#include "llvm/Support/raw_ostream.h"
#include "llvm/Target/TargetOptions.h"
#include "llvm/MC/MCAsmInfo.h"
-
#include <limits>
using namespace llvm;
X86InstrInfo::X86InstrInfo(X86TargetMachine &tm)
: TargetInstrInfoImpl(X86Insts, array_lengthof(X86Insts)),
TM(tm), RI(tm, *this) {
- SmallVector<unsigned,16> AmbEntries;
+ enum {
+ TB_NOT_REVERSABLE = 1U << 31,
+ TB_FLAGS = TB_NOT_REVERSABLE
+ };
+
static const unsigned OpTbl2Addr[][2] = {
{ X86::ADC32ri, X86::ADC32mi },
{ X86::ADC32ri8, X86::ADC32mi8 },
{ X86::ADC64rr, X86::ADC64mr },
{ X86::ADD16ri, X86::ADD16mi },
{ X86::ADD16ri8, X86::ADD16mi8 },
+ { X86::ADD16ri_DB, X86::ADD16mi | TB_NOT_REVERSABLE },
+ { X86::ADD16ri8_DB, X86::ADD16mi8 | TB_NOT_REVERSABLE },
{ X86::ADD16rr, X86::ADD16mr },
+ { X86::ADD16rr_DB, X86::ADD16mr | TB_NOT_REVERSABLE },
{ X86::ADD32ri, X86::ADD32mi },
{ X86::ADD32ri8, X86::ADD32mi8 },
+ { X86::ADD32ri_DB, X86::ADD32mi | TB_NOT_REVERSABLE },
+ { X86::ADD32ri8_DB, X86::ADD32mi8 | TB_NOT_REVERSABLE },
{ X86::ADD32rr, X86::ADD32mr },
+ { X86::ADD32rr_DB, X86::ADD32mr | TB_NOT_REVERSABLE },
{ X86::ADD64ri32, X86::ADD64mi32 },
{ X86::ADD64ri8, X86::ADD64mi8 },
+ { X86::ADD64ri32_DB,X86::ADD64mi32 | TB_NOT_REVERSABLE },
+ { X86::ADD64ri8_DB, X86::ADD64mi8 | TB_NOT_REVERSABLE },
{ X86::ADD64rr, X86::ADD64mr },
+ { X86::ADD64rr_DB, X86::ADD64mr | TB_NOT_REVERSABLE },
{ X86::ADD8ri, X86::ADD8mi },
{ X86::ADD8rr, X86::ADD8mr },
{ X86::AND16ri, X86::AND16mi },
for (unsigned i = 0, e = array_lengthof(OpTbl2Addr); i != e; ++i) {
unsigned RegOp = OpTbl2Addr[i][0];
- unsigned MemOp = OpTbl2Addr[i][1];
- if (!RegOp2MemOpTable2Addr.insert(std::make_pair((unsigned*)RegOp,
- std::make_pair(MemOp,0))).second)
- assert(false && "Duplicated entries?");
+ unsigned MemOp = OpTbl2Addr[i][1] & ~TB_FLAGS;
+ assert(!RegOp2MemOpTable2Addr.count(RegOp) && "Duplicated entries?");
+ RegOp2MemOpTable2Addr[RegOp] = std::make_pair(MemOp, 0U);
+
+ // If this is not a reversable operation (because there is a many->one)
+ // mapping, don't insert the reverse of the operation into MemOp2RegOpTable.
+ if (OpTbl2Addr[i][1] & TB_NOT_REVERSABLE)
+ continue;
+
// Index 0, folded load and store, no alignment requirement.
unsigned AuxInfo = 0 | (1 << 4) | (1 << 5);
- if (!MemOp2RegOpTable.insert(std::make_pair((unsigned*)MemOp,
- std::make_pair(RegOp,
- AuxInfo))).second)
- AmbEntries.push_back(MemOp);
+
+ assert(!MemOp2RegOpTable.count(MemOp) &&
+ "Duplicated entries in unfolding maps?");
+ MemOp2RegOpTable[MemOp] = std::make_pair(RegOp, AuxInfo);
}
// If the third value is 1, then it's folding either a load or a store.
