//===-- ARMBaseInstrInfo.cpp - ARM Instruction Information ----------------===// // // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. // See https://llvm.org/LICENSE.txt for license information. // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception // //===----------------------------------------------------------------------===// // // This file contains the Base ARM implementation of the TargetInstrInfo class. // //===----------------------------------------------------------------------===// #include "ARMBaseInstrInfo.h" #include "ARMBaseRegisterInfo.h" #include "ARMConstantPoolValue.h" #include "ARMFeatures.h" #include "ARMHazardRecognizer.h" #include "ARMMachineFunctionInfo.h" #include "ARMSubtarget.h" #include "MCTargetDesc/ARMAddressingModes.h" #include "MCTargetDesc/ARMBaseInfo.h" #include "MVETailPredUtils.h" #include "llvm/ADT/DenseMap.h" #include "llvm/ADT/STLExtras.h" #include "llvm/ADT/SmallSet.h" #include "llvm/ADT/SmallVector.h" #include "llvm/CodeGen/DFAPacketizer.h" #include "llvm/CodeGen/LiveVariables.h" #include "llvm/CodeGen/MachineBasicBlock.h" #include "llvm/CodeGen/MachineConstantPool.h" #include "llvm/CodeGen/MachineFrameInfo.h" #include "llvm/CodeGen/MachineFunction.h" #include "llvm/CodeGen/MachineInstr.h" #include "llvm/CodeGen/MachineInstrBuilder.h" #include "llvm/CodeGen/MachineMemOperand.h" #include "llvm/CodeGen/MachineModuleInfo.h" #include "llvm/CodeGen/MachineOperand.h" #include "llvm/CodeGen/MachinePipeliner.h" #include "llvm/CodeGen/MachineRegisterInfo.h" #include "llvm/CodeGen/MachineScheduler.h" #include "llvm/CodeGen/MultiHazardRecognizer.h" #include "llvm/CodeGen/ScoreboardHazardRecognizer.h" #include "llvm/CodeGen/SelectionDAGNodes.h" #include "llvm/CodeGen/TargetInstrInfo.h" #include "llvm/CodeGen/TargetRegisterInfo.h" #include "llvm/CodeGen/TargetSchedule.h" #include "llvm/IR/Attributes.h" #include "llvm/IR/Constants.h" #include "llvm/IR/DebugLoc.h" #include "llvm/IR/Function.h" #include "llvm/IR/GlobalValue.h" #include "llvm/MC/MCAsmInfo.h" #include "llvm/MC/MCInstrDesc.h" #include "llvm/MC/MCInstrItineraries.h" #include "llvm/Support/BranchProbability.h" #include "llvm/Support/Casting.h" #include "llvm/Support/CommandLine.h" #include "llvm/Support/Compiler.h" #include "llvm/Support/Debug.h" #include "llvm/Support/ErrorHandling.h" #include "llvm/Support/raw_ostream.h" #include "llvm/Target/TargetMachine.h" #include "llvm/TargetParser/Triple.h" #include #include #include #include #include #include #include using namespace llvm; #define DEBUG_TYPE "arm-instrinfo" #define GET_INSTRINFO_CTOR_DTOR #include "ARMGenInstrInfo.inc" static cl::opt EnableARM3Addr("enable-arm-3-addr-conv", cl::Hidden, cl::desc("Enable ARM 2-addr to 3-addr conv")); /// ARM_MLxEntry - Record information about MLA / MLS instructions. struct ARM_MLxEntry { uint16_t MLxOpc; // MLA / MLS opcode uint16_t MulOpc; // Expanded multiplication opcode uint16_t AddSubOpc; // Expanded add / sub opcode bool NegAcc; // True if the acc is negated before the add / sub. bool HasLane; // True if instruction has an extra "lane" operand. }; static const ARM_MLxEntry ARM_MLxTable[] = { // MLxOpc, MulOpc, AddSubOpc, NegAcc, HasLane // fp scalar ops { ARM::VMLAS, ARM::VMULS, ARM::VADDS, false, false }, { ARM::VMLSS, ARM::VMULS, ARM::VSUBS, false, false }, { ARM::VMLAD, ARM::VMULD, ARM::VADDD, false, false }, { ARM::VMLSD, ARM::VMULD, ARM::VSUBD, false, false }, { ARM::VNMLAS, ARM::VNMULS, ARM::VSUBS, true, false }, { ARM::VNMLSS, ARM::VMULS, ARM::VSUBS, true, false }, { ARM::VNMLAD, ARM::VNMULD, ARM::VSUBD, true, false }, { ARM::VNMLSD, ARM::VMULD, ARM::VSUBD, true, false }, // fp SIMD ops { ARM::VMLAfd, ARM::VMULfd, ARM::VADDfd, false, false }, { ARM::VMLSfd, ARM::VMULfd, ARM::VSUBfd, false, false }, { ARM::VMLAfq, ARM::VMULfq, ARM::VADDfq, false, false }, { ARM::VMLSfq, ARM::VMULfq, ARM::VSUBfq, false, false }, { ARM::VMLAslfd, ARM::VMULslfd, ARM::VADDfd, false, true }, { ARM::VMLSslfd, ARM::VMULslfd, ARM::VSUBfd, false, true }, { ARM::VMLAslfq, ARM::VMULslfq, ARM::VADDfq, false, true }, { ARM::VMLSslfq, ARM::VMULslfq, ARM::VSUBfq, false, true }, }; ARMBaseInstrInfo::ARMBaseInstrInfo(const ARMSubtarget& STI) : ARMGenInstrInfo(ARM::ADJCALLSTACKDOWN, ARM::ADJCALLSTACKUP), Subtarget(STI) { for (unsigned i = 0, e = std::size(ARM_MLxTable); i != e; ++i) { if (!MLxEntryMap.insert(std::make_pair(ARM_MLxTable[i].MLxOpc, i)).second) llvm_unreachable("Duplicated entries?"); MLxHazardOpcodes.insert(ARM_MLxTable[i].AddSubOpc); MLxHazardOpcodes.insert(ARM_MLxTable[i].MulOpc); } } // Use a ScoreboardHazardRecognizer for prepass ARM scheduling. TargetInstrImpl // currently defaults to no prepass hazard recognizer. ScheduleHazardRecognizer * ARMBaseInstrInfo::CreateTargetHazardRecognizer(const TargetSubtargetInfo *STI, const ScheduleDAG *DAG) const { if (usePreRAHazardRecognizer()) { const InstrItineraryData *II = static_cast(STI)->getInstrItineraryData(); return new ScoreboardHazardRecognizer(II, DAG, "pre-RA-sched"); } return TargetInstrInfo::CreateTargetHazardRecognizer(STI, DAG); } // Called during: // - pre-RA scheduling // - post-RA scheduling when FeatureUseMISched is set ScheduleHazardRecognizer *ARMBaseInstrInfo::CreateTargetMIHazardRecognizer( const InstrItineraryData *II, const ScheduleDAGMI *DAG) const { MultiHazardRecognizer *MHR = new MultiHazardRecognizer(); // We would like to restrict this hazard recognizer to only // post-RA scheduling; we can tell that we're post-RA because we don't // track VRegLiveness. // Cortex-M7: TRM indicates that there is a single ITCM bank and two DTCM // banks banked on bit 2. Assume that TCMs are in use. if (Subtarget.isCortexM7() && !DAG->hasVRegLiveness()) MHR->AddHazardRecognizer( std::make_unique(DAG, 0x4, true)); // Not inserting ARMHazardRecognizerFPMLx because that would change // legacy behavior auto BHR = TargetInstrInfo::CreateTargetMIHazardRecognizer(II, DAG); MHR->AddHazardRecognizer(std::unique_ptr(BHR)); return MHR; } // Called during post-RA scheduling when FeatureUseMISched is not set ScheduleHazardRecognizer *ARMBaseInstrInfo:: CreateTargetPostRAHazardRecognizer(const InstrItineraryData *II, const ScheduleDAG *DAG) const { MultiHazardRecognizer *MHR = new MultiHazardRecognizer(); if (Subtarget.isThumb2() || Subtarget.hasVFP2Base()) MHR->AddHazardRecognizer(std::make_unique()); auto BHR = TargetInstrInfo::CreateTargetPostRAHazardRecognizer(II, DAG); if (BHR) MHR->AddHazardRecognizer(std::unique_ptr(BHR)); return MHR; } MachineInstr * ARMBaseInstrInfo::convertToThreeAddress(MachineInstr &MI, LiveVariables *LV, LiveIntervals *LIS) const { // FIXME: Thumb2 support. if (!EnableARM3Addr) return nullptr; MachineFunction &MF = *MI.getParent()->getParent(); uint64_t TSFlags = MI.getDesc().TSFlags; bool isPre = false; switch ((TSFlags & ARMII::IndexModeMask) >> ARMII::IndexModeShift) { default: return nullptr; case ARMII::IndexModePre: isPre = true; break; case ARMII::IndexModePost: break; } // Try splitting an indexed load/store to an un-indexed one plus an add/sub // operation. unsigned MemOpc = getUnindexedOpcode(MI.getOpcode()); if (MemOpc == 0) return nullptr; MachineInstr *UpdateMI = nullptr; MachineInstr *MemMI = nullptr; unsigned AddrMode = (TSFlags & ARMII::AddrModeMask); const MCInstrDesc &MCID = MI.getDesc(); unsigned NumOps = MCID.getNumOperands(); bool isLoad = !MI.mayStore(); const MachineOperand &WB = isLoad ? MI.getOperand(1) : MI.getOperand(0); const MachineOperand &Base = MI.getOperand(2); const MachineOperand &Offset = MI.getOperand(NumOps - 3); Register WBReg = WB.getReg(); Register BaseReg = Base.getReg(); Register OffReg = Offset.getReg(); unsigned OffImm = MI.getOperand(NumOps - 2).getImm(); ARMCC::CondCodes Pred = (ARMCC::CondCodes)MI.getOperand(NumOps - 1).getImm(); switch (AddrMode) { default: llvm_unreachable("Unknown indexed op!"); case ARMII::AddrMode2: { bool isSub = ARM_AM::getAM2Op(OffImm) == ARM_AM::sub; unsigned Amt = ARM_AM::getAM2Offset(OffImm); if (OffReg == 0) { if (ARM_AM::getSOImmVal(Amt) == -1) // Can't encode it in a so_imm operand. This transformation will // add more than 1 instruction. Abandon! return nullptr; UpdateMI = BuildMI(MF, MI.getDebugLoc(), get(isSub ? ARM::SUBri : ARM::ADDri), WBReg) .addReg(BaseReg) .addImm(Amt) .add(predOps(Pred)) .add(condCodeOp()); } else if (Amt != 0) { ARM_AM::ShiftOpc ShOpc = ARM_AM::getAM2ShiftOpc(OffImm); unsigned SOOpc = ARM_AM::getSORegOpc(ShOpc, Amt); UpdateMI = BuildMI(MF, MI.getDebugLoc(), get(isSub ? ARM::SUBrsi : ARM::ADDrsi), WBReg) .addReg(BaseReg) .addReg(OffReg) .addReg(0) .addImm(SOOpc) .add(predOps(Pred)) .add(condCodeOp()); } else UpdateMI = BuildMI(MF, MI.getDebugLoc(), get(isSub ? ARM::SUBrr : ARM::ADDrr), WBReg) .addReg(BaseReg) .addReg(OffReg) .add(predOps(Pred)) .add(condCodeOp()); break; } case ARMII::AddrMode3 : { bool isSub = ARM_AM::getAM3Op(OffImm) == ARM_AM::sub; unsigned Amt = ARM_AM::getAM3Offset(OffImm); if (OffReg == 0) // Immediate is 8-bits. It's guaranteed to fit in a so_imm operand. UpdateMI = BuildMI(MF, MI.getDebugLoc(), get(isSub ? ARM::SUBri : ARM::ADDri), WBReg) .addReg(BaseReg) .addImm(Amt) .add(predOps(Pred)) .add(condCodeOp()); else UpdateMI = BuildMI(MF, MI.getDebugLoc(), get(isSub ? ARM::SUBrr : ARM::ADDrr), WBReg) .addReg(BaseReg) .addReg(OffReg) .add(predOps(Pred)) .add(condCodeOp()); break; } } std::vector NewMIs; if (isPre) { if (isLoad) MemMI = BuildMI(MF, MI.getDebugLoc(), get(MemOpc), MI.getOperand(0).getReg()) .addReg(WBReg) .addImm(0) .addImm(Pred); else MemMI = BuildMI(MF, MI.getDebugLoc(), get(MemOpc)) .addReg(MI.getOperand(1).getReg()) .addReg(WBReg) .addReg(0) .addImm(0) .addImm(Pred); NewMIs.push_back(MemMI); NewMIs.push_back(UpdateMI); } else { if (isLoad) MemMI = BuildMI(MF, MI.getDebugLoc(), get(MemOpc), MI.getOperand(0).getReg()) .addReg(BaseReg) .addImm(0) .addImm(Pred); else MemMI = BuildMI(MF, MI.getDebugLoc(), get(MemOpc)) .addReg(MI.getOperand(1).getReg()) .addReg(BaseReg) .addReg(0) .addImm(0) .addImm(Pred); if (WB.isDead()) UpdateMI->getOperand(0).setIsDead(); NewMIs.push_back(UpdateMI); NewMIs.push_back(MemMI); } // Transfer LiveVariables states, kill / dead info. if (LV) { for (const MachineOperand &MO : MI.operands()) { if (MO.isReg() && MO.getReg().isVirtual()) { Register Reg = MO.getReg(); LiveVariables::VarInfo &VI = LV->getVarInfo(Reg); if (MO.isDef()) { MachineInstr *NewMI = (Reg == WBReg) ? UpdateMI : MemMI; if (MO.isDead()) LV->addVirtualRegisterDead(Reg, *NewMI); } if (MO.isUse() && MO.isKill()) { for (unsigned j = 0; j < 2; ++j) { // Look at the two new MI's in reverse order. MachineInstr *NewMI = NewMIs[j]; if (!NewMI->readsRegister(Reg)) continue; LV->addVirtualRegisterKilled(Reg, *NewMI); if (VI.removeKill(MI)) VI.Kills.push_back(NewMI); break; } } } } } MachineBasicBlock &MBB = *MI.getParent(); MBB.insert(MI, NewMIs[1]); MBB.insert(MI, NewMIs[0]); return NewMIs[0]; } // Branch analysis. // Cond vector output format: // 0 elements indicates an unconditional branch // 2 elements indicates a conditional branch; the elements are // the condition to check and the CPSR. // 3 elements indicates a hardware loop end; the elements // are the opcode, the operand value to test, and a dummy // operand used to pad out to 3 operands. bool ARMBaseInstrInfo::analyzeBranch(MachineBasicBlock &MBB, MachineBasicBlock *&TBB, MachineBasicBlock *&FBB, SmallVectorImpl &Cond, bool AllowModify) const { TBB = nullptr; FBB = nullptr; MachineBasicBlock::instr_iterator I = MBB.instr_end(); if (I == MBB.instr_begin()) return false; // Empty blocks are easy. --I; // Walk backwards from the end of the basic block until the branch is // analyzed or we give up. while (isPredicated(*I) || I->isTerminator() || I->isDebugValue()) { // Flag to be raised on unanalyzeable instructions. This is useful in cases // where we want to clean up on the end of the basic block before we bail // out. bool CantAnalyze = false; // Skip over DEBUG values, predicated nonterminators and speculation // barrier terminators. while (I->isDebugInstr() || !I->isTerminator() || isSpeculationBarrierEndBBOpcode(I->getOpcode()) || I->getOpcode() == ARM::t2DoLoopStartTP){ if (I == MBB.instr_begin()) return false; --I; } if (isIndirectBranchOpcode(I->getOpcode()) || isJumpTableBranchOpcode(I->getOpcode())) { // Indirect branches and jump tables can't be analyzed, but we still want // to clean up any instructions at the tail of the basic block. CantAnalyze = true; } else if (isUncondBranchOpcode(I->getOpcode())) { TBB = I->getOperand(0).getMBB(); } else if (isCondBranchOpcode(I->getOpcode())) { // Bail out if we encounter multiple conditional branches. if (!Cond.empty()) return true; assert(!FBB && "FBB should have been null."); FBB = TBB; TBB = I->getOperand(0).getMBB(); Cond.push_back(I->getOperand(1)); Cond.push_back(I->getOperand(2)); } else if (I->isReturn()) { // Returns can't be analyzed, but we should run cleanup. CantAnalyze = true; } else if (I->getOpcode() == ARM::t2LoopEnd && MBB.getParent() ->getSubtarget() .enableMachinePipeliner()) { if (!Cond.empty()) return true; FBB = TBB; TBB = I->getOperand(1).getMBB(); Cond.push_back(MachineOperand::CreateImm(I->getOpcode())); Cond.push_back(I->getOperand(0)); Cond.push_back(MachineOperand::CreateImm(0)); } else { // We encountered other unrecognized terminator. Bail out immediately. return true; } // Cleanup code - to be run for unpredicated unconditional branches and // returns. if (!isPredicated(*I) && (isUncondBranchOpcode(I->getOpcode()) || isIndirectBranchOpcode(I->getOpcode()) || isJumpTableBranchOpcode(I->getOpcode()) || I->isReturn())) { // Forget any previous condition branch information - it no longer applies. Cond.clear(); FBB = nullptr; // If we can modify the function, delete everything below this // unconditional branch. if (AllowModify) { MachineBasicBlock::iterator DI = std::next(I); while (DI != MBB.instr_end()) { MachineInstr &InstToDelete = *DI; ++DI; // Speculation barriers must not be deleted. if (isSpeculationBarrierEndBBOpcode(InstToDelete.getOpcode())) continue; InstToDelete.eraseFromParent(); } } } if (CantAnalyze) { // We may not be able to analyze the block, but we could still have // an unconditional branch as the last instruction in the block, which // just branches to layout successor. If this is the case, then just // remove it if we're allowed to make modifications. if (AllowModify && !isPredicated(MBB.back()) && isUncondBranchOpcode(MBB.back().getOpcode()) && TBB && MBB.isLayoutSuccessor(TBB)) removeBranch(MBB); return true; } if (I == MBB.instr_begin()) return false; --I; } // We made it past the terminators without bailing out - we must have // analyzed this branch successfully. return false; } unsigned ARMBaseInstrInfo::removeBranch(MachineBasicBlock &MBB, int *BytesRemoved) const { assert(!BytesRemoved && "code size not handled"); MachineBasicBlock::iterator I = MBB.getLastNonDebugInstr(); if (I == MBB.end()) return 0; if (!isUncondBranchOpcode(I->getOpcode()) && !isCondBranchOpcode(I->getOpcode()) && I->getOpcode() != ARM::t2LoopEnd) return 0; // Remove the branch. I->eraseFromParent(); I = MBB.end(); if (I == MBB.begin()) return 1; --I; if (!isCondBranchOpcode(I->getOpcode()) && I->getOpcode() != ARM::t2LoopEnd) return 1; // Remove the branch. I->eraseFromParent(); return 2; } unsigned ARMBaseInstrInfo::insertBranch(MachineBasicBlock &MBB, MachineBasicBlock *TBB, MachineBasicBlock *FBB, ArrayRef Cond, const DebugLoc &DL, int *BytesAdded) const { assert(!BytesAdded && "code size not handled"); ARMFunctionInfo *AFI = MBB.getParent()->getInfo(); int BOpc = !AFI->isThumbFunction() ? ARM::B : (AFI->isThumb2Function() ? ARM::t2B : ARM::tB); int BccOpc = !AFI->isThumbFunction() ? ARM::Bcc : (AFI->isThumb2Function() ? ARM::t2Bcc : ARM::tBcc); bool isThumb = AFI->isThumbFunction() || AFI->isThumb2Function(); // Shouldn't be a fall through. assert(TBB && "insertBranch must not be told to insert a fallthrough"); assert((Cond.size() == 2 || Cond.size() == 0 || Cond.size() == 3) && "ARM branch conditions have two or three components!"); // For conditional branches, we use addOperand to preserve CPSR flags. if (!FBB) { if (Cond.empty()) { // Unconditional branch? if (isThumb) BuildMI(&MBB, DL, get(BOpc)).addMBB(TBB).add(predOps(ARMCC::AL)); else BuildMI(&MBB, DL, get(BOpc)).addMBB(TBB); } else if (Cond.size() == 2) { BuildMI(&MBB, DL, get(BccOpc)) .addMBB(TBB) .addImm(Cond[0].getImm()) .add(Cond[1]); } else BuildMI(&MBB, DL, get(Cond[0].getImm())).add(Cond[1]).addMBB(TBB); return 1; } // Two-way conditional branch. if (Cond.size() == 2) BuildMI(&MBB, DL, get(BccOpc)) .addMBB(TBB) .addImm(Cond[0].getImm()) .add(Cond[1]); else if (Cond.size() == 3) BuildMI(&MBB, DL, get(Cond[0].getImm())).add(Cond[1]).addMBB(TBB); if (isThumb) BuildMI(&MBB, DL, get(BOpc)).addMBB(FBB).add(predOps(ARMCC::AL)); else BuildMI(&MBB, DL, get(BOpc)).addMBB(FBB); return 2; } bool ARMBaseInstrInfo:: reverseBranchCondition(SmallVectorImpl &Cond) const { if (Cond.size() == 2) { ARMCC::CondCodes CC = (ARMCC::CondCodes)(int)Cond[0].getImm(); Cond[0].setImm(ARMCC::getOppositeCondition(CC)); return false; } return true; } bool ARMBaseInstrInfo::isPredicated(const MachineInstr &MI) const { if (MI.isBundle()) { MachineBasicBlock::const_instr_iterator I = MI.getIterator(); MachineBasicBlock::const_instr_iterator E = MI.getParent()->instr_end(); while (++I != E && I->isInsideBundle()) { int PIdx = I->findFirstPredOperandIdx(); if (PIdx != -1 && I->getOperand(PIdx).getImm() != ARMCC::AL) return true; } return false; } int PIdx = MI.findFirstPredOperandIdx(); return PIdx != -1 && MI.getOperand(PIdx).getImm() != ARMCC::AL; } std::string ARMBaseInstrInfo::createMIROperandComment( const MachineInstr &MI, const MachineOperand &Op, unsigned OpIdx, const TargetRegisterInfo *TRI) const { // First, let's see if there is a generic comment for this operand std::string GenericComment = TargetInstrInfo::createMIROperandComment(MI, Op, OpIdx, TRI); if (!GenericComment.empty()) return GenericComment; // If not, check if we have an immediate operand. if (!Op.isImm()) return std::string(); // And print its corresponding condition code if the immediate is a // predicate. int FirstPredOp = MI.findFirstPredOperandIdx(); if (FirstPredOp != (int) OpIdx) return std::string(); std::string CC = "CC::"; CC += ARMCondCodeToString((ARMCC::CondCodes)Op.getImm()); return CC; } bool ARMBaseInstrInfo::PredicateInstruction( MachineInstr &MI, ArrayRef Pred) const { unsigned Opc = MI.getOpcode(); if (isUncondBranchOpcode(Opc)) { MI.setDesc(get(getMatchingCondBranchOpcode(Opc))); MachineInstrBuilder(*MI.getParent()->getParent(), MI) .addImm(Pred[0].getImm()) .addReg(Pred[1].getReg()); return true; } int PIdx = MI.findFirstPredOperandIdx(); if (PIdx != -1) { MachineOperand &PMO = MI.getOperand(PIdx); PMO.setImm(Pred[0].getImm()); MI.getOperand(PIdx+1).setReg(Pred[1].getReg()); // Thumb 1 arithmetic instructions do not set CPSR when executed inside an // IT block. This affects how they are printed. const MCInstrDesc &MCID = MI.getDesc(); if (MCID.TSFlags & ARMII::ThumbArithFlagSetting) { assert(MCID.operands()[1].isOptionalDef() && "CPSR def isn't expected operand"); assert((MI.getOperand(1).isDead() || MI.getOperand(1).getReg() != ARM::CPSR) && "if conversion tried to stop defining used CPSR"); MI.getOperand(1).setReg(ARM::NoRegister); } return true; } return false; } bool ARMBaseInstrInfo::SubsumesPredicate(ArrayRef Pred1, ArrayRef Pred2) const { if (Pred1.size() > 2 || Pred2.size() > 2) return false; ARMCC::CondCodes CC1 = (ARMCC::CondCodes)Pred1[0].getImm(); ARMCC::CondCodes CC2 = (ARMCC::CondCodes)Pred2[0].getImm(); if (CC1 == CC2) return true; switch (CC1) { default: return false; case ARMCC::AL: return true; case ARMCC::HS: return CC2 == ARMCC::HI; case ARMCC::LS: return CC2 == ARMCC::LO || CC2 == ARMCC::EQ; case ARMCC::GE: return CC2 == ARMCC::GT; case ARMCC::LE: return CC2 == ARMCC::LT; } } bool ARMBaseInstrInfo::ClobbersPredicate(MachineInstr &MI, std::vector &Pred, bool SkipDead) const { bool Found = false; for (const MachineOperand &MO : MI.operands()) { bool ClobbersCPSR = MO.isRegMask() && MO.clobbersPhysReg(ARM::CPSR); bool IsCPSR = MO.isReg() && MO.isDef() && MO.getReg() == ARM::CPSR; if (ClobbersCPSR || IsCPSR) { // Filter out T1 instructions that have a dead CPSR, // allowing IT blocks to be generated containing T1 instructions const MCInstrDesc &MCID = MI.getDesc(); if (MCID.TSFlags & ARMII::ThumbArithFlagSetting && MO.isDead() && SkipDead) continue; Pred.push_back(MO); Found = true; } } return Found; } bool ARMBaseInstrInfo::isCPSRDefined(const MachineInstr &MI) { for (const auto &MO : MI.operands()) if (MO.isReg() && MO.getReg() == ARM::CPSR && MO.isDef() && !MO.isDead()) return true; return false; } static bool isEligibleForITBlock(const MachineInstr *MI) { switch (MI->getOpcode()) { default: return true; case ARM::tADC: // ADC (register) T1 case ARM::tADDi3: // ADD (immediate) T1 case ARM::tADDi8: // ADD (immediate) T2 case ARM::tADDrr: // ADD (register) T1 case ARM::tAND: // AND (register) T1 case ARM::tASRri: // ASR (immediate) T1 case ARM::tASRrr: // ASR (register) T1 case ARM::tBIC: // BIC (register) T1 case ARM::tEOR: // EOR (register) T1 case ARM::tLSLri: // LSL (immediate) T1 case ARM::tLSLrr: // LSL (register) T1 case ARM::tLSRri: // LSR (immediate) T1 case ARM::tLSRrr: // LSR (register) T1 case ARM::tMUL: // MUL T1 case ARM::tMVN: // MVN (register) T1 case ARM::tORR: // ORR (register) T1 case ARM::tROR: // ROR (register) T1 case ARM::tRSB: // RSB (immediate) T1 case ARM::tSBC: // SBC (register) T1 case ARM::tSUBi3: // SUB (immediate) T1 case ARM::tSUBi8: // SUB (immediate) T2 case ARM::tSUBrr: // SUB (register) T1 return !ARMBaseInstrInfo::isCPSRDefined(*MI); } } /// isPredicable - Return true if the specified instruction can be predicated. /// By default, this returns true for every instruction with a /// PredicateOperand. bool ARMBaseInstrInfo::isPredicable(const MachineInstr &MI) const { if (!MI.isPredicable()) return false; if (MI.isBundle()) return false; if (!isEligibleForITBlock(&MI)) return false; const MachineFunction *MF = MI.getParent()->getParent(); const ARMFunctionInfo *AFI = MF->getInfo(); // Neon instructions in Thumb2 IT blocks are deprecated, see ARMARM. // In their ARM encoding, they can't be encoded in a conditional form. if ((MI.getDesc().TSFlags & ARMII::DomainMask) == ARMII::DomainNEON) return false; // Make indirect control flow changes unpredicable when SLS mitigation is // enabled. const ARMSubtarget &ST = MF->getSubtarget(); if (ST.hardenSlsRetBr() && isIndirectControlFlowNotComingBack(MI)) return false; if (ST.hardenSlsBlr() && isIndirectCall(MI)) return false; if (AFI->isThumb2Function()) { if (getSubtarget().restrictIT()) return isV8EligibleForIT(&MI); } return true; } namespace llvm { template <> bool IsCPSRDead(const MachineInstr *MI) { for (const MachineOperand &MO : MI->operands()) { if (!MO.isReg() || MO.isUndef() || MO.isUse()) continue; if (MO.getReg() != ARM::CPSR) continue; if (!MO.isDead()) return false; } // all definitions of CPSR are dead return true; } } // end namespace llvm /// GetInstSize - Return the size of the specified MachineInstr. /// unsigned ARMBaseInstrInfo::getInstSizeInBytes(const MachineInstr &MI) const { const MachineBasicBlock &MBB = *MI.getParent(); const MachineFunction *MF = MBB.getParent(); const MCAsmInfo *MAI = MF->getTarget().getMCAsmInfo(); const MCInstrDesc &MCID = MI.getDesc(); switch (MI.getOpcode()) { default: // Return the size specified in .td file. If there's none, return 0, as we // can't define a default size (Thumb1 instructions are 2 bytes, Thumb2 // instructions are 2-4 bytes, and ARM instructions are 4 bytes), in // contrast to AArch64 instructions which have a default size of 4 bytes for // example. return MCID.getSize(); case TargetOpcode::BUNDLE: return getInstBundleLength(MI); case ARM::CONSTPOOL_ENTRY: case ARM::JUMPTABLE_INSTS: case ARM::JUMPTABLE_ADDRS: case ARM::JUMPTABLE_TBB: case ARM::JUMPTABLE_TBH: // If this machine instr is a constant pool entry, its size is recorded as // operand #2. return MI.getOperand(2).getImm(); case ARM::SPACE: return MI.getOperand(1).getImm(); case ARM::INLINEASM: case ARM::INLINEASM_BR: { // If this machine instr is an inline asm, measure it. unsigned Size = getInlineAsmLength(MI.getOperand(0).getSymbolName(), *MAI); if (!MF->getInfo()->isThumbFunction()) Size = alignTo(Size, 4); return Size; } } } unsigned ARMBaseInstrInfo::getInstBundleLength(const MachineInstr &MI) const { unsigned Size = 0; MachineBasicBlock::const_instr_iterator I = MI.getIterator(); MachineBasicBlock::const_instr_iterator E = MI.getParent()->instr_end(); while (++I != E && I->isInsideBundle()) { assert(!I->isBundle() && "No nested bundle!"); Size += getInstSizeInBytes(*I); } return Size; } void ARMBaseInstrInfo::copyFromCPSR(MachineBasicBlock &MBB, MachineBasicBlock::iterator I, unsigned DestReg, bool KillSrc, const ARMSubtarget &Subtarget) const { unsigned Opc = Subtarget.isThumb() ? (Subtarget.isMClass() ? ARM::t2MRS_M : ARM::t2MRS_AR) : ARM::MRS; MachineInstrBuilder MIB = BuildMI(MBB, I, I->getDebugLoc(), get(Opc), DestReg); // There is only 1 A/R class MRS instruction, and it always refers to // APSR. However, there are lots of other possibilities on M-class cores. if (Subtarget.isMClass()) MIB.addImm(0x800); MIB.add(predOps(ARMCC::AL)) .addReg(ARM::CPSR, RegState::Implicit | getKillRegState(KillSrc)); } void ARMBaseInstrInfo::copyToCPSR(MachineBasicBlock &MBB, MachineBasicBlock::iterator I, unsigned SrcReg, bool KillSrc, const ARMSubtarget &Subtarget) const { unsigned Opc = Subtarget.isThumb() ? (Subtarget.isMClass() ? ARM::t2MSR_M : ARM::t2MSR_AR) : ARM::MSR; MachineInstrBuilder MIB = BuildMI(MBB, I, I->getDebugLoc(), get(Opc)); if (Subtarget.isMClass()) MIB.addImm(0x800); else MIB.addImm(8); MIB.addReg(SrcReg, getKillRegState(KillSrc)) .add(predOps(ARMCC::AL)) .addReg(ARM::CPSR, RegState::Implicit | RegState::Define); } void llvm::addUnpredicatedMveVpredNOp(MachineInstrBuilder &MIB) { MIB.addImm(ARMVCC::None); MIB.addReg(0); MIB.addReg(0); // tp_reg } void llvm::addUnpredicatedMveVpredROp(MachineInstrBuilder &MIB, Register DestReg) { addUnpredicatedMveVpredNOp(MIB); MIB.addReg(DestReg, RegState::Undef); } void llvm::addPredicatedMveVpredNOp(MachineInstrBuilder &MIB, unsigned Cond) { MIB.addImm(Cond); MIB.addReg(ARM::VPR, RegState::Implicit); MIB.addReg(0); // tp_reg } void llvm::addPredicatedMveVpredROp(MachineInstrBuilder &MIB, unsigned Cond, unsigned Inactive) { addPredicatedMveVpredNOp(MIB, Cond); MIB.addReg(Inactive); } void ARMBaseInstrInfo::copyPhysReg(MachineBasicBlock &MBB, MachineBasicBlock::iterator I, const DebugLoc &DL, MCRegister DestReg, MCRegister SrcReg, bool KillSrc) const { bool GPRDest = ARM::GPRRegClass.contains(DestReg); bool GPRSrc = ARM::GPRRegClass.contains(SrcReg); if (GPRDest && GPRSrc) { BuildMI(MBB, I, DL, get(ARM::MOVr), DestReg) .addReg(SrcReg, getKillRegState(KillSrc)) .add(predOps(ARMCC::AL)) .add(condCodeOp()); return; } bool SPRDest = ARM::SPRRegClass.contains(DestReg); bool SPRSrc = ARM::SPRRegClass.contains(SrcReg); unsigned Opc = 0; if (SPRDest && SPRSrc) Opc = ARM::VMOVS; else if (GPRDest && SPRSrc) Opc = ARM::VMOVRS; else if (SPRDest && GPRSrc) Opc = ARM::VMOVSR; else if (ARM::DPRRegClass.contains(DestReg, SrcReg) && Subtarget.hasFP64()) Opc = ARM::VMOVD; else if (ARM::QPRRegClass.contains(DestReg, SrcReg)) Opc = Subtarget.hasNEON() ? ARM::VORRq : ARM::MQPRCopy; if (Opc) { MachineInstrBuilder MIB = BuildMI(MBB, I, DL, get(Opc), DestReg); MIB.addReg(SrcReg, getKillRegState(KillSrc)); if (Opc == ARM::VORRq || Opc == ARM::MVE_VORR) MIB.addReg(SrcReg, getKillRegState(KillSrc)); if (Opc == ARM::MVE_VORR) addUnpredicatedMveVpredROp(MIB, DestReg); else if (Opc != ARM::MQPRCopy) MIB.add(predOps(ARMCC::AL)); return; } // Handle register classes that require multiple instructions. unsigned BeginIdx = 0; unsigned SubRegs = 0; int Spacing = 1; // Use VORRq when possible. if (ARM::QQPRRegClass.contains(DestReg, SrcReg)) { Opc = Subtarget.hasNEON() ? ARM::VORRq : ARM::MVE_VORR; BeginIdx = ARM::qsub_0; SubRegs = 2; } else if (ARM::QQQQPRRegClass.contains(DestReg, SrcReg)) { Opc = Subtarget.hasNEON() ? ARM::VORRq : ARM::MVE_VORR; BeginIdx = ARM::qsub_0; SubRegs = 4; // Fall back to VMOVD. } else if (ARM::DPairRegClass.contains(DestReg, SrcReg)) { Opc = ARM::VMOVD; BeginIdx = ARM::dsub_0; SubRegs = 2; } else if (ARM::DTripleRegClass.contains(DestReg, SrcReg)) { Opc = ARM::VMOVD; BeginIdx = ARM::dsub_0; SubRegs = 3; } else if (ARM::DQuadRegClass.contains(DestReg, SrcReg)) { Opc = ARM::VMOVD; BeginIdx = ARM::dsub_0; SubRegs = 4; } else if (ARM::GPRPairRegClass.contains(DestReg, SrcReg)) { Opc = Subtarget.isThumb2() ? ARM::tMOVr : ARM::MOVr; BeginIdx = ARM::gsub_0; SubRegs = 2; } else if (ARM::DPairSpcRegClass.contains(DestReg, SrcReg)) { Opc = ARM::VMOVD; BeginIdx = ARM::dsub_0; SubRegs = 2; Spacing = 2; } else if (ARM::DTripleSpcRegClass.contains(DestReg, SrcReg)) { Opc = ARM::VMOVD; BeginIdx = ARM::dsub_0; SubRegs = 3; Spacing = 2; } else if (ARM::DQuadSpcRegClass.contains(DestReg, SrcReg)) { Opc = ARM::VMOVD; BeginIdx = ARM::dsub_0; SubRegs = 4; Spacing = 2; } else if (ARM::DPRRegClass.contains(DestReg, SrcReg) && !Subtarget.hasFP64()) { Opc = ARM::VMOVS; BeginIdx = ARM::ssub_0; SubRegs = 2; } else if (SrcReg == ARM::CPSR) { copyFromCPSR(MBB, I, DestReg, KillSrc, Subtarget); return; } else if (DestReg == ARM::CPSR) { copyToCPSR(MBB, I, SrcReg, KillSrc, Subtarget); return; } else if (DestReg == ARM::VPR) { assert(ARM::GPRRegClass.contains(SrcReg)); BuildMI(MBB, I, I->getDebugLoc(), get(ARM::VMSR_P0), DestReg) .addReg(SrcReg, getKillRegState(KillSrc)) .add(predOps(ARMCC::AL)); return; } else if (SrcReg == ARM::VPR) { assert(ARM::GPRRegClass.contains(DestReg)); BuildMI(MBB, I, I->getDebugLoc(), get(ARM::VMRS_P0), DestReg) .addReg(SrcReg, getKillRegState(KillSrc)) .add(predOps(ARMCC::AL)); return; } else if (DestReg == ARM::FPSCR_NZCV) { assert(ARM::GPRRegClass.contains(SrcReg)); BuildMI(MBB, I, I->getDebugLoc(), get(ARM::VMSR_FPSCR_NZCVQC), DestReg) .addReg(SrcReg, getKillRegState(KillSrc)) .add(predOps(ARMCC::AL)); return; } else if (SrcReg == ARM::FPSCR_NZCV) { assert(ARM::GPRRegClass.contains(DestReg)); BuildMI(MBB, I, I->getDebugLoc(), get(ARM::VMRS_FPSCR_NZCVQC), DestReg) .addReg(SrcReg, getKillRegState(KillSrc)) .add(predOps(ARMCC::AL)); return; } assert(Opc && "Impossible reg-to-reg copy"); const TargetRegisterInfo *TRI = &getRegisterInfo(); MachineInstrBuilder Mov; // Copy register tuples backward when the first Dest reg overlaps with SrcReg. if (TRI->regsOverlap(SrcReg, TRI->getSubReg(DestReg, BeginIdx))) { BeginIdx = BeginIdx + ((SubRegs - 1) * Spacing); Spacing = -Spacing; } #ifndef NDEBUG SmallSet DstRegs; #endif for (unsigned i = 0; i != SubRegs; ++i) { Register Dst = TRI->getSubReg(DestReg, BeginIdx + i * Spacing); Register Src = TRI->getSubReg(SrcReg, BeginIdx + i * Spacing); assert(Dst && Src && "Bad sub-register"); #ifndef NDEBUG assert(!DstRegs.count(Src) && "destructive vector copy"); DstRegs.insert(Dst); #endif Mov = BuildMI(MBB, I, I->getDebugLoc(), get(Opc), Dst).addReg(Src); // VORR (NEON or MVE) takes two source operands. if (Opc == ARM::VORRq || Opc == ARM::MVE_VORR) { Mov.addReg(Src); } // MVE VORR takes predicate operands in place of an ordinary condition. if (Opc == ARM::MVE_VORR) addUnpredicatedMveVpredROp(Mov, Dst); else Mov = Mov.add(predOps(ARMCC::AL)); // MOVr can set CC. if (Opc == ARM::MOVr) Mov = Mov.add(condCodeOp()); } // Add implicit super-register defs and kills to the last instruction. Mov->addRegisterDefined(DestReg, TRI); if (KillSrc) Mov->addRegisterKilled(SrcReg, TRI); } std::optional ARMBaseInstrInfo::isCopyInstrImpl(const MachineInstr &MI) const { // VMOVRRD is also a copy instruction but it requires // special way of handling. It is more complex copy version // and since that we are not considering it. For recognition // of such instruction isExtractSubregLike MI interface fuction // could be used. // VORRq is considered as a move only if two inputs are // the same register. if (!MI.isMoveReg() || (MI.getOpcode() == ARM::VORRq && MI.getOperand(1).getReg() != MI.getOperand(2).getReg())) return std::nullopt; return DestSourcePair{MI.getOperand(0), MI.getOperand(1)}; } std::optional ARMBaseInstrInfo::describeLoadedValue(const MachineInstr &MI, Register Reg) const { if (auto DstSrcPair = isCopyInstrImpl(MI)) { Register DstReg = DstSrcPair->Destination->getReg(); // TODO: We don't handle cases where the forwarding reg is narrower/wider // than the copy registers. Consider for example: // // s16 = VMOVS s0 // s17 = VMOVS s1 // call @callee(d0) // // We'd like to describe the call site value of d0 as d8, but this requires // gathering and merging the descriptions for the two VMOVS instructions. // // We also don't handle the reverse situation, where the forwarding reg is // narrower than the copy destination: // // d8 = VMOVD d0 // call @callee(s1) // // We need to produce a fragment description (the call site value of s1 is // /not/ just d8). if (DstReg != Reg) return std::nullopt; } return TargetInstrInfo::describeLoadedValue(MI, Reg); } const MachineInstrBuilder & ARMBaseInstrInfo::AddDReg(MachineInstrBuilder &MIB, unsigned Reg, unsigned SubIdx, unsigned State, const TargetRegisterInfo *TRI) const { if (!SubIdx) return MIB.addReg(Reg, State); if (Register::isPhysicalRegister(Reg)) return MIB.addReg(TRI->getSubReg(Reg, SubIdx), State); return MIB.addReg(Reg, State, SubIdx); } void ARMBaseInstrInfo::storeRegToStackSlot(MachineBasicBlock &MBB, MachineBasicBlock::iterator I, Register SrcReg, bool isKill, int FI, const TargetRegisterClass *RC, const TargetRegisterInfo *TRI, Register VReg) const { MachineFunction &MF = *MBB.getParent(); MachineFrameInfo &MFI = MF.getFrameInfo(); Align Alignment = MFI.getObjectAlign(FI); MachineMemOperand *MMO = MF.getMachineMemOperand( MachinePointerInfo::getFixedStack(MF, FI), MachineMemOperand::MOStore, MFI.getObjectSize(FI), Alignment); switch (TRI->getSpillSize(*RC)) { case 2: if (ARM::HPRRegClass.hasSubClassEq(RC)) { BuildMI(MBB, I, DebugLoc(), get(ARM::VSTRH)) .addReg(SrcReg, getKillRegState(isKill)) .addFrameIndex(FI) .addImm(0) .addMemOperand(MMO) .add(predOps(ARMCC::AL)); } else llvm_unreachable("Unknown reg class!"); break; case 4: if (ARM::GPRRegClass.hasSubClassEq(RC)) { BuildMI(MBB, I, DebugLoc(), get(ARM::STRi12)) .addReg(SrcReg, getKillRegState(isKill)) .addFrameIndex(FI) .addImm(0) .addMemOperand(MMO) .add(predOps(ARMCC::AL)); } else if (ARM::SPRRegClass.hasSubClassEq(RC)) { BuildMI(MBB, I, DebugLoc(), get(ARM::VSTRS)) .addReg(SrcReg, getKillRegState(isKill)) .addFrameIndex(FI) .addImm(0) .addMemOperand(MMO) .add(predOps(ARMCC::AL)); } else if (ARM::VCCRRegClass.hasSubClassEq(RC)) { BuildMI(MBB, I, DebugLoc(), get(ARM::VSTR_P0_off)) .addReg(SrcReg, getKillRegState(isKill)) .addFrameIndex(FI) .addImm(0) .addMemOperand(MMO) .add(predOps(ARMCC::AL)); } else llvm_unreachable("Unknown reg class!"); break; case 8: if (ARM::DPRRegClass.hasSubClassEq(RC)) { BuildMI(MBB, I, DebugLoc(), get(ARM::VSTRD)) .addReg(SrcReg, getKillRegState(isKill)) .addFrameIndex(FI) .addImm(0) .addMemOperand(MMO) .add(predOps(ARMCC::AL)); } else if (ARM::GPRPairRegClass.hasSubClassEq(RC)) { if (Subtarget.hasV5TEOps()) { MachineInstrBuilder MIB = BuildMI(MBB, I, DebugLoc(), get(ARM::STRD)); AddDReg(MIB, SrcReg, ARM::gsub_0, getKillRegState(isKill), TRI); AddDReg(MIB, SrcReg, ARM::gsub_1, 0, TRI); MIB.addFrameIndex(FI).addReg(0).addImm(0).addMemOperand(MMO) .add(predOps(ARMCC::AL)); } else { // Fallback to STM instruction, which has existed since the dawn of // time. MachineInstrBuilder MIB = BuildMI(MBB, I, DebugLoc(), get(ARM::STMIA)) .addFrameIndex(FI) .addMemOperand(MMO) .add(predOps(ARMCC::AL)); AddDReg(MIB, SrcReg, ARM::gsub_0, getKillRegState(isKill), TRI); AddDReg(MIB, SrcReg, ARM::gsub_1, 0, TRI); } } else llvm_unreachable("Unknown reg class!"); break; case 16: if (ARM::DPairRegClass.hasSubClassEq(RC) && Subtarget.hasNEON()) { // Use aligned spills if the stack can be realigned. if (Alignment >= 16 && getRegisterInfo().canRealignStack(MF)) { BuildMI(MBB, I, DebugLoc(), get(ARM::VST1q64)) .addFrameIndex(FI) .addImm(16) .addReg(SrcReg, getKillRegState(isKill)) .addMemOperand(MMO) .add(predOps(ARMCC::AL)); } else { BuildMI(MBB, I, DebugLoc(), get(ARM::VSTMQIA)) .addReg(SrcReg, getKillRegState(isKill)) .addFrameIndex(FI) .addMemOperand(MMO) .add(predOps(ARMCC::AL)); } } else if (ARM::QPRRegClass.hasSubClassEq(RC) && Subtarget.hasMVEIntegerOps()) { auto MIB = BuildMI(MBB, I, DebugLoc(), get(ARM::MVE_VSTRWU32)); MIB.addReg(SrcReg, getKillRegState(isKill)) .addFrameIndex(FI) .addImm(0) .addMemOperand(MMO); addUnpredicatedMveVpredNOp(MIB); } else llvm_unreachable("Unknown reg class!"); break; case 24: if (ARM::DTripleRegClass.hasSubClassEq(RC)) { // Use aligned spills if the stack can be realigned. if (Alignment >= 16 && getRegisterInfo().canRealignStack(MF) && Subtarget.hasNEON()) { BuildMI(MBB, I, DebugLoc(), get(ARM::VST1d64TPseudo)) .addFrameIndex(FI) .addImm(16) .addReg(SrcReg, getKillRegState(isKill)) .addMemOperand(MMO) .add(predOps(ARMCC::AL)); } else { MachineInstrBuilder MIB = BuildMI(MBB, I, DebugLoc(), get(ARM::VSTMDIA)) .addFrameIndex(FI) .add(predOps(ARMCC::AL)) .addMemOperand(MMO); MIB = AddDReg(MIB, SrcReg, ARM::dsub_0, getKillRegState(isKill), TRI); MIB = AddDReg(MIB, SrcReg, ARM::dsub_1, 0, TRI); AddDReg(MIB, SrcReg, ARM::dsub_2, 0, TRI); } } else llvm_unreachable("Unknown reg class!"); break; case 32: if (ARM::QQPRRegClass.hasSubClassEq(RC) || ARM::MQQPRRegClass.hasSubClassEq(RC) || ARM::DQuadRegClass.hasSubClassEq(RC)) { if (Alignment >= 16 && getRegisterInfo().canRealignStack(MF) && Subtarget.hasNEON()) { // FIXME: It's possible to only store part of the QQ register if the // spilled def has a sub-register index. BuildMI(MBB, I, DebugLoc(), get(ARM::VST1d64QPseudo)) .addFrameIndex(FI) .addImm(16) .addReg(SrcReg, getKillRegState(isKill)) .addMemOperand(MMO) .add(predOps(ARMCC::AL)); } else if (Subtarget.hasMVEIntegerOps()) { BuildMI(MBB, I, DebugLoc(), get(ARM::MQQPRStore)) .addReg(SrcReg, getKillRegState(isKill)) .addFrameIndex(FI) .addMemOperand(MMO); } else { MachineInstrBuilder MIB = BuildMI(MBB, I, DebugLoc(), get(ARM::VSTMDIA)) .addFrameIndex(FI) .add(predOps(ARMCC::AL)) .addMemOperand(MMO); MIB = AddDReg(MIB, SrcReg, ARM::dsub_0, getKillRegState(isKill), TRI); MIB = AddDReg(MIB, SrcReg, ARM::dsub_1, 0, TRI); MIB = AddDReg(MIB, SrcReg, ARM::dsub_2, 0, TRI); AddDReg(MIB, SrcReg, ARM::dsub_3, 0, TRI); } } else llvm_unreachable("Unknown reg class!"); break; case 64: if (ARM::MQQQQPRRegClass.hasSubClassEq(RC) && Subtarget.hasMVEIntegerOps()) { BuildMI(MBB, I, DebugLoc(), get(ARM::MQQQQPRStore)) .addReg(SrcReg, getKillRegState(isKill)) .addFrameIndex(FI) .addMemOperand(MMO); } else if (ARM::QQQQPRRegClass.hasSubClassEq(RC)) { MachineInstrBuilder MIB = BuildMI(MBB, I, DebugLoc(), get(ARM::VSTMDIA)) .addFrameIndex(FI) .add(predOps(ARMCC::AL)) .addMemOperand(MMO); MIB = AddDReg(MIB, SrcReg, ARM::dsub_0, getKillRegState(isKill), TRI); MIB = AddDReg(MIB, SrcReg, ARM::dsub_1, 0, TRI); MIB = AddDReg(MIB, SrcReg, ARM::dsub_2, 0, TRI); MIB = AddDReg(MIB, SrcReg, ARM::dsub_3, 0, TRI); MIB = AddDReg(MIB, SrcReg, ARM::dsub_4, 0, TRI); MIB = AddDReg(MIB, SrcReg, ARM::dsub_5, 0, TRI); MIB = AddDReg(MIB, SrcReg, ARM::dsub_6, 0, TRI); AddDReg(MIB, SrcReg, ARM::dsub_7, 0, TRI); } else llvm_unreachable("Unknown reg class!"); break; default: llvm_unreachable("Unknown reg class!"); } } unsigned ARMBaseInstrInfo::isStoreToStackSlot(const MachineInstr &MI, int &FrameIndex) const { switch (MI.getOpcode()) { default: break; case ARM::STRrs: case ARM::t2STRs: // FIXME: don't use t2STRs to access frame. if (MI.getOperand(1).isFI() && MI.getOperand(2).isReg() && MI.getOperand(3).isImm() && MI.getOperand(2).getReg() == 0 && MI.getOperand(3).getImm() == 0) { FrameIndex = MI.getOperand(1).getIndex(); return MI.getOperand(0).getReg(); } break; case ARM::STRi12: case ARM::t2STRi12: case ARM::tSTRspi: case ARM::VSTRD: case ARM::VSTRS: case ARM::VSTR_P0_off: case ARM::MVE_VSTRWU32: if (MI.getOperand(1).isFI() && MI.getOperand(2).isImm() && MI.getOperand(2).getImm() == 0) { FrameIndex = MI.getOperand(1).getIndex(); return MI.getOperand(0).getReg(); } break; case ARM::VST1q64: case ARM::VST1d64TPseudo: case ARM::VST1d64QPseudo: if (MI.getOperand(0).isFI() && MI.getOperand(2).getSubReg() == 0) { FrameIndex = MI.getOperand(0).getIndex(); return MI.getOperand(2).getReg(); } break; case ARM::VSTMQIA: if (MI.getOperand(1).isFI() && MI.getOperand(0).getSubReg() == 0) { FrameIndex = MI.getOperand(1).getIndex(); return MI.getOperand(0).getReg(); } break; case ARM::MQQPRStore: case ARM::MQQQQPRStore: if (MI.getOperand(1).isFI()) { FrameIndex = MI.getOperand(1).getIndex(); return MI.getOperand(0).getReg(); } break; } return 0; } unsigned ARMBaseInstrInfo::isStoreToStackSlotPostFE(const MachineInstr &MI, int &FrameIndex) const { SmallVector Accesses; if (MI.mayStore() && hasStoreToStackSlot(MI, Accesses) && Accesses.size() == 1) { FrameIndex = cast(Accesses.front()->getPseudoValue()) ->getFrameIndex(); return true; } return false; } void ARMBaseInstrInfo::loadRegFromStackSlot(MachineBasicBlock &MBB, MachineBasicBlock::iterator I, Register DestReg, int FI, const TargetRegisterClass *RC, const TargetRegisterInfo *TRI, Register VReg) const { DebugLoc DL; if (I != MBB.end()) DL = I->getDebugLoc(); MachineFunction &MF = *MBB.getParent(); MachineFrameInfo &MFI = MF.getFrameInfo(); const Align Alignment = MFI.getObjectAlign(FI); MachineMemOperand *MMO = MF.getMachineMemOperand( MachinePointerInfo::getFixedStack(MF, FI), MachineMemOperand::MOLoad, MFI.getObjectSize(FI), Alignment); switch (TRI->getSpillSize(*RC)) { case 2: if (ARM::HPRRegClass.hasSubClassEq(RC)) { BuildMI(MBB, I, DL, get(ARM::VLDRH), DestReg) .addFrameIndex(FI) .addImm(0) .addMemOperand(MMO) .add(predOps(ARMCC::AL)); } else llvm_unreachable("Unknown reg class!"); break; case 4: if (ARM::GPRRegClass.hasSubClassEq(RC)) { BuildMI(MBB, I, DL, get(ARM::LDRi12), DestReg) .addFrameIndex(FI) .addImm(0) .addMemOperand(MMO) .add(predOps(ARMCC::AL)); } else if (ARM::SPRRegClass.hasSubClassEq(RC)) { BuildMI(MBB, I, DL, get(ARM::VLDRS), DestReg) .addFrameIndex(FI) .addImm(0) .addMemOperand(MMO) .add(predOps(ARMCC::AL)); } else if (ARM::VCCRRegClass.hasSubClassEq(RC)) { BuildMI(MBB, I, DL, get(ARM::VLDR_P0_off), DestReg) .addFrameIndex(FI) .addImm(0) .addMemOperand(MMO) .add(predOps(ARMCC::AL)); } else llvm_unreachable("Unknown reg class!"); break; case 8: if (ARM::DPRRegClass.hasSubClassEq(RC)) { BuildMI(MBB, I, DL, get(ARM::VLDRD), DestReg) .addFrameIndex(FI) .addImm(0) .addMemOperand(MMO) .add(predOps(ARMCC::AL)); } else if (ARM::GPRPairRegClass.hasSubClassEq(RC)) { MachineInstrBuilder MIB; if (Subtarget.hasV5TEOps()) { MIB = BuildMI(MBB, I, DL, get(ARM::LDRD)); AddDReg(MIB, DestReg, ARM::gsub_0, RegState::DefineNoRead, TRI); AddDReg(MIB, DestReg, ARM::gsub_1, RegState::DefineNoRead, TRI); MIB.addFrameIndex(FI).addReg(0).addImm(0).addMemOperand(MMO) .add(predOps(ARMCC::AL)); } else { // Fallback to LDM instruction, which has existed since the dawn of // time. MIB = BuildMI(MBB, I, DL, get(ARM::LDMIA)) .addFrameIndex(FI) .addMemOperand(MMO) .add(predOps(ARMCC::AL)); MIB = AddDReg(MIB, DestReg, ARM::gsub_0, RegState::DefineNoRead, TRI); MIB = AddDReg(MIB, DestReg, ARM::gsub_1, RegState::DefineNoRead, TRI); } if (DestReg.isPhysical()) MIB.addReg(DestReg, RegState::ImplicitDefine); } else llvm_unreachable("Unknown reg class!"); break; case 16: if (ARM::DPairRegClass.hasSubClassEq(RC) && Subtarget.hasNEON()) { if (Alignment >= 16 && getRegisterInfo().canRealignStack(MF)) { BuildMI(MBB, I, DL, get(ARM::VLD1q64), DestReg) .addFrameIndex(FI) .addImm(16) .addMemOperand(MMO) .add(predOps(ARMCC::AL)); } else { BuildMI(MBB, I, DL, get(ARM::VLDMQIA), DestReg) .addFrameIndex(FI) .addMemOperand(MMO) .add(predOps(ARMCC::AL)); } } else if (ARM::QPRRegClass.hasSubClassEq(RC) && Subtarget.hasMVEIntegerOps()) { auto MIB = BuildMI(MBB, I, DL, get(ARM::MVE_VLDRWU32), DestReg); MIB.addFrameIndex(FI) .addImm(0) .addMemOperand(MMO); addUnpredicatedMveVpredNOp(MIB); } else llvm_unreachable("Unknown reg class!"); break; case 24: if (ARM::DTripleRegClass.hasSubClassEq(RC)) { if (Alignment >= 16 && getRegisterInfo().canRealignStack(MF) && Subtarget.hasNEON()) { BuildMI(MBB, I, DL, get(ARM::VLD1d64TPseudo), DestReg) .addFrameIndex(FI) .addImm(16) .addMemOperand(MMO) .add(predOps(ARMCC::AL)); } else { MachineInstrBuilder MIB = BuildMI(MBB, I, DL, get(ARM::VLDMDIA)) .addFrameIndex(FI) .addMemOperand(MMO) .add(predOps(ARMCC::AL)); MIB = AddDReg(MIB, DestReg, ARM::dsub_0, RegState::DefineNoRead, TRI); MIB = AddDReg(MIB, DestReg, ARM::dsub_1, RegState::DefineNoRead, TRI); MIB = AddDReg(MIB, DestReg, ARM::dsub_2, RegState::DefineNoRead, TRI); if (DestReg.isPhysical()) MIB.addReg(DestReg, RegState::ImplicitDefine); } } else llvm_unreachable("Unknown reg class!"); break; case 32: if (ARM::QQPRRegClass.hasSubClassEq(RC) || ARM::MQQPRRegClass.hasSubClassEq(RC) || ARM::DQuadRegClass.hasSubClassEq(RC)) { if (Alignment >= 16 && getRegisterInfo().canRealignStack(MF) && Subtarget.hasNEON()) { BuildMI(MBB, I, DL, get(ARM::VLD1d64QPseudo), DestReg) .addFrameIndex(FI) .addImm(16) .addMemOperand(MMO) .add(predOps(ARMCC::AL)); } else if (Subtarget.hasMVEIntegerOps()) { BuildMI(MBB, I, DL, get(ARM::MQQPRLoad), DestReg) .addFrameIndex(FI) .addMemOperand(MMO); } else { MachineInstrBuilder MIB = BuildMI(MBB, I, DL, get(ARM::VLDMDIA)) .addFrameIndex(FI) .add(predOps(ARMCC::AL)) .addMemOperand(MMO); MIB = AddDReg(MIB, DestReg, ARM::dsub_0, RegState::DefineNoRead, TRI); MIB = AddDReg(MIB, DestReg, ARM::dsub_1, RegState::DefineNoRead, TRI); MIB = AddDReg(MIB, DestReg, ARM::dsub_2, RegState::DefineNoRead, TRI); MIB = AddDReg(MIB, DestReg, ARM::dsub_3, RegState::DefineNoRead, TRI); if (DestReg.isPhysical()) MIB.addReg(DestReg, RegState::ImplicitDefine); } } else llvm_unreachable("Unknown reg class!"); break; case 64: if (ARM::MQQQQPRRegClass.hasSubClassEq(RC) && Subtarget.hasMVEIntegerOps()) { BuildMI(MBB, I, DL, get(ARM::MQQQQPRLoad), DestReg) .addFrameIndex(FI) .addMemOperand(MMO); } else if (ARM::QQQQPRRegClass.hasSubClassEq(RC)) { MachineInstrBuilder MIB = BuildMI(MBB, I, DL, get(ARM::VLDMDIA)) .addFrameIndex(FI) .add(predOps(ARMCC::AL)) .addMemOperand(MMO); MIB = AddDReg(MIB, DestReg, ARM::dsub_0, RegState::DefineNoRead, TRI); MIB = AddDReg(MIB, DestReg, ARM::dsub_1, RegState::DefineNoRead, TRI); MIB = AddDReg(MIB, DestReg, ARM::dsub_2, RegState::DefineNoRead, TRI); MIB = AddDReg(MIB, DestReg, ARM::dsub_3, RegState::DefineNoRead, TRI); MIB = AddDReg(MIB, DestReg, ARM::dsub_4, RegState::DefineNoRead, TRI); MIB = AddDReg(MIB, DestReg, ARM::dsub_5, RegState::DefineNoRead, TRI); MIB = AddDReg(MIB, DestReg, ARM::dsub_6, RegState::DefineNoRead, TRI); MIB = AddDReg(MIB, DestReg, ARM::dsub_7, RegState::DefineNoRead, TRI); if (DestReg.isPhysical()) MIB.addReg(DestReg, RegState::ImplicitDefine); } else llvm_unreachable("Unknown reg class!"); break; default: llvm_unreachable("Unknown regclass!"); } } unsigned ARMBaseInstrInfo::isLoadFromStackSlot(const MachineInstr &MI, int &FrameIndex) const { switch (MI.getOpcode()) { default: break; case ARM::LDRrs: case ARM::t2LDRs: // FIXME: don't use t2LDRs to access frame. if (MI.getOperand(1).isFI() && MI.getOperand(2).isReg() && MI.getOperand(3).isImm() && MI.getOperand(2).getReg() == 0 && MI.getOperand(3).getImm() == 0) { FrameIndex = MI.getOperand(1).getIndex(); return MI.getOperand(0).getReg(); } break; case ARM::LDRi12: case ARM::t2LDRi12: case ARM::tLDRspi: case ARM::VLDRD: case ARM::VLDRS: case ARM::VLDR_P0_off: case ARM::MVE_VLDRWU32: if (MI.getOperand(1).isFI() && MI.getOperand(2).isImm() && MI.getOperand(2).getImm() == 0) { FrameIndex = MI.getOperand(1).getIndex(); return MI.getOperand(0).getReg(); } break; case ARM::VLD1q64: case ARM::VLD1d8TPseudo: case ARM::VLD1d16TPseudo: case ARM::VLD1d32TPseudo: case ARM::VLD1d64TPseudo: case ARM::VLD1d8QPseudo: case ARM::VLD1d16QPseudo: case ARM::VLD1d32QPseudo: case ARM::VLD1d64QPseudo: if (MI.getOperand(1).isFI() && MI.getOperand(0).getSubReg() == 0) { FrameIndex = MI.getOperand(1).getIndex(); return MI.getOperand(0).getReg(); } break; case ARM::VLDMQIA: if (MI.getOperand(1).isFI() && MI.getOperand(0).getSubReg() == 0) { FrameIndex = MI.getOperand(1).getIndex(); return MI.getOperand(0).getReg(); } break; case ARM::MQQPRLoad: case ARM::MQQQQPRLoad: if (MI.getOperand(1).isFI()) { FrameIndex = MI.getOperand(1).getIndex(); return MI.getOperand(0).getReg(); } break; } return 0; } unsigned ARMBaseInstrInfo::isLoadFromStackSlotPostFE(const MachineInstr &MI, int &FrameIndex) const { SmallVector Accesses; if (MI.mayLoad() && hasLoadFromStackSlot(MI, Accesses) && Accesses.size() == 1) { FrameIndex = cast(Accesses.front()->getPseudoValue()) ->getFrameIndex(); return true; } return false; } /// Expands MEMCPY to either LDMIA/STMIA or LDMIA_UPD/STMID_UPD /// depending on whether the result is used. void ARMBaseInstrInfo::expandMEMCPY(MachineBasicBlock::iterator MI) const { bool isThumb1 = Subtarget.isThumb1Only(); bool isThumb2 = Subtarget.isThumb2(); const ARMBaseInstrInfo *TII = Subtarget.getInstrInfo(); DebugLoc dl = MI->getDebugLoc(); MachineBasicBlock *BB = MI->getParent(); MachineInstrBuilder LDM, STM; if (isThumb1 || !MI->getOperand(1).isDead()) { MachineOperand LDWb(MI->getOperand(1)); LDM = BuildMI(*BB, MI, dl, TII->get(isThumb2 ? ARM::t2LDMIA_UPD : isThumb1 ? ARM::tLDMIA_UPD : ARM::LDMIA_UPD)) .add(LDWb); } else { LDM = BuildMI(*BB, MI, dl, TII->get(isThumb2 ? ARM::t2LDMIA : ARM::LDMIA)); } if (isThumb1 || !MI->getOperand(0).isDead()) { MachineOperand STWb(MI->getOperand(0)); STM = BuildMI(*BB, MI, dl, TII->get(isThumb2 ? ARM::t2STMIA_UPD : isThumb1 ? ARM::tSTMIA_UPD : ARM::STMIA_UPD)) .add(STWb); } else { STM = BuildMI(*BB, MI, dl, TII->get(isThumb2 ? ARM::t2STMIA : ARM::STMIA)); } MachineOperand LDBase(MI->getOperand(3)); LDM.add(LDBase).add(predOps(ARMCC::AL)); MachineOperand STBase(MI->getOperand(2)); STM.add(STBase).add(predOps(ARMCC::AL)); // Sort the scratch registers into ascending order. const TargetRegisterInfo &TRI = getRegisterInfo(); SmallVector ScratchRegs; for (MachineOperand &MO : llvm::drop_begin(MI->operands(), 5)) ScratchRegs.push_back(MO.getReg()); llvm::sort(ScratchRegs, [&TRI](const unsigned &Reg1, const unsigned &Reg2) -> bool { return TRI.getEncodingValue(Reg1) < TRI.getEncodingValue(Reg2); }); for (const auto &Reg : ScratchRegs) { LDM.addReg(Reg, RegState::Define); STM.addReg(Reg, RegState::Kill); } BB->erase(MI); } bool ARMBaseInstrInfo::expandPostRAPseudo(MachineInstr &MI) const { if (MI.getOpcode() == TargetOpcode::LOAD_STACK_GUARD) { expandLoadStackGuard(MI); MI.getParent()->erase(MI); return true; } if (MI.getOpcode() == ARM::MEMCPY) { expandMEMCPY(MI); return true; } // This hook gets to expand COPY instructions before they become // copyPhysReg() calls. Look for VMOVS instructions that can legally be // widened to VMOVD. We prefer the VMOVD when possible because it may be // changed into a VORR that can go down the NEON pipeline. if (!MI.isCopy() || Subtarget.dontWidenVMOVS() || !Subtarget.hasFP64()) return false; // Look for a copy between even S-registers. That is where we keep floats // when using NEON v2f32 instructions for f32 arithmetic. Register DstRegS = MI.getOperand(0).getReg(); Register SrcRegS = MI.getOperand(1).getReg(); if (!ARM::SPRRegClass.contains(DstRegS, SrcRegS)) return false; const TargetRegisterInfo *TRI = &getRegisterInfo(); unsigned DstRegD = TRI->getMatchingSuperReg(DstRegS, ARM::ssub_0, &ARM::DPRRegClass); unsigned SrcRegD = TRI->getMatchingSuperReg(SrcRegS, ARM::ssub_0, &ARM::DPRRegClass); if (!DstRegD || !SrcRegD) return false; // We want to widen this into a DstRegD = VMOVD SrcRegD copy. This is only // legal if the COPY already defines the full DstRegD, and it isn't a // sub-register insertion. if (!MI.definesRegister(DstRegD, TRI) || MI.readsRegister(DstRegD, TRI)) return false; // A dead copy shouldn't show up here, but reject it just in case. if (MI.getOperand(0).isDead()) return false; // All clear, widen the COPY. LLVM_DEBUG(dbgs() << "widening: " << MI); MachineInstrBuilder MIB(*MI.getParent()->getParent(), MI); // Get rid of the old implicit-def of DstRegD. Leave it if it defines a Q-reg // or some other super-register. int ImpDefIdx = MI.findRegisterDefOperandIdx(DstRegD); if (ImpDefIdx != -1) MI.removeOperand(ImpDefIdx); // Change the opcode and operands. MI.setDesc(get(ARM::VMOVD)); MI.getOperand(0).setReg(DstRegD); MI.getOperand(1).setReg(SrcRegD); MIB.add(predOps(ARMCC::AL)); // We are now reading SrcRegD instead of SrcRegS. This may upset the // register scavenger and machine verifier, so we need to indicate that we // are reading an undefined value from SrcRegD, but a proper value from // SrcRegS. MI.getOperand(1).setIsUndef(); MIB.addReg(SrcRegS, RegState::Implicit); // SrcRegD may actually contain an unrelated value in the ssub_1 // sub-register. Don't kill it. Only kill the ssub_0 sub-register. if (MI.getOperand(1).isKill()) { MI.getOperand(1).setIsKill(false); MI.addRegisterKilled(SrcRegS, TRI, true); } LLVM_DEBUG(dbgs() << "replaced by: " << MI); return true; } /// Create a copy of a const pool value. Update CPI to the new index and return /// the label UID. static unsigned duplicateCPV(MachineFunction &MF, unsigned &CPI) { MachineConstantPool *MCP = MF.getConstantPool(); ARMFunctionInfo *AFI = MF.getInfo(); const MachineConstantPoolEntry &MCPE = MCP->getConstants()[CPI]; assert(MCPE.isMachineConstantPoolEntry() && "Expecting a machine constantpool entry!"); ARMConstantPoolValue *ACPV = static_cast(MCPE.Val.MachineCPVal); unsigned PCLabelId = AFI->createPICLabelUId(); ARMConstantPoolValue *NewCPV = nullptr; // FIXME: The below assumes PIC relocation model and that the function // is Thumb mode (t1 or t2). PCAdjustment would be 8 for ARM mode PIC, and // zero for non-PIC in ARM or Thumb. The callers are all of thumb LDR // instructions, so that's probably OK, but is PIC always correct when // we get here? if (ACPV->isGlobalValue()) NewCPV = ARMConstantPoolConstant::Create( cast(ACPV)->getGV(), PCLabelId, ARMCP::CPValue, 4, ACPV->getModifier(), ACPV->mustAddCurrentAddress()); else if (ACPV->isExtSymbol()) NewCPV = ARMConstantPoolSymbol:: Create(MF.getFunction().getContext(), cast(ACPV)->getSymbol(), PCLabelId, 4); else if (ACPV->isBlockAddress()) NewCPV = ARMConstantPoolConstant:: Create(cast(ACPV)->getBlockAddress(), PCLabelId, ARMCP::CPBlockAddress, 4); else if (ACPV->isLSDA()) NewCPV = ARMConstantPoolConstant::Create(&MF.getFunction(), PCLabelId, ARMCP::CPLSDA, 4); else if (ACPV->isMachineBasicBlock()) NewCPV = ARMConstantPoolMBB:: Create(MF.getFunction().getContext(), cast(ACPV)->getMBB(), PCLabelId, 4); else llvm_unreachable("Unexpected ARM constantpool value type!!"); CPI = MCP->getConstantPoolIndex(NewCPV, MCPE.getAlign()); return PCLabelId; } void ARMBaseInstrInfo::reMaterialize(MachineBasicBlock &MBB, MachineBasicBlock::iterator I, Register DestReg, unsigned SubIdx, const MachineInstr &Orig, const TargetRegisterInfo &TRI) const { unsigned Opcode = Orig.getOpcode(); switch (Opcode) { default: { MachineInstr *MI = MBB.getParent()->CloneMachineInstr(&Orig); MI->substituteRegister(Orig.getOperand(0).getReg(), DestReg, SubIdx, TRI); MBB.insert(I, MI); break; } case ARM::tLDRpci_pic: case ARM::t2LDRpci_pic: { MachineFunction &MF = *MBB.getParent(); unsigned CPI = Orig.getOperand(1).getIndex(); unsigned PCLabelId = duplicateCPV(MF, CPI); BuildMI(MBB, I, Orig.getDebugLoc(), get(Opcode), DestReg) .addConstantPoolIndex(CPI) .addImm(PCLabelId) .cloneMemRefs(Orig); break; } } } MachineInstr & ARMBaseInstrInfo::duplicate(MachineBasicBlock &MBB, MachineBasicBlock::iterator InsertBefore, const MachineInstr &Orig) const { MachineInstr &Cloned = TargetInstrInfo::duplicate(MBB, InsertBefore, Orig); MachineBasicBlock::instr_iterator I = Cloned.getIterator(); for (;;) { switch (I->getOpcode()) { case ARM::tLDRpci_pic: case ARM::t2LDRpci_pic: { MachineFunction &MF = *MBB.getParent(); unsigned CPI = I->getOperand(1).getIndex(); unsigned PCLabelId = duplicateCPV(MF, CPI); I->getOperand(1).setIndex(CPI); I->getOperand(2).setImm(PCLabelId); break; } } if (!I->isBundledWithSucc()) break; ++I; } return Cloned; } bool ARMBaseInstrInfo::produceSameValue(const MachineInstr &MI0, const MachineInstr &MI1, const MachineRegisterInfo *MRI) const { unsigned Opcode = MI0.getOpcode(); if (Opcode == ARM::t2LDRpci || Opcode == ARM::t2LDRpci_pic || Opcode == ARM::tLDRpci || Opcode == ARM::tLDRpci_pic || Opcode == ARM::LDRLIT_ga_pcrel || Opcode == ARM::LDRLIT_ga_pcrel_ldr || Opcode == ARM::tLDRLIT_ga_pcrel || Opcode == ARM::t2LDRLIT_ga_pcrel || Opcode == ARM::MOV_ga_pcrel || Opcode == ARM::MOV_ga_pcrel_ldr || Opcode == ARM::t2MOV_ga_pcrel) { if (MI1.getOpcode() != Opcode) return false; if (MI0.getNumOperands() != MI1.getNumOperands()) return false; const MachineOperand &MO0 = MI0.getOperand(1); const MachineOperand &MO1 = MI1.getOperand(1); if (MO0.getOffset() != MO1.getOffset()) return false; if (Opcode == ARM::LDRLIT_ga_pcrel || Opcode == ARM::LDRLIT_ga_pcrel_ldr || Opcode == ARM::tLDRLIT_ga_pcrel || Opcode == ARM::t2LDRLIT_ga_pcrel || Opcode == ARM::MOV_ga_pcrel || Opcode == ARM::MOV_ga_pcrel_ldr || Opcode == ARM::t2MOV_ga_pcrel) // Ignore the PC labels. return MO0.getGlobal() == MO1.getGlobal(); const MachineFunction *MF = MI0.getParent()->getParent(); const MachineConstantPool *MCP = MF->getConstantPool(); int CPI0 = MO0.getIndex(); int CPI1 = MO1.getIndex(); const MachineConstantPoolEntry &MCPE0 = MCP->getConstants()[CPI0]; const MachineConstantPoolEntry &MCPE1 = MCP->getConstants()[CPI1]; bool isARMCP0 = MCPE0.isMachineConstantPoolEntry(); bool isARMCP1 = MCPE1.isMachineConstantPoolEntry(); if (isARMCP0 && isARMCP1) { ARMConstantPoolValue *ACPV0 = static_cast(MCPE0.Val.MachineCPVal); ARMConstantPoolValue *ACPV1 = static_cast(MCPE1.Val.MachineCPVal); return ACPV0->hasSameValue(ACPV1); } else if (!isARMCP0 && !isARMCP1) { return MCPE0.Val.ConstVal == MCPE1.Val.ConstVal; } return false; } else if (Opcode == ARM::PICLDR) { if (MI1.getOpcode() != Opcode) return false; if (MI0.getNumOperands() != MI1.getNumOperands()) return false; Register Addr0 = MI0.getOperand(1).getReg(); Register Addr1 = MI1.getOperand(1).getReg(); if (Addr0 != Addr1) { if (!MRI || !Addr0.isVirtual() || !Addr1.isVirtual()) return false; // This assumes SSA form. MachineInstr *Def0 = MRI->getVRegDef(Addr0); MachineInstr *Def1 = MRI->getVRegDef(Addr1); // Check if the loaded value, e.g. a constantpool of a global address, are // the same. if (!produceSameValue(*Def0, *Def1, MRI)) return false; } for (unsigned i = 3, e = MI0.getNumOperands(); i != e; ++i) { // %12 = PICLDR %11, 0, 14, %noreg const MachineOperand &MO0 = MI0.getOperand(i); const MachineOperand &MO1 = MI1.getOperand(i); if (!MO0.isIdenticalTo(MO1)) return false; } return true; } return MI0.isIdenticalTo(MI1, MachineInstr::IgnoreVRegDefs); } /// areLoadsFromSameBasePtr - This is used by the pre-regalloc scheduler to /// determine if two loads are loading from the same base address. It should /// only return true if the base pointers are the same and the only differences /// between the two addresses is the offset. It also returns the offsets by /// reference. /// /// FIXME: remove this in favor of the MachineInstr interface once pre-RA-sched /// is permanently disabled. bool ARMBaseInstrInfo::areLoadsFromSameBasePtr(SDNode *Load1, SDNode *Load2, int64_t &Offset1, int64_t &Offset2) const { // Don't worry about Thumb: just ARM and Thumb2. if (Subtarget.isThumb1Only()) return false; if (!Load1->isMachineOpcode() || !Load2->isMachineOpcode()) return false; switch (Load1->getMachineOpcode()) { default: return false; case ARM::LDRi12: case ARM::LDRBi12: case ARM::LDRD: case ARM::LDRH: case ARM::LDRSB: case ARM::LDRSH: case ARM::VLDRD: case ARM::VLDRS: case ARM::t2LDRi8: case ARM::t2LDRBi8: case ARM::t2LDRDi8: case ARM::t2LDRSHi8: case ARM::t2LDRi12: case ARM::t2LDRBi12: case ARM::t2LDRSHi12: break; } switch (Load2->getMachineOpcode()) { default: return false; case ARM::LDRi12: case ARM::LDRBi12: case ARM::LDRD: case ARM::LDRH: case ARM::LDRSB: case ARM::LDRSH: case ARM::VLDRD: case ARM::VLDRS: case ARM::t2LDRi8: case ARM::t2LDRBi8: case ARM::t2LDRSHi8: case ARM::t2LDRi12: case ARM::t2LDRBi12: case ARM::t2LDRSHi12: break; } // Check if base addresses and chain operands match. if (Load1->getOperand(0) != Load2->getOperand(0) || Load1->getOperand(4) != Load2->getOperand(4)) return false; // Index should be Reg0. if (Load1->getOperand(3) != Load2->getOperand(3)) return false; // Determine the offsets. if (isa(Load1->getOperand(1)) && isa(Load2->getOperand(1))) { Offset1 = cast(Load1->getOperand(1))->getSExtValue(); Offset2 = cast(Load2->getOperand(1))->getSExtValue(); return true; } return false; } /// shouldScheduleLoadsNear - This is a used by the pre-regalloc scheduler to /// determine (in conjunction with areLoadsFromSameBasePtr) if two loads should /// be scheduled togther. On some targets if two loads are loading from /// addresses in the same cache line, it's better if they are scheduled /// together. This function takes two integers that represent the load offsets /// from the common base address. It returns true if it decides it's desirable /// to schedule the two loads together. "NumLoads" is the number of loads that /// have already been scheduled after Load1. /// /// FIXME: remove this in favor of the MachineInstr interface once pre-RA-sched /// is permanently disabled. bool ARMBaseInstrInfo::shouldScheduleLoadsNear(SDNode *Load1, SDNode *Load2, int64_t Offset1, int64_t Offset2, unsigned NumLoads) const { // Don't worry about Thumb: just ARM and Thumb2. if (Subtarget.isThumb1Only()) return false; assert(Offset2 > Offset1); if ((Offset2 - Offset1) / 8 > 64) return false; // Check if the machine opcodes are different. If they are different // then we consider them to not be of the same base address, // EXCEPT in the case of Thumb2 byte loads where one is LDRBi8 and the other LDRBi12. // In this case, they are considered to be the same because they are different // encoding forms of the same basic instruction. if ((Load1->getMachineOpcode() != Load2->getMachineOpcode()) && !((Load1->getMachineOpcode() == ARM::t2LDRBi8 && Load2->getMachineOpcode() == ARM::t2LDRBi12) || (Load1->getMachineOpcode() == ARM::t2LDRBi12 && Load2->getMachineOpcode() == ARM::t2LDRBi8))) return false; // FIXME: overly conservative? // Four loads in a row should be sufficient. if (NumLoads >= 3) return false; return true; } bool ARMBaseInstrInfo::isSchedulingBoundary(const MachineInstr &MI, const MachineBasicBlock *MBB, const MachineFunction &MF) const { // Debug info is never a scheduling boundary. It's necessary to be explicit // due to the special treatment of IT instructions below, otherwise a // dbg_value followed by an IT will result in the IT instruction being // considered a scheduling hazard, which is wrong. It should be the actual // instruction preceding the dbg_value instruction(s), just like it is // when debug info is not present. if (MI.isDebugInstr()) return false; // Terminators and labels can't be scheduled around. if (MI.isTerminator() || MI.isPosition()) return true; // INLINEASM_BR can jump to another block if (MI.getOpcode() == TargetOpcode::INLINEASM_BR) return true; if (isSEHInstruction(MI)) return true; // Treat the start of the IT block as a scheduling boundary, but schedule // t2IT along with all instructions following it. // FIXME: This is a big hammer. But the alternative is to add all potential // true and anti dependencies to IT block instructions as implicit operands // to the t2IT instruction. The added compile time and complexity does not // seem worth it. MachineBasicBlock::const_iterator I = MI; // Make sure to skip any debug instructions while (++I != MBB->end() && I->isDebugInstr()) ; if (I != MBB->end() && I->getOpcode() == ARM::t2IT) return true; // Don't attempt to schedule around any instruction that defines // a stack-oriented pointer, as it's unlikely to be profitable. This // saves compile time, because it doesn't require every single // stack slot reference to depend on the instruction that does the // modification. // Calls don't actually change the stack pointer, even if they have imp-defs. // No ARM calling conventions change the stack pointer. (X86 calling // conventions sometimes do). if (!MI.isCall() && MI.definesRegister(ARM::SP)) return true; return false; } bool ARMBaseInstrInfo:: isProfitableToIfCvt(MachineBasicBlock &MBB, unsigned NumCycles, unsigned ExtraPredCycles, BranchProbability Probability) const { if (!NumCycles) return false; // If we are optimizing for size, see if the branch in the predecessor can be // lowered to cbn?z by the constant island lowering pass, and return false if // so. This results in a shorter instruction sequence. if (MBB.getParent()->getFunction().hasOptSize()) { MachineBasicBlock *Pred = *MBB.pred_begin(); if (!Pred->empty()) { MachineInstr *LastMI = &*Pred->rbegin(); if (LastMI->getOpcode() == ARM::t2Bcc) { const TargetRegisterInfo *TRI = &getRegisterInfo(); MachineInstr *CmpMI = findCMPToFoldIntoCBZ(LastMI, TRI); if (CmpMI) return false; } } } return isProfitableToIfCvt(MBB, NumCycles, ExtraPredCycles, MBB, 0, 0, Probability); } bool ARMBaseInstrInfo:: isProfitableToIfCvt(MachineBasicBlock &TBB, unsigned TCycles, unsigned TExtra, MachineBasicBlock &FBB, unsigned FCycles, unsigned FExtra, BranchProbability Probability) const { if (!TCycles) return false; // In thumb code we often end up trading one branch for a IT block, and // if we are cloning the instruction can increase code size. Prevent // blocks with multiple predecesors from being ifcvted to prevent this // cloning. if (Subtarget.isThumb2() && TBB.getParent()->getFunction().hasMinSize()) { if (TBB.pred_size() != 1 || FBB.pred_size() != 1) return false; } // Attempt to estimate the relative costs of predication versus branching. // Here we scale up each component of UnpredCost to avoid precision issue when // scaling TCycles/FCycles by Probability. const unsigned ScalingUpFactor = 1024; unsigned PredCost = (TCycles + FCycles + TExtra + FExtra) * ScalingUpFactor; unsigned UnpredCost; if (!Subtarget.hasBranchPredictor()) { // When we don't have a branch predictor it's always cheaper to not take a // branch than take it, so we have to take that into account. unsigned NotTakenBranchCost = 1; unsigned TakenBranchCost = Subtarget.getMispredictionPenalty(); unsigned TUnpredCycles, FUnpredCycles; if (!FCycles) { // Triangle: TBB is the fallthrough TUnpredCycles = TCycles + NotTakenBranchCost; FUnpredCycles = TakenBranchCost; } else { // Diamond: TBB is the block that is branched to, FBB is the fallthrough TUnpredCycles = TCycles + TakenBranchCost; FUnpredCycles = FCycles + NotTakenBranchCost; // The branch at the end of FBB will disappear when it's predicated, so // discount it from PredCost. PredCost -= 1 * ScalingUpFactor; } // The total cost is the cost of each path scaled by their probabilites unsigned TUnpredCost = Probability.scale(TUnpredCycles * ScalingUpFactor); unsigned FUnpredCost = Probability.getCompl().scale(FUnpredCycles * ScalingUpFactor); UnpredCost = TUnpredCost + FUnpredCost; // When predicating assume that the first IT can be folded away but later // ones cost one cycle each if (Subtarget.isThumb2() && TCycles + FCycles > 4) { PredCost += ((TCycles + FCycles - 4) / 4) * ScalingUpFactor; } } else { unsigned TUnpredCost = Probability.scale(TCycles * ScalingUpFactor); unsigned FUnpredCost = Probability.getCompl().scale(FCycles * ScalingUpFactor); UnpredCost = TUnpredCost + FUnpredCost; UnpredCost += 1 * ScalingUpFactor; // The branch itself UnpredCost += Subtarget.getMispredictionPenalty() * ScalingUpFactor / 10; } return PredCost <= UnpredCost; } unsigned ARMBaseInstrInfo::extraSizeToPredicateInstructions(const MachineFunction &MF, unsigned NumInsts) const { // Thumb2 needs a 2-byte IT instruction to predicate up to 4 instructions. // ARM has a condition code field in every predicable instruction, using it // doesn't change code size. if (!Subtarget.isThumb2()) return 0; // It's possible that the size of the IT is restricted to a single block. unsigned MaxInsts = Subtarget.restrictIT() ? 1 : 4; return divideCeil(NumInsts, MaxInsts) * 2; } unsigned ARMBaseInstrInfo::predictBranchSizeForIfCvt(MachineInstr &MI) const { // If this branch is likely to be folded into the comparison to form a // CB(N)Z, then removing it won't reduce code size at all, because that will // just replace the CB(N)Z with a CMP. if (MI.getOpcode() == ARM::t2Bcc && findCMPToFoldIntoCBZ(&MI, &getRegisterInfo())) return 0; unsigned Size = getInstSizeInBytes(MI); // For Thumb2, all branches are 32-bit instructions during the if conversion // pass, but may be replaced with 16-bit instructions during size reduction. // Since the branches considered by if conversion tend to be forward branches // over small basic blocks, they are very likely to be in range for the // narrow instructions, so we assume the final code size will be half what it // currently is. if (Subtarget.isThumb2()) Size /= 2; return Size; } bool ARMBaseInstrInfo::isProfitableToUnpredicate(MachineBasicBlock &TMBB, MachineBasicBlock &FMBB) const { // Reduce false anti-dependencies to let the target's out-of-order execution // engine do its thing. return Subtarget.isProfitableToUnpredicate(); } /// getInstrPredicate - If instruction is predicated, returns its predicate /// condition, otherwise returns AL. It also returns the condition code /// register by reference. ARMCC::CondCodes llvm::getInstrPredicate(const MachineInstr &MI, Register &PredReg) { int PIdx = MI.findFirstPredOperandIdx(); if (PIdx == -1) { PredReg = 0; return ARMCC::AL; } PredReg = MI.getOperand(PIdx+1).getReg(); return (ARMCC::CondCodes)MI.getOperand(PIdx).getImm(); } unsigned llvm::getMatchingCondBranchOpcode(unsigned Opc) { if (Opc == ARM::B) return ARM::Bcc; if (Opc == ARM::tB) return ARM::tBcc; if (Opc == ARM::t2B) return ARM::t2Bcc; llvm_unreachable("Unknown unconditional branch opcode!"); } MachineInstr *ARMBaseInstrInfo::commuteInstructionImpl(MachineInstr &MI, bool NewMI, unsigned OpIdx1, unsigned OpIdx2) const { switch (MI.getOpcode()) { case ARM::MOVCCr: case ARM::t2MOVCCr: { // MOVCC can be commuted by inverting the condition. Register PredReg; ARMCC::CondCodes CC = getInstrPredicate(MI, PredReg); // MOVCC AL can't be inverted. Shouldn't happen. if (CC == ARMCC::AL || PredReg != ARM::CPSR) return nullptr; MachineInstr *CommutedMI = TargetInstrInfo::commuteInstructionImpl(MI, NewMI, OpIdx1, OpIdx2); if (!CommutedMI) return nullptr; // After swapping the MOVCC operands, also invert the condition. CommutedMI->getOperand(CommutedMI->findFirstPredOperandIdx()) .setImm(ARMCC::getOppositeCondition(CC)); return CommutedMI; } } return TargetInstrInfo::commuteInstructionImpl(MI, NewMI, OpIdx1, OpIdx2); } /// Identify instructions that can be folded into a MOVCC instruction, and /// return the defining instruction. MachineInstr * ARMBaseInstrInfo::canFoldIntoMOVCC(Register Reg, const MachineRegisterInfo &MRI, const TargetInstrInfo *TII) const { if (!Reg.isVirtual()) return nullptr; if (!MRI.hasOneNonDBGUse(Reg)) return nullptr; MachineInstr *MI = MRI.getVRegDef(Reg); if (!MI) return nullptr; // Check if MI can be predicated and folded into the MOVCC. if (!isPredicable(*MI)) return nullptr; // Check if MI has any non-dead defs or physreg uses. This also detects // predicated instructions which will be reading CPSR. for (const MachineOperand &MO : llvm::drop_begin(MI->operands(), 1)) { // Reject frame index operands, PEI can't handle the predicated pseudos. if (MO.isFI() || MO.isCPI() || MO.isJTI()) return nullptr; if (!MO.isReg()) continue; // MI can't have any tied operands, that would conflict with predication. if (MO.isTied()) return nullptr; if (MO.getReg().isPhysical()) return nullptr; if (MO.isDef() && !MO.isDead()) return nullptr; } bool DontMoveAcrossStores = true; if (!MI->isSafeToMove(/* AliasAnalysis = */ nullptr, DontMoveAcrossStores)) return nullptr; return MI; } bool ARMBaseInstrInfo::analyzeSelect(const MachineInstr &MI, SmallVectorImpl &Cond, unsigned &TrueOp, unsigned &FalseOp, bool &Optimizable) const { assert((MI.getOpcode() == ARM::MOVCCr || MI.getOpcode() == ARM::t2MOVCCr) && "Unknown select instruction"); // MOVCC operands: // 0: Def. // 1: True use. // 2: False use. // 3: Condition code. // 4: CPSR use. TrueOp = 1; FalseOp = 2; Cond.push_back(MI.getOperand(3)); Cond.push_back(MI.getOperand(4)); // We can always fold a def. Optimizable = true; return false; } MachineInstr * ARMBaseInstrInfo::optimizeSelect(MachineInstr &MI, SmallPtrSetImpl &SeenMIs, bool PreferFalse) const { assert((MI.getOpcode() == ARM::MOVCCr || MI.getOpcode() == ARM::t2MOVCCr) && "Unknown select instruction"); MachineRegisterInfo &MRI = MI.getParent()->getParent()->getRegInfo(); MachineInstr *DefMI = canFoldIntoMOVCC(MI.getOperand(2).getReg(), MRI, this); bool Invert = !DefMI; if (!DefMI) DefMI = canFoldIntoMOVCC(MI.getOperand(1).getReg(), MRI, this); if (!DefMI) return nullptr; // Find new register class to use. MachineOperand FalseReg = MI.getOperand(Invert ? 2 : 1); MachineOperand TrueReg = MI.getOperand(Invert ? 1 : 2); Register DestReg = MI.getOperand(0).getReg(); const TargetRegisterClass *FalseClass = MRI.getRegClass(FalseReg.getReg()); const TargetRegisterClass *TrueClass = MRI.getRegClass(TrueReg.getReg()); if (!MRI.constrainRegClass(DestReg, FalseClass)) return nullptr; if (!MRI.constrainRegClass(DestReg, TrueClass)) return nullptr; // Create a new predicated version of DefMI. // Rfalse is the first use. MachineInstrBuilder NewMI = BuildMI(*MI.getParent(), MI, MI.getDebugLoc(), DefMI->getDesc(), DestReg); // Copy all the DefMI operands, excluding its (null) predicate. const MCInstrDesc &DefDesc = DefMI->getDesc(); for (unsigned i = 1, e = DefDesc.getNumOperands(); i != e && !DefDesc.operands()[i].isPredicate(); ++i) NewMI.add(DefMI->getOperand(i)); unsigned CondCode = MI.getOperand(3).getImm(); if (Invert) NewMI.addImm(ARMCC::getOppositeCondition(ARMCC::CondCodes(CondCode))); else NewMI.addImm(CondCode); NewMI.add(MI.getOperand(4)); // DefMI is not the -S version that sets CPSR, so add an optional %noreg. if (NewMI->hasOptionalDef()) NewMI.add(condCodeOp()); // The output register value when the predicate is false is an implicit // register operand tied to the first def. // The tie makes the register allocator ensure the FalseReg is allocated the // same register as operand 0. FalseReg.setImplicit(); NewMI.add(FalseReg); NewMI->tieOperands(0, NewMI->getNumOperands() - 1); // Update SeenMIs set: register newly created MI and erase removed DefMI. SeenMIs.insert(NewMI); SeenMIs.erase(DefMI); // If MI is inside a loop, and DefMI is outside the loop, then kill flags on // DefMI would be invalid when tranferred inside the loop. Checking for a // loop is expensive, but at least remove kill flags if they are in different // BBs. if (DefMI->getParent() != MI.getParent()) NewMI->clearKillInfo(); // The caller will erase MI, but not DefMI. DefMI->eraseFromParent(); return NewMI; } /// Map pseudo instructions that imply an 'S' bit onto real opcodes. Whether the /// instruction is encoded with an 'S' bit is determined by the optional CPSR /// def operand. /// /// This will go away once we can teach tblgen how to set the optional CPSR def /// operand itself. struct AddSubFlagsOpcodePair { uint16_t PseudoOpc; uint16_t MachineOpc; }; static const AddSubFlagsOpcodePair AddSubFlagsOpcodeMap[] = { {ARM::ADDSri, ARM::ADDri}, {ARM::ADDSrr, ARM::ADDrr}, {ARM::ADDSrsi, ARM::ADDrsi}, {ARM::ADDSrsr, ARM::ADDrsr}, {ARM::SUBSri, ARM::SUBri}, {ARM::SUBSrr, ARM::SUBrr}, {ARM::SUBSrsi, ARM::SUBrsi}, {ARM::SUBSrsr, ARM::SUBrsr}, {ARM::RSBSri, ARM::RSBri}, {ARM::RSBSrsi, ARM::RSBrsi}, {ARM::RSBSrsr, ARM::RSBrsr}, {ARM::tADDSi3, ARM::tADDi3}, {ARM::tADDSi8, ARM::tADDi8}, {ARM::tADDSrr, ARM::tADDrr}, {ARM::tADCS, ARM::tADC}, {ARM::tSUBSi3, ARM::tSUBi3}, {ARM::tSUBSi8, ARM::tSUBi8}, {ARM::tSUBSrr, ARM::tSUBrr}, {ARM::tSBCS, ARM::tSBC}, {ARM::tRSBS, ARM::tRSB}, {ARM::tLSLSri, ARM::tLSLri}, {ARM::t2ADDSri, ARM::t2ADDri}, {ARM::t2ADDSrr, ARM::t2ADDrr}, {ARM::t2ADDSrs, ARM::t2ADDrs}, {ARM::t2SUBSri, ARM::t2SUBri}, {ARM::t2SUBSrr, ARM::t2SUBrr}, {ARM::t2SUBSrs, ARM::t2SUBrs}, {ARM::t2RSBSri, ARM::t2RSBri}, {ARM::t2RSBSrs, ARM::t2RSBrs}, }; unsigned llvm::convertAddSubFlagsOpcode(unsigned OldOpc) { for (const auto &Entry : AddSubFlagsOpcodeMap) if (OldOpc == Entry.PseudoOpc) return Entry.MachineOpc; return 0; } void llvm::emitARMRegPlusImmediate(MachineBasicBlock &MBB, MachineBasicBlock::iterator &MBBI, const DebugLoc &dl, Register DestReg, Register BaseReg, int NumBytes, ARMCC::CondCodes Pred, Register PredReg, const ARMBaseInstrInfo &TII, unsigned MIFlags) { if (NumBytes == 0 && DestReg != BaseReg) { BuildMI(MBB, MBBI, dl, TII.get(ARM::MOVr), DestReg) .addReg(BaseReg, RegState::Kill) .add(predOps(Pred, PredReg)) .add(condCodeOp()) .setMIFlags(MIFlags); return; } bool isSub = NumBytes < 0; if (isSub) NumBytes = -NumBytes; while (NumBytes) { unsigned RotAmt = ARM_AM::getSOImmValRotate(NumBytes); unsigned ThisVal = NumBytes & llvm::rotr(0xFF, RotAmt); assert(ThisVal && "Didn't extract field correctly"); // We will handle these bits from offset, clear them. NumBytes &= ~ThisVal; assert(ARM_AM::getSOImmVal(ThisVal) != -1 && "Bit extraction didn't work?"); // Build the new ADD / SUB. unsigned Opc = isSub ? ARM::SUBri : ARM::ADDri; BuildMI(MBB, MBBI, dl, TII.get(Opc), DestReg) .addReg(BaseReg, RegState::Kill) .addImm(ThisVal) .add(predOps(Pred, PredReg)) .add(condCodeOp()) .setMIFlags(MIFlags); BaseReg = DestReg; } } bool llvm::tryFoldSPUpdateIntoPushPop(const ARMSubtarget &Subtarget, MachineFunction &MF, MachineInstr *MI, unsigned NumBytes) { // This optimisation potentially adds lots of load and store // micro-operations, it's only really a great benefit to code-size. if (!Subtarget.hasMinSize()) return false; // If only one register is pushed/popped, LLVM can use an LDR/STR // instead. We can't modify those so make sure we're dealing with an // instruction we understand. bool IsPop = isPopOpcode(MI->getOpcode()); bool IsPush = isPushOpcode(MI->getOpcode()); if (!IsPush && !IsPop) return false; bool IsVFPPushPop = MI->getOpcode() == ARM::VSTMDDB_UPD || MI->getOpcode() == ARM::VLDMDIA_UPD; bool IsT1PushPop = MI->getOpcode() == ARM::tPUSH || MI->getOpcode() == ARM::tPOP || MI->getOpcode() == ARM::tPOP_RET; assert((IsT1PushPop || (MI->getOperand(0).getReg() == ARM::SP && MI->getOperand(1).getReg() == ARM::SP)) && "trying to fold sp update into non-sp-updating push/pop"); // The VFP push & pop act on D-registers, so we can only fold an adjustment // by a multiple of 8 bytes in correctly. Similarly rN is 4-bytes. Don't try // if this is violated. if (NumBytes % (IsVFPPushPop ? 8 : 4) != 0) return false; // ARM and Thumb2 push/pop insts have explicit "sp, sp" operands (+ // pred) so the list starts at 4. Thumb1 starts after the predicate. int RegListIdx = IsT1PushPop ? 