X86InstrInfo.td

//===----------------------------------------------------------------------===//
// 
//                     The LLVM Compiler Infrastructure
//
// This file is distributed under the University of Illinois Open Source
// License. See LICENSE.TXT for details.
// 
//===----------------------------------------------------------------------===//
//
// This file describes the X86 instruction set, defining the instructions, and
// properties of the instructions which are needed for code generation, machine
// code emission, and analysis.
//
//===----------------------------------------------------------------------===//

//===----------------------------------------------------------------------===//
// X86 specific DAG Nodes.
//

def SDTIntShiftDOp: SDTypeProfile<1, 3,
                                  [SDTCisSameAs<0, 1>, SDTCisSameAs<0, 2>,
                                   SDTCisInt<0>, SDTCisInt<3>]>;

def SDTX86CmpTest : SDTypeProfile<1, 2, [SDTCisVT<0, i32>, SDTCisSameAs<1, 2>]>;

def SDTX86Cmov    : SDTypeProfile<1, 4,
                                  [SDTCisSameAs<0, 1>, SDTCisSameAs<1, 2>,
                                   SDTCisVT<3, i8>, SDTCisVT<4, i32>]>;

// Unary and binary operator instructions that set EFLAGS as a side-effect.
def SDTUnaryArithWithFlags : SDTypeProfile<2, 1,
                                           [SDTCisInt<0>, SDTCisVT<1, i32>]>;

def SDTBinaryArithWithFlags : SDTypeProfile<2, 2,
                                            [SDTCisSameAs<0, 2>,
                                             SDTCisSameAs<0, 3>,
                                             SDTCisInt<0>, SDTCisVT<1, i32>]>;
def SDTX86BrCond  : SDTypeProfile<0, 3,
                                  [SDTCisVT<0, OtherVT>,
                                   SDTCisVT<1, i8>, SDTCisVT<2, i32>]>;

def SDTX86SetCC   : SDTypeProfile<1, 2,
                                  [SDTCisVT<0, i8>,
                                   SDTCisVT<1, i8>, SDTCisVT<2, i32>]>;
def SDTX86SetCC_C : SDTypeProfile<1, 2,
                                  [SDTCisInt<0>,
                                   SDTCisVT<1, i8>, SDTCisVT<2, i32>]>;

def SDTX86cas : SDTypeProfile<0, 3, [SDTCisPtrTy<0>, SDTCisInt<1>, 
                                     SDTCisVT<2, i8>]>;
def SDTX86cas8 : SDTypeProfile<0, 1, [SDTCisPtrTy<0>]>;

def SDTX86atomicBinary : SDTypeProfile<2, 3, [SDTCisInt<0>, SDTCisInt<1>,
                                SDTCisPtrTy<2>, SDTCisInt<3>,SDTCisInt<4>]>;
def SDTX86Ret     : SDTypeProfile<0, -1, [SDTCisVT<0, i16>]>;

def SDT_X86CallSeqStart : SDCallSeqStart<[SDTCisVT<0, i32>]>;
def SDT_X86CallSeqEnd   : SDCallSeqEnd<[SDTCisVT<0, i32>,
                                        SDTCisVT<1, i32>]>;

def SDT_X86Call   : SDTypeProfile<0, -1, [SDTCisVT<0, iPTR>]>;

def SDT_X86VASTART_SAVE_XMM_REGS : SDTypeProfile<0, -1, [SDTCisVT<0, i8>,
                                                         SDTCisVT<1, iPTR>,
                                                         SDTCisVT<2, iPTR>]>;

def SDTX86RepStr  : SDTypeProfile<0, 1, [SDTCisVT<0, OtherVT>]>;

def SDTX86Void    : SDTypeProfile<0, 0, []>;

def SDTX86Wrapper : SDTypeProfile<1, 1, [SDTCisSameAs<0, 1>, SDTCisPtrTy<0>]>;

def SDT_X86TLSADDR : SDTypeProfile<0, 1, [SDTCisInt<0>]>;

def SDT_X86TLSCALL : SDTypeProfile<0, 1, [SDTCisPtrTy<0>]>;

def SDT_X86SegmentBaseAddress : SDTypeProfile<1, 1, [SDTCisPtrTy<0>]>;

def SDT_X86EHRET : SDTypeProfile<0, 1, [SDTCisInt<0>]>;

def SDT_X86TCRET : SDTypeProfile<0, 2, [SDTCisPtrTy<0>, SDTCisVT<1, i32>]>;

def X86bsf     : SDNode<"X86ISD::BSF",      SDTUnaryArithWithFlags>;
def X86bsr     : SDNode<"X86ISD::BSR",      SDTUnaryArithWithFlags>;
def X86shld    : SDNode<"X86ISD::SHLD",     SDTIntShiftDOp>;
def X86shrd    : SDNode<"X86ISD::SHRD",     SDTIntShiftDOp>;

def X86cmp     : SDNode<"X86ISD::CMP" ,     SDTX86CmpTest>;
def X86bt      : SDNode<"X86ISD::BT",       SDTX86CmpTest>;

def X86cmov    : SDNode<"X86ISD::CMOV",     SDTX86Cmov>;
def X86brcond  : SDNode<"X86ISD::BRCOND",   SDTX86BrCond,
                        [SDNPHasChain]>;
def X86setcc   : SDNode<"X86ISD::SETCC",    SDTX86SetCC>;
def X86setcc_c : SDNode<"X86ISD::SETCC_CARRY", SDTX86SetCC_C>;

def X86cas : SDNode<"X86ISD::LCMPXCHG_DAG", SDTX86cas,
                        [SDNPHasChain, SDNPInFlag, SDNPOutFlag, SDNPMayStore,
                         SDNPMayLoad]>;
def X86cas8 : SDNode<"X86ISD::LCMPXCHG8_DAG", SDTX86cas8,
                        [SDNPHasChain, SDNPInFlag, SDNPOutFlag, SDNPMayStore,
                         SDNPMayLoad]>;
def X86AtomAdd64 : SDNode<"X86ISD::ATOMADD64_DAG", SDTX86atomicBinary,
                        [SDNPHasChain, SDNPMayStore, 
                         SDNPMayLoad, SDNPMemOperand]>;
def X86AtomSub64 : SDNode<"X86ISD::ATOMSUB64_DAG", SDTX86atomicBinary,
                        [SDNPHasChain, SDNPMayStore, 
                         SDNPMayLoad, SDNPMemOperand]>;
def X86AtomOr64 : SDNode<"X86ISD::ATOMOR64_DAG", SDTX86atomicBinary,
                        [SDNPHasChain, SDNPMayStore, 
                         SDNPMayLoad, SDNPMemOperand]>;
def X86AtomXor64 : SDNode<"X86ISD::ATOMXOR64_DAG", SDTX86atomicBinary,
                        [SDNPHasChain, SDNPMayStore, 
                         SDNPMayLoad, SDNPMemOperand]>;
def X86AtomAnd64 : SDNode<"X86ISD::ATOMAND64_DAG", SDTX86atomicBinary,
                        [SDNPHasChain, SDNPMayStore, 
                         SDNPMayLoad, SDNPMemOperand]>;
def X86AtomNand64 : SDNode<"X86ISD::ATOMNAND64_DAG", SDTX86atomicBinary,
                        [SDNPHasChain, SDNPMayStore, 
                         SDNPMayLoad, SDNPMemOperand]>;
def X86AtomSwap64 : SDNode<"X86ISD::ATOMSWAP64_DAG", SDTX86atomicBinary,
                        [SDNPHasChain, SDNPMayStore, 
                         SDNPMayLoad, SDNPMemOperand]>;
def X86retflag : SDNode<"X86ISD::RET_FLAG", SDTX86Ret,
                        [SDNPHasChain, SDNPOptInFlag, SDNPVariadic]>;