{ X86::BT64ri8, X86::BT64mi8, 1, 0 },
{ X86::CALL32r, X86::CALL32m, 1, 0 },
{ X86::CALL64r, X86::CALL64m, 1, 0 },
+ { X86::WINCALL64r, X86::WINCALL64m, 1, 0 },
{ X86::CMP16ri, X86::CMP16mi, 1, 0 },
{ X86::CMP16ri8, X86::CMP16mi8, 1, 0 },
{ X86::CMP16rr, X86::CMP16mr, 1, 0 },
{ X86::DIV64r, X86::DIV64m, 1, 0 },
{ X86::DIV8r, X86::DIV8m, 1, 0 },
{ X86::EXTRACTPSrr, X86::EXTRACTPSmr, 0, 16 },
- { X86::FsMOVAPDrr, X86::MOVSDmr, 0, 0 },
- { X86::FsMOVAPSrr, X86::MOVSSmr, 0, 0 },
+ { X86::FsMOVAPDrr, X86::MOVSDmr | TB_NOT_REVERSABLE , 0, 0 },
+ { X86::FsMOVAPSrr, X86::MOVSSmr | TB_NOT_REVERSABLE , 0, 0 },
{ X86::IDIV16r, X86::IDIV16m, 1, 0 },
{ X86::IDIV32r, X86::IDIV32m, 1, 0 },
{ X86::IDIV64r, X86::IDIV64m, 1, 0 },
};
for (unsigned i = 0, e = array_lengthof(OpTbl0); i != e; ++i) {
- unsigned RegOp = OpTbl0[i][0];
- unsigned MemOp = OpTbl0[i][1];
- unsigned Align = OpTbl0[i][3];
- if (!RegOp2MemOpTable0.insert(std::make_pair((unsigned*)RegOp,
- std::make_pair(MemOp,Align))).second)
- assert(false && "Duplicated entries?");
+ unsigned RegOp = OpTbl0[i][0];
+ unsigned MemOp = OpTbl0[i][1] & ~TB_FLAGS;
unsigned FoldedLoad = OpTbl0[i][2];
+ unsigned Align = OpTbl0[i][3];
+ assert(!RegOp2MemOpTable0.count(RegOp) && "Duplicated entries?");
+ RegOp2MemOpTable0[RegOp] = std::make_pair(MemOp, Align);
+
+ // If this is not a reversable operation (because there is a many->one)
+ // mapping, don't insert the reverse of the operation into MemOp2RegOpTable.
+ if (OpTbl0[i][1] & TB_NOT_REVERSABLE)
+ continue;
+
// Index 0, folded load or store.
unsigned AuxInfo = 0 | (FoldedLoad << 4) | ((FoldedLoad^1) << 5);
- if (RegOp != X86::FsMOVAPDrr && RegOp != X86::FsMOVAPSrr)
- if (!MemOp2RegOpTable.insert(std::make_pair((unsigned*)MemOp,
- std::make_pair(RegOp, AuxInfo))).second)
- AmbEntries.push_back(MemOp);
+ assert(!MemOp2RegOpTable.count(MemOp) && "Duplicated entries?");
+ MemOp2RegOpTable[MemOp] = std::make_pair(RegOp, AuxInfo);
}
static const unsigned OpTbl1[][3] = {
{ X86::CVTTSD2SIrr, X86::CVTTSD2SIrm, 0 },
{ X86::CVTTSS2SI64rr, X86::CVTTSS2SI64rm, 0 },
{ X86::CVTTSS2SIrr, X86::CVTTSS2SIrm, 0 },
- { X86::FsMOVAPDrr, X86::MOVSDrm, 0 },
- { X86::FsMOVAPSrr, X86::MOVSSrm, 0 },
+ { X86::FsMOVAPDrr, X86::MOVSDrm | TB_NOT_REVERSABLE , 0 },
+ { X86::FsMOVAPSrr, X86::MOVSSrm | TB_NOT_REVERSABLE , 0 },
{ X86::IMUL16rri, X86::IMUL16rmi, 0 },
{ X86::IMUL16rri8, X86::IMUL16rmi8, 0 },
{ X86::IMUL32rri, X86::IMUL32rmi, 0 },
{ X86::Int_CVTPD2PSrr, X86::Int_CVTPD2PSrm, 16 },
{ X86::Int_CVTPS2DQrr, X86::Int_CVTPS2DQrm, 16 },
{ X86::Int_CVTPS2PDrr, X86::Int_CVTPS2PDrm, 0 },
- { X86::Int_CVTSD2SI64rr,X86::Int_CVTSD2SI64rm, 0 },
- { X86::Int_CVTSD2SIrr, X86::Int_CVTSD2SIrm, 0 },
+ { X86::CVTSD2SI64rr, X86::CVTSD2SI64rm, 0 },
+ { X86::CVTSD2SIrr, X86::CVTSD2SIrm, 0 },
{ X86::Int_CVTSD2SSrr, X86::Int_CVTSD2SSrm, 0 },
