#define DEBUG_TYPE "arm-pseudo"
#include "ARM.h"
+#include "ARMAddressingModes.h"
#include "ARMBaseInstrInfo.h"
+#include "ARMBaseRegisterInfo.h"
+#include "ARMMachineFunctionInfo.h"
+#include "ARMRegisterInfo.h"
+#include "llvm/CodeGen/MachineFrameInfo.h"
#include "llvm/CodeGen/MachineFunctionPass.h"
#include "llvm/CodeGen/MachineInstrBuilder.h"
+#include "llvm/Target/TargetFrameInfo.h"
#include "llvm/Target/TargetRegisterInfo.h"
+#include "llvm/Support/raw_ostream.h" // FIXME: for debug only. remove!
using namespace llvm;
namespace {
static char ID;
ARMExpandPseudo() : MachineFunctionPass(ID) {}
- const TargetInstrInfo *TII;
+ const ARMBaseInstrInfo *TII;
const TargetRegisterInfo *TRI;
+ const ARMSubtarget *STI;
virtual bool runOnMachineFunction(MachineFunction &Fn);
void ExpandVLD(MachineBasicBlock::iterator &MBBI);
void ExpandVST(MachineBasicBlock::iterator &MBBI);
void ExpandLaneOp(MachineBasicBlock::iterator &MBBI);
+ void ExpandVTBL(MachineBasicBlock::iterator &MBBI,
+ unsigned Opc, bool IsExt, unsigned NumRegs);
};
char ARMExpandPseudo::ID = 0;
}
friend bool operator<(const NEONLdStTableEntry &TE, unsigned PseudoOpc) {
return TE.PseudoOpc < PseudoOpc;
}
- friend bool ATTRIBUTE_UNUSED operator<(unsigned PseudoOpc,
- const NEONLdStTableEntry &TE) {
+ friend bool LLVM_ATTRIBUTE_UNUSED operator<(unsigned PseudoOpc,
+ const NEONLdStTableEntry &TE) {
return PseudoOpc < TE.PseudoOpc;
}
};
}
static const NEONLdStTableEntry NEONLdStTable[] = {
+{ ARM::VLD1DUPq16Pseudo, ARM::VLD1DUPq16, true, false, SingleSpc, 2, 4},
+{ ARM::VLD1DUPq16Pseudo_UPD, ARM::VLD1DUPq16_UPD, true, true, SingleSpc, 2, 4},
+{ ARM::VLD1DUPq32Pseudo, ARM::VLD1DUPq32, true, false, SingleSpc, 2, 2},
+{ ARM::VLD1DUPq32Pseudo_UPD, ARM::VLD1DUPq32_UPD, true, true, SingleSpc, 2, 2},
+{ ARM::VLD1DUPq8Pseudo, ARM::VLD1DUPq8, true, false, SingleSpc, 2, 8},
+{ ARM::VLD1DUPq8Pseudo_UPD, ARM::VLD1DUPq8_UPD, true, true, SingleSpc, 2, 8},
+
+{ ARM::VLD1LNq16Pseudo, ARM::VLD1LNd16, true, false, EvenDblSpc, 1, 4 },
+{ ARM::VLD1LNq16Pseudo_UPD, ARM::VLD1LNd16_UPD, true, true, EvenDblSpc, 1, 4 },
+{ ARM::VLD1LNq32Pseudo, ARM::VLD1LNd32, true, false, EvenDblSpc, 1, 2 },
+{ ARM::VLD1LNq32Pseudo_UPD, ARM::VLD1LNd32_UPD, true, true, EvenDblSpc, 1, 2 },
+{ ARM::VLD1LNq8Pseudo, ARM::VLD1LNd8, true, false, EvenDblSpc, 1, 8 },
+{ ARM::VLD1LNq8Pseudo_UPD, ARM::VLD1LNd8_UPD, true, true, EvenDblSpc, 1, 8 },
+
{ ARM::VLD1d64QPseudo, ARM::VLD1d64Q, true, false, SingleSpc, 4, 1 },
{ ARM::VLD1d64QPseudo_UPD, ARM::VLD1d64Q_UPD, true, true, SingleSpc, 4, 1 },
{ ARM::VLD1d64TPseudo, ARM::VLD1d64T, true, false, SingleSpc, 3, 1 },
{ ARM::VLD1q8Pseudo, ARM::VLD1q8, true, false, SingleSpc, 2, 8 },
{ ARM::VLD1q8Pseudo_UPD, ARM::VLD1q8_UPD, true, true, SingleSpc, 2, 8 },
+{ ARM::VLD2DUPd16Pseudo, ARM::VLD2DUPd16, true, false, SingleSpc, 2, 4},
+{ ARM::VLD2DUPd16Pseudo_UPD, ARM::VLD2DUPd16_UPD, true, true, SingleSpc, 2, 4},
+{ ARM::VLD2DUPd32Pseudo, ARM::VLD2DUPd32, true, false, SingleSpc, 2, 2},
