconds that have elapsed since the fixed reference time posix_ns A 32 bit variable containing the number of nanoseconds that have elapsed since the last time of day calculation...

Страница 219: ...the saved time base value Computing and saving a new value for ticks_per_sec Later calls to compute the time of day can use the updated variables along with the current time base value to calculate t...

Страница 220: ...nction Description 0 1 WP Watchdog Period 00 217 clocks 01 221 clocks 10 225 clocks 11 229 clocks Specifies the period for a watchdog timer event 2 3 WRC Watchdog Reset Control 00 No reset 01 Processo...

Страница 221: ...rrupt is assigned to exception vector offset 0x1000 Fixed interval timer FIT interrupt This noncritical interrupt is assigned to exception vector offset 0x1010 0 1 2 3 4 5 6 31 ENW WIS WRS PIS FIS Fig...

Страница 222: ...l register TCR WP is used to select the TBL bit that controls the time out as shown in Table 8 5 Software cannot disable watchdog time outs This is because the time base register is always incrementin...

Страница 223: ...tchdog time out forces a reset if a reset condition is specified by TCR WRC The processor sets the TSR WIS bit but never clears it Only software can clear the bit Figure 8 6 shows the watchdog event s...

Страница 224: ...reset watchdog interrupts are disabled because MSR CE 0 However watchdog time outs continue to occur because the time base register is always incrementing and a valid watchdog interval is always spec...

Страница 225: ...a valid fixed interval is always specified by TCR FP A FIT event causes a FIT interrupt when both of the following bits are set to 1 The FIT interrupt enable bit in the timer control register TCR FIE...

Страница 226: ...534 www xilinx com March 2002 Release 1 800 255 7778 Virtex II Pro Platform FPGA Documentation Chapter 8 Timer Resources R...

Страница 227: ...ternal halt signal is de asserted Stop Stop can be used to halt the processor using the JTAG port rather than the external halt signal No instructions are executed during a halt but processor register...

Страница 228: ...pping the processor single stepping instruction execution setting breakpoints and monitoring processor status Access to processor resources is provided through the JTAG port External debug events stop...

Страница 229: ...from interrupt instruction the processor re enters the stopped state Debug wait mode is enabled by setting the debug wait mode bit in the MSR MSR DWE 1 Internal debug mode and external debug mode must...

Страница 230: ...not external debug mode is enabled 1 IDM Internal Debug Mode 0 Disabled 1 Enabled Specifies whether or not internal debug mode is enabled 2 3 RST Reset 00 No reset 01 Processor reset 10 Chip reset 11...

Страница 231: ...0 Disabled 1 Enabled Specifies whether or not the instruction address compare 4 IAC4 debug event is enabled 14 IA34 Instruction Address Range Compare 3 4 0 Disabled 1 Enabled Instruction address compa...

Страница 232: ...all bits 01 Ignore least significant bit 10 Ignore least significant two bits 11 Ignore least significant five bits Specifies the granularity of DAC2 exact address comparisons 00 Byte granular 01 Half...

Страница 233: ...h DV2BE bit corresponds to a byte in the DVC2 register DVC2 events are disabled when DV2BE 0b0000 24 31 Reserved Table 9 2 Debug Control Register 1 DBCR1 Field Definitions Continued Bit Name Function...

Страница 234: ...tes whether an IAC2 debug event occurred 7 DR1 Data Address Compare 1 Read Debug Event 0 Did not occur 1 Occurred Indicates whether a DAC1 read debug event occurred 8 DW1 Data Address Compare 1 Write...

Страница 235: ...data value compare registers DVC1 and DVC2 These registers are used by the data value compare debug event Figure 9 5 shows the format of the DVCn registers Any data value can be loaded in these regist...

Страница 236: ...rigger events are used by external tools to collect instruction trace information Debug status is recorded in the debug status register DBSR A debug event can set debug status bits even if all debug m...

Страница 237: ...r than the debug exception Instructions that cause an exception do not result in an IC debug event This sc instruction however causes a system call exception after it executes Here the debug event occ...

Страница 238: ...ption handler is executed It is enabled by setting DBCR0 EDE 1 and disabled by clearing DBCR0 EDE 0 The processor reports the occurrence of an EDE debug event by setting the EDE bit in the debug statu...

Страница 239: ...that would have executed had the UDE event not occurred Instruction Address Compare Debug Event An instruction address compare IAC debug event occurs immediately before executing an instruction The ef...

Страница 240: ...is detected the IAn enable bits in DBCR0 determine which DBSR status bits are set to 1 For example both DBSR IA1 IA2 are set to 1 if DBCR0 IA1 IA2 1 when an IA12 address range match is detected Howeve...

Страница 241: ...is in the exclusive IA12 range The processor clears IA12X to 0 The second IAC event occurs when an instruction address is in the inclusive IA12 range The processor sets IA12X to 0 The third IAC event...

Страница 242: ...DACn register used in the comparison and the debug status register bit set when the event occurs Any number of DAC exact address match conditions can be enabled simultaneously DAC address range compar...

Страница 243: ...by a string instruction In some cases software must use a word size granularity to produce a DAC match on a specific byte address DAC Address Range Match A DAC address range match causes a debug event...

Страница 244: ...es that are out of range Table 9 9 summarizes the DBCR1 bits used to control DAC address range comparisons and the DBSR bits used to report their status The processor does not clear the DBSR status bi...

Страница 245: ...ed Instructions in this category are dcbf and dcbst Cache control instructions that are speculative are treated as loads by the debug mechanism These instructions can cause DAC read events Instruction...

Страница 246: ...page 543 the byte enables specify which DVCn register bytes participate in the DVC comparison DVnBE0 controls participation of DVCn data value byte 0 DVnBE1 controls participation of DVCn data value...

Страница 247: ...DBSR status bits are set to reflect the event Read and write DVC events are recorded independently using the DRn and DWn status bits Table 9 12 summarizes how the status bits are used by DVC events St...

Страница 248: ...use imprecise debug events when MSR DE 0 Branch taken BT if DBCR0 IDM 0 If internal debug mode is enabled BT events cannot cause imprecise debug events when MSR DE 0 Exception taken EDE Trap instructi...

Страница 249: ...wing JTAG Signals The JTAG debug port implements the four required JTAG signals TCK TMS TDI and TDO It also implements the optional TRST signal JTAG Clock The frequency of the JTAG clock signal TCK ca...

Страница 250: ...n 11 I HALT Processor halt 12 NC Reserved no connection 13 NC Reserved no connection 14 KEY No pin should be placed at this position 15 NC Reserved no connection 16 GND Ground Notes 1 A 10K pull up re...

Страница 251: ...escribe different packages Every pin map within the BSDL description is given a unique name The instruction statements which describe bit patterns that must be shifted into the instruction register to...

Страница 252: ...560 www xilinx com March 2002 Release 1 800 255 7778 Virtex II Pro Platform FPGA Documentation R...

Страница 253: ...at are connected to the same system reset network The scope of a system reset depends on the system implementation The type of reset is recorded in the most recent reset field of the debug status regi...

Страница 254: ...reset on the processor state as controlled by the MSR Special Purpose Registers Table 10 2 shows the contents of the special purpose registers SPRs that have defined values following a reset The cont...

Страница 255: ...ization code should follow when performing this configuration 1 Configure the real mode memory system by updating the storage attribute control registers After reset all memory is marked as guarded st...

Страница 256: ...er is cleared Software should update this register as appropriate to enable instruction caching 4 Configure the data cache to improve data access performance Like the instruction cache the data cache...

Страница 257: ...code is presented as pseudocode Where appropriate function calls are given names similar to the PowerPC instruction mnemonics Specific chip level implementations containing the PPC405 might require a...

Страница 258: ...errupt handlers Initialize interrupt vector table Initialize exception vector prefix mtspr EVPR prefix_addr Prepare system for asynchronous interrupts Initialize and configure timer resources mtspr PI...

Страница 259: ...02 Release www xilinx com 567 Virtex II Pro Platform FPGA Documentation 1 800 255 7778 R Initialize other processor resources Initialize non processor resources Branch to operating system or applicati...

Страница 260: ...568 www xilinx com March 2002 Release 1 800 255 7778 Virtex II Pro Platform FPGA Documentation R...

Страница 261: ...is Shift Right Algebraic Word Immediate Figure 11 1 Instruction Description Format UG011_50_033101 add Add Description The sum of the contents of register rA and register rB is loaded into register rD...

Страница 262: ...age 317 Registers Altered A summary of the PowerPC registers that are modified by executing the instruction Exceptions A list of the exceptions that can occur as a result of executing the instruction...

Страница 263: ...F is the split field SPRF corresponds to SPRN as follows SPRF0 4 is equivalent to SPRN5 9 SPRF5 9 is equivalent to SPRN0 4 In the mfdcr and mtdcr instructions DCRN is the assembler operand and DCRF is...

Страница 264: ...CR CR0 LT GT EQ SO if Rc 1 XER SO OV if OE 1 If an overflow occurs it is possible that the contents of CR0 do not reflect the infinitely precise result Exceptions None Compatibility This instruction i...

Страница 265: ...magnitude of the resulting sum Pseudocode rD rA rB if rD 232 1 then XER CA 1 else XER CA 0 Registers Altered rD XER CA CR CR0 LT GT EQ SO if Rc 1 XER SO OV if OE 1 Exceptions None Compatibility This i...

