X-Git-Url: http://demsky.eecs.uci.edu/git/?a=blobdiff_plain;f=lib%2FTarget%2FX86%2FX86InstrArithmetic.td;h=225e9720da0c26bcb29515506958fab49b52099b;hb=a0de26ce34877fac86ef6d939d56bbebb3b4e14c;hp=7029b7139a57ba70de69a3da3ebf5cc1361e71d1;hpb=922d314e8f9f0d8e447c055485a2969ee9cf2dd2;p=oota-llvm.git diff --git a/lib/Target/X86/X86InstrArithmetic.td b/lib/Target/X86/X86InstrArithmetic.td index 7029b7139a5..225e9720da0 100644 --- a/lib/Target/X86/X86InstrArithmetic.td +++ b/lib/Target/X86/X86InstrArithmetic.td @@ -1,10 +1,10 @@ -//===- X86InstrArithmetic.td - Integer Arithmetic Instrs ---*- tablegen -*-===// -// +//===-- X86InstrArithmetic.td - Integer Arithmetic Instrs --*- tablegen -*-===// +// // The LLVM Compiler Infrastructure // // This file is distributed under the University of Illinois Open Source // License. See LICENSE.TXT for details. -// +// //===----------------------------------------------------------------------===// // // This file describes the integer arithmetic instructions in the X86 @@ -14,7 +14,7 @@ //===----------------------------------------------------------------------===// // LEA - Load Effective Address - +let SchedRW = [WriteLEA] in { let neverHasSideEffects = 1 in def LEA16r : I<0x8D, MRMSrcMem, (outs GR16:$dst), (ins i32mem:$src), @@ -29,48 +29,59 @@ def LEA32r : I<0x8D, MRMSrcMem, def LEA64_32r : I<0x8D, MRMSrcMem, (outs GR32:$dst), (ins lea64_32mem:$src), "lea{l}\t{$src|$dst}, {$dst|$src}", - [(set GR32:$dst, lea32addr:$src)], IIC_LEA>, + [(set GR32:$dst, lea64_32addr:$src)], IIC_LEA>, Requires<[In64BitMode]>; let isReMaterializable = 1 in -def LEA64r : RI<0x8D, MRMSrcMem, (outs GR64:$dst), (ins i64mem:$src), +def LEA64r : RI<0x8D, MRMSrcMem, (outs GR64:$dst), (ins lea64mem:$src), "lea{q}\t{$src|$dst}, {$dst|$src}", [(set GR64:$dst, lea64addr:$src)], IIC_LEA>; - - +} // SchedRW //===----------------------------------------------------------------------===// // Fixed-Register Multiplication and Division Instructions. // +// SchedModel info for instruction that loads one value and gets the second +// (and possibly third) value from a register. +// This is used for instructions that put the memory operands before other +// uses. +class SchedLoadReg : Sched<[SW, + // Memory operand. + ReadDefault, ReadDefault, ReadDefault, ReadDefault, ReadDefault, + // Register reads (implicit or explicit). + ReadAfterLd, ReadAfterLd]>; + // Extra precision multiplication // AL is really implied by AX, but the registers in Defs must match the // SDNode results (i8, i32). +// AL,AH = AL*GR8 let Defs = [AL,EFLAGS,AX], Uses = [AL] in def MUL8r : I<0xF6, MRM4r, (outs), (ins GR8:$src), "mul{b}\t$src", // FIXME: Used for 8-bit mul, ignore result upper 8 bits. // This probably ought to be moved to a def : Pat<> if the // syntax can be accepted. [(set AL, (mul AL, GR8:$src)), - (implicit EFLAGS)]>; // AL,AH = AL*GR8 - + (implicit EFLAGS)], IIC_MUL8>, Sched<[WriteIMul]>; +// AX,DX = AX*GR16 let Defs = [AX,DX,EFLAGS], Uses = [AX], neverHasSideEffects = 1 in def MUL16r : I<0xF7, MRM4r, (outs), (ins GR16:$src), - "mul{w}\t$src", - [], IIC_MUL16_REG>, OpSize; // AX,DX = AX*GR16 - + "mul{w}\t$src", + [], IIC_MUL16_REG>, OpSize, Sched<[WriteIMul]>; +// EAX,EDX = EAX*GR32 let Defs = [EAX,EDX,EFLAGS], Uses = [EAX], neverHasSideEffects = 1 in def MUL32r : I<0xF7, MRM4r, (outs), (ins GR32:$src), - "mul{l}\t$src", // EAX,EDX = EAX*GR32 + "mul{l}\t$src", [/*(set EAX, EDX, EFLAGS, (X86umul_flag EAX, GR32:$src))*/], - IIC_MUL32_REG>; + IIC_MUL32_REG>, Sched<[WriteIMul]>; +// RAX,RDX = RAX*GR64 let Defs = [RAX,RDX,EFLAGS], Uses = [RAX], neverHasSideEffects = 1 in def MUL64r : RI<0xF7, MRM4r, (outs), (ins GR64:$src), - "mul{q}\t$src", // RAX,RDX = RAX*GR64 + "mul{q}\t$src", [/*(set RAX, RDX, EFLAGS, (X86umul_flag RAX, GR64:$src))*/], - IIC_MUL64>; - + IIC_MUL64>, Sched<[WriteIMul]>; +// AL,AH = AL*[mem8] let Defs = [AL,EFLAGS,AX], Uses = [AL] in def MUL8m : I<0xF6, MRM4m, (outs), (ins i8mem :$src), "mul{b}\t$src", @@ -78,50 +89,60 @@ def MUL8m : I<0xF6, MRM4m, (outs), (ins i8mem :$src), // This probably ought to be moved to a def : Pat<> if the // syntax can be accepted. [(set AL, (mul AL, (loadi8 addr:$src))), - (implicit EFLAGS)], IIC_MUL8>; // AL,AH = AL*[mem8] - + (implicit EFLAGS)], IIC_MUL8>, SchedLoadReg; +// AX,DX = AX*[mem16] let mayLoad = 1, neverHasSideEffects = 1 in { let Defs = [AX,DX,EFLAGS], Uses = [AX] in def MUL16m : I<0xF7, MRM4m, (outs), (ins i16mem:$src), "mul{w}\t$src", - [], IIC_MUL16_MEM>, OpSize; // AX,DX = AX*[mem16] - + [], IIC_MUL16_MEM>, OpSize, SchedLoadReg; +// EAX,EDX = EAX*[mem32] let Defs = [EAX,EDX,EFLAGS], Uses = [EAX] in def MUL32m : I<0xF7, MRM4m, (outs), (ins i32mem:$src), "mul{l}\t$src", - [], IIC_MUL32_MEM>; // EAX,EDX = EAX*[mem32] + [], IIC_MUL32_MEM>, SchedLoadReg; +// RAX,RDX = RAX*[mem64] let Defs = [RAX,RDX,EFLAGS], Uses = [RAX] in def MUL64m : RI<0xF7, MRM4m, (outs), (ins i64mem:$src), - "mul{q}\t$src", [], IIC_MUL64>; // RAX,RDX = RAX*[mem64] + "mul{q}\t$src", [], IIC_MUL64>, SchedLoadReg; } let neverHasSideEffects = 1 in { +// AL,AH = AL*GR8 let Defs = [AL,EFLAGS,AX], Uses = [AL] in -def IMUL8r : I<0xF6, MRM5r, (outs), (ins GR8:$src), "imul{b}\t$src", []>; - // AL,AH = AL*GR8 +def IMUL8r : I<0xF6, MRM5r, (outs), (ins GR8:$src), "imul{b}\t$src", [], + IIC_IMUL8>, Sched<[WriteIMul]>; +// AX,DX = AX*GR16 let Defs = [AX,DX,EFLAGS], Uses = [AX] in -def IMUL16r : I<0xF7, MRM5r, (outs), (ins GR16:$src), "imul{w}\t$src", []>, - OpSize; // AX,DX = AX*GR16 +def IMUL16r : I<0xF7, MRM5r, (outs), (ins GR16:$src), "imul{w}\t$src", [], + IIC_IMUL16_RR>, OpSize, Sched<[WriteIMul]>; +// EAX,EDX = EAX*GR32 let Defs = [EAX,EDX,EFLAGS], Uses = [EAX] in -def IMUL32r : I<0xF7, MRM5r, (outs), (ins GR32:$src), "imul{l}\t$src", []>; - // EAX,EDX = EAX*GR32 +def IMUL32r : I<0xF7, MRM5r, (outs), (ins GR32:$src), "imul{l}\t$src", [], + IIC_IMUL32_RR>, Sched<[WriteIMul]>; +// RAX,RDX = RAX*GR64 let Defs = [RAX,RDX,EFLAGS], Uses = [RAX] in -def IMUL64r : RI<0xF7, MRM5r, (outs), (ins GR64:$src), "imul{q}\t$src", []>; - // RAX,RDX = RAX*GR64 +def IMUL64r : RI<0xF7, MRM5r, (outs), (ins GR64:$src), "imul{q}\t$src", [], + IIC_IMUL64_RR>, Sched<[WriteIMul]>; let mayLoad = 1 in { +// AL,AH = AL*[mem8] let Defs = [AL,EFLAGS,AX], Uses = [AL] in def IMUL8m : I<0xF6, MRM5m, (outs), (ins i8mem :$src), - "imul{b}\t$src", []>; // AL,AH = AL*[mem8] + "imul{b}\t$src", [], IIC_IMUL8>, SchedLoadReg; +// AX,DX = AX*[mem16] let Defs = [AX,DX,EFLAGS], Uses = [AX] in def IMUL16m : I<0xF7, MRM5m, (outs), (ins i16mem:$src), - "imul{w}\t$src", []>, OpSize; // AX,DX = AX*[mem16] + "imul{w}\t$src", [], IIC_IMUL16_MEM>, OpSize, + SchedLoadReg; +// EAX,EDX = EAX*[mem32] let Defs = [EAX,EDX,EFLAGS], Uses = [EAX] in def IMUL32m : I<0xF7, MRM5m, (outs), (ins i32mem:$src), - "imul{l}\t$src", []>; // EAX,EDX = EAX*[mem32] + "imul{l}\t$src", [], IIC_IMUL32_MEM>, SchedLoadReg; +// RAX,RDX = RAX*[mem64] let Defs = [RAX,RDX,EFLAGS], Uses = [RAX] in def IMUL64m : RI<0xF7, MRM5m, (outs), (ins i64mem:$src), - "imul{q}\t$src", []>; // RAX,RDX = RAX*[mem64] + "imul{q}\t$src", [], IIC_IMUL64>, SchedLoadReg; } } // neverHasSideEffects @@ -129,7 +150,8 @@ def IMUL64m : RI<0xF7, MRM5m, (outs), (ins i64mem:$src), let Defs = [EFLAGS] in { let Constraints = "$src1 = $dst" in { -let isCommutable = 1 in { // X = IMUL Y, Z --> X = IMUL Z, Y +let isCommutable = 1, SchedRW = [WriteIMul] in { +// X = IMUL Y, Z --> X = IMUL Z, Y // Register-Register Signed Integer Multiply def IMUL16rr : I<0xAF, MRMSrcReg, (outs GR16:$dst), (ins GR16:$src1,GR16:$src2), "imul{w}\t{$src2, $dst|$dst, $src2}", @@ -147,9 +169,10 @@ def IMUL64rr : RI<0xAF, MRMSrcReg, (outs GR64:$dst), [(set GR64:$dst, EFLAGS, (X86smul_flag GR64:$src1, GR64:$src2))], IIC_IMUL64_RR>, TB; -} +} // isCommutable, SchedRW // Register-Memory Signed Integer Multiply +let SchedRW = [WriteIMulLd, ReadAfterLd] in { def IMUL16rm : I<0xAF, MRMSrcMem, (outs GR16:$dst), (ins GR16:$src1, i16mem:$src2), "imul{w}\t{$src2, $dst|$dst, $src2}", @@ -157,7 +180,7 @@ def IMUL16rm : I<0xAF, MRMSrcMem, (outs GR16:$dst), (X86smul_flag GR16:$src1, (load addr:$src2)))], IIC_IMUL16_RM>, TB, OpSize; -def IMUL32rm : I<0xAF, MRMSrcMem, (outs GR32:$dst), +def IMUL32rm : I<0xAF, MRMSrcMem, (outs GR32:$dst), (ins GR32:$src1, i32mem:$src2), "imul{l}\t{$src2, $dst|$dst, $src2}", [(set GR32:$dst, EFLAGS, @@ -171,18 +194,20 @@ def IMUL64rm : RI<0xAF, MRMSrcMem, (outs GR64:$dst), (X86smul_flag GR64:$src1, (load addr:$src2)))], IIC_IMUL64_RM>, TB; +} // SchedRW } // Constraints = "$src1 = $dst" } // Defs = [EFLAGS] // Surprisingly enough, these are not two address instructions! let Defs = [EFLAGS] in { +let SchedRW = [WriteIMul] in { // Register-Integer Signed Integer Multiply def IMUL16rri : Ii16<0x69, MRMSrcReg, // GR16 = GR16*I16 (outs GR16:$dst), (ins GR16:$src1, i16imm:$src2), "imul{w}\t{$src2, $src1, $dst|$dst, $src1, $src2}", - [(set GR16:$dst, EFLAGS, - (X86smul_flag GR16:$src1, imm:$src2))], + [(set GR16:$dst, EFLAGS, + (X86smul_flag GR16:$src1, imm:$src2))], IIC_IMUL16_RRI>, OpSize; def IMUL16rri8 : Ii8<0x6B, MRMSrcReg, // GR16 = GR16*I8 (outs GR16:$dst), (ins GR16:$src1, i16i8imm:$src2), @@ -215,9 +240,10 @@ def IMUL64rri8 : RIi8<0x6B, MRMSrcReg, // GR64 = GR64*I8 [(set GR64:$dst, EFLAGS, (X86smul_flag GR64:$src1, i64immSExt8:$src2))], IIC_IMUL64_RRI>; - +} // SchedRW // Memory-Integer Signed Integer Multiply +let SchedRW = [WriteIMulLd] in { def IMUL16rmi : Ii16<0x69, MRMSrcMem, // GR16 = [mem16]*I16 (outs GR16:$dst), (ins i16mem:$src1, i16imm:$src2), "imul{w}\t{$src2, $src1, $dst|$dst, $src1, $src2}", @@ -259,12 +285,15 @@ def IMUL64rmi8 : RIi8<0x6B, MRMSrcMem, // GR64 = [mem64]*I8 (X86smul_flag (load addr:$src1), i64immSExt8:$src2))], IIC_IMUL64_RMI>; +} // SchedRW } // Defs = [EFLAGS] // unsigned division/remainder +let hasSideEffects = 1 in { // so that we don't speculatively execute +let SchedRW = [WriteIDiv] in { let Defs = [AL,EFLAGS,AX], Uses = [AX] in def DIV8r : I<0xF6, MRM6r, (outs), (ins GR8:$src), // AX/r8 = AL,AH "div{b}\t$src", [], IIC_DIV8_REG>; @@ -278,24 +307,30 @@ def DIV32r : I<0xF7, MRM6r, (outs), (ins GR32:$src), // EDX:EAX/r32 = EAX,EDX let Defs = [RAX,RDX,EFLAGS], Uses = [RAX,RDX] in def DIV64r : RI<0xF7, MRM6r, (outs), (ins GR64:$src), "div{q}\t$src", [], IIC_DIV64>; +} // SchedRW let mayLoad = 1 in { let Defs = [AL,EFLAGS,AX], Uses = [AX] in def DIV8m : I<0xF6, MRM6m, (outs), (ins i8mem:$src), // AX/[mem8] = AL,AH - "div{b}\t$src", [], IIC_DIV8_MEM>; + "div{b}\t$src", [], IIC_DIV8_MEM>, + SchedLoadReg; let Defs = [AX,DX,EFLAGS], Uses = [AX,DX] in def DIV16m : I<0xF7, MRM6m, (outs), (ins i16mem:$src), // DX:AX/[mem16] = AX,DX - "div{w}\t$src", [], IIC_DIV16>, OpSize; + "div{w}\t$src", [], IIC_DIV16>, OpSize, + SchedLoadReg; let Defs = [EAX,EDX,EFLAGS], Uses = [EAX,EDX] in // EDX:EAX/[mem32] = EAX,EDX def DIV32m : I<0xF7, MRM6m, (outs), (ins