+{ ARM::VLD2DUPd32Pseudo_UPD, ARM::VLD2DUPd32_UPD, true, true, SingleSpc, 2, 2},
+{ ARM::VLD2DUPd8Pseudo, ARM::VLD2DUPd8, true, false, SingleSpc, 2, 8},
+{ ARM::VLD2DUPd8Pseudo_UPD, ARM::VLD2DUPd8_UPD, true, true, SingleSpc, 2, 8},
+
{ ARM::VLD2LNd16Pseudo, ARM::VLD2LNd16, true, false, SingleSpc, 2, 4 },
{ ARM::VLD2LNd16Pseudo_UPD, ARM::VLD2LNd16_UPD, true, true, SingleSpc, 2, 4 },
{ ARM::VLD2LNd32Pseudo, ARM::VLD2LNd32, true, false, SingleSpc, 2, 2 },
{ ARM::VLD2q8Pseudo, ARM::VLD2q8, true, false, SingleSpc, 4, 8 },
{ ARM::VLD2q8Pseudo_UPD, ARM::VLD2q8_UPD, true, true, SingleSpc, 4, 8 },
+{ ARM::VLD3DUPd16Pseudo, ARM::VLD3DUPd16, true, false, SingleSpc, 3, 4},
+{ ARM::VLD3DUPd16Pseudo_UPD, ARM::VLD3DUPd16_UPD, true, true, SingleSpc, 3, 4},
+{ ARM::VLD3DUPd32Pseudo, ARM::VLD3DUPd32, true, false, SingleSpc, 3, 2},
+{ ARM::VLD3DUPd32Pseudo_UPD, ARM::VLD3DUPd32_UPD, true, true, SingleSpc, 3, 2},
+{ ARM::VLD3DUPd8Pseudo, ARM::VLD3DUPd8, true, false, SingleSpc, 3, 8},
+{ ARM::VLD3DUPd8Pseudo_UPD, ARM::VLD3DUPd8_UPD, true, true, SingleSpc, 3, 8},
+
{ ARM::VLD3LNd16Pseudo, ARM::VLD3LNd16, true, false, SingleSpc, 3, 4 },
{ ARM::VLD3LNd16Pseudo_UPD, ARM::VLD3LNd16_UPD, true, true, SingleSpc, 3, 4 },
{ ARM::VLD3LNd32Pseudo, ARM::VLD3LNd32, true, false, SingleSpc, 3, 2 },
{ ARM::VLD3q8Pseudo_UPD, ARM::VLD3q8_UPD, true, true, EvenDblSpc, 3, 8 },
{ ARM::VLD3q8oddPseudo_UPD, ARM::VLD3q8_UPD, true, true, OddDblSpc, 3, 8 },
+{ ARM::VLD4DUPd16Pseudo, ARM::VLD4DUPd16, true, false, SingleSpc, 4, 4},
+{ ARM::VLD4DUPd16Pseudo_UPD, ARM::VLD4DUPd16_UPD, true, true, SingleSpc, 4, 4},
+{ ARM::VLD4DUPd32Pseudo, ARM::VLD4DUPd32, true, false, SingleSpc, 4, 2},
+{ ARM::VLD4DUPd32Pseudo_UPD, ARM::VLD4DUPd32_UPD, true, true, SingleSpc, 4, 2},
+{ ARM::VLD4DUPd8Pseudo, ARM::VLD4DUPd8, true, false, SingleSpc, 4, 8},
+{ ARM::VLD4DUPd8Pseudo_UPD, ARM::VLD4DUPd8_UPD, true, true, SingleSpc, 4, 8},
+
{ ARM::VLD4LNd16Pseudo, ARM::VLD4LNd16, true, false, SingleSpc, 4, 4 },
{ ARM::VLD4LNd16Pseudo_UPD, ARM::VLD4LNd16_UPD, true, true, SingleSpc, 4, 4 },
{ ARM::VLD4LNd32Pseudo, ARM::VLD4LNd32, true, false, SingleSpc, 4, 2 },
{ ARM::VLD4q8Pseudo_UPD, ARM::VLD4q8_UPD, true, true, EvenDblSpc, 4, 8 },
{ ARM::VLD4q8oddPseudo_UPD, ARM::VLD4q8_UPD, true, true, OddDblSpc, 4, 8 },
+{ ARM::VST1LNq16Pseudo, ARM::VST1LNd16, false, false, EvenDblSpc, 1, 4 },
+{ ARM::VST1LNq16Pseudo_UPD, ARM::VST1LNd16_UPD,false, true, EvenDblSpc, 1, 4 },
+{ ARM::VST1LNq32Pseudo, ARM::VST1LNd32, false, false, EvenDblSpc, 1, 2 },
+{ ARM::VST1LNq32Pseudo_UPD, ARM::VST1LNd32_UPD,false, true, EvenDblSpc, 1, 2 },
+{ ARM::VST1LNq8Pseudo, ARM::VST1LNd8, false, false, EvenDblSpc, 1, 8 },
+{ ARM::VST1LNq8Pseudo_UPD, ARM::VST1LNd8_UPD, false, true, EvenDblSpc, 1, 8 },
+
{ ARM::VST1d64QPseudo, ARM::VST1d64Q, false, false, SingleSpc, 4, 1 },
{ ARM::VST1d64QPseudo_UPD, ARM::VST1d64Q_UPD, false, true, SingleSpc, 4, 1 },