Страница 266: ...n be used to perform addition on integers larger than 32 bits as described on page 390 Pseudocode rD rA rB XER CA if rD 232 1 then XER CA 1 else XER CA 0 Registers Altered rD XER CA CR CR0 LT GT EQ SO...

Страница 267: ...and added to the contents of register rA The resulting sum is loaded into register rD Simplified mnemonics defined for this instruction are described in the following sections Load Address page 834 Lo...

Страница 268: ...D XER CA is updated to reflect the unsigned magnitude of the resulting sum Simplified mnemonics defined for this instruction are described in Subtract Instructions page 831 Pseudocode rD rA EXTS SIMM...

Страница 269: ...ng sum addic is one of three instructions that implicitly update CR CR0 without having an RC field The other instructions are andi and andis Simplified mnemonics defined for this instruction are descr...

Страница 270: ...sum is loaded into register rD Simplified mnemonics defined for this instruction are described in the following sections Load Immediate page 834 Subtract Instructions page 831 An addis instruction fol...

Страница 271: ...ed on page 390 Pseudocode rD rA XER CA 1 if rD 232 1 then XER CA 1 else XER CA 0 Registers Altered rD XER CA CR CR0 LT GT EQ SO if Rc 1 XER SO OV if OE 1 Exceptions None Execution of any of the follow...

Страница 272: ...Pseudocode rD rA XER CA if rD 232 1 then XER CA 1 else XER CA 0 Registers Altered rD XER CA CR CR0 LT GT EQ SO if Rc 1 XER SO OV if OE 1 Exceptions None Execution of any of the following invalid instr...

Страница 273: ...th the contents of register rB and the result is loaded into register rA Pseudocode rA rS rB Registers Altered rA CR CR0 LT GT EQ SO if Rc 1 Exceptions None Compatibility This instruction is defined b...

Страница 274: ...ne s complement of the contents of register rB and the result is loaded into register rA Pseudocode rA rS rB Registers Altered rA CR CR0 LT GT EQ SO if Rc 1 Exceptions None Compatibility This instruct...

Страница 275: ...gister rA andi is one of three instructions that implicitly update CR CR0 without having an Rc field The other instructions are addic and andis The andi instruction can be used to test whether any of...

Страница 276: ...register rA andis is one of three instructions that implicitly update CR CR0 without having an Rc field The other instructions are addic and andi The andis instruction can be used to test whether any...

Страница 277: ...bits If the AA field contains 0 relative addressing the branch instruction address is used as the base address The branch instruction address is the current instruction address CIA If the AA field co...

Страница 278: ...ogram flow is transferred to the NIA If the LK field contains 1 then the address of the instruction following the branch instruction CIA 4 is loaded into the LR The BO field specifies whether the bran...

Страница 279: ...Pro Platform FPGA Documentation 1 800 255 7778 Alphabetical Instruction Listing R LR if LK 1 Exceptions None Compatibility This instruction is defined by the PowerPC user instruction set architecture...

Страница 280: ...is decremented The encoding of the BO field is described in Conditional Branch Control page 367 The BI field specifies which CR bit is tested if the branch is conditional on the CR register Simplifie...

Страница 281: ...tform FPGA Documentation 1 800 255 7778 Alphabetical Instruction Listing R the decremented CTR are used as the NIA Compatibility This instruction is defined by the PowerPC user instruction set archite...

Страница 282: ...The BO field also specifies whether the CTR is decremented The encoding of the BO field is described in Conditional Branch Control page 367 The BI field specifies which CR bit is tested if the branch...

Страница 283: ...91 Virtex II Pro Platform FPGA Documentation 1 800 255 7778 Alphabetical Instruction Listing R Compatibility This instruction is defined by the PowerPC user instruction set architecture UISA It is imp...

Страница 284: ...cribed in Compare Instructions page 828 Pseudocode c0 3 0b0000 if rA rB then c0 1 if rA rB then c1 1 if rA rB then c2 1 c3 XER SO n crfD CR CRn c0 3 Registers Altered CR CRn as specified by the crfD f...

Страница 285: ...defined for this instruction are described in Compare Instructions page 828 Pseudocode c0 3 0b0000 if rA EXTS SIMM then c0 1 if rA EXTS SIMM then c1 1 if rA EXTS SIMM then c2 1 c3 XER SO n crfD CR CRn...

Страница 286: ...are described in Compare Instructions page 828 Pseudocode c0 3 0b0000 if rA rB then c0 1 if rA rB then c1 1 if rA rB then c2 1 c3 XER SO n crfD CR CRn c0 3 Registers Altered CR CRn as specified by th...

Страница 287: ...implified mnemonics defined for this instruction are described in Compare Instructions page 828 Pseudocode c0 3 0b0000 if rA 160 UIMM then c0 1 if rA 160 UIMM then c1 1 if rA 160 UIMM then c2 1 c3 XER...

Страница 288: ...0 do while n 32 if rS n 1 then leave n n 1 rA n Registers Altered rA CR CR0 LT GT EQ SO if Rc 1 Exceptions None Execution of any of the following invalid instruction forms results in a boundedly undef...

Страница 289: ...ecified by crbD Pseudocode CR crbD CR crbA CR crbB Registers Altered CR Exceptions None Execution of any of the following invalid instruction forms results in a boundedly undefined result rather than...

Страница 290: ...nto the CR bit specified by crbD Pseudocode CR crbD CR crbA CR crbB Registers Altered CR Exceptions None Execution of any of the following invalid instruction forms results in a boundedly undefined re...

Страница 291: ...s defined for this instruction are described in CR Logical Instructions page 828 Pseudocode CR crbD CR crbA CR crbB Registers Altered CR Exceptions None Execution of any of the following invalid instr...

Страница 292: ...CR bit specified by crbD Pseudocode CR crbD CR crbA CR crbB Registers Altered CR Exceptions None Execution of any of the following invalid instruction forms results in a boundedly undefined result rat...

Страница 293: ...efined for this instruction are described in CR Logical Instructions page 828 Pseudocode CR crbD CR crbA CR crbB Registers Altered CR Exceptions None Execution of any of the following invalid instruct...

Страница 294: ...instruction are described in CR Logical Instructions page 828 Pseudocode CR crbD CR crbA CR crbB Registers Altered CR Exceptions None Execution of any of the following invalid instruction forms result...

Страница 295: ...into the CR bit specified by crbD Pseudocode CR crbD CR crbA CR crbB Registers Altered CR Exceptions None Execution of any of the following invalid instruction forms results in a boundedly undefined...

Страница 296: ...instruction are described in CR Logical Instructions page 828 Pseudocode CR crbD CR crbA CR crbB Registers Altered CR Exceptions None Execution of any of the following invalid instruction forms resul...

Страница 297: ...a cacheline is allocated and the value of the bytes in that line are undefined If EA is cachable and has a write through caching policy a no operation occurs This is true whether or not EA is cached b...

Страница 298: ...dered a store with respect to data address compare DAC debug exceptions Debug exceptions can occur as a result of executing this instruction Execution of any of the following invalid instruction forms...

Страница 299: ...attribute indicates EA is cachable If EA is not cached no operation is performed Pseudocode EA rA 0 rB Flush data cacheline corresponding to EA Registers Altered None Exceptions Data storage if the ac...

Страница 300: ...Pro Platform FPGA Documentation Chapter 11 Instruction Set R Compatibility This instruction is defined by the virtual environment architecture level VEA of the PowerPC architecture the PowerPC embedde...

Страница 301: ...Pseudocode EA rA 0 rB Invalidate data cacheline corresponding to EA Registers Altered None Exceptions Data storage if the access is prevented by zone protection when data relocation is enabled No acc...

Страница 302: ...lowing invalid instruction forms results in a boundedly undefined result rather than a program exception Reserved bits containing a non zero value Compatibility This instruction is defined by the oper...

Страница 303: ...peration occurs if the data cacheline is unmodified or if EA is not cached Pseudocode EA rA 0 rB Store modified data cacheline corresponding to EA Registers Altered None Exceptions Data storage if the...

Страница 304: ...Pro Platform FPGA Documentation Chapter 11 Instruction Set R Compatibility This instruction is defined by the virtual environment architecture level VEA of the PowerPC architecture the PowerPC embedde...

Страница 305: ...m will likely load data from EA in the near future The processor can potentially improve performance by loading the cacheline into the data cache Pseudocode EA rA 0 rB Prefetch data cacheline correspo...

Страница 306: ...Pro Platform FPGA Documentation Chapter 11 Instruction Set R Compatibility This instruction is defined by the virtual environment architecture level VEA of the PowerPC architecture the PowerPC embedde...

Страница 307: ...oading the cacheline into the data cache In the PPC405 this instruction operates identically to dcbt In other PowerPC implementations this instruction can cause unique bus cycles to occur and addition...

Страница 308: ...Pro Platform FPGA Documentation Chapter 11 Instruction Set R Compatibility This instruction is defined by the virtual environment architecture level VEA of the PowerPC architecture the PowerPC embedde...

Страница 309: ...eline is allocated and the value of the bytes in that line are cleared to 0 The data cacheline is marked modified If EA is cachable and has a write through caching policy an alignment exception occurs...

Страница 310: ...tion is enabled and a valid translation entry corresponding to the EA is not found in the TLB This instruction is considered a store with respect to the above data access exceptions It is also conside...

Страница 311: ...chelines it is lost This instruction is intended for use during initialization to invalidate the entire data cache before is enabled A sequence of dccci instructions should be executed one for each co...

Страница 312: ...en though the instruction is not address specific multiple addresses are selected by a single EA This instruction does not cause data address compare DAC debug exceptions Execution of any of the follo...

