diff options
Diffstat (limited to 'llvm/lib')
51 files changed, 170 insertions, 159 deletions
diff --git a/llvm/lib/CodeGen/GlobalISel/LegalizerHelper.cpp b/llvm/lib/CodeGen/GlobalISel/LegalizerHelper.cpp index 178529f..52c43a4 100644 --- a/llvm/lib/CodeGen/GlobalISel/LegalizerHelper.cpp +++ b/llvm/lib/CodeGen/GlobalISel/LegalizerHelper.cpp @@ -9084,6 +9084,8 @@ LegalizerHelper::lowerShuffleVector(MachineInstr &MI) { SmallVector<Register, 32> BuildVec; LLT EltTy = DstTy.getScalarType(); + DenseMap<unsigned, Register> CachedExtract; + for (int Idx : Mask) { if (Idx < 0) { if (!Undef.isValid()) @@ -9097,9 +9099,13 @@ LegalizerHelper::lowerShuffleVector(MachineInstr &MI) { int NumElts = Src0Ty.getNumElements(); Register SrcVec = Idx < NumElts ? Src0Reg : Src1Reg; int ExtractIdx = Idx < NumElts ? Idx : Idx - NumElts; - auto IdxK = MIRBuilder.buildConstant(IdxTy, ExtractIdx); - auto Extract = MIRBuilder.buildExtractVectorElement(EltTy, SrcVec, IdxK); - BuildVec.push_back(Extract.getReg(0)); + auto [It, Inserted] = CachedExtract.try_emplace(Idx); + if (Inserted) { + auto IdxK = MIRBuilder.buildConstant(IdxTy, ExtractIdx); + It->second = + MIRBuilder.buildExtractVectorElement(EltTy, SrcVec, IdxK).getReg(0); + } + BuildVec.push_back(It->second); } assert(DstTy.isVector() && "Unexpected scalar G_SHUFFLE_VECTOR"); diff --git a/llvm/lib/CodeGen/RegisterUsageInfo.cpp b/llvm/lib/CodeGen/RegisterUsageInfo.cpp index 7a4628a..2ef380f 100644 --- a/llvm/lib/CodeGen/RegisterUsageInfo.cpp +++ b/llvm/lib/CodeGen/RegisterUsageInfo.cpp @@ -44,7 +44,7 @@ void PhysicalRegisterUsageInfo::setTargetMachine(const TargetMachine &TM) { } bool PhysicalRegisterUsageInfo::doInitialization(Module &M) { - RegMasks.grow(M.size()); + RegMasks.reserve(M.size()); return false; } diff --git a/llvm/lib/CodeGen/SelectionDAG/SelectionDAGBuilder.cpp b/llvm/lib/CodeGen/SelectionDAG/SelectionDAGBuilder.cpp index dee0909..a522650 100644 --- a/llvm/lib/CodeGen/SelectionDAG/SelectionDAGBuilder.cpp +++ b/llvm/lib/CodeGen/SelectionDAG/SelectionDAGBuilder.cpp @@ -2015,9 +2015,9 @@ SDValue SelectionDAGBuilder::getValueImpl(const Value *V) { Register InReg = FuncInfo.InitializeRegForValue(Inst); std::optional<CallingConv::ID> CallConv; - auto *CI = dyn_cast<CallInst>(Inst); - if (CI && !CI->isInlineAsm()) - CallConv = CI->getCallingConv(); + auto *CB = dyn_cast<CallBase>(Inst); + if (CB && !CB->isInlineAsm()) + CallConv = CB->getCallingConv(); RegsForValue RFV(*DAG.getContext(), TLI, DAG.getDataLayout(), InReg, Inst->getType(), CallConv); diff --git a/llvm/lib/CodeGen/SelectionDAG/TargetLowering.cpp b/llvm/lib/CodeGen/SelectionDAG/TargetLowering.cpp index 920dff9..da4e409 100644 --- a/llvm/lib/CodeGen/SelectionDAG/TargetLowering.cpp +++ b/llvm/lib/CodeGen/SelectionDAG/TargetLowering.cpp @@ -9899,6 +9899,18 @@ SDValue TargetLowering::expandBSWAP(SDNode *N, SelectionDAG &DAG) const { // Use a rotate by 8. This can