{ ARM::VST1d64TPseudo, ARM::VST1d64T, false, false, SingleSpc, 3, 1 },
D1 = TRI->getSubReg(Reg, ARM::dsub_3);
D2 = TRI->getSubReg(Reg, ARM::dsub_5);
D3 = TRI->getSubReg(Reg, ARM::dsub_7);
- }
+ }
}
/// ExpandVLD - Translate VLD pseudo instructions with Q, QQ or QQQQ register
if (TableEntry->HasWriteBack)
MIB.addOperand(MI.getOperand(OpIdx++));
- MIB = AddDefaultPred(MIB);
// For an instruction writing double-spaced subregs, the pseudo instruction
- // has an extra operand that is a use of the super-register. Copy that over
+ // has an extra operand that is a use of the super-register. Record the
+ // operand index and skip over it.
+ unsigned SrcOpIdx = 0;
+ if (RegSpc == EvenDblSpc || RegSpc == OddDblSpc)
+ SrcOpIdx = OpIdx++;
+
+ // Copy the predicate operands.
+ MIB.addOperand(MI.getOperand(OpIdx++));
+ MIB.addOperand(MI.getOperand(OpIdx++));
+
+ // Copy the super-register source operand used for double-spaced subregs over
// to the new instruction as an implicit operand.
- if (RegSpc == EvenDblSpc || RegSpc == OddDblSpc) {
- MachineOperand MO = MI.getOperand(OpIdx);
+ if (SrcOpIdx != 0) {
+ MachineOperand MO = MI.getOperand(SrcOpIdx);
MO.setImplicit(true);
MIB.addOperand(MO);
}
MIB.addOperand(MI.getOperand(OpIdx++));
bool SrcIsKill = MI.getOperand(OpIdx).isKill();
- unsigned SrcReg = MI.getOperand(OpIdx).getReg();
+ unsigned SrcReg = MI.getOperand(OpIdx++).getReg();
unsigned D0, D1, D2, D3;
GetDSubRegs(SrcReg, RegSpc, TRI, D0, D1, D2, D3);
MIB.addReg(D0).addReg(D1);
MIB.addReg(D2);
if (NumRegs > 3)
MIB.addReg(D3);
- MIB = AddDefaultPred(MIB);
- TransferImpOps(MI, MIB, MIB);
+
+ // Copy the predicate operands.
+ MIB.addOperand(MI.getOperand(OpIdx++));
+ MIB.addOperand(MI.getOperand(OpIdx++));
+
if (SrcIsKill)
// Add an implicit kill for the super-reg.
(*MIB).addRegisterKilled(SrcReg, TRI, true);
+ TransferImpOps(MI, MIB, MIB);
MI.eraseFromParent();
}
}
assert(Lane < RegElts && "out of range lane for VLD/VST-lane");
- unsigned DstReg, D0, D1, D2, D3;
- bool DstIsDead;
+ unsigned D0, D1, D2, D3;
+ unsigned DstReg = 0;
+ bool DstIsDead = false;
if (TableEntry->IsLoad) {
DstIsDead = MI.getOperand(OpIdx).isDead();
DstReg = MI.getOperand(OpIdx++).getReg();
GetDSubRegs(DstReg, RegSpc, TRI, D0, D1, D2, D3);
- MIB.addReg(D0, RegState::Define | getDeadRegState(DstIsDead))
- .addReg(D1, RegState::Define | getDeadRegState(DstIsDead));
+ MIB.addReg(D0, RegState::Define | getDeadRegState(DstIsDead));
+ if (NumRegs > 1)
+ MIB.addReg(D1, RegState::Define | getDeadRegState(DstIsDead));
if (NumRegs > 2)
MIB.addReg(D2, RegState::Define | getDeadRegState(DstIsDead));
if (NumRegs > 3)
// Add the subregs as sources of the new instruction.
unsigned SrcFlags = (getUndefRegState(MO.isUndef()) |
getKillRegState(MO.isKill()));
- MIB.addReg(D0, SrcFlags).addReg(D1, SrcFlags);
+ MIB.addReg(D0, SrcFlags);
+ if (NumRegs > 1)
+ MIB.addReg(D1, SrcFlags);
if (NumRegs > 2)
MIB.addReg(D2, SrcFlags);
if (NumRegs > 3)
// Add the lane number operand.