Страница 313: ...congruence class If CCR0 CWS 0 the line in way A is selected If CCR0 CWS 1 the line in way B is selected If CCR0 CIS 0 the information read is a word of data from the selected cacheline EA27 29 is use...

Страница 314: ...data access exceptions It can cause data TLB miss exceptions related to EA even though the instruction is not address specific multiple addresses are selected by a single EA This instruction cannot ca...

Страница 315: ...dividend is negative The 32 bit remainder can be calculated using the following sequence of instructions divw rD rA rB rD quotient mullw rD rD rB rD quotient divisor subf rD rD rA rD remainder The con...

Страница 316: ...lease 1 800 255 7778 Virtex II Pro Platform FPGA Documentation Chapter 11 Instruction Set R Compatibility This instruction is defined by the PowerPC user instruction set architecture UISA It is implem...

Страница 317: ...rD remainder If Rc 1 CR CR0 LT GT EQ are set using a signed comparison of the result to 0 even though the instruction produces an unsigned integer as a quotient The contents of register rD are undefin...

Страница 318: ...struction completion Non memory instructions following eieio can complete before the memory operations ordered by eieio The sync instruction is used to guarantee ordering of both instruction completio...

Страница 319: ...nts of register rB A one s complement of the result is calculated and loaded in register rA Pseudocode rA rS rB Registers Altered rA CR CR0 LT GT EQ SO if Rc 1 Exceptions None Compatibility This instr...

Страница 320: ...register rA Pseudocode rA EXTS rS24 31 Registers Altered rA CR CR0 LT GT EQ SO if Rc 1 Exceptions None Execution of any of the following invalid instruction forms results in a boundedly undefined resu...

Страница 321: ...d into register rA Pseudocode rA EXTS rS16 31 Registers Altered rA CR CR0 LT GT EQ SO if Rc 1 Exceptions None Execution of any of the following invalid instruction forms results in a boundedly undefin...

Страница 322: ...f the access is prevented by no access allowed zone protection This only applies to accesses in user mode when data relocation is enabled Data TLB miss if data relocation is enabled and a valid transl...

Страница 323: ...Documentation 1 800 255 7778 Alphabetical Instruction Listing R Compatibility This instruction is defined by the virtual environment architecture level VEA of the PowerPC architecture the PowerPC embe...

Страница 324: ...the cacheline into the instruction cache Pseudocode EA rA 0 rB Prefetch instruction cacheline corresponding to EA Registers Altered None Exceptions This instruction is considered a load with respect t...

Страница 325: ...l Instruction Listing R Compatibility This instruction is defined by the virtual environment architecture level VEA of the PowerPC embedded environment architecture and the PowerPC Book E architecture...

Страница 326: ...use during initialization to invalidate the entire instruction cache before is enabled Pseudocode Invalidate the instruction cache Registers Altered None Exceptions Program Attempted execution of thi...

Страница 327: ...CCR0 CWS is used to select one of the two cachelines within the congruence class If CCR0 CWS 0 the line in way A is selected If CCR0 CWS 1 the line in way B is selected If CCR0 CIS 0 the information r...

Страница 328: ...use data TLB miss exceptions related to the EA even though the instruction is not address specific multiple addresses are selected by a single EA This instruction cannot cause data storage exceptions...

Страница 329: ...etes execution Prefetched instructions are discarded by the execution of isync All instructions following isync are executed in the context established by the instructions preceding the isync isync do...

Страница 330: ...ress The byte referenced by EA is extended to 32 bits by concatenating 24 0 bits on the left The result is loaded into register rD Pseudocode EA rA 0 EXTS d rD 240 MS EA 1 Registers Altered rD Excepti...

Страница 331: ...into register rD The EA is loaded into rA Pseudocode EA rA EXTS d rD 240 MS EA 1 rA EA Registers Altered rA rD Exceptions Data storage if the access is prevented by no access allowed zone protection...

Страница 332: ...eudocode EA rA rB rD 240 MS EA 1 rA EA Registers Altered rA rD Exceptions Data storage if the access is prevented by no access allowed zone protection This only applies to accesses in user mode when d...

Страница 333: ...s loaded into register rD Pseudocode EA rA 0 rB rD 240 MS EA 1 Registers Altered rD Exceptions Data storage if the access is prevented by no access allowed zone protection This only applies to accesse...

Страница 334: ...e used as the base address The halfword referenced by EA is sign extended to 32 bits and loaded into register rD Pseudocode EA rA 0 EXTS d rD EXTS MS EA 2 Registers Altered rD Exceptions Data storage...

Страница 335: ...The EA is loaded into rA Pseudocode EA rA EXTS d rD EXTS MS EA 2 rA EA Registers Altered rA rD Exceptions Data storage if the access is prevented by no access allowed zone protection This only applies...

Страница 336: ...rD EXTS MS EA 2 rA EA Registers Altered rA rD Exceptions Data storage if the access is prevented by no access allowed zone protection This only applies to accesses in user mode when data relocation is...

Страница 337: ...ter rD Pseudocode EA rA 0 rB rD EXTS MS EA 2 Registers Altered rD Exceptions Data storage if the access is prevented by no access allowed zone protection This only applies to accesses in user mode whe...

Страница 338: ...memory word are loaded into rD 24 31 Bits 8 15 of the memory word are loaded into rD 16 23 16 0 bits are loaded into rD 0 15 Pseudocode EA rA 0 rB rD 160 MS EA 1 1 MS EA 1 Registers Altered rD Except...

Страница 339: ...address The halfword referenced by EA is extended to 32 bits by concatenating 16 0 bits on the left The result is loaded into register rD Pseudocode EA rA 0 EXTS d rD 160 MS EA 2 Registers Altered rD...

Страница 340: ...d into register rD The EA is loaded into rA Pseudocode EA rA EXTS d rD 160 MS EA 2 rA EA Registers Altered rA rD Exceptions Data storage if the access is prevented by no access allowed zone protection...

Страница 341: ...o rA Pseudocode EA rA rB rD 160 MS EA 2 rA EA Registers Altered rA rD Exceptions Data storage if the access is prevented by no access allowed zone protection This only applies to accesses in user mode...

Страница 342: ...s loaded into register rD Pseudocode EA rA 0 rB rD 160 MS EA 2 Registers Altered rD Exceptions Data storage if the access is prevented by no access allowed zone protection This only applies to accesse...

Страница 343: ...are loaded into GPRs rD through r31 Pseudocode EA rA 0 EXTS d n rD do while n 31 if n rA n 31 then GPR n MS EA 4 n n 1 EA EA 4 Registers Altered rD through r31 Exceptions Data storage if the access i...

Страница 344: ...lease 1 800 255 7778 Virtex II Pro Platform FPGA Documentation Chapter 11 Instruction Set R Compatibility This instruction is defined by the PowerPC user instruction set architecture UISA It is implem...

Страница 345: ...tive bytes starting at the memory address referenced by EA are loaded into GPRs rD through RFINAL The sequence of registers wraps around to r0 if necessary RFINAL rD nr 1 modulo 32 Bytes are loaded in...

Страница 346: ...miss if data relocation is enabled and a valid translation entry corresponding to the EA is not found in the TLB Execution of any of the following invalid instruction forms results in a boundedly und...

Страница 347: ...data n consecutive bytes starting at the memory address referenced by EA are loaded into GPRs rD through RFINAL The sequence of registers wraps around to r0 if necessary RFINAL rD nr 1 modulo 32 Bytes...

Страница 348: ...r However a data machine check exception can occur when XER TBC 0 if the following conditions are true The instruction access passes all protection checks The data address is cachable Access of the da...

Страница 349: ...or accessing a semaphore See Semaphore Synchronization page 426 for more information Pseudocode EA rA 0 rB rD MS EA 4 RESERVE 1 Registers Altered rD Exceptions Alignment if the EA is not aligned on a...

Страница 350: ...into rD 16 23 Bits 16 23 of the memory word are loaded into rD 8 15 Bits 23 31 of the memory word are loaded into rD 0 7 Pseudocode EA rA 0 rB rD MS EA 3 1 MS EA 2 1 MS EA 1 1 MS EA 1 Registers Altere...

Страница 351: ...s of register rA are used as the base address The word referenced by EA is loaded into register rD Pseudocode EA rA 0 EXTS d rD MS EA 4 Registers Altered rD Exceptions Data storage if the access is pr...

Страница 352: ...to rA Pseudocode EA rA EXTS d rD MS EA 4 rA EA Registers Altered rA rD Exceptions Data storage if the access is prevented by no access allowed zone protection This only applies to accesses in user mod...

Страница 353: ...EA Registers Altered rA rD Exceptions Data storage if the access is prevented by no access allowed zone protection This only applies to accesses in user mode when data relocation is enabled Data TLB...

Страница 354: ...A rA 0 rB rD MS EA 4 Registers Altered rD Exceptions Data storage if the access is prevented by no access allowed zone protection This only applies to accesses in user mode when data relocation is ena...

Страница 355: ...tents of rD are replaced by the low order 32 bits of the temporary result An example of this operation is shown in Figure 3 28 page 408 Pseudocode prod0 31 rA 16 31 rB 0 15 signed temp0 32 prod0 31 rD...

Страница 356: ...earest representable value If the result is less than 231 the value stored in rD is 231 If the result is greater than 231 1 the value stored in rD is 231 1 An example of this operation is shown in Fig...