2 : 4; // Calculate the space we'll need in terms of registers. unsigned RegsNeeded; const TargetRegisterClass *RegClass; if (IsVFPPushPop) { RegsNeeded = NumBytes / 8; RegClass = &ARM::DPRRegClass; } else { RegsNeeded = NumBytes / 4; RegClass = &ARM::GPRRegClass; } // We're going to have to strip all list operands off before // re-adding them since the order matters, so save the existing ones // for later. SmallVector RegList; // We're also going to need the first register transferred by this // instruction, which won't necessarily be the first register in the list. unsigned FirstRegEnc = -1; const TargetRegisterInfo *TRI = MF.getRegInfo().getTargetRegisterInfo(); for (int i = MI->getNumOperands() - 1; i >= RegListIdx; --i) { MachineOperand &MO = MI->getOperand(i); RegList.push_back(MO); if (MO.isReg() && !MO.isImplicit() && TRI->getEncodingValue(MO.getReg()) < FirstRegEnc) FirstRegEnc = TRI->getEncodingValue(MO.getReg()); } const MCPhysReg *CSRegs = TRI->getCalleeSavedRegs(&MF); // Now try to find enough space in the reglist to allocate NumBytes. for (int CurRegEnc = FirstRegEnc - 1; CurRegEnc >= 0 && RegsNeeded; --CurRegEnc) { unsigned CurReg = RegClass->getRegister(CurRegEnc); if (IsT1PushPop && CurRegEnc > TRI->getEncodingValue(ARM::R7)) continue; if (!IsPop) { // Pushing any register is completely harmless, mark the register involved // as undef since we don't care about its value and must not restore it // during stack unwinding. RegList.push_back(MachineOperand::CreateReg(CurReg, false, false, false, false, true)); --RegsNeeded; continue; } // However, we can only pop an extra register if it's not live. For // registers live within the function we might clobber a return value // register; the other way a register can be live here is if it's // callee-saved. if (isCalleeSavedRegister(CurReg, CSRegs) || MI->getParent()->computeRegisterLiveness(TRI, CurReg, MI) != MachineBasicBlock::LQR_Dead) { // VFP pops don't allow holes in the register list, so any skip is fatal // for our transformation. GPR pops do, so we should just keep looking. if (IsVFPPushPop) return false; else continue; } // Mark the unimportant registers as in the POP. RegList.push_back(MachineOperand::CreateReg(CurReg, true, false, false, true)); --RegsNeeded; } if (RegsNeeded > 0) return false; // Finally we know we can profitably perform the optimisation so go // ahead: strip all existing registers off and add them back again // in the right order. for (int i = MI->getNumOperands() - 1; i >= RegListIdx; --i) MI->removeOperand(i); // Add the complete list back in. MachineInstrBuilder MIB(MF, &*MI); for (const MachineOperand &MO : llvm::reverse(RegList)) MIB.add(MO); return true; } bool llvm::rewriteARMFrameIndex(MachineInstr &MI, unsigned FrameRegIdx, Register FrameReg, int &Offset, const ARMBaseInstrInfo &TII) { unsigned Opcode = MI.getOpcode(); const MCInstrDesc &Desc = MI.getDesc(); unsigned AddrMode = (Desc.TSFlags & ARMII::AddrModeMask); bool isSub = false; // Memory operands in inline assembly always use AddrMode2. if (Opcode == ARM::INLINEASM || Opcode == ARM::INLINEASM_BR) AddrMode = ARMII::AddrMode2; if (Opcode == ARM::ADDri) { Offset += MI.getOperand(FrameRegIdx+1).getImm(); if (Offset == 0) { // Turn it into a move. MI.setDesc(TII.get(ARM::MOVr)); MI.getOperand(FrameRegIdx).ChangeToRegister(FrameReg, false); MI.removeOperand(FrameRegIdx+1); Offset = 0; return true; } else if (Offset < 0) { Offset = -Offset; isSub = true; MI.setDesc(TII.get(ARM::SUBri)); } // Common case: small offset, fits into instruction. if (ARM_AM::getSOImmVal(Offset) != -1) { // Replace the FrameIndex with sp / fp MI.getOperand(FrameRegIdx).ChangeToRegister(FrameReg, false); MI.getOperand(FrameRegIdx+1).ChangeToImmediate(Offset); Offset = 0; return true; } // Otherwise, pull as much of the immedidate into this ADDri/SUBri // as possible. unsigned RotAmt = ARM_AM::getSOImmValRotate(Offset); unsigned ThisImmVal = Offset & llvm::rotr(0xFF, RotAmt); // We will handle these bits from offset, clear them. Offset &= ~ThisImmVal; // Get the properly encoded SOImmVal field. assert(ARM_AM::getSOImmVal(ThisImmVal) != -1 && "Bit extraction didn't work?"); MI.getOperand(FrameRegIdx+1).ChangeToImmediate(ThisImmVal); } else { unsigned ImmIdx = 0; int InstrOffs = 0; unsigned NumBits = 0; unsigned Scale = 1; switch (AddrMode) { case ARMII::AddrMode_i12: ImmIdx = FrameRegIdx + 1; InstrOffs = MI.getOperand(ImmIdx).getImm(); NumBits = 12; break; case ARMII::AddrMode2: ImmIdx = FrameRegIdx+2; InstrOffs = ARM_AM::getAM2Offset(MI.getOperand(ImmIdx).getImm()); if (ARM_AM::getAM2Op(MI.getOperand(ImmIdx).getImm()) == ARM_AM::sub) InstrOffs *= -1; NumBits = 12; break; case ARMII::AddrMode3: ImmIdx = FrameRegIdx+2; InstrOffs = ARM_AM::getAM3Offset(MI.getOperand(ImmIdx).getImm()); if (ARM_AM::getAM3Op(MI.getOperand(ImmIdx).getImm()) == ARM_AM::sub) InstrOffs *= -1; NumBits = 8; break; case ARMII::AddrMode4: case ARMII::AddrMode6: // Can't fold any offset even if it's zero. return false; case ARMII::AddrMode5: ImmIdx = FrameRegIdx+1; InstrOffs = ARM_AM::getAM5Offset(MI.getOperand(ImmIdx).getImm()); if (ARM_AM::getAM5Op(MI.getOperand(ImmIdx).getImm()) == ARM_AM::sub) InstrOffs *= -1; NumBits = 8; Scale = 4; break; case ARMII::AddrMode5FP16: ImmIdx = FrameRegIdx+1; InstrOffs = ARM_AM::getAM5Offset(MI.getOperand(ImmIdx).getImm()); if (ARM_AM::getAM5Op(MI.getOperand(ImmIdx).getImm()) == ARM_AM::sub) InstrOffs *= -1; NumBits = 8; Scale = 2; break; case ARMII::AddrModeT2_i7: case ARMII::AddrModeT2_i7s2: case ARMII::AddrModeT2_i7s4: ImmIdx = FrameRegIdx+1; InstrOffs = MI.getOperand(ImmIdx).getImm(); NumBits = 7; Scale = (AddrMode == ARMII::AddrModeT2_i7s2 ? 2 : AddrMode == ARMII::AddrModeT2_i7s4 ? 4 : 1); break; default: llvm_unreachable("Unsupported addressing mode!"); } Offset += InstrOffs * Scale; assert((Offset & (Scale-1)) == 0 && "Can't encode this offset!"); if (Offset < 0) { Offset = -Offset; isSub = true; } // Attempt to fold address comp. if opcode has offset bits if (NumBits > 0) { // Common case: small offset, fits into instruction. MachineOperand &ImmOp = MI.getOperand(ImmIdx); int ImmedOffset = Offset / Scale; unsigned Mask = (1 << NumBits) - 1; if ((unsigned)Offset <= Mask * Scale) { // Replace the FrameIndex with sp MI.getOperand(FrameRegIdx).ChangeToRegister(FrameReg, false); // FIXME: When addrmode2 goes away, this will simplify (like the // T2 version), as the LDR.i12 versions don't need the encoding // tricks for the offset value. if (isSub) { if (AddrMode == ARMII::AddrMode_i12) ImmedOffset = -ImmedOffset; else ImmedOffset |= 1 << NumBits; } ImmOp.ChangeToImmediate(ImmedOffset); Offset = 0; return true; } // Otherwise, it didn't fit. Pull in what we can to simplify the immed. ImmedOffset = ImmedOffset & Mask; if (isSub) { if (AddrMode == ARMII::AddrMode_i12) ImmedOffset = -ImmedOffset; else ImmedOffset |= 1 << NumBits; } ImmOp.ChangeToImmediate(ImmedOffset); Offset &= ~(Mask*Scale); } } Offset = (isSub) ? -Offset : Offset; return Offset == 0; } /// analyzeCompare - For a comparison instruction, return the source registers /// in SrcReg and SrcReg2 if having two register operands, and the value it /// compares against in CmpValue. Return true if the comparison instruction /// can be analyzed. bool ARMBaseInstrInfo::analyzeCompare(const MachineInstr &MI, Register &SrcReg, Register &SrcReg2, int64_t &CmpMask, int64_t &CmpValue) const { switch (MI.getOpcode()) { default: break; case ARM::CMPri: case ARM::t2CMPri: case ARM::tCMPi8: SrcReg = MI.getOperand(0).getReg(); SrcReg2 = 0; CmpMask = ~0; CmpValue = MI.getOperand(1).getImm(); return true; case ARM::CMPrr: case ARM::t2CMPrr: case ARM::tCMPr: SrcReg = MI.getOperand(0).getReg(); SrcReg2 = MI.getOperand(1).getReg(); CmpMask = ~0; CmpValue = 0; return true; case ARM::TSTri: case ARM::t2TSTri: SrcReg = MI.getOperand(0).getReg(); SrcReg2 = 0; CmpMask = MI.getOperand(1).getImm(); CmpValue = 0; return true; } return false; } /// isSuitableForMask - Identify a suitable 'and' instruction that /// operates on the given source register and applies the same mask /// as a 'tst' instruction. Provide a limited look-through for copies. /// When successful, MI will hold the found instruction. static bool isSuitableForMask(MachineInstr *&MI, Register SrcReg, int CmpMask, bool CommonUse) { switch (MI->getOpcode()) { case ARM::ANDri: case ARM::t2ANDri: if (CmpMask != MI->getOperand(2).getImm()) return false; if (SrcReg == MI->getOperand(CommonUse ? 1 : 0).getReg()) return true; break; } return false; } /// getCmpToAddCondition - assume the flags are set by CMP(a,b), return /// the condition code if we modify the instructions such that flags are /// set by ADD(a,b,X). inline static ARMCC::CondCodes getCmpToAddCondition(ARMCC::CondCodes CC) { switch (CC) { default: return ARMCC::AL; case ARMCC::HS: return ARMCC::LO; case ARMCC::LO: return ARMCC::HS; case ARMCC::VS: return ARMCC::VS; case ARMCC::VC: return ARMCC::VC; } } /// isRedundantFlagInstr - check whether the first instruction, whose only /// purpose is to update flags, can be made redundant. /// CMPrr can be made redundant by SUBrr if the operands are the same. /// CMPri can be made redundant by SUBri if the operands are the same. /// CMPrr(r0, r1) can be made redundant by ADDr[ri](r0, r1, X). /// This function can be extended later on. inline static bool isRedundantFlagInstr(const MachineInstr *CmpI, Register SrcReg, Register SrcReg2, int64_t ImmValue, const MachineInstr *OI, bool &IsThumb1) { if ((CmpI->getOpcode() == ARM::CMPrr || CmpI->getOpcode() == ARM::t2CMPrr) && (OI->getOpcode() == ARM::SUBrr || OI->getOpcode() == ARM::t2SUBrr) && ((OI->getOperand(1).getReg() == SrcReg && OI->getOperand(2).getReg() == SrcReg2) || (OI->getOperand(1).getReg() == SrcReg2 && OI->getOperand(2).getReg() == SrcReg))) { IsThumb1 = false; return true; } if (CmpI->getOpcode() == ARM::tCMPr && OI->getOpcode() == ARM::tSUBrr && ((OI->getOperand(2).getReg() == SrcReg && OI->getOperand(3).getReg() == SrcReg2) || (OI->getOperand(2).getReg() == SrcReg2 && OI->getOperand(3).getReg() == SrcReg))) { IsThumb1 = true; return true; } if ((CmpI->getOpcode() == ARM::CMPri || CmpI->getOpcode() == ARM::t2CMPri) && (OI->getOpcode() == ARM::SUBri || OI->getOpcode() == ARM::t2SUBri) && OI->getOperand(1).getReg() == SrcReg && OI->getOperand(2).getImm() == ImmValue) { IsThumb1 = false; return true; } if (CmpI->getOpcode() == ARM::tCMPi8 && (OI->getOpcode() == ARM::tSUBi8 || OI->getOpcode() == ARM::tSUBi3) && OI->getOperand(2).getReg() == SrcReg && OI->getOperand(3).getImm() == ImmValue) { IsThumb1 = true; return true; } if ((CmpI->getOpcode() == ARM::CMPrr || CmpI->getOpcode() == ARM::t2CMPrr) && (OI->getOpcode() == ARM::ADDrr || OI->getOpcode() == ARM::t2ADDrr || OI->getOpcode() == ARM::ADDri || OI->getOpcode() == ARM::t2ADDri) && OI->getOperand(0).isReg() && OI->getOperand(1).isReg() && OI->getOperand(0).getReg() == SrcReg && OI->getOperand(1).getReg() == SrcReg2) { IsThumb1 = false; return true; } if (CmpI->getOpcode() == ARM::tCMPr && (OI->getOpcode() == ARM::tADDi3 || OI->getOpcode() == ARM::tADDi8 || OI->getOpcode() == ARM::tADDrr) && OI->getOperand(0).getReg() == SrcReg && OI->getOperand(2).getReg() == SrcReg2) { IsThumb1 = true; return true; } return false; } static bool isOptimizeCompareCandidate(MachineInstr *MI, bool &IsThumb1) { switch (MI->getOpcode()) { default: return false; case ARM::tLSLri: case ARM::tLSRri: case ARM::tLSLrr: case ARM::tLSRrr: case ARM::tSUBrr: case ARM::tADDrr: case ARM::tADDi3: case ARM::tADDi8: case ARM::tSUBi3: case ARM::tSUBi8: case ARM::tMUL: case ARM::tADC: case ARM::tSBC: case ARM::tRSB: case ARM::tAND: case ARM::tORR: case ARM::tEOR: case ARM::tBIC: case ARM::tMVN: case ARM::tASRri: case ARM::tASRrr: case ARM::tROR: IsThumb1 = true; [[fallthrough]]; case ARM::RSBrr: case ARM::RSBri: case ARM::RSCrr: case ARM::RSCri: case ARM::ADDrr: case ARM::ADDri: case ARM::ADCrr: case ARM::ADCri: case ARM::SUBrr: case ARM::SUBri: case ARM::SBCrr: case ARM::SBCri: case ARM::t2RSBri: case ARM::t2ADDrr: case ARM::t2ADDri: case ARM::t2ADCrr: case ARM::t2ADCri: case ARM::t2SUBrr: case ARM::t2SUBri: case ARM::t2SBCrr: case ARM::t2SBCri: case ARM::ANDrr: case ARM::ANDri: case ARM::ANDrsr: case ARM::ANDrsi: case ARM::t2ANDrr: case ARM::t2ANDri: case ARM::t2ANDrs: case ARM::ORRrr: case ARM::ORRri: case ARM::ORRrsr: case ARM::ORRrsi: case ARM::t2ORRrr: case ARM::t2ORRri: case ARM::t2ORRrs: case ARM::EORrr: case ARM::EORri: case ARM::EORrsr: case ARM::EORrsi: case ARM::t2EORrr: case ARM::t2EORri: case ARM::t2EORrs: case ARM::BICri: case ARM::BICrr: case ARM::BICrsi: case ARM::BICrsr: case ARM::t2BICri: case ARM::t2BICrr: case ARM::t2BICrs: case ARM::t2LSRri: case ARM::t2LSRrr: case ARM::t2LSLri: case ARM::t2LSLrr: case ARM::MOVsr: case ARM::MOVsi: return true; } } /// optimizeCompareInstr - Convert the instruction supplying the argument to the /// comparison into one that sets the zero bit in the flags register; /// Remove a redundant Compare instruction if an earlier instruction can set the /// flags in the same way as Compare. /// E.g. SUBrr(r1,r2) and CMPrr(r1,r2). We also handle the case where two /// operands are swapped: SUBrr(r1,r2) and CMPrr(r2,r1), by updating the /// condition code of instructions which use the flags. bool ARMBaseInstrInfo::optimizeCompareInstr( MachineInstr &CmpInstr, Register SrcReg, Register SrcReg2, int64_t CmpMask, int64_t CmpValue, const MachineRegisterInfo *MRI) const { // Get the unique definition of SrcReg. MachineInstr *MI = MRI->getUniqueVRegDef(SrcReg); if (!MI) return false; // Masked compares sometimes use the same register as the corresponding 'and'. if (CmpMask != ~0) { if (!isSuitableForMask(MI, SrcReg, CmpMask, false) || isPredicated(*MI)) { MI = nullptr; for (MachineRegisterInfo::use_instr_iterator UI = MRI->use_instr_begin(SrcReg), UE = MRI->use_instr_end(); UI != UE; ++UI) { if (UI->getParent() != CmpInstr.getParent()) continue; MachineInstr *PotentialAND = &*UI; if (!isSuitableForMask(PotentialAND, SrcReg, CmpMask, true) || isPredicated(*PotentialAND)) continue; MI = PotentialAND; break; } if (!MI) return false; } } // Get ready to iterate backward from CmpInstr. MachineBasicBlock::iterator I = CmpInstr, E = MI, B = CmpInstr.getParent()->begin(); // Early exit if CmpInstr is at the beginning of the BB. if (I == B) return false; // There are two possible candidates which can be changed to set CPSR: // One is MI, the other is a SUB or ADD instruction. // For CMPrr(r1,r2), we are looking for SUB(r1,r2), SUB(r2,r1), or // ADDr[ri](r1, r2, X). // For CMPri(r1, CmpValue), we are looking for SUBri(r1, CmpValue). MachineInstr *SubAdd = nullptr; if (SrcReg2 != 0) // MI is not a candidate for CMPrr. MI = nullptr; else if (MI->getParent() != CmpInstr.getParent() || CmpValue != 0) { // Conservatively refuse to convert an instruction which isn't in the same // BB as the comparison. // For CMPri w/ CmpValue != 0, a SubAdd may still be a candidate. // Thus we cannot return here. if (CmpInstr.getOpcode() == ARM::CMPri || CmpInstr.getOpcode() == ARM::t2CMPri || CmpInstr.getOpcode() == ARM::tCMPi8) MI = nullptr; else return false; } bool IsThumb1 = false; if (MI && !isOptimizeCompareCandidate(MI, IsThumb1)) return false; // We also want to do this peephole for cases like this: if (a*b == 0), // and optimise away the CMP instruction from the generated code sequence: // MULS, MOVS, MOVS, CMP. Here the MOVS instructions load the boolean values // resulting from the select instruction, but these MOVS instructions for // Thumb1 (V6M) are flag setting and are thus preventing this optimisation. // However, if we only have MOVS instructions in between the CMP and the // other instruction (the MULS in this example), then the CPSR is dead so we // can safely reorder the sequence into: MOVS, MOVS, MULS, CMP. We do this // reordering and then continue the analysis hoping we can eliminate the // CMP. This peephole works on the vregs, so is still in SSA form. As a // consequence, the movs won't redefine/kill the MUL operands which would // make this reordering illegal. const TargetRegisterInfo *TRI = &getRegisterInfo(); if (MI && IsThumb1) { --I; if (I != E && !MI->readsRegister(ARM::CPSR, TRI)) { bool CanReorder = true; for (; I != E; --I) { if (I->getOpcode() != ARM::tMOVi8) { CanReorder = false; break; } } if (CanReorder) { MI = MI->removeFromParent(); E = CmpInstr; CmpInstr.getParent()->insert(E, MI); } } I = CmpInstr; E = MI; } // Check that CPSR isn't set between the comparison instruction and the one we // want to change. At the same time, search for SubAdd. bool SubAddIsThumb1 = false; do { const MachineInstr &Instr = *--I; // Check whether CmpInstr can be made redundant by the current instruction. if (isRedundantFlagInstr(&CmpInstr, SrcReg, SrcReg2, CmpValue, &Instr, SubAddIsThumb1)) { SubAdd = &*I; break; } // Allow E (which was initially MI) to be SubAdd but do not search before E. if (I == E) break; if (Instr.modifiesRegister(ARM::CPSR, TRI) || Instr.readsRegister(ARM::CPSR, TRI)) // This instruction modifies or uses CPSR after the one we want to // change. We can't do this transformation. return false; if (I == B) { // In some cases, we scan the use-list of an instruction for an AND; // that AND is in the same BB, but may not be scheduled before the // corresponding TST. In that case, bail out. // // FIXME: We could try to reschedule the AND. return false; } } while (true); // Return false if no candidates exist. if (!MI && !SubAdd) return false; // If we found a SubAdd, use it as it will be closer to the CMP if (SubAdd) { MI = SubAdd; IsThumb1 = SubAddIsThumb1; } // We can't use a predicated instruction - it doesn't always write the flags. if (isPredicated(*MI)) return false; // Scan forward for the use of CPSR // When checking against MI: if it's a conditional code that requires // checking of the V bit or C bit, then this is not safe to do. // It is safe to remove CmpInstr if CPSR is redefined or killed. // If we are done with the basic block, we need to check whether CPSR is // live-out. SmallVector, 4> OperandsToUpdate; bool isSafe = false; I = CmpInstr; E = CmpInstr.getParent()->end(); while (!isSafe && ++I != E) { const MachineInstr &Instr = *I; for (unsigned IO = 0, EO = Instr.getNumOperands(); !isSafe && IO != EO; ++IO) { const MachineOperand &MO = Instr.getOperand(IO); if (MO.isRegMask() && MO.clobbersPhysReg(ARM::CPSR)) { isSafe = true; break; } if (!MO.isReg() || MO.getReg() != ARM::CPSR) continue; if (MO.isDef()) { isSafe = true; break; } // Condition code is after the operand before CPSR except for VSELs. ARMCC::CondCodes CC; bool IsInstrVSel = true; switch (Instr.getOpcode()) { default: IsInstrVSel = false; CC = (ARMCC::CondCodes)Instr.getOperand(IO - 1).getImm(); break; case ARM::VSELEQD: case ARM::VSELEQS: case ARM::VSELEQH: CC = ARMCC::EQ; break; case ARM::VSELGTD: case ARM::VSELGTS: case ARM::VSELGTH: CC = ARMCC::GT; break; case ARM::VSELGED: case ARM::VSELGES: case ARM::VSELGEH: CC = ARMCC::GE; break; case ARM::VSELVSD: case ARM::VSELVSS: case ARM::VSELVSH: CC = ARMCC::VS; break; } if (SubAdd) { // If we have SUB(r1, r2) and CMP(r2, r1), the condition code based // on CMP needs to be updated to be based on SUB. // If we have ADD(r1, r2, X) and CMP(r1, r2), the condition code also // needs to be modified. // Push the condition code operands to OperandsToUpdate. // If it is safe to remove CmpInstr, the condition code of these // operands will be modified. unsigned Opc = SubAdd->getOpcode(); bool IsSub = Opc == ARM::SUBrr || Opc == ARM::t2SUBrr || Opc == ARM::SUBri || Opc == ARM::t2SUBri || Opc == ARM::tSUBrr || Opc == ARM::tSUBi3 || Opc == ARM::tSUBi8; unsigned OpI = Opc != ARM::tSUBrr ? 1 : 2; if (!IsSub || (SrcReg2 != 0 && SubAdd->getOperand(OpI).getReg() == SrcReg2 && SubAdd->getOperand(OpI + 1).getReg() == SrcReg)) { // VSel doesn't support condition code update. if (IsInstrVSel) return false; // Ensure we can swap the condition. ARMCC::CondCodes NewCC = (IsSub ? getSwappedCondition(CC) : getCmpToAddCondition(CC)); if (NewCC == ARMCC::AL) return false; OperandsToUpdate.push_back( std::make_pair(&((*I).getOperand(IO - 1)), NewCC)); } } else { // No SubAdd, so this is x = y, z; cmp x, 0. switch (CC) { case ARMCC::EQ: // Z case ARMCC::NE: // Z case ARMCC::MI: // N case ARMCC::PL: // N case ARMCC::AL: // none // CPSR can be used multiple times, we should continue. break; case ARMCC::HS: // C case ARMCC::LO: // C case ARMCC::VS: // V case ARMCC::VC: // V case ARMCC::HI: // C Z case ARMCC::LS: // C Z case ARMCC::GE: // N V case ARMCC::LT: // N V case ARMCC::GT: // Z N V case ARMCC::LE: // Z N V // The instruction uses the V bit or C bit which is not safe. return false; } } } } // If CPSR is not killed nor re-defined, we should check whether it is // live-out. If it is live-out, do not optimize. if (!isSafe) { MachineBasicBlock *MBB = CmpInstr.getParent(); for (MachineBasicBlock *Succ : MBB->successors()) if (Succ->isLiveIn(ARM::CPSR)) return false; } // Toggle the optional operand to CPSR (if it exists - in Thumb1 we always // set CPSR so this is represented as an explicit output) if (!IsThumb1) { unsigned CPSRRegNum = MI->getNumExplicitOperands() - 1; MI->getOperand(CPSRRegNum).setReg(ARM::CPSR); MI->getOperand(CPSRRegNum).setIsDef(true); } assert(!isPredicated(*MI) && "Can't use flags from predicated instruction"); CmpInstr.eraseFromParent(); // Modify the condition code of operands in OperandsToUpdate. // Since we have SUB(r1, r2) and CMP(r2, r1), the condition code needs to // be changed from r2 > r1 to r1 < r2, from r2 < r1 to r1 > r2, etc. for (unsigned i = 0, e = OperandsToUpdate.size(); i < e; i++) OperandsToUpdate[i].first->setImm(OperandsToUpdate[i].second); MI->clearRegisterDeads(ARM::CPSR); return true; } bool ARMBaseInstrInfo::shouldSink(const MachineInstr &MI) const { // Do not sink MI if it might be used to optimize a redundant compare. // We heuristically only look at the instruction immediately following MI to // avoid potentially searching the entire basic block. if (isPredicated(MI)) return true; MachineBasicBlock::const_iterator Next = &MI; ++Next; Register SrcReg, SrcReg2; int64_t CmpMask, CmpValue; bool IsThumb1; if (Next != MI.getParent()->end() && analyzeCompare(*Next, SrcReg, SrcReg2, CmpMask, CmpValue) && isRedundantFlagInstr(&*Next, SrcReg, SrcReg2, CmpValue, &MI, IsThumb1)) return false; return true; } bool ARMBaseInstrInfo::FoldImmediate(MachineInstr &UseMI, MachineInstr &DefMI, Register Reg, MachineRegisterInfo *MRI) const { // Fold large immediates into add, sub, or, xor. unsigned DefOpc = DefMI.getOpcode(); if (DefOpc != ARM::t2MOVi32imm && DefOpc != ARM::MOVi32imm && DefOpc != ARM::tMOVi32imm) return false; if (!DefMI.getOperand(1).isImm()) // Could be t2MOVi32imm @xx return false; if (!MRI->hasOneNonDBGUse(Reg)) return false; const MCInstrDesc &DefMCID = DefMI.getDesc(); if (DefMCID.hasOptionalDef()) { unsigned NumOps = DefMCID.getNumOperands(); const MachineOperand &MO = DefMI.getOperand(NumOps - 1); if (MO.getReg() == ARM::CPSR && !MO.isDead()) // If DefMI defines CPSR and it is not dead, it's obviously not safe // to delete DefMI. return false; } const MCInstrDesc &UseMCID = UseMI.getDesc(); if (UseMCID.hasOptionalDef()) { unsigned NumOps = UseMCID.getNumOperands(); if (UseMI.getOperand(NumOps - 1).getReg() == ARM::CPSR) // If the instruction sets the flag, do not attempt this optimization // since it may change the semantics of the code. return false; } unsigned UseOpc = UseMI.getOpcode(); unsigned NewUseOpc = 0; uint32_t ImmVal = (uint32_t)DefMI.getOperand(1).getImm(); uint32_t SOImmValV1 = 0, SOImmValV2 = 0; bool Commute = false; switch (UseOpc) { default: return false; case ARM::SUBrr: case ARM::ADDrr: case ARM::ORRrr: case ARM::EORrr: case ARM::t2SUBrr: case ARM::t2ADDrr: case ARM::t2ORRrr: case ARM::t2EORrr: { Commute = UseMI.getOperand(2).getReg() != Reg; switch (UseOpc) { default: break; case ARM::ADDrr: case ARM::SUBrr: if (UseOpc == ARM::SUBrr && Commute) return false; // ADD/SUB are special because they're essentially the same operation, so // we can handle a larger range of immediates. if (ARM_AM::isSOImmTwoPartVal(ImmVal)) NewUseOpc = UseOpc == ARM::ADDrr ? ARM::ADDri : ARM::SUBri; else if (ARM_AM::isSOImmTwoPartVal(-ImmVal)) { ImmVal = -ImmVal; NewUseOpc = UseOpc == ARM::ADDrr ? ARM::SUBri : ARM::ADDri; } else return false; SOImmValV1 = (uint32_t)ARM_AM::getSOImmTwoPartFirst(ImmVal); SOImmValV2 = (uint32_t)ARM_AM::getSOImmTwoPartSecond(ImmVal); break; case ARM::ORRrr: case ARM::EORrr: if (!ARM_AM::isSOImmTwoPartVal(ImmVal)) return false; SOImmValV1 = (uint32_t)ARM_AM::getSOImmTwoPartFirst(ImmVal); SOImmValV2 = (uint32_t)ARM_AM::getSOImmTwoPartSecond(ImmVal); switch (UseOpc) { default: break; case ARM::ORRrr: NewUseOpc = ARM::ORRri; break; case ARM::EORrr: NewUseOpc = ARM::EORri; break; } break; case ARM::t2ADDrr: case ARM::t2SUBrr: { if (UseOpc == ARM::t2SUBrr && Commute) return false; // ADD/SUB are special because they're essentially the same operation, so // we can handle a larger range of immediates. const bool ToSP = DefMI.getOperand(0).getReg() == ARM::SP; const unsigned t2ADD = ToSP ? ARM::t2ADDspImm : ARM::t2ADDri; const unsigned t2SUB = ToSP ? ARM::t2SUBspImm : ARM::t2SUBri; if (ARM_AM::isT2SOImmTwoPartVal(ImmVal)) NewUseOpc = UseOpc == ARM::t2ADDrr ? t2ADD : t2SUB; else if (ARM_AM::isT2SOImmTwoPartVal(-ImmVal)) { ImmVal = -ImmVal; NewUseOpc = UseOpc == ARM::t2ADDrr ? t2SUB : t2ADD; } else return false; SOImmValV1 = (uint32_t)ARM_AM::getT2SOImmTwoPartFirst(ImmVal); SOImmValV2 = (uint32_t)ARM_AM::getT2SOImmTwoPartSecond(ImmVal); break; } case ARM::t2ORRrr: case ARM::t2EORrr: if (!ARM_AM::isT2SOImmTwoPartVal(ImmVal)) return false; SOImmValV1 = (uint32_t)ARM_AM::getT2SOImmTwoPartFirst(ImmVal); SOImmValV2 = (uint32_t)ARM_AM::getT2SOImmTwoPartSecond(ImmVal); switch (UseOpc) { default: break; case ARM::t2ORRrr: NewUseOpc = ARM::t2ORRri; break; case ARM::t2EORrr: NewUseOpc = ARM::t2EORri; break; } break; } } } unsigned OpIdx = Commute ? 2 : 1; Register Reg1 = UseMI.getOperand(OpIdx).getReg(); bool isKill = UseMI.getOperand(OpIdx).isKill(); const TargetRegisterClass *TRC = MRI->getRegClass(Reg); Register NewReg = MRI->createVirtualRegister(TRC); BuildMI(*UseMI.getParent(), UseMI, UseMI.getDebugLoc(), get(NewUseOpc), NewReg) .addReg(Reg1, getKillRegState(isKill)) .addImm(SOImmValV1) .add(predOps(ARMCC::AL)) .add(condCodeOp()); UseMI.setDesc(get(NewUseOpc)); UseMI.getOperand(1).setReg(NewReg); UseMI.getOperand(1).setIsKill(); UseMI.getOperand(2).ChangeToImmediate(SOImmValV2); DefMI.eraseFromParent(); // FIXME: t2ADDrr should be split, as different rulles apply when writing to SP. // Just as t2ADDri, that was split to [t2ADDri, t2ADDspImm]. // Then the below code will not be needed, as the input/output register // classes will be rgpr or gprSP. // For now, we fix the UseMI operand explicitly here: switch(NewUseOpc){ case ARM::t2ADDspImm: case ARM::t2SUBspImm: case ARM::t2ADDri: case ARM::t2SUBri: MRI->constrainRegClass(UseMI.getOperand(0).getReg(), TRC); } return true; } static unsigned getNumMicroOpsSwiftLdSt(const InstrItineraryData *ItinData, const MachineInstr &MI) { switch (MI.getOpcode()) { default: { const MCInstrDesc &Desc = MI.getDesc(); int UOps = ItinData->getNumMicroOps(Desc.getSchedClass()); assert(UOps >= 0 && "bad # UOps"); return UOps; } case ARM::LDRrs: case ARM::LDRBrs: case ARM::STRrs: case ARM::STRBrs: { unsigned ShOpVal = MI.getOperand(3).getImm(); bool isSub = ARM_AM::getAM2Op(ShOpVal) == ARM_AM::sub; unsigned ShImm = ARM_AM::getAM2Offset(ShOpVal); if (!isSub && (ShImm == 0 || ((ShImm == 1 || ShImm == 2 || ShImm == 3) && ARM_AM::getAM2ShiftOpc(ShOpVal) == ARM_AM::lsl))) return 1; return 2; } case ARM::LDRH: case ARM::STRH: { if (!MI.getOperand(2).getReg()) return 1; unsigned ShOpVal = MI.getOperand(3).getImm(); bool isSub = ARM_AM::getAM2Op(ShOpVal) == ARM_AM::sub; unsigned ShImm = ARM_AM::getAM2Offset(ShOpVal); if (!isSub && (ShImm == 0 || ((ShImm == 1 || ShImm == 2 || ShImm == 3) && ARM_AM::getAM2ShiftOpc(ShOpVal) == ARM_AM::lsl))) return 1; return 2; } case ARM::LDRSB: case ARM::LDRSH: return (ARM_AM::getAM3Op(MI.getOperand(3).getImm()) == ARM_AM::sub) ? 