{ X86::Int_CVTSI2SD64rr,X86::Int_CVTSI2SD64rm, 0 },
{ X86::Int_CVTSI2SDrr, X86::Int_CVTSI2SDrm, 0 },
{ X86::Int_CVTSS2SDrr, X86::Int_CVTSS2SDrm, 0 },
{ X86::Int_CVTSS2SI64rr,X86::Int_CVTSS2SI64rm, 0 },
{ X86::Int_CVTSS2SIrr, X86::Int_CVTSS2SIrm, 0 },
- { X86::Int_CVTTPD2DQrr, X86::Int_CVTTPD2DQrm, 16 },
- { X86::Int_CVTTPS2DQrr, X86::Int_CVTTPS2DQrm, 16 },
+ { X86::CVTTPD2DQrr, X86::CVTTPD2DQrm, 16 },
+ { X86::CVTTPS2DQrr, X86::CVTTPS2DQrm, 16 },
{ X86::Int_CVTTSD2SI64rr,X86::Int_CVTTSD2SI64rm, 0 },
{ X86::Int_CVTTSD2SIrr, X86::Int_CVTTSD2SIrm, 0 },
{ X86::Int_CVTTSS2SI64rr,X86::Int_CVTTSS2SI64rm, 0 },
for (unsigned i = 0, e = array_lengthof(OpTbl1); i != e; ++i) {
unsigned RegOp = OpTbl1[i][0];
- unsigned MemOp = OpTbl1[i][1];
+ unsigned MemOp = OpTbl1[i][1] & ~TB_FLAGS;
unsigned Align = OpTbl1[i][2];
- if (!RegOp2MemOpTable1.insert(std::make_pair((unsigned*)RegOp,
- std::make_pair(MemOp,Align))).second)
- assert(false && "Duplicated entries?");
+ assert(!RegOp2MemOpTable1.count(RegOp) && "Duplicate entries");
+ RegOp2MemOpTable1[RegOp] = std::make_pair(MemOp, Align);
+
+ // If this is not a reversable operation (because there is a many->one)
+ // mapping, don't insert the reverse of the operation into MemOp2RegOpTable.
+ if (OpTbl1[i][1] & TB_NOT_REVERSABLE)
+ continue;
+
// Index 1, folded load
unsigned AuxInfo = 1 | (1 << 4);
- if (RegOp != X86::FsMOVAPDrr && RegOp != X86::FsMOVAPSrr)
- if (!MemOp2RegOpTable.insert(std::make_pair((unsigned*)MemOp,
- std::make_pair(RegOp, AuxInfo))).second)
- AmbEntries.push_back(MemOp);
+ assert(!MemOp2RegOpTable.count(MemOp) && "Duplicate entries");
+ MemOp2RegOpTable[MemOp] = std::make_pair(RegOp, AuxInfo);
}
static const unsigned OpTbl2[][3] = {
{ X86::ADC32rr, X86::ADC32rm, 0 },
{ X86::ADC64rr, X86::ADC64rm, 0 },
{ X86::ADD16rr, X86::ADD16rm, 0 },
+ { X86::ADD16rr_DB, X86::ADD16rm | TB_NOT_REVERSABLE, 0 },
{ X86::ADD32rr, X86::ADD32rm, 0 },
+ { X86::ADD32rr_DB, X86::ADD32rm | TB_NOT_REVERSABLE, 0 },
{ X86::ADD64rr, X86::ADD64rm, 0 },
+ { X86::ADD64rr_DB, X86::ADD64rm | TB_NOT_REVERSABLE, 0 },
{ X86::ADD8rr, X86::ADD8rm, 0 },
{ X86::ADDPDrr, X86::ADDPDrm, 16 },
{ X86::ADDPSrr, X86::ADDPSrm, 16 },
for (unsigned i = 0, e = array_lengthof(OpTbl2); i != e; ++i) {
unsigned RegOp = OpTbl2[i][0];
- unsigned MemOp = OpTbl2[i][1];
+ unsigned MemOp = OpTbl2[i][1] & ~TB_FLAGS;
unsigned Align = OpTbl2[i][2];
- if (!RegOp2MemOpTable2.insert(std::make_pair((unsigned*)RegOp,
- std::make_pair(MemOp,Align))).second)
- assert(false && "Duplicated entries?");
+
+ assert(!RegOp2MemOpTable2.count(RegOp) && "Duplicate entry!");
+ RegOp2MemOpTable2[RegOp] = std::make_pair(MemOp, Align);
+
+ // If this is not a reversable operation (because there is a many->one)
+ // mapping, don't insert the reverse of the operation into MemOp2RegOpTable.