def X86vastart_save_xmm_regs :
                 SDNode<"X86ISD::VASTART_SAVE_XMM_REGS",
                        SDT_X86VASTART_SAVE_XMM_REGS,
                        [SDNPHasChain, SDNPVariadic]>;

def X86callseq_start :
                 SDNode<"ISD::CALLSEQ_START", SDT_X86CallSeqStart,
                        [SDNPHasChain, SDNPOutFlag]>;
def X86callseq_end :
                 SDNode<"ISD::CALLSEQ_END",   SDT_X86CallSeqEnd,
                        [SDNPHasChain, SDNPOptInFlag, SDNPOutFlag]>;       

def X86call    : SDNode<"X86ISD::CALL",     SDT_X86Call,
                        [SDNPHasChain, SDNPOutFlag, SDNPOptInFlag,
                         SDNPVariadic]>;

def X86rep_stos: SDNode<"X86ISD::REP_STOS", SDTX86RepStr,
                        [SDNPHasChain, SDNPInFlag, SDNPOutFlag, SDNPMayStore]>;
def X86rep_movs: SDNode<"X86ISD::REP_MOVS", SDTX86RepStr,
                        [SDNPHasChain, SDNPInFlag, SDNPOutFlag, SDNPMayStore,
                         SDNPMayLoad]>;

def X86rdtsc   : SDNode<"X86ISD::RDTSC_DAG", SDTX86Void,
                        [SDNPHasChain, SDNPOutFlag, SDNPSideEffect]>;

def X86Wrapper    : SDNode<"X86ISD::Wrapper",     SDTX86Wrapper>;
def X86WrapperRIP : SDNode<"X86ISD::WrapperRIP",  SDTX86Wrapper>;

def X86tlsaddr : SDNode<"X86ISD::TLSADDR", SDT_X86TLSADDR,
                        [SDNPHasChain, SDNPOptInFlag, SDNPOutFlag]>;
def X86SegmentBaseAddress : SDNode<"X86ISD::SegmentBaseAddress",
                                 SDT_X86SegmentBaseAddress, []>;

def X86ehret : SDNode<"X86ISD::EH_RETURN", SDT_X86EHRET,
                        [SDNPHasChain]>;

def X86tcret : SDNode<"X86ISD::TC_RETURN", SDT_X86TCRET, 
                        [SDNPHasChain,  SDNPOptInFlag, SDNPVariadic]>;

def X86add_flag  : SDNode<"X86ISD::ADD",  SDTBinaryArithWithFlags,
                          [SDNPCommutative]>;
def X86sub_flag  : SDNode<"X86ISD::SUB",  SDTBinaryArithWithFlags>;
def X86smul_flag : SDNode<"X86ISD::SMUL", SDTBinaryArithWithFlags,
                          [SDNPCommutative]>;
def X86umul_flag : SDNode<"X86ISD::UMUL", SDTUnaryArithWithFlags,
                          [SDNPCommutative]>;
                          
def X86inc_flag  : SDNode<"X86ISD::INC",  SDTUnaryArithWithFlags>;
def X86dec_flag  : SDNode<"X86ISD::DEC",  SDTUnaryArithWithFlags>;
def X86or_flag   : SDNode<"X86ISD::OR",   SDTBinaryArithWithFlags,
                          [SDNPCommutative]>;
def X86xor_flag  : SDNode<"X86ISD::XOR",  SDTBinaryArithWithFlags,
                          [SDNPCommutative]>;
def X86and_flag  : SDNode<"X86ISD::AND",  SDTBinaryArithWithFlags,
                          [SDNPCommutative]>;

def X86mul_imm : SDNode<"X86ISD::MUL_IMM", SDTIntBinOp>;

def X86MingwAlloca : SDNode<"X86ISD::MINGW_ALLOCA", SDTX86Void,
                            [SDNPHasChain, SDNPInFlag, SDNPOutFlag]>;
                            
def X86TLSCall : SDNode<"X86ISD::TLSCALL", SDT_X86TLSCALL,
                        []>;

//===----------------------------------------------------------------------===//
// X86 Operand Definitions.
//

// A version of ptr_rc which excludes SP, ESP, and RSP. This is used for
// the index operand of an address, to conform to x86 encoding restrictions.
def ptr_rc_nosp : PointerLikeRegClass<1>;

// *mem - Operand definitions for the funky X86 addressing mode operands.
//
def X86MemAsmOperand : AsmOperandClass {
  let Name = "Mem";
  let SuperClasses = [];
}
def X86NoSegMemAsmOperand : AsmOperandClass {
  let Name = "NoSegMem";
  let SuperClasses = [X86MemAsmOperand];
}
def X86AbsMemAsmOperand : AsmOperandClass {
  let Name = "AbsMem";
  let SuperClasses = [X86NoSegMemAsmOperand];
}
class X86MemOperand<string printMethod> : Operand<iPTR> {
  let PrintMethod = printMethod;
  let MIOperandInfo = (ops ptr_rc, i8imm, ptr_rc_nosp, i32imm, i8imm);
  let ParserMatchClass = X86MemAsmOperand;
}

def opaque32mem : X86MemOperand<"printopaquemem">;
def opaque48mem : X86MemOperand<"printopaquemem">;
def opaque80mem : X86MemOperand<"printopaquemem">;
def opaque512mem : X86MemOperand<"printopaquemem">;

def i8mem   : X86MemOperand<"printi8mem">;
def i16mem  : X86MemOperand<"printi16mem">;
def i32mem  : X86MemOperand<"printi32mem">;
def i64mem  : X86MemOperand<"printi64mem">;
def i128mem : X86MemOperand<"printi128mem">;
//def i256mem : X86MemOperand<"printi256mem">;
def f32mem  : X86MemOperand<"printf32mem">;
def f64mem  : X86MemOperand<"printf64mem">;
def f80mem  : X86MemOperand<"printf80mem">;
def f128mem : X86MemOperand<"printf128mem">;
//def f256mem : X86MemOperand<"printf256mem">;

// A version of i8mem for use on x86-64 that uses GR64_NOREX instead of
// plain GR64, so that it doesn't potentially require a REX prefix.
def i8mem_NOREX : Operand<i64> {
  let PrintMethod = "printi8mem";
  let MIOperandInfo = (ops GR64_NOREX, i8imm, GR64_NOREX_NOSP, i32imm, i8imm);
  let ParserMatchClass = X86MemAsmOperand;
}

// Special i32mem for addresses of load folding tail calls. These are not
// allowed to use callee-saved registers since they must be scheduled
// after callee-saved register are popped.
def i32mem_TC : Operand<i32> {
  let PrintMethod = "printi32mem";
  let MIOperandInfo = (ops GR32_TC, i8imm, GR32_TC, i32imm, i8imm);
  let ParserMatchClass = X86MemAsmOperand;
}

def lea32mem : Operand<i32> {
  let PrintMethod = "printlea32mem";
  let MIOperandInfo = (ops GR32, i8imm, GR32_NOSP, i32imm);
  let ParserMatchClass = X86NoSegMemAsmOperand;
}

let ParserMatchClass = X86AbsMemAsmOperand,
    PrintMethod = "print_pcrel_imm" in {
def i32imm_pcrel : Operand<i32>;

def offset8 : Operand<i64>;
def offset16 : Operand<i64>;
def offset32 : Operand<i64>;
def offset64 : Operand<i64>;

