From 9e18844d8a257561189ffda7ebdf3a21ba061fcd Mon Sep 17 00:00:00 2001 From: Tanner Gooding Date: Thu, 23 Aug 2018 16:18:33 -0700 Subject: [PATCH] Merging the instrsxarch fp, rf, and wf parameters into a single flags parameter --- src/jit/codegeninterface.h | 6 + src/jit/emitxarch.cpp | 103 ++-- src/jit/instr.cpp | 16 +- src/jit/instr.h | 35 +- src/jit/instrsxarch.h | 1166 ++++++++++++++++++++++---------------------- 5 files changed, 669 insertions(+), 657 deletions(-) diff --git a/src/jit/codegeninterface.h b/src/jit/codegeninterface.h index 72beee4..9c941bf 100644 --- a/src/jit/codegeninterface.h +++ b/src/jit/codegeninterface.h @@ -98,7 +98,13 @@ protected: bool m_genAlignLoops; private: +#if defined(_TARGET_XARCH_) + static const insFlags instInfo[INS_count]; +#elif defined(_TARGET_ARM_) || defined(_TARGET_ARM64_) static const BYTE instInfo[INS_count]; +#else +#error Unsupported target architecture +#endif #define INST_FP 0x01 // is it a FP instruction? public: diff --git a/src/jit/emitxarch.cpp b/src/jit/emitxarch.cpp index b879571..1b1af43 100644 --- a/src/jit/emitxarch.cpp +++ b/src/jit/emitxarch.cpp @@ -1196,18 +1196,15 @@ inline ssize_t emitter::emitGetInsCIdisp(instrDesc* id) * The following table is used by the instIsFP()/instUse/DefFlags() helpers. */ -#define INST_DEF_FL 0x20 // does the instruction set flags? -#define INST_USE_FL 0x40 // does the instruction use flags? - // clang-format off -const BYTE CodeGenInterface::instInfo[] = -{ - #define INST0(id, nm, fp, um, rf, wf, mr ) (INST_USE_FL*rf|INST_DEF_FL*wf|INST_FP*fp), - #define INST1(id, nm, fp, um, rf, wf, mr ) (INST_USE_FL*rf|INST_DEF_FL*wf|INST_FP*fp), - #define INST2(id, nm, fp, um, rf, wf, mr, mi ) (INST_USE_FL*rf|INST_DEF_FL*wf|INST_FP*fp), - #define INST3(id, nm, fp, um, rf, wf, mr, mi, rm ) (INST_USE_FL*rf|INST_DEF_FL*wf|INST_FP*fp), - #define INST4(id, nm, fp, um, rf, wf, mr, mi, rm, a4 ) (INST_USE_FL*rf|INST_DEF_FL*wf|INST_FP*fp), - #define INST5(id, nm, fp, um, rf, wf, mr, mi, rm, a4, rr ) (INST_USE_FL*rf|INST_DEF_FL*wf|INST_FP*fp), +const insFlags CodeGenInterface::instInfo[] = +{ + #define INST0(id, nm, um, mr, flags) static_cast(flags), + #define INST1(id, nm, um, mr, flags) static_cast(flags), + #define INST2(id, nm, um, mr, mi, flags) static_cast(flags), + #define INST3(id, nm, um, mr, mi, rm, flags) static_cast(flags), + #define INST4(id, nm, um, mr, mi, rm, a4, flags) static_cast(flags), + #define INST5(id, nm, um, mr, mi, rm, a4, rr, flags) static_cast(flags), #include "instrs.h" #undef INST0 #undef INST1 @@ -1226,12 +1223,12 @@ const BYTE CodeGenInterface::instInfo[] = // clang-format off const BYTE emitter::emitInsModeFmtTab[] = { - #define INST0(id, nm, fp, um, rf, wf, mr ) um, - #define INST1(id, nm, fp, um, rf, wf, mr ) um, - #define INST2(id, nm, fp, um, rf, wf, mr, mi ) um, - #define INST3(id, nm, fp, um, rf, wf, mr, mi, rm ) um, - #define INST4(id, nm, fp, um, rf, wf, mr, mi, rm, a4 ) um, - #define INST5(id, nm, fp, um, rf, wf, mr, mi, rm, a4, rr) um, + #define INST0(id, nm, um, mr, flags) um, + #define INST1(id, nm, um, mr, flags) um, + #define INST2(id, nm, um, mr, mi, flags) um, + #define INST3(id, nm, um, mr, mi, rm, flags) um, + #define INST4(id, nm, um, mr, mi, rm, a4, flags) um, + #define INST5(id, nm, um, mr, mi, rm, a4, rr, flags) um, #include "instrs.h" #undef INST0 #undef INST1 @@ -1317,12 +1314,12 @@ inline size_t insCode(instruction ins) const static size_t insCodes[] = { - #define INST0(id, nm, fp, um, rf, wf, mr ) mr, - #define INST1(id, nm, fp, um, rf, wf, mr ) mr, - #define INST2(id, nm, fp, um, rf, wf, mr, mi ) mr, - #define INST3(id, nm, fp, um, rf, wf, mr, mi, rm ) mr, - #define INST4(id, nm, fp, um, rf, wf, mr, mi, rm, a4 ) mr, - #define INST5(id, nm, fp, um, rf, wf, mr, mi, rm, a4, rr) mr, + #define INST0(id, nm, um, mr, flags) mr, + #define INST1(id, nm, um, mr, flags) mr, + #define INST2(id, nm, um, mr, mi, flags) mr, + #define INST3(id, nm, um, mr, mi, rm, flags) mr, + #define INST4(id, nm, um, mr, mi, rm, a4, flags) mr, + #define INST5(id, nm, um, mr, mi, rm, a4, rr, flags) mr, #include "instrs.h" #undef INST0 #undef INST1 @@ -1350,12 +1347,12 @@ inline size_t insCodeACC(instruction ins) const static size_t insCodesACC[] = { - #define INST0(id, nm, fp, um, rf, wf, mr ) - #define INST1(id, nm, fp, um, rf, wf, mr ) - #define INST2(id, nm, fp, um, rf, wf, mr, mi ) - #define INST3(id, nm, fp, um, rf, wf, mr, mi, rm ) - #define INST4(id, nm, fp, um, rf, wf, mr, mi, rm, a4 ) a4, - #define INST5(id, nm, fp, um, rf, wf, mr, mi, rm, a4, rr) a4, + #define INST0(id, nm, um, mr, flags) + #define INST1(id, nm, um, mr, flags) + #define INST2(id, nm, um, mr, mi, flags) + #define INST3(id, nm, um, mr, mi, rm, flags) + #define INST4(id, nm, um, mr, mi, rm, a4, flags) a4, + #define INST5(id, nm, um, mr, mi, rm, a4, rr, flags) a4, #include "instrs.h" #undef INST0 #undef INST1 @@ -1383,12 +1380,12 @@ inline size_t insCodeRR(instruction ins) const static size_t insCodesRR[] = { - #define INST0(id, nm, fp, um, rf, wf, mr ) - #define INST1(id, nm, fp, um, rf, wf, mr ) - #define INST2(id, nm, fp, um, rf, wf, mr, mi ) - #define INST3(id, nm, fp, um, rf, wf, mr, mi, rm ) - #define INST4(id, nm, fp, um, rf, wf, mr, mi, rm, a4 ) - #define INST5(id, nm, fp, um, rf, wf, mr, mi, rm, a4, rr) rr, + #define INST0(id, nm, um, mr, flags) + #define INST1(id, nm, um, mr, flags) + #define INST2(id, nm, um, mr, mi, flags) + #define INST3(id, nm, um, mr, mi, rm, flags) + #define INST4(id, nm, um, mr, mi, rm, a4, flags) + #define INST5(id, nm, um, mr, mi, rm, a4, rr, flags) rr, #include "instrs.h" #undef INST0 #undef INST1 @@ -1409,12 +1406,12 @@ inline size_t insCodeRR(instruction ins) const static size_t insCodesRM[] = { - #define INST0(id, nm, fp, um, rf, wf, mr ) - #define INST1(id, nm, fp, um, rf, wf, mr ) - #define INST2(id, nm, fp, um, rf, wf, mr, mi ) - #define INST3(id, nm, fp, um, rf, wf, mr, mi, rm ) rm, - #define INST4(id, nm, fp, um, rf, wf, mr, mi, rm, a4 ) rm, - #define INST5(id, nm, fp, um, rf, wf, mr, mi, rm, a4, rr) rm, + #define INST0(id, nm, um, mr, flags) + #define INST1(id, nm, um, mr, flags) + #define INST2(id, nm, um, mr, mi, flags) + #define INST3(id, nm, um, mr, mi, rm, flags) rm, + #define INST4(id, nm, um, mr, mi, rm, a4, flags) rm, + #define INST5(id, nm, um, mr, mi, rm, a4, rr, flags) rm, #include "instrs.h" #undef INST0 #undef INST1 @@ -1449,12 +1446,12 @@ inline size_t insCodeRM(instruction ins) const static size_t insCodesMI[] = { - #define INST0(id, nm, fp, um, rf, wf, mr ) - #define INST1(id, nm, fp, um, rf, wf, mr ) - #define INST2(id, nm, fp, um, rf, wf, mr, mi ) mi, - #define INST3(id, nm, fp, um, rf, wf, mr, mi, rm ) mi, - #define INST4(id, nm, fp, um, rf, wf, mr, mi, rm, a4 ) mi, - #define INST5(id, nm, fp, um, rf, wf, mr, mi, rm, a4, rr) mi, + #define INST0(id, nm, um, mr, flags) + #define INST1(id, nm, um, mr, flags) + #define INST2(id, nm, um, mr, mi, flags) mi, + #define INST3(id, nm, um, mr, mi, rm, flags) mi, + #define INST4(id, nm, um, mr, mi, rm, a4, flags) mi, + #define INST5(id, nm, um, mr, mi, rm, a4, rr, flags) mi, #include "instrs.h" #undef INST0 #undef INST1 @@ -1489,12 +1486,12 @@ inline size_t insCodeMI(instruction ins) const static size_t insCodesMR[] = { - #define INST0(id, nm, fp, um, rf, wf, mr ) - #define INST1(id, nm, fp, um, rf, wf, mr ) mr, - #define INST2(id, nm, fp, um, rf, wf, mr, mi ) mr, - #define INST3(id, nm, fp, um, rf, wf, mr, mi, rm ) mr, - #define INST4(id, nm, fp, um, rf, wf, mr, mi, rm, a4 ) mr, - #define INST5(id, nm, fp, um, rf, wf, mr, mi, rm, a4, rr) mr, + #define INST0(id, nm, um, mr, flags) + #define INST1(id, nm, um, mr, flags) mr, + #define INST2(id, nm, um, mr, mi, flags) mr, + #define INST3(id, nm, um, mr, mi, rm, flags) mr, + #define INST4(id, nm, um, mr, mi, rm, a4, flags) mr, + #define INST5(id, nm, um, mr, mi, rm, a4, rr, flags) mr, #include "instrs.h" #undef INST0 #undef INST1 diff --git a/src/jit/instr.cpp b/src/jit/instr.cpp index 763a403..1724f82 100644 --- a/src/jit/instr.cpp +++ b/src/jit/instr.cpp @@ -37,12 +37,12 @@ const char* CodeGen::genInsName(instruction ins) const char * const insNames[] = { #if defined(_TARGET_XARCH_) - #define INST0(id, nm, fp, um, rf, wf, mr ) nm, - #define INST1(id, nm, fp, um, rf, wf, mr ) nm, - #define INST2(id, nm, fp, um, rf, wf, mr, mi ) nm, - #define INST3(id, nm, fp, um, rf, wf, mr, mi, rm ) nm, - #define INST4(id, nm, fp, um, rf, wf, mr, mi, rm, a4 ) nm, - #define INST5(id, nm, fp, um, rf, wf, mr, mi, rm, a4, rr ) nm, + #define INST0(id, nm, um, mr, flags) nm, + #define INST1(id, nm, um, mr, flags) nm, + #define INST2(id, nm, um, mr, mi, flags) nm, + #define INST3(id, nm, um, mr, mi, rm, flags) nm, + #define INST4(id, nm, um, mr, mi, rm, a4, flags) nm, + #define INST5(id, nm, um, mr, mi, rm, a4, rr, flags) nm, #include "instrs.h" #elif defined(_TARGET_ARM_) @@ -215,7 +215,11 @@ bool CodeGenInterface::instIsFP(instruction ins) { assert((unsigned)ins < _countof(instInfo)); +#ifdef _TARGET_XARCH_ + return (instInfo[ins] & INS_FLAGS_x87Instr) != 0; +#else return (instInfo[ins] & INST_FP) != 0; +#endif } #ifdef _TARGET_XARCH_ diff --git a/src/jit/instr.h b/src/jit/instr.h index efa9412..c00a61c 100644 --- a/src/jit/instr.h +++ b/src/jit/instr.h @@ -15,12 +15,12 @@ enum instruction : unsigned { #if defined(_TARGET_XARCH_) - #define INST0(id, nm, fp, um, rf, wf, mr ) INS_##id, - #define INST1(id, nm, fp, um, rf, wf, mr ) INS_##id, - #define INST2(id, nm, fp, um, rf, wf, mr, mi ) INS_##id, - #define INST3(id, nm, fp, um, rf, wf, mr, mi, rm ) INS_##id, - #define INST4(id, nm, fp, um, rf, wf, mr, mi, rm, a4 ) INS_##id, - #define INST5(id, nm, fp, um, rf, wf, mr, mi, rm, a4, rr) INS_##id, + #define INST0(id, nm, um, mr, flags) INS_##id, + #define INST1(id, nm, um, mr, flags) INS_##id, + #define INST2(id, nm, um, mr, mi, flags) INS_##id, + #define INST3(id, nm, um, mr, mi, rm, flags) INS_##id, + #define INST4(id, nm, um, mr, mi, rm, a4, flags) INS_##id, + #define INST5(id, nm, um, mr, mi, rm, a4, rr, flags) INS_##id, #include "instrs.h" #elif defined(_TARGET_ARM_) @@ -86,13 +86,28 @@ enum GCtype : unsigned GCT_BYREF }; -// TODO-Cleanup: Move 'insFlags' under _TARGET_ARM_ +#if defined(_TARGET_XARCH_) +enum insFlags: uint8_t +{ + INS_FLAGS_None = 0x00, + INS_FLAGS_ReadsFlags = 0x01, + INS_FLAGS_WritesFlags = 0x02, + INS_FLAGS_x87Instr = 0x04, + + // TODO-Cleanup: Remove this flag and its usage from _TARGET_XARCH_ + INS_FLAGS_DONT_CARE = 0x00, +}; +#elif defined(_TARGET_ARM_) || defined(_TARGET_ARM64_) +// TODO-Cleanup: Move 'insFlags' under _TARGET_ARM_ enum insFlags: unsigned { - INS_FLAGS_NOT_SET, - INS_FLAGS_SET, - INS_FLAGS_DONT_CARE + INS_FLAGS_NOT_SET = 0x00, + INS_FLAGS_SET = 0x01, + INS_FLAGS_DONT_CARE = 0x02, }; +#else +#error Unsupported target architecture +#endif #if defined(_TARGET_ARM_) enum insOpts: unsigned diff --git a/src/jit/instrsxarch.h b/src/jit/instrsxarch.h index e46134e..d4278f7 100644 --- a/src/jit/instrsxarch.h +++ b/src/jit/instrsxarch.h @@ -10,15 +10,13 @@ * * id -- the enum name for the instruction * nm -- textual name (for assembly dipslay) - * fp -- 1 = floating point instruction, 0 = not floating point instruction * um -- update mode, see IUM_xx enum (rd, wr, or rw) - * rf -- 1 = reads flags, 0 = doesn't read flags - * wf -- 1 = writes flags, 0 = doesn't write flags * mr -- base encoding for R/M[reg] addressing mode * mi -- base encoding for R/M,icon addressing mode * rm -- base encoding for reg,R/M addressing mode * a4 -- base encoding for eax,i32 addressing mode * rr -- base encoding for register addressing mode + * flags -- flags, see INS_FLAGS_* enum * ******************************************************************************/ @@ -32,111 +30,110 @@ #endif /*****************************************************************************/ #ifndef INST0 -#define INST0(id, nm, fp, um, rf, wf, mr ) +#define INST0(id, nm, um, mr, flags) #endif #ifndef INST2 -#define INST2(id, nm, fp, um, rf, wf, mr, mi ) +#define INST2(id, nm, um, mr, mi, flags) #endif #ifndef INST3 -#define INST3(id, nm, fp, um, rf, wf, mr, mi, rm ) +#define INST3(id, nm, um, mr, mi, rm, flags) #endif #ifndef INST4 -#define INST4(id, nm, fp, um, rf, wf, mr, mi, rm, a4 ) +#define INST4(id, nm, um, mr, mi, rm, a4, flags) #endif #ifndef INST5 -#define INST5(id, nm, fp, um, rf, wf, mr, mi, rm, a4, rr) +#define INST5(id, nm, um, mr, mi, rm, a4, rr, flags) #endif /*****************************************************************************/ /* The following is x86-specific */ /*****************************************************************************/ -// enum name FP updmode rf wf R/M[reg] R/M,icon reg,R/M eax,i32 register -INST5(invalid, "INVALID" , 0, IUM_RD, 0, 0, BAD_CODE, BAD_CODE, BAD_CODE, BAD_CODE, BAD_CODE) +// id nm um mr mi rm a4 rr flags +INST5(invalid, "INVALID", IUM_RD, BAD_CODE, BAD_CODE, BAD_CODE, BAD_CODE, BAD_CODE, INS_FLAGS_None) -INST5(push , "push" , 0, IUM_RD, 0, 0, 0x0030FE, 0x000068, BAD_CODE, BAD_CODE, 0x000050) -INST5(pop , "pop" , 0, IUM_WR, 0, 0, 0x00008E, BAD_CODE, BAD_CODE, BAD_CODE, 0x000058) +INST5(push, "push", IUM_RD, 0x0030FE, 0x000068, BAD_CODE, BAD_CODE, 0x000050, INS_FLAGS_None) +INST5(pop, "pop", IUM_WR, 0x00008E, BAD_CODE, BAD_CODE, BAD_CODE, 0x000058, INS_FLAGS_None) // Does not affect the stack tracking in the emitter -INST5(push_hide, "push" , 0, IUM_RD, 0, 0, 0x0030FE, 0x000068, BAD_CODE, BAD_CODE, 0x000050) -INST5(pop_hide, "pop" , 0, IUM_WR, 0, 0, 0x00008E, BAD_CODE, BAD_CODE, BAD_CODE, 0x000058) - -INST5(inc , "inc" , 0, IUM_RW, 0, 1, 0x0000FE, BAD_CODE, BAD_CODE, BAD_CODE, 0x000040) -INST5(inc_l , "inc" , 0, IUM_RW, 0, 1, 0x0000FE, BAD_CODE, BAD_CODE, BAD_CODE, 0x00C0FE) -INST5(dec , "dec" , 0, IUM_RW, 0, 1, 0x0008FE, BAD_CODE, BAD_CODE, BAD_CODE, 0x000048) -INST5(dec_l , "dec" , 0, IUM_RW, 0, 1, 0x0008FE, BAD_CODE, BAD_CODE, BAD_CODE, 0x00C8FE) - -// enum name FP updmode rf wf R/M,R/M[reg] R/M,icon reg,R/M eax,i32 - -INST4(add , "add" , 0, IUM_RW, 0, 1, 0x000000, 0x000080, 0x000002, 0x000004) -INST4(or , "or" , 0, IUM_RW, 0, 1, 0x000008, 0x000880, 0x00000A, 0x00000C) -INST4(adc , "adc" , 0, IUM_RW, 1, 1, 0x000010, 0x001080, 0x000012, 0x000014) -INST4(sbb , "sbb" , 0, IUM_RW, 1, 1, 0x000018, 0x001880, 0x00001A, 0x00001C) -INST4(and , "and" , 0, IUM_RW, 0, 1, 0x000020, 0x002080, 0x000022, 0x000024) -INST4(sub , "sub" , 0, IUM_RW, 0, 1, 0x000028, 0x002880, 0x00002A, 0x00002C) -INST4(xor , "xor" , 0, IUM_RW, 0, 1, 0x000030, 0x003080, 0x000032, 0x000034) -INST4(cmp , "cmp" , 0, IUM_RD, 0, 1, 0x000038, 0x003880, 0x00003A, 0x00003C) -INST4(test , "test" , 0, IUM_RD, 0, 1, 0x000084, 0x0000F6, 0x000084, 0x0000A8) -INST4(mov , "mov" , 0, IUM_WR, 0, 0, 0x000088, 0x0000C6, 0x00008A, 0x0000B0) - -INST4(lea , "lea" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, 0x00008D, BAD_CODE) - -// enum name FP updmode rf wf R/M,R/M[reg] R/M,icon reg,R/M +INST5(push_hide, "push", IUM_RD, 0x0030FE, 0x000068, BAD_CODE, BAD_CODE, 0x000050, INS_FLAGS_None) +INST5(pop_hide, "pop", IUM_WR, 0x00008E, BAD_CODE, BAD_CODE, BAD_CODE, 0x000058, INS_FLAGS_None) + +INST5(inc, "inc", IUM_RW, 0x0000FE, BAD_CODE, BAD_CODE, BAD_CODE, 0x000040, INS_FLAGS_WritesFlags) +INST5(inc_l, "inc", IUM_RW, 0x0000FE, BAD_CODE, BAD_CODE, BAD_CODE, 0x00C0FE, INS_FLAGS_WritesFlags) +INST5(dec, "dec", IUM_RW, 0x0008FE, BAD_CODE, BAD_CODE, BAD_CODE, 0x000048, INS_FLAGS_WritesFlags) +INST5(dec_l, "dec", IUM_RW, 0x0008FE, BAD_CODE, BAD_CODE, BAD_CODE, 0x00C8FE, INS_FLAGS_WritesFlags) + +// id nm um mr mi rm a4 flags +INST4(add, "add", IUM_RW, 0x000000, 0x000080, 0x000002, 0x000004, INS_FLAGS_WritesFlags) +INST4(or, "or", IUM_RW, 0x000008, 0x000880, 0x00000A, 0x00000C, INS_FLAGS_WritesFlags) +INST4(adc, "adc", IUM_RW, 0x000010, 0x001080, 0x000012, 0x000014, INS_FLAGS_ReadsFlags | INS_FLAGS_WritesFlags) +INST4(sbb, "sbb", IUM_RW, 0x000018, 0x001880, 0x00001A, 0x00001C, INS_FLAGS_ReadsFlags | INS_FLAGS_WritesFlags) +INST4(and, "and", IUM_RW, 0x000020, 0x002080, 0x000022, 0x000024, INS_FLAGS_WritesFlags) +INST4(sub, "sub", IUM_RW, 0x000028, 0x002880, 0x00002A, 0x00002C, INS_FLAGS_WritesFlags) +INST4(xor, "xor", IUM_RW, 0x000030, 0x003080, 0x000032, 0x000034, INS_FLAGS_WritesFlags) +INST4(cmp, "cmp", IUM_RD, 0x000038, 0x003880, 0x00003A, 0x00003C, INS_FLAGS_WritesFlags) +INST4(test, "test", IUM_RD, 0x000084, 0x0000F6, 0x000084, 0x0000A8, INS_FLAGS_WritesFlags) +INST4(mov, "mov", IUM_WR, 0x000088, 0x0000C6, 0x00008A, 0x0000B0, INS_FLAGS_None) + +INST4(lea, "lea", IUM_WR, BAD_CODE, BAD_CODE, 