+ if (OpTbl2[i][1] & TB_NOT_REVERSABLE)
+ continue;
+
// Index 2, folded load
unsigned AuxInfo = 2 | (1 << 4);
- if (!MemOp2RegOpTable.insert(std::make_pair((unsigned*)MemOp,
- std::make_pair(RegOp, AuxInfo))).second)
- AmbEntries.push_back(MemOp);
+ assert(!MemOp2RegOpTable.count(MemOp) &&
+ "Duplicated entries in unfolding maps?");
+ MemOp2RegOpTable[MemOp] = std::make_pair(RegOp, AuxInfo);
}
-
- // Remove ambiguous entries.
- assert(AmbEntries.empty() && "Duplicated entries in unfolding maps?");
}
bool
unsigned Opc = TM.getSubtarget<X86Subtarget>().is64Bit()
? X86::LEA64_32r : X86::LEA32r;
MachineRegisterInfo &RegInfo = MFI->getParent()->getRegInfo();
- unsigned leaInReg = RegInfo.createVirtualRegister(&X86::GR32RegClass);
+ unsigned leaInReg = RegInfo.createVirtualRegister(&X86::GR32_NOSPRegClass);
unsigned leaOutReg = RegInfo.createVirtualRegister(&X86::GR32RegClass);
// Build and insert into an implicit UNDEF value. This is OK because
break;
case X86::ADD16ri:
case X86::ADD16ri8:
+ case X86::ADD16ri_DB:
+ case X86::ADD16ri8_DB:
addRegOffset(MIB, leaInReg, true, MI->getOperand(2).getImm());
break;
- case X86::ADD16rr: {
+ case X86::ADD16rr:
+ case X86::ADD16rr_DB: {
unsigned Src2 = MI->getOperand(2).getReg();
bool isKill2 = MI->getOperand(2).isKill();
unsigned leaInReg2 = 0;
// just a single insert_subreg.
addRegReg(MIB, leaInReg, true, leaInReg, false);
} else {
- leaInReg2 = RegInfo.createVirtualRegister(&X86::GR32RegClass);
+ leaInReg2 = RegInfo.createVirtualRegister(&X86::GR32_NOSPRegClass);
// Build and insert into an implicit UNDEF value. This is OK because
// well be shifting and then extracting the lower 16-bits.
BuildMI(*MFI, MIB, MI->getDebugLoc(), get(X86::IMPLICIT_DEF), leaInReg2);
unsigned ShAmt = MI->getOperand(2).getImm();
if (ShAmt == 0 || ShAmt >= 4) return 0;
+ // LEA can't handle RSP.
+ if (TargetRegisterInfo::isVirtualRegister(Src) &&
+ !MF.getRegInfo().constrainRegClass(Src, &X86::GR64_NOSPRegClass))
+ return 0;
+
NewMI = BuildMI(MF, MI->getDebugLoc(), get(X86::LEA64r))
.addReg(Dest, RegState::Define | getDeadRegState(isDead))
.addReg(0).addImm(1 << ShAmt)
unsigned ShAmt = MI->getOperand(2).getImm();
if (ShAmt == 0 || ShAmt >= 4) return 0;
+ // LEA can't handle ESP.
+ if (TargetRegisterInfo::isVirtualRegister(Src) &&
+ !MF.getRegInfo().constrainRegClass(Src, &X86::GR32_NOSPRegClass))
+ return 0;
+
unsigned Opc = is64Bit ? X86::LEA64_32r : X86::LEA32r;
NewMI = BuildMI(MF, MI->getDebugLoc(), get(Opc))
.addReg(Dest, RegState::Define | getDeadRegState(isDead))
assert(MI->getNumOperands() >= 2 && "Unknown inc instruction!");
unsigned Opc = MIOpc == X86::INC64r ? X86::LEA64r
: (is64Bit ? X86::LEA64_32r : X86::LEA32r);
+
+ // LEA can't handle RSP.