3 : 2; case ARM::LDRSB_POST: case ARM::LDRSH_POST: { Register Rt = MI.getOperand(0).getReg(); Register Rm = MI.getOperand(3).getReg(); return (Rt == Rm) ? 4 : 3; } case ARM::LDR_PRE_REG: case ARM::LDRB_PRE_REG: { Register Rt = MI.getOperand(0).getReg(); Register Rm = MI.getOperand(3).getReg(); if (Rt == Rm) return 3; unsigned ShOpVal = MI.getOperand(4).getImm(); bool isSub = ARM_AM::getAM2Op(ShOpVal) == ARM_AM::sub; unsigned ShImm = ARM_AM::getAM2Offset(ShOpVal); if (!isSub && (ShImm == 0 || ((ShImm == 1 || ShImm == 2 || ShImm == 3) && ARM_AM::getAM2ShiftOpc(ShOpVal) == ARM_AM::lsl))) return 2; return 3; } case ARM::STR_PRE_REG: case ARM::STRB_PRE_REG: { unsigned ShOpVal = MI.getOperand(4).getImm(); bool isSub = ARM_AM::getAM2Op(ShOpVal) == ARM_AM::sub; unsigned ShImm = ARM_AM::getAM2Offset(ShOpVal); if (!isSub && (ShImm == 0 || ((ShImm == 1 || ShImm == 2 || ShImm == 3) && ARM_AM::getAM2ShiftOpc(ShOpVal) == ARM_AM::lsl))) return 2; return 3; } case ARM::LDRH_PRE: case ARM::STRH_PRE: { Register Rt = MI.getOperand(0).getReg(); Register Rm = MI.getOperand(3).getReg(); if (!Rm) return 2; if (Rt == Rm) return 3; return (ARM_AM::getAM3Op(MI.getOperand(4).getImm()) == ARM_AM::sub) ? 3 : 2; } case ARM::LDR_POST_REG: case ARM::LDRB_POST_REG: case ARM::LDRH_POST: { Register Rt = MI.getOperand(0).getReg(); Register Rm = MI.getOperand(3).getReg(); return (Rt == Rm) ? 3 : 2; } case ARM::LDR_PRE_IMM: case ARM::LDRB_PRE_IMM: case ARM::LDR_POST_IMM: case ARM::LDRB_POST_IMM: case ARM::STRB_POST_IMM: case ARM::STRB_POST_REG: case ARM::STRB_PRE_IMM: case ARM::STRH_POST: case ARM::STR_POST_IMM: case ARM::STR_POST_REG: case ARM::STR_PRE_IMM: return 2; case ARM::LDRSB_PRE: case ARM::LDRSH_PRE: { Register Rm = MI.getOperand(3).getReg(); if (Rm == 0) return 3; Register Rt = MI.getOperand(0).getReg(); if (Rt == Rm) return 4; unsigned ShOpVal = MI.getOperand(4).getImm(); bool isSub = ARM_AM::getAM2Op(ShOpVal) == ARM_AM::sub; unsigned ShImm = ARM_AM::getAM2Offset(ShOpVal); if (!isSub && (ShImm == 0 || ((ShImm == 1 || ShImm == 2 || ShImm == 3) && ARM_AM::getAM2ShiftOpc(ShOpVal) == ARM_AM::lsl))) return 3; return 4; } case ARM::LDRD: { Register Rt = MI.getOperand(0).getReg(); Register Rn = MI.getOperand(2).getReg(); Register Rm = MI.getOperand(3).getReg(); if (Rm) return (ARM_AM::getAM3Op(MI.getOperand(4).getImm()) == ARM_AM::sub) ? 4 : 3; return (Rt == Rn) ? 3 : 2; } case ARM::STRD: { Register Rm = MI.getOperand(3).getReg(); if (Rm) return (ARM_AM::getAM3Op(MI.getOperand(4).getImm()) == ARM_AM::sub) ? 4 : 3; return 2; } case ARM::LDRD_POST: case ARM::t2LDRD_POST: return 3; case ARM::STRD_POST: case ARM::t2STRD_POST: return 4; case ARM::LDRD_PRE: { Register Rt = MI.getOperand(0).getReg(); Register Rn = MI.getOperand(3).getReg(); Register Rm = MI.getOperand(4).getReg(); if (Rm) return (ARM_AM::getAM3Op(MI.getOperand(5).getImm()) == ARM_AM::sub) ? 5 : 4; return (Rt == Rn) ? 4 : 3; } case ARM::t2LDRD_PRE: { Register Rt = MI.getOperand(0).getReg(); Register Rn = MI.getOperand(3).getReg(); return (Rt == Rn) ? 4 : 3; } case ARM::STRD_PRE: { Register Rm = MI.getOperand(4).getReg(); if (Rm) return (ARM_AM::getAM3Op(MI.getOperand(5).getImm()) == ARM_AM::sub) ? 5 : 4; return 3; } case ARM::t2STRD_PRE: return 3; case ARM::t2LDR_POST: case ARM::t2LDRB_POST: case ARM::t2LDRB_PRE: case ARM::t2LDRSBi12: case ARM::t2LDRSBi8: case ARM::t2LDRSBpci: case ARM::t2LDRSBs: case ARM::t2LDRH_POST: case ARM::t2LDRH_PRE: case ARM::t2LDRSBT: case ARM::t2LDRSB_POST: case ARM::t2LDRSB_PRE: case ARM::t2LDRSH_POST: case ARM::t2LDRSH_PRE: case ARM::t2LDRSHi12: case ARM::t2LDRSHi8: case ARM::t2LDRSHpci: case ARM::t2LDRSHs: return 2; case ARM::t2LDRDi8: { Register Rt = MI.getOperand(0).getReg(); Register Rn = MI.getOperand(2).getReg(); return (Rt == Rn) ? 3 : 2; } case ARM::t2STRB_POST: case ARM::t2STRB_PRE: case ARM::t2STRBs: case ARM::t2STRDi8: case ARM::t2STRH_POST: case ARM::t2STRH_PRE: case ARM::t2STRHs: case ARM::t2STR_POST: case ARM::t2STR_PRE: case ARM::t2STRs: return 2; } } // Return the number of 32-bit words loaded by LDM or stored by STM. If this // can't be easily determined return 0 (missing MachineMemOperand). // // FIXME: The current MachineInstr design does not support relying on machine // mem operands to determine the width of a memory access. Instead, we expect // the target to provide this information based on the instruction opcode and // operands. However, using MachineMemOperand is the best solution now for // two reasons: // // 1) getNumMicroOps tries to infer LDM memory width from the total number of MI // operands. This is much more dangerous than using the MachineMemOperand // sizes because CodeGen passes can insert/remove optional machine operands. In // fact, it's totally incorrect for preRA passes and appears to be wrong for // postRA passes as well. // // 2) getNumLDMAddresses is only used by the scheduling machine model and any // machine model that calls this should handle the unknown (zero size) case. // // Long term, we should require a target hook that verifies MachineMemOperand // sizes during MC lowering. That target hook should be local to MC lowering // because we can't ensure that it is aware of other MI forms. Doing this will // ensure that MachineMemOperands are correctly propagated through all passes. unsigned ARMBaseInstrInfo::getNumLDMAddresses(const MachineInstr &MI) const { unsigned Size = 0; for (MachineInstr::mmo_iterator I = MI.memoperands_begin(), E = MI.memoperands_end(); I != E; ++I) { Size += (*I)->getSize(); } // FIXME: The scheduler currently can't handle values larger than 16. But // the values can actually go up to 32 for floating-point load/store // multiple (VLDMIA etc.). Also, the way this code is reasoning about memory // operations isn't right; we could end up with "extra" memory operands for // various reasons, like tail merge merging two memory operations. return std::min(Size / 4, 16U); } static unsigned getNumMicroOpsSingleIssuePlusExtras(unsigned Opc, unsigned NumRegs) { unsigned UOps = 1 + NumRegs; // 1 for address computation. switch (Opc) { default: break; case ARM::VLDMDIA_UPD: case ARM::VLDMDDB_UPD: case ARM::VLDMSIA_UPD: case ARM::VLDMSDB_UPD: case ARM::VSTMDIA_UPD: case ARM::VSTMDDB_UPD: case ARM::VSTMSIA_UPD: case ARM::VSTMSDB_UPD: case ARM::LDMIA_UPD: case ARM::LDMDA_UPD: case ARM::LDMDB_UPD: case ARM::LDMIB_UPD: case ARM::STMIA_UPD: case ARM::STMDA_UPD: case ARM::STMDB_UPD: case ARM::STMIB_UPD: case ARM::tLDMIA_UPD: case ARM::tSTMIA_UPD: case ARM::t2LDMIA_UPD: case ARM::t2LDMDB_UPD: case ARM::t2STMIA_UPD: case ARM::t2STMDB_UPD: ++UOps; // One for base register writeback. break; case ARM::LDMIA_RET: case ARM::tPOP_RET: case ARM::t2LDMIA_RET: UOps += 2; // One for base reg wb, one for write to pc. break; } return UOps; } unsigned ARMBaseInstrInfo::getNumMicroOps(const InstrItineraryData *ItinData, const MachineInstr &MI) const { if (!ItinData || ItinData->isEmpty()) return 1; const MCInstrDesc &Desc = MI.getDesc(); unsigned Class = Desc.getSchedClass(); int ItinUOps = ItinData->getNumMicroOps(Class); if (ItinUOps >= 0) { if (Subtarget.isSwift() && (Desc.mayLoad() || Desc.mayStore())) return getNumMicroOpsSwiftLdSt(ItinData, MI); return ItinUOps; } unsigned Opc = MI.getOpcode(); switch (Opc) { default: llvm_unreachable("Unexpected multi-uops instruction!"); case ARM::VLDMQIA: case ARM::VSTMQIA: return 2; // The number of uOps for load / store multiple are determined by the number // registers. // // On Cortex-A8, each pair of register loads / stores can be scheduled on the // same cycle. The scheduling for the first load / store must be done // separately by assuming the address is not 64-bit aligned. // // On Cortex-A9, the formula is simply (#reg / 2) + (#reg % 2). If the address // is not 64-bit aligned, then AGU would take an extra cycle. For VFP / NEON // load / store multiple, the formula is (#reg / 2) + (#reg % 2) + 1. case ARM::VLDMDIA: case ARM::VLDMDIA_UPD: case ARM::VLDMDDB_UPD: case ARM::VLDMSIA: case ARM::VLDMSIA_UPD: case ARM::VLDMSDB_UPD: case ARM::VSTMDIA: case ARM::VSTMDIA_UPD: case ARM::VSTMDDB_UPD: case ARM::VSTMSIA: case ARM::VSTMSIA_UPD: case ARM::VSTMSDB_UPD: { unsigned NumRegs = MI.getNumOperands() - Desc.getNumOperands(); return (NumRegs / 2) + (NumRegs % 2) + 1; } case ARM::LDMIA_RET: case ARM::LDMIA: case ARM::LDMDA: case ARM::LDMDB: case ARM::LDMIB: case ARM::LDMIA_UPD: case ARM::LDMDA_UPD: case ARM::LDMDB_UPD: case ARM::LDMIB_UPD: case ARM::STMIA: case ARM::STMDA: case ARM::STMDB: case ARM::STMIB: case ARM::STMIA_UPD: case ARM::STMDA_UPD: case ARM::STMDB_UPD: case ARM::STMIB_UPD: case ARM::tLDMIA: case ARM::tLDMIA_UPD: case ARM::tSTMIA_UPD: case ARM::tPOP_RET: case ARM::tPOP: case ARM::tPUSH: case ARM::t2LDMIA_RET: case ARM::t2LDMIA: case ARM::t2LDMDB: case ARM::t2LDMIA_UPD: case ARM::t2LDMDB_UPD: case ARM::t2STMIA: case ARM::t2STMDB: case ARM::t2STMIA_UPD: case ARM::t2STMDB_UPD: { unsigned NumRegs = MI.getNumOperands() - Desc.getNumOperands() + 1; switch (Subtarget.getLdStMultipleTiming()) { case ARMSubtarget::SingleIssuePlusExtras: return getNumMicroOpsSingleIssuePlusExtras(Opc, NumRegs); case ARMSubtarget::SingleIssue: // Assume the worst. return NumRegs; case ARMSubtarget::DoubleIssue: { if (NumRegs < 4) return 2; // 4 registers would be issued: 2, 2. // 5 registers would be issued: 2, 2, 1. unsigned UOps = (NumRegs / 2); if (NumRegs % 2) ++UOps; return UOps; } case ARMSubtarget::DoubleIssueCheckUnalignedAccess: { unsigned UOps = (NumRegs / 2); // If there are odd number of registers or if it's not 64-bit aligned, // then it takes an extra AGU (Address Generation Unit) cycle. if ((NumRegs % 2) || !MI.hasOneMemOperand() || (*MI.memoperands_begin())->getAlign() < Align(8)) ++UOps; return UOps; } } } } llvm_unreachable("Didn't find the number of microops"); } int ARMBaseInstrInfo::getVLDMDefCycle(const InstrItineraryData *ItinData, const MCInstrDesc &DefMCID, unsigned DefClass, unsigned DefIdx, unsigned DefAlign) const { int RegNo = (int)(DefIdx+1) - DefMCID.getNumOperands() + 1; if (RegNo <= 0) // Def is the address writeback. return ItinData->getOperandCycle(DefClass, DefIdx); int DefCycle; if (Subtarget.isCortexA8() || Subtarget.isCortexA7()) { // (regno / 2) + (regno % 2) + 1 DefCycle = RegNo / 2 + 1; if (RegNo % 2) ++DefCycle; } else if (Subtarget.isLikeA9() || Subtarget.isSwift()) { DefCycle = RegNo; bool isSLoad = false; switch (DefMCID.getOpcode()) { default: break; case ARM::VLDMSIA: case ARM::VLDMSIA_UPD: case ARM::VLDMSDB_UPD: isSLoad = true; break; } // If there are odd number of 'S' registers or if it's not 64-bit aligned, // then it takes an extra cycle. if ((isSLoad && (RegNo % 2)) || DefAlign < 8) ++DefCycle; } else { // Assume the worst. DefCycle = RegNo + 2; } return DefCycle; } int ARMBaseInstrInfo::getLDMDefCycle(const InstrItineraryData *ItinData, const MCInstrDesc &DefMCID, unsigned DefClass, unsigned DefIdx, unsigned DefAlign) const { int RegNo = (int)(DefIdx+1) - DefMCID.getNumOperands() + 1; if (RegNo <= 0) // Def is the address writeback. return ItinData->getOperandCycle(DefClass, DefIdx); int DefCycle; if (Subtarget.isCortexA8() || Subtarget.isCortexA7()) { // 4 registers would be issued: 1, 2, 1. // 5 registers would be issued: 1, 2, 2. DefCycle = RegNo / 2; if (DefCycle < 1) DefCycle = 1; // Result latency is issue cycle + 2: E2. DefCycle += 2; } else if (Subtarget.isLikeA9() || Subtarget.isSwift()) { DefCycle = (RegNo / 2); // If there are odd number of registers or if it's not 64-bit aligned, // then it takes an extra AGU (Address Generation Unit) cycle. if ((RegNo % 2) || DefAlign < 8) ++DefCycle; // Result latency is AGU cycles + 2. DefCycle += 2; } else { // Assume the worst. DefCycle = RegNo + 2; } return DefCycle; } int ARMBaseInstrInfo::getVSTMUseCycle(const InstrItineraryData *ItinData, const MCInstrDesc &UseMCID, unsigned UseClass, unsigned UseIdx, unsigned UseAlign) const { int RegNo = (int)(UseIdx+1) - UseMCID.getNumOperands() + 1; if (RegNo <= 0) return ItinData->getOperandCycle(UseClass, UseIdx); int UseCycle; if (Subtarget.isCortexA8() || Subtarget.isCortexA7()) { // (regno / 2) + (regno % 2) + 1 UseCycle = RegNo / 2 + 1; if (RegNo % 2) ++UseCycle; } else if (Subtarget.isLikeA9() || Subtarget.isSwift()) { UseCycle = RegNo; bool isSStore = false; switch (UseMCID.getOpcode()) { default: break; case ARM::VSTMSIA: case ARM::VSTMSIA_UPD: case ARM::VSTMSDB_UPD: isSStore = true; break; } // If there are odd number of 'S' registers or if it's not 64-bit aligned, // then it takes an extra cycle. if ((isSStore && (RegNo % 2)) || UseAlign < 8) ++UseCycle; } else { // Assume the worst. UseCycle = RegNo + 2; } return UseCycle; } int ARMBaseInstrInfo::getSTMUseCycle(const InstrItineraryData *ItinData, const MCInstrDesc &UseMCID, unsigned UseClass, unsigned UseIdx, unsigned UseAlign) const { int RegNo = (int)(UseIdx+1) - UseMCID.getNumOperands() + 1; if (RegNo <= 0) return ItinData->getOperandCycle(UseClass, UseIdx); int UseCycle; if (Subtarget.isCortexA8() || Subtarget.isCortexA7()) { UseCycle = RegNo / 2; if (UseCycle < 2) UseCycle = 2; // Read in E3. UseCycle += 2; } else if (Subtarget.isLikeA9() || Subtarget.isSwift()) { UseCycle = (RegNo / 2); // If there are odd number of registers or if it's not 64-bit aligned, // then it takes an extra AGU (Address Generation Unit) cycle. if ((RegNo % 2) || UseAlign < 8) ++UseCycle; } else { // Assume the worst. UseCycle = 1; } return UseCycle; } int ARMBaseInstrInfo::getOperandLatency(const InstrItineraryData *ItinData, const MCInstrDesc &DefMCID, unsigned DefIdx, unsigned DefAlign, const MCInstrDesc &UseMCID, unsigned UseIdx, unsigned UseAlign) const { unsigned DefClass = DefMCID.getSchedClass(); unsigned UseClass = UseMCID.getSchedClass(); if (DefIdx < DefMCID.getNumDefs() && UseIdx < UseMCID.getNumOperands()) return ItinData->getOperandLatency(DefClass, DefIdx, UseClass, UseIdx); // This may be a def / use of a variable_ops instruction, the operand // latency might be determinable dynamically. Let the target try to // figure it out. int DefCycle = -1; bool LdmBypass = false; switch (DefMCID.getOpcode()) { default: DefCycle = ItinData->getOperandCycle(DefClass, DefIdx); break; case ARM::VLDMDIA: case ARM::VLDMDIA_UPD: case ARM::VLDMDDB_UPD: case ARM::VLDMSIA: case ARM::VLDMSIA_UPD: case ARM::VLDMSDB_UPD: DefCycle = getVLDMDefCycle(ItinData, DefMCID, DefClass, DefIdx, DefAlign); break; case ARM::LDMIA_RET: case ARM::LDMIA: case ARM::LDMDA: case ARM::LDMDB: case ARM::LDMIB: case ARM::LDMIA_UPD: case ARM::LDMDA_UPD: case ARM::LDMDB_UPD: case ARM::LDMIB_UPD: case ARM::tLDMIA: case ARM::tLDMIA_UPD: case ARM::tPUSH: case ARM::t2LDMIA_RET: case ARM::t2LDMIA: case ARM::t2LDMDB: case ARM::t2LDMIA_UPD: case ARM::t2LDMDB_UPD: LdmBypass = true; DefCycle = getLDMDefCycle(ItinData, DefMCID, DefClass, DefIdx, DefAlign); break; } if (DefCycle == -1) // We can't seem to determine the result latency of the def, assume it's 2. DefCycle = 2; int UseCycle = -1; switch (UseMCID.getOpcode()) { default: UseCycle = ItinData->getOperandCycle(UseClass, UseIdx); break; case ARM::VSTMDIA: case ARM::VSTMDIA_UPD: case ARM::VSTMDDB_UPD: case ARM::VSTMSIA: case ARM::VSTMSIA_UPD: case ARM::VSTMSDB_UPD: UseCycle = getVSTMUseCycle(ItinData, UseMCID, UseClass, UseIdx, UseAlign); break; case ARM::STMIA: case ARM::STMDA: case ARM::STMDB: case ARM::STMIB: case ARM::STMIA_UPD: case ARM::STMDA_UPD: case ARM::STMDB_UPD: case ARM::STMIB_UPD: case ARM::tSTMIA_UPD: case ARM::tPOP_RET: case ARM::tPOP: case ARM::t2STMIA: case ARM::t2STMDB: case ARM::t2STMIA_UPD: case ARM::t2STMDB_UPD: UseCycle = getSTMUseCycle(ItinData, UseMCID, UseClass, UseIdx, UseAlign); break; } if (UseCycle == -1) // Assume it's read in the first stage. UseCycle = 1; UseCycle = DefCycle - UseCycle + 1; if (UseCycle > 0) { if (LdmBypass) { // It's a variable_ops instruction so we can't use DefIdx here. Just use // first def operand. if (ItinData->hasPipelineForwarding(DefClass, DefMCID.getNumOperands()-1, UseClass, UseIdx)) --UseCycle; } else if (ItinData->hasPipelineForwarding(DefClass, DefIdx, UseClass, UseIdx)) { --UseCycle; } } return UseCycle; } static const MachineInstr *getBundledDefMI(const TargetRegisterInfo *TRI, const MachineInstr *MI, unsigned Reg, unsigned &DefIdx, unsigned &Dist) { Dist = 0; MachineBasicBlock::const_iterator I = MI; ++I; MachineBasicBlock::const_instr_iterator II = std::prev(I.getInstrIterator()); assert(II->isInsideBundle() && "Empty bundle?"); int Idx = -1; while (II->isInsideBundle()) { Idx = II->findRegisterDefOperandIdx(Reg, false, true, TRI); if (Idx != -1) break; --II; ++Dist; } assert(Idx != -1 && "Cannot find bundled definition!"); DefIdx = Idx; return &*II; } static const MachineInstr *getBundledUseMI(const TargetRegisterInfo *TRI, const MachineInstr &MI, unsigned Reg, unsigned &UseIdx, unsigned &Dist) { Dist = 0; MachineBasicBlock::const_instr_iterator II = ++MI.getIterator(); assert(II->isInsideBundle() && "Empty bundle?"); MachineBasicBlock::const_instr_iterator E = MI.getParent()->instr_end(); // FIXME: This doesn't properly handle multiple uses. int Idx = -1; while (II != E && II->isInsideBundle()) { Idx = II->findRegisterUseOperandIdx(Reg, false, TRI); if (Idx != -1) break; if (II->getOpcode() != ARM::t2IT) ++Dist; ++II; } if (Idx == -1) { Dist = 0; return nullptr; } UseIdx = Idx; return &*II; } /// Return the number of cycles to add to (or subtract from) the static /// itinerary based on the def opcode and alignment. The caller will ensure that /// adjusted latency is at least one cycle. static int adjustDefLatency(const ARMSubtarget &Subtarget, const MachineInstr &DefMI, const MCInstrDesc &DefMCID, unsigned DefAlign) { int Adjust = 0; if (Subtarget.isCortexA8() || Subtarget.isLikeA9() || Subtarget.isCortexA7()) { // FIXME: Shifter op hack: no shift (i.e. [r +/- r]) or [r + r << 2] // variants are one cycle cheaper. switch (DefMCID.getOpcode()) { default: break; case ARM::LDRrs: case ARM::LDRBrs: { unsigned ShOpVal = DefMI.getOperand(3).getImm(); unsigned ShImm = ARM_AM::getAM2Offset(ShOpVal); if (ShImm == 0 || (ShImm == 2 && ARM_AM::getAM2ShiftOpc(ShOpVal) == ARM_AM::lsl)) --Adjust; break; } case ARM::t2LDRs: case ARM::t2LDRBs: case ARM::t2LDRHs: case ARM::t2LDRSHs: { // Thumb2 mode: lsl only. unsigned ShAmt = DefMI.getOperand(3).getImm(); if (ShAmt == 0 || ShAmt == 2) --Adjust; break; } } } else if (Subtarget.isSwift()) { // FIXME: Properly handle all of the latency adjustments for address // writeback. switch (DefMCID.getOpcode()) { default: break; case ARM::LDRrs: case ARM::LDRBrs: { unsigned ShOpVal = DefMI.getOperand(3).getImm(); bool isSub = ARM_AM::getAM2Op(ShOpVal) == ARM_AM::sub; unsigned ShImm = ARM_AM::getAM2Offset(ShOpVal); if (!isSub && (ShImm == 0 || ((ShImm == 1 || ShImm == 2 || ShImm == 3) && ARM_AM::getAM2ShiftOpc(ShOpVal) == ARM_AM::lsl))) Adjust -= 2; else if (!isSub && ShImm == 1 && ARM_AM::getAM2ShiftOpc(ShOpVal) == ARM_AM::lsr) --Adjust; break; } case ARM::t2LDRs: case ARM::t2LDRBs: case ARM::t2LDRHs: case ARM::t2LDRSHs: { // Thumb2 mode: lsl only. unsigned ShAmt = DefMI.getOperand(3).getImm(); if (ShAmt == 0 || ShAmt == 1 || ShAmt == 2 || ShAmt == 3) Adjust -= 2; break; } } } if (DefAlign < 8 && Subtarget.checkVLDnAccessAlignment()) { switch (DefMCID.getOpcode()) { default: break; case ARM::VLD1q8: case ARM::VLD1q16: case ARM::VLD1q32: case ARM::VLD1q64: case ARM::VLD1q8wb_fixed: case ARM::VLD1q16wb_fixed: case ARM::VLD1q32wb_fixed: case ARM::VLD1q64wb_fixed: case ARM::VLD1q8wb_register: case ARM::VLD1q16wb_register: case ARM::VLD1q32wb_register: case ARM::VLD1q64wb_register: case ARM::VLD2d8: case ARM::VLD2d16: case ARM::VLD2d32: case ARM::VLD2q8: case ARM::VLD2q16: case ARM::VLD2q32: case ARM::VLD2d8wb_fixed: case ARM::VLD2d16wb_fixed: case ARM::VLD2d32wb_fixed: case ARM::VLD2q8wb_fixed: case ARM::VLD2q16wb_fixed: case ARM::VLD2q32wb_fixed: case ARM::VLD2d8wb_register: case ARM::VLD2d16wb_register: case ARM::VLD2d32wb_register: case ARM::VLD2q8wb_register: case ARM::VLD2q16wb_register: case ARM::VLD2q32wb_register: case ARM::VLD3d8: case ARM::VLD3d16: case ARM::VLD3d32: case ARM::VLD1d64T: case ARM::VLD3d8_UPD: case ARM::VLD3d16_UPD: case ARM::VLD3d32_UPD: case ARM::VLD1d64Twb_fixed: case ARM::VLD1d64Twb_register: case ARM::VLD3q8_UPD: case ARM::VLD3q16_UPD: case ARM::VLD3q32_UPD: case ARM::VLD4d8: case ARM::VLD4d16: case ARM::VLD4d32: case ARM::VLD1d64Q: case ARM::VLD4d8_UPD: case ARM::VLD4d16_UPD: case ARM::VLD4d32_UPD: case ARM::VLD1d64Qwb_fixed: case ARM::VLD1d64Qwb_register: case ARM::VLD4q8_UPD: case ARM::VLD4q16_UPD: case ARM::VLD4q32_UPD: case ARM::VLD1DUPq8: case ARM::VLD1DUPq16: case ARM::VLD1DUPq32: case ARM::VLD1DUPq8wb_fixed: case ARM::VLD1DUPq16wb_fixed: case ARM::VLD1DUPq32wb_fixed: case ARM::VLD1DUPq8wb_register: case ARM::VLD1DUPq16wb_register: case ARM::VLD1DUPq32wb_register: case ARM::VLD2DUPd8: case ARM::VLD2DUPd16: case ARM::VLD2DUPd32: case ARM::VLD2DUPd8wb_fixed: case ARM::VLD2DUPd16wb_fixed: case ARM::VLD2DUPd32wb_fixed: case ARM::VLD2DUPd8wb_register: case ARM::VLD2DUPd16wb_register: case ARM::VLD2DUPd32wb_register: case ARM::VLD4DUPd8: case ARM::VLD4DUPd16: case ARM::VLD4DUPd32: case ARM::VLD4DUPd8_UPD: case ARM::VLD4DUPd16_UPD: case ARM::VLD4DUPd32_UPD: case ARM::VLD1LNd8: case ARM::VLD1LNd16: case ARM::VLD1LNd32: case ARM::VLD1LNd8_UPD: case ARM::VLD1LNd16_UPD: case ARM::VLD1LNd32_UPD: case ARM::VLD2LNd8: case ARM::VLD2LNd16: case ARM::VLD2LNd32: case ARM::VLD2LNq16: case ARM::VLD2LNq32: case ARM::VLD2LNd8_UPD: case ARM::VLD2LNd16_UPD: case ARM::VLD2LNd32_UPD: case ARM::VLD2LNq16_UPD: case ARM::VLD2LNq32_UPD: case ARM::VLD4LNd8: case ARM::VLD4LNd16: case ARM::VLD4LNd32: case ARM::VLD4LNq16: case ARM::VLD4LNq32: case ARM::VLD4LNd8_UPD: case ARM::VLD4LNd16_UPD: case ARM::VLD4LNd32_UPD: case ARM::VLD4LNq16_UPD: case ARM::VLD4LNq32_UPD: // If the address is not 64-bit aligned, the latencies of these // instructions increases by one. ++Adjust; break; } } return Adjust; } int ARMBaseInstrInfo::getOperandLatency(const InstrItineraryData *ItinData, const MachineInstr &DefMI, unsigned DefIdx, const MachineInstr &UseMI, unsigned UseIdx) const { // No operand latency. The caller may fall back to getInstrLatency. if (!ItinData || ItinData->isEmpty()) return -1; const MachineOperand &DefMO = DefMI.getOperand(DefIdx); Register Reg = DefMO.getReg(); const MachineInstr *ResolvedDefMI = &DefMI; unsigned DefAdj = 0; if (DefMI.isBundle()) ResolvedDefMI = getBundledDefMI(&getRegisterInfo(), &DefMI, Reg, DefIdx, DefAdj); if (ResolvedDefMI->isCopyLike() || ResolvedDefMI->isInsertSubreg() || ResolvedDefMI->isRegSequence() || ResolvedDefMI->isImplicitDef()) { return 1; } const MachineInstr *ResolvedUseMI = &UseMI; unsigned UseAdj = 0; if (UseMI.isBundle()) { ResolvedUseMI = getBundledUseMI(&getRegisterInfo(), UseMI, Reg, UseIdx, UseAdj); if (!ResolvedUseMI) return -1; } return getOperandLatencyImpl( ItinData, *ResolvedDefMI, DefIdx, ResolvedDefMI->getDesc(), DefAdj, DefMO, Reg, *ResolvedUseMI, UseIdx, ResolvedUseMI->getDesc(), UseAdj); } int ARMBaseInstrInfo::getOperandLatencyImpl( const InstrItineraryData *ItinData, const MachineInstr &DefMI, unsigned DefIdx, const MCInstrDesc &DefMCID, unsigned DefAdj, const MachineOperand &DefMO, unsigned Reg, const MachineInstr &UseMI, unsigned UseIdx, const MCInstrDesc &UseMCID, unsigned UseAdj) const { if (Reg == ARM::CPSR) { if (DefMI.getOpcode() == ARM::FMSTAT) { // fpscr -> cpsr stalls over 20 cycles on A8 (and earlier?) return Subtarget.isLikeA9() ? 1 : 20; } // CPSR set and branch can be paired in the same cycle. if (UseMI.isBranch()) return 0; // Otherwise it takes the instruction latency (generally one). unsigned Latency = getInstrLatency(ItinData, DefMI); // For Thumb2 and -Os, prefer scheduling CPSR setting instruction close to // its uses. Instructions which are otherwise scheduled between them may // incur a code size penalty (not able to use the CPSR setting 16-bit // instructions). if (Latency > 0 && Subtarget.isThumb2()) { const MachineFunction *MF = DefMI.getParent()->getParent(); // FIXME: Use Function::hasOptSize(). if (MF->getFunction().hasFnAttribute(Attribute::OptimizeForSize)) --Latency; } return Latency; } if (DefMO.isImplicit() || UseMI.getOperand(UseIdx).isImplicit()) return -1; unsigned DefAlign = DefMI.hasOneMemOperand() ? (*DefMI.memoperands_begin())->getAlign().value() : 0; unsigned UseAlign = UseMI.hasOneMemOperand() ? (*UseMI.memoperands_begin())->getAlign().value() : 0; // Get the itinerary's latency if possible, and handle variable_ops. int Latency = getOperandLatency(ItinData, DefMCID, DefIdx, DefAlign, UseMCID, UseIdx, UseAlign); // Unable to find operand latency. The caller may resort to getInstrLatency. if (Latency < 0) return Latency; // Adjust for IT block position. int Adj = DefAdj + UseAdj; // Adjust for dynamic def-side opcode variants not captured by the itinerary. Adj += adjustDefLatency(Subtarget, DefMI, DefMCID, DefAlign); if (Adj >= 0 || (int)Latency > -Adj) { return Latency + Adj; } // Return the itinerary latency, which may be zero but not less than zero. return Latency; } int ARMBaseInstrInfo::getOperandLatency(const InstrItineraryData *ItinData, SDNode *DefNode, unsigned DefIdx, SDNode *UseNode, unsigned UseIdx) const { if (!DefNode->isMachineOpcode()) return 1; const MCInstrDesc &DefMCID = get(DefNode->getMachineOpcode()); if (isZeroCost(DefMCID.Opcode)) return 0; if (!ItinData || ItinData->isEmpty()) return DefMCID.mayLoad() ? 3 : 1; if (!UseNode->isMachineOpcode()) { int Latency = ItinData->getOperandCycle(DefMCID.getSchedClass(), DefIdx); int Adj = Subtarget.getPreISelOperandLatencyAdjustment(); int Threshold = 1 + Adj; return Latency <= Threshold ? 