be further expanded if necessary. return DAG.getNode(ISD::ROTL, dl, VT, Op, DAG.getConstant(8, dl, SHVT)); case MVT::i32: + // This is meant for ARM speficially, which has ROTR but no ROTL. + if (isOperationLegalOrCustom(ISD::ROTR, VT)) { + SDValue Mask = DAG.getConstant(0x00FF00FF, dl, VT); + // (x & 0x00FF00FF) rotr 8 | (x rotl 8) & 0x00FF00FF + SDValue And = DAG.getNode(ISD::AND, dl, VT, Op, Mask); + SDValue Rotr = + DAG.getNode(ISD::ROTR, dl, VT, And, DAG.getConstant(8, dl, SHVT)); + SDValue Rotl = + DAG.getNode(ISD::ROTR, dl, VT, Op, DAG.getConstant(24, dl, SHVT)); + SDValue And2 = DAG.getNode(ISD::AND, dl, VT, Rotl, Mask); + return DAG.getNode(ISD::OR, dl, VT, Rotr, And2); + } Tmp4 = DAG.getNode(ISD::SHL, dl, VT, Op, DAG.getConstant(24, dl, SHVT)); Tmp3 = DAG.getNode(ISD::AND, dl, VT, Op, DAG.getConstant(0xFF00, dl, VT)); diff --git a/llvm/lib/CodeGen/TargetOptionsImpl.cpp b/llvm/lib/CodeGen/TargetOptionsImpl.cpp index 049efe8..c33bf8b 100644 --- a/llvm/lib/CodeGen/TargetOptionsImpl.cpp +++ b/llvm/lib/CodeGen/TargetOptionsImpl.cpp @@ -44,7 +44,7 @@ bool TargetOptions::FramePointerIsReserved(const MachineFunction &MF) const { return false; return StringSwitch<bool>(FPAttr.getValueAsString()) - .Cases("all", "non-leaf", "reserved", true) + .Cases({"all", "non-leaf", "reserved"}, true) .Case("none", false); } diff --git a/llvm/lib/IR/DebugProgramInstruction.cpp b/llvm/lib/IR/DebugProgramInstruction.cpp index 2b9b0f9..d9357bb 100644 --- a/llvm/lib/IR/DebugProgramInstruction.cpp +++ b/llvm/lib/IR/DebugProgramInstruction.cpp @@ -665,11 +665,11 @@ void DbgMarker::eraseFromParent() { } iterator_range<DbgRecord::self_iterator> DbgMarker::getDbgRecordRange() { - return make_range(StoredDbgRecords.begin(), StoredDbgRecords.end()); + return StoredDbgRecords; } iterator_range<DbgRecord::const_self_iterator> DbgMarker::getDbgRecordRange() const { - return make_range(StoredDbgRecords.begin(), StoredDbgRecords.end()); + return StoredDbgRecords; } void DbgRecord::removeFromParent() { diff --git a/llvm/lib/MC/MCParser/MasmParser.cpp b/llvm/lib/MC/MCParser/MasmParser.cpp index d4901d9..8a8f111 100644 --- a/llvm/lib/MC/MCParser/MasmParser.cpp +++ b/llvm/lib/MC/MCParser/MasmParser.cpp @@ -5844,11 +5844,11 @@ bool MasmParser::lookUpField(const StructInfo &Structure, StringRef Member, bool MasmParser::lookUpType(StringRef Name, AsmTypeInfo &Info) const { unsigned Size = StringSwitch<unsigned>(Name) - .CasesLower("byte", "db", "sbyte", 1) - .CasesLower("word", "dw", "sword", 2) - .CasesLower("dword", "dd", "sdword", 4) - .CasesLower("fword", "df", 6) - .CasesLower("qword", "dq", "sqword", 8) + .CasesLower({"byte", "db", "sbyte"}, 1) + .CasesLower({"word", "dw", "sword"}, 2) + .CasesLower({"dword", "dd", "sdword"}, 4) + .CasesLower({"fword", "df"}, 6) + .CasesLower({"qword", "dq", "sqword"}, 8) .CaseLower("real4", 4) .CaseLower("real8", 8) .CaseLower("real10", 10) diff --git