+ if (TargetRegisterInfo::isVirtualRegister(Src) &&
+ !MF.getRegInfo().constrainRegClass(Src,
+ MIOpc == X86::INC64r ? X86::GR64_NOSPRegisterClass :
+ X86::GR32_NOSPRegisterClass))
+ return 0;
+
NewMI = addRegOffset(BuildMI(MF, MI->getDebugLoc(), get(Opc))
.addReg(Dest, RegState::Define |
getDeadRegState(isDead)),
assert(MI->getNumOperands() >= 2 && "Unknown dec instruction!");
unsigned Opc = MIOpc == X86::DEC64r ? X86::LEA64r
: (is64Bit ? X86::LEA64_32r : X86::LEA32r);
+ // LEA can't handle RSP.
+ if (TargetRegisterInfo::isVirtualRegister(Src) &&
+ !MF.getRegInfo().constrainRegClass(Src,
+ MIOpc == X86::DEC64r ? X86::GR64_NOSPRegisterClass :
+ X86::GR32_NOSPRegisterClass))
+ return 0;
+
NewMI = addRegOffset(BuildMI(MF, MI->getDebugLoc(), get(Opc))
.addReg(Dest, RegState::Define |
getDeadRegState(isDead)),
Src, isKill, -1);
break;
case X86::ADD64rr:
- case X86::ADD32rr: {
+ case X86::ADD64rr_DB:
+ case X86::ADD32rr:
+ case X86::ADD32rr_DB: {
assert(MI->getNumOperands() >= 3 && "Unknown add instruction!");
- unsigned Opc = MIOpc == X86::ADD64rr ? X86::LEA64r
- : (is64Bit ? X86::LEA64_32r : X86::LEA32r);
+ unsigned Opc;
+ TargetRegisterClass *RC;
+ if (MIOpc == X86::ADD64rr || MIOpc == X86::ADD64rr_DB) {
+ Opc = X86::LEA64r;
+ RC = X86::GR64_NOSPRegisterClass;
+ } else {
+ Opc = is64Bit ? X86::LEA64_32r : X86::LEA32r;
+ RC = X86::GR32_NOSPRegisterClass;
+ }
+
+
unsigned Src2 = MI->getOperand(2).getReg();
bool isKill2 = MI->getOperand(2).isKill();
+
+ // LEA can't handle RSP.
+ if (TargetRegisterInfo::isVirtualRegister(Src2) &&
+ !MF.getRegInfo().constrainRegClass(Src2, RC))
+ return 0;
+
NewMI = addRegReg(BuildMI(MF, MI->getDebugLoc(), get(Opc))
.addReg(Dest, RegState::Define |
getDeadRegState(isDead)),
LV->replaceKillInstruction(Src2, MI, NewMI);
break;
}
- case X86::ADD16rr: {
+ case X86::ADD16rr:
+ case X86::ADD16rr_DB: {
if (DisableLEA16)
return is64Bit ? convertToThreeAddressWithLEA(MIOpc, MFI, MBBI, LV) : 0;
assert(MI->getNumOperands() >= 3 && "Unknown add instruction!");
}
case X86::ADD64ri32:
case X86::ADD64ri8:
+ case X86::ADD64ri32_DB:
+ case X86::ADD64ri8_DB:
assert(MI->getNumOperands() >= 3 && "Unknown add instruction!");
NewMI = addRegOffset(BuildMI(MF, MI->getDebugLoc(), get(X86::LEA64r))
.addReg(Dest, RegState::Define |
Src, isKill, MI->getOperand(2).getImm());
break;
case X86::ADD32ri:
- case X86::ADD32ri8: {
+ case X86::ADD32ri8:
+ case X86::ADD32ri_DB:
+ case X86::ADD32ri8_DB: {
assert(MI->getNumOperands() >= 3 && "Unknown add instruction!");
unsigned Opc = is64Bit ? X86::LEA64_32r : X86::LEA32r;
NewMI = addRegOffset(BuildMI(MF, MI->getDebugLoc(), get(Opc))
}
case X86::ADD16ri:
case X86::ADD16ri8:
+ case X86::ADD16ri_DB:
+ case X86::ADD16ri8_DB:
if (DisableLEA16)
return is64Bit ? convertToThreeAddressWithLEA(MIOpc, MFI, MBBI, LV) : 0;
assert(MI->getNumOperands() >= 3 && "Unknown add instruction!");
return X86::GR8_ABCD_HRegClass.contains(Reg);
}
+// Try and copy between VR128/VR64 and GR64 registers.