// Branch targets have OtherVT type and print as pc-relative values.
def brtarget : Operand<OtherVT>;
def brtarget8 : Operand<OtherVT>;

}

def SSECC : Operand<i8> {
  let PrintMethod = "printSSECC";
}

class ImmSExtAsmOperandClass : AsmOperandClass {
  let SuperClasses = [ImmAsmOperand];
  let RenderMethod = "addImmOperands";
}

// Sign-extended immediate classes. We don't need to define the full lattice
// here because there is no instruction with an ambiguity between ImmSExti64i32
// and ImmSExti32i8.
//
// The strange ranges come from the fact that the assembler always works with
// 64-bit immediates, but for a 16-bit target value we want to accept both "-1"
// (which will be a -1ULL), and "0xFF" (-1 in 16-bits).

// [0, 0x7FFFFFFF]                                            | [0xFFFFFFFF80000000, 0xFFFFFFFFFFFFFFFF]
def ImmSExti64i32AsmOperand : ImmSExtAsmOperandClass {
  let Name = "ImmSExti64i32";
}

// [0, 0x0000007F] | [0x000000000000FF80, 0x000000000000FFFF] | [0xFFFFFFFFFFFFFF80, 0xFFFFFFFFFFFFFFFF]
def ImmSExti16i8AsmOperand : ImmSExtAsmOperandClass {
  let Name = "ImmSExti16i8";
  let SuperClasses = [ImmSExti64i32AsmOperand];
}

// [0, 0x0000007F] | [0x00000000FFFFFF80, 0x00000000FFFFFFFF] | [0xFFFFFFFFFFFFFF80, 0xFFFFFFFFFFFFFFFF]
def ImmSExti32i8AsmOperand : ImmSExtAsmOperandClass {
  let Name = "ImmSExti32i8";
}

// [0, 0x0000007F]                                            | [0xFFFFFFFFFFFFFF80, 0xFFFFFFFFFFFFFFFF]
def ImmSExti64i8AsmOperand : ImmSExtAsmOperandClass {
  let Name = "ImmSExti64i8";
  let SuperClasses = [ImmSExti16i8AsmOperand, ImmSExti32i8AsmOperand, ImmSExti64i32AsmOperand];
}

// A couple of more descriptive operand definitions.
// 16-bits but only 8 bits are significant.
def i16i8imm  : Operand<i16> {
  let ParserMatchClass = ImmSExti16i8AsmOperand;
}
// 32-bits but only 8 bits are significant.
def i32i8imm  : Operand<i32> {
  let ParserMatchClass = ImmSExti32i8AsmOperand;
}

//===----------------------------------------------------------------------===//
// X86 Complex Pattern Definitions.
//

// Define X86 specific addressing mode.
def addr      : ComplexPattern<iPTR, 5, "SelectAddr", [], []>;
def lea32addr : ComplexPattern<i32, 4, "SelectLEAAddr",
                               [add, sub, mul, X86mul_imm, shl, or, frameindex],
                               []>;
def tls32addr : ComplexPattern<i32, 4, "SelectTLSADDRAddr",
                               [tglobaltlsaddr], []>;

//===----------------------------------------------------------------------===//
// X86 Instruction Predicate Definitions.
def HasCMov      : Predicate<"Subtarget->hasCMov()">;
def NoCMov       : Predicate<"!Subtarget->hasCMov()">;
def HasMMX       : Predicate<"Subtarget->hasMMX()">;
def HasSSE1      : Predicate<"Subtarget->hasSSE1()">;
def HasSSE2      : Predicate<"Subtarget->hasSSE2()">;
def HasSSE3      : Predicate<"Subtarget->hasSSE3()">;
def HasSSSE3     : Predicate<"Subtarget->hasSSSE3()">;
def HasSSE41     : Predicate<"Subtarget->hasSSE41()">;
def HasSSE42     : Predicate<"Subtarget->hasSSE42()">;
def HasSSE4A     : Predicate<"Subtarget->hasSSE4A()">;
def HasAVX       : Predicate<"Subtarget->hasAVX()">;
def HasFMA3      : Predicate<"Subtarget->hasFMA3()">;
def HasFMA4      : Predicate<"Subtarget->hasFMA4()">;
def FPStackf32   : Predicate<"!Subtarget->hasSSE1()">;
def FPStackf64   : Predicate<"!Subtarget->hasSSE2()">;
def In32BitMode  : Predicate<"!Subtarget->is64Bit()">;
def In64BitMode  : Predicate<"Subtarget->is64Bit()">;
def IsWin64      : Predicate<"Subtarget->isTargetWin64()">;
def NotWin64     : Predicate<"!Subtarget->isTargetWin64()">;
def SmallCode    : Predicate<"TM.getCodeModel() == CodeModel::Small">;
def KernelCode   : Predicate<"TM.getCodeModel() == CodeModel::Kernel">;
def FarData      : Predicate<"TM.getCodeModel() != CodeModel::Small &&"
                             "TM.getCodeModel() != CodeModel::Kernel">;
def NearData     : Predicate<"TM.getCodeModel() == CodeModel::Small ||"
                             "TM.getCodeModel() == CodeModel::Kernel">;
def IsStatic     : Predicate<"TM.getRelocationModel() == Reloc::Static">;
def IsNotPIC     : Predicate<"TM.getRelocationModel() != Reloc::PIC_">;
def OptForSize   : Predicate<"OptForSize">;
def OptForSpeed  : Predicate<"!OptForSize">;
def FastBTMem    : Predicate<"!Subtarget->isBTMemSlow()">;
def CallImmAddr  : Predicate<"Subtarget->IsLegalToCallImmediateAddr(TM)">;
def HasAES       : Predicate<"Subtarget->hasAES()">;

//===----------------------------------------------------------------------===//
// X86 Instruction Format Definitions.
//

include "X86InstrFormats.td"

//===----------------------------------------------------------------------===//
// Pattern fragments...
//

// X86 specific condition code. These correspond to CondCode in
// X86InstrInfo.h. They must be kept in synch.
def X86_COND_A   : PatLeaf<(i8 0)>;  // alt. COND_NBE
def X86_COND_AE  : PatLeaf<(i8 1)>;  // alt. COND_NC
def X86_COND_B   : PatLeaf<(i8 2)>;  // alt. COND_C
def X86_COND_BE  : PatLeaf<(i8 3)>;  // alt. COND_NA
def X86_COND_E   : PatLeaf<(i8 4)>;  // alt. COND_Z
def X86_COND_G   : PatLeaf<(i8 5)>;  // alt. COND_NLE
def X86_COND_GE  : PatLeaf<(i8 6)>;  // alt. COND_NL
def X86_COND_L   : PatLeaf<(i8 7)>;  // alt. COND_NGE
def X86_COND_LE  : PatLeaf<(i8 8)>;  // alt. COND_NG
def X86_COND_NE  : PatLeaf<(i8 9)>;  // alt. COND_NZ
def X86_COND_NO  : PatLeaf<(i8 10)>;
def X86_COND_NP  : PatLeaf<(i8 11)>; // alt. COND_PO
def X86_COND_NS  : PatLeaf<(i8 12)>;
def X86_COND_O   : PatLeaf<(i8 13)>;
def X86_COND_P   : PatLeaf<(i8 14)>; // alt. COND_PE
def X86_COND_S   : PatLeaf<(i8 15)>;

def immSext8 : PatLeaf<(imm), [{
  return N->getSExtValue() == (int8_t)N->getSExtValue();
}]>;

def i16immSExt8  : PatLeaf<(i16 immSext8)>;
def i32immSExt8  : PatLeaf<(i32 immSext8)>;