MIB.addImm(Lane);
+ OpIdx += 1;
+
+ // Copy the predicate operands.
+ MIB.addOperand(MI.getOperand(OpIdx++));
+ MIB.addOperand(MI.getOperand(OpIdx++));
- MIB = AddDefaultPred(MIB);
// Copy the super-register source to be an implicit source.
MO.setImplicit(true);
MIB.addOperand(MO);
MI.eraseFromParent();
}
+/// ExpandVTBL - Translate VTBL and VTBX pseudo instructions with Q or QQ
+/// register operands to real instructions with D register operands.
+void ARMExpandPseudo::ExpandVTBL(MachineBasicBlock::iterator &MBBI,
+ unsigned Opc, bool IsExt, unsigned NumRegs) {
+ MachineInstr &MI = *MBBI;
+ MachineBasicBlock &MBB = *MI.getParent();
+
+ MachineInstrBuilder MIB = BuildMI(MBB, MBBI, MI.getDebugLoc(), TII->get(Opc));
+ unsigned OpIdx = 0;
+
+ // Transfer the destination register operand.
+ MIB.addOperand(MI.getOperand(OpIdx++));
+ if (IsExt)
+ MIB.addOperand(MI.getOperand(OpIdx++));
+
+ bool SrcIsKill = MI.getOperand(OpIdx).isKill();
+ unsigned SrcReg = MI.getOperand(OpIdx++).getReg();
+ unsigned D0, D1, D2, D3;
+ GetDSubRegs(SrcReg, SingleSpc, TRI, D0, D1, D2, D3);
+ MIB.addReg(D0).addReg(D1);
+ if (NumRegs > 2)
+ MIB.addReg(D2);
+ if (NumRegs > 3)
+ MIB.addReg(D3);
+
+ // Copy the other source register operand.
+ MIB.addOperand(MI.getOperand(OpIdx++));
+
+ // Copy the predicate operands.
+ MIB.addOperand(MI.getOperand(OpIdx++));
+ MIB.addOperand(MI.getOperand(OpIdx++));
+
+ if (SrcIsKill)
+ // Add an implicit kill for the super-reg.
+ (*MIB).addRegisterKilled(SrcReg, TRI, true);
+ TransferImpOps(MI, MIB, MIB);
+ MI.eraseFromParent();
+}
+
bool ARMExpandPseudo::ExpandMBB(MachineBasicBlock &MBB) {
bool Modified = false;
ModifiedOp = false;
break;
- case ARM::tLDRpci_pic:
+ case ARM::Int_eh_sjlj_dispatchsetup: {
+ MachineFunction &MF = *MI.getParent()->getParent();
+ const ARMBaseInstrInfo *AII =
+ static_cast<const ARMBaseInstrInfo*>(TII);
+ const ARMBaseRegisterInfo &RI = AII->getRegisterInfo();
+ // For functions using a base pointer, we rematerialize it (via the frame
+ // pointer) here since eh.sjlj.setjmp and eh.sjlj.longjmp don't do it
+ // for us. Otherwise, expand to nothing.
+ if (RI.hasBasePointer(MF)) {
+ ARMFunctionInfo *AFI = MF.getInfo<ARMFunctionInfo>();
+ int32_t NumBytes = AFI->getFramePtrSpillOffset();
+ unsigned FramePtr = RI.getFrameRegister(MF);
+ assert(MF.getTarget().getFrameInfo()->hasFP(MF) &&
+ "base pointer without frame pointer?");
+
+ if (AFI->isThumb2Function()) {
+ llvm::emitT2RegPlusImmediate(MBB, MBBI, MI.getDebugLoc(), ARM::R6,
+ FramePtr, -NumBytes, ARMCC::AL, 0, *TII);
+ } else if (AFI->isThumbFunction()) {
+ llvm::emitThumbRegPlusImmediate(MBB, MBBI, ARM::R6,
+ FramePtr, -NumBytes,
+ *TII, RI, MI.getDebugLoc());
+ } else {
+ llvm::emitARMRegPlusImmediate(MBB, MBBI, MI.getDebugLoc(), ARM::R6,
+ FramePtr, -NumBytes, ARMCC::AL, 0,
+ *TII);
+ }
+ // If there's dynamic realignment, adjust for it.