i32mem:$src), - "div{l}\t$src", [], IIC_DIV32>; + "div{l}\t$src", [], IIC_DIV32>, + SchedLoadReg; // RDX:RAX/[mem64] = RAX,RDX let Defs = [RAX,RDX,EFLAGS], Uses = [RAX,RDX] in def DIV64m : RI<0xF7, MRM6m, (outs), (ins i64mem:$src), - "div{q}\t$src", [], IIC_DIV64>; + "div{q}\t$src", [], IIC_DIV64>, + SchedLoadReg; } // Signed division/remainder. +let SchedRW = [WriteIDiv] in { let Defs = [AL,EFLAGS,AX], Uses = [AX] in def IDIV8r : I<0xF6, MRM7r, (outs), (ins GR8:$src), // AX/r8 = AL,AH "idiv{b}\t$src", [], IIC_IDIV8>; @@ -309,21 +344,27 @@ def IDIV32r: I<0xF7, MRM7r, (outs), (ins GR32:$src), // EDX:EAX/r32 = EAX,EDX let Defs = [RAX,RDX,EFLAGS], Uses = [RAX,RDX] in def IDIV64r: RI<0xF7, MRM7r, (outs), (ins GR64:$src), "idiv{q}\t$src", [], IIC_IDIV64>; +} // SchedRW let mayLoad = 1 in { let Defs = [AL,EFLAGS,AX], Uses = [AX] in def IDIV8m : I<0xF6, MRM7m, (outs), (ins i8mem:$src), // AX/[mem8] = AL,AH - "idiv{b}\t$src", [], IIC_IDIV8>; + "idiv{b}\t$src", [], IIC_IDIV8>, + SchedLoadReg; let Defs = [AX,DX,EFLAGS], Uses = [AX,DX] in def IDIV16m: I<0xF7, MRM7m, (outs), (ins i16mem:$src), // DX:AX/[mem16] = AX,DX - "idiv{w}\t$src", [], IIC_IDIV16>, OpSize; + "idiv{w}\t$src", [], IIC_IDIV16>, OpSize, + SchedLoadReg; let Defs = [EAX,EDX,EFLAGS], Uses = [EAX,EDX] in // EDX:EAX/[mem32] = EAX,EDX -def IDIV32m: I<0xF7, MRM7m, (outs), (ins i32mem:$src), - "idiv{l}\t$src", [], IIC_IDIV32>; +def IDIV32m: I<0xF7, MRM7m, (outs), (ins i32mem:$src), + "idiv{l}\t$src", [], IIC_IDIV32>, + SchedLoadReg; let Defs = [RAX,RDX,EFLAGS], Uses = [RAX,RDX] in // RDX:RAX/[mem64] = RAX,RDX def IDIV64m: RI<0xF7, MRM7m, (outs), (ins i64mem:$src), - "idiv{q}\t$src", [], IIC_IDIV64>; + "idiv{q}\t$src", [], IIC_IDIV64>, + SchedLoadReg; } +} // hasSideEffects = 0 //===----------------------------------------------------------------------===// // Two address Instructions. @@ -332,7 +373,7 @@ def IDIV64m: RI<0xF7, MRM7m, (outs), (ins i64mem:$src), // unary instructions let CodeSize = 2 in { let Defs = [EFLAGS] in { -let Constraints = "$src1 = $dst" in { +let Constraints = "$src1 = $dst", SchedRW = [WriteALU] in { def NEG8r : I<0xF6, MRM3r, (outs GR8 :$dst), (ins GR8 :$src1), "neg{b}\t$dst", [(set GR8:$dst, (ineg GR8:$src1)), @@ -348,8 +389,10 @@ def NEG32r : I<0xF7, MRM3r, (outs GR32:$dst), (ins GR32:$src1), def NEG64r : RI<0xF7, MRM3r, (outs GR64:$dst), (ins GR64:$src1), "neg{q}\t$dst", [(set GR64:$dst, (ineg GR64:$src1)), (implicit EFLAGS)], IIC_UNARY_REG>; -} // Constraints = "$src1 = $dst" +} // Constraints = "$src1 = $dst", SchedRW +// Read-modify-write negate. +let SchedRW = [WriteALULd, WriteRMW] in { def NEG8m : I<0xF6, MRM3m, (outs), (ins i8mem :$dst), "neg{b}\t$dst", [(store (ineg (loadi8 addr:$dst)), addr:$dst), @@ -365,12 +408,13 @@ def NEG32m : I<0xF7, MRM3m, (outs), (ins i32mem:$dst), def NEG64m : RI<0xF7, MRM3m, (outs), (ins i64mem:$dst), "neg{q}\t$dst", [(store (ineg (loadi64 addr:$dst)), addr:$dst), (implicit EFLAGS)], IIC_UNARY_MEM>; +} // SchedRW } // Defs = [EFLAGS] // Note: NOT does not set EFLAGS! -let Constraints = "$src1 = $dst" in { +let Constraints = "$src1 = $dst", SchedRW = [WriteALU] in { // Match xor -1 to not. Favors these over a move imm + xor to save code size. let AddedComplexity = 15 in { def NOT8r : I<0xF6, MRM2r, (outs GR8 :$dst), (ins GR8 :$src1), @@ -385,8 +429,9 @@ def NOT32r : I<0xF7, MRM2r, (outs GR32:$dst), (ins GR32:$src1), def NOT64r : RI<0xF7, MRM2r, (outs GR64:$dst), (ins GR64:$src1), "not{q}\t$dst", [(set GR64:$dst, (not GR64:$src1))], IIC_UNARY_REG>; } -} // Constraints = "$src1 = $dst" +} // Constraints = "$src1 = $dst", SchedRW +let SchedRW = [WriteALULd, WriteRMW] in { def NOT8m : I<0xF6, MRM2m, (outs), (ins i8mem :$dst), "not{b}\t$dst", [(store (not (loadi8 addr:$dst)), addr:$dst)], IIC_UNARY_MEM>; @@ -399,11 +444,12 @@ def NOT32m : I<0xF7, MRM2m, (outs), (ins i32mem:$dst), [(store (not (loadi32 addr:$dst)), addr:$dst)], IIC_UNARY_MEM>; def NOT64m : RI<0xF7, MRM2m, (outs), (ins i64mem:$dst), "not{q}\t$dst", [(store (not (loadi64 addr:$dst)), addr:$dst)], IIC_UNARY_MEM>; +} // SchedRW } // CodeSize // TODO: inc/dec is slow for P4, but fast for Pentium-M. let Defs = [EFLAGS] in { -let Constraints = "$src1 = $dst" in { +let Constraints = "$src1 = $dst", SchedRW = [WriteALU] in { let CodeSize = 2 in def INC8r : I<0xFE, MRM0r, (outs GR8 :$dst), (ins GR8 :$src1), "inc{b}\t$dst", @@ -411,11 +457,11 @@ def INC8r : I<0xFE, MRM0r, (outs GR8 :$dst), (ins GR8 :$src1), IIC_UNARY_REG>; let isConvertibleToThreeAddress = 1, CodeSize = 1 in { // Can xform into LEA. -def INC16r : I<0x40, AddRegFrm, (outs GR16:$dst), (ins GR16:$src1), +def INC16r : I<0x40, AddRegFrm, (outs GR16:$dst), (ins GR16:$src1), "inc{w}\t$dst", [(set GR16:$dst, EFLAGS, (X86inc_flag GR16:$src1))], IIC_UNARY_REG>, OpSize, Requires<[In32BitMode]>; -def INC32r : I<0x40, AddRegFrm, (outs GR32:$dst), (ins GR32:$src1), +def INC32r : I<0x40, AddRegFrm, (outs GR32:$dst), (ins GR32:$src1), "inc{l}\t$dst", [(set GR32:$dst, EFLAGS, (X86inc_flag GR32:$src1))], IIC_UNARY_REG>, @@ -429,31 +475,31 @@ def INC64r : RI<0xFF, MRM0r, (outs GR64:$dst), (ins GR64:$src1), "inc{q}\t$dst", // In 64-bit mode, single byte INC and DEC cannot be encoded. let isConvertibleToThreeAddress = 1, CodeSize = 2 in { // Can transform into LEA. -def INC64_16r : I<0xFF, MRM0r, (outs GR16:$dst), (ins GR16:$src1), +def INC64_16r : I<0xFF, MRM0r, (outs GR16:$dst), (ins GR16:$src1), "inc{w}\t$dst", [(set GR16:$dst, EFLAGS, (X86inc_flag GR16:$src1))], IIC_UNARY_REG>, OpSize, Requires<[In64BitMode]>; -def