Страница 357: ...n rD If the result overflows rD is loaded with the nearest representable value If the result is greater than 232 1 the value stored in rD is 232 1 An example of this operation is shown in Figure 3 28...

Страница 358: ...ts of rD are replaced by the low order 32 bits of the temporary result An example of this operation is shown in Figure 3 28 page 408 Pseudocode prod0 31 rA 16 31 rB 0 15 unsigned temp0 32 prod0 31 rD...

Страница 359: ...ntents of rD are replaced by the low order 32 bits of the temporary result An example of this operation is shown in Figure 3 29 page 410 Pseudocode prod0 31 rA 0 15 rB 0 15 signed temp0 32 prod0 31 rD...

Страница 360: ...earest representable value If the result is less than 231 the value stored in rD is 231 If the result is greater than 231 1 the value stored in rD is 231 1 An example of this operation is shown in Fig...

Страница 361: ...in rD If the result overflows rD is loaded with the nearest representable value If the result is greater than 232 1 the value stored in rD is 232 1 An example of this operation is shown in Figure 3 29...

Страница 362: ...nts of rD are replaced by the low order 32 bits of the temporary result An example of this operation is shown in Figure 3 29 page 410 Pseudocode prod0 31 rA 0 15 rB 0 15 unsigned temp0 32 prod0 31 rD...

Страница 363: ...nts of rD are replaced by the low order 32 bits of the temporary result An example of this operation is shown in Figure 3 30 page 413 Pseudocode prod0 31 rA 16 31 rB 16 31 signed temp0 32 prod0 31 rD...

Страница 364: ...rest representable value If the result is less than 231 the value stored in rD is 231 If the result is greater than 231 1 the value stored in rD is 231 1 An example of this operation is shown in Figur...

Страница 365: ...rD If the result overflows rD is loaded with the nearest representable value If the result is greater than 232 1 the value stored in rD is 232 1 An example of this operation is shown in Figure 3 30 pa...

Страница 366: ...of rD are replaced by the low order 32 bits of the temporary result An example of this operation is shown in Figure 3 30 page 413 Pseudocode prod0 31 rA 16 31 rB 16 31 unsigned temp0 32 prod0 31 rD r...

Страница 367: ...fD CR CRn CR CRm Registers Altered CR CRn where n is specified by crfD Exceptions None Execution of any of the following invalid instruction forms results in a boundedly undefined result rather than a...

Страница 368: ...CR CRn XER0 3 XER0 3 0b0000 Registers Altered CR CRn where n is specified by the crfD field XER0 3 Exceptions None Execution of any of the following invalid instruction forms results in a boundedly un...

Страница 369: ...ode rD CR Registers Altered rD Exceptions None Execution of any of the following invalid instruction forms results in a boundedly undefined result rather than a program exception Reserved bits contain...

Страница 370: ...d rD Exceptions Program Attempted execution of this instruction from user mode Execution of any of the following invalid instruction forms results in a boundedly undefined result rather than a program...

Страница 371: ...ution of this instruction from user mode Execution of any of the following invalid instruction forms results in a boundedly undefined result rather than a program exception Reserved bits containing a...

Страница 372: ...SPRF0 4 rD SPR SPRN Registers Altered rD Exceptions Program Attempted execution of this instruction from user mode if SPRF 0 bit 11 of the instruction opcode is 1 Execution of any of the following in...

Страница 373: ...pose Registers page 830 Pseudocode TBRN TBRF5 9 TBRF0 4 rD TBR TBRN Registers Altered rD Exceptions Execution of any of the following invalid instruction forms results in a boundedly undefined result...

Страница 374: ...not copied and the corresponding CR bits are unchanged The following table shows the relationship between the CRM field and the rS and CR registers The CRn field is shown for completeness See mtcrf Fi...

Страница 375: ...Exceptions None Execution of any of the following invalid instruction forms results in a boundedly undefined result rather than a program exception Reserved bits containing a non zero value Compatibil...

Страница 376: ...R DCRN Exceptions Program Attempted execution of this instruction from user mode Execution of any of the following invalid instruction forms results in a boundedly undefined result rather than a progr...

Страница 377: ...tion of this instruction from user mode Execution of any of the following invalid instruction forms results in a boundedly undefined result rather than a program exception Reserved bits containing a n...

Страница 378: ...SPRF0 4 SPR SPRN rS Registers Altered SPR SPRN Exceptions Program Attempted execution of this instruction from user mode if SPRF 0 bit 11 of the instruction is 1 Execution of any of the following inv...

Страница 379: ...f rB The resulting signed 32 bit product is loaded into register rD An example of this operation is shown in Figure 3 34 page 420 Pseudocode rD 0 31 rA 16 31 rB 0 15 signed Registers Altered rD CR CR0...

Страница 380: ...The resulting unsigned 32 bit product is loaded into register rD An example of this operation is shown in Figure 3 34 page 420 Pseudocode rD 0 31 rA 16 31 rB 0 15 unsigned Registers Altered rD CR CR0...

Страница 381: ...of rB The resulting signed 32 bit product is loaded into register rD An example of this operation is shown in Figure 3 35 page 421 Pseudocode rD 0 31 rA 0 15 rB 0 15 signed Registers Altered rD CR CR0...

Страница 382: ...B The resulting unsigned 32 bit product is loaded into register rD An example of this operation is shown in Figure 3 35 page 421 Pseudocode rD 0 31 rA 0 15 rB 0 15 unsigned Registers Altered rD CR CR0...

Страница 383: ...quantities This instruction can be used with mullw or mulli to calculate a full 64 bit product Pseudocode prod0 63 rA rB signed rD prod0 31 Registers Altered rD CR CR0 LT GT EQ SO if Rc 1 Exceptions...

Страница 384: ...he operands are to be interpreted as signed quantities Pseudocode prod0 63 rA rB unsigned rD prod0 31 Registers Altered rD CR CR0 LT GT EQ SO if Rc 1 Exceptions Execution of any of the following inval...

Страница 385: ...rB The resulting signed 32 bit product is loaded into register rD An example of this operation is shown in Figure 3 36 page 422 Pseudocode rD 0 31 rA 16 31 rB 16 31 signed Registers Altered rD CR CR0...

Страница 386: ...ulting unsigned 32 bit product is loaded into register rD An example of this operation is shown in Figure 3 36 page 422 Pseudocode rD 0 31 rA 16 31 rB 16 31 unsigned Registers Altered rD CR CR0 LT GT...

Страница 387: ...loaded into register rD The result loaded into register rD is always correct regardless of whether the operands are interpreted as signed or unsigned integers This instruction can be used with mulhw t...

Страница 388: ...s are interpreted as signed integers The result loaded into register rD is always correct regardless of whether the operands are interpreted as signed or unsigned integers This instruction can be used...

Страница 389: ...plement of the result is loaded into register rA The one s complement of a number can be obtained using nand with rS rB Pseudocode rA rS rB Registers Altered rA CR CR0 LT GT EQ SO if Rc 1 Exceptions N...

Страница 390: ...de rD rA 1 Registers Altered rD CR CR0 LT GT EQ SO if Rc 1 XER SO OV if OE 1 Exceptions Execution of any of the following invalid instruction forms results in a boundedly undefined result rather than...

Страница 391: ...f rD are replaced by the low order 32 bits of the temporary result An example of this operation is shown in Figure 3 31 page 415 Pseudocode prod0 31 rA 16 31 rB 0 15 signed nprod0 31 1 prod0 31 signed...

Страница 392: ...epresentable value If the result is less than 231 the value stored in rD is 231 If the result is greater than 231 1 the value stored in rD is 231 1 An example of this operation is shown in Figure 3 31...

Страница 393: ...of rD are replaced by the low order 32 bits of the temporary result An example of this operation is shown in Figure 3 32 page 417 Pseudocode prod0 31 rA 0 15 rB 0 15 signed nprod0 31 1 prod0 31 signed...

Страница 394: ...epresentable value If the result is less than 231 the value stored in rD is 231 If the result is greater than 231 1 the value stored in rD is 231 1 An example of this operation is shown in Figure 3 32...

Страница 395: ...rD are replaced by the low order 32 bits of the temporary result An example of this operation is shown in Figure 3 33 page 419 Pseudocode prod0 31 rA 16 31 rB 16 31 signed nprod0 31 1 prod0 31 signed...

Страница 396: ...resentable value If the result is less than 231 the value stored in rD is 231 If the result is greater than 231 1 the value stored in rD is 231 1 An example of this operation is shown in Figure 3 33 p...

Страница 397: ...e one s complement of a number can be obtained using nor with rS rB Simplified mnemonics defined for this instruction are described in Other Simplified Mnemonics page 834 Pseudocode rA rS rB Registers...

Страница 398: ...ter can be copied into another register using or with rS rB Simplified mnemonics defined for this instruction are described in Other Simplified Mnemonics page 834 Pseudocode rA rS rB Registers Altered...

Страница 399: ...e one s complement of the contents of register rB and the result is loaded into register rA Pseudocode rA rS rB Registers Altered rA CR CR0 LT GT EQ SO if Rc 1 Exceptions None Compatibility This instr...

Страница 400: ...M field and the result is loaded into register rA Simplified mnemonics defined for this instruction are described in Other Simplified Mnemonics page 834 The preferred no operation an instruction that...

Страница 401: ...atenating 16 0 bits on the right The contents of the register rS are ORed with the extended UIMM field and the result is loaded into register rA Pseudocode rA rS UIMM 160 Registers Altered rA Exceptio...