/// Load patterns: these constraint the match to the right address space.
def dsload : PatFrag<(ops node:$ptr), (load node:$ptr), [{
  if (const Value *Src = cast<LoadSDNode>(N)->getSrcValue())
    if (const PointerType *PT = dyn_cast<PointerType>(Src->getType()))
      if (PT->getAddressSpace() > 255)
        return false;
  return true;
}]>;

def gsload : PatFrag<(ops node:$ptr), (load node:$ptr), [{
  if (const Value *Src = cast<LoadSDNode>(N)->getSrcValue())
    if (const PointerType *PT = dyn_cast<PointerType>(Src->getType()))
      return PT->getAddressSpace() == 256;
  return false;
}]>;

def fsload : PatFrag<(ops node:$ptr), (load node:$ptr), [{
  if (const Value *Src = cast<LoadSDNode>(N)->getSrcValue())
    if (const PointerType *PT = dyn_cast<PointerType>(Src->getType()))
      return PT->getAddressSpace() == 257;
  return false;
}]>;


// Helper fragments for loads.
// It's always safe to treat a anyext i16 load as a i32 load if the i16 is
// known to be 32-bit aligned or better. Ditto for i8 to i16.
def loadi16 : PatFrag<(ops node:$ptr), (i16 (unindexedload node:$ptr)), [{
  LoadSDNode *LD = cast<LoadSDNode>(N);
  if (const Value *Src = LD->getSrcValue())
    if (const PointerType *PT = dyn_cast<PointerType>(Src->getType()))
      if (PT->getAddressSpace() > 255)
        return false;
  ISD::LoadExtType ExtType = LD->getExtensionType();
  if (ExtType == ISD::NON_EXTLOAD)
    return true;
  if (ExtType == ISD::EXTLOAD)
    return LD->getAlignment() >= 2 && !LD->isVolatile();
  return false;
}]>;

def loadi16_anyext : PatFrag<(ops node:$ptr), (i32 (unindexedload node:$ptr)),[{
  LoadSDNode *LD = cast<LoadSDNode>(N);
  if (const Value *Src = LD->getSrcValue())
    if (const PointerType *PT = dyn_cast<PointerType>(Src->getType()))
      if (PT->getAddressSpace() > 255)
        return false;
  ISD::LoadExtType ExtType = LD->getExtensionType();
  if (ExtType == ISD::EXTLOAD)
    return LD->getAlignment() >= 2 && !LD->isVolatile();
  return false;
}]>;

def loadi32 : PatFrag<(ops node:$ptr), (i32 (unindexedload node:$ptr)), [{
  LoadSDNode *LD = cast<LoadSDNode>(N);
  if (const Value *Src = LD->getSrcValue())
    if (const PointerType *PT = dyn_cast<PointerType>(Src->getType()))
      if (PT->getAddressSpace() > 255)
        return false;
  ISD::LoadExtType ExtType = LD->getExtensionType();
  if (ExtType == ISD::NON_EXTLOAD)
    return true;
  if (ExtType == ISD::EXTLOAD)
    return LD->getAlignment() >= 4 && !LD->isVolatile();
  return false;
}]>;

def loadi8  : PatFrag<(ops node:$ptr), (i8  (dsload node:$ptr))>;
def loadi64 : PatFrag<(ops node:$ptr), (i64 (dsload node:$ptr))>;
def loadf32 : PatFrag<(ops node:$ptr), (f32 (dsload node:$ptr))>;
def loadf64 : PatFrag<(ops node:$ptr), (f64 (dsload node:$ptr))>;
def loadf80 : PatFrag<(ops node:$ptr), (f80 (dsload node:$ptr))>;

def sextloadi16i8  : PatFrag<(ops node:$ptr), (i16 (sextloadi8 node:$ptr))>;
def sextloadi32i8  : PatFrag<(ops node:$ptr), (i32 (sextloadi8 node:$ptr))>;
def sextloadi32i16 : PatFrag<(ops node:$ptr), (i32 (sextloadi16 node:$ptr))>;

def zextloadi8i1   : PatFrag<(ops node:$ptr), (i8  (zextloadi1 node:$ptr))>;
def zextloadi16i1  : PatFrag<(ops node:$ptr), (i16 (zextloadi1 node:$ptr))>;
def zextloadi32i1  : PatFrag<(ops node:$ptr), (i32 (zextloadi1 node:$ptr))>;
def zextloadi16i8  : PatFrag<(ops node:$ptr), (i16 (zextloadi8 node:$ptr))>;
def zextloadi32i8  : PatFrag<(ops node:$ptr), (i32 (zextloadi8 node:$ptr))>;
def zextloadi32i16 : PatFrag<(ops node:$ptr), (i32 (zextloadi16 node:$ptr))>;

def extloadi8i1    : PatFrag<(ops node:$ptr), (i8  (extloadi1 node:$ptr))>;
def extloadi16i1   : PatFrag<(ops node:$ptr), (i16 (extloadi1 node:$ptr))>;
def extloadi32i1   : PatFrag<(ops node:$ptr), (i32 (extloadi1 node:$ptr))>;
def extloadi16i8   : PatFrag<(ops node:$ptr), (i16 (extloadi8 node:$ptr))>;
def extloadi32i8   : PatFrag<(ops node:$ptr), (i32 (extloadi8 node:$ptr))>;
def extloadi32i16  : PatFrag<(ops node:$ptr), (i32 (extloadi16 node:$ptr))>;


// An 'and' node with a single use.
def and_su : PatFrag<(ops node:$lhs, node:$rhs), (and node:$lhs, node:$rhs), [{
  return N->hasOneUse();
}]>;
// An 'srl' node with a single use.
def srl_su : PatFrag<(ops node:$lhs, node:$rhs), (srl node:$lhs, node:$rhs), [{
  return N->hasOneUse();
}]>;
// An 'trunc' node with a single use.
def trunc_su : PatFrag<(ops node:$src), (trunc node:$src), [{
  return N->hasOneUse();
}]>;

// Treat an 'or' node is as an 'add' if the or'ed bits are known to be zero.
def or_is_add : PatFrag<(ops node:$lhs, node:$rhs), (or node:$lhs, node:$rhs),[{
  if (ConstantSDNode *CN = dyn_cast<ConstantSDNode>(N->getOperand(1)))
    return CurDAG->MaskedValueIsZero(N->getOperand(0), CN->getAPIntValue());

  unsigned BitWidth = N->getValueType(0).getScalarType().getSizeInBits();
  APInt Mask = APInt::getAllOnesValue(BitWidth);
  APInt KnownZero0, KnownOne0;
  CurDAG->ComputeMaskedBits(N->getOperand(0), Mask, KnownZero0, KnownOne0, 0);
  APInt KnownZero1, KnownOne1;
  CurDAG->ComputeMaskedBits(N->getOperand(1), Mask, KnownZero1, KnownOne1, 0);
  return (~KnownZero0 & ~KnownZero1) == 0;
}]>;