+static unsigned CopyToFromAsymmetricReg(unsigned DestReg, unsigned SrcReg) {
+ // SrcReg(VR128) -> DestReg(GR64)
+ // SrcReg(VR64) -> DestReg(GR64)
+ // SrcReg(GR64) -> DestReg(VR128)
+ // SrcReg(GR64) -> DestReg(VR64)
+
+ if (X86::GR64RegClass.contains(DestReg)) {
+ if (X86::VR128RegClass.contains(SrcReg)) {
+ // Copy from a VR128 register to a GR64 register.
+ return X86::MOVPQIto64rr;
+ } else if (X86::VR64RegClass.contains(SrcReg)) {
+ // Copy from a VR64 register to a GR64 register.
+ return X86::MOVSDto64rr;
+ }
+ } else if (X86::GR64RegClass.contains(SrcReg)) {
+ // Copy from a GR64 register to a VR128 register.
+ if (X86::VR128RegClass.contains(DestReg))
+ return X86::MOV64toPQIrr;
+ // Copy from a GR64 register to a VR64 register.
+ else if (X86::VR64RegClass.contains(DestReg))
+ return X86::MOV64toSDrr;
+ }
+
+ return 0;
+}
+
void X86InstrInfo::copyPhysReg(MachineBasicBlock &MBB,
MachineBasicBlock::iterator MI, DebugLoc DL,
unsigned DestReg, unsigned SrcReg,
Opc = X86::MOVAPSrr;
else if (X86::VR64RegClass.contains(DestReg, SrcReg))
Opc = X86::MMX_MOVQ64rr;
+ else
+ Opc = CopyToFromAsymmetricReg(DestReg, SrcReg);
if (Opc) {
BuildMI(MBB, MI, DL, get(Opc), DestReg)
MachineInstr *MI, unsigned i,
const SmallVectorImpl<MachineOperand> &MOs,
unsigned Size, unsigned Align) const {
- const DenseMap<unsigned*, std::pair<unsigned,unsigned> > *OpcodeTablePtr=NULL;
+ const DenseMap<unsigned, std::pair<unsigned,unsigned> > *OpcodeTablePtr = 0;
bool isTwoAddrFold = false;
unsigned NumOps = MI->getDesc().getNumOperands();
bool isTwoAddr = NumOps > 1 &&
// If table selected...
if (OpcodeTablePtr) {
// Find the Opcode to fuse
- DenseMap<unsigned*, std::pair<unsigned,unsigned> >::const_iterator I =
- OpcodeTablePtr->find((unsigned*)MI->getOpcode());
+ DenseMap<unsigned, std::pair<unsigned,unsigned> >::const_iterator I =
+ OpcodeTablePtr->find(MI->getOpcode());
if (I != OpcodeTablePtr->end()) {
unsigned Opcode = I->second.first;
unsigned MinAlign = I->second.second;
case X86::Int_CVTSS2SDrr:
case X86::RCPSSr:
case X86::RCPSSr_Int:
- case X86::ROUNDSDr_Int:
- case X86::ROUNDSSr_Int:
+ case X86::ROUNDSDr:
+ case X86::ROUNDSSr:
case X86::RSQRTSSr:
case X86::RSQRTSSr_Int:
case X86::SQRTSSr:
case X86::Int_CVTSS2SDrr:
case X86::RCPSSr:
case X86::RCPSSr_Int:
- case X86::ROUNDSDr_Int:
- case X86::ROUNDSSr_Int:
+ case X86::ROUNDSDr:
+ case X86::ROUNDSSr:
case X86::RSQRTSSr:
case X86::RSQRTSSr_Int:
case X86::SQRTSSr:
Alignment = (*LoadMI->memoperands_begin())->getAlignment();
else
switch (LoadMI->getOpcode()) {
+ case X86::AVX_SET0PSY:
+ case X86::AVX_SET0PDY:
+ Alignment = 32;
+ break;
case X86::V_SET0PS:
case X86::V_SET0PD:
case X86::V_SET0PI:
case X86::V_SETALLONES:
+ case X86::AVX_SET0PS:
+ case X86::AVX_SET0PD:
+ case X86::AVX_SET0PI:
Alignment = 16;
break;
case X86::FsFLD0SD:
case X86::V_SET0PD:
case X86::V_SET0PI:
case X86::V_SETALLONES:
+ case X86::AVX_SET0PS:
+ case X86::AVX_SET0PD:
+ case X86::AVX_SET0PI:
+ case X86::AVX_SET0PSY:
+ case X86::AVX_SET0PDY:
case X86::FsFLD0SD:
case X86::FsFLD0SS: {
// Folding a V_SET0P? or V_SETALLONES as a load, to ease register pressure.