a/llvm/lib/Support/SpecialCaseList.cpp b/llvm/lib/Support/SpecialCaseList.cpp index c27f627..3a97185 100644 --- a/llvm/lib/Support/SpecialCaseList.cpp +++ b/llvm/lib/Support/SpecialCaseList.cpp @@ -92,8 +92,10 @@ void SpecialCaseList::GlobMatcher::preprocess(bool BySize) { for (const auto &G : reverse(Globs)) { StringRef Prefix = G.Pattern.prefix(); + StringRef Suffix = G.Pattern.suffix(); - auto &V = PrefixToGlob.emplace(Prefix).first->second; + auto &SToGlob = PrefixSuffixToGlob.emplace(Prefix).first->second; + auto &V = SToGlob.emplace(reverse(Suffix)).first->second; V.emplace_back(&G); } } @@ -101,16 +103,18 @@ void SpecialCaseList::GlobMatcher::preprocess(bool BySize) { void SpecialCaseList::GlobMatcher::match( StringRef Query, llvm::function_ref<void(StringRef Rule, unsigned LineNo)> Cb) const { - if (!PrefixToGlob.empty()) { - for (const auto &[_, V] : PrefixToGlob.find_prefixes(Query)) { - for (const auto *G : V) { - if (G->Pattern.match(Query)) { - Cb(G->Name, G->LineNo); - // As soon as we find a match in the vector, we can break for this - // vector, since the globs are already sorted by priority within the - // prefix group. However, we continue searching other prefix groups in - // the map, as they may contain a better match overall. - break; + if (!PrefixSuffixToGlob.empty()) { + for (const auto &[_, SToGlob] : PrefixSuffixToGlob.find_prefixes(Query)) { + for (const auto &[_, V] : SToGlob.find_prefixes(reverse(Query))) { + for (const auto *G : V) { + if (G->Pattern.match(Query)) { + Cb(G->Name, G->LineNo); + // As soon as we find a match in the vector, we can break for this + // vector, since the globs are already sorted by priority within the + // prefix group. However, we continue searching other prefix groups + // in the map, as they may contain a better match overall. + break; + } } } } diff --git a/llvm/lib/Target/AArch64/AArch64MachineFunctionInfo.h b/llvm/lib/Target/AArch64/AArch64MachineFunctionInfo.h index d1832f4..f680a5e 100644 --- a/llvm/lib/Target/AArch64/AArch64MachineFunctionInfo.h +++ b/llvm/lib/Target/AArch64/AArch64MachineFunctionInfo.h @@ -651,7 +651,7 @@ struct AArch64FunctionInfo final : public yaml::MachineFunctionInfo { AArch64FunctionInfo(const llvm::AArch64FunctionInfo &MFI); void mappingImpl(yaml::IO &YamlIO) override; - ~AArch64FunctionInfo() = default; + ~AArch64FunctionInfo() override = default; }; template <> struct MappingTraits<AArch64FunctionInfo> { diff --git a/llvm/lib/Target/AArch64/AArch64RegisterInfo.h b/llvm/lib/Target/AArch64/AArch64RegisterInfo.h index 72a7676..47d76f3 100644 --- a/llvm/lib/Target/AArch64/AArch64RegisterInfo.h +++ b/llvm/lib/Target/AArch64/AArch64RegisterInfo.h @@ -154,7 +154,7 @@ public: bool shouldAnalyzePhysregInMachineLoopInfo(MCRegister R) const override; - virtual bool isIgnoredCVReg(MCRegister LLVMReg) const override; + bool isIgnoredCVReg(MCRegister LLVMReg) const override; }; } // end