//===----------------------------------------------------------------------===//
// Instruction list...
//

// ADJCALLSTACKDOWN/UP implicitly use/def ESP because they may be expanded into
// a stack adjustment and the codegen must know that they may modify the stack
// pointer before prolog-epilog rewriting occurs.
// Pessimistically assume ADJCALLSTACKDOWN / ADJCALLSTACKUP will become
// sub / add which can clobber EFLAGS.
let Defs = [ESP, EFLAGS], Uses = [ESP] in {
def ADJCALLSTACKDOWN32 : I<0, Pseudo, (outs), (ins i32imm:$amt),
                           "#ADJCALLSTACKDOWN",
                           [(X86callseq_start timm:$amt)]>,
                          Requires<[In32BitMode]>;
def ADJCALLSTACKUP32   : I<0, Pseudo, (outs), (ins i32imm:$amt1, i32imm:$amt2),
                           "#ADJCALLSTACKUP",
                           [(X86callseq_end timm:$amt1, timm:$amt2)]>,
                          Requires<[In32BitMode]>;
}

// x86-64 va_start lowering magic.
let usesCustomInserter = 1 in {
def VASTART_SAVE_XMM_REGS : I<0, Pseudo,
                              (outs),
                              (ins GR8:$al,
                                   i64imm:$regsavefi, i64imm:$offset,
                                   variable_ops),
                              "#VASTART_SAVE_XMM_REGS $al, $regsavefi, $offset",
                              [(X86vastart_save_xmm_regs GR8:$al,
                                                         imm:$regsavefi,
                                                         imm:$offset)]>;

// Dynamic stack allocation yields _alloca call for Cygwin/Mingw targets.  Calls
// to _alloca is needed to probe the stack when allocating more than 4k bytes in
// one go. Touching the stack at 4K increments is necessary to ensure that the
// guard pages used by the OS virtual memory manager are allocated in correct
// sequence.
// The main point of having separate instruction are extra unmodelled effects
// (compared to ordinary calls) like stack pointer change.

def MINGW_ALLOCA : I<0, Pseudo, (outs), (ins),
                     "# dynamic stack allocation",
                     [(X86MingwAlloca)]>;
}

// Nop
let neverHasSideEffects = 1 in {
  def NOOP : I<0x90, RawFrm, (outs), (ins), "nop", []>;
  def NOOPW : I<0x1f, MRM0m, (outs), (ins i16mem:$zero),
                "nop{w}\t$zero", []>, TB, OpSize;
  def NOOPL : I<0x1f, MRM0m, (outs), (ins i32mem:$zero),
                "nop{l}\t$zero", []>, TB;
}

// Trap
def INTO : I<0xce, RawFrm, (outs), (ins), "into", []>;
def INT3 : I<0xcc, RawFrm, (outs), (ins), "int3", []>;
// FIXME: need to make sure that "int $3" matches int3
def INT : Ii8<0xcd, RawFrm, (outs), (ins i8imm:$trap), "int\t$trap", []>;
def IRET16 : I<0xcf, RawFrm, (outs), (ins), "iret{w}", []>, OpSize;
def IRET32 : I<0xcf, RawFrm, (outs), (ins), "iret{l}", []>;

// PIC base construction.  This expands to code that looks like this:
//     call  $next_inst
//     popl %destreg"
let neverHasSideEffects = 1, isNotDuplicable = 1, Uses = [ESP] in
  def MOVPC32r : Ii32<0xE8, Pseudo, (outs GR32:$reg), (ins i32imm:$label),
                      "", []>;

//===----------------------------------------------------------------------===//
//  Control Flow Instructions.
//

// Return instructions.
let isTerminator = 1, isReturn = 1, isBarrier = 1,
    hasCtrlDep = 1, FPForm = SpecialFP in {
  def RET    : I   <0xC3, RawFrm, (outs), (ins variable_ops),
                    "ret",
                    [(X86retflag 0)]>;
  def RETI   : Ii16<0xC2, RawFrm, (outs), (ins i16imm:$amt, variable_ops),
                    "ret\t$amt",
                    [(X86retflag timm:$amt)]>;
  def LRET   : I   <0xCB, RawFrm, (outs), (ins),
                    "lret", []>;
  def LRETI  : Ii16<0xCA, RawFrm, (outs), (ins i16imm:$amt),
                    "lret\t$amt", []>;
}

// Unconditional branches.
let isBarrier = 1, isBranch = 1, isTerminator = 1 in {
  def JMP_4 : Ii32PCRel<0xE9, RawFrm, (outs), (ins brtarget:$dst),
                        "jmp\t$dst", [(br bb:$dst)]>;
  def JMP_1 : Ii8PCRel<0xEB, RawFrm, (outs), (ins brtarget8:$dst),
                       "jmp\t$dst", []>;
}

// Conditional Branches.
let isBranch = 1, isTerminator = 1, Uses = [EFLAGS] in {
  multiclass ICBr<bits<8> opc1, bits<8> opc4, string asm, PatFrag Cond> {
    def _1 : Ii8PCRel <opc1, RawFrm, (outs), (ins brtarget8:$dst), asm, []>;
    def _4 : Ii32PCRel<opc4, RawFrm, (outs), (ins brtarget:$dst), asm,
                       [(X86brcond bb:$dst, Cond, EFLAGS)]>, TB;
  }
}

defm JO  : ICBr<0x70, 0x80, "jo\t$dst" , X86_COND_O>;
defm JNO : ICBr<0x71, 0x81, "jno\t$dst" , X86_COND_NO>;
defm JB  : ICBr<0x72, 0x82, "jb\t$dst" , X86_COND_B>;
defm JAE : ICBr<0x73, 0x83, "jae\t$dst", X86_COND_AE>;
defm JE  : ICBr<0x74, 0x84, "je\t$dst" , X86_COND_E>;
defm JNE : ICBr<0x75, 0x85, "jne\t$dst", X86_COND_NE>;
defm JBE : ICBr<0x76, 0x86, "jbe\t$dst", X86_COND_BE>;
defm JA  : ICBr<0x77, 0x87, "ja\t$dst" , X86_COND_A>;
defm JS  : ICBr<0x78, 0x88, "js\t$dst" , X86_COND_S>;
defm JNS : ICBr<0x79, 0x89, "jns\t$dst", X86_COND_NS>;
defm JP  : ICBr<0x7A, 0x8A, "jp\t$dst" , X86_COND_P>;
defm JNP : ICBr<0x7B, 0x8B, "jnp\t$dst", X86_COND_NP>;
defm JL  : ICBr<0x7C, 0x8C, "jl\t$dst" , X86_COND_L>;
defm JGE : ICBr<0x7D, 0x8D, "jge\t$dst", X86_COND_GE>;
defm JLE : ICBr<0x7E, 0x8E, "jle\t$dst", X86_COND_LE>;
defm JG  : ICBr<0x7F, 0x8F, "jg\t$dst" , X86_COND_G>;

// FIXME: What about the CX/RCX versions of this instruction?
let Uses = [ECX], isBranch = 1, isTerminator = 1 in
  def JCXZ8 : Ii8PCRel<0xE3, RawFrm, (outs), (ins brtarget8:$dst),
                       "jcxz\t$dst", []>;