INC64_32r : I<0xFF, MRM0r, (outs GR32:$dst), (ins GR32:$src1), +def INC64_32r : I<0xFF, MRM0r, (outs GR32:$dst), (ins GR32:$src1), "inc{l}\t$dst", [(set GR32:$dst, EFLAGS, (X86inc_flag GR32:$src1))], IIC_UNARY_REG>, Requires<[In64BitMode]>; -def DEC64_16r : I<0xFF, MRM1r, (outs GR16:$dst), (ins GR16:$src1), +def DEC64_16r : I<0xFF, MRM1r, (outs GR16:$dst), (ins GR16:$src1), "dec{w}\t$dst", [(set GR16:$dst, EFLAGS, (X86dec_flag GR16:$src1))], IIC_UNARY_REG>, OpSize, Requires<[In64BitMode]>; -def DEC64_32r : I<0xFF, MRM1r, (outs GR32:$dst), (ins GR32:$src1), +def DEC64_32r : I<0xFF, MRM1r, (outs GR32:$dst), (ins GR32:$src1), "dec{l}\t$dst", [(set GR32:$dst, EFLAGS, (X86dec_flag GR32:$src1))], IIC_UNARY_REG>, Requires<[In64BitMode]>; } // isConvertibleToThreeAddress = 1, CodeSize = 2 -} // Constraints = "$src1 = $dst" +} // Constraints = "$src1 = $dst", SchedRW -let CodeSize = 2 in { +let CodeSize = 2, SchedRW = [WriteALULd, WriteRMW] in { def INC8m : I<0xFE, MRM0m, (outs), (ins i8mem :$dst), "inc{b}\t$dst", [(store (add (loadi8 addr:$dst), 1), addr:$dst), (implicit EFLAGS)], IIC_UNARY_MEM>; @@ -468,7 +514,7 @@ let CodeSize = 2 in { def INC64m : RI<0xFF, MRM0m, (outs), (ins i64mem:$dst), "inc{q}\t$dst", [(store (add (loadi64 addr:$dst), 1), addr:$dst), (implicit EFLAGS)], IIC_UNARY_MEM>; - + // These are duplicates of their 32-bit counterparts. Only needed so X86 knows // how to unfold them. // FIXME: What is this for?? @@ -488,21 +534,21 @@ def DEC64_32m : I<0xFF, MRM1m, (outs), (ins i32mem:$dst), "dec{l}\t$dst", [(store (add (loadi32 addr:$dst), -1), addr:$dst), (implicit EFLAGS)], IIC_UNARY_MEM>, Requires<[In64BitMode]>; -} // CodeSize = 2 +} // CodeSize = 2, SchedRW -let Constraints = "$src1 = $dst" in { +let Constraints = "$src1 = $dst", SchedRW = [WriteALU] in { let CodeSize = 2 in def DEC8r : I<0xFE, MRM1r, (outs GR8 :$dst), (ins GR8 :$src1), "dec{b}\t$dst", [(set GR8:$dst, EFLAGS, (X86dec_flag GR8:$src1))], IIC_UNARY_REG>; let isConvertibleToThreeAddress = 1, CodeSize = 1 in { // Can xform into LEA. -def DEC16r : I<0x48, AddRegFrm, (outs GR16:$dst), (ins GR16:$src1), +def DEC16r : I<0x48, AddRegFrm, (outs GR16:$dst), (ins GR16:$src1), "dec{w}\t$dst", [(set GR16:$dst, EFLAGS, (X86dec_flag GR16:$src1))], IIC_UNARY_REG>, OpSize, Requires<[In32BitMode]>; -def DEC32r : I<0x48, AddRegFrm, (outs GR32:$dst), (ins GR32:$src1), +def DEC32r : I<0x48, AddRegFrm, (outs GR32:$dst), (ins GR32:$src1), "dec{l}\t$dst", [(set GR32:$dst, EFLAGS, (X86dec_flag GR32:$src1))], IIC_UNARY_REG>, @@ -511,10 +557,10 @@ def DEC64r : RI<0xFF, MRM1r, (outs GR64:$dst), (ins GR64:$src1), "dec{q}\t$dst", [(set GR64:$dst, EFLAGS, (X86dec_flag GR64:$src1))], IIC_UNARY_REG>; } // CodeSize = 2 -} // Constraints = "$src1 = $dst" +} // Constraints = "$src1 = $dst", SchedRW -let CodeSize = 2 in { +let CodeSize = 2, SchedRW = [WriteALULd, WriteRMW] in { def DEC8m : I<0xFE, MRM1m, (outs), (ins i8mem :$dst), "dec{b}\t$dst", [(store (add (loadi8 addr:$dst), -1), addr:$dst), (implicit EFLAGS)], IIC_UNARY_MEM>; @@ -529,7 +575,7 @@ let CodeSize = 2 in { def DEC64m : RI<0xFF, MRM1m, (outs), (ins i64mem:$dst), "dec{q}\t$dst", [(store (add (loadi64 addr:$dst), -1), addr:$dst), (implicit EFLAGS)], IIC_UNARY_MEM>; -} // CodeSize = 2 +} // CodeSize = 2, SchedRW } // Defs = [EFLAGS] @@ -543,57 +589,57 @@ class X86TypeInfo { /// VT - This is the value type itself. ValueType VT = vt; - + /// InstrSuffix - This is the suffix used on instructions with this type. For /// example, i8 -> "b", i16 -> "w", i32 -> "l", i64 -> "q". string InstrSuffix = instrsuffix; - + /// RegClass - This is the register class associated with this type. For /// example, i8 -> GR8, i16 -> GR16, i32 -> GR32, i64 -> GR64. RegisterClass RegClass = regclass; - + /// LoadNode - This is the load node associated with this type. For /// example, i8 -> loadi8, i16 -> loadi16, i32 -> loadi32, i64 -> loadi64. PatFrag LoadNode = loadnode; - + /// MemOperand - This is the memory operand associated with this type. For /// example, i8 -> i8mem, i16 -> i16mem, i32 -> i32mem, i64 -> i64mem. X86MemOperand MemOperand = memoperand; - + /// ImmEncoding - This is the encoding of an immediate of this type. For /// example, i8 -> Imm8, i16 -> Imm16, i32 -> Imm32. Note that i64 -> Imm32 /// since the immediate fields of i64 instructions is a 32-bit sign extended /// value. ImmType ImmEncoding = immkind; - + /// ImmOperand - This is the operand kind of an immediate of this type. For /// example, i8 -> i8imm, i16 -> i16imm, i32 -> i32imm. Note that i64 -> /// i64i32imm since the immediate fields of i64 instructions is a 32-bit sign /// extended value. Operand ImmOperand = immoperand; - + /// ImmOperator - This is the operator that should be used to match an /// immediate of this kind in a pattern (e.g. imm, or i64immSExt32). SDPatternOperator ImmOperator = immoperator; - + /// Imm8Operand - This is the operand kind to use for an imm8 of this type. /// For example, i8 -> , i16 -> i16i8imm, i32 -> i32i8imm. This is /// only used for instructions that have a sign-extended imm8 field form. Operand Imm8Operand = imm8operand; - + /// Imm8Operator - This is the operator that should be used to match an 8-bit /// sign extended immediate of this kind in a pattern (e.g. imm16immSExt8). SDPatternOperator Imm8Operator = imm8operator; - + /// HasOddOpcode - This bit is true if the instruction should have an odd (as /// opposed to even) opcode. Operations on i8 