0x00008D, BAD_CODE, INS_FLAGS_None) + +// id nm um mr mi rm flags // Note that emitter has only partial support for BT. It can only emit the reg,reg form // and the registers need to be reversed to get the correct encoding. -INST3(bt , "bt" , 0, IUM_RD, 0, 1, 0x0F00A3, BAD_CODE, 0x0F00A3) +INST3(bt, "bt", IUM_RD, 0x0F00A3, BAD_CODE, 0x0F00A3, INS_FLAGS_WritesFlags) -INST3(movsx , "movsx" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, 0x0F00BE) +INST3(movsx, "movsx", IUM_WR, BAD_CODE, BAD_CODE, 0x0F00BE, INS_FLAGS_None) #ifdef _TARGET_AMD64_ -INST3(movsxd , "movsxd" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, 0x4800000063LL ) +INST3(movsxd, "movsxd", IUM_WR, BAD_CODE, BAD_CODE, 0x4800000063, INS_FLAGS_None) #endif -INST3(movzx , "movzx" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, 0x0F00B6) - -INST3(cmovo , "cmovo" , 0, IUM_WR, 1, 0, BAD_CODE, BAD_CODE, 0x0F0040) -INST3(cmovno , "cmovno" , 0, IUM_WR, 1, 0, BAD_CODE, BAD_CODE, 0x0F0041) -INST3(cmovb , "cmovb" , 0, IUM_WR, 1, 0, BAD_CODE, BAD_CODE, 0x0F0042) -INST3(cmovae , "cmovae" , 0, IUM_WR, 1, 0, BAD_CODE, BAD_CODE, 0x0F0043) -INST3(cmove , "cmove" , 0, IUM_WR, 1, 0, BAD_CODE, BAD_CODE, 0x0F0044) -INST3(cmovne , "cmovne" , 0, IUM_WR, 1, 0, BAD_CODE, BAD_CODE, 0x0F0045) -INST3(cmovbe , "cmovbe" , 0, IUM_WR, 1, 0, BAD_CODE, BAD_CODE, 0x0F0046) -INST3(cmova , "cmova" , 0, IUM_WR, 1, 0, BAD_CODE, BAD_CODE, 0x0F0047) -INST3(cmovs , "cmovs" , 0, IUM_WR, 1, 0, BAD_CODE, BAD_CODE, 0x0F0048) -INST3(cmovns , "cmovns" , 0, IUM_WR, 1, 0, BAD_CODE, BAD_CODE, 0x0F0049) -INST3(cmovpe , "cmovpe" , 0, IUM_WR, 1, 0, BAD_CODE, BAD_CODE, 0x0F004A) -INST3(cmovpo , "cmovpo" , 0, IUM_WR, 1, 0, BAD_CODE, BAD_CODE, 0x0F004B) -INST3(cmovl , "cmovl" , 0, IUM_WR, 1, 0, BAD_CODE, BAD_CODE, 0x0F004C) -INST3(cmovge , "cmovge" , 0, IUM_WR, 1, 0, BAD_CODE, BAD_CODE, 0x0F004D) -INST3(cmovle , "cmovle" , 0, IUM_WR, 1, 0, BAD_CODE, BAD_CODE, 0x0F004E) -INST3(cmovg , "cmovg" , 0, IUM_WR, 1, 0, BAD_CODE, BAD_CODE, 0x0F004F) - -INST3(xchg , "xchg" , 0, IUM_RW, 0, 0, 0x000086, BAD_CODE, 0x000086) -INST3(imul , "imul" , 0, IUM_RW, 0, 1, 0x0F00AC, BAD_CODE, 0x0F00AF) // op1 *= op2 - -// enum name FP updmode rf wf R/M,R/M[reg] R/M,icon reg,R/M +INST3(movzx, "movzx", IUM_WR, BAD_CODE, BAD_CODE, 0x0F00B6, INS_FLAGS_None) + +INST3(cmovo, "cmovo", IUM_WR, BAD_CODE, BAD_CODE, 0x0F0040, INS_FLAGS_ReadsFlags) +INST3(cmovno, "cmovno", IUM_WR, BAD_CODE, BAD_CODE, 0x0F0041, INS_FLAGS_ReadsFlags) +INST3(cmovb, "cmovb", IUM_WR, BAD_CODE, BAD_CODE, 0x0F0042, INS_FLAGS_ReadsFlags) +INST3(cmovae, "cmovae", IUM_WR, BAD_CODE, BAD_CODE, 0x0F0043, INS_FLAGS_ReadsFlags) +INST3(cmove, "cmove", IUM_WR, BAD_CODE, BAD_CODE, 0x0F0044, INS_FLAGS_ReadsFlags) +INST3(cmovne, "cmovne", IUM_WR, BAD_CODE, BAD_CODE, 0x0F0045, INS_FLAGS_ReadsFlags) +INST3(cmovbe, "cmovbe", IUM_WR, BAD_CODE, BAD_CODE, 0x0F0046, INS_FLAGS_ReadsFlags) +INST3(cmova, "cmova", IUM_WR, BAD_CODE, BAD_CODE, 0x0F0047, INS_FLAGS_ReadsFlags) +INST3(cmovs, "cmovs", IUM_WR, BAD_CODE, BAD_CODE, 0x0F0048, INS_FLAGS_ReadsFlags) +INST3(cmovns, "cmovns", IUM_WR, BAD_CODE, BAD_CODE, 0x0F0049, INS_FLAGS_ReadsFlags) +INST3(cmovpe, "cmovpe", IUM_WR, BAD_CODE, BAD_CODE, 0x0F004A, INS_FLAGS_ReadsFlags) +INST3(cmovpo, "cmovpo", IUM_WR, BAD_CODE, BAD_CODE, 0x0F004B, INS_FLAGS_ReadsFlags) +INST3(cmovl, "cmovl", IUM_WR, BAD_CODE, BAD_CODE, 0x0F004C, INS_FLAGS_ReadsFlags) +INST3(cmovge, "cmovge", IUM_WR, BAD_CODE, BAD_CODE, 0x0F004D, INS_FLAGS_ReadsFlags) +INST3(cmovle, "cmovle", IUM_WR, BAD_CODE, BAD_CODE, 0x0F004E, INS_FLAGS_ReadsFlags) +INST3(cmovg, "cmovg", IUM_WR, BAD_CODE, BAD_CODE, 0x0F004F, INS_FLAGS_ReadsFlags) + +INST3(xchg, "xchg", IUM_RW, 0x000086, BAD_CODE, 0x000086, INS_FLAGS_None) +INST3(imul, "imul", IUM_RW, 0x0F00AC, BAD_CODE, 0x0F00AF, INS_FLAGS_WritesFlags) // op1 *= op2 + +// id nm um mr mi rm flags // Instead of encoding these as 3-operand instructions, we encode them // as 2-operand instructions with the target register being implicit // implicit_reg = op1*op2_icon #define INSTMUL INST3 -INSTMUL(imul_AX, "imul", 0, IUM_RD, 0, 1, BAD_CODE, 0x000068, BAD_CODE) -INSTMUL(imul_CX, "imul", 0, IUM_RD, 0, 1, BAD_CODE, 0x000868, BAD_CODE) -INSTMUL(imul_DX, "imul", 0, IUM_RD, 0, 1, BAD_CODE, 0x001068, BAD_CODE) -INSTMUL(imul_BX, "imul", 0, IUM_RD, 0, 1, BAD_CODE, 0x001868, BAD_CODE) -INSTMUL(imul_SP, "imul", 0, IUM_RD, 0, 1, BAD_CODE, BAD_CODE, BAD_CODE) -INSTMUL(imul_BP, "imul", 0, IUM_RD, 0, 1, BAD_CODE, 0x002868, BAD_CODE) -INSTMUL(imul_SI, "imul", 0, IUM_RD, 0, 1, BAD_CODE, 0x003068, BAD_CODE) -INSTMUL(imul_DI, "imul", 0, IUM_RD, 0, 1, BAD_CODE, 0x003868, BAD_CODE) +INSTMUL(imul_AX, "imul", IUM_RD, BAD_CODE, 0x000068, BAD_CODE, INS_FLAGS_WritesFlags) +INSTMUL(imul_CX, "imul", IUM_RD, BAD_CODE, 0x000868, BAD_CODE, INS_FLAGS_WritesFlags) +INSTMUL(imul_DX, "imul", IUM_RD, BAD_CODE, 0x001068, BAD_CODE, INS_FLAGS_WritesFlags) +INSTMUL(imul_BX, "imul", IUM_RD, BAD_CODE, 0x001868, BAD_CODE, INS_FLAGS_WritesFlags) +INSTMUL(imul_SP, "imul", IUM_RD, BAD_CODE, BAD_CODE, BAD_CODE, INS_FLAGS_WritesFlags) +INSTMUL(imul_BP, "imul", IUM_RD, BAD_CODE, 0x002868, BAD_CODE, INS_FLAGS_WritesFlags) +INSTMUL(imul_SI, "imul", IUM_RD, BAD_CODE, 0x003068, BAD_CODE, INS_FLAGS_WritesFlags) +INSTMUL(imul_DI, "imul", IUM_RD, BAD_CODE, 0x003868, BAD_CODE, INS_FLAGS_WritesFlags) #ifdef _TARGET_AMD64_ -INSTMUL(imul_08, "imul", 0, IUM_RD, 0, 1, BAD_CODE, 0x4400000068, BAD_CODE) -INSTMUL(imul_09, "imul", 0, IUM_RD, 0, 1, BAD_CODE, 0x4400000868, BAD_CODE) -INSTMUL(imul_10, "imul", 0, IUM_RD, 0, 1, BAD_CODE, 0x4400001068, BAD_CODE) -INSTMUL(imul_11, "imul", 0, IUM_RD, 0, 1, BAD_CODE, 0x4400001868, BAD_CODE) -INSTMUL(imul_12, "imul", 0, IUM_RD, 0, 1, BAD_CODE, 0x4400002068, BAD_CODE) -INSTMUL(imul_13, "imul", 0, IUM_RD, 0, 1, BAD_CODE, 0x4400002868, BAD_CODE) -INSTMUL(imul_14, "imul", 0, IUM_RD, 0, 1, BAD_CODE, 0x4400003068, BAD_CODE) -INSTMUL(imul_15, "imul", 0, IUM_RD, 0, 1, BAD_CODE, 0x4400003868, BAD_CODE) +INSTMUL(imul_08, "imul", IUM_RD, BAD_CODE, 0x4400000068, BAD_CODE, INS_FLAGS_WritesFlags) +INSTMUL(imul_09, "imul", IUM_RD, BAD_CODE, 0x4400000868, BAD_CODE, INS_FLAGS_WritesFlags) +INSTMUL(imul_10, "imul", IUM_RD, BAD_CODE, 0x4400001068, BAD_CODE, INS_FLAGS_WritesFlags) +INSTMUL(imul_11, "imul", IUM_RD, BAD_CODE, 0x4400001868, BAD_CODE, INS_FLAGS_WritesFlags) +INSTMUL(imul_12, "imul", IUM_RD, BAD_CODE, 0x4400002068, BAD_CODE, INS_FLAGS_WritesFlags) +INSTMUL(imul_13, "imul", IUM_RD, BAD_CODE, 0x4400002868, BAD_CODE, INS_FLAGS_WritesFlags) +INSTMUL(imul_14, "imul", IUM_RD, BAD_CODE, 0x4400003068, BAD_CODE, INS_FLAGS_WritesFlags) +INSTMUL(imul_15, "imul", IUM_RD, BAD_CODE, 0x4400003868, BAD_CODE, INS_FLAGS_WritesFlags) #endif // _TARGET_AMD64_ @@ -176,578 +173,571 @@ INSTMUL(imul_15, "imul", 0, IUM_RD, 0, 1, BAD_CODE, 0x4400003868, BAD_CODE) #define VEX3INT(c1,c2) PACK4(c1, 0xc5, 0x02, c2) #define VEX3FLT(c1,c2) PACK4(c1, 0xc5, 0x02, c2) -// Please insert any SSE2 instructions between FIRST_SSE2_INSTRUCTION and LAST_SSE2_INSTRUCTION -INST3(FIRST_SSE2_INSTRUCTION, "FIRST_SSE2_INSTRUCTION", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, BAD_CODE) - +INST3(FIRST_SSE2_INSTRUCTION, "FIRST_SSE2_INSTRUCTION", IUM_WR, BAD_CODE, BAD_CODE, BAD_CODE, INS_FLAGS_None) // These are the SSE instructions used on x86 -INST3( mov_i2xmm, "movd" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0x6E)) // Move int reg to a xmm reg. reg1=xmm reg, reg2=int reg -INST3( mov_xmm2i, "movd" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0x7E)) // Move xmm reg to an int reg. reg1=xmm reg, reg2=int reg -INST3( pmovmskb, "pmovmskb" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0xD7)) // Move the MSB bits of all bytes in a xmm reg to an int reg -INST3( movmskpd, "movmskpd" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0x50)) // Extract 2-bit sign mask from xmm and store in reg. The upper bits of r32 or r64 are filled with zeros. -INST3( movd, "movd" , 0, IUM_WR, 0, 0, PCKDBL(0x7E), BAD_CODE, PCKDBL(0x6E)) -INST3( movq, "movq" , 0, IUM_WR, 0, 0, PCKDBL(0xD6), BAD_CODE, SSEFLT(0x7E)) -INST3( movsdsse2, "movsd" , 0, IUM_WR, 0, 0, SSEDBL(0x11), BAD_CODE, SSEDBL(0x10)) - -INST3( punpckldq, "punpckldq" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0x62)) - -INST3( xorps, "xorps" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKFLT(0x57)) // XOR packed singles - -INST3( cvttsd2si, "cvttsd2si" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSEDBL(0x2C)) // cvt with trunc scalar double to signed DWORDs - -INST3( movntdq, "movntdq" , 0, IUM_WR, 0, 0, PCKDBL(0xE7), BAD_CODE, BAD_CODE) -INST3( movnti, "movnti" , 0, IUM_WR, 0, 0, PCKFLT(0xC3), BAD_CODE, BAD_CODE) -INST3( movntpd, "movntpd" , 0, IUM_WR, 0, 0, PCKDBL(0x2B), BAD_CODE, BAD_CODE) -INST3( movntps, "movntps" , 0, IUM_WR, 0, 0, PCKFLT(0x2B), BAD_CODE, BAD_CODE) -INST3( movdqu, "movdqu" , 0, IUM_WR, 0, 0, SSEFLT(0x7F), BAD_CODE, SSEFLT(0x6F)) -INST3( movdqa, "movdqa" , 0, IUM_WR, 0, 0, PCKDBL(0x7F), BAD_CODE, PCKDBL(0x6F)) -INST3( movlpd, "movlpd" , 0, IUM_WR, 0, 0, PCKDBL(0x13), BAD_CODE, PCKDBL(0x12)) -INST3( movlps, "movlps" , 0, IUM_WR, 0, 0, PCKFLT(0x13), BAD_CODE, PCKFLT(0x12)) -INST3( movhpd, "movhpd" , 0, IUM_WR, 0, 0, PCKDBL(0x17), BAD_CODE, PCKDBL(0x16)) -INST3( movhps, "movhps" , 0, IUM_WR, 0, 0, PCKFLT(0x17), BAD_CODE, PCKFLT(0x16)) -INST3( movss, "movss" , 0, IUM_WR, 0, 0, SSEFLT(0x11), BAD_CODE, SSEFLT(0x10)) -INST3( movapd, "movapd" , 0, IUM_WR, 0, 0, PCKDBL(0x29), BAD_CODE, PCKDBL(0x28)) -INST3( movaps, "movaps" , 0, IUM_WR, 0, 0, PCKFLT(0x29), BAD_CODE, PCKFLT(0x28)) -INST3( movupd, "movupd" , 0, IUM_WR, 0, 0, PCKDBL(0x11), BAD_CODE, PCKDBL(0x10)) -INST3( movups, "movups" , 0, IUM_WR, 0, 0, PCKFLT(0x11), BAD_CODE, PCKFLT(0x10)) -INST3( movhlps, "movhlps" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKFLT(0x12)) -INST3( movlhps, "movlhps" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKFLT(0x16)) -INST3( movmskps, "movmskps" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKFLT(0x50)) -INST3( unpckhps, "unpckhps" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKFLT(0x15)) -INST3( unpcklps, "unpcklps" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKFLT(0x14)) -INST3( maskmovdqu, "maskmovdqu" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0xF7)) - -INST3( shufps, "shufps" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKFLT(0xC6)) -INST3( shufpd, "shufpd" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0xC6)) - -INST3( punpckhdq, "punpckhdq" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0x6A)) - -INST3( lfence, "lfence" , 0, IUM_RD, 0, 0, 0x000FE8AE, BAD_CODE, BAD_CODE) -INST3( mfence, "mfence" , 0, IUM_RD, 0, 0, 0x000FF0AE, BAD_CODE, BAD_CODE) -INST3( prefetchnta, "prefetchnta" , 0, IUM_RD, 0, 0, 0x000F0018, BAD_CODE, BAD_CODE) -INST3( prefetcht0, "prefetcht0" , 0, IUM_RD, 0, 0, 0x000F0818, BAD_CODE, BAD_CODE) -INST3( prefetcht1, "prefetcht1" , 0, IUM_RD, 0, 0, 0x000F1018, BAD_CODE, BAD_CODE) -INST3( prefetcht2, "prefetcht2" , 0, IUM_RD, 0, 0, 0x000F1818, BAD_CODE, BAD_CODE) -INST3( sfence, "sfence" , 0, IUM_RD, 0, 0, 0x000FF8AE, BAD_CODE, BAD_CODE) +INST3(mov_i2xmm, "movd", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0x6E), INS_FLAGS_None) // Move int reg to a xmm reg. reg1=xmm reg, reg2=int reg +INST3(mov_xmm2i, "movd", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0x7E), INS_FLAGS_None) // Move xmm reg to an int reg. reg1=xmm reg, reg2=int reg +INST3(pmovmskb, "pmovmskb", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0xD7), INS_FLAGS_None) // Move the MSB bits of all bytes in a xmm reg to an int reg +INST3(movmskpd, "movmskpd", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0x50), INS_FLAGS_None) // Extract 2-bit sign mask from xmm and store in reg. The upper bits of r32 or r64 are filled with zeros. +INST3(movd, "movd", IUM_WR, PCKDBL(0x7E), BAD_CODE, PCKDBL(0x6E), INS_FLAGS_None) +INST3(movq, "movq", IUM_WR, PCKDBL(0xD6), BAD_CODE, SSEFLT(0x7E), INS_FLAGS_None) +INST3(movsdsse2, "movsd", IUM_WR, SSEDBL(0x11), BAD_CODE, SSEDBL(0x10), INS_FLAGS_None) + +INST3(punpckldq, "punpckldq", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0x62), INS_FLAGS_None) + +INST3(xorps, "xorps", IUM_WR, BAD_CODE, BAD_CODE, PCKFLT(0x57), INS_FLAGS_None) // XOR packed singles + +INST3(cvttsd2si, "cvttsd2si", IUM_WR, BAD_CODE, BAD_CODE, SSEDBL(0x2C), INS_FLAGS_None) // cvt with trunc scalar double to signed DWORDs + +INST3(movntdq, "movntdq", IUM_WR, PCKDBL(0xE7), BAD_CODE, BAD_CODE, INS_FLAGS_None) +INST3(movnti, "movnti", IUM_WR, PCKFLT(0xC3), BAD_CODE, BAD_CODE, INS_FLAGS_None) +INST3(movntpd, "movntpd", IUM_WR, PCKDBL(0x2B), BAD_CODE, BAD_CODE, INS_FLAGS_None) +INST3(movntps, "movntps", IUM_WR, PCKFLT(0x2B), BAD_CODE, BAD_CODE, INS_FLAGS_None) +INST3(movdqu, "movdqu", IUM_WR, SSEFLT(0x7F), BAD_CODE, SSEFLT(0x6F), INS_FLAGS_None) +INST3(movdqa, "movdqa", IUM_WR, PCKDBL(0x7F), BAD_CODE, PCKDBL(0x6F), INS_FLAGS_None) +INST3(movlpd, "movlpd", IUM_WR, PCKDBL(0x13), BAD_CODE, PCKDBL(0x12), INS_FLAGS_None) +INST3(movlps, "movlps", IUM_WR, PCKFLT(0x13), BAD_CODE, PCKFLT(0x12), INS_FLAGS_None) +INST3(movhpd, "movhpd", IUM_WR, PCKDBL(0x17), BAD_CODE, PCKDBL(0x16), INS_FLAGS_None) +INST3(movhps, "movhps", IUM_WR, PCKFLT(0x17), BAD_CODE, PCKFLT(0x16), INS_FLAGS_None) +INST3(movss, "movss", IUM_WR, SSEFLT(0x11), BAD_CODE, SSEFLT(0x10), INS_FLAGS_None) +INST3(movapd, "movapd", IUM_WR, PCKDBL(0x29), BAD_CODE, PCKDBL(0x28), INS_FLAGS_None) +INST3(movaps, "movaps", IUM_WR, PCKFLT(0x29), BAD_CODE, PCKFLT(0x28), INS_FLAGS_None) +INST3(movupd, "movupd", IUM_WR, PCKDBL(0x11), BAD_CODE, PCKDBL(0x10), INS_FLAGS_None) +INST3(movups, "movups", IUM_WR, PCKFLT(0x11), BAD_CODE, PCKFLT(0x10), INS_FLAGS_None) +INST3(movhlps, "movhlps", IUM_WR, BAD_CODE, BAD_CODE, PCKFLT(0x12), INS_FLAGS_None) +INST3(movlhps, "movlhps", IUM_WR, BAD_CODE, BAD_CODE, PCKFLT(0x16), INS_FLAGS_None) +INST3(movmskps, "movmskps", IUM_WR, BAD_CODE, BAD_CODE, PCKFLT(0x50), INS_FLAGS_None) +INST3(unpckhps, "unpckhps", IUM_WR, BAD_CODE, BAD_CODE, PCKFLT(0x15), INS_FLAGS_None) +INST3(unpcklps, "unpcklps", IUM_WR, BAD_CODE, BAD_CODE, PCKFLT(0x14), INS_FLAGS_None) +INST3(maskmovdqu, "maskmovdqu", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0xF7), INS_FLAGS_None) + +INST3(shufps, "shufps", IUM_WR, BAD_CODE, BAD_CODE, PCKFLT(0xC6), INS_FLAGS_None) +INST3(shufpd, "shufpd", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0xC6), INS_FLAGS_None) + +INST3(punpckhdq, "punpckhdq", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0x6A), INS_FLAGS_None) + +INST3(lfence, "lfence", IUM_RD, 0x000FE8AE, BAD_CODE, BAD_CODE, INS_FLAGS_None) +INST3(mfence, "mfence", IUM_RD, 0x000FF0AE, BAD_CODE, BAD_CODE, INS_FLAGS_None) +INST3(prefetchnta, "prefetchnta", IUM_RD, 0x000F0018, BAD_CODE, BAD_CODE, INS_FLAGS_None) +INST3(prefetcht0, "prefetcht0", IUM_RD, 0x000F0818, BAD_CODE, BAD_CODE, INS_FLAGS_None) +INST3(prefetcht1, "prefetcht1", IUM_RD, 0x000F1018, BAD_CODE, BAD_CODE, INS_FLAGS_None) +INST3(prefetcht2, "prefetcht2", IUM_RD, 0x000F1818, BAD_CODE, BAD_CODE, INS_FLAGS_None) +INST3(sfence, "sfence", IUM_RD, 0x000FF8AE, BAD_CODE, BAD_CODE, INS_FLAGS_None) // SSE 2 arith -INST3( addps, "addps", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKFLT(0x58)) // Add packed singles -INST3( addss, "addss", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSEFLT(0x58)) // Add scalar singles -INST3( addpd, "addpd", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0x58)) // Add packed doubles -INST3( addsd, "addsd", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSEDBL(0x58)) // Add scalar doubles -INST3( mulps, "mulps", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKFLT(0x59)) // Multiply packed singles -INST3( mulss, "mulss", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSEFLT(0x59)) // Multiply scalar single -INST3( mulpd, "mulpd", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0x59)) // Multiply packed doubles -INST3( mulsd, "mulsd", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSEDBL(0x59)) // Multiply scalar doubles -INST3( subps, "subps", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKFLT(0x5C)) // Subtract packed singles -INST3( subss, "subss", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSEFLT(0x5C)) // Subtract scalar singles -INST3( subpd, "subpd", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0x5C)) // Subtract packed doubles -INST3( subsd, "subsd", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSEDBL(0x5C)) // Subtract scalar