+ if (RI.needsStackRealignment(MF)) {
+ MachineFrameInfo *MFI = MF.getFrameInfo();
+ unsigned MaxAlign = MFI->getMaxAlignment();
+ assert (!AFI->isThumb1OnlyFunction());
+ // Emit bic r6, r6, MaxAlign
+ unsigned bicOpc = AFI->isThumbFunction() ?
+ ARM::t2BICri : ARM::BICri;
+ AddDefaultCC(AddDefaultPred(BuildMI(MBB, MBBI, MI.getDebugLoc(),
+ TII->get(bicOpc), ARM::R6)
+ .addReg(ARM::R6, RegState::Kill)
+ .addImm(MaxAlign-1)));
+ }
+
+ }
+ MI.eraseFromParent();
+ break;
+ }
+
+ case ARM::MOVsrl_flag:
+ case ARM::MOVsra_flag: {
+ // These are just fancy MOVs insructions.
+ AddDefaultPred(BuildMI(MBB, MBBI, MI.getDebugLoc(), TII->get(ARM::MOVs),
+ MI.getOperand(0).getReg())
+ .addOperand(MI.getOperand(1))
+ .addReg(0)
+ .addImm(ARM_AM::getSORegOpc((Opcode == ARM::MOVsrl_flag ? ARM_AM::lsr
+ : ARM_AM::asr), 1)))
+ .addReg(ARM::CPSR, RegState::Define);
+ MI.eraseFromParent();
+ break;
+ }
+ case ARM::RRX: {
+ // This encodes as "MOVs Rd, Rm, rrx
+ MachineInstrBuilder MIB =
+ AddDefaultPred(BuildMI(MBB, MBBI, MI.getDebugLoc(), TII->get(ARM::MOVs),
+ MI.getOperand(0).getReg())
+ .addOperand(MI.getOperand(1))
+ .addOperand(MI.getOperand(1))
+ .addImm(ARM_AM::getSORegOpc(ARM_AM::rrx, 0)))
+ .addReg(0);
+ TransferImpOps(MI, MIB, MIB);
+ MI.eraseFromParent();
+ break;
+ }
+ case ARM::TPsoft: {
+ unsigned PredReg = 0;
+ ARMCC::CondCodes Pred = llvm::getInstrPredicate(&MI, PredReg);
+ MachineInstrBuilder MIB =
+ BuildMI(MBB, MBBI, MI.getDebugLoc(),
+ TII->get(ARM::BL))
+ .addExternalSymbol("__aeabi_read_tp", 0);
+
+ (*MIB).setMemRefs(MI.memoperands_begin(), MI.memoperands_end());
+ TransferImpOps(MI, MIB, MIB);
+ MI.eraseFromParent();
+ break;
+ };
+ case ARM::t2LDRHpci:
+ case ARM::t2LDRBpci:
+ case ARM::t2LDRSHpci:
+ case ARM::t2LDRSBpci:
+ case ARM::t2LDRpci: {
+ unsigned NewLdOpc;
+ if (Opcode == ARM::t2LDRpci)
+ NewLdOpc = ARM::t2LDRi12;
+ else if (Opcode == ARM::t2LDRHpci)
+ NewLdOpc = ARM::t2LDRHi12;
+ else if (Opcode == ARM::t2LDRBpci)
+ NewLdOpc = ARM::t2LDRBi12;
+ else if (Opcode == ARM::t2LDRSHpci)
+ NewLdOpc = ARM::t2LDRSHi12;
+ else if (Opcode == ARM::t2LDRSBpci)
+ NewLdOpc = ARM::t2LDRSBi12;
+ else
+ llvm_unreachable("Not a known opcode?");
+
+ unsigned DstReg = MI.getOperand(0).getReg();
+ MachineInstrBuilder MIB =
+ AddDefaultPred(BuildMI(MBB, MBBI, MI.getDebugLoc(),
+ TII->get(NewLdOpc), DstReg)
+ .addReg(ARM::PC)
+ .addOperand(MI.getOperand(1)));
+ (*MIB).setMemRefs(MI.memoperands_begin(), MI.memoperands_end());
+ TransferImpOps(MI, MIB, MIB);
+ MI.eraseFromParent();
+ break;
+ }
+ case ARM::tLDRpci_pic:
case ARM::t2LDRpci_pic: {
unsigned NewLdOpc = (Opcode == ARM::tLDRpci_pic)
- ? ARM::tLDRpci : ARM::t2LDRpci;