Страница 402: ...docode MSR SRR3 Synchronize context NIA SRR2 Registers Altered MSR Exceptions Program Attempted execution of this instruction from user mode Execution of any of the following invalid instruction forms...

Страница 403: ...into the MSR Pseudocode MSR SRR1 Synchronize context NIA SRR0 Registers Altered MSR Exceptions Program Attempted execution of this instruction from user mode Execution of any of the following invalid...

Страница 404: ...ed data is inserted into register rA under control of the mask If a mask bit contains a 1 the corresponding bit in the rotated data is inserted into the corresponding bit of register rA If a mask bit...

Страница 405: ...The contents of register rS are rotated left by the number of bit positions specified by the SH field The rotated data is ANDed with the mask and the result is loaded into register rA This instructio...

Страница 406: ...of register rS are rotated left by the number of bit positions specified by the contents of register rB27 31 The rotated data is ANDed with the mask and the result is loaded into register rA This inst...

Страница 407: ...truction is context synchronizing Instructions fetched from the NIA use the new context loaded into the MSR Pseudocode SRR1 MSR MSR WE EE PR DR IR 0 SRR0 CIA 4 Synchronize context NIA EVPR0 15 0x0C00...

Страница 408: ...are lost and 0 bits fill vacated bit positions on the right The result is loaded into register rA If rB26 1 register rA is cleared to zero Pseudocode n rB 27 31 r ROTL rS n if rB 26 0 then m MASK 0 3...

Страница 409: ...ded into register rA If rS contains a negative number and any 1 bits are shifted out of the least significant bit position XER CA is set to 1 Otherwise XER CA is cleared to 0 If rB26 1 XER CA and all...

Страница 410: ...d bit positions on the left The result is loaded into register rA If rS contains a negative number and any 1 bits are shifted out of the least significant bit position XER CA is set to 1 Otherwise XER...

Страница 411: ...bit are lost and 0 bits fill the vacated bit positions on the left The result is loaded into register rA If rB26 1 register rA is cleared to 0 Pseudocode n rB 27 31 r ROTL rS 32 n if rB 26 0 then m M...

Страница 412: ...cant byte of register rS is stored into the byte referenced by EA Pseudocode EA rA 0 EXTS d MS EA 1 rS 24 31 Registers Altered None Exceptions Data storage if the access is prevented by zone protectio...

Страница 413: ...EXTS d MS EA 1 rS 24 31 rA EA Registers Altered rA Exceptions Data storage if the access is prevented by zone protection when data relocation is enabled No access allowed zone protection applies only...

Страница 414: ...rA Exceptions Data storage if the access is prevented by zone protection when data relocation is enabled No access allowed zone protection applies only to accesses in user mode Read only zone protecti...

Страница 415: ...Registers Altered None Exceptions Data storage if the access is prevented by zone protection when data relocation is enabled No access allowed zone protection applies only to accesses in user mode Rea...

Страница 416: ...nt halfword of register rS is stored into the halfword referenced by EA Pseudocode EA rA 0 EXTS d MS EA 2 rS 16 31 Registers Altered None Exceptions Data storage if the access is prevented by zone pro...

Страница 417: ...d into the byte referenced by EA 1 Pseudocode EA rA 0 rB MS EA 2 rS 24 31 rS 16 23 Registers Altered None Exceptions Data storage if the access is prevented by zone protection when data relocation is...

Страница 418: ...rA EXTS d MS EA 2 rS 16 31 rA EA Registers Altered rA Exceptions Data storage if the access is prevented by zone protection when data relocation is enabled No access allowed zone protection applies o...

Страница 419: ...Altered rA Exceptions Data storage if the access is prevented by zone protection when data relocation is enabled No access allowed zone protection applies only to accesses in user mode Read only zone...

Страница 420: ...31 Registers Altered None Exceptions Data storage if the access is prevented by zone protection when data relocation is enabled No access allowed zone protection applies only to accesses in user mode...

Страница 421: ...d into n consecutive words starting at the memory address referenced by EA Pseudocode EA rA 0 EXTS d r rS do while r 31 MS EA 4 GPR r r r 1 EA EA 4 Registers Altered None Exceptions Data storage if th...

Страница 422: ...4 GPRs rS through rS nr 1 are stored into n consecutive bytes starting at the memory address referenced by EA The sequence of registers wraps around to r0 if necessary The bytes within each register...

Страница 423: ...ivileged modes Data TLB miss if data relocation is enabled and a valid translation entry corresponding to the EA is not found in the TLB Execution of any of the following invalid instruction forms res...

Страница 424: ...TBC Let nr specify the number of registers to load with data nr CEIL n 4 GPRs rS through rS nr 1 are stored into n consecutive bytes starting at the memory address referenced by EA The sequence of re...

Страница 425: ...C 0 data storage and data TLB miss exceptions do not occur However a data machine check exception can occur when XER TBC 0 if the following conditions are true The instruction access passes all protec...

Страница 426: ...tents of register rS are stored into the word referenced by EA Pseudocode EA rA 0 EXTS d MS EA 4 rS Registers Altered None Exceptions Data storage if the access is prevented by zone protection when da...

Страница 427: ...are stored into the byte referenced by EA 1 rS 8 15 are stored into the byte referenced by EA 2 rS 0 7 are stored into the byte referenced by EA 3 Pseudocode EA rA 0 rB MS EA 4 rS 24 31 rS 16 23 rS 8...

Страница 428: ...lease 1 800 255 7778 Virtex II Pro Platform FPGA Documentation Chapter 11 Instruction Set R Compatibility This instruction is defined by the PowerPC user instruction set architecture UISA It is implem...

Страница 429: ...vation bit is cleared to 0 when the instruction is executed no store operation is performed Execution of this instruction always clears the reservation bit CR CR0 is updated as follows CR CR0 LT GT ar...

Страница 430: ...enabled No access allowed zone protection applies only to accesses in user mode Read only zone protection applies to user and privileged modes Data TLB miss if data relocation is enabled and a valid...

Страница 431: ...TS d MS EA 4 rS rA EA Registers Altered rA Exceptions Data storage if the access is prevented by zone protection when data relocation is enabled No access allowed zone protection applies only to acces...

Страница 432: ...ons Data storage if the access is prevented by zone protection when data relocation is enabled No access allowed zone protection applies only to accesses in user mode Read only zone protection applies...

Страница 433: ...Altered None Exceptions Data storage if the access is prevented by zone protection when data relocation is enabled No access allowed zone protection applies only to accesses in user mode Read only zon...

Страница 434: ...rB to the one s complement of register rA and adding 1 to the result Simplified mnemonics defined for this instruction are described in Subtract Instructions page 831 Pseudocode rD rA rB 1 Registers A...

Страница 435: ...g 1 to the result XER CA is updated to reflect the unsigned magnitude of the result Simplified mnemonics defined for this instruction are described in Subtract Instructions page 831 Pseudocode rD rA r...

Страница 436: ...e subtract from extended instructions can be used to perform subtraction on integers larger than 32 bits as described on page 392 Pseudocode rD rA rB XER CA if rD 232 1 then XER CA 1 else XER CA 0 Reg...

Страница 437: ...operation is equivalent to adding the contents of the SIMM field sign extended to 32 bits to the one s complement of register rA and adding 1 to the result XER CA is updated to reflect the unsigned ma...

Страница 438: ...than 32 bits as described on page 392 Pseudocode rD rA 0xFFFF_FFFF XER CA if rD 232 1 then XER CA 1 else XER CA 0 Registers Altered rD XER CA CR CR0 LT GT EQ SO if Rc 1 XER SO OV if OE 1 Exceptions Ex...

Страница 439: ...escribed on page 392 Pseudocode rD rA XER CA if rD 232 1 then XER CA 1 else XER CA 0 Registers Altered rD XER CA CR CR0 LT GT EQ SO if Rc 1 XER SO OV if OE 1 Exceptions Execution of any of the followi...

Страница 440: ...antee ordering of both instruction completion and storage access The eieio instruction orders memory access not instruction completion Non memory instructions following eieio can complete before the m...

Страница 441: ...follow the tlbia instruction to guarantee that the effect of invalidating the TLB is visible to subsequent instructions Registers Altered None Exceptions Program Attempted execution of this instructio...

Страница 442: ...s read regardless of whether address translation is enabled Simplified mnemonics defined for this instruction are described in TLB Management Instructions page 832 Pseudocode tlb_entry rA26 31 if WS4...

Страница 443: ...com 751 Virtex II Pro Platform FPGA Documentation 1 800 255 7778 Alphabetical Instruction Listing R architecture Because it is optional and not defined by the PowerPC architecture it is not implement...

Страница 444: ...PID24 31 If a valid entry is found the corresponding TLB index is loaded into rD The TLB is searched regardless of whether address translation is enabled If Rc 1 CR CR0 is updated to reflect the sear...

Страница 445: ...n Listing R Compatibility This instruction is defined as optional by the operating environment architecture level OEA of the PowerPC embedded environment architecture and the PowerPC Book E architectu...

Страница 446: ...Registers Altered None Exceptions Program Attempted execution of this instruction from user mode Execution of any of the following invalid instruction forms results in a boundedly undefined result ra...

Страница 447: ...ronizing instruction should follow the tlbwe instruction to guarantee that the effect of writing a TLB entry is visible to subsequent instructions Simplified mnemonics defined for this instruction are...

Страница 448: ...002 Release 1 800 255 7778 Virtex II Pro Platform FPGA Documentation Chapter 11 Instruction Set R architecture Because it is optional and not defined by the PowerPC architecture it is not implemented...