1 : Latency - Adj; } const MCInstrDesc &UseMCID = get(UseNode->getMachineOpcode()); auto *DefMN = cast(DefNode); unsigned DefAlign = !DefMN->memoperands_empty() ? (*DefMN->memoperands_begin())->getAlign().value() : 0; auto *UseMN = cast(UseNode); unsigned UseAlign = !UseMN->memoperands_empty() ? (*UseMN->memoperands_begin())->getAlign().value() : 0; int Latency = getOperandLatency(ItinData, DefMCID, DefIdx, DefAlign, UseMCID, UseIdx, UseAlign); if (Latency > 1 && (Subtarget.isCortexA8() || Subtarget.isLikeA9() || Subtarget.isCortexA7())) { // FIXME: Shifter op hack: no shift (i.e. [r +/- r]) or [r + r << 2] // variants are one cycle cheaper. switch (DefMCID.getOpcode()) { default: break; case ARM::LDRrs: case ARM::LDRBrs: { unsigned ShOpVal = cast(DefNode->getOperand(2))->getZExtValue(); unsigned ShImm = ARM_AM::getAM2Offset(ShOpVal); if (ShImm == 0 || (ShImm == 2 && ARM_AM::getAM2ShiftOpc(ShOpVal) == ARM_AM::lsl)) --Latency; break; } case ARM::t2LDRs: case ARM::t2LDRBs: case ARM::t2LDRHs: case ARM::t2LDRSHs: { // Thumb2 mode: lsl only. unsigned ShAmt = cast(DefNode->getOperand(2))->getZExtValue(); if (ShAmt == 0 || ShAmt == 2) --Latency; break; } } } else if (DefIdx == 0 && Latency > 2 && Subtarget.isSwift()) { // FIXME: Properly handle all of the latency adjustments for address // writeback. switch (DefMCID.getOpcode()) { default: break; case ARM::LDRrs: case ARM::LDRBrs: { unsigned ShOpVal = cast(DefNode->getOperand(2))->getZExtValue(); unsigned ShImm = ARM_AM::getAM2Offset(ShOpVal); if (ShImm == 0 || ((ShImm == 1 || ShImm == 2 || ShImm == 3) && ARM_AM::getAM2ShiftOpc(ShOpVal) == ARM_AM::lsl)) Latency -= 2; else if (ShImm == 1 && ARM_AM::getAM2ShiftOpc(ShOpVal) == ARM_AM::lsr) --Latency; break; } case ARM::t2LDRs: case ARM::t2LDRBs: case ARM::t2LDRHs: case ARM::t2LDRSHs: // Thumb2 mode: lsl 0-3 only. Latency -= 2; break; } } if (DefAlign < 8 && Subtarget.checkVLDnAccessAlignment()) switch (DefMCID.getOpcode()) { default: break; case ARM::VLD1q8: case ARM::VLD1q16: case ARM::VLD1q32: case ARM::VLD1q64: case ARM::VLD1q8wb_register: case ARM::VLD1q16wb_register: case ARM::VLD1q32wb_register: case ARM::VLD1q64wb_register: case ARM::VLD1q8wb_fixed: case ARM::VLD1q16wb_fixed: case ARM::VLD1q32wb_fixed: case ARM::VLD1q64wb_fixed: case ARM::VLD2d8: case ARM::VLD2d16: case ARM::VLD2d32: case ARM::VLD2q8Pseudo: case ARM::VLD2q16Pseudo: case ARM::VLD2q32Pseudo: case ARM::VLD2d8wb_fixed: case ARM::VLD2d16wb_fixed: case ARM::VLD2d32wb_fixed: case ARM::VLD2q8PseudoWB_fixed: case ARM::VLD2q16PseudoWB_fixed: case ARM::VLD2q32PseudoWB_fixed: case ARM::VLD2d8wb_register: case ARM::VLD2d16wb_register: case ARM::VLD2d32wb_register: case ARM::VLD2q8PseudoWB_register: case ARM::VLD2q16PseudoWB_register: case ARM::VLD2q32PseudoWB_register: case ARM::VLD3d8Pseudo: case ARM::VLD3d16Pseudo: case ARM::VLD3d32Pseudo: case ARM::VLD1d8TPseudo: case ARM::VLD1d16TPseudo: case ARM::VLD1d32TPseudo: case ARM::VLD1d64TPseudo: case ARM::VLD1d64TPseudoWB_fixed: case ARM::VLD1d64TPseudoWB_register: case ARM::VLD3d8Pseudo_UPD: case ARM::VLD3d16Pseudo_UPD: case ARM::VLD3d32Pseudo_UPD: case ARM::VLD3q8Pseudo_UPD: case ARM::VLD3q16Pseudo_UPD: case ARM::VLD3q32Pseudo_UPD: case ARM::VLD3q8oddPseudo: case ARM::VLD3q16oddPseudo: case ARM::VLD3q32oddPseudo: case ARM::VLD3q8oddPseudo_UPD: case ARM::VLD3q16oddPseudo_UPD: case ARM::VLD3q32oddPseudo_UPD: case ARM::VLD4d8Pseudo: case ARM::VLD4d16Pseudo: case ARM::VLD4d32Pseudo: case ARM::VLD1d8QPseudo: case ARM::VLD1d16QPseudo: case ARM::VLD1d32QPseudo: case ARM::VLD1d64QPseudo: case ARM::VLD1d64QPseudoWB_fixed: case ARM::VLD1d64QPseudoWB_register: case ARM::VLD1q8HighQPseudo: case ARM::VLD1q8LowQPseudo_UPD: case ARM::VLD1q8HighTPseudo: case ARM::VLD1q8LowTPseudo_UPD: case ARM::VLD1q16HighQPseudo: case ARM::VLD1q16LowQPseudo_UPD: case ARM::VLD1q16HighTPseudo: case ARM::VLD1q16LowTPseudo_UPD: case ARM::VLD1q32HighQPseudo: case ARM::VLD1q32LowQPseudo_UPD: case ARM::VLD1q32HighTPseudo: case ARM::VLD1q32LowTPseudo_UPD: case ARM::VLD1q64HighQPseudo: case ARM::VLD1q64LowQPseudo_UPD: case ARM::VLD1q64HighTPseudo: case ARM::VLD1q64LowTPseudo_UPD: case ARM::VLD4d8Pseudo_UPD: case ARM::VLD4d16Pseudo_UPD: case ARM::VLD4d32Pseudo_UPD: case ARM::VLD4q8Pseudo_UPD: case ARM::VLD4q16Pseudo_UPD: case ARM::VLD4q32Pseudo_UPD: case ARM::VLD4q8oddPseudo: case ARM::VLD4q16oddPseudo: case ARM::VLD4q32oddPseudo: case ARM::VLD4q8oddPseudo_UPD: case ARM::VLD4q16oddPseudo_UPD: case ARM::VLD4q32oddPseudo_UPD: case ARM::VLD1DUPq8: case ARM::VLD1DUPq16: case ARM::VLD1DUPq32: case ARM::VLD1DUPq8wb_fixed: case ARM::VLD1DUPq16wb_fixed: case ARM::VLD1DUPq32wb_fixed: case ARM::VLD1DUPq8wb_register: case ARM::VLD1DUPq16wb_register: case ARM::VLD1DUPq32wb_register: case ARM::VLD2DUPd8: case ARM::VLD2DUPd16: case ARM::VLD2DUPd32: case ARM::VLD2DUPd8wb_fixed: case ARM::VLD2DUPd16wb_fixed: case ARM::VLD2DUPd32wb_fixed: case ARM::VLD2DUPd8wb_register: case ARM::VLD2DUPd16wb_register: case ARM::VLD2DUPd32wb_register: case ARM::VLD2DUPq8EvenPseudo: case ARM::VLD2DUPq8OddPseudo: case ARM::VLD2DUPq16EvenPseudo: case ARM::VLD2DUPq16OddPseudo: case ARM::VLD2DUPq32EvenPseudo: case ARM::VLD2DUPq32OddPseudo: case ARM::VLD3DUPq8EvenPseudo: case ARM::VLD3DUPq8OddPseudo: case ARM::VLD3DUPq16EvenPseudo: case ARM::VLD3DUPq16OddPseudo: case ARM::VLD3DUPq32EvenPseudo: case ARM::VLD3DUPq32OddPseudo: case ARM::VLD4DUPd8Pseudo: case ARM::VLD4DUPd16Pseudo: case ARM::VLD4DUPd32Pseudo: case ARM::VLD4DUPd8Pseudo_UPD: case ARM::VLD4DUPd16Pseudo_UPD: case ARM::VLD4DUPd32Pseudo_UPD: case ARM::VLD4DUPq8EvenPseudo: case ARM::VLD4DUPq8OddPseudo: case ARM::VLD4DUPq16EvenPseudo: case ARM::VLD4DUPq16OddPseudo: case ARM::VLD4DUPq32EvenPseudo: case ARM::VLD4DUPq32OddPseudo: 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::VLD2LNq16Pseudo: case ARM::VLD2LNq32Pseudo: case ARM::VLD2LNd8Pseudo_UPD: case ARM::VLD2LNd16Pseudo_UPD: case ARM::VLD2LNd32Pseudo_UPD: case ARM::VLD2LNq16Pseudo_UPD: case ARM::VLD2LNq32Pseudo_UPD: case ARM::VLD4LNd8Pseudo: case ARM::VLD4LNd16Pseudo: case ARM::VLD4LNd32Pseudo: case ARM::VLD4LNq16Pseudo: case ARM::VLD4LNq32Pseudo: case ARM::VLD4LNd8Pseudo_UPD: case ARM::VLD4LNd16Pseudo_UPD: case ARM::VLD4LNd32Pseudo_UPD: case ARM::VLD4LNq16Pseudo_UPD: case ARM::VLD4LNq32Pseudo_UPD: // If the address is not 64-bit aligned, the latencies of these // instructions increases by one. ++Latency; break; } return Latency; } unsigned ARMBaseInstrInfo::getPredicationCost(const MachineInstr &MI) const { if (MI.isCopyLike() || MI.isInsertSubreg() || MI.isRegSequence() || MI.isImplicitDef()) return 0; if (MI.isBundle()) return 0; const MCInstrDesc &MCID = MI.getDesc(); if (MCID.isCall() || (MCID.hasImplicitDefOfPhysReg(ARM::CPSR) && !Subtarget.cheapPredicableCPSRDef())) { // When predicated, CPSR is an additional source operand for CPSR updating // instructions, this apparently increases their latencies. return 1; } return 0; } unsigned ARMBaseInstrInfo::getInstrLatency(const InstrItineraryData *ItinData, const MachineInstr &MI, unsigned *PredCost) const { if (MI.isCopyLike() || MI.isInsertSubreg() || MI.isRegSequence() || MI.isImplicitDef()) return 1; // An instruction scheduler typically runs on unbundled instructions, however // other passes may query the latency of a bundled instruction. if (MI.isBundle()) { unsigned Latency = 0; MachineBasicBlock::const_instr_iterator I = MI.getIterator(); MachineBasicBlock::const_instr_iterator E = MI.getParent()->instr_end(); while (++I != E && I->isInsideBundle()) { if (I->getOpcode() != ARM::t2IT) Latency += getInstrLatency(ItinData, *I, PredCost); } return Latency; } const MCInstrDesc &MCID = MI.getDesc(); if (PredCost && (MCID.isCall() || (MCID.hasImplicitDefOfPhysReg(ARM::CPSR) && !Subtarget.cheapPredicableCPSRDef()))) { // When predicated, CPSR is an additional source operand for CPSR updating // instructions, this apparently increases their latencies. *PredCost = 1; } // Be sure to call getStageLatency for an empty itinerary in case it has a // valid MinLatency property. if (!ItinData) return MI.mayLoad() ? 3 : 1; unsigned Class = MCID.getSchedClass(); // For instructions with variable uops, use uops as latency. if (!ItinData->isEmpty() && ItinData->getNumMicroOps(Class) < 0) return getNumMicroOps(ItinData, MI); // For the common case, fall back on the itinerary's latency. unsigned Latency = ItinData->getStageLatency(Class); // Adjust for dynamic def-side opcode variants not captured by the itinerary. unsigned DefAlign = MI.hasOneMemOperand() ? (*MI.memoperands_begin())->getAlign().value() : 0; int Adj = adjustDefLatency(Subtarget, MI, MCID, DefAlign); if (Adj >= 0 || (int)Latency > -Adj) { return Latency + Adj; } return Latency; } int ARMBaseInstrInfo::getInstrLatency(const InstrItineraryData *ItinData, SDNode *Node) const { if (!Node->isMachineOpcode()) return 1; if (!ItinData || ItinData->isEmpty()) return 1; unsigned Opcode = Node->getMachineOpcode(); switch (Opcode) { default: return ItinData->getStageLatency(get(Opcode).getSchedClass()); case ARM::VLDMQIA: case ARM::VSTMQIA: return 2; } } bool ARMBaseInstrInfo::hasHighOperandLatency(const TargetSchedModel &SchedModel, const MachineRegisterInfo *MRI, const MachineInstr &DefMI, unsigned DefIdx, const MachineInstr &UseMI, unsigned UseIdx) const { unsigned DDomain = DefMI.getDesc().TSFlags & ARMII::DomainMask; unsigned UDomain = UseMI.getDesc().TSFlags & ARMII::DomainMask; if (Subtarget.nonpipelinedVFP() && (DDomain == ARMII::DomainVFP || UDomain == ARMII::DomainVFP)) return true; // Hoist VFP / NEON instructions with 4 or higher latency. unsigned Latency = SchedModel.computeOperandLatency(&DefMI, DefIdx, &UseMI, UseIdx); if (Latency <= 3) return false; return DDomain == ARMII::DomainVFP || DDomain == ARMII::DomainNEON || UDomain == ARMII::DomainVFP || UDomain == ARMII::DomainNEON; } bool ARMBaseInstrInfo::hasLowDefLatency(const TargetSchedModel &SchedModel, const MachineInstr &DefMI, unsigned DefIdx) const { const InstrItineraryData *ItinData = SchedModel.getInstrItineraries(); if (!ItinData || ItinData->isEmpty()) return false; unsigned DDomain = DefMI.getDesc().TSFlags & ARMII::DomainMask; if (DDomain == ARMII::DomainGeneral) { unsigned DefClass = DefMI.getDesc().getSchedClass(); int DefCycle = ItinData->getOperandCycle(DefClass, DefIdx); return (DefCycle != -1 && DefCycle <= 2); } return false; } bool ARMBaseInstrInfo::verifyInstruction(const MachineInstr &MI, StringRef &ErrInfo) const { if (convertAddSubFlagsOpcode(MI.getOpcode())) { ErrInfo = "Pseudo flag setting opcodes only exist in Selection DAG"; return false; } if (MI.getOpcode() == ARM::tMOVr && !Subtarget.hasV6Ops()) { // Make sure we don't generate a lo-lo mov that isn't supported. if (!ARM::hGPRRegClass.contains(MI.getOperand(0).getReg()) && !ARM::hGPRRegClass.contains(MI.getOperand(1).getReg())) { ErrInfo = "Non-flag-setting Thumb1 mov is v6-only"; return false; } } if (MI.getOpcode() == ARM::tPUSH || MI.getOpcode() == ARM::tPOP || MI.getOpcode() == ARM::tPOP_RET) { for (const MachineOperand &MO : llvm::drop_begin(MI.operands(), 2)) { if (MO.isImplicit() || !MO.isReg()) continue; Register Reg = MO.getReg(); if (Reg < ARM::R0 || Reg > ARM::R7) { if (!(MI.getOpcode() == ARM::tPUSH && Reg == ARM::LR) && !(MI.getOpcode() == ARM::tPOP_RET && Reg == ARM::PC)) { ErrInfo = "Unsupported register in Thumb1 push/pop"; return false; } } } } if (MI.getOpcode() == ARM::MVE_VMOV_q_rr) { assert(MI.getOperand(4).isImm() && MI.getOperand(5).isImm()); if ((MI.getOperand(4).getImm() != 2 && MI.getOperand(4).getImm() != 3) || MI.getOperand(4).getImm() != MI.getOperand(5).getImm() + 2) { ErrInfo = "Incorrect array index for MVE_VMOV_q_rr"; return false; } } // Check the address model by taking the first Imm operand and checking it is // legal for that addressing mode. ARMII::AddrMode AddrMode = (ARMII::AddrMode)(MI.getDesc().TSFlags & ARMII::AddrModeMask); switch (AddrMode) { default: break; case ARMII::AddrModeT2_i7: case ARMII::AddrModeT2_i7s2: case ARMII::AddrModeT2_i7s4: case ARMII::AddrModeT2_i8: case ARMII::AddrModeT2_i8pos: case ARMII::AddrModeT2_i8neg: case ARMII::AddrModeT2_i8s4: case ARMII::AddrModeT2_i12: { uint32_t Imm = 0; for (auto Op : MI.operands()) { if (Op.isImm()) { Imm = Op.getImm(); break; } } if (!isLegalAddressImm(MI.getOpcode(), Imm, this)) { ErrInfo = "Incorrect AddrMode Imm for instruction"; return false; } break; } } return true; } void ARMBaseInstrInfo::expandLoadStackGuardBase(MachineBasicBlock::iterator MI, unsigned LoadImmOpc, unsigned LoadOpc) const { assert(!Subtarget.isROPI() && !Subtarget.isRWPI() && "ROPI/RWPI not currently supported with stack guard"); MachineBasicBlock &MBB = *MI->getParent(); DebugLoc DL = MI->getDebugLoc(); Register Reg = MI->getOperand(0).getReg(); MachineInstrBuilder MIB; unsigned int Offset = 0; if (LoadImmOpc == ARM::MRC || LoadImmOpc == ARM::t2MRC) { assert(!Subtarget.isReadTPSoft() && "TLS stack protector requires hardware TLS register"); BuildMI(MBB, MI, DL, get(LoadImmOpc), Reg) .addImm(15) .addImm(0) .addImm(13) .addImm(0) .addImm(3) .add(predOps(ARMCC::AL)); Module &M = *MBB.getParent()->getFunction().getParent(); Offset = M.getStackProtectorGuardOffset(); if (Offset & ~0xfffU) { // The offset won't fit in the LDR's 12-bit immediate field, so emit an // extra ADD to cover the delta. This gives us a guaranteed 8 additional // bits, resulting in a range of 0 to +1 MiB for the guard offset. unsigned AddOpc = (LoadImmOpc == ARM::MRC) ? ARM::ADDri : ARM::t2ADDri; BuildMI(MBB, MI, DL, get(AddOpc), Reg) .addReg(Reg, RegState::Kill) .addImm(Offset & ~0xfffU) .add(predOps(ARMCC::AL)) .addReg(0); Offset &= 0xfffU; } } else { const GlobalValue *GV = cast((*MI->memoperands_begin())->getValue()); bool IsIndirect = Subtarget.isGVIndirectSymbol(GV); unsigned TargetFlags = ARMII::MO_NO_FLAG; if (Subtarget.isTargetMachO()) { TargetFlags |= ARMII::MO_NONLAZY; } else if (Subtarget.isTargetCOFF()) { if (GV->hasDLLImportStorageClass()) TargetFlags |= ARMII::MO_DLLIMPORT; else if (IsIndirect) TargetFlags |= ARMII::MO_COFFSTUB; } else if (Subtarget.isGVInGOT(GV)) { TargetFlags |= ARMII::MO_GOT; } BuildMI(MBB, MI, DL, get(LoadImmOpc), Reg) .addGlobalAddress(GV, 0, TargetFlags); if (IsIndirect) { MIB = BuildMI(MBB, MI, DL, get(LoadOpc), Reg); MIB.addReg(Reg, RegState::Kill).addImm(0); auto Flags = MachineMemOperand::MOLoad | MachineMemOperand::MODereferenceable | MachineMemOperand::MOInvariant; MachineMemOperand *MMO = MBB.getParent()->getMachineMemOperand( MachinePointerInfo::getGOT(*MBB.getParent()), Flags, 4, Align(4)); MIB.addMemOperand(MMO).add(predOps(ARMCC::AL)); } } MIB = BuildMI(MBB, MI, DL, get(LoadOpc), Reg); MIB.addReg(Reg, RegState::Kill) .addImm(Offset) .cloneMemRefs(*MI) .add(predOps(ARMCC::AL)); } bool ARMBaseInstrInfo::isFpMLxInstruction(unsigned Opcode, unsigned &MulOpc, unsigned &AddSubOpc, bool &NegAcc, bool &HasLane) const { DenseMap::const_iterator I = MLxEntryMap.find(Opcode); if (I == MLxEntryMap.end()) return false; const ARM_MLxEntry &Entry = ARM_MLxTable[I->second]; MulOpc = Entry.MulOpc; AddSubOpc = Entry.AddSubOpc; NegAcc = Entry.NegAcc; HasLane = Entry.HasLane; return true; } //===----------------------------------------------------------------------===// // Execution domains. //===----------------------------------------------------------------------===// // // Some instructions go down the NEON pipeline, some go down the VFP pipeline, // and some can go down both. The vmov instructions go down the VFP pipeline, // but they can be changed to vorr equivalents that are executed by the NEON // pipeline. // // We use the following execution domain numbering: // enum ARMExeDomain { ExeGeneric = 0, ExeVFP = 1, ExeNEON = 2 }; // // Also see ARMInstrFormats.td and Domain* enums in ARMBaseInfo.h // std::pair ARMBaseInstrInfo::getExecutionDomain(const MachineInstr &MI) const { // If we don't have access to NEON instructions then we won't be able // to swizzle anything to the NEON domain. Check to make sure. if (Subtarget.hasNEON()) { // VMOVD, VMOVRS and VMOVSR are VFP instructions, but can be changed to NEON // if they are not predicated. if (MI.getOpcode() == ARM::VMOVD && !isPredicated(MI)) return std::make_pair(ExeVFP, (1 << ExeVFP) | (1 << ExeNEON)); // CortexA9 is particularly picky about mixing the two and wants these // converted. if (Subtarget.useNEONForFPMovs() && !isPredicated(MI) && (MI.getOpcode() == ARM::VMOVRS || MI.getOpcode() == ARM::VMOVSR || MI.getOpcode() == ARM::VMOVS)) return std::make_pair(ExeVFP, (1 << ExeVFP) | (1 << ExeNEON)); } // No other instructions can be swizzled, so just determine their domain. unsigned Domain = MI.getDesc().TSFlags & ARMII::DomainMask; if (Domain & ARMII::DomainNEON) return std::make_pair(ExeNEON, 0); // Certain instructions can go either way on Cortex-A8. // Treat them as NEON instructions. if ((Domain & ARMII::DomainNEONA8) && Subtarget.isCortexA8()) return std::make_pair(ExeNEON, 0); if (Domain & ARMII::DomainVFP) return std::make_pair(ExeVFP, 0); return std::make_pair(ExeGeneric, 0); } static unsigned getCorrespondingDRegAndLane(const TargetRegisterInfo *TRI, unsigned SReg, unsigned &Lane) { unsigned DReg = TRI->getMatchingSuperReg(SReg, ARM::ssub_0, &ARM::DPRRegClass); Lane = 0; if (DReg != ARM::NoRegister) return DReg; Lane = 1; DReg = TRI->getMatchingSuperReg(SReg, ARM::ssub_1, &ARM::DPRRegClass); assert(DReg && "S-register with no D super-register?"); return DReg; } /// getImplicitSPRUseForDPRUse - Given a use of a DPR register and lane, /// set ImplicitSReg to a register number that must be marked as implicit-use or /// zero if no register needs to be defined as implicit-use. /// /// If the function cannot determine if an SPR should be marked implicit use or /// not, it returns false. /// /// This function handles cases where an instruction is being modified from taking /// an SPR to a DPR[Lane]. A use of the DPR is being added, which may conflict /// with an earlier def of an SPR corresponding to DPR[Lane^1] (i.e. the other /// lane of the DPR). /// /// If the other SPR is defined, an implicit-use of it should be added. Else, /// (including the case where the DPR itself is defined), it should not. /// static bool getImplicitSPRUseForDPRUse(const TargetRegisterInfo *TRI, MachineInstr &MI, unsigned DReg, unsigned Lane, unsigned &ImplicitSReg) { // If the DPR is defined or used already, the other SPR lane will be chained // correctly, so there is nothing to be done. if (MI.definesRegister(DReg, TRI) || MI.readsRegister(DReg, TRI)) { ImplicitSReg = 0; return true; } // Otherwise we need to go searching to see if the SPR is set explicitly. ImplicitSReg = TRI->getSubReg(DReg, (Lane & 1) ? ARM::ssub_0 : ARM::ssub_1); MachineBasicBlock::LivenessQueryResult LQR = MI.getParent()->computeRegisterLiveness(TRI, ImplicitSReg, MI); if (LQR == MachineBasicBlock::LQR_Live) return true; else if (LQR == MachineBasicBlock::LQR_Unknown) return false; // If the register is known not to be live, there is no need to add an // implicit-use. ImplicitSReg = 0; return true; } void ARMBaseInstrInfo::setExecutionDomain(MachineInstr &MI, unsigned Domain) const { unsigned DstReg, SrcReg, DReg; unsigned Lane; MachineInstrBuilder MIB(*MI.getParent()->getParent(), MI); const TargetRegisterInfo *TRI = &getRegisterInfo(); switch (MI.getOpcode()) { default: llvm_unreachable("cannot handle opcode!"); break; case ARM::VMOVD: if (Domain != ExeNEON) break; // Zap the predicate operands. assert(!isPredicated(MI) && "Cannot predicate a VORRd"); // Make sure we've got NEON instructions. assert(Subtarget.hasNEON() && "VORRd requires NEON"); // Source instruction is %DDst = VMOVD %DSrc, 14, %noreg (; implicits) DstReg = MI.getOperand(0).getReg(); SrcReg = MI.getOperand(1).getReg(); for (unsigned i = MI.getDesc().getNumOperands(); i; --i) MI.removeOperand(i - 1); // Change to a %DDst = VORRd %DSrc, %DSrc, 14, %noreg (; implicits) MI.setDesc(get(ARM::VORRd)); MIB.addReg(DstReg, RegState::Define) .addReg(SrcReg) .addReg(SrcReg) .add(predOps(ARMCC::AL)); break; case ARM::VMOVRS: if (Domain != ExeNEON) break; assert(!isPredicated(MI) && "Cannot predicate a VGETLN"); // Source instruction is %RDst = VMOVRS %SSrc, 14, %noreg (; implicits) DstReg = MI.getOperand(0).getReg(); SrcReg = MI.getOperand(1).getReg(); for (unsigned i = MI.getDesc().getNumOperands(); i; --i) MI.removeOperand(i - 1); DReg = getCorrespondingDRegAndLane(TRI, SrcReg, Lane); // Convert to %RDst = VGETLNi32 %DSrc, Lane, 14, %noreg (; imps) // Note that DSrc has been widened and the other lane may be undef, which // contaminates the entire register. MI.setDesc(get(ARM::VGETLNi32)); MIB.addReg(DstReg, RegState::Define) .addReg(DReg, RegState::Undef) .addImm(Lane) .add(predOps(ARMCC::AL)); // The old source should be an implicit use, otherwise we might think it // was dead before here. MIB.addReg(SrcReg, RegState::Implicit); break; case ARM::VMOVSR: { if (Domain != ExeNEON) break; assert(!isPredicated(MI) && "Cannot predicate a VSETLN"); // Source instruction is %SDst = VMOVSR %RSrc, 14, %noreg (; implicits) DstReg = MI.getOperand(0).getReg(); SrcReg = MI.getOperand(1).getReg(); DReg = getCorrespondingDRegAndLane(TRI, DstReg, Lane); unsigned ImplicitSReg; if (!getImplicitSPRUseForDPRUse(TRI, MI, DReg, Lane, ImplicitSReg)) break; for (unsigned i = MI.getDesc().getNumOperands(); i; --i) MI.removeOperand(i - 1); // Convert to %DDst = VSETLNi32 %DDst, %RSrc, Lane, 14, %noreg (; imps) // Again DDst may be undefined at the beginning of this instruction. MI.setDesc(get(ARM::VSETLNi32)); MIB.addReg(DReg, RegState::Define) .addReg(DReg, getUndefRegState(!MI.readsRegister(DReg, TRI))) .addReg(SrcReg) .addImm(Lane) .add(predOps(ARMCC::AL)); // The narrower destination must be marked as set to keep previous chains // in place. MIB.addReg(DstReg, RegState::Define | RegState::Implicit); if (ImplicitSReg != 0) MIB.addReg(ImplicitSReg, RegState::Implicit); break; } case ARM::VMOVS: { if (Domain != ExeNEON) break; // Source instruction is %SDst = VMOVS %SSrc, 14, %noreg (; implicits) DstReg = MI.getOperand(0).getReg(); SrcReg = MI.getOperand(1).getReg(); unsigned DstLane = 0, SrcLane = 0, DDst, DSrc; DDst = getCorrespondingDRegAndLane(TRI, DstReg, DstLane); DSrc = getCorrespondingDRegAndLane(TRI, SrcReg, SrcLane); unsigned ImplicitSReg; if (!getImplicitSPRUseForDPRUse(TRI, MI, DSrc, SrcLane, ImplicitSReg)) break; for (unsigned i = MI.getDesc().getNumOperands(); i; --i) MI.removeOperand(i - 1); if (DSrc == DDst) { // Destination can be: // %DDst = VDUPLN32d %DDst, Lane, 14, %noreg (; implicits) MI.setDesc(get(ARM::VDUPLN32d)); MIB.addReg(DDst, RegState::Define) .addReg(DDst, getUndefRegState(!MI.readsRegister(DDst, TRI))) .addImm(SrcLane) .add(predOps(ARMCC::AL)); // Neither the source or the destination are naturally represented any // more, so add them in manually. MIB.addReg(DstReg, RegState::Implicit | RegState::Define); MIB.addReg(SrcReg, RegState::Implicit); if (ImplicitSReg != 0) MIB.addReg(ImplicitSReg, RegState::Implicit); break; } // In general there's no single instruction that can perform an S <-> S // move in NEON space, but a pair of VEXT instructions *can* do the // job. It turns out that the VEXTs needed will only use DSrc once, with // the position based purely on the combination of lane-0 and lane-1 // involved. For example // vmov s0, s2 -> vext.32 d0, d0, d1, #1 vext.32 d0, d0, d0, #1 // vmov s1, s3 -> vext.32 d0, d1, d0, #1 vext.32 d0, d0, d0, #1 // vmov s0, s3 -> vext.32 d0, d0, d0, #1 vext.32 d0, d1, d0, #1 // vmov s1, s2 -> vext.32 d0, d0, d0, #1 vext.32 d0, d0, d1, #1 // // Pattern of the MachineInstrs is: // %DDst = VEXTd32 %DSrc1, %DSrc2, Lane, 14, %noreg (;implicits) MachineInstrBuilder NewMIB; NewMIB = BuildMI(*MI.getParent(), MI, MI.getDebugLoc(), get(ARM::VEXTd32), DDst); // On the first instruction, both DSrc and DDst may be undef if present. // Specifically when the original instruction didn't have them as an // . unsigned CurReg = SrcLane == 1 && DstLane == 1 ? DSrc : DDst; bool CurUndef = !MI.readsRegister(CurReg, TRI); NewMIB.addReg(CurReg, getUndefRegState(CurUndef)); CurReg = SrcLane == 0 && DstLane == 0 ? DSrc : DDst; CurUndef = !MI.readsRegister(CurReg, TRI); NewMIB.addReg(CurReg, getUndefRegState(CurUndef)) .addImm(1) .add(predOps(ARMCC::AL)); if (SrcLane == DstLane) NewMIB.addReg(SrcReg, RegState::Implicit); MI.setDesc(get(ARM::VEXTd32)); MIB.addReg(DDst, RegState::Define); // On the second instruction, DDst has definitely been defined above, so // it is not undef. DSrc, if present, can be undef as above. CurReg = SrcLane == 1 && DstLane == 0 ? DSrc : DDst; CurUndef = CurReg == DSrc && !MI.readsRegister(CurReg, TRI); MIB.addReg(CurReg, getUndefRegState(CurUndef)); CurReg = SrcLane == 0 && DstLane == 1 ? DSrc : DDst; CurUndef = CurReg == DSrc && !MI.readsRegister(CurReg, TRI); MIB.addReg(CurReg, getUndefRegState(CurUndef)) .addImm(1) .add(predOps(ARMCC::AL)); if (SrcLane != DstLane) MIB.addReg(SrcReg, RegState::Implicit); // As before, the original destination is no longer represented, add it // implicitly. MIB.addReg(DstReg, RegState::Define | RegState::Implicit); if (ImplicitSReg != 0) MIB.addReg(ImplicitSReg, RegState::Implicit); break; } } } //===----------------------------------------------------------------------===// // Partial register updates //===----------------------------------------------------------------------===// // // Swift renames NEON registers with 64-bit granularity. That means any // instruction writing an S-reg implicitly reads the containing D-reg. The // problem is mostly avoided by translating f32 operations to v2f32 operations // on D-registers, but f32 loads are still a problem. // // These instructions can load an f32 into a NEON register: // // VLDRS - Only writes S, partial D update. // VLD1LNd32 - Writes all D-regs, explicit partial D update, 2 uops. // VLD1DUPd32 - Writes all D-regs, no partial reg update, 2 uops. // // FCONSTD can be used as a dependency-breaking instruction. unsigned ARMBaseInstrInfo::getPartialRegUpdateClearance( const MachineInstr &MI, unsigned OpNum, const TargetRegisterInfo *TRI) const { auto PartialUpdateClearance = Subtarget.getPartialUpdateClearance(); if (!PartialUpdateClearance) return 0; assert(TRI && "Need TRI instance"); const MachineOperand &MO = MI.getOperand(OpNum); if (MO.readsReg()) return 0; Register Reg = MO.getReg(); int UseOp = -1; switch (MI.getOpcode()) { // Normal instructions writing only an S-register. case ARM::VLDRS: case ARM::FCONSTS: case ARM::VMOVSR: case ARM::VMOVv8i8: case ARM::VMOVv4i16: case ARM::VMOVv2i32: case ARM::VMOVv2f32: case ARM::VMOVv1i64: UseOp = MI.findRegisterUseOperandIdx(Reg, false, TRI); break; // Explicitly reads the dependency. case ARM::VLD1LNd32: UseOp = 3; break; default: return 0; } // If this instruction actually reads a value from Reg, there is no unwanted // dependency. if (UseOp != -1 && MI.getOperand(UseOp).readsReg()) return 0; // We must be able to clobber the whole D-reg. if (Reg.isVirtual()) { // Virtual register must be a def undef foo:ssub_0 operand. if (!MO.getSubReg() || MI.readsVirtualRegister(Reg)) return 0; } else if (ARM::SPRRegClass.contains(Reg)) { // Physical register: MI must define the full D-reg. unsigned DReg = TRI->getMatchingSuperReg(Reg, ARM::ssub_0, &ARM::DPRRegClass); if (!DReg || !MI.definesRegister(DReg, TRI)) return 0; } // MI has an unwanted D-register dependency. // Avoid defs in the previous N instructrions. return PartialUpdateClearance; } // Break a partial register dependency after getPartialRegUpdateClearance // returned non-zero. void ARMBaseInstrInfo::breakPartialRegDependency( MachineInstr &MI, unsigned OpNum, const TargetRegisterInfo *TRI) const { assert(OpNum < MI.getDesc().getNumDefs() && "OpNum is not a def"); assert(TRI && "Need TRI instance"); const MachineOperand &MO = MI.getOperand(OpNum); Register Reg = MO.getReg(); assert(Reg.isPhysical() && "Can't break virtual register dependencies."); unsigned DReg = Reg; // If MI defines an S-reg, find the corresponding D super-register. if (ARM::SPRRegClass.contains(Reg)) { DReg = ARM::D0 + (Reg - ARM::S0) / 2; assert(TRI->isSuperRegister(Reg, DReg) && "Register