+ ? ARM::tLDRpci : ARM::t2LDRi12;
unsigned DstReg = MI.getOperand(0).getReg();
bool DstIsDead = MI.getOperand(0).isDead();
MachineInstrBuilder MIB1 =
- AddDefaultPred(BuildMI(MBB, MBBI, MI.getDebugLoc(),
- TII->get(NewLdOpc), DstReg)
- .addOperand(MI.getOperand(1)));
+ BuildMI(MBB, MBBI, MI.getDebugLoc(),
+ TII->get(NewLdOpc), DstReg);
+ if (Opcode == ARM::t2LDRpci_pic) MIB1.addReg(ARM::PC);
+ MIB1.addOperand(MI.getOperand(1));
+ AddDefaultPred(MIB1);
(*MIB1).setMemRefs(MI.memoperands_begin(), MI.memoperands_end());
MachineInstrBuilder MIB2 = BuildMI(MBB, MBBI, MI.getDebugLoc(),
TII->get(ARM::tPICADD))
- .addReg(DstReg, getDefRegState(true) | getDeadRegState(DstIsDead))
+ .addReg(DstReg, RegState::Define | getDeadRegState(DstIsDead))
.addReg(DstReg)
.addOperand(MI.getOperand(2));
TransferImpOps(MI, MIB1, MIB2);
}
case ARM::MOVi32imm:
- case ARM::t2MOVi32imm: {
+ case ARM::MOVCCi32imm:
+ case ARM::t2MOVi32imm:
+ case ARM::t2MOVCCi32imm: {
unsigned PredReg = 0;
ARMCC::CondCodes Pred = llvm::getInstrPredicate(&MI, PredReg);
unsigned DstReg = MI.getOperand(0).getReg();
bool DstIsDead = MI.getOperand(0).isDead();
- const MachineOperand &MO = MI.getOperand(1);
+ bool isCC = Opcode == ARM::MOVCCi32imm || Opcode == ARM::t2MOVCCi32imm;
+ const MachineOperand &MO = MI.getOperand(isCC ? 2 : 1);
MachineInstrBuilder LO16, HI16;
+ if (!STI->hasV6T2Ops() &&
+ (Opcode == ARM::MOVi32imm || Opcode == ARM::MOVCCi32imm)) {
+ // Expand into a movi + orr.
+ LO16 = BuildMI(MBB, MBBI, MI.getDebugLoc(), TII->get(ARM::MOVi), DstReg);
+ HI16 = BuildMI(MBB, MBBI, MI.getDebugLoc(), TII->get(ARM::ORRri))
+ .addReg(DstReg, RegState::Define | getDeadRegState(DstIsDead))
+ .addReg(DstReg);
+
+ assert (MO.isImm() && "MOVi32imm w/ non-immediate source operand!");
+ unsigned ImmVal = (unsigned)MO.getImm();
+ unsigned SOImmValV1 = ARM_AM::getSOImmTwoPartFirst(ImmVal);
+ unsigned SOImmValV2 = ARM_AM::getSOImmTwoPartSecond(ImmVal);
+ LO16 = LO16.addImm(SOImmValV1);
+ HI16 = HI16.addImm(SOImmValV2);
+ (*LO16).setMemRefs(MI.memoperands_begin(), MI.memoperands_end());
+ (*HI16).setMemRefs(MI.memoperands_begin(), MI.memoperands_end());
+ LO16.addImm(Pred).addReg(PredReg).addReg(0);
+ HI16.addImm(Pred).addReg(PredReg).addReg(0);
+ TransferImpOps(MI, LO16, HI16);
+ MI.eraseFromParent();
+ break;
+ }
+
+ bool isThumb =
+ (Opcode == ARM::t2MOVi32imm || Opcode == ARM::t2MOVCCi32imm);
+
LO16 = BuildMI(MBB, MBBI, MI.getDebugLoc(),
- TII->get(Opcode == ARM::MOVi32imm ?
- ARM::MOVi16 : ARM::t2MOVi16),
+ TII->get(isThumb ? ARM::t2MOVi16 : ARM::MOVi16),
DstReg);
HI16 = BuildMI(MBB, MBBI, MI.getDebugLoc(),
- TII->get(Opcode == ARM::MOVi32imm ?