Страница 449: ...eptions are disabled MSR DE 0 an imprecise debug event is reported by setting DBSR IDE to 1 If the trap instruction debug event is enabled as an internal debug event DBCR TDE 1 DBCR IDM 1 and DBCR EDM...

Страница 450: ...Execution of any of the following invalid instruction forms results in a boundedly undefined result rather than a program exception Reserved bits containing a non zero value Compatibility This instruc...

Страница 451: ...1 and debug exceptions are disabled MSR DE 0 an imprecise debug event is reported by setting DBSR IDE to 1 If the trap instruction debug event is enabled as an internal debug event DBCR TDE 1 DBCR IDM...

Страница 452: ...Execution of any of the following invalid instruction forms results in a boundedly undefined result rather than a program exception Reserved bits containing a non zero value Compatibility This instruc...

Страница 453: ...instruction from user mode Execution of any of the following invalid instruction forms results in a boundedly undefined result rather than a program exception Reserved bits containing a non zero value...

Страница 454: ...tion from user mode Execution of any of the following invalid instruction forms results in a boundedly undefined result rather than a program exception Reserved bits containing a non zero value Compat...

Страница 455: ...th the contents of register rB and the result is loaded into register rA Pseudocode rA rS rB Registers Altered rA CR CR0 LT GT EQ SO if Rc 1 Exceptions None Compatibility This instruction is defined b...

Страница 456: ...ating 16 0 bits on the left The contents of register rS are XORed with the extended UIMM field and the result is loaded into register rA Pseudocode rA rS 160 UIMM Registers Altered rA Exceptions None...

Страница 457: ...ncatenating 16 0 bits on the right The contents of register rS are XORed with the extended UIMM field and the result is loaded into register rA Pseudocode rA rS UIMM 160 Registers Altered rA Exception...

Страница 458: ...766 www xilinx com March 2002 Release 1 800 255 7778 Virtex II Pro Platform FPGA Documentation Chapter 11 Instruction Set R...

Страница 459: ...Registers GPRs page 360 CR Condition Register Condition Register CR page 361 MSR Machine State Register Machine State Register page 431 CCR0 Core Configuration Register 0 Core Configuration Register...

Страница 460: ...GR page 455 SLER Storage Little Endian Register Storage Little Endian Register SLER page 455 SPRG0 SPR General Purpose Register 0 SPR General Purpose Registers page 432 SPRG1 SPR General Purpose Regis...

Страница 461: ...l Purpose Register 13 13 0x0D 0b01101 No Read Write Undefined r14 General Purpose Register 14 14 0x0E 0b01110 No Read Write Undefined r15 General Purpose Register 15 15 0x0F 0b01111 No Read Write Unde...

Страница 462: ...ble Yes Read Write 0x0000_0000 Table A 4 Special Purpose Registers Sorted by Name Name Descriptive Name SPRN SPRF Privileged Access Reset Value Dec Hex Hex Binary CCR0 Core Configuration Register 0 94...

Страница 463: ...01000 No Read Only Undefined SPRG4 SPR General Purpose Register 4 276 0x114 0x288 0b10100_01000 Yes Read Write Undefined SPRG5 SPR General Purpose Register 5 261 0x105 0x0A8 0b00101_01000 No Read Only...

Страница 464: ...277 0x115 0x2A8 0b10101_01000 Yes Read Write Undefined SPRG6 SPR General Purpose Register 6 278 0x116 0x2C8 0b10110_01000 Yes Read Write Undefined SPRG7 SPR General Purpose Register 7 279 0x117 0x2E8...

Страница 465: ...0x3FB 0x37F 0b11011_11111 Yes Read Write 0x0000_0000 Table A 5 Special Purpose Registers Sorted by SPRN Continued Name Descriptive Name SPRN SPRF Privileged Access Reset Value Dec Hex Hex Binary Table...

Страница 466: ...101 Yes Read Write Undefined DAC2 Data Address Compare 2 1015 0x3F7 0x2FF 0b10111_11111 Yes Read Write Undefined TSR Timer Status Register 984 0x3D8 0x31E 0b11000_11110 Yes Read Clear Undefined1 SGR S...

Страница 467: ...BRN that appears in the instruction encoding Device Control Registers Device control registers DCRs are not architecturally part of the PPC405 DCRs are used to control configure and record status for...

Страница 468: ...776 www xilinx com March 2002 Release 1 800 255 7778 Virtex II Pro Platform FPGA Documentation Appendix A Register Summary R...

Страница 469: ...by Mnemonic Table B 1 lists the PPC405 instruction set in alphabetical order by mnemonic Table B 1 Instructions Sorted by Mnemonic 0 6 9 11 12 14 16 17 20 21 22 26 30 31 add 31 rD rA rB OE 266 Rc add...

Страница 470: ...31 00000 rA rB 278 0 dcbtst 31 00000 rA rB 246 0 dcbz 31 00000 rA rB 1014 0 dccci 31 00000 rA rB 454 0 dcread 31 rD rA rB 486 0 divw 31 rD rA rB OE 491 Rc divwu 31 rD rA rB OE 459 Rc eieio 31 00000 00...

Страница 471: ...rD rA rB OE 140 Rc machhw 4 rD rA rB OE 44 Rc machhws 4 rD rA rB OE 108 Rc machhwsu 4 rD rA rB OE 76 Rc machhwu 4 rD rA rB OE 12 Rc maclhw 4 rD rA rB OE 428 Rc maclhws 4 rD rA rB OE 492 Rc maclhwsu 4...

Страница 472: ...rA rB OE 110 Rc nmaclhw 4 rD rA rB OE 430 Rc nmaclhws 4 rD rA rB OE 494 Rc nor 31 rS rA rB 124 Rc or 31 rS rA rB 444 Rc orc 31 rS rA rB 412 Rc ori 24 rS rA UIMM oris 25 rS rA UIMM rfci 19 00000 00000...

Страница 473: ...stwux 31 rS rA rB 183 0 stwx 31 rS rA rB 151 0 subf 31 rD rA rB OE 40 Rc subfc 31 rD rA rB OE 8 Rc subfe 31 rD rA rB OE 136 Rc subfic 8 rD rA SIMM subfme 31 rD rA 00000 OE 232 Rc subfze 31 rD rA 0000...

Страница 474: ...rB OE 140 Rc mulchw 4 rD rA rB 168 Rc macchw 4 rD rA rB OE 172 Rc nmacchw 4 rD rA rB OE 174 Rc macchwsu 4 rD rA rB OE 204 Rc macchws 4 rD rA rB OE 236 Rc nmacchws 4 rD rA rB OE 238 Rc mullhwu 4 rD rA...

Страница 475: ...O BI 00000 528 LK rlwimi 20 rS rA SH MB ME Rc rlwinm 21 rS rA SH MB ME Rc rlwnm 23 rS rA rB MB ME Rc ori 24 rS rA UIMM oris 25 rS rA UIMM xori 26 rS rA UIMM xoris 27 rS rA UIMM andi 28 rS rA UIMM andi...

Страница 476: ...000 00000 E 0000 163 0 stwux 31 rS rA rB 183 0 subfze 31 rD rA 00000 OE 200 Rc addze 31 rD rA 00000 OE 202 Rc stbx 31 rS rA rB 215 0 subfme 31 rD rA 00000 OE 232 Rc addme 31 rD rA 00000 OE 234 Rc mull...

Страница 477: ...rA rB 536 Rc tlbsync 31 00000 00000 00000 566 0 lswi 31 rD rA NB 597 0 sync 31 00000 00000 00000 598 0 stswx 31 rS rA rB 661 0 stwbrx 31 rS rA rB 662 0 stswi 31 rS rA NB 725 0 dcba 31 00000 rA rB 758...

Страница 478: ...rA d lhau 43 rD rA d sth 44 rS rA d sthu 45 rS rA d lmw 46 rD rA d stmw 47 rS rA d Table B 2 Instructions Sorted by Opcode Continued 0 6 9 11 12 14 16 17 20 21 22 26 30 31 Table B 3 Integer Add and Su...

Страница 479: ...rB 392 Rc mulli 7 rD rA SIMM mullw 31 rD rA rB OE 235 Rc Table B 5 Integer Multiply Accumulate Instructions 0 6 11 16 21 22 31 macchw 4 rD rA rB OE 172 Rc macchws 4 rD rA rB OE 236 Rc macchwsu 4 rD rA...

Страница 480: ...1 rS rA rB 60 Rc andi 28 rS rA UIMM andis 29 rS rA UIMM cntlzw 31 rS rA 00000 26 Rc eqv 31 rS rA rB 284 Rc extsb 31 rS rA 00000 954 Rc extsh 31 rS rA 00000 922 Rc nand 31 rS rA rB 476 Rc nor 31 rS rA...

Страница 481: ...rD rA rB 119 0 lbzx 31 rD rA rB 87 0 lha 42 rD rA d lhau 43 rD rA d lhaux 31 rD rA rB 375 0 lhax 31 rD rA rB 343 0 lhz 40 rD rA d lhzu 41 rD rA d lhzux 31 rD rA rB 311 0 lhzx 31 rD rA rB 279 0 lwz 32...

Страница 482: ...ad and Store String Instructions 0 6 11 16 21 31 lswi 31 rD rA NB 597 0 lswx 31 rD rA rB 533 0 stswi 31 rS rA NB 725 0 stswx 31 rS rA rB 661 0 Table B 15 Branch Instructions 0 6 11 16 21 30 31 b 18 LI...