enums broken"); } assert(ARM::DPRRegClass.contains(DReg) && "Can only break D-reg deps"); assert(MI.definesRegister(DReg, TRI) && "MI doesn't clobber full D-reg"); // FIXME: In some cases, VLDRS can be changed to a VLD1DUPd32 which defines // the full D-register by loading the same value to both lanes. The // instruction is micro-coded with 2 uops, so don't do this until we can // properly schedule micro-coded instructions. The dispatcher stalls cause // too big regressions. // Insert the dependency-breaking FCONSTD before MI. // 96 is the encoding of 0.5, but the actual value doesn't matter here. BuildMI(*MI.getParent(), MI, MI.getDebugLoc(), get(ARM::FCONSTD), DReg) .addImm(96) .add(predOps(ARMCC::AL)); MI.addRegisterKilled(DReg, TRI, true); } bool ARMBaseInstrInfo::hasNOP() const { return Subtarget.hasFeature(ARM::HasV6KOps); } bool ARMBaseInstrInfo::isSwiftFastImmShift(const MachineInstr *MI) const { if (MI->getNumOperands() < 4) return true; unsigned ShOpVal = MI->getOperand(3).getImm(); unsigned ShImm = ARM_AM::getSORegOffset(ShOpVal); // Swift supports faster shifts for: lsl 2, lsl 1, and lsr 1. if ((ShImm == 1 && ARM_AM::getSORegShOp(ShOpVal) == ARM_AM::lsr) || ((ShImm == 1 || ShImm == 2) && ARM_AM::getSORegShOp(ShOpVal) == ARM_AM::lsl)) return true; return false; } bool ARMBaseInstrInfo::getRegSequenceLikeInputs( const MachineInstr &MI, unsigned DefIdx, SmallVectorImpl &InputRegs) const { assert(DefIdx < MI.getDesc().getNumDefs() && "Invalid definition index"); assert(MI.isRegSequenceLike() && "Invalid kind of instruction"); switch (MI.getOpcode()) { case ARM::VMOVDRR: // dX = VMOVDRR rY, rZ // is the same as: // dX = REG_SEQUENCE rY, ssub_0, rZ, ssub_1 // Populate the InputRegs accordingly. // rY const MachineOperand *MOReg = &MI.getOperand(1); if (!MOReg->isUndef()) InputRegs.push_back(RegSubRegPairAndIdx(MOReg->getReg(), MOReg->getSubReg(), ARM::ssub_0)); // rZ MOReg = &MI.getOperand(2); if (!MOReg->isUndef()) InputRegs.push_back(RegSubRegPairAndIdx(MOReg->getReg(), MOReg->getSubReg(), ARM::ssub_1)); return true; } llvm_unreachable("Target dependent opcode missing"); } bool ARMBaseInstrInfo::getExtractSubregLikeInputs( const MachineInstr &MI, unsigned DefIdx, RegSubRegPairAndIdx &InputReg) const { assert(DefIdx < MI.getDesc().getNumDefs() && "Invalid definition index"); assert(MI.isExtractSubregLike() && "Invalid kind of instruction"); switch (MI.getOpcode()) { case ARM::VMOVRRD: // rX, rY = VMOVRRD dZ // is the same as: // rX = EXTRACT_SUBREG dZ, ssub_0 // rY = EXTRACT_SUBREG dZ, ssub_1 const MachineOperand &MOReg = MI.getOperand(2); if (MOReg.isUndef()) return false; InputReg.Reg = MOReg.getReg(); InputReg.SubReg = MOReg.getSubReg(); InputReg.SubIdx = DefIdx == 0 ? ARM::ssub_0 : ARM::ssub_1; return true; } llvm_unreachable("Target dependent opcode missing"); } bool ARMBaseInstrInfo::getInsertSubregLikeInputs( const MachineInstr &MI, unsigned DefIdx, RegSubRegPair &BaseReg, RegSubRegPairAndIdx &InsertedReg) const { assert(DefIdx < MI.getDesc().getNumDefs() && "Invalid definition index"); assert(MI.isInsertSubregLike() && "Invalid kind of instruction"); switch (MI.getOpcode()) { case ARM::VSETLNi32: case ARM::MVE_VMOV_to_lane_32: // dX = VSETLNi32 dY, rZ, imm // qX = MVE_VMOV_to_lane_32 qY, rZ, imm const MachineOperand &MOBaseReg = MI.getOperand(1); const MachineOperand &MOInsertedReg = MI.getOperand(2); if (MOInsertedReg.isUndef()) return false; const MachineOperand &MOIndex = MI.getOperand(3); BaseReg.Reg = MOBaseReg.getReg(); BaseReg.SubReg = MOBaseReg.getSubReg(); InsertedReg.Reg = MOInsertedReg.getReg(); InsertedReg.SubReg = MOInsertedReg.getSubReg(); InsertedReg.SubIdx = ARM::ssub_0 + MOIndex.getImm(); return true; } llvm_unreachable("Target dependent opcode missing"); } std::pair ARMBaseInstrInfo::decomposeMachineOperandsTargetFlags(unsigned TF) const { const unsigned Mask = ARMII::MO_OPTION_MASK; return std::make_pair(TF & Mask, TF & ~Mask); } ArrayRef> ARMBaseInstrInfo::getSerializableDirectMachineOperandTargetFlags() const { using namespace ARMII; static const std::pair TargetFlags[] = { {MO_LO16, "arm-lo16"}, {MO_HI16, "arm-hi16"}, {MO_LO_0_7, "arm-lo-0-7"}, {MO_HI_0_7, "arm-hi-0-7"}, {MO_LO_8_15, "arm-lo-8-15"}, {MO_HI_8_15, "arm-hi-8-15"}, }; return ArrayRef(TargetFlags); } ArrayRef> ARMBaseInstrInfo::getSerializableBitmaskMachineOperandTargetFlags() const { using namespace ARMII; static const std::pair TargetFlags[] = { {MO_COFFSTUB, "arm-coffstub"}, {MO_GOT, "arm-got"}, {MO_SBREL, "arm-sbrel"}, {MO_DLLIMPORT, "arm-dllimport"}, {MO_SECREL, "arm-secrel"}, {MO_NONLAZY, "arm-nonlazy"}}; return ArrayRef(TargetFlags); } std::optional ARMBaseInstrInfo::isAddImmediate(const MachineInstr &MI, Register Reg) const { int Sign = 1; unsigned Opcode = MI.getOpcode(); int64_t Offset = 0; // TODO: Handle cases where Reg is a super- or sub-register of the // destination register. const MachineOperand &Op0 = MI.getOperand(0); if (!Op0.isReg() || Reg != Op0.getReg()) return std::nullopt; // We describe SUBri or ADDri instructions. if (Opcode == ARM::SUBri) Sign = -1; else if (Opcode != ARM::ADDri) return std::nullopt; // TODO: Third operand can be global address (usually some string). Since // strings can be relocated we cannot calculate their offsets for // now. if (!MI.getOperand(1).isReg() || !MI.getOperand(2).isImm()) return std::nullopt; Offset = MI.getOperand(2).getImm() * Sign; return RegImmPair{MI.getOperand(1).getReg(), Offset}; } bool llvm::registerDefinedBetween(unsigned Reg, MachineBasicBlock::iterator From, MachineBasicBlock::iterator To, const TargetRegisterInfo *TRI) { for (auto I = From; I != To; ++I) if (I->modifiesRegister(Reg, TRI)) return true; return false; } MachineInstr *llvm::findCMPToFoldIntoCBZ(MachineInstr *Br, const TargetRegisterInfo *TRI) { // Search backwards to the instruction that defines CSPR. This may or not // be a CMP, we check that after this loop. If we find another instruction // that reads cpsr, we return nullptr. MachineBasicBlock::iterator CmpMI = Br; while (CmpMI != Br->getParent()->begin()) { --CmpMI; if (CmpMI->modifiesRegister(ARM::CPSR, TRI)) break; if (CmpMI->readsRegister(ARM::CPSR, TRI)) break; } // Check that this inst is a CMP r[0-7], #0 and that the register // is not redefined between the cmp and the br. if (CmpMI->getOpcode() != ARM::tCMPi8 && CmpMI->getOpcode() != ARM::t2CMPri) return nullptr; Register Reg = CmpMI->getOperand(0).getReg(); Register PredReg; ARMCC::CondCodes Pred = getInstrPredicate(*CmpMI, PredReg); if (Pred != ARMCC::AL || CmpMI->getOperand(1).getImm() != 0) return nullptr; if (!isARMLowRegister(Reg)) return nullptr; if (registerDefinedBetween(Reg, CmpMI->getNextNode(), Br, TRI)) return nullptr; return &*CmpMI; } unsigned llvm::ConstantMaterializationCost(unsigned Val, const ARMSubtarget *Subtarget, bool ForCodesize) { if (Subtarget->isThumb()) { if (Val <= 255) // MOV return ForCodesize ? 2 : 1; if (Subtarget->hasV6T2Ops() && (Val <= 0xffff || // MOV ARM_AM::getT2SOImmVal(Val) != -1 || // MOVW ARM_AM::getT2SOImmVal(~Val) != -1)) // MVN return ForCodesize ? 4 : 1; if (Val <= 510) // MOV + ADDi8 return ForCodesize ? 4 : 2; if (~Val <= 255) // MOV + MVN return ForCodesize ? 4 : 2; if (ARM_AM::isThumbImmShiftedVal(Val)) // MOV + LSL return ForCodesize ? 4 : 2; } else { if (ARM_AM::getSOImmVal(Val) != -1) // MOV return ForCodesize ? 4 : 1; if (ARM_AM::getSOImmVal(~Val) != -1) // MVN return ForCodesize ? 4 : 1; if (Subtarget->hasV6T2Ops() && Val <= 0xffff) // MOVW return ForCodesize ? 4 : 1; if (ARM_AM::isSOImmTwoPartVal(Val)) // two instrs return ForCodesize ? 8 : 2; if (ARM_AM::isSOImmTwoPartValNeg(Val)) // two instrs return ForCodesize ? 8 : 2; } if (Subtarget->useMovt()) // MOVW + MOVT return ForCodesize ? 8 : 2; return ForCodesize ? 8 : 3; // Literal pool load } bool llvm::HasLowerConstantMaterializationCost(unsigned Val1, unsigned Val2, const ARMSubtarget *Subtarget, bool ForCodesize) { // Check with ForCodesize unsigned Cost1 = ConstantMaterializationCost(Val1, Subtarget, ForCodesize); unsigned Cost2 = ConstantMaterializationCost(Val2, Subtarget, ForCodesize); if (Cost1 < Cost2) return true; if (Cost1 > Cost2) return false; // If they are equal, try with !ForCodesize return ConstantMaterializationCost(Val1, Subtarget, !ForCodesize) < ConstantMaterializationCost(Val2, Subtarget, !ForCodesize); } /// Constants defining how certain sequences should be outlined. /// This encompasses how an outlined function should be called, and what kind of /// frame should be emitted for that outlined function. /// /// \p MachineOutlinerTailCall implies that the function is being created from /// a sequence of instructions ending in a return. /// /// That is, /// /// I1 OUTLINED_FUNCTION: /// I2 --> B OUTLINED_FUNCTION I1 /// BX LR I2 /// BX LR /// /// +-------------------------+--------+-----+ /// | | Thumb2 | ARM | /// +-------------------------+--------+-----+ /// | Call overhead in Bytes | 4 | 4 | /// | Frame overhead in Bytes | 0 | 0 | /// | Stack fixup required | No | No | /// +-------------------------+--------+-----+ /// /// \p MachineOutlinerThunk implies that the function is being created from /// a sequence of instructions ending in a call. The outlined function is /// called with a BL instruction, and the outlined function tail-calls the /// original call destination. /// /// That is, /// /// I1 OUTLINED_FUNCTION: /// I2 --> BL OUTLINED_FUNCTION I1 /// BL f I2 /// B f /// /// +-------------------------+--------+-----+ /// | | Thumb2 | ARM | /// +-------------------------+--------+-----+ /// | Call overhead in Bytes | 4 | 4 | /// | Frame overhead in Bytes | 0 | 0 | /// | Stack fixup required | No | No | /// +-------------------------+--------+-----+ /// /// \p MachineOutlinerNoLRSave implies that the function should be called using /// a BL instruction, but doesn't require LR to be saved and restored. This /// happens when LR is known to be dead. /// /// That is, /// /// I1 OUTLINED_FUNCTION: /// I2 --> BL OUTLINED_FUNCTION I1 /// I3 I2 /// I3 /// BX LR /// /// +-------------------------+--------+-----+ /// | | Thumb2 | ARM | /// +-------------------------+--------+-----+ /// | Call overhead in Bytes | 4 | 4 | /// | Frame overhead in Bytes | 2 | 4 | /// | Stack fixup required | No | No | /// +-------------------------+--------+-----+ /// /// \p MachineOutlinerRegSave implies that the function should be called with a /// save and restore of LR to an available register. This allows us to avoid /// stack fixups. Note that this outlining variant is compatible with the /// NoLRSave case. /// /// That is, /// /// I1 Save LR OUTLINED_FUNCTION: /// I2 --> BL OUTLINED_FUNCTION I1 /// I3 Restore LR I2 /// I3 /// BX LR /// /// +-------------------------+--------+-----+ /// | | Thumb2 | ARM | /// +-------------------------+--------+-----+ /// | Call overhead in Bytes | 8 | 12 | /// | Frame overhead in Bytes | 2 | 4 | /// | Stack fixup required | No | No | /// +-------------------------+--------+-----+ /// /// \p MachineOutlinerDefault implies that the function should be called with /// a save and restore of LR to the stack. /// /// That is, /// /// I1 Save LR OUTLINED_FUNCTION: /// I2 --> BL OUTLINED_FUNCTION I1 /// I3 Restore LR I2 /// I3 /// BX LR /// /// +-------------------------+--------+-----+ /// | | Thumb2 | ARM | /// +-------------------------+--------+-----+ /// | Call overhead in Bytes | 8 | 12 | /// | Frame overhead in Bytes | 2 | 4 | /// | Stack fixup required | Yes | Yes | /// +-------------------------+--------+-----+ enum MachineOutlinerClass { MachineOutlinerTailCall, MachineOutlinerThunk, MachineOutlinerNoLRSave, MachineOutlinerRegSave, MachineOutlinerDefault }; enum MachineOutlinerMBBFlags { LRUnavailableSomewhere = 0x2, HasCalls = 0x4, UnsafeRegsDead = 0x8 }; struct OutlinerCosts { int CallTailCall; int FrameTailCall; int CallThunk; int FrameThunk; int CallNoLRSave; int FrameNoLRSave; int CallRegSave; int FrameRegSave; int CallDefault; int FrameDefault; int SaveRestoreLROnStack; OutlinerCosts(const ARMSubtarget &target) : CallTailCall(target.isThumb() ? 4 : 4), FrameTailCall(target.isThumb() ? 0 : 0), CallThunk(target.isThumb() ? 4 : 4), FrameThunk(target.isThumb() ? 0 : 0), CallNoLRSave(target.isThumb() ? 4 : 4), FrameNoLRSave(target.isThumb() ? 2 : 4), CallRegSave(target.isThumb() ? 8 : 12), FrameRegSave(target.isThumb() ? 2 : 4), CallDefault(target.isThumb() ? 8 : 12), FrameDefault(target.isThumb() ? 2 : 4), SaveRestoreLROnStack(target.isThumb() ? 8 : 8) {} }; Register ARMBaseInstrInfo::findRegisterToSaveLRTo(outliner::Candidate &C) const { MachineFunction *MF = C.getMF(); const TargetRegisterInfo &TRI = *MF->getSubtarget().getRegisterInfo(); const ARMBaseRegisterInfo *ARI = static_cast(&TRI); BitVector regsReserved = ARI->getReservedRegs(*MF); // Check if there is an available register across the sequence that we can // use. for (Register Reg : ARM::rGPRRegClass) { if (!(Reg < regsReserved.size() && regsReserved.test(Reg)) && Reg != ARM::LR && // LR is not reserved, but don't use it. Reg != ARM::R12 && // R12 is not guaranteed to be preserved. C.isAvailableAcrossAndOutOfSeq(Reg, TRI) && C.isAvailableInsideSeq(Reg, TRI)) return Reg; } return Register(); } // Compute liveness of LR at the point after the interval [I, E), which // denotes a *backward* iteration through instructions. Used only for return // basic blocks, which do not end with a tail call. static bool isLRAvailable(const TargetRegisterInfo &TRI, MachineBasicBlock::reverse_iterator I, MachineBasicBlock::reverse_iterator E) { // At the end of the function LR dead. bool Live = false; for (; I != E; ++I) { const MachineInstr &MI = *I; // Check defs of LR. if (MI.modifiesRegister(ARM::LR, &TRI)) Live = false; // Check uses of LR. unsigned Opcode = MI.getOpcode(); if (Opcode == ARM::BX_RET || Opcode == ARM::MOVPCLR || Opcode == ARM::SUBS_PC_LR || Opcode == ARM::tBX_RET || Opcode == ARM::tBXNS_RET) { // These instructions use LR, but it's not an (explicit or implicit) // operand. Live = true; continue; } if (MI.readsRegister(ARM::LR, &TRI)) Live = true; } return !Live; } std::optional ARMBaseInstrInfo::getOutliningCandidateInfo( std::vector &RepeatedSequenceLocs) const { outliner::Candidate &FirstCand = RepeatedSequenceLocs[0]; unsigned SequenceSize = std::accumulate(FirstCand.front(), std::next(FirstCand.back()), 0, [this](unsigned Sum, const MachineInstr &MI) { return Sum + getInstSizeInBytes(MI); }); // Properties about candidate MBBs that hold for all of them. unsigned FlagsSetInAll = 0xF; // Compute liveness information for each candidate, and set FlagsSetInAll. const TargetRegisterInfo &TRI = getRegisterInfo(); for (outliner::Candidate &C : RepeatedSequenceLocs) FlagsSetInAll &= C.Flags; // According to the ARM Procedure Call Standard, the following are // undefined on entry/exit from a function call: // // * Register R12(IP), // * Condition codes (and thus the CPSR register) // // Since we control the instructions which are part of the outlined regions // we don't need to be fully compliant with the AAPCS, but we have to // guarantee that if a veneer is inserted at link time the code is still // correct. Because of this, we can't outline any sequence of instructions // where one of these registers is live into/across it. Thus, we need to // delete those candidates. auto CantGuaranteeValueAcrossCall = [&TRI](outliner::Candidate &C) { // If the unsafe registers in this block are all dead, then we don't need // to compute liveness here. if (C.Flags & UnsafeRegsDead) return false; return C.isAnyUnavailableAcrossOrOutOfSeq({ARM::R12, ARM::CPSR}, TRI); }; // Are there any candidates where those registers are live? if (!(FlagsSetInAll & UnsafeRegsDead)) { // Erase every candidate that violates the restrictions above. (It could be // true that we have viable candidates, so it's not worth bailing out in // the case that, say, 1 out of 20 candidates violate the restructions.) llvm::erase_if(RepeatedSequenceLocs, CantGuaranteeValueAcrossCall); // If the sequence doesn't have enough candidates left, then we're done. if (RepeatedSequenceLocs.size() < 2) return std::nullopt; } // We expect the majority of the outlining candidates to be in consensus with // regard to return address sign and authentication, and branch target // enforcement, in other words, partitioning according to all the four // possible combinations of PAC-RET and BTI is going to yield one big subset // and three small (likely empty) subsets. That allows us to cull incompatible // candidates separately for PAC-RET and BTI. // Partition the candidates in two sets: one with BTI enabled and one with BTI // disabled. Remove the candidates from the smaller set. If they are the same // number prefer the non-BTI ones for outlining, since they have less // overhead. auto NoBTI = llvm::partition(RepeatedSequenceLocs, [](const outliner::Candidate &C) { const ARMFunctionInfo &AFI = *C.getMF()->getInfo(); return AFI.branchTargetEnforcement(); }); if (std::distance(RepeatedSequenceLocs.begin(), NoBTI) > std::distance(NoBTI, RepeatedSequenceLocs.end())) RepeatedSequenceLocs.erase(NoBTI, RepeatedSequenceLocs.end()); else RepeatedSequenceLocs.erase(RepeatedSequenceLocs.begin(), NoBTI); if (RepeatedSequenceLocs.size() < 2) return std::nullopt; // Likewise, partition the candidates according to PAC-RET enablement. auto NoPAC = llvm::partition(RepeatedSequenceLocs, [](const outliner::Candidate &C) { const ARMFunctionInfo &AFI = *C.getMF()->getInfo(); // If the function happens to not spill the LR, do not disqualify it // from the outlining. return AFI.shouldSignReturnAddress(true); }); if (std::distance(RepeatedSequenceLocs.begin(), NoPAC) > std::distance(NoPAC, RepeatedSequenceLocs.end())) RepeatedSequenceLocs.erase(NoPAC, RepeatedSequenceLocs.end()); else RepeatedSequenceLocs.erase(RepeatedSequenceLocs.begin(), NoPAC); if (RepeatedSequenceLocs.size() < 2) return std::nullopt; // At this point, we have only "safe" candidates to outline. Figure out // frame + call instruction information. unsigned LastInstrOpcode = RepeatedSequenceLocs[0].back()->getOpcode(); // Helper lambda which sets call information for every candidate. auto SetCandidateCallInfo = [&RepeatedSequenceLocs](unsigned CallID, unsigned NumBytesForCall) { for (outliner::Candidate &C : RepeatedSequenceLocs) C.setCallInfo(CallID, NumBytesForCall); }; OutlinerCosts Costs(Subtarget); const auto &SomeMFI = *RepeatedSequenceLocs.front().getMF()->getInfo(); // Adjust costs to account for the BTI instructions. if (SomeMFI.branchTargetEnforcement()) { Costs.FrameDefault += 4; Costs.FrameNoLRSave += 4; Costs.FrameRegSave += 4; Costs.FrameTailCall += 4; Costs.FrameThunk += 4; } // Adjust costs to account for sign and authentication instructions. if (SomeMFI.shouldSignReturnAddress(true)) { Costs.CallDefault += 8; // +PAC instr, +AUT instr Costs.SaveRestoreLROnStack += 8; // +PAC instr, +AUT instr } unsigned FrameID = MachineOutlinerDefault; unsigned NumBytesToCreateFrame = Costs.FrameDefault; // If the last instruction in any candidate is a terminator, then we should // tail call all of the candidates. if (RepeatedSequenceLocs[0].back()->isTerminator()) { FrameID = MachineOutlinerTailCall; NumBytesToCreateFrame = Costs.FrameTailCall; SetCandidateCallInfo(MachineOutlinerTailCall, Costs.CallTailCall); } else if (LastInstrOpcode == ARM::BL || LastInstrOpcode == ARM::BLX || LastInstrOpcode == ARM::BLX_noip || LastInstrOpcode == ARM::tBL || LastInstrOpcode == ARM::tBLXr || LastInstrOpcode == ARM::tBLXr_noip || LastInstrOpcode == ARM::tBLXi) { FrameID = MachineOutlinerThunk; NumBytesToCreateFrame = Costs.FrameThunk; SetCandidateCallInfo(MachineOutlinerThunk, Costs.CallThunk); } else { // We need to decide how to emit calls + frames. We can always emit the same // frame if we don't need to save to the stack. If we have to save to the // stack, then we need a different frame. unsigned NumBytesNoStackCalls = 0; std::vector CandidatesWithoutStackFixups; for (outliner::Candidate &C : RepeatedSequenceLocs) { // LR liveness is overestimated in return blocks, unless they end with a // tail call. const auto Last = C.getMBB()->rbegin(); const bool LRIsAvailable = C.getMBB()->isReturnBlock() && !Last->isCall() ? isLRAvailable(TRI, Last, (MachineBasicBlock::reverse_iterator)C.front()) : C.isAvailableAcrossAndOutOfSeq(ARM::LR, TRI); if (LRIsAvailable) { FrameID = MachineOutlinerNoLRSave; NumBytesNoStackCalls += Costs.CallNoLRSave; C.setCallInfo(MachineOutlinerNoLRSave, Costs.CallNoLRSave); CandidatesWithoutStackFixups.push_back(C); } // Is an unused register available? If so, we won't modify the stack, so // we can outline with the same frame type as those that don't save LR. else if (findRegisterToSaveLRTo(C)) { FrameID = MachineOutlinerRegSave; NumBytesNoStackCalls += Costs.CallRegSave; C.setCallInfo(MachineOutlinerRegSave, Costs.CallRegSave); CandidatesWithoutStackFixups.push_back(C); } // Is SP used in the sequence at all? If not, we don't have to modify // the stack, so we are guaranteed to get the same frame. else if (C.isAvailableInsideSeq(ARM::SP, TRI)) { NumBytesNoStackCalls += Costs.CallDefault; C.setCallInfo(MachineOutlinerDefault, Costs.CallDefault); CandidatesWithoutStackFixups.push_back(C); } // If we outline this, we need to modify the stack. Pretend we don't // outline this by saving all of its bytes. else NumBytesNoStackCalls += SequenceSize; } // If there are no places where we have to save LR, then note that we don't // have to update the stack. Otherwise, give every candidate the default // call type if (NumBytesNoStackCalls <= RepeatedSequenceLocs.size() * Costs.CallDefault) { RepeatedSequenceLocs = CandidatesWithoutStackFixups; FrameID = MachineOutlinerNoLRSave; } else SetCandidateCallInfo(MachineOutlinerDefault, Costs.CallDefault); } // Does every candidate's MBB contain a call? If so, then we might have a // call in the range. if (FlagsSetInAll & MachineOutlinerMBBFlags::HasCalls) { // check if the range contains a call. These require a save + restore of // the link register. if (std::any_of(FirstCand.front(), FirstCand.back(), [](const MachineInstr &MI) { return MI.isCall(); })) NumBytesToCreateFrame += Costs.SaveRestoreLROnStack; // Handle the last instruction separately. If it is tail call, then the // last instruction is a call, we don't want to save + restore in this // case. However, it could be possible that the last instruction is a // call without it being valid to tail call this sequence. We should // consider this as well. else if (FrameID != MachineOutlinerThunk && FrameID != MachineOutlinerTailCall && FirstCand.back()->isCall()) NumBytesToCreateFrame += Costs.SaveRestoreLROnStack; } return outliner::OutlinedFunction(RepeatedSequenceLocs, SequenceSize, NumBytesToCreateFrame, FrameID); } bool ARMBaseInstrInfo::checkAndUpdateStackOffset(MachineInstr *MI, int64_t Fixup, bool Updt) const { int SPIdx = MI->findRegisterUseOperandIdx(ARM::SP); unsigned AddrMode = (MI->getDesc().TSFlags & ARMII::AddrModeMask); if (SPIdx < 0) // No SP operand return true; else if (SPIdx != 1 && (AddrMode != ARMII::AddrModeT2_i8s4 || SPIdx != 2)) // If SP is not the base register we can't do much return false; // Stack might be involved but addressing mode doesn't handle any offset. // Rq: AddrModeT1_[1|2|4] don't operate on SP if (AddrMode == ARMII::AddrMode1 || // Arithmetic instructions AddrMode == ARMII::AddrMode4 || // Load/Store Multiple AddrMode == ARMII::AddrMode6 || // Neon Load/Store Multiple AddrMode == ARMII::AddrModeT2_so || // SP can't be used as based register AddrMode == ARMII::AddrModeT2_pc || // PCrel access AddrMode == ARMII::AddrMode2 || // Used by PRE and POST indexed LD/ST AddrMode == ARMII::AddrModeT2_i7 || // v8.1-M MVE AddrMode == ARMII::AddrModeT2_i7s2 || // v8.1-M MVE AddrMode == ARMII::AddrModeT2_i7s4 || // v8.1-M sys regs VLDR/VSTR AddrMode == ARMII::AddrModeNone || AddrMode == ARMII::AddrModeT2_i8 || // Pre/Post inc instructions AddrMode == ARMII::AddrModeT2_i8neg) // Always negative imm return false; unsigned NumOps = MI->getDesc().getNumOperands(); unsigned ImmIdx = NumOps - 3; const MachineOperand &Offset = MI->getOperand(ImmIdx); assert(Offset.isImm() && "Is not an immediate"); int64_t OffVal = Offset.getImm(); if (OffVal < 0) // Don't override data if the are below SP. return false; unsigned NumBits = 0; unsigned Scale = 1; switch (AddrMode) { case ARMII::AddrMode3: if (ARM_AM::getAM3Op(OffVal) == ARM_AM::sub) return false; OffVal = ARM_AM::getAM3Offset(OffVal); NumBits = 8; break; case ARMII::AddrMode5: if (ARM_AM::getAM5Op(OffVal) == ARM_AM::sub) return false; OffVal = ARM_AM::getAM5Offset(OffVal); NumBits = 8; Scale = 4; break; case ARMII::AddrMode5FP16: if (ARM_AM::getAM5FP16Op(OffVal) == ARM_AM::sub) return false; OffVal = ARM_AM::getAM5FP16Offset(OffVal); NumBits = 8; Scale = 2; break; case ARMII::AddrModeT2_i8pos: NumBits = 8; break; case ARMII::AddrModeT2_i8s4: // FIXME: Values are already scaled in this addressing mode. assert((Fixup & 3) == 0 && "Can't encode this offset!"); NumBits = 10; break; case ARMII::AddrModeT2_ldrex: NumBits = 8; Scale = 4; break; case ARMII::AddrModeT2_i12: case ARMII::AddrMode_i12: NumBits = 12; break; case ARMII::AddrModeT1_s: // SP-relative LD/ST NumBits = 8; Scale = 4; break; default: llvm_unreachable("Unsupported addressing mode!"); } // Make sure the offset is encodable for instructions that scale the // immediate. assert(((OffVal * Scale + Fixup) & (Scale - 1)) == 0 && "Can't encode this offset!"); OffVal += Fixup / Scale; unsigned Mask = (1 << NumBits) - 1; if (OffVal <= Mask) { if (Updt) MI->getOperand(ImmIdx).setImm(OffVal); return true; } return false; } void ARMBaseInstrInfo::mergeOutliningCandidateAttributes( Function &F, std::vector &Candidates) const { outliner::Candidate &C = Candidates.front(); // branch-target-enforcement is guaranteed to be consistent between all // candidates, so we only need to look at one. const Function &CFn = C.getMF()->getFunction(); if (CFn.hasFnAttribute("branch-target-enforcement")) F.addFnAttr(CFn.getFnAttribute("branch-target-enforcement")); ARMGenInstrInfo::mergeOutliningCandidateAttributes(F, Candidates); } bool ARMBaseInstrInfo::isFunctionSafeToOutlineFrom( MachineFunction &MF, bool OutlineFromLinkOnceODRs) const { const Function &F = MF.getFunction(); // Can F be deduplicated by the linker? If it can, don't outline from it. if (!OutlineFromLinkOnceODRs && F.hasLinkOnceODRLinkage()) return false; // Don't outline from functions with section markings; the program could // expect that all the code is in the named section. // FIXME: Allow outlining from multiple functions with the same section // marking. if (F.hasSection()) return false; // FIXME: Thumb1 outlining is not handled if (MF.getInfo()->isThumb1OnlyFunction()) return false; // It's safe to outline from MF. return true; } bool ARMBaseInstrInfo::isMBBSafeToOutlineFrom(MachineBasicBlock &MBB, unsigned &Flags) const { // Check if LR is available through all of the MBB. If it's not, then set // a flag. assert(MBB.getParent()->getRegInfo().tracksLiveness() && "Suitable Machine Function for outlining must track liveness"); LiveRegUnits LRU(getRegisterInfo()); for (MachineInstr &MI : llvm::reverse(MBB)) LRU.accumulate(MI); // Check if each of the unsafe registers are available... bool R12AvailableInBlock = LRU.available(ARM::R12); bool CPSRAvailableInBlock = LRU.available(ARM::CPSR); // If all of these are dead (and not live out), we know we don't have to check // them later. if (R12AvailableInBlock && CPSRAvailableInBlock) Flags |= MachineOutlinerMBBFlags::UnsafeRegsDead; // Now, add the live outs to the set. LRU.addLiveOuts(MBB); // If any of these registers is available in the MBB, but also a live out of // the block, then we know outlining is unsafe. if (R12AvailableInBlock && !LRU.available(ARM::R12)) return false; if (CPSRAvailableInBlock && !LRU.available(ARM::CPSR)) return false; // Check if there's a call inside this MachineBasicBlock. If there is, then // set a flag. if (any_of(MBB, [](MachineInstr &MI) { return MI.isCall(); })) Flags |= MachineOutlinerMBBFlags::HasCalls; // LR liveness is overestimated in return blocks. bool LRIsAvailable = MBB.isReturnBlock() && !MBB.back().isCall() ? isLRAvailable(getRegisterInfo(), MBB.rbegin(), MBB.rend()) : LRU.available(ARM::LR); if (!LRIsAvailable) Flags |= MachineOutlinerMBBFlags::LRUnavailableSomewhere; return true; } outliner::InstrType ARMBaseInstrInfo::getOutliningTypeImpl(MachineBasicBlock::iterator &MIT, unsigned Flags) const { MachineInstr &MI = *MIT; const TargetRegisterInfo *TRI = &getRegisterInfo(); // PIC instructions contain labels, outlining them would break offset // computing. unsigned Opc = MI.getOpcode(); unsigned Opc = MI.getOpcode(); if (Opc == ARM::tPICADD || Opc == ARM::PICADD || Opc == ARM::PICSTR || Opc == ARM::PICSTRB || Opc == ARM::PICSTRH || Opc == ARM::PICLDR || Opc == ARM::PICLDRB || Opc == ARM::PICLDRH || Opc == ARM::PICLDRSB || Opc == ARM::PICLDRSH || Opc == ARM::t2LDRpci_pic || Opc == ARM::t2MOVi16_ga_pcrel || Opc == ARM::t2MOVTi16_ga_pcrel || Opc == ARM::t2MOV_ga_pcrel) return outliner::InstrType::Illegal; // Be conservative with ARMv8.1 MVE instructions. if (Opc == ARM::t2BF_LabelPseudo || Opc == ARM::t2DoLoopStart || Opc == ARM::t2DoLoopStartTP || Opc == ARM::t2WhileLoopStart || Opc == ARM::t2WhileLoopStartLR || Opc == ARM::t2WhileLoopStartTP || Opc == ARM::t2LoopDec || Opc == ARM::t2LoopEnd || Opc == ARM::t2LoopEndDec) return outliner::InstrType::Illegal; const MCInstrDesc &MCID = MI.getDesc(); uint64_t MIFlags = MCID.TSFlags; if ((MIFlags & ARMII::DomainMask) == ARMII::DomainMVE) return outliner::InstrType::Illegal; // Is this a terminator for a basic block? if (MI.isTerminator()) // TargetInstrInfo::getOutliningType has already filtered out anything // that would break this, so we can allow it here. return outliner::InstrType::Legal; // Don't outline if link register or program counter value are used. if (MI.readsRegister(ARM::LR, TRI) || MI.readsRegister(ARM::PC, TRI)) return outliner::InstrType::Illegal; if (MI.isCall()) { // Get the function associated with the call. Look at each operand and find // the one that represents the calle and get its name. const Function *Callee = nullptr; for (const MachineOperand &MOP : MI.operands()) { if (MOP.isGlobal()) { Callee = dyn_cast(MOP.getGlobal()); break; } } // Dont't outline calls to "mcount" like functions, in particular Linux // kernel function tracing relies on it. if (Callee && (Callee->getName() == "\01__gnu_mcount_nc" || Callee->getName() == "\01mcount" || Callee->getName() == "__mcount")) return outliner::InstrType::Illegal; // If we don't know anything about the callee, assume it depends on the // stack layout of the caller. In that case, it's only legal to outline // as a tail-call. Explicitly list the call instructions we know about so // we don't get unexpected results with call pseudo-instructions. auto UnknownCallOutlineType = outliner::InstrType::Illegal; if (Opc == ARM::BL || Opc == ARM::tBL || Opc == ARM::BLX || Opc == ARM::BLX_noip || Opc == ARM::tBLXr || Opc == ARM::tBLXr_noip || Opc == ARM::tBLXi) UnknownCallOutlineType = outliner::InstrType::LegalTerminator; if (!Callee) return UnknownCallOutlineType; // We have a function we have information about. Check if it's something we // can safely outline. MachineFunction *MF = MI.getParent()->getParent(); MachineFunction *CalleeMF = MF->getMMI().getMachineFunction(*Callee); // We don't know what's going on with the callee at all. Don't touch it. if (!CalleeMF) return UnknownCallOutlineType; // Check if we know anything about the callee saves on the function. If we // don't, then don't touch it, since that implies that we haven't computed // anything about its stack frame yet. MachineFrameInfo &MFI = CalleeMF->getFrameInfo(); if (!MFI.isCalleeSavedInfoValid() || MFI.getStackSize() > 0 || MFI.getNumObjects() > 0) return UnknownCallOutlineType; // At this point, we can say that CalleeMF ought to not pass anything on the // stack. Therefore, we can outline it. return outliner::InstrType::Legal; } // Since calls are handled, don't touch LR or PC if (MI.modifiesRegister(ARM::LR, TRI) || MI.modifiesRegister(ARM::PC, TRI)) return outliner::InstrType::Illegal; // Does this use the stack? if (MI.modifiesRegister(ARM::SP, TRI) || MI.readsRegister(ARM::SP, TRI)) { // True if there is no chance that any outlined candidate from this range // could require stack fixups. That is, both // * LR is available in the range (No save/restore around call) // * The range doesn't include calls (No save/restore in outlined frame) // are true. // These conditions also ensure correctness of the return address // authentication - we insert sign and authentication instructions only if // we save/restore LR on stack, but then this condition ensures that the // outlined range does not modify the SP, therefore the SP value used for // signing is the same as the one used for authentication. // FIXME: This is very restrictive; the flags check the whole block, // not just the bit we will try to outline. bool MightNeedStackFixUp = (Flags & (MachineOutlinerMBBFlags::LRUnavailableSomewhere | MachineOutlinerMBBFlags::HasCalls)); if (!MightNeedStackFixUp) return outliner::InstrType::Legal; // Any modification of SP will break our code to save/restore LR. // FIXME: We could handle some instructions which add a constant offset to // SP, with a bit more work. if (MI.modifiesRegister(ARM::SP, TRI)) return outliner::InstrType::Illegal; // At this point, we have a stack instruction that we might need to fix up. // up. We'll handle it if it's a load or store. if (checkAndUpdateStackOffset(&MI, Subtarget.getStackAlignment().value(), false)) return outliner::InstrType::Legal; // We can't fix it up, so don't outline it. return outliner::InstrType::Illegal; } // Be conservative with IT blocks. if (MI.readsRegister(ARM::ITSTATE, TRI) || MI.modifiesRegister(ARM::ITSTATE, TRI)) return outliner::InstrType::Illegal; // Don't outline CFI instructions. if (MI.isCFIInstruction()) return outliner::InstrType::Illegal; return outliner::InstrType::Legal; } void ARMBaseInstrInfo::fixupPostOutline(MachineBasicBlock &MBB) const { for (MachineInstr &MI : MBB) { checkAndUpdateStackOffset(&MI, Subtarget.getStackAlignment().value(), true); } } void ARMBaseInstrInfo::saveLROnStack(MachineBasicBlock &MBB, MachineBasicBlock::iterator It, bool CFI, bool Auth) const { int Align = std::max(Subtarget.getStackAlignment().value(), uint64_t(8)); assert(Align >= 8 && Align <= 256); if (Auth) { assert(Subtarget.isThumb2()); // Compute PAC in R12. Outlining ensures R12 is dead across the outlined // sequence. BuildMI(MBB, It, DebugLoc(), get(ARM::t2PAC)) .setMIFlags(MachineInstr::FrameSetup); BuildMI(MBB, It, DebugLoc(), get(ARM::t2STRD_PRE), ARM::SP) .addReg(ARM::R12, RegState::Kill) .addReg(ARM::LR, RegState::Kill) .addReg(ARM::SP) .addImm(-Align) .add(predOps(ARMCC::AL)) .setMIFlags(MachineInstr::FrameSetup); } else { unsigned Opc = Subtarget.isThumb() ? ARM::t2STR_PRE : ARM::STR_PRE_IMM; BuildMI(MBB, It, DebugLoc(), get(Opc), ARM::SP) .addReg(ARM::LR, RegState::Kill) .addReg(ARM::SP) .addImm(-Align) .add(predOps(ARMCC::AL)) .setMIFlags(MachineInstr::FrameSetup); } if (!CFI) return; MachineFunction &MF = *MBB.getParent(); // Add a CFI, saying CFA is offset by Align bytes from SP. int64_t StackPosEntry = MF.addFrameInst(MCCFIInstruction::cfiDefCfaOffset(nullptr, Align)); BuildMI(MBB, It, DebugLoc(), get(ARM::CFI_INSTRUCTION)) .addCFIIndex(StackPosEntry) .setMIFlags(MachineInstr::FrameSetup); // Add a CFI saying that the LR that we want to find is now higher than // before. int LROffset = Auth ? Align - 4 : Align; const MCRegisterInfo *MRI = Subtarget.getRegisterInfo(); unsigned DwarfLR = MRI->getDwarfRegNum(ARM::LR, true); int64_t LRPosEntry = MF.addFrameInst( MCCFIInstruction::createOffset(nullptr, DwarfLR, -LROffset)); BuildMI(MBB, It, DebugLoc(), get(ARM::CFI_INSTRUCTION)) .addCFIIndex(LRPosEntry) .setMIFlags(MachineInstr::FrameSetup); if (Auth) { // Add a CFI for the location of the return adddress PAC. unsigned DwarfRAC = MRI->getDwarfRegNum(ARM::RA_AUTH_CODE, true); int64_t RACPosEntry = MF.addFrameInst( MCCFIInstruction::createOffset(nullptr, DwarfRAC, -Align)); BuildMI(MBB, It, DebugLoc(), get(ARM::CFI_INSTRUCTION)) .addCFIIndex(RACPosEntry) .setMIFlags(MachineInstr::FrameSetup); } } void ARMBaseInstrInfo::emitCFIForLRSaveToReg(MachineBasicBlock &MBB, MachineBasicBlock::iterator It, Register Reg) const { MachineFunction &MF = *MBB.getParent(); const MCRegisterInfo *MRI = Subtarget.getRegisterInfo(); unsigned DwarfLR = MRI->getDwarfRegNum(ARM::LR, true); unsigned DwarfReg = MRI->getDwarfRegNum(Reg, true); int64_t LRPosEntry = MF.addFrameInst( MCCFIInstruction::createRegister(nullptr, DwarfLR, DwarfReg)); BuildMI(MBB, It, DebugLoc(), get(ARM::CFI_INSTRUCTION)) .addCFIIndex(LRPosEntry) .setMIFlags(MachineInstr::FrameSetup); } void ARMBaseInstrInfo::restoreLRFromStack(MachineBasicBlock &MBB, MachineBasicBlock::iterator It, bool CFI, bool Auth) const { int Align = Subtarget.getStackAlignment().value(); if (Auth) { assert(Subtarget.isThumb2()); // Restore return address PAC and LR. BuildMI(MBB, It, DebugLoc(), get(ARM::t2LDRD_POST)) .addReg(ARM::R12, RegState::Define) .addReg(ARM::LR, RegState::Define) .addReg(ARM::SP, RegState::Define) .addReg(ARM::SP) .addImm(Align) .add(predOps(ARMCC::AL)) .setMIFlags(MachineInstr::FrameDestroy); // LR authentication is after the CFI instructions, below. } else { unsigned Opc = Subtarget.isThumb() ? ARM::t2LDR_POST : ARM::LDR_POST_IMM; MachineInstrBuilder MIB = BuildMI(MBB, It, DebugLoc(), get(Opc), ARM::LR) .addReg(ARM::SP, RegState::Define) .addReg(ARM::SP); if (!Subtarget.isThumb()) MIB.addReg(0); MIB.addImm(Subtarget.getStackAlignment().value()) .add(predOps(ARMCC::AL)) .setMIFlags(MachineInstr::FrameDestroy); } if (CFI) { // Now stack has moved back up... MachineFunction &MF = *MBB.getParent(); const MCRegisterInfo *MRI = Subtarget.getRegisterInfo(); unsigned DwarfLR = MRI->getDwarfRegNum(ARM::LR, true); int64_t StackPosEntry = MF.addFrameInst(MCCFIInstruction::cfiDefCfaOffset(nullptr, 0)); BuildMI(MBB, It, DebugLoc(), get(ARM::CFI_INSTRUCTION)) .addCFIIndex(StackPosEntry) .setMIFlags(MachineInstr::FrameDestroy); // ... and we have restored LR. int64_t LRPosEntry = MF.addFrameInst(MCCFIInstruction::createRestore(nullptr, DwarfLR)); BuildMI(MBB, It, DebugLoc(), get(ARM::CFI_INSTRUCTION)) .addCFIIndex(LRPosEntry) .setMIFlags(MachineInstr::FrameDestroy); if (Auth) { unsigned DwarfRAC = MRI->getDwarfRegNum(ARM::RA_AUTH_CODE, true); int64_t Entry = MF.addFrameInst(MCCFIInstruction::createUndefined(nullptr, DwarfRAC)); BuildMI(MBB, It, DebugLoc(), get(ARM::CFI_INSTRUCTION)) .addCFIIndex(Entry) .setMIFlags(MachineInstr::FrameDestroy); } } if (Auth) BuildMI(MBB, It, DebugLoc(), get(ARM::t2AUT)); } void ARMBaseInstrInfo::emitCFIForLRRestoreFromReg( MachineBasicBlock &MBB, MachineBasicBlock::iterator It) const { MachineFunction &MF = *MBB.getParent(); const MCRegisterInfo *MRI = Subtarget.getRegisterInfo(); unsigned DwarfLR = MRI->getDwarfRegNum(ARM::LR, true); int64_t LRPosEntry = MF.addFrameInst(MCCFIInstruction::createRestore(nullptr, DwarfLR)); BuildMI(MBB, It, DebugLoc(), get(ARM::CFI_INSTRUCTION)) .addCFIIndex(LRPosEntry) .setMIFlags(MachineInstr::FrameDestroy); } void ARMBaseInstrInfo::buildOutlinedFrame( MachineBasicBlock &MBB, MachineFunction &MF, const outliner::OutlinedFunction &OF) const { // For thunk outlining, rewrite the last instruction from a call to a // tail-call. if (OF.FrameConstructionID == MachineOutlinerThunk) { MachineInstr *Call = &*--MBB.instr_end(); bool isThumb = Subtarget.isThumb(); unsigned FuncOp = isThumb ? 2 : 0; unsigned Opc = Call->getOperand(FuncOp).isReg() ? isThumb ? ARM::tTAILJMPr : ARM::TAILJMPr : isThumb ? Subtarget.isTargetMachO() ? ARM::tTAILJMPd : ARM::tTAILJMPdND : ARM::TAILJMPd; MachineInstrBuilder MIB = BuildMI(MBB, MBB.end(), DebugLoc(), get(Opc)) .add(Call->getOperand(FuncOp)); if (isThumb && !Call->getOperand(FuncOp).isReg()) MIB.add(predOps(ARMCC::AL)); Call->eraseFromParent(); } // Is there a call in the outlined range? auto IsNonTailCall = [](MachineInstr &MI) { return MI.isCall() && !MI.isReturn(); }; if (llvm::any_of(MBB.instrs(), IsNonTailCall)) { MachineBasicBlock::iterator It = MBB.begin(); MachineBasicBlock::iterator Et = MBB.end(); if (OF.FrameConstructionID == MachineOutlinerTailCall || OF.FrameConstructionID == MachineOutlinerThunk) Et = std::prev(MBB.end()); // We have to save and restore LR, we need to add it to the liveins if it // is not already part of the set. This is suffient since outlined // functions only have one block. if (!MBB.isLiveIn(ARM::LR)) MBB.addLiveIn(ARM::LR); // Insert a save before the outlined region bool Auth = OF.Candidates.front() .getMF() ->getInfo() ->shouldSignReturnAddress(true); saveLROnStack(MBB, It, true, Auth); // Fix up the instructions in the range, since we're going to modify the // stack. assert(OF.FrameConstructionID != MachineOutlinerDefault && "Can only fix up stack references once"); fixupPostOutline(MBB); // Insert a restore before the terminator for the function. Restore LR. restoreLRFromStack(MBB, Et, true, Auth); } // If this is a tail call outlined function, then there's already a return. if (OF.FrameConstructionID == MachineOutlinerTailCall || OF.FrameConstructionID == MachineOutlinerThunk) return; // Here we have to insert the return ourselves. Get the correct opcode from // current feature set. BuildMI(MBB, MBB.end(), DebugLoc(), get(Subtarget.getReturnOpcode())) .add(predOps(ARMCC::AL)); // Did we have to modify the stack by saving the link register? if (OF.FrameConstructionID != MachineOutlinerDefault && OF.Candidates[0].CallConstructionID != MachineOutlinerDefault) return; // We modified the stack. // Walk over the basic block and fix up all the stack accesses. fixupPostOutline(MBB); } MachineBasicBlock::iterator ARMBaseInstrInfo::insertOutlinedCall( Module &M, MachineBasicBlock &MBB, MachineBasicBlock::iterator &It, MachineFunction &MF, outliner::Candidate &C) const { MachineInstrBuilder MIB; MachineBasicBlock::iterator CallPt; unsigned Opc; bool isThumb = Subtarget.isThumb(); // Are we tail calling? if (C.CallConstructionID == MachineOutlinerTailCall) { // If yes, then we can just branch to the label. Opc = isThumb ? Subtarget.isTargetMachO() ? ARM::tTAILJMPd : ARM::tTAILJMPdND : ARM::TAILJMPd; MIB = BuildMI(MF, DebugLoc(), get(Opc)) .addGlobalAddress(M.getNamedValue(MF.getName())); if (isThumb) MIB.add(predOps(ARMCC::AL)); It = MBB.insert(It, MIB); return It; } // Create the call instruction. Opc = isThumb ? ARM::tBL : ARM::BL; MachineInstrBuilder CallMIB = BuildMI(MF, DebugLoc(), get(Opc)); if (isThumb) CallMIB.add(predOps(ARMCC::AL)); CallMIB.addGlobalAddress(M.getNamedValue(MF.getName())); if (C.CallConstructionID == MachineOutlinerNoLRSave || C.CallConstructionID == MachineOutlinerThunk) { // No, so just insert the call. It = MBB.insert(It, CallMIB); return It; } const ARMFunctionInfo &AFI = *C.getMF()->getInfo(); // Can we save to a register? if (C.CallConstructionID == MachineOutlinerRegSave) { Register Reg = findRegisterToSaveLRTo(C); assert(Reg != 0 && "No callee-saved register available?"); // Save and restore LR from that register. copyPhysReg(MBB, It, DebugLoc(), Reg, ARM::LR, true); if (!AFI.isLRSpilled()) emitCFIForLRSaveToReg(MBB, It, Reg); CallPt = MBB.insert(It, CallMIB); copyPhysReg(MBB, It, DebugLoc(), ARM::LR, Reg, true); if (!AFI.isLRSpilled()) emitCFIForLRRestoreFromReg(MBB, It); It--; return CallPt; } // We have the default case. Save and restore from SP. if (!MBB.isLiveIn(ARM::LR)) MBB.addLiveIn(ARM::LR); bool Auth = !AFI.isLRSpilled() && AFI.shouldSignReturnAddress(true); saveLROnStack(MBB, It, !AFI.isLRSpilled(), Auth); CallPt = MBB.insert(It, CallMIB); restoreLRFromStack(MBB, It, !AFI.isLRSpilled(), Auth); It--; return CallPt; } bool ARMBaseInstrInfo::shouldOutlineFromFunctionByDefault( MachineFunction &MF) const { return Subtarget.isMClass() && MF.getFunction().hasMinSize(); } bool ARMBaseInstrInfo::isReallyTriviallyReMaterializable( const MachineInstr &MI) const { // Try hard to rematerialize any VCTPs because if we spill P0, it will block // the tail predication conversion. This means that the element count // register has to be live for longer, but that has to be better than // spill/restore and VPT predication. return isVCTP(&MI) && !isPredicated(MI); } unsigned llvm::getBLXOpcode(const MachineFunction &MF) { return (MF.getSubtarget().hardenSlsBlr()) ? ARM::BLX_noip : ARM::BLX; } unsigned llvm::gettBLXrOpcode(const MachineFunction &MF) { return (MF.getSubtarget().hardenSlsBlr()) ? ARM::tBLXr_noip : ARM::tBLXr; } unsigned llvm::getBLXpredOpcode(const MachineFunction &MF) { return (MF.getSubtarget().hardenSlsBlr()) ? ARM::BLX_pred_noip : ARM::BLX_pred; } namespace { class ARMPipelinerLoopInfo : public TargetInstrInfo::PipelinerLoopInfo { MachineInstr *EndLoop, *LoopCount; MachineFunction *MF; const TargetInstrInfo *TII; // Bitset[0 .. MAX_STAGES-1] ... iterations needed // [LAST_IS_USE] : last reference to register in schedule is a use // [SEEN_AS_LIVE] : Normal pressure algorithm believes register is live static int constexpr MAX_STAGES = 30; static int constexpr LAST_IS_USE = MAX_STAGES; static int constexpr SEEN_AS_LIVE = MAX_STAGES + 1; typedef std::bitset IterNeed; typedef std::map IterNeeds; void bumpCrossIterationPressure(RegPressureTracker &RPT, const IterNeeds &CIN); bool tooMuchRegisterPressure(SwingSchedulerDAG &SSD, SMSchedule &SMS); // Meanings of the various stuff with loop types: // t2Bcc: // EndLoop = branch at end of original BB that will become a kernel // LoopCount = CC setter live into branch // t2LoopEnd: // EndLoop = branch at end of original BB // LoopCount = t2LoopDec public: ARMPipelinerLoopInfo(MachineInstr *EndLoop, MachineInstr *LoopCount) : EndLoop(EndLoop), LoopCount(LoopCount), MF(EndLoop->getParent()->getParent()), TII(MF->getSubtarget().getInstrInfo()) {} bool shouldIgnoreForPipelining(const MachineInstr *MI) const override { // Only ignore the terminator. return MI == EndLoop || MI == LoopCount; } bool shouldUseSchedule(SwingSchedulerDAG &SSD, SMSchedule &SMS) override { if (tooMuchRegisterPressure(SSD, SMS)) return false; return true; } std::optional createTripCountGreaterCondition( int TC, MachineBasicBlock &MBB, SmallVectorImpl &Cond) override { if (isCondBranchOpcode(EndLoop->getOpcode())) { Cond.push_back(EndLoop->getOperand(1)); Cond.push_back(EndLoop->getOperand(2)); if (EndLoop->getOperand(0).getMBB() == EndLoop->getParent()) { TII->reverseBranchCondition(Cond); } return {}; } else if (EndLoop->getOpcode() == ARM::t2LoopEnd) { // General case just lets the unrolled t2LoopDec do the subtraction and // therefore just needs to check if zero has been reached. MachineInstr *LoopDec = nullptr; for (auto &I : MBB.instrs()) if (I.getOpcode() == ARM::t2LoopDec) LoopDec = &I; assert(LoopDec && "Unable to find copied LoopDec"); // Check if we're done with the loop. BuildMI(&MBB, LoopDec->getDebugLoc(), TII->get(ARM::t2CMPri)) .addReg(LoopDec->getOperand(0).getReg()) .addImm(0) .addImm(ARMCC::AL) .addReg(ARM::NoRegister); Cond.push_back(MachineOperand::CreateImm(ARMCC::EQ)); Cond.push_back(MachineOperand::CreateReg(ARM::CPSR, false)); return {}; } else llvm_unreachable("Unknown EndLoop"); } void setPreheader(MachineBasicBlock *NewPreheader) override {} void adjustTripCount(int TripCountAdjust) override {} void disposed() override {} }; void ARMPipelinerLoopInfo::bumpCrossIterationPressure(RegPressureTracker &RPT, const IterNeeds &CIN) { // Increase pressure by the amounts in CrossIterationNeeds for (const auto &N : CIN) { int Cnt = N.second.count() - N.second[SEEN_AS_LIVE] * 2; for (int I = 0; I < Cnt; ++I) RPT.increaseRegPressure(Register(N.first), LaneBitmask::getNone(), LaneBitmask::getAll()); } // Decrease pressure by the amounts in CrossIterationNeeds for (const auto &N : CIN) { int Cnt = N.second.count() - N.second[SEEN_AS_LIVE] * 2; for (int I = 0; I < Cnt; ++I) RPT.decreaseRegPressure(Register(N.first), LaneBitmask::getAll(), LaneBitmask::getNone()); } } bool ARMPipelinerLoopInfo::tooMuchRegisterPressure(SwingSchedulerDAG &SSD, SMSchedule &SMS) { IterNeeds CrossIterationNeeds; // Determine which values will be loop-carried after the schedule is // applied for (auto &SU : SSD.SUnits) { const MachineInstr *MI = SU.getInstr(); int Stg = SMS.stageScheduled(const_cast(&SU)); for (auto &S : SU.Succs) if (MI->isPHI() && S.getKind() == SDep::Anti) { Register Reg = S.getReg(); if (Reg.isVirtual()) CrossIterationNeeds.insert(std::make_pair(Reg.id(), IterNeed())) .first->second.set(0); } else if (S.isAssignedRegDep()) { int OStg = SMS.stageScheduled(S.getSUnit()); if (OStg >= 0 && OStg != Stg) { Register Reg = S.getReg(); if (Reg.isVirtual()) CrossIterationNeeds.insert(std::make_pair(Reg.id(), IterNeed())) .first->second |= ((1 << (OStg - Stg)) - 1); } } } // Determine more-or-less what the proposed schedule (reversed) is going to // be; it might not be quite the same because the within-cycle ordering // created by SMSchedule depends upon changes to help with address offsets and // the like. std::vector ProposedSchedule; for (int Cycle = SMS.getFinalCycle(); Cycle >= SMS.getFirstCycle(); --Cycle) for (int Stage = 0, StageEnd = SMS.getMaxStageCount(); Stage <= StageEnd; ++Stage) { std::deque Instrs = SMS.getInstructions(Cycle + Stage * SMS.getInitiationInterval()); std::sort(Instrs.begin(), Instrs.end(), [](SUnit *A, SUnit *B) { return A->NodeNum > B->NodeNum; }); for (SUnit *SU : Instrs) ProposedSchedule.push_back(SU); } // Learn whether the last use/def of each cross-iteration register is a use or // def. If it is a def, RegisterPressure will implicitly increase max pressure // and we do not have to add the pressure. for (auto *SU : ProposedSchedule) for (ConstMIBundleOperands OperI(*SU->getInstr()); OperI.isValid(); ++OperI) { auto MO = *OperI; if (!MO.isReg() || !MO.getReg()) continue; Register Reg = MO.getReg(); auto CIter = CrossIterationNeeds.find(Reg.id()); if (CIter == CrossIterationNeeds.end() || CIter->second[LAST_IS_USE] || CIter->second[SEEN_AS_LIVE]) continue; if (MO.isDef() && !MO.isDead()) CIter->second.set(SEEN_AS_LIVE); else if (MO.isUse()) CIter->second.set(LAST_IS_USE); } for (auto &CI : CrossIterationNeeds) CI.second.reset(LAST_IS_USE); RegionPressure RecRegPressure; RegPressureTracker RPTracker(RecRegPressure); RegisterClassInfo RegClassInfo; RegClassInfo.runOnMachineFunction(*MF); RPTracker.init(MF, &RegClassInfo, nullptr, EndLoop->getParent(), EndLoop->getParent()->end(), false, false); const TargetRegisterInfo *TRI = MF->getSubtarget().getRegisterInfo(); bumpCrossIterationPressure(RPTracker, CrossIterationNeeds); for (auto *SU : ProposedSchedule) { MachineBasicBlock::const_iterator CurInstI = SU->getInstr(); RPTracker.setPos(std::next(CurInstI)); RPTracker.recede(); // Track what cross-iteration registers would be seen as live for (ConstMIBundleOperands OperI(*CurInstI); OperI.isValid(); ++OperI) { auto MO = *OperI; if (!MO.isReg() || !MO.getReg()) continue; Register Reg = MO.getReg(); if (MO.isDef() && !MO.isDead()) { auto CIter = CrossIterationNeeds.find(Reg.id()); if (CIter != CrossIterationNeeds.end()) { CIter->second.reset(0); CIter->second.reset(SEEN_AS_LIVE); } } } for (auto &S : SU->Preds) { auto Stg = SMS.stageScheduled(SU); if (S.isAssignedRegDep()) { Register Reg = S.getReg(); auto CIter = CrossIterationNeeds.find(Reg.id()); if (CIter != CrossIterationNeeds.end()) { auto Stg2 = SMS.stageScheduled(const_cast(S.getSUnit())); assert(Stg2 <= Stg && "Data dependence upon earlier stage"); if (Stg - Stg2 < MAX_STAGES) CIter->second.set(Stg - Stg2); CIter->second.set(SEEN_AS_LIVE); } } } bumpCrossIterationPressure(RPTracker, CrossIterationNeeds); } auto &P = RPTracker.getPressure().MaxSetPressure; for (unsigned I = 0, E = P.size(); I < E; ++I) if (P[I] > TRI->getRegPressureSetLimit(*MF, I)) { return true; } return false; } } // namespace std::unique_ptr ARMBaseInstrInfo::analyzeLoopForPipelining(MachineBasicBlock *LoopBB) const { MachineBasicBlock::iterator I = LoopBB->getFirstTerminator(); MachineBasicBlock *Preheader = *LoopBB->pred_begin(); if (Preheader == LoopBB) Preheader = *std::next(LoopBB->pred_begin()); if (I != LoopBB->end() && I->getOpcode() == ARM::t2Bcc) { // If the branch is a Bcc, then the CPSR should be set somewhere within the // block. We need to determine the reaching definition of CPSR so that // it can be marked as non-pipelineable, allowing the pipeliner to force // it into stage 0 or give up if it cannot or will not do so. MachineInstr *CCSetter = nullptr; for (auto &L : LoopBB->instrs()) { if (L.isCall()) return nullptr; if (isCPSRDefined(L)) CCSetter = &L; } if (CCSetter) return std::make_unique(&*I, CCSetter); else return nullptr; // Unable to find the CC setter, so unable to guarantee // that pipeline will work } // Recognize: // preheader: // %1 = t2DoopLoopStart %0 // loop: // %2 = phi %1, , %..., %loop // %3 = t2LoopDec %2, // t2LoopEnd %3, %loop if (I != LoopBB->end() && I->getOpcode() == ARM::t2LoopEnd) { for (auto &L : LoopBB->instrs()) if (L.isCall()) return nullptr; else if (isVCTP(&L)) return nullptr; Register LoopDecResult = I->getOperand(0).getReg(); MachineRegisterInfo &MRI = LoopBB->getParent()->getRegInfo(); MachineInstr *LoopDec = MRI.getUniqueVRegDef(LoopDecResult); if (!LoopDec || LoopDec->getOpcode() != ARM::t2LoopDec) return nullptr; MachineInstr *LoopStart = nullptr; for (auto &J : Preheader->instrs()) if (J.getOpcode() == ARM::t2DoLoopStart) LoopStart = &J; if (!LoopStart) return nullptr; return std::make_unique(&*I, LoopDec); } return nullptr; }