- ARM::MOVTi16 : ARM::t2MOVTi16))
- .addReg(DstReg, getDefRegState(true) | getDeadRegState(DstIsDead))
+ TII->get(isThumb ? ARM::t2MOVTi16 : ARM::MOVTi16))
+ .addReg(DstReg, RegState::Define | getDeadRegState(DstIsDead))
.addReg(DstReg);
if (MO.isImm()) {
AddDefaultPred(BuildMI(MBB, MBBI, MI.getDebugLoc(),
TII->get(ARM::VMOVQ))
.addReg(EvenDst,
- getDefRegState(true) | getDeadRegState(DstIsDead))
+ RegState::Define | getDeadRegState(DstIsDead))
.addReg(EvenSrc, getKillRegState(SrcIsKill)));
MachineInstrBuilder Odd =
AddDefaultPred(BuildMI(MBB, MBBI, MI.getDebugLoc(),
TII->get(ARM::VMOVQ))
.addReg(OddDst,
- getDefRegState(true) | getDeadRegState(DstIsDead))
+ RegState::Define | getDeadRegState(DstIsDead))
.addReg(OddSrc, getKillRegState(SrcIsKill)));
TransferImpOps(MI, Even, Odd);
MI.eraseFromParent();
+ break;
+ }
+
+ case ARM::VLDMQIA:
+ case ARM::VLDMQDB: {
+ unsigned NewOpc = (Opcode == ARM::VLDMQIA) ? ARM::VLDMDIA : ARM::VLDMDDB;
+ MachineInstrBuilder MIB =
+ BuildMI(MBB, MBBI, MI.getDebugLoc(), TII->get(NewOpc));
+ unsigned OpIdx = 0;
+
+ // Grab the Q register destination.
+ bool DstIsDead = MI.getOperand(OpIdx).isDead();
+ unsigned DstReg = MI.getOperand(OpIdx++).getReg();
+
+ // Copy the source register.
+ MIB.addOperand(MI.getOperand(OpIdx++));
+
+ // Copy the predicate operands.
+ MIB.addOperand(MI.getOperand(OpIdx++));
+ MIB.addOperand(MI.getOperand(OpIdx++));
+
+ // Add the destination operands (D subregs).
+ unsigned D0 = TRI->getSubReg(DstReg, ARM::dsub_0);
+ unsigned D1 = TRI->getSubReg(DstReg, ARM::dsub_1);
+ MIB.addReg(D0, RegState::Define | getDeadRegState(DstIsDead))
+ .addReg(D1, RegState::Define | getDeadRegState(DstIsDead));
+
+ // Add an implicit def for the super-register.
+ MIB.addReg(DstReg, RegState::ImplicitDefine | getDeadRegState(DstIsDead));
+ TransferImpOps(MI, MIB, MIB);
+ MI.eraseFromParent();
+ break;
+ }
+
+ case ARM::VSTMQIA:
+ case ARM::VSTMQDB: {
+ unsigned NewOpc = (Opcode == ARM::VSTMQIA) ? ARM::VSTMDIA : ARM::VSTMDDB;
+ MachineInstrBuilder MIB =
+ BuildMI(MBB, MBBI, MI.getDebugLoc(), TII->get(NewOpc));
+ unsigned OpIdx = 0;
+
+ // Grab the Q register source.
+ bool SrcIsKill = MI.getOperand(OpIdx).isKill();
+ unsigned SrcReg = MI.getOperand(OpIdx++).getReg();
+
+ // Copy the destination register.
+ MIB.addOperand(MI.getOperand(OpIdx++));
+
+ // Copy the predicate operands.
+ MIB.addOperand(MI.getOperand(OpIdx++));
+ MIB.addOperand(MI.getOperand(OpIdx++));
+
+ // Add the source operands (D subregs).
+ unsigned D0 = TRI->getSubReg(SrcReg, ARM::dsub_0);
+ unsigned D1 = TRI->getSubReg(SrcReg, ARM::dsub_1);
+ MIB.addReg(D0).addReg(D1);
+
+ if (SrcIsKill)
+ // Add an implicit kill for the Q register.
+ (*MIB).addRegisterKilled(SrcReg, TRI, true);
+
+ TransferImpOps(MI, MIB, MIB);
+ MI.eraseFromParent();
+ break;
+ }
+ case ARM::VDUPfqf:
+ case ARM::VDUPfdf:{
+ unsigned NewOpc = Opcode == ARM::VDUPfqf ? ARM::VDUPLNfq : ARM::VDUPLNfd;
+ MachineInstrBuilder MIB =
+ BuildMI(MBB, MBBI, MI.getDebugLoc(), TII->get(NewOpc));
+ unsigned OpIdx = 0;
+ unsigned SrcReg = MI.getOperand(1).getReg();
+ unsigned Lane = getARMRegisterNumbering(SrcReg) & 1;
+ unsigned DReg = TRI->getMatchingSuperReg(SrcReg,
+ Lane & 1 ? ARM::ssub_1 : ARM::ssub_0, &ARM::DPR_VFP2RegClass);
+ // The lane is [0,1] for the containing DReg superregister.
+ // Copy the dst/src register operands.
+ MIB.addOperand(MI.getOperand(OpIdx++));
+ MIB.addReg(DReg);
+ ++OpIdx;
+ // Add the lane select operand.
+ MIB.addImm(Lane);
+ // Add the predicate operands.