Страница 483: ...ynchronization Instructions 0 6 11 16 21 31 eieio 31 00000 00000 00000 854 0 isync 19 00000 00000 00000 150 0 lwarx 31 rD rA rB 20 0 stwcx 31 rS rA rB 150 1 sync 31 00000 00000 00000 598 0 Table B 20...

Страница 484: ...0 rA rB 470 0 dcbst 31 00000 rA rB 54 0 dcbt 31 00000 rA rB 278 0 dcbtst 31 00000 rA rB 246 0 dcbz 31 00000 rA rB 1014 0 dccci 31 00000 rA rB 454 0 dcread 31 rD rA rB 486 0 icbi 31 00000 rA rB 982 0 i...

Страница 485: ...A UIMM andi 28 rS rA UIMM andis 29 rS rA UIMM lwz 32 rD rA d lwzu 33 rD rA d lbz 34 rD rA d lbzu 35 rD rA d stw 36 rS rA d stwu 37 rS rA d stb 38 rS rA d stbu 39 rS rA d lhz 40 rD rA d lhzu 41 rD rA d...

Страница 486: ...4 0 mfcr 31 rD 00000 00000 19 0 lwarx 31 rD rA rB 20 0 lwzx 31 rD rA rB 23 0 slw 31 rS rA rB 24 Rc cntlzw 31 rS rA 00000 26 Rc and 31 rS rA rB 28 Rc cmpl 31 crfD 00 rA rB 32 0 dcbst 31 00000 rA rB 54...

Страница 487: ...31 rS rA rB 476 Rc dcread 31 rD rA rB 486 0 mcrxr 31 crfD 00 00000 00000 512 0 lswx 31 rD rA rB 533 0 lwbrx 31 rD rA rB 534 0 srw 31 rS rA rB 536 Rc tlbsync 31 00000 00000 00000 566 0 lswi 31 rD rA N...

Страница 488: ...ble B 30 XL Form 0 6 9 11 14 16 21 31 mcrf 19 crfD 00 crfS 00 00000 0 0 bclr 19 BO BI 00000 16 LK crnor 19 crbD crbA crbB 33 0 rfi 19 00000 00000 00000 50 0 rfci 19 00000 00000 00000 51 0 crandc 19 cr...

Страница 489: ...36 Rc nmacchws 4 rD rA rB OE 238 Rc maclhwu 4 rD rA rB OE 396 Rc maclhw 4 rD rA rB OE 428 Rc nmaclhw 4 rD rA rB OE 430 Rc maclhwsu 4 rD rA rB OE 460 Rc maclhws 4 rD rA rB OE 492 Rc nmaclhws 4 rD rA rB...

Страница 490: ...x x x UISA D addic x x x UISA D addic x x x UISA D addis x x x UISA D addme x x x UISA XO addze x x x UISA XO and x x x UISA X andc x x x UISA X andi x x x UISA D andis x x x UISA D b x x x UISA I bc...

Страница 491: ...x x x UISA D lbzu x x x UISA D lbzux x x x UISA X lbzx x x x UISA X lha x x x UISA D lhau x x x UISA D lhaux x x x UISA X lhax x x x UISA X lhbrx x x x UISA X lhz x x x UISA D lhzu x x x UISA D lhzux...

Страница 492: ...ISA X mfcr x x x UISA X mfdcr x x OEA x XFX mfmsr x x x OEA x X mfspr x x x UISA XFX OEA x1 mftb x x VEA XFX mtcrf x x x UISA XFX mtdcr x x OEA x XFX mtmsr x x x OEA x X mtspr x x x UISA XFX OEA x1 mu...

Страница 493: ...imi x x x UISA M rlwinm x x x UISA M rlwnm x x x UISA M sc x x x UISA SC slw x x x UISA X sraw x x x UISA X srawi x x x UISA X srw x x x UISA X stb x x x UISA D stbu x x x UISA D stbux x x x UISA X st...

Страница 494: ...lified mnemonic its equivalent mnemonic is listed in the column headed Equivalent Mnemonic Otherwise the column is shaded gray stwbrx x x x UISA X stwcx x x x UISA X stwu x x x UISA D stwux x x x UISA...

Страница 495: ...page 576 addic Add Immediate Carrying and Record page 577 addis Add Immediate Shifted page 578 addme Add to Minus One Extended page 579 addme Add to Minus One Extended and Record addmeo Add to Minus O...

Страница 496: ...CTR Not Zero to Link Register bclr page 823 bdnzlrl Branch if Decremented CTR Not Zero to Link Register and Link bclrl page 824 bdnzt Branch if Decremented CTR Not Zero and Condition True bc page 822...

Страница 497: ...827 bf Branch if Condition False bc page 822 bfa Branch if Condition False Absolute bca page 822 bfctr Branch if Condition False to Count Register bcctr page 823 bfctrl Branch if Condition False to C...

Страница 498: ...ess Than bc page 825 blta Branch if Less Than Absolute bca page 825 bltctr Branch if Less Than to Count Register bcctr page 826 bltctrl Branch if Less Than to Count Register and Link bcctrl page 827 b...

Страница 499: ...verflow and Link bcl page 826 bnsla Branch if Not Summary Overflow Absolute and Link bcla page 826 bnslr Branch if Not Summary Overflow to Link Register bclr page 826 bnslrl Branch if Not Summary Over...

Страница 500: ...ister AND page 597 crandc Condition Register AND with Complement page 598 crclr Condition Register Clear crxor page 828 creqv Condition Register Equivalent page 599 crmove Condition Register Move cror...

Страница 501: ...nd Record rlwinm extsb Extend Sign Byte page 628 extsb Extend Sign Byte and Record extsh Extend Sign Halfword page 629 extsh Extend Sign Halfword and Record icbi Instruction Cache Block Invalidate pag...

Страница 502: ...Modulo Signed with Overflow Enabled macchwo Multiply Accumulate Cross Halfword to Word Modulo Signed with Overflow Enabled and Record macchws Multiply Accumulate Cross Halfword to Word Saturate Signed...

Страница 503: ...Enabled machhwuo Multiply Accumulate High Halfword to Word Modulo Unsigned with Overflow Enabled and Record maclhw Multiply Accumulate Low Halfword to Word Modulo Signed page 671 maclhw Multiply Accu...

Страница 504: ...ss Register mfdvc1 Move From Data Value Compare 1 mfdvc2 Move From Data Value Compare 2 mfesr Move From Exception Syndrome Register mfevpr Move From Exception Vector Prefix Register mfiac1 Move From I...

Страница 505: ...age 681 mftbl Move From Time Base Lower mfspr page 830 mftbu Move From Time Base Upper mftcr Move From Timer Control Register mftsr Move From Timer Status Register mfusprg0 Move From User SPR General...

Страница 506: ...rogrammable Interval Timer mtsgr Move to Storage Guarded Register mtsler Move to Storage Little Endian Register mtspr Move to Special Purpose Register page 686 mtsprg0 Move to SPR General Purpose Regi...

Страница 507: ...mulhhwu Multiply High Halfword to Word Unsigned and Record mulhw Multiply High Word page 691 mulhw Multiply High Word and Record mulhwu Multiply High Word Unsigned page 692 mulhwu Multiply High Word...

Страница 508: ...ow Enabled and Record nmachhws Negative Multiply Accumulate High Halfword to Word Saturate Signed page 702 nmachhws Negative Multiply Accumulate High Halfword to Word Saturate Signed and Record nmachh...

Страница 509: ...29 rotlw Rotate Left and Record rlwinm rotlwi Rotate Left Immediate rlwinm rotlwi Rotate Left Immediate and Record rlwinm rotrwi Rotate Right Immediate rlwinm rotrwi Rotate Right Immediate and Record...

Страница 510: ...act Carrying and Record subfc subco Subtract Carrying with Overflow Enabled subfco page 832 subco Subtract Carrying with Overflow Enabled and Record subfco subf Subtract from page 742 subf Subtract fr...

Страница 511: ...Enabled and Record subfo sync Synchronize page 748 tlbia TLB Invalidate All page 749 tlbre TLB Read Entry page 750 tlbrehi Read TLBHI Portion of TLB Entry tlbre page 832 tlbrelo Read TLBLO Portion of...

Страница 512: ...te twi twlng Trap if Logically Not Greater Than tw twlngi Trap if Logically Not Greater Than Immediate twi twlnl Trap if Logically Not Less Than tw twlnli Trap if Logically Not Less Than Immediate twi...

Страница 513: ...met The condition tested can include a specific bit b in the CR whether or not the contents of the CTR are zero or both The simplified mnemonics in Table C 2 through Table C 6 are formed using the fo...

Страница 514: ...nt CTR Branch if CTR 0 bdz bdza bdzlr bdzl bdzla bdzlrl Decrement CTR Branch if CTR 0 and Condition True CRb 1 bdzt bdzta bdztlr bdztl bdztla bdztlrl Decrement CTR Branch if CTR 0 and Condition False...

Страница 515: ...bdnztlr b bclr 8 b Decrement CTR Branch if CTR 0 and Condition False CRb 0 bdnzflr b bclr 0 b Decrement CTR Branch if CTR 0 bdzlr bclr 18 0 Decrement CTR Branch if CTR 0 and Condition True CRb 1 bdztl...

Страница 516: ...lse CRb 0 bdzfl b target bcl 2 b target bdzfla b target bcla 2 b target Table C 5 Branch True False to Relative Absolute LK 1 Continued Operation LR Updated Branch Relative Branch Absolute Simplified...