// Create a constant-pool entry.
MachineConstantPool &MCP = *MF.getConstantPool();
const Type *Ty;
- if (LoadMI->getOpcode() == X86::FsFLD0SS)
+ unsigned Opc = LoadMI->getOpcode();
+ if (Opc == X86::FsFLD0SS)
Ty = Type::getFloatTy(MF.getFunction()->getContext());
- else if (LoadMI->getOpcode() == X86::FsFLD0SD)
+ else if (Opc == X86::FsFLD0SD)
Ty = Type::getDoubleTy(MF.getFunction()->getContext());
+ else if (Opc == X86::AVX_SET0PSY || Opc == X86::AVX_SET0PDY)
+ Ty = VectorType::get(Type::getFloatTy(MF.getFunction()->getContext()), 8);
else
Ty = VectorType::get(Type::getInt32Ty(MF.getFunction()->getContext()), 4);
const Constant *C = LoadMI->getOpcode() == X86::V_SETALLONES ?
// Folding a memory location into the two-address part of a two-address
// instruction is different than folding it other places. It requires
// replacing the *two* registers with the memory location.
- const DenseMap<unsigned*, std::pair<unsigned,unsigned> > *OpcodeTablePtr=NULL;
+ const DenseMap<unsigned, std::pair<unsigned,unsigned> > *OpcodeTablePtr = 0;
if (isTwoAddr && NumOps >= 2 && OpNum < 2) {
OpcodeTablePtr = &RegOp2MemOpTable2Addr;
} else if (OpNum == 0) { // If operand 0
case X86::MOV8r0:
case X86::MOV16r0:
case X86::MOV32r0:
- case X86::MOV64r0:
- return true;
+ case X86::MOV64r0: return true;
default: break;
}
OpcodeTablePtr = &RegOp2MemOpTable0;
OpcodeTablePtr = &RegOp2MemOpTable2;
}
- if (OpcodeTablePtr) {
- // Find the Opcode to fuse
- DenseMap<unsigned*, std::pair<unsigned,unsigned> >::const_iterator I =
- OpcodeTablePtr->find((unsigned*)Opc);
- if (I != OpcodeTablePtr->end())
- return true;
- }
+ if (OpcodeTablePtr && OpcodeTablePtr->count(Opc))
+ return true;
return TargetInstrInfoImpl::canFoldMemoryOperand(MI, Ops);
}
bool X86InstrInfo::unfoldMemoryOperand(MachineFunction &MF, MachineInstr *MI,
unsigned Reg, bool UnfoldLoad, bool UnfoldStore,
SmallVectorImpl<MachineInstr*> &NewMIs) const {
- DenseMap<unsigned*, std::pair<unsigned,unsigned> >::const_iterator I =
- MemOp2RegOpTable.find((unsigned*)MI->getOpcode());
+ DenseMap<unsigned, std::pair<unsigned,unsigned> >::const_iterator I =
+ MemOp2RegOpTable.find(MI->getOpcode());
if (I == MemOp2RegOpTable.end())
return false;
unsigned Opc = I->second.first;
if (!N->isMachineOpcode())
return false;
- DenseMap<unsigned*, std::pair<unsigned,unsigned> >::const_iterator I =
- MemOp2RegOpTable.find((unsigned*)N->getMachineOpcode());
+ DenseMap<unsigned, std::pair<unsigned,unsigned> >::const_iterator I =
+ MemOp2RegOpTable.find(N->getMachineOpcode());
if (I == MemOp2RegOpTable.end())
return false;
unsigned Opc = I->second.first;
unsigned X86InstrInfo::getOpcodeAfterMemoryUnfold(unsigned Opc,
bool UnfoldLoad, bool UnfoldStore,
unsigned *LoadRegIndex) const {
- DenseMap<unsigned*, std::pair<unsigned,unsigned> >::const_iterator I =