+ MIB.addOperand(MI.getOperand(OpIdx++));
+ MIB.addOperand(MI.getOperand(OpIdx++));
+
+ TransferImpOps(MI, MIB, MIB);
+ MI.eraseFromParent();
+ break;
}
case ARM::VLD1q8Pseudo:
case ARM::VLD4q8oddPseudo_UPD:
case ARM::VLD4q16oddPseudo_UPD:
case ARM::VLD4q32oddPseudo_UPD:
+ case ARM::VLD1DUPq8Pseudo:
+ case ARM::VLD1DUPq16Pseudo:
+ case ARM::VLD1DUPq32Pseudo:
+ case ARM::VLD1DUPq8Pseudo_UPD:
+ case ARM::VLD1DUPq16Pseudo_UPD:
+ case ARM::VLD1DUPq32Pseudo_UPD:
+ case ARM::VLD2DUPd8Pseudo:
+ case ARM::VLD2DUPd16Pseudo:
+ case ARM::VLD2DUPd32Pseudo:
+ case ARM::VLD2DUPd8Pseudo_UPD:
+ case ARM::VLD2DUPd16Pseudo_UPD:
+ case ARM::VLD2DUPd32Pseudo_UPD:
+ case ARM::VLD3DUPd8Pseudo:
+ case ARM::VLD3DUPd16Pseudo:
+ case ARM::VLD3DUPd32Pseudo:
+ case ARM::VLD3DUPd8Pseudo_UPD:
+ case ARM::VLD3DUPd16Pseudo_UPD:
+ case ARM::VLD3DUPd32Pseudo_UPD:
+ case ARM::VLD4DUPd8Pseudo:
+ case ARM::VLD4DUPd16Pseudo:
+ case ARM::VLD4DUPd32Pseudo:
+ case ARM::VLD4DUPd8Pseudo_UPD:
+ case ARM::VLD4DUPd16Pseudo_UPD:
+ case ARM::VLD4DUPd32Pseudo_UPD:
ExpandVLD(MBBI);
break;
ExpandVST(MBBI);
break;
+ case ARM::VLD1LNq8Pseudo:
+ case ARM::VLD1LNq16Pseudo:
+ case ARM::VLD1LNq32Pseudo:
+ case ARM::VLD1LNq8Pseudo_UPD:
+ case ARM::VLD1LNq16Pseudo_UPD:
+ case ARM::VLD1LNq32Pseudo_UPD:
case ARM::VLD2LNd8Pseudo:
case ARM::VLD2LNd16Pseudo:
case ARM::VLD2LNd32Pseudo:
case ARM::VLD4LNd32Pseudo_UPD:
case ARM::VLD4LNq16Pseudo_UPD:
case ARM::VLD4LNq32Pseudo_UPD:
+ case ARM::VST1LNq8Pseudo:
+ case ARM::VST1LNq16Pseudo:
+ case ARM::VST1LNq32Pseudo:
+ case ARM::VST1LNq8Pseudo_UPD:
+ case ARM::VST1LNq16Pseudo_UPD:
+ case ARM::VST1LNq32Pseudo_UPD:
case ARM::VST2LNd8Pseudo:
case ARM::VST2LNd16Pseudo:
case ARM::VST2LNd32Pseudo:
case ARM::VST4LNq32Pseudo_UPD:
ExpandLaneOp(MBBI);
break;
+
+ case ARM::VTBL2Pseudo:
+ ExpandVTBL(MBBI, ARM::VTBL2, false, 2); break;
+ case ARM::VTBL3Pseudo:
+ ExpandVTBL(MBBI, ARM::VTBL3, false, 3); break;
+ case ARM::VTBL4Pseudo:
+ ExpandVTBL(MBBI, ARM::VTBL4, false, 4); break;
+ case ARM::VTBX2Pseudo:
+ ExpandVTBL(MBBI, ARM::VTBX2, true, 2); break;
+ case ARM::VTBX3Pseudo:
+ ExpandVTBL(MBBI, ARM::VTBX3, true, 3); break;
+ case ARM::VTBX4Pseudo:
+ ExpandVTBL(MBBI, ARM::VTBX4, true, 4); break;
}
if (ModifiedOp)
}
bool ARMExpandPseudo::runOnMachineFunction(MachineFunction &MF) {
- TII = MF.getTarget().getInstrInfo();
+ TII = static_cast<const ARMBaseInstrInfo*>(MF.getTarget().getInstrInfo());
TRI = MF.getTarget().getRegisterInfo();
+ STI = &MF.getTarget().getSubtarget<ARMSubtarget>();
bool Modified = false;
for (MachineFunction::iterator MFI = MF.begin(), E = MF.end(); MFI != E;
projects / oota-llvm.git / blobdiff