// Indirect branches
let isBranch = 1, isTerminator = 1, isBarrier = 1, isIndirectBranch = 1 in {
  def JMP32r     : I<0xFF, MRM4r, (outs), (ins GR32:$dst), "jmp{l}\t{*}$dst",
                     [(brind GR32:$dst)]>;
  def JMP32m     : I<0xFF, MRM4m, (outs), (ins i32mem:$dst), "jmp{l}\t{*}$dst",
                     [(brind (loadi32 addr:$dst))]>;
                     
  def FARJMP16i  : Iseg16<0xEA, RawFrm, (outs), 
                          (ins i16imm:$seg, i16imm:$off),
                          "ljmp{w}\t$seg, $off", []>, OpSize;
  def FARJMP32i  : Iseg32<0xEA, RawFrm, (outs),
                          (ins i16imm:$seg, i32imm:$off),
                          "ljmp{l}\t$seg, $off", []>;                     

  def FARJMP16m  : I<0xFF, MRM5m, (outs), (ins opaque32mem:$dst), 
                     "ljmp{w}\t{*}$dst", []>, OpSize;
  def FARJMP32m  : I<0xFF, MRM5m, (outs), (ins opaque48mem:$dst),
                     "ljmp{l}\t{*}$dst", []>;
}


// Loop instructions

def LOOP   : I<0xE2, RawFrm, (outs), (ins brtarget8:$dst), "loop\t$dst", []>;
def LOOPE  : I<0xE1, RawFrm, (outs), (ins brtarget8:$dst), "loope\t$dst", []>;
def LOOPNE : I<0xE0, RawFrm, (outs), (ins brtarget8:$dst), "loopne\t$dst", []>;

//===----------------------------------------------------------------------===//
//  Call Instructions...
//
let isCall = 1 in
  // All calls clobber the non-callee saved registers. ESP is marked as
  // a use to prevent stack-pointer assignments that appear immediately
  // before calls from potentially appearing dead. Uses for argument
  // registers are added manually.
  let Defs = [EAX, ECX, EDX, FP0, FP1, FP2, FP3, FP4, FP5, FP6, ST0,
              MM0, MM1, MM2, MM3, MM4, MM5, MM6, MM7,
              XMM0, XMM1, XMM2, XMM3, XMM4, XMM5, XMM6, XMM7,
              XMM8, XMM9, XMM10, XMM11, XMM12, XMM13, XMM14, XMM15, EFLAGS],
      Uses = [ESP] in {
    def CALLpcrel32 : Ii32PCRel<0xE8, RawFrm,
                           (outs), (ins i32imm_pcrel:$dst,variable_ops),
                           "call\t$dst", []>;
    def CALL32r     : I<0xFF, MRM2r, (outs), (ins GR32:$dst, variable_ops),
                        "call\t{*}$dst", [(X86call GR32:$dst)]>;
    def CALL32m     : I<0xFF, MRM2m, (outs), (ins i32mem:$dst, variable_ops),
                        "call\t{*}$dst", [(X86call (loadi32 addr:$dst))]>;
  
    def FARCALL16i  : Iseg16<0x9A, RawFrm, (outs), 
                             (ins i16imm:$seg, i16imm:$off),
                             "lcall{w}\t$seg, $off", []>, OpSize;
    def FARCALL32i  : Iseg32<0x9A, RawFrm, (outs),
                             (ins i16imm:$seg, i32imm:$off),
                             "lcall{l}\t$seg, $off", []>;
                             
    def FARCALL16m  : I<0xFF, MRM3m, (outs), (ins opaque32mem:$dst),
                        "lcall{w}\t{*}$dst", []>, OpSize;
    def FARCALL32m  : I<0xFF, MRM3m, (outs), (ins opaque48mem:$dst),
                        "lcall{l}\t{*}$dst", []>;
  }

// Constructing a stack frame.

def ENTER : I<0xC8, RawFrm, (outs), (ins i16imm:$len, i8imm:$lvl),
              "enter\t$len, $lvl", []>;

// Tail call stuff.

let isCall = 1, isTerminator = 1, isReturn = 1, isBarrier = 1 in
  let Defs = [EAX, ECX, EDX, FP0, FP1, FP2, FP3, FP4, FP5, FP6, ST0,
              MM0, MM1, MM2, MM3, MM4, MM5, MM6, MM7,
              XMM0, XMM1, XMM2, XMM3, XMM4, XMM5, XMM6, XMM7,
              XMM8, XMM9, XMM10, XMM11, XMM12, XMM13, XMM14, XMM15, EFLAGS],
      Uses = [ESP] in {
  def TCRETURNdi : I<0, Pseudo, (outs), 
                     (ins i32imm_pcrel:$dst, i32imm:$offset, variable_ops),
                   "#TC_RETURN $dst $offset", []>;
  def TCRETURNri : I<0, Pseudo, (outs), 
                     (ins GR32_TC:$dst, i32imm:$offset, variable_ops),
                     "#TC_RETURN $dst $offset", []>;
  let mayLoad = 1 in
  def TCRETURNmi : I<0, Pseudo, (outs), 
                     (ins i32mem_TC:$dst, i32imm:$offset, variable_ops),
                     "#TC_RETURN $dst $offset", []>;

  // FIXME: The should be pseudo instructions that are lowered when going to
  // mcinst.
  def TAILJMPd : Ii32PCRel<0xE9, RawFrm, (outs),
                           (ins i32imm_pcrel:$dst, variable_ops),
                 "jmp\t$dst  # TAILCALL",
                 []>;
  def TAILJMPr : I<0xFF, MRM4r, (outs), (ins GR32_TC:$dst, variable_ops), 
                   "jmp{l}\t{*}$dst  # TAILCALL",
                 []>;     
  let mayLoad = 1 in
  def TAILJMPm : I<0xFF, MRM4m, (outs), (ins i32mem_TC:$dst, variable_ops),
                   "jmp{l}\t{*}$dst  # TAILCALL", []>;

  // FIXME: This is a hack so that MCInst lowering can preserve the TAILCALL
  // marker on instructions, while still being able to relax.
  let isCodeGenOnly = 1 in {
    def TAILJMP_1 : Ii8PCRel<0xEB, RawFrm, (outs), (ins brtarget8:$dst),
                         "jmp\t$dst  # TAILCALL", []>;
  }
}