- MemOp2RegOpTable.find((unsigned*)Opc);
+ DenseMap<unsigned, std::pair<unsigned,unsigned> >::const_iterator I =
+ MemOp2RegOpTable.find(Opc);
if (I == MemOp2RegOpTable.end())
return 0;
bool FoldedLoad = I->second.second & (1 << 4);
case X86::XMM12: case X86::XMM13: case X86::XMM14: case X86::XMM15:
case X86::YMM8: case X86::YMM9: case X86::YMM10: case X86::YMM11:
case X86::YMM12: case X86::YMM13: case X86::YMM14: case X86::YMM15:
+ case X86::CR8: case X86::CR9: case X86::CR10: case X86::CR11:
+ case X86::CR12: case X86::CR13: case X86::CR14: case X86::CR15:
return true;
}
return false;
{ X86::V_SET0PS, X86::V_SET0PD, X86::V_SET0PI },
{ X86::XORPSrm, X86::XORPDrm, X86::PXORrm },
{ X86::XORPSrr, X86::XORPDrr, X86::PXORrr },
+ // AVX 128-bit support
+ { X86::VMOVAPSmr, X86::VMOVAPDmr, X86::VMOVDQAmr },
+ { X86::VMOVAPSrm, X86::VMOVAPDrm, X86::VMOVDQArm },
+ { X86::VMOVAPSrr, X86::VMOVAPDrr, X86::VMOVDQArr },
+ { X86::VMOVUPSmr, X86::VMOVUPDmr, X86::VMOVDQUmr },
+ { X86::VMOVUPSrm, X86::VMOVUPDrm, X86::VMOVDQUrm },
+ { X86::VMOVNTPSmr, X86::VMOVNTPDmr, X86::VMOVNTDQmr },
+ { X86::VANDNPSrm, X86::VANDNPDrm, X86::VPANDNrm },
+ { X86::VANDNPSrr, X86::VANDNPDrr, X86::VPANDNrr },
+ { X86::VANDPSrm, X86::VANDPDrm, X86::VPANDrm },
+ { X86::VANDPSrr, X86::VANDPDrr, X86::VPANDrr },
+ { X86::VORPSrm, X86::VORPDrm, X86::VPORrm },
+ { X86::VORPSrr, X86::VORPDrr, X86::VPORrr },
+ { X86::AVX_SET0PS, X86::AVX_SET0PD, X86::AVX_SET0PI },
+ { X86::VXORPSrm, X86::VXORPDrm, X86::VPXORrm },
+ { X86::VXORPSrr, X86::VXORPDrr, X86::VPXORrr },
};
// FIXME: Some shuffle and unpack instructions have equivalents in different
if (TM->getRelocationModel() != Reloc::PIC_)
return false;
+ X86MachineFunctionInfo *X86FI = MF.getInfo<X86MachineFunctionInfo>();
+ unsigned GlobalBaseReg = X86FI->getGlobalBaseReg();
+
+ // If we didn't need a GlobalBaseReg, don't insert code.
+ if (GlobalBaseReg == 0)
+ return false;
+
// Insert the set of GlobalBaseReg into the first MBB of the function
MachineBasicBlock &FirstMBB = MF.front();
MachineBasicBlock::iterator MBBI = FirstMBB.begin();
if (TM->getSubtarget<X86Subtarget>().isPICStyleGOT())
PC = RegInfo.createVirtualRegister(X86::GR32RegisterClass);
else
- PC = TII->getGlobalBaseReg(&MF);
+ PC = GlobalBaseReg;
// Operand of MovePCtoStack is completely ignored by asm printer. It's
// only used in JIT code emission as displacement to pc.
// If we're using vanilla 'GOT' PIC style, we should use relative addressing
// not to pc, but to _GLOBAL_OFFSET_TABLE_ external.
if (TM->getSubtarget<X86Subtarget>().isPICStyleGOT()) {
- unsigned GlobalBaseReg = TII->getGlobalBaseReg(&MF);
// Generate addl $__GLOBAL_OFFSET_TABLE_ + [.-piclabel], %some_register
BuildMI(FirstMBB, MBBI, DL, TII->get(X86::ADD32ri), GlobalBaseReg)
.addReg(PC).addExternalSymbol("_GLOBAL_OFFSET_TABLE_",
projects / oota-llvm.git / blobdiff