are usually even, operations on /// other datatypes are odd. bit HasOddOpcode = hasOddOpcode; - + /// HasOpSizePrefix - This bit is set to true if the instruction should have /// the 0x66 operand size prefix. This is set for i16 types. bit HasOpSizePrefix = hasOpSizePrefix; - + /// HasREX_WPrefix - This bit is set to true if the instruction should have /// the 0x40 REX prefix. This is set for i64 types. bit HasREX_WPrefix = hasREX_WPrefix; @@ -623,12 +669,12 @@ def Xi64 : X86TypeInfo opcode, Format f, X86TypeInfo typeinfo, dag outs, dag ins, +class ITy opcode, Format f, X86TypeInfo typeinfo, dag outs, dag ins, string mnemonic, string args, list pattern, InstrItinClass itin = IIC_BIN_NONMEM> : I<{opcode{7}, opcode{6}, opcode{5}, opcode{4}, opcode{3}, opcode{2}, opcode{1}, typeinfo.HasOddOpcode }, - f, outs, ins, + f, outs, ins, !strconcat(mnemonic, "{", typeinfo.InstrSuffix, "}\t", args), pattern, itin> { @@ -639,10 +685,12 @@ class ITy opcode, Format f, X86TypeInfo typeinfo, dag outs, dag ins, // BinOpRR - Instructions like "add reg, reg, reg". class BinOpRR opcode, string mnemonic, X86TypeInfo typeinfo, - dag outlist, list pattern, Format f = MRMDestReg> + dag outlist, list pattern, InstrItinClass itin, + Format f = MRMDestReg> : ITy; + mnemonic, "{$src2, $src1|$src1, $src2}", pattern, itin>, + Sched<[WriteALU]>; // BinOpRR_R - Instructions like "add reg, reg, reg", where the pattern has // just a regclass (no eflags) as a result. @@ -650,7 +698,8 @@ class BinOpRR_R opcode, string mnemonic, X86TypeInfo typeinfo, SDNode opnode> : BinOpRR; + (opnode typeinfo.RegClass:$src1, typeinfo.RegClass:$src2))], + IIC_BIN_NONMEM>; // BinOpRR_F - Instructions like "cmp reg, Reg", where the pattern has // just a EFLAGS as a result. @@ -659,7 +708,7 @@ class BinOpRR_F opcode, string mnemonic, X86TypeInfo typeinfo, : BinOpRR; + IIC_BIN_NONMEM, f>; // BinOpRR_RF - Instructions like "add reg, reg, reg", where the pattern has // both a regclass and EFLAGS as a result. @@ -667,7 +716,8 @@ class BinOpRR_RF opcode, string mnemonic, X86TypeInfo typeinfo, SDNode opnode> : BinOpRR; + (opnode typeinfo.RegClass:$src1, typeinfo.RegClass:$src2))], + IIC_BIN_NONMEM>; // BinOpRR_RFF - Instructions like "adc reg, reg, reg", where the pattern has // both a regclass and EFLAGS as a result, and has EFLAGS as input. @@ -676,25 +726,29 @@ class BinOpRR_RFF opcode, string mnemonic, X86TypeInfo typeinfo, : BinOpRR; + EFLAGS))], IIC_BIN_NONMEM>; // BinOpRR_Rev - Instructions like "add reg, reg, reg" (reversed encoding). class BinOpRR_Rev opcode, string mnemonic, X86TypeInfo typeinfo> : ITy { + mnemonic, "{$src2, $dst|$dst, $src2}", [], IIC_BIN_NONMEM>, + Sched<[WriteALU]> { // The disassembler should know about this, but not the asmparser. let isCodeGenOnly = 1; + let hasSideEffects = 0; } // BinOpRR_F_Rev - Instructions like "cmp reg, reg" (reversed encoding). class BinOpRR_F_Rev opcode, string mnemonic, X86TypeInfo typeinfo> : ITy { + mnemonic, "{$src2, $src1|$src1, $src2}", [], IIC_BIN_NONMEM>, + Sched<[WriteALU]> { // The disassembler should know about this, but not the asmparser. let isCodeGenOnly = 1; + let hasSideEffects = 0; } // BinOpRM - Instructions like "add reg, reg, [mem]". @@ -702,7 +756,8 @@ class BinOpRM opcode, string mnemonic, X86TypeInfo typeinfo, dag outlist, list pattern> : ITy; + mnemonic, "{$src2, $src1|$src1, $src2}", pattern, IIC_BIN_NONMEM>, + Sched<[WriteALULd, ReadAfterLd]>; // BinOpRM_R - Instructions like "add reg, reg, [mem]". class BinOpRM_R opcode, string mnemonic, X86TypeInfo typeinfo, @@ -738,7 +793,8 @@ class BinOpRI opcode, string mnemonic, X86TypeInfo typeinfo, Format f, dag outlist, list pattern> : ITy { + mnemonic, "{$src2, $src1|$src1, $src2}", pattern, IIC_BIN_NONMEM>, + Sched<[WriteALU]> { let ImmT = typeinfo.ImmEncoding; } @@ -760,14 +816,13 @@ class BinOpRI_F opcode, string mnemonic, X86TypeInfo typeinfo, class BinOpRI_RF opcode, string mnemonic, X86TypeInfo typeinfo, SDNode opnode, Format f> : BinOpRI; - // BinOpRI_RFF - Instructions like "adc reg, reg, imm". class BinOpRI_RFF opcode, string mnemonic, X86TypeInfo typeinfo, SDNode opnode, Format f> : BinOpRI; @@ -776,7 +831,8 @@ class BinOpRI8 opcode, string mnemonic, X86TypeInfo typeinfo, Format f, dag outlist, list pattern> : ITy { + mnemonic, "{$src2, $src1|$src1, $src2}", pattern, IIC_BIN_NONMEM>, + Sched<[WriteALU]> { let ImmT = Imm8; // Always 8-bit immediate. } @@ -786,7 +842,7 @@ class BinOpRI8_R opcode, string mnemonic, X86TypeInfo typeinfo, : BinOpRI8; - + // BinOpRI8_F - Instructions like "cmp reg, imm8". class BinOpRI8_F opcode, string mnemonic, X86TypeInfo typeinfo, SDNode opnode, Format f> @@ -814,7 +870,8 @@ class BinOpMR opcode, string mnemonic, X86TypeInfo typeinfo, list pattern> : ITy; + mnemonic, "{$src, $dst|$dst, $src}", pattern, IIC_BIN_MEM>, + Sched<[WriteALULd, WriteRMW]>; // BinOpMR_RMW - Instructions like "add [mem], reg". class BinOpMR_RMW opcode, string mnemonic, X86TypeInfo typeinfo, @@ -842,22 +899,22 @@ class BinOpMI pattern, bits<8> opcode = 0x80> : ITy { + mnemonic, "{$src, $dst|$dst, $src}", pattern, IIC_BIN_MEM>, + Sched<[WriteALULd, WriteRMW]> { let ImmT = typeinfo.ImmEncoding; } // BinOpMI_RMW - Instructions like "add [mem], imm". class BinOpMI_RMW - : BinOpMI; - // BinOpMI_RMW_FF - Instructions like "adc [mem], imm". class BinOpMI_RMW_FF - : BinOpMI; @@ -865,7 +922,7 @@ class BinOpMI_RMW_FF opcode = 0x80> - : BinOpMI; @@ -875,7 +932,8 @@ class BinOpMI8 