//===----------------------------------------------------------------------===//
//  Miscellaneous Instructions...
//
let Defs = [EBP, ESP], Uses = [EBP, ESP], mayLoad = 1, neverHasSideEffects=1 in
def LEAVE    : I<0xC9, RawFrm,
                 (outs), (ins), "leave", []>;

def POPCNT16rr : I<0xB8, MRMSrcReg, (outs GR16:$dst), (ins GR16:$src),
                   "popcnt{w}\t{$src, $dst|$dst, $src}", []>, OpSize, XS;
let mayLoad = 1 in
def POPCNT16rm : I<0xB8, MRMSrcMem, (outs GR16:$dst), (ins i16mem:$src),
                   "popcnt{w}\t{$src, $dst|$dst, $src}", []>, OpSize, XS;
def POPCNT32rr : I<0xB8, MRMSrcReg, (outs GR32:$dst), (ins GR32:$src),
                   "popcnt{l}\t{$src, $dst|$dst, $src}", []>, XS;
let mayLoad = 1 in
def POPCNT32rm : I<0xB8, MRMSrcMem, (outs GR32:$dst), (ins i32mem:$src),
                   "popcnt{l}\t{$src, $dst|$dst, $src}", []>, XS;

let Defs = [ESP], Uses = [ESP], neverHasSideEffects=1 in {
let mayLoad = 1 in {
def POP16r  : I<0x58, AddRegFrm, (outs GR16:$reg), (ins), "pop{w}\t$reg", []>,
  OpSize;
def POP32r  : I<0x58, AddRegFrm, (outs GR32:$reg), (ins), "pop{l}\t$reg", []>;
def POP16rmr: I<0x8F, MRM0r, (outs GR16:$reg), (ins), "pop{w}\t$reg", []>,
  OpSize;
def POP16rmm: I<0x8F, MRM0m, (outs i16mem:$dst), (ins), "pop{w}\t$dst", []>,
  OpSize;
def POP32rmr: I<0x8F, MRM0r, (outs GR32:$reg), (ins), "pop{l}\t$reg", []>;
def POP32rmm: I<0x8F, MRM0m, (outs i32mem:$dst), (ins), "pop{l}\t$dst", []>;
}

let mayStore = 1 in {
def PUSH16r  : I<0x50, AddRegFrm, (outs), (ins GR16:$reg), "push{w}\t$reg",[]>,
  OpSize;
def PUSH32r  : I<0x50, AddRegFrm, (outs), (ins GR32:$reg), "push{l}\t$reg",[]>;
def PUSH16rmr: I<0xFF, MRM6r, (outs), (ins GR16:$reg), "push{w}\t$reg",[]>,
  OpSize;
def PUSH16rmm: I<0xFF, MRM6m, (outs), (ins i16mem:$src), "push{w}\t$src",[]>,
  OpSize;
def PUSH32rmr: I<0xFF, MRM6r, (outs), (ins GR32:$reg), "push{l}\t$reg",[]>;
def PUSH32rmm: I<0xFF, MRM6m, (outs), (ins i32mem:$src), "push{l}\t$src",[]>;
}
}

let Defs = [ESP], Uses = [ESP], neverHasSideEffects = 1, mayStore = 1 in {
def PUSHi8   : Ii8<0x6a, RawFrm, (outs), (ins i32i8imm:$imm), 
                      "push{l}\t$imm", []>;
def PUSHi16  : Ii16<0x68, RawFrm, (outs), (ins i16imm:$imm), 
                      "push{w}\t$imm", []>, OpSize;
def PUSHi32  : Ii32<0x68, RawFrm, (outs), (ins i32imm:$imm), 
                      "push{l}\t$imm", []>;
}

let Defs = [ESP, EFLAGS], Uses = [ESP], mayLoad = 1, neverHasSideEffects=1 in {
def POPF16   : I<0x9D, RawFrm, (outs), (ins), "popf{w}", []>, OpSize;
def POPF32   : I<0x9D, RawFrm, (outs), (ins), "popf{l|d}", []>,
               Requires<[In32BitMode]>;
}
let Defs = [ESP], Uses = [ESP, EFLAGS], mayStore = 1, neverHasSideEffects=1 in {
def PUSHF16  : I<0x9C, RawFrm, (outs), (ins), "pushf{w}", []>, OpSize;
def PUSHF32  : I<0x9C, RawFrm, (outs), (ins), "pushf{l|d}", []>,
               Requires<[In32BitMode]>;
}

let Defs = [EDI, ESI, EBP, EBX, EDX, ECX, EAX, ESP], Uses = [ESP],
    mayLoad=1, neverHasSideEffects=1 in {
def POPA32   : I<0x61, RawFrm, (outs), (ins), "popa{l}", []>,
               Requires<[In32BitMode]>;
}
let Defs = [ESP], Uses = [EDI, ESI, EBP, EBX, EDX, ECX, EAX, ESP],
    mayStore=1, neverHasSideEffects=1 in {
def PUSHA32  : I<0x60, RawFrm, (outs), (ins), "pusha{l}", []>,
               Requires<[In32BitMode]>;
}

let Uses = [EFLAGS], Constraints = "$src = $dst" in     // GR32 = bswap GR32
  def BSWAP32r : I<0xC8, AddRegFrm,
                   (outs GR32:$dst), (ins GR32:$src),
                   "bswap{l}\t$dst", 
                   [(set GR32:$dst, (bswap GR32:$src))]>, TB;


// Bit scan instructions.
let Defs = [EFLAGS] in {
def BSF16rr  : I<0xBC, MRMSrcReg, (outs GR16:$dst), (ins GR16:$src),
                 "bsf{w}\t{$src, $dst|$dst, $src}",
                 [(set GR16:$dst, EFLAGS, (X86bsf GR16:$src))]>, TB, OpSize;
def BSF16rm  : I<0xBC, MRMSrcMem, (outs GR16:$dst), (ins i16mem:$src),
                 "bsf{w}\t{$src, $dst|$dst, $src}",
                 [(set GR16:$dst, EFLAGS, (X86bsf (loadi16 addr:$src)))]>, TB,
                 OpSize;
def BSF32rr  : I<0xBC, MRMSrcReg, (outs GR32:$dst), (ins GR32:$src),
                 "bsf{l}\t{$src, $dst|$dst, $src}",
                 [(set GR32:$dst, EFLAGS, (X86bsf GR32:$src))]>, TB;
def BSF32rm  : I<0xBC, MRMSrcMem, (outs GR32:$dst), (ins i32mem:$src),
                 "bsf{l}\t{$src, $dst|$dst, $src}",
                 [(set GR32:$dst, EFLAGS, (X86bsf (loadi32 addr:$src)))]>, TB;

def BSR16rr  : I<0xBD, MRMSrcReg, (outs GR16:$dst), (ins GR16:$src),
                 "bsr{w}\t{$src, $dst|$dst, $src}",
                 [(set GR16:$dst, EFLAGS, (X86bsr GR16:$src))]>, TB, OpSize;
def BSR16rm  : I<0xBD, MRMSrcMem, (outs GR16:$dst), (ins i16mem:$src),
                 "bsr{w}\t{$src, $dst|$dst, $src}",
                 [(set GR16:$dst, EFLAGS, (X86bsr (loadi16 addr:$src)))]>, TB,
                 OpSize;
def BSR32rr  : I<0xBD, MRMSrcReg, (outs GR32:$dst), (ins GR32:$src),
                 "bsr{l}\t{$src, $dst|$dst, $src}",
                 [(set GR32:$dst, EFLAGS, (X86bsr GR32:$src))]>, TB;
def BSR32rm  : I<0xBD, MRMSrcMem, (outs GR32:$dst), (ins i32mem:$src),
                 "bsr{l}\t{$src, $dst|$dst, $src}",
                 [(set GR32:$dst, EFLAGS, (X86bsr (loadi32 addr:$src)))]>, TB;
} // Defs = [EFLAGS]

let neverHasSideEffects = 1 in
def LEA16r   : I<0x8D, MRMSrcMem,
                 (outs GR16:$dst), (ins lea32mem:$src),
                 "lea{w}\t{$src|$dst}, {$dst|$src}", []>, OpSize;
let isReMaterializable = 1 in
def LEA32r   : I<0x8D, MRMSrcMem,
                 (outs GR32:$dst), (ins lea32mem:$src),
                 "lea{l}\t{$src|$dst}, {$dst|$src}",
                 [(set GR32:$dst, lea32addr:$src)]>, Requires<[In32BitMode]>;

let Defs = [ECX,EDI,ESI], Uses = [ECX,EDI,ESI], isCodeGenOnly = 1 in {
def REP_MOVSB : I<0xA4, RawFrm, (outs), (ins), "{rep;movsb|rep movsb}",
                  [(X86rep_movs i8)]>, REP;
def REP_MOVSW : I<0xA5, RawFrm, (outs), (ins), "{rep;movsw|rep movsw}",
                  [(X86rep_movs i16)]>, REP, OpSize;
def REP_MOVSD : I<0xA5, RawFrm, (outs), (ins), "{rep;movsl|rep movsd}",
                  [(X86rep_movs i32)]>, REP;
}