doubles -INST3( minps, "minps", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKFLT(0x5D)) // Return Minimum packed singles -INST3( minss, "minss", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSEFLT(0x5D)) // Return Minimum scalar single -INST3( minpd, "minpd", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0x5D)) // Return Minimum packed doubles -INST3( minsd, "minsd", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSEDBL(0x5D)) // Return Minimum scalar double -INST3( divps, "divps", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKFLT(0x5E)) // Divide packed singles -INST3( divss, "divss", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSEFLT(0x5E)) // Divide scalar singles -INST3( divpd, "divpd", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0x5E)) // Divide packed doubles -INST3( divsd, "divsd", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSEDBL(0x5E)) // Divide scalar doubles -INST3( maxps, "maxps", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKFLT(0x5F)) // Return Maximum packed singles -INST3( maxss, "maxss", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSEFLT(0x5F)) // Return Maximum scalar single -INST3( maxpd, "maxpd", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0x5F)) // Return Maximum packed doubles -INST3( maxsd, "maxsd", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSEDBL(0x5F)) // Return Maximum scalar double -INST3( xorpd, "xorpd", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0x57)) // XOR packed doubles -INST3( andps, "andps", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKFLT(0x54)) // AND packed singles -INST3( andpd, "andpd", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0x54)) // AND packed doubles -INST3( sqrtps, "sqrtps", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKFLT(0x51)) // Sqrt of packed singles -INST3( sqrtss, "sqrtss", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSEFLT(0x51)) // Sqrt of scalar single -INST3( sqrtpd, "sqrtpd", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0x51)) // Sqrt of packed doubles -INST3( sqrtsd, "sqrtsd", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSEDBL(0x51)) // Sqrt of scalar double -INST3( andnps, "andnps", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKFLT(0x55)) // And-Not packed singles -INST3( andnpd, "andnpd", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0x55)) // And-Not packed doubles -INST3( orps, "orps", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKFLT(0x56)) // Or packed singles -INST3( orpd, "orpd", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0x56)) // Or packed doubles -INST3( haddpd, "haddpd", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0x7C)) // Horizontal add packed doubles -INST3( haddps, "haddps", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSEDBL(0x7C)) // Horizontal add packed floats -INST3( hsubpd, "hsubpd", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0x7D)) // Horizontal subtract packed doubles -INST3( hsubps, "hsubps", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSEDBL(0x7D)) // Horizontal subtract packed floats -INST3( addsubps, "addsubps", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSEDBL(0xD0)) // Add/Subtract packed singles -INST3( addsubpd, "addsubpd", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0xD0)) // Add/Subtract packed doubles +INST3(addps, "addps", IUM_WR, BAD_CODE, BAD_CODE, PCKFLT(0x58), INS_FLAGS_None) // Add packed singles +INST3(addss, "addss", IUM_WR, BAD_CODE, BAD_CODE, SSEFLT(0x58), INS_FLAGS_None) // Add scalar singles +INST3(addpd, "addpd", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0x58), INS_FLAGS_None) // Add packed doubles +INST3(addsd, "addsd", IUM_WR, BAD_CODE, BAD_CODE, SSEDBL(0x58), INS_FLAGS_None) // Add scalar doubles +INST3(mulps, "mulps", IUM_WR, BAD_CODE, BAD_CODE, PCKFLT(0x59), INS_FLAGS_None) // Multiply packed singles +INST3(mulss, "mulss", IUM_WR, BAD_CODE, BAD_CODE, SSEFLT(0x59), INS_FLAGS_None) // Multiply scalar single +INST3(mulpd, "mulpd", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0x59), INS_FLAGS_None) // Multiply packed doubles +INST3(mulsd, "mulsd", IUM_WR, BAD_CODE, BAD_CODE, SSEDBL(0x59), INS_FLAGS_None) // Multiply scalar doubles +INST3(subps, "subps", IUM_WR, BAD_CODE, BAD_CODE, PCKFLT(0x5C), INS_FLAGS_None) // Subtract packed singles +INST3(subss, "subss", IUM_WR, BAD_CODE, BAD_CODE, SSEFLT(0x5C), INS_FLAGS_None) // Subtract scalar singles +INST3(subpd, "subpd", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0x5C), INS_FLAGS_None) // Subtract packed doubles +INST3(subsd, "subsd", IUM_WR, BAD_CODE, BAD_CODE, SSEDBL(0x5C), INS_FLAGS_None) // Subtract scalar doubles +INST3(minps, "minps", IUM_WR, BAD_CODE, BAD_CODE, PCKFLT(0x5D), INS_FLAGS_None) // Return Minimum packed singles +INST3(minss, "minss", IUM_WR, BAD_CODE, BAD_CODE, SSEFLT(0x5D), INS_FLAGS_None) // Return Minimum scalar single +INST3(minpd, "minpd", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0x5D), INS_FLAGS_None) // Return Minimum packed doubles +INST3(minsd, "minsd", IUM_WR, BAD_CODE, BAD_CODE, SSEDBL(0x5D), INS_FLAGS_None) // Return Minimum scalar double +INST3(divps, "divps", IUM_WR, BAD_CODE, BAD_CODE, PCKFLT(0x5E), INS_FLAGS_None) // Divide packed singles +INST3(divss, "divss", IUM_WR, BAD_CODE, BAD_CODE, SSEFLT(0x5E), INS_FLAGS_None) // Divide scalar singles +INST3(divpd, "divpd", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0x5E), INS_FLAGS_None) // Divide packed doubles +INST3(divsd, "divsd", IUM_WR, BAD_CODE, BAD_CODE, SSEDBL(0x5E), INS_FLAGS_None) // Divide scalar doubles +INST3(maxps, "maxps", IUM_WR, BAD_CODE, BAD_CODE, PCKFLT(0x5F), INS_FLAGS_None) // Return Maximum packed singles +INST3(maxss, "maxss", IUM_WR, BAD_CODE, BAD_CODE, SSEFLT(0x5F), INS_FLAGS_None) // Return Maximum scalar single +INST3(maxpd, "maxpd", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0x5F), INS_FLAGS_None) // Return Maximum packed doubles +INST3(maxsd, "maxsd", IUM_WR, BAD_CODE, BAD_CODE, SSEDBL(0x5F), INS_FLAGS_None) // Return Maximum scalar double +INST3(xorpd, "xorpd", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0x57), INS_FLAGS_None) // XOR packed doubles +INST3(andps, "andps", IUM_WR, BAD_CODE, BAD_CODE, PCKFLT(0x54), INS_FLAGS_None) // AND packed singles +INST3(andpd, "andpd", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0x54), INS_FLAGS_None) // AND packed doubles +INST3(sqrtps, "sqrtps", IUM_WR, BAD_CODE, BAD_CODE, PCKFLT(0x51), INS_FLAGS_None) // Sqrt of packed singles +INST3(sqrtss, "sqrtss", IUM_WR, BAD_CODE, BAD_CODE, SSEFLT(0x51), INS_FLAGS_None) // Sqrt of scalar single +INST3(sqrtpd, "sqrtpd", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0x51), INS_FLAGS_None) // Sqrt of packed doubles +INST3(sqrtsd, "sqrtsd", IUM_WR, BAD_CODE, BAD_CODE, SSEDBL(0x51), INS_FLAGS_None) // Sqrt of scalar double +INST3(andnps, "andnps", IUM_WR, BAD_CODE, BAD_CODE, PCKFLT(0x55), INS_FLAGS_None) // And-Not packed singles +INST3(andnpd, "andnpd", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0x55), INS_FLAGS_None) // And-Not packed doubles +INST3(orps, "orps", IUM_WR, BAD_CODE, BAD_CODE, PCKFLT(0x56), INS_FLAGS_None) // Or packed singles +INST3(orpd, "orpd", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0x56), INS_FLAGS_None) // Or packed doubles +INST3(haddpd, "haddpd", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0x7C), INS_FLAGS_None) // Horizontal add packed doubles +INST3(haddps, "haddps", IUM_WR, BAD_CODE, BAD_CODE, SSEDBL(0x7C), INS_FLAGS_None) // Horizontal add packed floats +INST3(hsubpd, "hsubpd", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0x7D), INS_FLAGS_None) // Horizontal subtract packed doubles +INST3(hsubps, "hsubps", IUM_WR, BAD_CODE, BAD_CODE, SSEDBL(0x7D), INS_FLAGS_None) // Horizontal subtract packed floats +INST3(addsubps, "addsubps", IUM_WR, BAD_CODE, BAD_CODE, SSEDBL(0xD0), INS_FLAGS_None) // Add/Subtract packed singles +INST3(addsubpd, "addsubpd", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0xD0), INS_FLAGS_None) // Add/Subtract packed doubles // SSE 2 approx arith -INST3( rcpps, "rcpps", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKFLT(0x53)) // Reciprocal of packed singles -INST3( rcpss, "rcpss", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSEFLT(0x53)) // Reciprocal of scalar single -INST3( rsqrtps, "rsqrtps", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKFLT(0x52)) // Reciprocal Sqrt of packed singles -INST3( rsqrtss, "rsqrtss", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSEFLT(0x52)) // Reciprocal Sqrt of scalar single +INST3(rcpps, "rcpps", IUM_WR, BAD_CODE, BAD_CODE, PCKFLT(0x53), INS_FLAGS_None) // Reciprocal of packed singles +INST3(rcpss, "rcpss", IUM_WR, BAD_CODE, BAD_CODE, SSEFLT(0x53), INS_FLAGS_None) // Reciprocal of scalar single +INST3(rsqrtps, "rsqrtps", IUM_WR, BAD_CODE, BAD_CODE, PCKFLT(0x52), INS_FLAGS_None) // Reciprocal Sqrt of packed singles +INST3(rsqrtss, "rsqrtss", IUM_WR, BAD_CODE, BAD_CODE, SSEFLT(0x52), INS_FLAGS_None) // Reciprocal Sqrt of scalar single // SSE2 conversions -INST3( cvtpi2ps, "cvtpi2ps", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKFLT(0x2A)) // cvt packed DWORDs to singles -INST3( cvtsi2ss, "cvtsi2ss", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSEFLT(0x2A)) // cvt DWORD to scalar single -INST3( cvtpi2pd, "cvtpi2pd", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0x2A)) // cvt packed DWORDs to doubles -INST3( cvtsi2sd, "cvtsi2sd", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSEDBL(0x2A)) // cvt DWORD to scalar double -INST3( cvttps2pi, "cvttps2pi", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKFLT(0x2C)) // cvt with trunc packed singles to DWORDs -INST3( cvttss2si, "cvttss2si", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSEFLT(0x2C)) // cvt with trunc scalar single to DWORD -INST3( cvttpd2pi, "cvttpd2pi", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0x2C)) // cvt with trunc packed doubles to DWORDs -INST3( cvtps2pi, "cvtps2pi", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKFLT(0x2D)) // cvt packed singles to DWORDs -INST3( cvtss2si, "cvtss2si", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSEFLT(0x2D)) // cvt scalar single to DWORD -INST3( cvtpd2pi, "cvtpd2pi", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0x2D)) // cvt packed doubles to DWORDs -INST3( cvtsd2si, "cvtsd2si", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSEDBL(0x2D)) // cvt scalar double to DWORD -INST3( cvtps2pd, "cvtps2pd", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKFLT(0x5A)) // cvt packed singles to doubles -INST3( cvtpd2ps, "cvtpd2ps", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0x5A)) // cvt packed doubles to singles -INST3( cvtss2sd, "cvtss2sd", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSEFLT(0x5A)) // cvt scalar single to scalar doubles -INST3( cvtsd2ss, "cvtsd2ss", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSEDBL(0x5A)) // cvt scalar double to scalar singles -INST3( cvtdq2ps, "cvtdq2ps", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKFLT(0x5B)) // cvt packed DWORDs to singles -INST3( cvtps2dq, "cvtps2dq", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0x5B)) // cvt packed singles to DWORDs -INST3( cvttps2dq, "cvttps2dq", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSEFLT(0x5B)) // cvt with trunc packed singles to DWORDs -INST3( cvtpd2dq, "cvtpd2dq", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSEDBL(0xE6)) // cvt packed doubles to DWORDs -INST3( cvttpd2dq, "cvttpd2dq", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0xE6)) // cvt with trunc packed doubles to DWORDs -INST3( cvtdq2pd, "cvtdq2pd", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSEFLT(0xE6)) // cvt packed DWORDs to doubles +INST3(cvtpi2ps, "cvtpi2ps", IUM_WR, BAD_CODE, BAD_CODE, PCKFLT(0x2A), INS_FLAGS_None) // cvt packed DWORDs to singles +INST3(cvtsi2ss, "cvtsi2ss", IUM_WR, BAD_CODE, BAD_CODE, SSEFLT(0x2A), INS_FLAGS_None) // cvt DWORD to scalar single +INST3(cvtpi2pd, "cvtpi2pd", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0x2A), INS_FLAGS_None) // cvt packed DWORDs to doubles +INST3(cvtsi2sd, "cvtsi2sd", IUM_WR, BAD_CODE, BAD_CODE, SSEDBL(0x2A), INS_FLAGS_None) // cvt DWORD to scalar double +INST3(cvttps2pi, "cvttps2pi", IUM_WR, BAD_CODE, BAD_CODE, PCKFLT(0x2C), INS_FLAGS_None) // cvt with trunc packed singles to DWORDs +INST3(cvttss2si, "cvttss2si", IUM_WR, BAD_CODE, BAD_CODE, SSEFLT(0x2C), INS_FLAGS_None) // cvt with trunc scalar single to DWORD +INST3(cvttpd2pi, "cvttpd2pi", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0x2C), INS_FLAGS_None) // cvt with trunc packed doubles to DWORDs +INST3(cvtps2pi, "cvtps2pi", IUM_WR, BAD_CODE, BAD_CODE, PCKFLT(0x2D), INS_FLAGS_None) // cvt packed singles to DWORDs +INST3(cvtss2si, "cvtss2si", IUM_WR, BAD_CODE, BAD_CODE, SSEFLT(0x2D), INS_FLAGS_None) // cvt scalar single to DWORD +INST3(cvtpd2pi, "cvtpd2pi", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0x2D), INS_FLAGS_None) // cvt packed doubles to DWORDs +INST3(cvtsd2si, "cvtsd2si", IUM_WR, BAD_CODE, BAD_CODE, SSEDBL(0x2D), INS_FLAGS_None) // cvt scalar double to DWORD +INST3(cvtps2pd, "cvtps2pd", IUM_WR, BAD_CODE, BAD_CODE, PCKFLT(0x5A), INS_FLAGS_None) // cvt packed singles to doubles +INST3(cvtpd2ps, "cvtpd2ps", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0x5A), INS_FLAGS_None) // cvt packed doubles to singles +INST3(cvtss2sd, "cvtss2sd", IUM_WR, BAD_CODE, BAD_CODE, SSEFLT(0x5A), INS_FLAGS_None) // cvt scalar single to scalar doubles +INST3(cvtsd2ss, "cvtsd2ss", IUM_WR, BAD_CODE, BAD_CODE, SSEDBL(0x5A), INS_FLAGS_None) // cvt scalar double to scalar singles +INST3(cvtdq2ps, "cvtdq2ps", IUM_WR, BAD_CODE, BAD_CODE, PCKFLT(0x5B), INS_FLAGS_None) // cvt packed DWORDs to singles +INST3(cvtps2dq, "cvtps2dq", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0x5B), INS_FLAGS_None) // cvt packed singles to DWORDs +INST3(cvttps2dq, "cvttps2dq", IUM_WR, BAD_CODE, BAD_CODE, SSEFLT(0x5B), INS_FLAGS_None) // cvt with trunc packed singles to DWORDs +INST3(cvtpd2dq, "cvtpd2dq", IUM_WR, BAD_CODE, BAD_CODE, SSEDBL(0xE6), INS_FLAGS_None) // cvt packed doubles to DWORDs +INST3(cvttpd2dq, "cvttpd2dq", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0xE6), INS_FLAGS_None) // cvt with trunc packed doubles to DWORDs +INST3(cvtdq2pd, "cvtdq2pd", IUM_WR, BAD_CODE, BAD_CODE, SSEFLT(0xE6), INS_FLAGS_None) // cvt packed DWORDs to doubles // SSE2 comparison instructions -INST3( comiss, "comiss", 0, IUM_RD, 0, 0, BAD_CODE, BAD_CODE, PCKFLT(0x2F)) // ordered compare singles -INST3( comisd, "comisd", 0, IUM_RD, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0x2F)) // ordered compare doubles -INST3( ucomiss, "ucomiss", 0, IUM_RD, 0, 0, BAD_CODE, BAD_CODE, PCKFLT(0x2E)) // unordered compare singles -INST3( ucomisd, "ucomisd", 0, IUM_RD, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0x2E)) // unordered compare doubles +INST3(comiss, "comiss", IUM_RD, BAD_CODE, BAD_CODE, PCKFLT(0x2F), INS_FLAGS_None) // ordered compare singles +INST3(comisd, "comisd", IUM_RD, BAD_CODE, BAD_CODE, PCKDBL(0x2F), INS_FLAGS_None) // ordered compare doubles +INST3(ucomiss, "ucomiss", IUM_RD, BAD_CODE, BAD_CODE, PCKFLT(0x2E), INS_FLAGS_None) // unordered compare singles +INST3(ucomisd, "ucomisd", IUM_RD, BAD_CODE, BAD_CODE, PCKDBL(0x2E), INS_FLAGS_None) // unordered compare doubles // SSE2 packed single/double comparison operations. // Note that these instructions not only compare but also overwrite the first source. -INST3( cmpps, "cmpps", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKFLT(0xC2)) // compare packed singles -INST3( cmppd, "cmppd", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0xC2)) // compare packed doubles -INST3( cmpss, "cmpss", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSEFLT(0xC2)) // compare scalar singles -INST3( cmpsd, "cmpsd", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSEDBL(0xC2)) // compare scalar doubles +INST3(cmpps, "cmpps", IUM_WR, BAD_CODE, BAD_CODE, PCKFLT(0xC2), INS_FLAGS_None) // compare packed singles +INST3(cmppd, "cmppd", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0xC2), INS_FLAGS_None) // compare packed doubles +INST3(cmpss, "cmpss", IUM_WR, BAD_CODE, BAD_CODE, SSEFLT(0xC2), INS_FLAGS_None) // compare scalar singles +INST3(cmpsd, "cmpsd", IUM_WR, BAD_CODE, BAD_CODE, SSEDBL(0xC2), INS_FLAGS_None) // compare scalar doubles //SSE2 packed integer operations -INST3( paddb, "paddb" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0xFC)) // Add packed byte integers -INST3( paddw, "paddw" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0xFD)) // Add packed word (16-bit) integers -INST3( paddd, "paddd" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0xFE)) // Add packed double-word (32-bit) integers -INST3( paddq, "paddq" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0xD4)) // Add packed quad-word (64-bit) integers -INST3( paddsb, "paddsb" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0xEC)) // Add packed signed byte integers and saturate the results -INST3( paddsw, "paddsw" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0xED)) // Add packed signed word integers and saturate the results -INST3( paddusb, "paddusb" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0xDC)) // Add packed unsigned byte integers and saturate the results -INST3( paddusw, "paddusw" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0xDD)) // Add packed unsigned word integers and saturate the results -INST3( pavgb, "pavgb" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0xE0)) // Average of packed byte integers -INST3( pavgw, "pavgw" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0xE3)) // Average of packed word integers -INST3( psubb, "psubb" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0xF8)) // Subtract packed word (16-bit) integers -INST3( psubw, "psubw" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0xF9)) // Subtract packed word (16-bit) integers -INST3( psubd, "psubd" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0xFA)) // Subtract packed double-word (32-bit) integers -INST3( psubq, "psubq" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0xFB)) // subtract packed quad-word (64-bit) integers -INST3( pmaddwd, "pmaddwd" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0xF5)) // Multiply packed signed 16-bit integers in a and b, producing intermediate signed 32-bit integers. Horizontally add adjacent pairs of intermediate 32-bit integers, and pack the results in dst -INST3( pmulhw, "pmulhw" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0xE5)) // Multiply high the packed 16-bit signed integers -INST3( pmulhuw, "pmulhuw" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0xE4)) // Multiply high the packed 16-bit unsigned integers -INST3( pmuludq, "pmuludq" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0xF4)) // packed multiply 32-bit unsigned integers and store 64-bit result -INST3( pmullw, "pmullw" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0xD5)) // Packed multiply 16 bit unsigned integers and store lower 16 bits of each result -INST3( pand, "pand" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0xDB)) // Packed bit-wise AND of two xmm regs -INST3( pandn, "pandn" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0xDF)) // Packed bit-wise AND NOT of two xmm regs -INST3( por, "por" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0xEB)) // Packed bit-wise OR of two xmm regs -INST3( pxor, "pxor" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0xEF)) // Packed bit-wise XOR of two xmm regs -INST3( psadbw, "psadbw" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0xF6)) // Compute the sum of absolute differences of packed unsigned 8-bit integers -INST3( psubsb, "psubsb" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0xE8)) // Subtract packed 8-bit integers in b from packed 8-bit integers in a using saturation -INST3( psubusb, "psubusb" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0xD8)) // Subtract packed unsigned 8-bit integers in b from packed unsigned 8-bit integers in a using saturation -INST3( psubsw, "psubsw" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0xE9)) // Subtract packed 16-bit integers in b from packed 16-bit integers in a using saturation -INST3( psubusw, "psubusw" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0xD9)) // Subtract packed unsigned 16-bit integers in b from packed unsigned 16-bit integers in a using saturation +INST3(paddb, "paddb", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0xFC), INS_FLAGS_None) // Add packed byte integers +INST3(paddw, "paddw", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0xFD), INS_FLAGS_None) // Add packed word (16-bit) integers +INST3(paddd, "paddd", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0xFE), INS_FLAGS_None) // Add packed double-word (32-bit) integers +INST3(paddq, "paddq", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0xD4), INS_FLAGS_None) // Add packed quad-word (64-bit) integers +INST3(paddsb, "paddsb", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0xEC), INS_FLAGS_None) // Add packed signed byte integers and saturate the results +INST3(paddsw, "paddsw", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0xED), INS_FLAGS_None) // Add packed signed word integers and saturate the results +INST3(paddusb, "paddusb", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0xDC), INS_FLAGS_None) // Add packed unsigned byte integers and saturate the results +INST3(paddusw, "paddusw", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0xDD), INS_FLAGS_None) // Add packed unsigned word integers and saturate the results +INST3(pavgb, "pavgb", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0xE0), INS_FLAGS_None) // Average of packed byte integers +INST3(pavgw, "pavgw", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0xE3), INS_FLAGS_None) // Average of packed word integers +INST3(psubb, "psubb", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0xF8), INS_FLAGS_None) // Subtract packed word (16-bit) integers +INST3(psubw, "psubw", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0xF9), INS_FLAGS_None) // Subtract packed word (16-bit) integers +INST3(psubd, "psubd", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0xFA), INS_FLAGS_None) // Subtract packed double-word (32-bit) integers +INST3(psubq, "psubq", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0xFB), INS_FLAGS_None) // subtract packed quad-word (64-bit) integers +INST3(pmaddwd, "pmaddwd", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0xF5), INS_FLAGS_None) // Multiply packed signed 16-bit integers in a and b, producing intermediate signed 32-bit integers. Horizontally add adjacent pairs of intermediate 32-bit integers, and pack the results in dst +INST3(pmulhw, "pmulhw", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0xE5), INS_FLAGS_None) // Multiply high the packed 16-bit signed integers +INST3(pmulhuw, "pmulhuw", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0xE4), INS_FLAGS_None) // Multiply high the packed 16-bit unsigned integers +INST3(pmuludq, "pmuludq", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0xF4), INS_FLAGS_None) // packed multiply 32-bit unsigned integers and store 64-bit result +INST3(pmullw, "pmullw", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0xD5), INS_FLAGS_None) // Packed multiply 16 bit unsigned integers and store lower 16 bits of each result +INST3(pand, "pand", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0xDB), INS_FLAGS_None) // Packed bit-wise AND of two xmm regs +INST3(pandn, "pandn", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0xDF), INS_FLAGS_None) // Packed bit-wise AND NOT of two xmm regs +INST3(por, "por", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0xEB), INS_FLAGS_None) // Packed bit-wise OR of two xmm regs +INST3(pxor, "pxor", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0xEF), INS_FLAGS_None) // Packed bit-wise XOR of two xmm regs +INST3(psadbw, "psadbw", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0xF6), INS_FLAGS_None) // Compute the sum of absolute differences of packed unsigned 8-bit integers +INST3(psubsb, "psubsb", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0xE8), INS_FLAGS_None) // Subtract packed 8-bit integers in b from packed 8-bit integers in a using saturation +INST3(psubusb, "psubusb", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0xD8), INS_FLAGS_None) // Subtract packed unsigned 8-bit integers in b from packed unsigned 8-bit integers in a using saturation +INST3(psubsw, "psubsw", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0xE9), INS_FLAGS_None) // Subtract packed 16-bit integers in b from packed 16-bit integers in a using saturation +INST3(psubusw, "psubusw", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0xD9), INS_FLAGS_None) // Subtract packed unsigned 16-bit integers in b from packed unsigned 16-bit integers in a using saturation // Note that the shift immediates share the same encoding between left and right-shift, and are distinguished by the Reg/Opcode, // which is handled in emitxarch.cpp. -INST3( psrldq, "psrldq" , 0, IUM_WR, 0, 0, BAD_CODE, PCKDBL(0x73), BAD_CODE ) // Shift right logical of xmm reg by given number of bytes -INST3( pslldq, "pslldq" , 0, IUM_WR, 0, 0, BAD_CODE, PCKDBL(0x73), BAD_CODE ) // Shift left logical of xmm reg by given number of bytes -INST3( psllw, "psllw" , 0, IUM_WR, 0, 0, BAD_CODE, PCKDBL(0x71), PCKDBL(0xF1)) // Packed shift left logical of 16-bit integers -INST3( pslld, "pslld" , 0, IUM_WR, 0, 0, BAD_CODE, PCKDBL(0x72), PCKDBL(0xF2)) // Packed shift left logical of 32-bit integers -INST3( psllq, "psllq" , 0, IUM_WR, 0, 0, BAD_CODE, PCKDBL(0x73), PCKDBL(0xF3)) // Packed shift left logical of 64-bit integers -INST3( psrlw, "psrlw" , 0, IUM_WR, 0, 0, BAD_CODE, PCKDBL(0x71), PCKDBL(0xD1)) // Packed shift right logical of 16-bit integers -INST3( psrld, "psrld" , 0, IUM_WR, 0, 0, BAD_CODE, PCKDBL(0x72), PCKDBL(0xD2)) // Packed shift right logical of 32-bit integers -INST3( psrlq, "psrlq" , 0, IUM_WR, 0, 0, BAD_CODE, PCKDBL(0x73), PCKDBL(0xD3)) // Packed shift right logical of 64-bit integers -INST3( psraw, "psraw" , 0, IUM_WR, 0, 0, BAD_CODE, PCKDBL(0x71), PCKDBL(0xE1)) // Packed shift right arithmetic of 16-bit integers -INST3( psrad, "psrad" , 0, IUM_WR, 0, 0, BAD_CODE, PCKDBL(0x72), PCKDBL(0xE2)) // Packed shift right arithmetic of 32-bit integers - -INST3( pmaxub, "pmaxub" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0xDE)) // packed maximum unsigned bytes -INST3( pminub, "pminub" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0xDA)) // packed minimum unsigned bytes -INST3( pmaxsw, "pmaxsw" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0xEE)) // packed maximum signed words -INST3( pminsw, "pminsw" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0xEA)) // packed minimum signed words -INST3( pcmpeqd, "pcmpeqd" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0x76)) // Packed compare 32-bit integers for equality -INST3( pcmpgtd, "pcmpgtd" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0x66)) // Packed compare 32-bit signed integers for greater than -INST3( pcmpeqw, "pcmpeqw" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0x75)) // Packed compare 16-bit integers for equality -INST3( pcmpgtw, "pcmpgtw" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0x65)) // Packed compare 16-bit signed integers for greater than -INST3( pcmpeqb, "pcmpeqb" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0x74)) // Packed compare 8-bit integers for equality -INST3( pcmpgtb, "pcmpgtb" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0x64)) // Packed compare 8-bit signed integers for greater than - -INST3( pshufd, "pshufd" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0x70)) // Packed shuffle of 32-bit integers -INST3( pshufhw, "pshufhw" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSEFLT(0x70)) // Shuffle the high words in xmm2/m128 based on the encoding in imm8 and store the result in xmm1. -INST3( pshuflw, "pshuflw" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSEDBL(0x70)) // Shuffle the low words in xmm2/m128 based on the encoding in imm8 and store the result in xmm1. -INST3( pextrw, "pextrw" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0xC5)) // Extract 16-bit value into a r32 with zero extended to 32-bits -INST3( pinsrw, "pinsrw" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0xC4)) // Insert word at index - -INST3( punpckhbw, "punpckhbw" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0x68)) // Packed logical (unsigned) widen ubyte to ushort (hi) -INST3( punpcklbw, "punpcklbw" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0x60)) // Packed logical (unsigned) widen ubyte to ushort (lo) -INST3( punpckhqdq, "punpckhqdq" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0x6D)) // Packed logical (unsigned) widen uint to ulong (hi) -INST3( punpcklqdq, "punpcklqdq" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0x6C)) // Packed logical (unsigned) widen uint to ulong (lo) -INST3( punpckhwd, "punpckhwd" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0x69)) // Packed logical (unsigned) widen ushort to uint (hi) -INST3( punpcklwd, "punpcklwd" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0x61)) // Packed logical (unsigned) widen ushort to uint (lo) -INST3( unpckhpd, "unpckhpd" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0x15)) // Packed logical (unsigned) widen ubyte to ushort (hi) -INST3( unpcklpd, "unpcklpd" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0x14)) // Packed logical (unsigned) widen ubyte to ushort (hi) - -INST3( packssdw, "packssdw" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0x6B)) // Pack (narrow) int to short with saturation -INST3( packsswb, "packsswb" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0x63)) // Pack (narrow) short to byte with saturation -INST3( packuswb, "packuswb" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PCKDBL(0x67)) // Pack (narrow) short to unsigned byte with saturation - -INST3(LAST_SSE2_INSTRUCTION, "LAST_SSE2_INSTRUCTION", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, BAD_CODE) - -INST3(FIRST_SSE4_INSTRUCTION, "FIRST_SSE4_INSTRUCTION", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, BAD_CODE) - -// enum name FP updmode rf wf MR MI RM -INST3( dpps, "dpps" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE3A(0x40)) // Packed dot product of two float vector regs -INST3( dppd, "dppd" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE3A(0x41)) // Packed dot product of two double vector regs -INST3( insertps, "insertps" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE3A(0x21)) // Insert packed single precision float value -INST3( pcmpeqq, "pcmpeqq" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x29)) // Packed compare 64-bit integers for equality -INST3( pcmpgtq, "pcmpgtq" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x37)) // Packed compare 64-bit integers for equality -INST3( pmulld, "pmulld" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x40)) // Packed multiply 32 bit unsigned integers and store lower 32 bits of each result -INST3( ptest, "ptest" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x17)) // Packed logical compare -INST3( phaddd, "phaddd" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x02)) // Packed horizontal add -INST3( pabsb, "pabsb" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x1C)) // Packed absolute value of bytes -INST3( pabsw, "pabsw" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x1D)) // Packed absolute value of 16-bit integers -INST3( pabsd, "pabsd" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x1E)) // Packed absolute value of 32-bit integers -INST3( palignr, "palignr" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE3A(0x0F)) // Packed Align Right -INST3( pmaddubsw, "pmaddubsw" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x04)) // Multiply and Add Packed Signed and Unsigned Bytes -INST3( pmulhrsw, "pmulhrsw" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x0B)) // Packed Multiply High with Round and Scale -INST3( pshufb, "pshufb" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x00)) // Packed Shuffle Bytes -INST3( psignb, "psignb" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x08)) // Packed SIGN -INST3( psignw, "psignw" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x09)) // Packed SIGN -INST3( psignd, "psignd" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x0A)) // Packed SIGN -INST3( pminsb, "pminsb" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x38)) // packed minimum signed bytes -INST3( pminsd, "pminsd" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x39)) // packed minimum 32-bit signed integers -INST3( pminuw, "pminuw" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x3A)) // packed minimum 16-bit unsigned integers -INST3( pminud, "pminud" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x3B)) // packed minimum 32-bit unsigned integers -INST3( pmaxsb, "pmaxsb" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x3C)) // packed maximum signed bytes -INST3( pmaxsd, "pmaxsd" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x3D)) // packed maximum 32-bit signed integers -INST3( pmaxuw, "pmaxuw" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x3E)) // packed maximum 16-bit unsigned integers -INST3( pmaxud, "pmaxud" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x3F)) // packed maximum 32-bit unsigned integers -INST3( pmovsxbw, "pmovsxbw" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x20)) // Packed sign extend byte to short -INST3( pmovsxbd, "pmovsxbd" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x21)) // Packed sign extend byte to int -INST3( pmovsxbq, "pmovsxbq" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x22)) // Packed sign extend byte to long -INST3( pmovsxwd, "pmovsxwd" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x23)) // Packed sign extend short to int -INST3( pmovsxwq, "pmovsxwq" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x24)) // Packed sign extend short to long -INST3( pmovsxdq, "pmovsxdq" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x25)) // Packed sign extend int to long -INST3( pmovzxbw, "pmovzxbw" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x30)) // Packed zero extend byte to short -INST3( pmovzxbd, "pmovzxbd" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x31)) // Packed zero extend byte to intg -INST3( pmovzxbq, "pmovzxbq" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x32)) // Packed zero extend byte to lon -INST3( pmovzxwd, "pmovzxwd" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x33)) // Packed zero extend short to int -INST3( pmovzxwq, "pmovzxwq" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x34)) // Packed zero extend short to long -INST3( pmovzxdq, "pmovzxdq" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x35)) // Packed zero extend int to long -INST3( packusdw, "packusdw" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x2B)) // Pack (narrow) int to unsigned short with saturation -INST3( roundps, "roundps" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE3A(0x08)) // Round packed single precision floating-point values -INST3( roundss, "roundss" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE3A(0x0A)) // Round scalar single precision floating-point values -INST3( roundpd, "roundpd" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE3A(0x09)) // Round packed double precision floating-point values -INST3( roundsd, "roundsd" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE3A(0x0B)) // Round scalar double precision floating-point values -INST3( pmuldq, "pmuldq" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x28)) // packed multiply 32-bit signed integers and store 64-bit result -INST3( blendps, "blendps" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE3A(0x0C)) // Blend Packed Single Precision Floating-Point Values -INST3( blendvps, "blendvps" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x14)) // Variable Blend Packed Singles -INST3( blendpd, "blendpd" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE3A(0x0D)) // Blend Packed Double Precision Floating-Point Values -INST3( blendvpd, "blendvpd" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x15)) // Variable Blend Packed Doubles -INST3( pblendw, "pblendw" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE3A(0x0E)) // Blend Packed Words -INST3( pblendvb, "pblendvb" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x10)) // Variable Blend Packed Bytes -INST3( phaddw, "phaddw" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x01)) // Packed horizontal add of 16-bit integers -INST3( phsubw, "phsubw" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x05)) // Packed horizontal subtract of 16-bit integers -INST3( phsubd, "phsubd" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x06)) // Packed horizontal subtract of 32-bit integers -INST3( phaddsw, "phaddsw" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x03)) // Packed horizontal add of 16-bit integers with saturation -INST3( phsubsw, "phsubsw" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x07)) // Packed horizontal subtract of 16-bit integers with saturation -INST3( lddqu, "lddqu" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSEDBL(0xF0)) // Load Unaligned integer -INST3( movntdqa, "movntdqa" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x2A)) // Load Double Quadword Non-Temporal Aligned Hint -INST3( movddup, "movddup" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSEDBL(0x12)) // Replicate Double FP Values -INST3( movsldup, "movsldup" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSEFLT(0x12)) // Replicate even-indexed Single FP Values -INST3( movshdup, "movshdup" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSEFLT(0x16)) // Replicate odd-indexed Single FP Values -INST3( phminposuw, "phminposuw" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x41)) // Packed Horizontal Word Minimum -INST3( mpsadbw, "mpsadbw" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE3A(0x42)) // Compute Multiple Packed Sums of Absolute Difference -INST3( pinsrb, "pinsrb" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE3A(0x20)) // Insert Byte -INST3( pinsrd, "pinsrd" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE3A(0x22)) // Insert Dword -INST3( pinsrq, "pinsrq" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE3A(0x22)) // Insert Qword -INST3( pextrb, "pextrb" , 0, IUM_WR, 0, 0, SSE3A(0x14), BAD_CODE, BAD_CODE) // Extract Byte -INST3( pextrd, "pextrd" , 0, IUM_WR, 0, 0, SSE3A(0x16), BAD_CODE, BAD_CODE) // Extract Dword -INST3( pextrq, "pextrq" , 0, IUM_WR, 0, 0, SSE3A(0x16), BAD_CODE, BAD_CODE) // Extract Qword -INST3( pextrw_sse41, "pextrw" , 0, IUM_WR, 0, 0, SSE3A(0x15), BAD_CODE, BAD_CODE) // Extract Word -INST3( extractps, "extractps" , 0, IUM_WR, 0, 0, SSE3A(0x17), BAD_CODE, BAD_CODE) // Extract Packed Floating-Point Values +INST3(psrldq, "psrldq", IUM_WR, BAD_CODE, PCKDBL(0x73), BAD_CODE, INS_FLAGS_None) // Shift right logical of xmm reg by given number of bytes +INST3(pslldq, "pslldq", IUM_WR, BAD_CODE, PCKDBL(0x73), BAD_CODE, INS_FLAGS_None) // Shift left logical of xmm reg by given number of bytes +INST3(psllw, "psllw", IUM_WR, BAD_CODE, PCKDBL(0x71), PCKDBL(0xF1), INS_FLAGS_None) // Packed shift left logical of 16-bit integers +INST3(pslld, "pslld", IUM_WR, BAD_CODE, PCKDBL(0x72), PCKDBL(0xF2), INS_FLAGS_None) // Packed shift left logical of 32-bit integers +INST3(psllq, "psllq", IUM_WR, BAD_CODE, PCKDBL(0x73), PCKDBL(0xF3), INS_FLAGS_None) // Packed shift left logical of 64-bit integers +INST3(psrlw, "psrlw", IUM_WR, BAD_CODE, PCKDBL(0x71), PCKDBL(0xD1), INS_FLAGS_None) // Packed shift right logical of 16-bit integers +INST3(psrld, "psrld", IUM_WR, BAD_CODE, PCKDBL(0x72), PCKDBL(0xD2), INS_FLAGS_None) // Packed shift right logical of 32-bit integers +INST3(psrlq, "psrlq", IUM_WR, BAD_CODE, PCKDBL(0x73), PCKDBL(0xD3), INS_FLAGS_None) // Packed shift right logical of 64-bit integers +INST3(psraw, "psraw", IUM_WR, BAD_CODE, PCKDBL(0x71), PCKDBL(0xE1), INS_FLAGS_None) // Packed shift right arithmetic of 16-bit integers +INST3(psrad, "psrad", IUM_WR, BAD_CODE, PCKDBL(0x72), PCKDBL(0xE2), INS_FLAGS_None) // Packed shift right arithmetic of 32-bit integers + +INST3(pmaxub, "pmaxub", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0xDE), INS_FLAGS_None) // packed maximum unsigned bytes +INST3(pminub, "pminub", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0xDA), INS_FLAGS_None) // packed minimum unsigned bytes +INST3(pmaxsw, "pmaxsw", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0xEE), INS_FLAGS_None) // packed maximum signed words +INST3(pminsw, "pminsw", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0xEA), INS_FLAGS_None) // packed minimum signed words +INST3(pcmpeqd, "pcmpeqd", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0x76), INS_FLAGS_None) // Packed compare 32-bit integers for equality +INST3(pcmpgtd, "pcmpgtd", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0x66), INS_FLAGS_None) // Packed compare 32-bit signed integers for greater than +INST3(pcmpeqw, "pcmpeqw", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0x75), INS_FLAGS_None) // Packed compare 16-bit integers for equality +INST3(pcmpgtw, "pcmpgtw", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0x65), INS_FLAGS_None) // Packed compare 16-bit signed integers for greater than +INST3(pcmpeqb, "pcmpeqb", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0x74), INS_FLAGS_None) // Packed compare 8-bit integers for equality +INST3(pcmpgtb, "pcmpgtb", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0x64), INS_FLAGS_None) // Packed compare 8-bit signed integers for greater than + +INST3(pshufd, "pshufd", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0x70), INS_FLAGS_None) // Packed shuffle of 32-bit integers +INST3(pshufhw, "pshufhw", IUM_WR, BAD_CODE, BAD_CODE, SSEFLT(0x70), INS_FLAGS_None) // Shuffle the high words in xmm2/m128 based on the encoding in imm8 and store the result in xmm1. +INST3(pshuflw, "pshuflw", IUM_WR, BAD_CODE, BAD_CODE, SSEDBL(0x70), INS_FLAGS_None) // Shuffle the low words in xmm2/m128 based on the encoding in imm8 and store the result in xmm1. +INST3(pextrw, "pextrw", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0xC5), INS_FLAGS_None) // Extract 16-bit value into a r32 with zero extended to 32-bits +INST3(pinsrw, "pinsrw", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0xC4), INS_FLAGS_None) // Insert word at index + +INST3(punpckhbw, "punpckhbw", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0x68), INS_FLAGS_None) // Packed logical (unsigned) widen ubyte to ushort (hi) +INST3(punpcklbw, "punpcklbw", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0x60), INS_FLAGS_None) // Packed logical (unsigned) widen ubyte to ushort (lo) +INST3(punpckhqdq, "punpckhqdq", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0x6D), INS_FLAGS_None) // Packed logical (unsigned) widen uint to ulong (hi) +INST3(punpcklqdq, "punpcklqdq", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0x6C), INS_FLAGS_None) // Packed logical (unsigned) widen uint to ulong (lo) +INST3(punpckhwd, "punpckhwd", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0x69), INS_FLAGS_None) // Packed logical (unsigned) widen ushort to uint (hi) +INST3(punpcklwd, "punpcklwd", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0x61), INS_FLAGS_None) // Packed logical (unsigned) widen ushort to uint (lo) +INST3(unpckhpd, "unpckhpd", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0x15), INS_FLAGS_None) // Packed logical (unsigned) widen ubyte to ushort (hi) +INST3(unpcklpd, "unpcklpd", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0x14), INS_FLAGS_None) // Packed logical (unsigned) widen ubyte to ushort (hi) + +INST3(packssdw, "packssdw", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0x6B), INS_FLAGS_None) // Pack (narrow) int to short with saturation +INST3(packsswb, "packsswb", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0x63), INS_FLAGS_None) // Pack (narrow) short to byte with saturation +INST3(packuswb, "packuswb", IUM_WR, BAD_CODE, BAD_CODE, PCKDBL(0x67), INS_FLAGS_None) // Pack (narrow) short to unsigned byte with saturation +INST3(LAST_SSE2_INSTRUCTION, "LAST_SSE2_INSTRUCTION", IUM_WR, BAD_CODE, BAD_CODE, BAD_CODE, INS_FLAGS_None) + +INST3(FIRST_SSE4_INSTRUCTION, "FIRST_SSE4_INSTRUCTION", IUM_WR, BAD_CODE, BAD_CODE, BAD_CODE, INS_FLAGS_None) +// id nm um mr mi rm flags +INST3(dpps, "dpps", IUM_WR, BAD_CODE, BAD_CODE, SSE3A(0x40), INS_FLAGS_None) // Packed dot product of two float vector regs +INST3(dppd, "dppd", IUM_WR, BAD_CODE, BAD_CODE, SSE3A(0x41), INS_FLAGS_None) // Packed dot product of two double vector regs +INST3(insertps, "insertps", IUM_WR, BAD_CODE, BAD_CODE, SSE3A(0x21), INS_FLAGS_None) // Insert packed single precision float value +INST3(pcmpeqq, "pcmpeqq", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x29), INS_FLAGS_None) // Packed compare 64-bit integers for equality +INST3(pcmpgtq, "pcmpgtq", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x37), INS_FLAGS_None) // Packed compare 64-bit integers for equality +INST3(pmulld, "pmulld", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x40), INS_FLAGS_None) // Packed multiply 32 bit unsigned integers and store lower 32 bits of each result +INST3(ptest, "ptest", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x17), INS_FLAGS_None) // Packed logical compare +INST3(phaddd, "phaddd", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x02), INS_FLAGS_None) // Packed horizontal add +INST3(pabsb, "pabsb", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x1C), INS_FLAGS_None) // Packed absolute value of bytes +INST3(pabsw, "pabsw", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x1D), INS_FLAGS_None) // Packed absolute value of 16-bit integers +INST3(pabsd, "pabsd", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x1E), INS_FLAGS_None) // Packed absolute value of 32-bit integers +INST3(palignr, "palignr", IUM_WR, BAD_CODE, BAD_CODE, SSE3A(0x0F), INS_FLAGS_None) // Packed Align Right +INST3(pmaddubsw, "pmaddubsw", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x04), INS_FLAGS_None) // Multiply and Add Packed Signed and Unsigned Bytes +INST3(pmulhrsw, "pmulhrsw", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x0B), INS_FLAGS_None) // Packed Multiply High with Round and Scale +INST3(pshufb, "pshufb", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x00), INS_FLAGS_None) // Packed Shuffle Bytes +INST3(psignb, "psignb", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x08), INS_FLAGS_None) // Packed SIGN +INST3(psignw, "psignw", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x09), INS_FLAGS_None) // Packed SIGN +INST3(psignd, "psignd", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x0A), INS_FLAGS_None) // Packed SIGN +INST3(pminsb, "pminsb", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x38), INS_FLAGS_None) // packed minimum signed bytes +INST3(pminsd, "pminsd", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x39), INS_FLAGS_None) // packed minimum 32-bit signed integers +INST3(pminuw, "pminuw", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x3A), INS_FLAGS_None) // packed minimum 16-bit unsigned integers +INST3(pminud, "pminud", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x3B), INS_FLAGS_None) // packed minimum 32-bit unsigned integers +INST3(pmaxsb, "pmaxsb", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x3C), INS_FLAGS_None) // packed maximum signed bytes +INST3(pmaxsd, "pmaxsd", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x3D), INS_FLAGS_None) // packed maximum 32-bit signed integers +INST3(pmaxuw, "pmaxuw", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x3E), INS_FLAGS_None) // packed maximum 16-bit unsigned integers +INST3(pmaxud, "pmaxud", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x3F), INS_FLAGS_None) // packed maximum 32-bit unsigned integers +INST3(pmovsxbw, "pmovsxbw", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x20), INS_FLAGS_None) // Packed sign extend byte to short +INST3(pmovsxbd, "pmovsxbd", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x21), INS_FLAGS_None) // Packed sign extend byte to int +INST3(pmovsxbq, "pmovsxbq", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x22), INS_FLAGS_None) // Packed sign extend byte to long +INST3(pmovsxwd, "pmovsxwd", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x23), INS_FLAGS_None) // Packed sign extend short to int +INST3(pmovsxwq, "pmovsxwq", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x24), INS_FLAGS_None) // Packed sign extend short to long +INST3(pmovsxdq, "pmovsxdq", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x25), INS_FLAGS_None) // Packed sign extend int to long +INST3(pmovzxbw, "pmovzxbw", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x30), INS_FLAGS_None) // Packed zero extend byte to short +INST3(pmovzxbd, "pmovzxbd", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x31), INS_FLAGS_None) // Packed zero extend byte to intg +INST3(pmovzxbq, "pmovzxbq", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x32), INS_FLAGS_None) // Packed zero extend byte to lon +INST3(pmovzxwd, "pmovzxwd", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x33), INS_FLAGS_None) // Packed zero extend short to int +INST3(pmovzxwq, "pmovzxwq", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x34), INS_FLAGS_None) // Packed zero extend short to long +INST3(pmovzxdq, "pmovzxdq", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x35), INS_FLAGS_None) // Packed zero extend int to long +INST3(packusdw, "packusdw", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x2B), INS_FLAGS_None) // Pack (narrow) int to unsigned short with saturation +INST3(roundps, "roundps", IUM_WR, BAD_CODE, BAD_CODE, SSE3A(0x08), INS_FLAGS_None) // Round packed single precision floating-point values +INST3(roundss, "roundss", IUM_WR, BAD_CODE, BAD_CODE, SSE3A(0x0A), INS_FLAGS_None) // Round scalar single precision floating-point values +INST3(roundpd, "roundpd", IUM_WR, BAD_CODE, BAD_CODE, SSE3A(0x09), INS_FLAGS_None) // Round packed double precision floating-point values +INST3(roundsd, "roundsd", IUM_WR, BAD_CODE, BAD_CODE, SSE3A(0x0B), INS_FLAGS_None) // Round scalar double precision floating-point values +INST3(pmuldq, "pmuldq", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x28), INS_FLAGS_None) // packed multiply 32-bit signed integers and store 64-bit result +INST3(blendps, "blendps", IUM_WR, BAD_CODE, BAD_CODE, SSE3A(0x0C), INS_FLAGS_None) // Blend Packed Single Precision Floating-Point Values +INST3(blendvps, "blendvps", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x14), INS_FLAGS_None) // Variable Blend Packed Singles +INST3(blendpd, "blendpd", IUM_WR, BAD_CODE, BAD_CODE, SSE3A(0x0D), INS_FLAGS_None) // Blend Packed Double Precision Floating-Point Values +INST3(blendvpd, "blendvpd", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x15), INS_FLAGS_None) // Variable Blend Packed Doubles +INST3(pblendw, "pblendw", IUM_WR, BAD_CODE, BAD_CODE, SSE3A(0x0E), INS_FLAGS_None) // Blend Packed Words +INST3(pblendvb, "pblendvb", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x10), INS_FLAGS_None) // Variable Blend Packed Bytes +INST3(phaddw, "phaddw", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x01), INS_FLAGS_None) // Packed horizontal add of 16-bit integers +INST3(phsubw, "phsubw", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x05), INS_FLAGS_None) // Packed horizontal subtract of 16-bit integers +INST3(phsubd, "phsubd", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x06), INS_FLAGS_None) // Packed horizontal subtract of 32-bit integers +INST3(phaddsw, "phaddsw", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x03), INS_FLAGS_None) // Packed horizontal add of 16-bit integers with saturation +INST3(phsubsw, "phsubsw", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x07), INS_FLAGS_None) // Packed horizontal subtract of 16-bit integers with saturation +INST3(lddqu, "lddqu", IUM_WR, BAD_CODE, BAD_CODE, SSEDBL(0xF0), INS_FLAGS_None) // Load Unaligned integer +INST3(movntdqa, "movntdqa", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x2A), INS_FLAGS_None) // Load Double Quadword Non-Temporal Aligned Hint +INST3(movddup, "movddup", IUM_WR, BAD_CODE, BAD_CODE, SSEDBL(0x12), INS_FLAGS_None) // Replicate Double FP Values +INST3(movsldup, "movsldup", IUM_WR, BAD_CODE, BAD_CODE, SSEFLT(0x12), INS_FLAGS_None) // Replicate even-indexed Single FP Values +INST3(movshdup, "movshdup", IUM_WR, BAD_CODE, BAD_CODE, SSEFLT(0x16), INS_FLAGS_None) // Replicate odd-indexed Single FP Values +INST3(phminposuw, "phminposuw", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x41), INS_FLAGS_None) // Packed Horizontal Word Minimum +INST3(mpsadbw, "mpsadbw", IUM_WR, BAD_CODE, BAD_CODE, SSE3A(0x42), INS_FLAGS_None) // Compute Multiple Packed Sums of Absolute Difference +INST3(pinsrb, "pinsrb", IUM_WR, BAD_CODE, BAD_CODE, SSE3A(0x20), INS_FLAGS_None) // Insert Byte +INST3(pinsrd, "pinsrd", IUM_WR, BAD_CODE, BAD_CODE, SSE3A(0x22), INS_FLAGS_None) // Insert Dword +INST3(pinsrq, "pinsrq", IUM_WR, BAD_CODE, BAD_CODE, SSE3A(0x22), INS_FLAGS_None) // Insert Qword +INST3(pextrb, "pextrb", IUM_WR, SSE3A(0x14), BAD_CODE, BAD_CODE, INS_FLAGS_None) // Extract Byte +INST3(pextrd, "pextrd", IUM_WR, SSE3A(0x16), BAD_CODE, BAD_CODE, INS_FLAGS_None) // Extract Dword +INST3(pextrq, "pextrq", IUM_WR, SSE3A(0x16), BAD_CODE, BAD_CODE, INS_FLAGS_None) // Extract Qword +INST3(pextrw_sse41, "pextrw", IUM_WR, SSE3A(0x15), BAD_CODE, BAD_CODE, INS_FLAGS_None) // Extract Word +INST3(extractps, "extractps", IUM_WR, SSE3A(0x17), BAD_CODE, BAD_CODE, INS_FLAGS_None) // Extract Packed Floating-Point Values //AES instructions -INST3(aesdec, "aesdec" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0xDE)) // Perform one round of an AES decryption flow -INST3(aesdeclast, "aesdeclast" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0xDF)) // Perform last round of an AES decryption flow -INST3(aesenc, "aesenc" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0xDC)) // Perform one round of an AES encryption flow -INST3(aesenclast, "aesenclast" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0xDD)) // Perform last round of an AES encryption flow -INST3(aesimc, "aesimc" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0xDB)) // Perform the AES InvMixColumn Transformation -INST3(aeskeygenassist,"aeskeygenassist" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE3A(0xDF)) // AES Round Key Generation Assist - -INST3(LAST_SSE4_INSTRUCTION, "LAST_SSE4_INSTRUCTION", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, BAD_CODE) - -INST3(FIRST_AVX_INSTRUCTION, "FIRST_AVX_INSTRUCTION", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, BAD_CODE) +INST3(aesdec, "aesdec", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0xDE), INS_FLAGS_None) // Perform one round of an AES decryption flow +INST3(aesdeclast, "aesdeclast", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0xDF), INS_FLAGS_None) // Perform last round of an AES decryption flow +INST3(aesenc, "aesenc", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0xDC), INS_FLAGS_None) // Perform one round of an AES encryption flow +INST3(aesenclast, "aesenclast", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0xDD), INS_FLAGS_None) // Perform last round of an AES encryption flow +INST3(aesimc, "aesimc", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0xDB), INS_FLAGS_None) // Perform the AES InvMixColumn Transformation +INST3(aeskeygenassist, "aeskeygenassist", IUM_WR, BAD_CODE, BAD_CODE, SSE3A(0xDF), INS_FLAGS_None) // AES Round Key Generation Assist +INST3(LAST_SSE4_INSTRUCTION, "LAST_SSE4_INSTRUCTION", IUM_WR, BAD_CODE, BAD_CODE, BAD_CODE, INS_FLAGS_None) + +INST3(FIRST_AVX_INSTRUCTION, "FIRST_AVX_INSTRUCTION", IUM_WR, BAD_CODE, BAD_CODE, BAD_CODE, INS_FLAGS_None) // AVX only instructions -INST3( vbroadcastss, "broadcastss" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x18)) // Broadcast float value read from memory to entire ymm register -INST3( vbroadcastsd, "broadcastsd" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x19)) // Broadcast float value read from memory to entire ymm register -INST3( vpbroadcastb, "pbroadcastb" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x78)) // Broadcast int8 value from reg/memory to entire ymm register -INST3( vpbroadcastw, "pbroadcastw" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x79)) // Broadcast int16 value from reg/memory to entire ymm register -INST3( vpbroadcastd, "pbroadcastd" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x58)) // Broadcast int32 value from reg/memory to entire ymm register -INST3( vpbroadcastq, "pbroadcastq" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x59)) // Broadcast int64 value from reg/memory to entire ymm register -INST3( vextractf128, "extractf128" , 0, IUM_WR, 0, 0, SSE3A(0x19), BAD_CODE, BAD_CODE) // Extract 128-bit packed floating point values -INST3( vextracti128, "extracti128" , 0, IUM_WR, 0, 0, SSE3A(0x39), BAD_CODE, BAD_CODE) // Extract 128-bit packed integer values -INST3( vinsertf128, "insertf128" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE3A(0x18)) // Insert 128-bit packed floating point values -INST3( vinserti128, "inserti128" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE3A(0x38)) // Insert 128-bit packed integer values -INST3( vzeroupper, "zeroupper" , 0, IUM_WR, 0, 0, 0xC577F8, BAD_CODE, BAD_CODE) // Zero upper 128-bits of all YMM regs (includes 2-byte fixed VEX prefix) -INST3( vperm2i128, "perm2i128" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE3A(0x46)) // Permute 128-bit halves of input register -INST3( vpermq, "permq" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE3A(0x00)) // Permute 64-bit of input register -INST3( vpblendd, "pblendd" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE3A(0x02)) // Blend Packed DWORDs -INST3( vblendvps, "blendvps" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE3A(0x4A)) // Variable Blend Packed Singles -INST3( vblendvpd, "blendvpd" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE3A(0x4B)) // Variable Blend Packed Doubles -INST3( vpblendvb, "pblendvb" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE3A(0x4C)) // Variable Blend Packed Bytes -INST3( vtestps, "testps" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x0E)) // Packed Bit Test -INST3( vtestpd, "testpd" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x0F)) // Packed Bit Test -INST3( vpsrlvd, "psrlvd" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x45)) // Variable Bit Shift Right Logical -INST3( vpsrlvq, "psrlvq" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x45)) // Variable Bit Shift Right Logical -INST3( vpsravd, "psravd" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x46)) // Variable Bit Shift Right Arithmetic -INST3( vpsllvd, "psllvd" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x47)) // Variable Bit Shift Left Logical -INST3( vpsllvq, "psllvq" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x47)) // Variable Bit Shift Left Logical -INST3( vpermilps, "permilps" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE3A(0x04)) // Permute In-Lane of Quadruples of Single-Precision Floating-Point Values -INST3( vpermilpd, "permilpd" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE3A(0x05)) // Permute In-Lane of Quadruples of Double-Precision Floating-Point Values -INST3( vpermilpsvar, "permilpsvar" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x0C)) // Permute In-Lane of Quadruples of Single-Precision Floating-Point Values -INST3( vpermilpdvar, "permilpdvar" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x0D)) // Permute In-Lane of Quadruples of Double-Precision Floating-Point Values -INST3( vperm2f128, "perm2f128" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE3A(0x06)) // Permute Floating-Point Values -INST3(vbroadcastf128,"broadcastf128",0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x1A)) // Broadcast packed float values read from memory to entire ymm register -INST3(vbroadcasti128,"broadcasti128",0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x5A)) // Broadcast packed integer values read from memory to entire ymm register -INST3(vmaskmovps, "maskmovps" ,0, IUM_WR, 0, 0, SSE38(0x2E), BAD_CODE, SSE38(0x2C)) // Conditional SIMD Packed Single-Precision Floating-Point Loads and Stores -INST3(vmaskmovpd, "maskmovpd" ,0, IUM_WR, 0, 0, SSE38(0x2F), BAD_CODE, SSE38(0x2D)) // Conditional SIMD Packed Double-Precision Floating-Point Loads and Stores -INST3(vpmaskmovd, "pmaskmovd" ,0, IUM_WR, 0, 0, SSE38(0x8E), BAD_CODE, SSE38(0x8C)) // Conditional SIMD Integer Packed Dword Loads and Stores -INST3(vpmaskmovq, "pmaskmovq" ,0, IUM_WR, 0, 0, SSE38(0x8E), BAD_CODE, SSE38(0x8C)) // Conditional SIMD Integer Packed Qword Loads and Stores - -INST3(FIRST_FMA_INSTRUCTION, "FIRST_FMA_INSTRUCTION", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, BAD_CODE) -// enum name FP updmode rf wf MR MI RM -INST3(vfmadd132pd, "fmadd132pd", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x98)) // Fused Multiply-Add of Packed Double-Precision Floating-Point Values -INST3(vfmadd213pd, "fmadd213pd", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0xA8)) // -INST3(vfmadd231pd, "fmadd231pd", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0xB8)) // -INST3(vfmadd132ps, "fmadd132ps", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x98)) // Fused Multiply-Add of Packed Single-Precision Floating-Point Values -INST3(vfmadd213ps, "fmadd213ps", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0xA8)) // -INST3(vfmadd231ps, "fmadd231ps", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0xB8)) // -INST3(vfmadd132sd, "fmadd132sd", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x99)) // Fused Multiply-Add of Scalar Double-Precision Floating-Point Values -INST3(vfmadd213sd, "fmadd213sd", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0xA9)) // -INST3(vfmadd231sd, "fmadd231sd", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0xB9)) // -INST3(vfmadd132ss, "fmadd132ss", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x99)) // Fused Multiply-Add of Scalar Single-Precision Floating-Point Values -INST3(vfmadd213ss, "fmadd213ss", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0xA9)) // -INST3(vfmadd231ss, "fmadd231ss", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0xB9)) // -INST3(vfmaddsub132pd, "fmaddsub132pd", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x96)) // Fused Multiply-Alternating Add/Subtract of Packed Double-Precision Floating-Point Values -INST3(vfmaddsub213pd, "fmaddsub213pd", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0xA6)) // -INST3(vfmaddsub231pd, "fmaddsub231pd", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0xB6)) // -INST3(vfmaddsub132ps, "fmaddsub132ps", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x96)) // Fused Multiply-Alternating Add/Subtract of Packed Single-Precision Floating-Point Values -INST3(vfmaddsub213ps, "fmaddsub213ps", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0xA6)) // -INST3(vfmaddsub231ps, "fmaddsub231ps", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0xB6)) // -INST3(vfmsubadd132pd, "fmsubadd132pd", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x97)) // Fused Multiply-Alternating Subtract/Add of Packed Double-Precision Floating-Point Values -INST3(vfmsubadd213pd, "fmsubadd213pd", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0xA7)) // -INST3(vfmsubadd231pd, "fmsubadd231pd", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0xB7)) // -INST3(vfmsubadd132ps, "fmsubadd132ps", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x97)) // Fused Multiply-Alternating Subtract/Add of Packed Single-Precision Floating-Point Values -INST3(vfmsubadd213ps, "fmsubadd213ps", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0xA7)) // -INST3(vfmsubadd231ps, "fmsubadd231ps", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0xB7)) // -INST3(vfmsub132pd, "fmsub132pd", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x9A)) // Fused Multiply-Subtract of Packed Double-Precision Floating-Point Values -INST3(vfmsub213pd, "fmsub213pd", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0xAA)) // -INST3(vfmsub231pd, "fmsub231pd", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0xBA)) // -INST3(vfmsub132ps, "fmsub132ps", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x9A)) // Fused Multiply-Subtract of Packed Single-Precision Floating-Point Values -INST3(vfmsub213ps, "fmsub213ps", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0xAA)) // -INST3(vfmsub231ps, "fmsub231ps", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0xBA)) // -INST3(vfmsub132sd, "fmsub132sd", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x9B)) // Fused Multiply-Subtract of Scalar Double-Precision Floating-Point Values -INST3(vfmsub213sd, "fmsub213sd", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0xAB)) // -INST3(vfmsub231sd, "fmsub231sd", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0xBB)) // -INST3(vfmsub132ss, "fmsub132ss", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x9B)) // Fused Multiply-Subtract of Scalar Single-Precision Floating-Point Values -INST3(vfmsub213ss, "fmsub213ss", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0xAB)) // -INST3(vfmsub231ss, "fmsub231ss", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0xBB)) // -INST3(vfnmadd132pd, "fmnadd132pd", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x9C)) // Fused Negative Multiply-Add of Packed Double-Precision Floating-Point Values -INST3(vfnmadd213pd, "fmnadd213pd", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0xAC)) // -INST3(vfnmadd231pd, "fmnadd231pd", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0xBC)) // -INST3(vfnmadd132ps, "fmnadd132ps", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x9C)) // Fused Negative Multiply-Add of Packed Single-Precision Floating-Point Values -INST3(vfnmadd213ps, "fmnadd213ps", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0xAC)) // -INST3(vfnmadd231ps, "fmnadd231ps", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0xBC)) // -INST3(vfnmadd132sd, "fmnadd132sd", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x9D)) // Fused Negative Multiply-Add of Scalar Double-Precision Floating-Point Values -INST3(vfnmadd213sd, "fmnadd213sd", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0xAD)) // -INST3(vfnmadd231sd, "fmnadd231sd", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0xBD)) // -INST3(vfnmadd132ss, "fmnadd132ss", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x9D)) // Fused Negative Multiply-Add of Scalar Single-Precision Floating-Point Values -INST3(vfnmadd213ss, "fmnadd213ss", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0xAD)) // -INST3(vfnmadd231ss, "fmnadd231ss", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0xBD)) // -INST3(vfnmsub132pd, "fmnsub132pd", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x9E)) // Fused Negative Multiply-Subtract of Packed Double-Precision Floating-Point Values -INST3(vfnmsub213pd, "fmnsub213pd", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0xAE)) // -INST3(vfnmsub231pd, "fmnsub231pd", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0xBE)) // -INST3(vfnmsub132ps, "fmnsub132ps", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x9E)) // Fused Negative Multiply-Subtract of Packed Single-Precision Floating-Point Values -INST3(vfnmsub213ps, "fmnsub213ps", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0xAE)) // -INST3(vfnmsub231ps, "fmnsub231ps", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0xBE)) // -INST3(vfnmsub132sd, "fmnsub132sd", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x9F)) // Fused Negative Multiply-Subtract of Scalar Double-Precision Floating-Point Values -INST3(vfnmsub213sd, "fmnsub213sd", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0xAF)) // -INST3(vfnmsub231sd, "fmnsub231sd", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0xBF)) // -INST3(vfnmsub132ss, "fmnsub132ss", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0x9F)) // Fused Negative Multiply-Subtract of Scalar Single-Precision Floating-Point Values -INST3(vfnmsub213ss, "fmnsub213ss", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0xAF)) // -INST3(vfnmsub231ss, "fmnsub231ss", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0xBF)) // -INST3(LAST_FMA_INSTRUCTION, "LAST_FMA_INSTRUCTION", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, BAD_CODE) +INST3(vbroadcastss, "broadcastss", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x18), INS_FLAGS_None) // Broadcast float value read from memory to entire ymm register +INST3(vbroadcastsd, "broadcastsd", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x19), INS_FLAGS_None) // Broadcast float value read from memory to entire ymm register +INST3(vpbroadcastb, "pbroadcastb", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x78), INS_FLAGS_None) // Broadcast int8 value from reg/memory to entire ymm register +INST3(vpbroadcastw, "pbroadcastw", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x79), INS_FLAGS_None) // Broadcast int16 value from reg/memory to entire ymm register +INST3(vpbroadcastd, "pbroadcastd", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x58), INS_FLAGS_None) // Broadcast int32 value from reg/memory to entire ymm register +INST3(vpbroadcastq, "pbroadcastq", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x59), INS_FLAGS_None) // Broadcast int64 value from reg/memory to entire ymm register +INST3(vextractf128, "extractf128", IUM_WR, SSE3A(0x19), BAD_CODE, BAD_CODE, INS_FLAGS_None) // Extract 128-bit packed floating point values +INST3(vextracti128, "extracti128", IUM_WR, SSE3A(0x39), BAD_CODE, BAD_CODE, INS_FLAGS_None) // Extract 128-bit packed integer values +INST3(vinsertf128, "insertf128", IUM_WR, BAD_CODE, BAD_CODE, SSE3A(0x18), INS_FLAGS_None) // Insert 128-bit packed floating point values +INST3(vinserti128, "inserti128", IUM_WR, BAD_CODE, BAD_CODE, SSE3A(0x38), INS_FLAGS_None) // Insert 128-bit packed integer values +INST3(vzeroupper, "zeroupper", IUM_WR, 0xC577F8, BAD_CODE, BAD_CODE, INS_FLAGS_None) // Zero upper 128-bits of all YMM regs (includes 2-byte fixed VEX prefix) +INST3(vperm2i128, "perm2i128", IUM_WR, BAD_CODE, BAD_CODE, SSE3A(0x46), INS_FLAGS_None) // Permute 128-bit halves of input register +INST3(vpermq, "permq", IUM_WR, BAD_CODE, BAD_CODE, SSE3A(0x00), INS_FLAGS_None) // Permute 64-bit of input register +INST3(vpblendd, "pblendd", IUM_WR, BAD_CODE, BAD_CODE, SSE3A(0x02), INS_FLAGS_None) // Blend Packed DWORDs +INST3(vblendvps, "blendvps", IUM_WR, BAD_CODE, BAD_CODE, SSE3A(0x4A), INS_FLAGS_None) // Variable Blend Packed Singles +INST3(vblendvpd, "blendvpd", IUM_WR, BAD_CODE, BAD_CODE, SSE3A(0x4B), INS_FLAGS_None) // Variable Blend Packed Doubles +INST3(vpblendvb, "pblendvb", IUM_WR, BAD_CODE, BAD_CODE, SSE3A(0x4C), INS_FLAGS_None) // Variable Blend Packed Bytes +INST3(vtestps, "testps", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x0E), INS_FLAGS_None) // Packed Bit Test +INST3(vtestpd, "testpd", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x0F), INS_FLAGS_None) // Packed Bit Test +INST3(vpsrlvd, "psrlvd", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x45), INS_FLAGS_None) // Variable Bit Shift Right Logical +INST3(vpsrlvq, "psrlvq", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x45), INS_FLAGS_None) // Variable Bit Shift Right Logical +INST3(vpsravd, "psravd", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x46), INS_FLAGS_None) // Variable Bit Shift Right Arithmetic +INST3(vpsllvd, "psllvd", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x47), INS_FLAGS_None) // Variable Bit Shift Left Logical +INST3(vpsllvq, "psllvq", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x47), INS_FLAGS_None) // Variable Bit Shift Left Logical +INST3(vpermilps, "permilps", IUM_WR, BAD_CODE, BAD_CODE, SSE3A(0x04), INS_FLAGS_None) // Permute In-Lane of Quadruples of Single-Precision Floating-Point Values +INST3(vpermilpd, "permilpd", IUM_WR, BAD_CODE, BAD_CODE, SSE3A(0x05), INS_FLAGS_None) // Permute In-Lane of Quadruples of Double-Precision Floating-Point Values +INST3(vpermilpsvar, "permilpsvar", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x0C), INS_FLAGS_None) // Permute In-Lane of Quadruples of Single-Precision Floating-Point Values +INST3(vpermilpdvar, "permilpdvar", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x0D), INS_FLAGS_None) // Permute In-Lane of Quadruples of Double-Precision Floating-Point Values +INST3(vperm2f128, "perm2f128", IUM_WR, BAD_CODE, BAD_CODE, SSE3A(0x06), INS_FLAGS_None) // Permute Floating-Point Values +INST3(vbroadcastf128, "broadcastf128", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x1A), INS_FLAGS_None) // Broadcast packed float values read from memory to entire ymm register +INST3(vbroadcasti128, "broadcasti128", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x5A), INS_FLAGS_None) // Broadcast packed integer values read from memory to entire ymm register +INST3(vmaskmovps, "maskmovps", IUM_WR, SSE38(0x2E), BAD_CODE, SSE38(0x2C), INS_FLAGS_None) // Conditional SIMD Packed Single-Precision Floating-Point Loads and Stores +INST3(vmaskmovpd, "maskmovpd", IUM_WR, SSE38(0x2F), BAD_CODE, SSE38(0x2D), INS_FLAGS_None) // Conditional SIMD Packed Double-Precision Floating-Point Loads and Stores +INST3(vpmaskmovd, "pmaskmovd", IUM_WR, SSE38(0x8E), BAD_CODE, SSE38(0x8C), INS_FLAGS_None) // Conditional SIMD Integer Packed Dword Loads and Stores +INST3(vpmaskmovq, "pmaskmovq", IUM_WR, SSE38(0x8E), BAD_CODE, SSE38(0x8C), INS_FLAGS_None) // Conditional SIMD Integer Packed Qword Loads and Stores + +INST3(FIRST_FMA_INSTRUCTION, "FIRST_FMA_INSTRUCTION", IUM_WR, BAD_CODE, BAD_CODE, BAD_CODE, INS_FLAGS_None) +// id nm um mr mi rm flags +INST3(vfmadd132pd, "fmadd132pd", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x98), INS_FLAGS_None) // Fused Multiply-Add of Packed Double-Precision Floating-Point Values +INST3(vfmadd213pd, "fmadd213pd", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0xA8), INS_FLAGS_None) // +INST3(vfmadd231pd, "fmadd231pd", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0xB8), INS_FLAGS_None) // +INST3(vfmadd132ps, "fmadd132ps", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x98), INS_FLAGS_None) // Fused Multiply-Add of Packed Single-Precision Floating-Point Values +INST3(vfmadd213ps, "fmadd213ps", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0xA8), INS_FLAGS_None) // +INST3(vfmadd231ps, "fmadd231ps", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0xB8), INS_FLAGS_None) // +INST3(vfmadd132sd, "fmadd132sd", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x99), INS_FLAGS_None) // Fused Multiply-Add of Scalar Double-Precision Floating-Point Values +INST3(vfmadd213sd, "fmadd213sd", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0xA9), INS_FLAGS_None) // +INST3(vfmadd231sd, "fmadd231sd", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0xB9), INS_FLAGS_None) // +INST3(vfmadd132ss, "fmadd132ss", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x99), INS_FLAGS_None) // Fused Multiply-Add of Scalar Single-Precision Floating-Point Values +INST3(vfmadd213ss, "fmadd213ss", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0xA9), INS_FLAGS_None) // +INST3(vfmadd231ss, "fmadd231ss", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0xB9), INS_FLAGS_None) // +INST3(vfmaddsub132pd, "fmaddsub132pd", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x96), INS_FLAGS_None) // Fused Multiply-Alternating Add/Subtract of Packed Double-Precision Floating-Point Values +INST3(vfmaddsub213pd, "fmaddsub213pd", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0xA6), INS_FLAGS_None) // +INST3(vfmaddsub231pd, "fmaddsub231pd", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0xB6), INS_FLAGS_None) // +INST3(vfmaddsub132ps, "fmaddsub132ps", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x96), INS_FLAGS_None) // Fused Multiply-Alternating Add/Subtract of Packed Single-Precision Floating-Point Values +INST3(vfmaddsub213ps, "fmaddsub213ps", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0xA6), INS_FLAGS_None) // +INST3(vfmaddsub231ps, "fmaddsub231ps", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0xB6), INS_FLAGS_None) // +INST3(vfmsubadd132pd, "fmsubadd132pd", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x97), INS_FLAGS_None) // Fused Multiply-Alternating Subtract/Add of Packed Double-Precision Floating-Point Values +INST3(vfmsubadd213pd, "fmsubadd213pd", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0xA7), INS_FLAGS_None) // +INST3(vfmsubadd231pd, "fmsubadd231pd", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0xB7), INS_FLAGS_None) // +INST3(vfmsubadd132ps, "fmsubadd132ps", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x97), INS_FLAGS_None) // Fused Multiply-Alternating Subtract/Add of Packed Single-Precision Floating-Point Values +INST3(vfmsubadd213ps, "fmsubadd213ps", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0xA7), INS_FLAGS_None) // +INST3(vfmsubadd231ps, "fmsubadd231ps", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0xB7), INS_FLAGS_None) // +INST3(vfmsub132pd, "fmsub132pd", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x9A), INS_FLAGS_None) // Fused Multiply-Subtract of Packed Double-Precision Floating-Point Values +INST3(vfmsub213pd, "fmsub213pd", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0xAA), INS_FLAGS_None) // +INST3(vfmsub231pd, "fmsub231pd", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0xBA), INS_FLAGS_None) // +INST3(vfmsub132ps, "fmsub132ps", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x9A), INS_FLAGS_None) // Fused Multiply-Subtract of Packed Single-Precision Floating-Point Values +INST3(vfmsub213ps, "fmsub213ps", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0xAA), INS_FLAGS_None) // +INST3(vfmsub231ps, "fmsub231ps", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0xBA), INS_FLAGS_None) // +INST3(vfmsub132sd, "fmsub132sd", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x9B), INS_FLAGS_None) // Fused Multiply-Subtract of Scalar Double-Precision Floating-Point Values +INST3(vfmsub213sd, "fmsub213sd", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0xAB), INS_FLAGS_None) // +INST3(vfmsub231sd, "fmsub231sd", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0xBB), INS_FLAGS_None) // +INST3(vfmsub132ss, "fmsub132ss", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x9B), INS_FLAGS_None) // Fused Multiply-Subtract of Scalar Single-Precision Floating-Point Values +INST3(vfmsub213ss, "fmsub213ss", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0xAB), INS_FLAGS_None) // +INST3(vfmsub231ss, "fmsub231ss", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0xBB), INS_FLAGS_None) // +INST3(vfnmadd132pd, "fmnadd132pd", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x9C), INS_FLAGS_None) // Fused Negative Multiply-Add of Packed Double-Precision Floating-Point Values +INST3(vfnmadd213pd, "fmnadd213pd", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0xAC), INS_FLAGS_None) // +INST3(vfnmadd231pd, "fmnadd231pd", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0xBC), INS_FLAGS_None) // +INST3(vfnmadd132ps, "fmnadd132ps", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x9C), INS_FLAGS_None) // Fused Negative Multiply-Add of Packed Single-Precision Floating-Point Values +INST3(vfnmadd213ps, "fmnadd213ps", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0xAC), INS_FLAGS_None) // +INST3(vfnmadd231ps, "fmnadd231ps", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0xBC), INS_FLAGS_None) // +INST3(vfnmadd132sd, "fmnadd132sd", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x9D), INS_FLAGS_None) // Fused Negative Multiply-Add of Scalar Double-Precision Floating-Point Values +INST3(vfnmadd213sd, "fmnadd213sd", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0xAD), INS_FLAGS_None) // +INST3(vfnmadd231sd, "fmnadd231sd", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0xBD), INS_FLAGS_None) // +INST3(vfnmadd132ss, "fmnadd132ss", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x9D), INS_FLAGS_None) // Fused Negative Multiply-Add of Scalar Single-Precision Floating-Point Values +INST3(vfnmadd213ss, "fmnadd213ss", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0xAD), INS_FLAGS_None) // +INST3(vfnmadd231ss, "fmnadd231ss", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0xBD), INS_FLAGS_None) // +INST3(vfnmsub132pd, "fmnsub132pd", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x9E), INS_FLAGS_None) // Fused Negative Multiply-Subtract of Packed Double-Precision Floating-Point Values +INST3(vfnmsub213pd, "fmnsub213pd", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0xAE), INS_FLAGS_None) // +INST3(vfnmsub231pd, "fmnsub231pd", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0xBE), INS_FLAGS_None) // +INST3(vfnmsub132ps, "fmnsub132ps", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x9E), INS_FLAGS_None) // Fused Negative Multiply-Subtract of Packed Single-Precision Floating-Point Values +INST3(vfnmsub213ps, "fmnsub213ps", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0xAE), INS_FLAGS_None) // +INST3(vfnmsub231ps, "fmnsub231ps", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0xBE), INS_FLAGS_None) // +INST3(vfnmsub132sd, "fmnsub132sd", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x9F), INS_FLAGS_None) // Fused Negative Multiply-Subtract of Scalar Double-Precision Floating-Point Values +INST3(vfnmsub213sd, "fmnsub213sd", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0xAF), INS_FLAGS_None) // +INST3(vfnmsub231sd, "fmnsub231sd", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0xBF), INS_FLAGS_None) // +INST3(vfnmsub132ss, "fmnsub132ss", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0x9F), INS_FLAGS_None) // Fused Negative Multiply-Subtract of Scalar Single-Precision Floating-Point Values +INST3(vfnmsub213ss, "fmnsub213ss", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0xAF), INS_FLAGS_None) // +INST3(vfnmsub231ss, "fmnsub231ss", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0xBF), INS_FLAGS_None) // +INST3(LAST_FMA_INSTRUCTION, "LAST_FMA_INSTRUCTION", IUM_WR, BAD_CODE, BAD_CODE, BAD_CODE, INS_FLAGS_None) // BMI1 -INST3(FIRST_BMI_INSTRUCTION, "FIRST_BMI_INSTRUCTION", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, BAD_CODE) -INST3(andn, "andn", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0xF2)) // Logical AND NOT -INST3(blsi, "blsi", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0xF3)) // Extract Lowest Set Isolated Bit -INST3(blsmsk, "blsmsk", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0xF3)) // Get Mask Up to Lowest Set Bit -INST3(blsr, "blsr", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0xF3)) // Reset Lowest Set Bit +INST3(FIRST_BMI_INSTRUCTION, "FIRST_BMI_INSTRUCTION", IUM_WR, BAD_CODE, BAD_CODE, BAD_CODE, INS_FLAGS_None) +INST3(andn, "andn", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0xF2), INS_FLAGS_None) // Logical AND NOT +INST3(blsi, "blsi", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0xF3), INS_FLAGS_None) // Extract Lowest Set Isolated Bit +INST3(blsmsk, "blsmsk", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0xF3), INS_FLAGS_None) // Get Mask Up to Lowest Set Bit +INST3(blsr, "blsr", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0xF3), INS_FLAGS_None) // Reset Lowest Set Bit // BMI2 -INST3(pdep, "pdep", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0xF5)) // Parallel Bits Deposit -INST3(pext, "pext", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSE38(0xF5)) // Parallel Bits Extract -INST3(LAST_BMI_INSTRUCTION, "LAST_BMI_INSTRUCTION", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, BAD_CODE) +INST3(pdep, "pdep", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0xF5), INS_FLAGS_None) // Parallel Bits Deposit +INST3(pext, "pext", IUM_WR, BAD_CODE, BAD_CODE, SSE38(0xF5), INS_FLAGS_None) // Parallel Bits Extract +INST3(LAST_BMI_INSTRUCTION, "LAST_BMI_INSTRUCTION", IUM_WR, BAD_CODE, BAD_CODE, BAD_CODE, INS_FLAGS_None) -INST3(LAST_AVX_INSTRUCTION, "LAST_AVX_INSTRUCTION", 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, BAD_CODE) +INST3(LAST_AVX_INSTRUCTION, "LAST_AVX_INSTRUCTION", IUM_WR, BAD_CODE, BAD_CODE, BAD_CODE, INS_FLAGS_None) // Scalar instructions in SSE4.2 -INST3( crc32, "crc32" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, PACK4(0xF2, 0x0F, 0x38, 0xF0)) +INST3(crc32, "crc32", IUM_WR, BAD_CODE, BAD_CODE, PACK4(0xF2, 0x0F, 0x38, 0xF0), INS_FLAGS_None) // BMI1 -INST3( tzcnt, "tzcnt" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSEFLT(0xBC)) // Count the Number of Trailing Zero Bits +INST3(tzcnt, "tzcnt", IUM_WR, BAD_CODE, BAD_CODE, SSEFLT(0xBC), INS_FLAGS_None) // Count the Number of Trailing Zero Bits // LZCNT -INST3( lzcnt, "lzcnt" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSEFLT(0xBD)) +INST3(lzcnt, "lzcnt", IUM_WR, BAD_CODE, BAD_CODE, SSEFLT(0xBD), INS_FLAGS_None) // POPCNT -INST3( popcnt, "popcnt" , 0, IUM_WR, 0, 0, BAD_CODE, BAD_CODE, SSEFLT(0xB8)) - -// enum name FP updmode rf wf R/M,R/M[reg] R/M,icon - -INST2(ret , "ret" , 0, IUM_RD, 0, 0, 0x0000C3, 0x0000C2) -INST2(loop , "loop" , 0, IUM_RD, 0, 0, BAD_CODE, 0x0000E2) -INST2(call , "call" , 0, IUM_RD, 0, 1, 0x0010FF, 0x0000E8) - -INST2(rol , "rol" , 0, IUM_RW, 0, 1, 0x0000D2, BAD_CODE) -INST2(rol_1 , "rol" , 0, IUM_RW, 0, 1, 0x0000D0, 0x0000D0) -INST2(rol_N , "rol" , 0, IUM_RW, 0, 1, 0x0000C0, 0x0000C0) -INST2(ror , "ror" , 0, IUM_RW, 0, 1, 0x0008D2, BAD_CODE) -INST2(ror_1 , "ror" , 0, IUM_RW, 0, 1, 0x0008D0, 0x0008D0) -INST2(ror_N , "ror" , 0, IUM_RW, 0, 1, 0x0008C0, 0x0008C0) - -INST2(rcl , "rcl" , 0, IUM_RW, 1, 1, 0x0010D2, BAD_CODE) -INST2(rcl_1 , "rcl" , 0, IUM_RW, 1, 1, 0x0010D0, 0x0010D0) -INST2(rcl_N , "rcl" , 0, IUM_RW, 1, 1, 0x0010C0, 0x0010C0) -INST2(rcr , "rcr" , 0, IUM_RW, 1, 1, 0x0018D2, BAD_CODE) -INST2(rcr_1 , "rcr" , 0, IUM_RW, 1, 1, 0x0018D0, 0x0018D0) -INST2(rcr_N , "rcr" , 0, IUM_RW, 1, 1, 0x0018C0, 0x0018C0) -INST2(shl , "shl" , 0, IUM_RW, 0, 1, 0x0020D2, BAD_CODE) -INST2(shl_1 , "shl" , 0, IUM_RW, 0, 1, 0x0020D0, 0x0020D0) -INST2(shl_N , "shl" , 0, IUM_RW, 0, 1, 0x0020C0, 0x0020C0) -INST2(shr , "shr" , 0, IUM_RW, 0, 1, 0x0028D2, BAD_CODE) -INST2(shr_1 , "shr" , 0, IUM_RW, 0, 1, 0x0028D0, 0x0028D0) -INST2(shr_N , "shr" , 0, IUM_RW, 0, 1, 0x0028C0, 0x0028C0) -INST2(sar , "sar" , 0, IUM_RW, 0, 1, 0x0038D2, BAD_CODE) -INST2(sar_1 , "sar" , 0, IUM_RW, 0, 1, 0x0038D0, 0x0038D0) -INST2(sar_N , "sar" , 0, IUM_RW, 0, 1, 0x0038C0, 0x0038C0) - - -// enum name FP updmode rf wf R/M,R/M[reg] - -INST1(r_movsb, "rep movsb" , 0, IUM_RD, 0, 0, 0x00A4F3) -INST1(r_movsd, "rep movsd" , 0, IUM_RD, 0, 0, 0x00A5F3) +INST3(popcnt, "popcnt", IUM_WR, BAD_CODE, BAD_CODE, SSEFLT(0xB8), INS_FLAGS_None) + +// id nm um mr mi flags +INST2(ret, "ret", IUM_RD, 0x0000C3, 0x0000C2, INS_FLAGS_None) +INST2(loop, "loop", IUM_RD, BAD_CODE, 0x0000E2, INS_FLAGS_None) +INST2(call, "call", IUM_RD, 0x0010FF, 0x0000E8, INS_FLAGS_WritesFlags) + +INST2(rol, "rol", IUM_RW, 0x0000D2, BAD_CODE, INS_FLAGS_WritesFlags) +INST2(rol_1, "rol", IUM_RW, 0x0000D0, 0x0000D0, INS_FLAGS_WritesFlags) +INST2(rol_N, "rol", IUM_RW, 0x0000C0, 0x0000C0, INS_FLAGS_WritesFlags) +INST2(ror, "ror", IUM_RW, 0x0008D2, BAD_CODE, INS_FLAGS_WritesFlags) +INST2(ror_1, "ror", IUM_RW, 0x0008D0, 0x0008D0, INS_FLAGS_WritesFlags) +INST2(ror_N, "ror", IUM_RW, 0x0008C0, 0x0008C0, INS_FLAGS_WritesFlags) + +INST2(rcl, "rcl", IUM_RW, 0x0010D2, BAD_CODE, INS_FLAGS_ReadsFlags | INS_FLAGS_WritesFlags) +INST2(rcl_1, "rcl", IUM_RW, 0x0010D0, 0x0010D0, INS_FLAGS_ReadsFlags | INS_FLAGS_WritesFlags) +INST2(rcl_N, "rcl", IUM_RW, 0x0010C0, 0x0010C0, INS_FLAGS_ReadsFlags | INS_FLAGS_WritesFlags) +INST2(rcr, "rcr", IUM_RW, 0x0018D2, BAD_CODE, INS_FLAGS_ReadsFlags | INS_FLAGS_WritesFlags) +INST2(rcr_1, "rcr", IUM_RW, 0x0018D0, 0x0018D0, INS_FLAGS_ReadsFlags | INS_FLAGS_WritesFlags) +INST2(rcr_N, "rcr", IUM_RW, 0x0018C0, 0x0018C0, INS_FLAGS_ReadsFlags | INS_FLAGS_WritesFlags) +INST2(shl, "shl", IUM_RW, 0x0020D2, BAD_CODE, INS_FLAGS_WritesFlags) +INST2(shl_1, "shl", IUM_RW, 0x0020D0, 0x0020D0, INS_FLAGS_WritesFlags) +INST2(shl_N, "shl", IUM_RW, 0x0020C0, 0x0020C0, INS_FLAGS_WritesFlags) +INST2(shr, "shr", IUM_RW, 0x0028D2, BAD_CODE, INS_FLAGS_WritesFlags) +INST2(shr_1, "shr", IUM_RW, 0x0028D0, 0x0028D0, INS_FLAGS_WritesFlags) +INST2(shr_N, "shr", IUM_RW, 0x0028C0, 0x0028C0, INS_FLAGS_WritesFlags) +INST2(sar, "sar", IUM_RW, 0x0038D2, BAD_CODE, INS_FLAGS_WritesFlags) +INST2(sar_1, "sar", IUM_RW, 0x0038D0, 0x0038D0, INS_FLAGS_WritesFlags) +INST2(sar_N, "sar", IUM_RW, 0x0038C0, 0x0038C0, INS_FLAGS_WritesFlags) + + +// id nm um mr flags +INST1(r_movsb, "rep movsb", IUM_RD, 0x00A4F3, INS_FLAGS_None) +INST1(r_movsd, "rep movsd", IUM_RD, 0x00A5F3, INS_FLAGS_None) #if defined(_TARGET_AMD64_) -INST1(r_movsq, "rep movsq" , 0, IUM_RD, 0, 0, 0xF3A548) +INST1(r_movsq, "rep movsq", IUM_RD, 0xF3A548, INS_FLAGS_None) #endif // defined(_TARGET_AMD64_) -INST1(movsb , "movsb" , 0, IUM_RD, 0, 0, 0x0000A4) -INST1(movsd , "movsd" , 0, IUM_RD, 0, 0, 0x0000A5) +INST1(movsb, "movsb", IUM_RD, 0x0000A4, INS_FLAGS_None) +INST1(movsd, "movsd", IUM_RD, 0x0000A5, INS_FLAGS_None) #if defined(_TARGET_AMD64_) -INST1(movsq, "movsq" , 0, IUM_RD, 0, 0, 0x00A548) +INST1(movsq, "movsq", IUM_RD, 0x00A548, INS_FLAGS_None) #endif // defined(_TARGET_AMD64_) -INST1(r_stosb, "rep stosb" , 0, IUM_RD, 0, 0, 0x00AAF3) -INST1(r_stosd, "rep stosd" , 0, IUM_RD, 0, 0, 0x00ABF3) +INST1(r_stosb, "rep stosb", IUM_RD, 0x00AAF3, INS_FLAGS_None) +INST1(r_stosd, "rep stosd", IUM_RD, 0x00ABF3, INS_FLAGS_None) #if defined(_TARGET_AMD64_) -INST1(r_stosq, "rep stosq" , 0, IUM_RD, 0, 0, 0xF3AB48) +INST1(r_stosq, "rep stosq", IUM_RD, 0xF3AB48, INS_FLAGS_None) #endif // defined(_TARGET_AMD64_) -INST1(stosb, "stosb" , 0, IUM_RD, 0, 0, 0x0000AA) -INST1(stosd, "stosd" , 0, IUM_RD, 0, 0, 0x0000AB) +INST1(stosb, "stosb", IUM_RD, 0x0000AA, INS_FLAGS_None) +INST1(stosd, "stosd", IUM_RD, 0x0000AB, INS_FLAGS_None) #if defined(_TARGET_AMD64_) -INST1(stosq, "stosq" , 0, IUM_RD, 0, 0, 0x00AB48) +INST1(stosq, "stosq", IUM_RD, 0x00AB48, INS_FLAGS_None) #endif // defined(_TARGET_AMD64_) -INST1(int3 , "int3" , 0, IUM_RD, 0, 0, 0x0000CC) -INST1(nop , "nop" , 0, IUM_RD, 0, 0, 0x000090) -INST1(lock , "lock" , 0, IUM_RD, 0, 0, 0x0000F0) -INST1(leave , "leave" , 0, IUM_RD, 0, 0, 0x0000C9) +INST1(int3, "int3", IUM_RD, 0x0000CC, INS_FLAGS_None) +INST1(nop, "nop", IUM_RD, 0x000090, INS_FLAGS_None) +INST1(lock, "lock", IUM_RD, 0x0000F0, INS_FLAGS_None) +INST1(leave, "leave", IUM_RD, 0x0000C9, INS_FLAGS_None) -INST1(neg , "neg" , 0, IUM_RW, 0, 1, 0x0018F6) -INST1(not , "not" , 0, IUM_RW, 0, 1, 0x0010F6) +INST1(neg, "neg", IUM_RW, 0x0018F6, INS_FLAGS_WritesFlags) +INST1(not, "not", IUM_RW, 0x0010F6, INS_FLAGS_WritesFlags) -INST1(cdq , "cdq" , 0, IUM_RD, 0, 1, 0x000099) -INST1(idiv , "idiv" , 0, IUM_RD, 0, 1, 0x0038F6) -INST1(imulEAX, "imul" , 0, IUM_RD, 0, 1, 0x0028F6) // edx:eax = eax*op1 -INST1(div , "div" , 0, IUM_RD, 0, 1, 0x0030F6) -INST1(mulEAX , "mul" , 0, IUM_RD, 0, 1, 0x0020F6) +INST1(cdq, "cdq", IUM_RD, 0x000099, INS_FLAGS_WritesFlags) +INST1(idiv, "idiv", IUM_RD, 0x0038F6, INS_FLAGS_WritesFlags) +INST1(imulEAX, "imul", IUM_RD, 0x0028F6, INS_FLAGS_WritesFlags) // edx:eax = eax*op1 +INST1(div, "div", IUM_RD, 0x0030F6, INS_FLAGS_WritesFlags) +INST1(mulEAX, "mul", IUM_RD, 0x0020F6, INS_FLAGS_WritesFlags) -INST1(sahf , "sahf" , 0, IUM_RD, 0, 1, 0x00009E) +INST1(sahf, "sahf", IUM_RD, 0x00009E, INS_FLAGS_WritesFlags) -INST1(xadd , "xadd" , 0, IUM_RW, 0, 1, 0x0F00C0) -INST1(cmpxchg, "cmpxchg" , 0, IUM_RW, 0, 1, 0x0F00B0) +INST1(xadd, "xadd", IUM_RW, 0x0F00C0, INS_FLAGS_WritesFlags) +INST1(cmpxchg, "cmpxchg", IUM_RW, 0x0F00B0, INS_FLAGS_WritesFlags) -INST1(shld , "shld" , 0, IUM_RW, 0, 1, 0x0F00A4) -INST1(shrd , "shrd" , 0, IUM_RW, 0, 1, 0x0F00AC) +INST1(shld, "shld", IUM_RW, 0x0F00A4, INS_FLAGS_WritesFlags) +INST1(shrd, "shrd", IUM_RW, 0x0F00AC, INS_FLAGS_WritesFlags) // For RyuJIT/x86, we follow the x86 calling convention that requires // us to return floating point value on the x87 FP stack, so we need // these instructions regardless of whether we're using full stack fp. #ifdef _TARGET_X86_ -INST1(fld , "fld" , 1, IUM_WR, 0, 0, 0x0000D9) -INST1(fstp , "fstp" , 1, IUM_WR, 0, 0, 0x0018D9) +INST1(fld, "fld", IUM_WR, 0x0000D9, INS_FLAGS_x87Instr) +INST1(fstp, "fstp", IUM_WR, 0x0018D9, INS_FLAGS_x87Instr) #endif // _TARGET_X86 -INST1(seto , "seto" , 0, IUM_WR, 1, 0, 0x0F0090) -INST1(setno , "setno" , 0, IUM_WR, 1, 0, 0x0F0091) -INST1(setb , "setb" , 0, IUM_WR, 1, 0, 0x0F0092) -INST1(setae , "setae" , 0, IUM_WR, 1, 0, 0x0F0093) -INST1(sete , "sete" , 0, IUM_WR, 1, 0, 0x0F0094) -INST1(setne , "setne" , 0, IUM_WR, 1, 0, 0x0F0095) -INST1(setbe , "setbe" , 0, IUM_WR, 1, 0, 0x0F0096) -INST1(seta , "seta" , 0, IUM_WR, 1, 0, 0x0F0097) -INST1(sets , "sets" , 0, IUM_WR, 1, 0, 0x0F0098) -INST1(setns , "setns" , 0, IUM_WR, 1, 0, 0x0F0099) -INST1(setpe , "setpe" , 0, IUM_WR, 1, 0, 0x0F009A) -INST1(setpo , "setpo" , 0, IUM_WR, 1, 0, 0x0F009B) -INST1(setl , "setl" , 0, IUM_WR, 1, 0, 0x0F009C) -INST1(setge , "setge" , 0, IUM_WR, 1, 0, 0x0F009D) -INST1(setle , "setle" , 0, IUM_WR, 1, 0, 0x0F009E) -INST1(setg , "setg" , 0, IUM_WR, 1, 0, 0x0F009F) +INST1(seto, "seto", IUM_WR, 0x0F0090, INS_FLAGS_ReadsFlags) +INST1(setno, "setno", IUM_WR, 0x0F0091, INS_FLAGS_ReadsFlags) +INST1(setb, "setb", IUM_WR, 0x0F0092, INS_FLAGS_ReadsFlags) +INST1(setae, "setae", IUM_WR, 0x0F0093, INS_FLAGS_ReadsFlags) +INST1(sete, "sete", IUM_WR, 0x0F0094, INS_FLAGS_ReadsFlags) +INST1(setne, "setne", IUM_WR, 0x0F0095, INS_FLAGS_ReadsFlags) +INST1(setbe, "setbe", IUM_WR, 0x0F0096, INS_FLAGS_ReadsFlags) +INST1(seta, "seta", IUM_WR, 0x0F0097, INS_FLAGS_ReadsFlags) +INST1(sets, "sets", IUM_WR, 0x0F0098, INS_FLAGS_ReadsFlags) +INST1(setns, "setns", IUM_WR, 0x0F0099, INS_FLAGS_ReadsFlags) +INST1(setpe, "setpe", IUM_WR, 0x0F009A, INS_FLAGS_ReadsFlags) +INST1(setpo, "setpo", IUM_WR, 0x0F009B, INS_FLAGS_ReadsFlags) +INST1(setl, "setl", IUM_WR, 0x0F009C, INS_FLAGS_ReadsFlags) +INST1(setge, "setge", IUM_WR, 0x0F009D, INS_FLAGS_ReadsFlags) +INST1(setle, "setle", IUM_WR, 0x0F009E, INS_FLAGS_ReadsFlags) +INST1(setg, "setg", IUM_WR, 0x0F009F, INS_FLAGS_ReadsFlags) #ifdef _TARGET_AMD64_ // A jump with rex prefix. This is used for register indirect // tail calls. -INST1(rex_jmp, "rex.jmp" , 0, IUM_RD, 0, 0, 0x0020FE) +INST1(rex_jmp, "rex.jmp", IUM_RD, 0x0020FE, INS_FLAGS_None) #endif -INST1(i_jmp , "jmp" , 0, IUM_RD, 0, 0, 0x0020FE) - -INST0(jmp , "jmp" , 0, IUM_RD, 0, 0, 0x0000EB) -INST0(jo , "jo" , 0, IUM_RD, 1, 0, 0x000070) -INST0(jno , "jno" , 0, IUM_RD, 1, 0, 0x000071) -INST0(jb , "jb" , 0, IUM_RD, 1, 0, 0x000072) -INST0(jae , "jae" , 0, IUM_RD, 1, 0, 0x000073) -INST0(je , "je" , 0, IUM_RD, 1, 0, 0x000074) -INST0(jne , "jne" , 0, IUM_RD, 1, 0, 0x000075) -INST0(jbe , "jbe" , 0, IUM_RD, 1, 0, 0x000076) -INST0(ja , "ja" , 0, IUM_RD, 1, 0, 0x000077) -INST0(js , "js" , 0, IUM_RD, 1, 0, 0x000078) -INST0(jns , "jns" , 0, IUM_RD, 1, 0, 0x000079) -INST0(jpe , "jpe" , 0, IUM_RD, 1, 0, 0x00007A) -INST0(jpo , "jpo" , 0, IUM_RD, 1, 0, 0x00007B) -INST0(jl , "jl" , 0, IUM_RD, 1, 0, 0x00007C) -INST0(jge , "jge" , 0, IUM_RD, 1, 0, 0x00007D) -INST0(jle , "jle" , 0, IUM_RD, 1, 0, 0x00007E) -INST0(jg , "jg" , 0, IUM_RD, 1, 0, 0x00007F) - -INST0(l_jmp , "jmp" , 0, IUM_RD, 0, 0, 0x0000E9) -INST0(l_jo , "jo" , 0, IUM_RD, 1, 0, 0x00800F) -INST0(l_jno , "jno" , 0, IUM_RD, 1, 0, 0x00810F) -INST0(l_jb , "jb" , 0, IUM_RD, 1, 0, 0x00820F) -INST0(l_jae , "jae" , 0, IUM_RD, 1, 0, 0x00830F) -INST0(l_je , "je" , 0, IUM_RD, 1, 0, 0x00840F) -INST0(l_jne , "jne" , 0, IUM_RD, 1, 0, 0x00850F) -INST0(l_jbe , "jbe" , 0, IUM_RD, 1, 0, 0x00860F) -INST0(l_ja , "ja" , 0, IUM_RD, 1, 0, 0x00870F) -INST0(l_js , "js" , 0, IUM_RD, 1, 0, 0x00880F) -INST0(l_jns , "jns" , 0, IUM_RD, 1, 0, 0x00890F) -INST0(l_jpe , "jpe" , 0, IUM_RD, 1, 0, 0x008A0F) -INST0(l_jpo , "jpo" , 0, IUM_RD, 1, 0, 0x008B0F) -INST0(l_jl , "jl" , 0, IUM_RD, 1, 0, 0x008C0F) -INST0(l_jge , "jge" , 0, IUM_RD, 1, 0, 0x008D0F) -INST0(l_jle , "jle" , 0, IUM_RD, 1, 0, 0x008E0F) -INST0(l_jg , "jg" , 0, IUM_RD, 1, 0, 0x008F0F) - -INST0(align , "align" , 0, IUM_RD, 0, 0, BAD_CODE) +INST1(i_jmp, "jmp", IUM_RD, 0x0020FE, INS_FLAGS_None) + +INST0(jmp, "jmp", IUM_RD, 0x0000EB, INS_FLAGS_None) +INST0(jo, "jo", IUM_RD, 0x000070, INS_FLAGS_ReadsFlags) +INST0(jno, "jno", IUM_RD, 0x000071, INS_FLAGS_ReadsFlags) +INST0(jb, "jb", IUM_RD, 0x000072, INS_FLAGS_ReadsFlags) +INST0(jae, "jae", IUM_RD, 0x000073, INS_FLAGS_ReadsFlags) +INST0(je, "je", IUM_RD, 0x000074, INS_FLAGS_ReadsFlags) +INST0(jne, "jne", IUM_RD, 0x000075, INS_FLAGS_ReadsFlags) +INST0(jbe, "jbe", IUM_RD, 0x000076, INS_FLAGS_ReadsFlags) +INST0(ja, "ja", IUM_RD, 0x000077, INS_FLAGS_ReadsFlags) +INST0(js, "js", IUM_RD, 0x000078, INS_FLAGS_ReadsFlags) +INST0(jns, "jns", IUM_RD, 0x000079, INS_FLAGS_ReadsFlags) +INST0(jpe, "jpe", IUM_RD, 0x00007A, INS_FLAGS_ReadsFlags) +INST0(jpo, "jpo", IUM_RD, 0x00007B, INS_FLAGS_ReadsFlags) +INST0(jl, "jl", IUM_RD, 0x00007C, INS_FLAGS_ReadsFlags) +INST0(jge, "jge", IUM_RD, 0x00007D, INS_FLAGS_ReadsFlags) +INST0(jle, "jle", IUM_RD, 0x00007E, INS_FLAGS_ReadsFlags) +INST0(jg, "jg", IUM_RD, 0x00007F, INS_FLAGS_ReadsFlags) + +INST0(l_jmp, "jmp", IUM_RD, 0x0000E9, INS_FLAGS_None) +INST0(l_jo, "jo", IUM_RD, 0x00800F, INS_FLAGS_ReadsFlags) +INST0(l_jno, "jno", IUM_RD, 0x00810F, INS_FLAGS_ReadsFlags) +INST0(l_jb, "jb", IUM_RD, 0x00820F, INS_FLAGS_ReadsFlags) +INST0(l_jae, "jae", IUM_RD, 0x00830F, INS_FLAGS_ReadsFlags) +INST0(l_je, "je", IUM_RD, 0x00840F, INS_FLAGS_ReadsFlags) +INST0(l_jne, "jne", IUM_RD, 0x00850F, INS_FLAGS_ReadsFlags) +INST0(l_jbe, "jbe", IUM_RD, 0x00860F, INS_FLAGS_ReadsFlags) +INST0(l_ja, "ja", IUM_RD, 0x00870F, INS_FLAGS_ReadsFlags) +INST0(l_js, "js", IUM_RD, 0x00880F, INS_FLAGS_ReadsFlags) +INST0(l_jns, "jns", IUM_RD, 0x00890F, INS_FLAGS_ReadsFlags) +INST0(l_jpe, "jpe", IUM_RD, 0x008A0F, INS_FLAGS_ReadsFlags) +INST0(l_jpo, "jpo", IUM_RD, 0x008B0F, INS_FLAGS_ReadsFlags) +INST0(l_jl, "jl", IUM_RD, 0x008C0F, INS_FLAGS_ReadsFlags) +INST0(l_jge, "jge", IUM_RD, 0x008D0F, INS_FLAGS_ReadsFlags) +INST0(l_jle, "jle", IUM_RD, 0x008E0F, INS_FLAGS_ReadsFlags) +INST0(l_jg, "jg", IUM_RD, 0x008F0F, INS_FLAGS_ReadsFlags) + +INST0(align, "align", IUM_RD, BAD_CODE, INS_FLAGS_None) /*****************************************************************************/ #undef INST0 -- 2.7.4