Страница 517: ...ns Operation LR not Updated LR Updated Relative Absolute to LR to CTR Relative Absolute to LR to CTR Table C 9 Table C 10 Table C 11 Table C 12 Branch if Less Than blt blta bltlr bltctr bltl bltla blt...

Страница 518: ...son to Relative Absolute LK 0 Continued Operation LR not Updated Branch Relative Branch Absolute Simplified Mnemonic Equivalent Mnemonic Simplified Mnemonic Equivalent Mnemonic Table C 10 Branch Compa...

Страница 519: ...target Branch if Not Equal bnel n target bcl 4 4 n 2 target bnela n target bcla 4 4 n 2 target Branch if Not Greater Than bngl n target bcl 4 4 n 1 target bngla n target bcla 4 4 n 1 target Branch if...

Страница 520: ...o 1 A suffix clears the y bit to 0 Compare Instructions The PowerPC compare instructions include an L opcode field that specifies whether the comparison is performed on a word or doubleword operand In...

Страница 521: ...rray index by the width of an element Condition Register Clear crclr bx crxor bx bx bx Condition Register Move crmove bx by cror bx by by Condition Register Not crnot bx by crnor bx by by Table C 14 S...

Страница 522: ...mtdac1 rS mtspr 1014 rS mfdac1 rD mfspr rD 1014 Data Address Compare 2 mtdac2 rS mtspr 1015 rS mfdac2 rD mfspr rD 1015 Debug Control Register 0 mtdbcr0 rS mtspr 1010 rS mfdbcr0 rD mfspr rD 1010 Debug...

Страница 523: ...General Purpose Register 5 mtsprg5 rS mtspr 277 rS SPR General Purpose Register 6 mfsprg6 rD mfspr rD 262 SPR General Purpose Register 6 mtsprg6 rS mtspr 278 rS SPR General Purpose Register 7 mfsprg7...

Страница 524: ...ison and the column headed U indicates an unsigned greater than comparison Table C 17 Simplified Mnemonics for Subtract Instructions Operation Simplified Mnemonic Equivalent Mnemonic Subtract rA rB su...

Страница 525: ...Mnemonic Trap if less than twlt rA rB tw 16 rA rB twlti rA SIMM twi 16 rA SIMM Trap if less than or equal twle rA rB tw 20 rA rB twlei rA SIMM twi 20 rA SIMM Trap if equal tweq rA rB tw 4 rA rB tweqi...

Страница 526: ...ress syntax d rA Move Register The simplified mnemonics in Table C 24 provide a shorthand for moving the contents of a GPR to another GPR Complement Register The simplified mnemonics in Table C 25 pro...

Страница 527: ...s a shorthand for copying the contents of a GPR into the CR Table C 25 Simplified Mnemonics for Complement Register Operation Simplified Mnemonic Equivalent Mnemonic Complement Not Register not rA rS...

Страница 528: ...836 www xilinx com March 2002 Release 1 800 255 7778 Virtex II Pro Platform FPGA Documentation Appendix C Simplified Mnemonics R...

Страница 529: ...g on an lwarx instruction that fails to return a desired value Performance can also be improved by using an ordinary load instruction to do the initial value check as follows loop lwz r5 0 r3 load the...

Страница 530: ...d into r5 loop lwarx r5 0 r3 load and reserve stwcx r4 0 r3 store new value if still reserved bne loop loop if reservation lost Fetch and Add The fetch and add primitive atomically increments a memory...

Страница 531: ...r6 were not equal by exiting the primitive with CR0 EQ 0 If this indication is not required by the application the second bne can be omitted The mr is used only when the application requires that the...

Страница 532: ...ress is in r4 and the next element pointers are at offset zero in the respective element data structure It is also assumed that the next element pointer of each list element is in a reservation granul...

Страница 533: ...ft amount Shift left immediate n 3 shift amount 32 rlwinm r2 r2 sh 0 31 sh rlwimi r2 r3 sh 32 sh 31 rlwinm r3 r3 sh 0 31 sh rlwimi r3 r4 sh 32 sh 31 rlwinm r4 r4 sh 0 31 sh Shift left n 2 shift amount...

Страница 534: ...tions and the instruction timings and latencies Usually the implementation involves a combination of conditional branches and unconditional branches Conditional branches require the evaluation of cond...

Страница 535: ...structions that can benefit from this action are Integer arithmetic compare and logical instructions that have the Rc opcode field set The addic andi and andis instructions CR logical instructions The...

Страница 536: ...s includes PLB contention between the instruction and data caches and the time associated with performing cache line fills and flushes Unless stated otherwise the number of cycles described applies to...

Страница 537: ...ion occurs If the instruction is predicted early at or before prefetch it executes in one cycle If the instruction is predicted during decode it executes in two cycles If the prediction is incorrect t...

Страница 538: ...uction that uses the loaded data a load use dependency exists When the loaded data is available it is forwarded to the operand register of the dependent instruction This prevents a processor stall fro...

Страница 539: ...bytes are not aligned on a word boundary Access to intermediate bytes consume one cycle for each word accessed Access to trailing bytes consume one cycle Unused bytes are discarded if the trailing by...

Страница 540: ...isses Cacheable instruction fetch misses and non cacheable instruction fetches incur penalty cycles for accessing memory over the PLB These penalty cycles depend on the speed of the PLB and when the a...

Страница 541: ...tions is usually limited to the operating system kernel and other privileged mode software Applications usually require no modification Software porting can be simplified through the use of structured...

Страница 542: ...ntrol Register TSR Timer Status Register DACn Data address compare registers Debugging DBCRn Debug control registers DBSR Debug status register DVCn Data value compare registers IACn Instruction addre...

Страница 543: ...e control register THRMn Thermal assist unit registers Thermal management Table E 2 6xx 7xx Registers Not Supported by 40x Processors Name Description Purpose Table E 3 Comparison of MSR Bit Definitio...

Страница 544: ...late a 52 bit virtual address within a page to a 32 bit physical address The page translation tables are created by software and stored in system memory The processor uses a translation look aside buf...

Страница 545: ...entries Zone protection can be used to override the access protection specified in a TLB entry Fields within the zone protection register ZPR define the protection level of a page or set of pages Mem...

Страница 546: ...erPC 401 processors cachelines can be individually locked PowerPC 403 processors not supported PowerPC 405 processors not supported PowerPC 6xx 7xx processors the instruction and data caches can be lo...

Страница 547: ...ports two locations for the interrupt handler table 0x000n_nnnn or 0xFFFn_nnnn selected by using the MSR IP bit New exceptions and interrupts are defined Some exceptions and interrupts supported by th...

Страница 548: ...798 for a list of implementation dependent PPC405 instructions This table also shows which PPC405 instructions are not supported by the PowerPC architecture Endian Support The default memory access o...

Страница 549: ...ment cache management memory synchronization exceptions timer resources and others Many of the differences are reflected the deletion modification and introduction of special purpose registers Porting...

Страница 550: ...cache write through register ICCR Instruction cache cacheability register SGR Storage Guarded Register SLER Storage Little Endian Register SU0R Storage User Defined 0 Register ZPR Zone Protection Regi...

Страница 551: ...s DECAR Decrementer Auto Reload DNVn1 Data cache normal victim register Cache control DTVn1 Data cache transient victim register DVLIM1 Data cache victim limit INVn1 Instruction cache normal victim re...

Страница 552: ...ode is not supported by PowerPC Book E implementations Address translation is always enabled and one or more TLB entries are initialized by the processor during reset so that instructions can be fetch...

Страница 553: ...rted by the PowerPC Book E architecture Memory Attributes The PowerPC 40x family and PowerPC Book E processors support the following memory attributes Write through W Caching inhibited I Memory cohere...

Страница 554: ...eieio instruction orders all memory accesses This guarantees that existing software that uses eieio works properly in PowerPC Book E implementations The memory synchronize msync instruction replaces...

Страница 555: ...registers with the appropriate vector offsets as shown in Table F 7 Some bits in the exception syndrome register ESR are redefined to support different exception conditions These changes are shown in...

Страница 556: ...processors can support implementation specific instructions For example the multiply accumulate MAC instructions are considered implementation dependent and are not guaranteed to be supported by othe...

Страница 557: ...t See reset clear register instructions 829 compare instructions 398 828 complement register instruction 834 condition register 361 CR mask CRM 423 CR0 361 CR1 362 CR logical instructions 376 828 effe...

Страница 558: ...data TLB miss exception 519 data storage exception 507 instruction storage exception 508 machine check exception 504 program exception 512 exception vector prefix register 500 execution model See als...

Страница 559: ...chine check enable 504 reset state 562 wait state enable 436 masking interrupts 496 memory coherency 448 memory management 345 memory synchronization See synchronization storage memory control instruc...

Страница 560: ...otation 571 SPR See special purpose register SPR general purpose register privileged mode 432 user mode 365 SPRGn See SPR general purpose register SRRn See save restore registers static branch predict...

Страница 561: ...torage attribute user defined UISA See PowerPC unconditional UDE See debug events user mode 344 user registers 359 user SPR general purpose register 364 USPRG0 See user SPR general purpose register UT...

Страница 562: ...870 www xilinx com March 2002 Release 1 800 255 7778 Virtex II Pro Platform FPGA Documentation R...

Отзывы:

Нет отзывов

Бренды по названию

Популярные бренды

Загрузить еще бренды