// These uses the DF flag in the EFLAGS register to inc or dec EDI and ESI
let Defs = [EDI,ESI], Uses = [EDI,ESI,EFLAGS] in {
def MOVSB : I<0xA4, RawFrm, (outs), (ins), "{movsb}", []>;
def MOVSW : I<0xA5, RawFrm, (outs), (ins), "{movsw}", []>, OpSize;
def MOVSD : I<0xA5, RawFrm, (outs), (ins), "{movsl|movsd}", []>;
}

let Defs = [ECX,EDI], Uses = [AL,ECX,EDI], isCodeGenOnly = 1 in
def REP_STOSB : I<0xAA, RawFrm, (outs), (ins), "{rep;stosb|rep stosb}",
                  [(X86rep_stos i8)]>, REP;
let Defs = [ECX,EDI], Uses = [AX,ECX,EDI], isCodeGenOnly = 1 in
def REP_STOSW : I<0xAB, RawFrm, (outs), (ins), "{rep;stosw|rep stosw}",
                  [(X86rep_stos i16)]>, REP, OpSize;
let Defs = [ECX,EDI], Uses = [EAX,ECX,EDI], isCodeGenOnly = 1 in
def REP_STOSD : I<0xAB, RawFrm, (outs), (ins), "{rep;stosl|rep stosd}",
                  [(X86rep_stos i32)]>, REP;

// These uses the DF flag in the EFLAGS register to inc or dec EDI and ESI
let Defs = [EDI], Uses = [AL,EDI,EFLAGS] in
def STOSB : I<0xAA, RawFrm, (outs), (ins), "{stosb}", []>;
let Defs = [EDI], Uses = [AX,EDI,EFLAGS] in
def STOSW : I<0xAB, RawFrm, (outs), (ins), "{stosw}", []>, OpSize;
let Defs = [EDI], Uses = [EAX,EDI,EFLAGS] in
def STOSD : I<0xAB, RawFrm, (outs), (ins), "{stosl|stosd}", []>;

def SCAS8 : I<0xAE, RawFrm, (outs), (ins), "scas{b}", []>;
def SCAS16 : I<0xAF, RawFrm, (outs), (ins), "scas{w}", []>, OpSize;
def SCAS32 : I<0xAF, RawFrm, (outs), (ins), "scas{l}", []>;

def CMPS8 : I<0xA6, RawFrm, (outs), (ins), "cmps{b}", []>;
def CMPS16 : I<0xA7, RawFrm, (outs), (ins), "cmps{w}", []>, OpSize;
def CMPS32 : I<0xA7, RawFrm, (outs), (ins), "cmps{l}", []>;

let Defs = [RAX, RDX] in
def RDTSC : I<0x31, RawFrm, (outs), (ins), "rdtsc", [(X86rdtsc)]>,
            TB;

let Defs = [RAX, RCX, RDX] in
def RDTSCP : I<0x01, MRM_F9, (outs), (ins), "rdtscp", []>, TB;

let isTerminator = 1, isBarrier = 1, hasCtrlDep = 1 in {
def TRAP    : I<0x0B, RawFrm, (outs), (ins), "ud2", [(trap)]>, TB;
}

def SYSCALL  : I<0x05, RawFrm,
                 (outs), (ins), "syscall", []>, TB;
def SYSRET   : I<0x07, RawFrm,
                 (outs), (ins), "sysret", []>, TB;
def SYSENTER : I<0x34, RawFrm,
                 (outs), (ins), "sysenter", []>, TB;
def SYSEXIT  : I<0x35, RawFrm,
                 (outs), (ins), "sysexit", []>, TB;

def WAIT : I<0x9B, RawFrm, (outs), (ins), "wait", []>;


//===----------------------------------------------------------------------===//
//  Input/Output Instructions...
//
let Defs = [AL], Uses = [DX] in
def IN8rr  : I<0xEC, RawFrm, (outs), (ins),
               "in{b}\t{%dx, %al|%AL, %DX}", []>;
let Defs = [AX], Uses = [DX] in
def IN16rr : I<0xED, RawFrm, (outs), (ins),
               "in{w}\t{%dx, %ax|%AX, %DX}", []>,  OpSize;
let Defs = [EAX], Uses = [DX] in
def IN32rr : I<0xED, RawFrm, (outs), (ins),
               "in{l}\t{%dx, %eax|%EAX, %DX}", []>;

let Defs = [AL] in
def IN8ri  : Ii8<0xE4, RawFrm, (outs), (ins i16i8imm:$port),
                  "in{b}\t{$port, %al|%AL, $port}", []>;
let Defs = [AX] in
def IN16ri : Ii8<0xE5, RawFrm, (outs), (ins i16i8imm:$port),
                  "in{w}\t{$port, %ax|%AX, $port}", []>, OpSize;
let Defs = [EAX] in
def IN32ri : Ii8<0xE5, RawFrm, (outs), (ins i16i8imm:$port),
                  "in{l}\t{$port, %eax|%EAX, $port}", []>;

let Uses = [DX, AL] in
def OUT8rr  : I<0xEE, RawFrm, (outs), (ins),
                "out{b}\t{%al, %dx|%DX, %AL}", []>;
let Uses = [DX, AX] in
def OUT16rr : I<0xEF, RawFrm, (outs), (ins),
                "out{w}\t{%ax, %dx|%DX, %AX}", []>, OpSize;
let Uses = [DX, EAX] in
def OUT32rr : I<0xEF, RawFrm, (outs), (ins),
                "out{l}\t{%eax, %dx|%DX, %EAX}", []>;

let Uses = [AL] in
def OUT8ir  : Ii8<0xE6, RawFrm, (outs), (ins i16i8imm:$port),
                   "out{b}\t{%al, $port|$port, %AL}", []>;
let Uses = [AX] in
def OUT16ir : Ii8<0xE7, RawFrm, (outs), (ins i16i8imm:$port),
                   "out{w}\t{%ax, $port|$port, %AX}", []>, OpSize;
let Uses = [EAX] in
def OUT32ir : Ii8<0xE7, RawFrm, (outs), (ins i16i8imm:$port),
                   "out{l}\t{%eax, $port|$port, %EAX}", []>;

def IN8  : I<0x6C, RawFrm, (outs), (ins),
             "ins{b}", []>;
def IN16 : I<0x6D, RawFrm, (outs), (ins),
             "ins{w}", []>,  OpSize;
def IN32 : I<0x6D, RawFrm, (outs), (ins),
             "ins{l}", []>;

//===----------------------------------------------------------------------===//
//  Move Instructions...
//
let neverHasSideEffects = 1 in {