namespace llvm diff --git a/llvm/lib/Target/AMDGPU/AMDGPUHSAMetadataStreamer.h b/llvm/lib/Target/AMDGPU/AMDGPUHSAMetadataStreamer.h index 1b4b113..6bad4dbd 100644 --- a/llvm/lib/Target/AMDGPU/AMDGPUHSAMetadataStreamer.h +++ b/llvm/lib/Target/AMDGPU/AMDGPUHSAMetadataStreamer.h @@ -131,7 +131,7 @@ protected: public: MetadataStreamerMsgPackV4() = default; - ~MetadataStreamerMsgPackV4() = default; + ~MetadataStreamerMsgPackV4() override = default; bool emitTo(AMDGPUTargetStreamer &TargetStreamer) override; @@ -154,7 +154,7 @@ protected: public: MetadataStreamerMsgPackV5() = default; - ~MetadataStreamerMsgPackV5() = default; + ~MetadataStreamerMsgPackV5() override = default; }; class MetadataStreamerMsgPackV6 final : public MetadataStreamerMsgPackV5 { @@ -163,7 +163,7 @@ protected: public: MetadataStreamerMsgPackV6() = default; - ~MetadataStreamerMsgPackV6() = default; + ~MetadataStreamerMsgPackV6() override = default; void emitKernelAttrs(const AMDGPUTargetMachine &TM, const MachineFunction &MF, msgpack::MapDocNode Kern) override; diff --git a/llvm/lib/Target/AMDGPU/AMDGPUMIRFormatter.h b/llvm/lib/Target/AMDGPU/AMDGPUMIRFormatter.h index c5c9473..0804133 100644 --- a/llvm/lib/Target/AMDGPU/AMDGPUMIRFormatter.h +++ b/llvm/lib/Target/AMDGPU/AMDGPUMIRFormatter.h @@ -26,20 +26,19 @@ struct PerFunctionMIParsingState; class AMDGPUMIRFormatter final : public MIRFormatter { public: AMDGPUMIRFormatter() = default; - virtual ~AMDGPUMIRFormatter() = default; + ~AMDGPUMIRFormatter() override = default; /// Implement target specific printing for machine operand immediate value, so /// that we can have more meaningful mnemonic than a 64-bit integer. Passing /// None to OpIdx means the index is unknown. - virtual void printImm(raw_ostream &OS, const MachineInstr &MI, - std::optional<unsigned> OpIdx, - int64_t Imm) const override; + void printImm(raw_ostream &OS, const MachineInstr &MI, + std::optional<unsigned> OpIdx, int64_t Imm) const override; /// Implement target specific parsing of immediate mnemonics. The mnemonic is /// a string with a leading dot. - virtual bool parseImmMnemonic(const unsigned OpCode, const unsigned OpIdx, - StringRef Src, int64_t &Imm, - ErrorCallbackType ErrorCallback) const override; + bool parseImmMnemonic(const unsigned OpCode, const unsigned OpIdx, + StringRef Src, int64_t &Imm, + ErrorCallbackType ErrorCallback) const override; /// Implement target specific parsing of target custom pseudo source value. bool diff --git a/llvm/lib/Target/AMDGPU/AMDGPUTargetMachine.cpp b/llvm/lib/Target/AMDGPU/AMDGPUTargetMachine.cpp index 02c5390..6214f4d 100644 --- a/llvm/lib/Target/AMDGPU/AMDGPUTargetMachine.cpp +++ b/llvm/lib/Target/AMDGPU/AMDGPUTargetMachine.cpp @@ -740,7 +740,7 @@ static StringRef getGPUOrDefault(const Triple &TT, StringRef GPU) { return "r600"; } -static Reloc::Model getEffectiveRelocModel(std::optional<Reloc::Model> RM) { +static