pattern> : ITy<0x82, f, typeinfo, (outs), (ins typeinfo.MemOperand:$dst, typeinfo.Imm8Operand:$src), - mnemonic, "{$src, $dst|$dst, $src}", pattern, IIC_BIN_MEM> { + mnemonic, "{$src, $dst|$dst, $src}", pattern, IIC_BIN_MEM>, + Sched<[WriteALULd, WriteRMW]> { let ImmT = Imm8; // Always 8-bit immediate. } @@ -907,10 +965,11 @@ class BinOpAI opcode, string mnemonic, X86TypeInfo typeinfo, Register areg, string operands> : ITy { + mnemonic, operands, []>, Sched<[WriteALU]> { let ImmT = typeinfo.ImmEncoding; let Uses = [areg]; let Defs = [areg]; + let hasSideEffects = 0; } /// ArithBinOp_RF - This is an arithmetic binary operator where the pattern is @@ -926,61 +985,61 @@ multiclass ArithBinOp_RF BaseOpc, bits<8> BaseOpc2, bits<8> BaseOpc4, let Constraints = "$src1 = $dst" in { let isCommutable = CommutableRR, isConvertibleToThreeAddress = ConvertibleToThreeAddress in { - def #NAME#8rr : BinOpRR_RF; - def #NAME#16rr : BinOpRR_RF; - def #NAME#32rr : BinOpRR_RF; - def #NAME#64rr : BinOpRR_RF; + def NAME#8rr : BinOpRR_RF; + def NAME#16rr : BinOpRR_RF; + def NAME#32rr : BinOpRR_RF; + def NAME#64rr : BinOpRR_RF; } // isCommutable - def #NAME#8rr_REV : BinOpRR_Rev; - def #NAME#16rr_REV : BinOpRR_Rev; - def #NAME#32rr_REV : BinOpRR_Rev; - def #NAME#64rr_REV : BinOpRR_Rev; + def NAME#8rr_REV : BinOpRR_Rev; + def NAME#16rr_REV : BinOpRR_Rev; + def NAME#32rr_REV : BinOpRR_Rev; + def NAME#64rr_REV : BinOpRR_Rev; - def #NAME#8rm : BinOpRM_RF; - def #NAME#16rm : BinOpRM_RF; - def #NAME#32rm : BinOpRM_RF; - def #NAME#64rm : BinOpRM_RF; + def NAME#8rm : BinOpRM_RF; + def NAME#16rm : BinOpRM_RF; + def NAME#32rm : BinOpRM_RF; + def NAME#64rm : BinOpRM_RF; let isConvertibleToThreeAddress = ConvertibleToThreeAddress in { // NOTE: These are order specific, we want the ri8 forms to be listed // first so that they are slightly preferred to the ri forms. - def #NAME#16ri8 : BinOpRI8_RF<0x82, mnemonic, Xi16, opnodeflag, RegMRM>; - def #NAME#32ri8 : BinOpRI8_RF<0x82, mnemonic, Xi32, opnodeflag, RegMRM>; - def #NAME#64ri8 : BinOpRI8_RF<0x82, mnemonic, Xi64, opnodeflag, RegMRM>; - - def #NAME#8ri : BinOpRI_RF<0x80, mnemonic, Xi8 , opnodeflag, RegMRM>; - def #NAME#16ri : BinOpRI_RF<0x80, mnemonic, Xi16, opnodeflag, RegMRM>; - def #NAME#32ri : BinOpRI_RF<0x80, mnemonic, Xi32, opnodeflag, RegMRM>; - def #NAME#64ri32: BinOpRI_RF<0x80, mnemonic, Xi64, opnodeflag, RegMRM>; + def NAME#16ri8 : BinOpRI8_RF<0x82, mnemonic, Xi16, opnodeflag, RegMRM>; + def NAME#32ri8 : BinOpRI8_RF<0x82, mnemonic, Xi32, opnodeflag, RegMRM>; + def NAME#64ri8 : BinOpRI8_RF<0x82, mnemonic, Xi64, opnodeflag, RegMRM>; + + def NAME#8ri : BinOpRI_RF<0x80, mnemonic, Xi8 , opnodeflag, RegMRM>; + def NAME#16ri : BinOpRI_RF<0x80, mnemonic, Xi16, opnodeflag, RegMRM>; + def NAME#32ri : BinOpRI_RF<0x80, mnemonic, Xi32, opnodeflag, RegMRM>; + def NAME#64ri32: BinOpRI_RF<0x80, mnemonic, Xi64, opnodeflag, RegMRM>; } } // Constraints = "$src1 = $dst" - def #NAME#8mr : BinOpMR_RMW; - def #NAME#16mr : BinOpMR_RMW; - def #NAME#32mr : BinOpMR_RMW; - def #NAME#64mr : BinOpMR_RMW; + def NAME#8mr : BinOpMR_RMW; + def NAME#16mr : BinOpMR_RMW; + def NAME#32mr : BinOpMR_RMW; + def NAME#64mr : BinOpMR_RMW; // NOTE: These are order specific, we want the mi8 forms to be listed // first so that they are slightly preferred to the mi forms. - def #NAME#16mi8 : BinOpMI8_RMW; - def #NAME#32mi8 : BinOpMI8_RMW; - def #NAME#64mi8 : BinOpMI8_RMW; - - def #NAME#8mi : BinOpMI_RMW; - def #NAME#16mi : BinOpMI_RMW; - def #NAME#32mi : BinOpMI_RMW; - def #NAME#64mi32 : BinOpMI_RMW; - - def #NAME#8i8 : BinOpAI; - def #NAME#16i16 : BinOpAI; - def #NAME#32i32 : BinOpAI; - def #NAME#64i32 : BinOpAI; - } + def NAME#16mi8 : BinOpMI8_RMW; + def NAME#32mi8 : BinOpMI8_RMW; + def NAME#64mi8 : BinOpMI8_RMW; + + def NAME#8mi : BinOpMI_RMW; + def NAME#16mi : BinOpMI_RMW; + def NAME#32mi : BinOpMI_RMW; + def NAME#64mi32 : BinOpMI_RMW; + + def NAME#8i8 : BinOpAI; + def NAME#16i16 : BinOpAI; + def NAME#32i32 : BinOpAI; + def NAME#64i32 : BinOpAI; + } } /// ArithBinOp_RFF - This is an arithmetic binary operator where the pattern is @@ -997,61 +1056,61 @@ multiclass ArithBinOp_RFF BaseOpc, bits<8> BaseOpc2, bits<8> BaseOpc4, let Constraints = "$src1 = $dst" in { let isCommutable = CommutableRR, isConvertibleToThreeAddress = ConvertibleToThreeAddress in { - def #NAME#8rr : BinOpRR_RFF; - def #NAME#16rr : BinOpRR_RFF; - def #NAME#32rr : BinOpRR_RFF; - def #NAME#64rr : BinOpRR_RFF; + def NAME#8rr : BinOpRR_RFF; + def NAME#16rr : BinOpRR_RFF; + def NAME#32rr : BinOpRR_RFF; + def NAME#64rr : BinOpRR_RFF; } // isCommutable - def #NAME#8rr_REV : BinOpRR_Rev; - def #NAME#16rr_REV : BinOpRR_Rev; - def #NAME#32rr_REV : BinOpRR_Rev; - def #NAME#64rr_REV : BinOpRR_Rev; + def NAME#8rr_REV : BinOpRR_Rev; + def NAME#16rr_REV : BinOpRR_Rev; + def NAME#32rr_REV : BinOpRR_Rev; + def NAME#64rr_REV : BinOpRR_Rev; - def #NAME#8rm : BinOpRM_RFF; - def #NAME#16rm : BinOpRM_RFF; - def #NAME#32rm : BinOpRM_RFF; - def #NAME#64rm : BinOpRM_RFF; + def NAME#8rm : BinOpRM_RFF; + def NAME#16rm : BinOpRM_RFF; + def NAME#32rm : BinOpRM_RFF; + def NAME#64rm : BinOpRM_RFF; let isConvertibleToThreeAddress = ConvertibleToThreeAddress in { // NOTE: These are order specific, we want the ri8 forms to be listed // first so that they are slightly preferred to the ri forms. - def #NAME#16ri8 : BinOpRI8_RFF<0x82, mnemonic, Xi16, opnode, RegMRM>; - def #NAME#32ri8 : BinOpRI8_RFF<0x82, mnemonic, Xi32, opnode, RegMRM>; - def #NAME#64ri8 : BinOpRI8_RFF<0x82, mnemonic, Xi64, opnode, RegMRM>; - - def #NAME#8ri : BinOpRI_RFF<0x80, mnemonic, Xi8 , opnode, RegMRM>; - def #NAME#16ri : BinOpRI_RFF<0x80, mnemonic, Xi16, opnode, RegMRM>; - def #NAME#32ri : BinOpRI_RFF<0x80, mnemonic, Xi32, opnode, RegMRM>; - def #NAME#64ri32: BinOpRI_RFF<0x80, mnemonic, Xi64, opnode, RegMRM>; + def NAME#16ri8 : BinOpRI8_RFF<0x82, mnemonic, Xi16, opnode, RegMRM>; + def NAME#32ri8 : BinOpRI8_RFF<0x82, mnemonic, Xi32, opnode, RegMRM>; + def NAME#64ri8 : BinOpRI8_RFF<0x82, mnemonic, Xi64, opnode, RegMRM>; + + def NAME#8ri : BinOpRI_RFF<0x80, mnemonic, Xi8 , opnode, RegMRM>; + def NAME#16ri : BinOpRI_RFF<0x80, mnemonic, Xi16, opnode, RegMRM>; + def NAME#32ri : BinOpRI_RFF<0x80, mnemonic, Xi32, opnode, RegMRM>; + def NAME#64ri32: BinOpRI_RFF<0x80, mnemonic, Xi64, opnode, RegMRM>; } } // Constraints = "$src1 = $dst" - def #NAME#8mr : BinOpMR_RMW_FF; - def #NAME#16mr : BinOpMR_RMW_FF; - def #NAME#32mr : BinOpMR_RMW_FF; - def #NAME#64mr : BinOpMR_RMW_FF; + def NAME#8mr : BinOpMR_RMW_FF; + def NAME#16mr : BinOpMR_RMW_FF; + def NAME#32mr : BinOpMR_RMW_FF; + def NAME#64mr : BinOpMR_RMW_FF; // NOTE: These are order specific, we want the mi8 forms to be listed // first so that they are slightly preferred to the mi forms. - def #NAME#16mi8 : BinOpMI8_RMW_FF; - def #NAME#32mi8 : BinOpMI8_RMW_FF; - def #NAME#64mi8 : BinOpMI8_RMW_FF; - - def #NAME#8mi : BinOpMI_RMW_FF; - def #NAME#16mi : BinOpMI_RMW_FF; - def #NAME#32mi : BinOpMI_RMW_FF; - def #NAME#64mi32 : BinOpMI_RMW_FF; - - def #NAME#8i8 : BinOpAI; - def #NAME#16i16 : BinOpAI; - def #NAME#32i32 : BinOpAI; - def #NAME#64i32 : BinOpAI; - } + def NAME#16mi8 : BinOpMI8_RMW_FF; + def NAME#32mi8 : BinOpMI8_RMW_FF; + def NAME#64mi8 : BinOpMI8_RMW_FF; + + def NAME#8mi : BinOpMI_RMW_FF; + def NAME#16mi : BinOpMI_RMW_FF; + def NAME#32mi : BinOpMI_RMW_FF; + def NAME#64mi32 : BinOpMI_RMW_FF; + + def NAME#8i8 : BinOpAI; + def NAME#16i16 : BinOpAI; + def NAME#32i32 : BinOpAI; + def NAME#64i32 : BinOpAI; + } } /// ArithBinOp_F - This is an arithmetic binary operator where the pattern is @@ -1065,60 +1124,60 @@ multiclass ArithBinOp_F BaseOpc, bits<8> BaseOpc2, bits<8> BaseOpc4, let Defs = [EFLAGS] in { let isCommutable = CommutableRR, isConvertibleToThreeAddress = ConvertibleToThreeAddress in { - def #NAME#8rr : BinOpRR_F; - def #NAME#16rr : BinOpRR_F; - def #NAME#32rr : BinOpRR_F; - def #NAME#64rr : BinOpRR_F; + def NAME#8rr : BinOpRR_F; + def NAME#16rr : BinOpRR_F; + def NAME#32rr : BinOpRR_F; + def NAME#64rr : BinOpRR_F; } // isCommutable - def #NAME#8rr_REV : BinOpRR_F_Rev; - def #NAME#16rr_REV : BinOpRR_F_Rev; - def #NAME#32rr_REV : BinOpRR_F_Rev; - def #NAME#64rr_REV : BinOpRR_F_Rev; + def NAME#8rr_REV : BinOpRR_F_Rev; + def NAME#16rr_REV : BinOpRR_F_Rev; + def NAME#32rr_REV : BinOpRR_F_Rev; + def NAME#64rr_REV : BinOpRR_F_Rev; - def #NAME#8rm : BinOpRM_F; - def #NAME#16rm : BinOpRM_F; - def #NAME#32rm : BinOpRM_F; - def #NAME#64rm : BinOpRM_F; + def NAME#8rm : BinOpRM_F; + def NAME#16rm : BinOpRM_F; + def NAME#32rm : BinOpRM_F; + def NAME#64rm : BinOpRM_F; let isConvertibleToThreeAddress = ConvertibleToThreeAddress in { // NOTE: These are order specific, we want the ri8 forms to be listed // first so that they are slightly preferred to the ri forms. - def #NAME#16ri8 : BinOpRI8_F<0x82, mnemonic, Xi16, opnode, RegMRM>; - def #NAME#32ri8 : BinOpRI8_F<0x82, mnemonic, Xi32, opnode, RegMRM>; - def #NAME#64ri8 : BinOpRI8_F<0x82, mnemonic, Xi64, opnode, RegMRM>; - - def #NAME#8ri : BinOpRI_F<0x80, mnemonic, Xi8 , opnode, RegMRM>; - def #NAME#16ri : BinOpRI_F<0x80, mnemonic, Xi16, opnode, RegMRM>; - def #NAME#32ri : BinOpRI_F<0x80, mnemonic, Xi32, opnode, RegMRM>; - def #NAME#64ri32: BinOpRI_F<0x80, mnemonic, Xi64, opnode, RegMRM>; + def NAME#16ri8 : BinOpRI8_F<0x82, mnemonic, Xi16, opnode, RegMRM>; + def NAME#32ri8 : BinOpRI8_F<0x82, mnemonic, Xi32, opnode, RegMRM>; + def NAME#64ri8 : BinOpRI8_F<0x82, mnemonic, Xi64, opnode, RegMRM>; + + def NAME#8ri : BinOpRI_F<0x80, mnemonic, Xi8 , opnode, RegMRM>; + def NAME#16ri : BinOpRI_F<0x80, mnemonic, Xi16, opnode, RegMRM>; + def NAME#32ri : BinOpRI_F<0x80, mnemonic, Xi32, opnode, RegMRM>; + def NAME#64ri32: BinOpRI_F<0x80, mnemonic, Xi64, opnode, RegMRM>; } - def #NAME#8mr : BinOpMR_F; - def #NAME#16mr : BinOpMR_F; - def #NAME#32mr : BinOpMR_F; - def #NAME#64mr : BinOpMR_F; + def NAME#8mr : BinOpMR_F; + def NAME#16mr : BinOpMR_F; + def NAME#32mr : BinOpMR_F; + def NAME#64mr : BinOpMR_F; // NOTE: These are order specific, we want the mi8 forms to be listed // first so that they are slightly preferred to the mi forms. - def #NAME#16mi8 : BinOpMI8_F; - def #NAME#32mi8 : BinOpMI8_F; - def #NAME#64mi8 : BinOpMI8_F; - - def #NAME#8mi : BinOpMI_F; - def #NAME#16mi : BinOpMI_F; - def #NAME#32mi : BinOpMI_F; - def #NAME#64mi32 : BinOpMI_F; - - def #NAME#8i8 : BinOpAI; - def #NAME#16i16 : BinOpAI; - def #NAME#32i32 : BinOpAI; - def #NAME#64i32 : BinOpAI; - } + def NAME#16mi8 : BinOpMI8_F; + def NAME#32mi8 : BinOpMI8_F; + def NAME#64mi8 : BinOpMI8_F; + + def NAME#8mi : BinOpMI_F; + def NAME#16mi : BinOpMI_F; + def NAME#32mi : BinOpMI_F; + def NAME#64mi32 : BinOpMI_F; + + def NAME#8i8 : BinOpAI; + def NAME#16i16 : BinOpAI; + def NAME#32i32 : BinOpAI; + def NAME#64i32 : BinOpAI; + } } @@ -1130,8 +1189,10 @@ defm XOR : ArithBinOp_RF<0x30, 0x32, 0x34, "xor", MRM6r, MRM6m, X86xor_flag, xor, 1, 0>; defm ADD : ArithBinOp_RF<0x00, 0x02, 0x04, "add", MRM0r, MRM0m, X86add_flag, add, 1, 1>; +let isCompare = 1 in { defm SUB : ArithBinOp_RF<0x28, 0x2A, 0x2C, "sub", MRM5r, MRM5m, X86sub_flag, sub, 0, 0>; +} // Arithmetic. let Uses = [EFLAGS] in { @@ -1141,7 +1202,9 @@ let Uses = [EFLAGS] in { 0, 0>; } +let isCompare = 1 in { defm CMP : ArithBinOp_F<0x38, 0x3A, 0x3C, "cmp", MRM7r, MRM7m, X86cmp, 0, 0>; +} //===----------------------------------------------------------------------===// @@ -1152,7 +1215,7 @@ defm CMP : ArithBinOp_F<0x38, 0x3A, 0x3C, "cmp", MRM7r, MRM7m, X86cmp, 0, 0>; def X86testpat : PatFrag<(ops node:$lhs, node:$rhs), (X86cmp (and_su node:$lhs, node:$rhs), 0)>; -let Defs = [EFLAGS] in { +let isCompare = 1, Defs = [EFLAGS] in { let isCommutable = 1 in { def TEST8rr : BinOpRR_F<0x84, "test", Xi8 , X86testpat, MRMSrcReg>; def TEST16rr : BinOpRR_F<0x84, "test", Xi16, X86testpat, MRMSrcReg>; @@ -1174,7 +1237,7 @@ let Defs = [EFLAGS] in { def TEST16mi : BinOpMI_F<"test", Xi16, X86testpat, MRM0m, 0xF6>; def TEST32mi : BinOpMI_F<"test", Xi32, X86testpat, MRM0m, 0xF6>; def TEST64mi32 : BinOpMI_F<"test", Xi64, X86testpat, MRM0m, 0xF6>; - + def TEST8i8 : BinOpAI<0xA8, "test", Xi8 , AL, "{$src, %al|AL, $src}">; def TEST16i16 : BinOpAI<0xA8, "test", Xi16, AX, @@ -1188,7 +1251,7 @@ let Defs = [EFLAGS] in { // register class is constrained to GR8_NOREX. let isPseudo = 1 in def TEST8ri_NOREX : I<0, Pseudo, (outs), (ins GR8_NOREX:$src, i8imm:$mask), - "", [], IIC_BIN_NONMEM>; + "", [], IIC_BIN_NONMEM>, Sched<[WriteALU]>; } //===----------------------------------------------------------------------===// @@ -1198,12 +1261,13 @@ multiclass bmi_andn { def rr : I<0xF2, MRMSrcReg, (outs RC:$dst), (ins RC:$src1, RC:$src2), !strconcat(mnemonic, "\t{$src2, $src1, $dst|$dst, $src1, $src2}"), - [(set RC:$dst, EFLAGS, (X86andn_flag RC:$src1, RC:$src2))], - IIC_BIN_NONMEM>; + [(set RC:$dst, EFLAGS, (X86and_flag (not RC:$src1), RC:$src2))], + IIC_BIN_NONMEM>, Sched<[WriteALU]>; def rm : I<0xF2, MRMSrcMem, (outs RC:$dst), (ins RC:$src1, x86memop:$src2), !strconcat(mnemonic, "\t{$src2, $src1, $dst|$dst, $src1, $src2}"), [(set RC:$dst, EFLAGS, - (X86andn_flag RC:$src1, (ld_frag addr:$src2)))], IIC_BIN_MEM>; + (X86and_flag (not RC:$src1), (ld_frag addr:$src2)))], IIC_BIN_MEM>, + Sched<[WriteALULd, ReadAfterLd]>; } let Predicates = [HasBMI], Defs = [EFLAGS] in { @@ -1211,6 +1275,17 @@ let Predicates = [HasBMI], Defs = [EFLAGS] in { defm ANDN64 : bmi_andn<"andn{q}", GR64, i64mem, loadi64>, T8, VEX_4V, VEX_W; } +let Predicates = [HasBMI] in { + def : Pat<(and (not GR32:$src1), GR32:$src2), + (ANDN32rr GR32:$src1, GR32:$src2)>; + def : Pat<(and (not GR64:$src1), GR64:$src2), + (ANDN64rr GR64:$src1, GR64:$src2)>; + def : Pat<(and (not GR32:$src1), (loadi32 addr:$src2)), + (ANDN32rm GR32:$src1, addr:$src2)>; + def : Pat<(and (not GR64:$src1), (loadi64 addr:$src2)), + (ANDN64rm GR64:$src1, addr:$src2)>; +} + //===----------------------------------------------------------------------===// // MULX Instruction // @@ -1219,12 +1294,12 @@ let neverHasSideEffects = 1 in { let isCommutable = 1 in def rr : I<0xF6, MRMSrcReg, (outs RC:$dst1, RC:$dst2), (ins RC:$src), !strconcat(mnemonic, "\t{$src, $dst2, $dst1|$dst1, $dst2, $src}"), - []>, T8XD, VEX_4V; + [], IIC_MUL8>, T8XD, VEX_4V, Sched<[WriteIMul]>; let mayLoad = 1 in def rm : I<0xF6, MRMSrcMem, (outs RC:$dst1, RC:$dst2), (ins x86memop:$src), !strconcat(mnemonic, "\t{$src, $dst2, $dst1|$dst1, $dst2, $src}"), - []>, T8XD, VEX_4V; + [], IIC_MUL8>, T8XD, VEX_4V, Sched<[WriteIMulLd]>; } } @@ -1234,3 +1309,53 @@ let Predicates = [HasBMI2] in { let Uses = [RDX] in defm MULX64 : bmi_mulx<"mulx{q}", GR64, i64mem>, VEX_W; } + +//===----------------------------------------------------------------------===// +// ADCX Instruction +// +let hasSideEffects = 0, Predicates = [HasADX], Defs = [EFLAGS] in { + let SchedRW = [WriteALU] in { + def ADCX32rr : I<0xF6, MRMSrcReg, (outs GR32:$dst), (ins GR32:$src), + "adcx{l}\t{$src, $dst|$dst, $src}", + [], IIC_BIN_NONMEM>, T8, OpSize; + + def ADCX64rr : I<0xF6, MRMSrcReg, (outs GR64:$dst), (ins GR64:$src), + "adcx{q}\t{$src, $dst|$dst, $src}", + [], IIC_BIN_NONMEM>, T8, OpSize, REX_W, Requires<[In64BitMode]>; + } // SchedRW + + let mayLoad = 1, SchedRW = [WriteALULd] in { + def ADCX32rm : I<0xF6, MRMSrcMem, (outs GR32:$dst), (ins i32mem:$src), + "adcx{l}\t{$src, $dst|$dst, $src}", + [], IIC_BIN_MEM>, T8, OpSize; + + def ADCX64rm : I<0xF6, MRMSrcMem, (outs GR64:$dst), (ins i64mem:$src), + "adcx{q}\t{$src, $dst|$dst, $src}", + [], IIC_BIN_MEM>, T8, OpSize, REX_W, Requires<[In64BitMode]>; + } +} + +//===----------------------------------------------------------------------===// +// ADOX Instruction +// +let hasSideEffects = 0, Predicates = [HasADX], Defs = [EFLAGS] in { + let SchedRW = [WriteALU] in { + def ADOX32rr : I<0xF6, MRMSrcReg, (outs GR32:$dst), (ins GR32:$src), + "adox{l}\t{$src, $dst|$dst, $src}", + [], IIC_BIN_NONMEM>, T8XS; + + def ADOX64rr : I<0xF6, MRMSrcReg, (outs GR64:$dst), (ins GR64:$src), + "adox{q}\t{$src, $dst|$dst, $src}", + [], IIC_BIN_NONMEM>, T8XS, REX_W, Requires<[In64BitMode]>; + } // SchedRW + + let mayLoad = 1, SchedRW = [WriteALULd] in { + def ADOX32rm : I<0xF6, MRMSrcMem, (outs GR32:$dst), (ins i32mem:$src), + "adox{l}\t{$src, $dst|$dst, $src}", + [], IIC_BIN_MEM>, T8XS; + + def ADOX64rm : I<0xF6, MRMSrcMem, (outs GR64:$dst), (ins i64mem:$src), + "adox{q}\t{$src, $dst|$dst, $src}", + [], IIC_BIN_MEM>, T8XS, REX_W, Requires<[In64BitMode]>; + } +}