Reloc::Model getEffectiveRelocModel() { // The AMDGPU toolchain only supports generating shared objects, so we // must always use PIC. return Reloc::PIC_; @@ -754,8 +754,8 @@ AMDGPUTargetMachine::AMDGPUTargetMachine(const Target &T, const Triple &TT, CodeGenOptLevel OptLevel) : CodeGenTargetMachineImpl( T, TT.computeDataLayout(), TT, getGPUOrDefault(TT, CPU), FS, Options, - getEffectiveRelocModel(RM), - getEffectiveCodeModel(CM, CodeModel::Small), OptLevel), + getEffectiveRelocModel(), getEffectiveCodeModel(CM, CodeModel::Small), + OptLevel), TLOF(createTLOF(getTargetTriple())) { initAsmInfo(); if (TT.isAMDGCN()) { diff --git a/llvm/lib/Target/AMDGPU/MCA/AMDGPUCustomBehaviour.h b/llvm/lib/Target/AMDGPU/MCA/AMDGPUCustomBehaviour.h index cbc7427..4d0c163 100644 --- a/llvm/lib/Target/AMDGPU/MCA/AMDGPUCustomBehaviour.h +++ b/llvm/lib/Target/AMDGPU/MCA/AMDGPUCustomBehaviour.h @@ -32,7 +32,7 @@ public: AMDGPUInstrPostProcess(const MCSubtargetInfo &STI, const MCInstrInfo &MCII) : InstrPostProcess(STI, MCII) {} - ~AMDGPUInstrPostProcess() = default; + ~AMDGPUInstrPostProcess() override = default; void postProcessInstruction(Instruction &Inst, const MCInst &MCI) override; }; @@ -88,7 +88,7 @@ public: AMDGPUCustomBehaviour(const MCSubtargetInfo &STI, const mca::SourceMgr &SrcMgr, const MCInstrInfo &MCII); - ~AMDGPUCustomBehaviour() = default; + ~AMDGPUCustomBehaviour() override = default; /// This method is used to determine if an instruction /// should be allowed to be dispatched. The return value is /// how many cycles until the instruction can be dispatched. diff --git a/llvm/lib/Target/AMDGPU/MCTargetDesc/AMDGPUMCExpr.h b/llvm/lib/Target/AMDGPU/MCTargetDesc/AMDGPUMCExpr.h index 54fcd2a..246a3f8 100644 --- a/llvm/lib/Target/AMDGPU/MCTargetDesc/AMDGPUMCExpr.h +++ b/llvm/lib/Target/AMDGPU/MCTargetDesc/AMDGPUMCExpr.h @@ -64,7 +64,7 @@ private: ArrayRef<const MCExpr *> Args; AMDGPUMCExpr(VariantKind Kind, ArrayRef<const MCExpr *> Args, MCContext &Ctx); - ~AMDGPUMCExpr(); + ~AMDGPUMCExpr() override; bool evaluateExtraSGPRs(MCValue &Res, const MCAssembler *Asm) const; bool evaluateTotalNumVGPR(MCValue &Res, const MCAssembler *Asm) const; diff --git a/llvm/lib/Target/AMDGPU/SIInstrInfo.h b/llvm/lib/Target/AMDGPU/SIInstrInfo.h index 5fdedda..dc23a21 100644 --- a/llvm/lib/Target/AMDGPU/SIInstrInfo.h +++ b/llvm/lib/Target/AMDGPU/SIInstrInfo.h @@ -1640,7 +1640,7 @@ public: unsigned *PredCost = nullptr) const override; InstructionUniformity - getInstructionUniformity(const MachineInstr &MI) const override final; + getInstructionUniformity(const MachineInstr &MI) const final; InstructionUniformity getGenericInstructionUniformity(const MachineInstr &MI) const; diff --git a/llvm/lib/Target/AMDGPU/SIMachineFunctionInfo.h b/llvm/lib/Target/AMDGPU/SIMachineFunctionInfo.h index 2c1a13c..019c3b7 100644 --- a/llvm/lib/Target/AMDGPU/SIMachineFunctionInfo.h +++ b/llvm/lib/Target/AMDGPU/SIMachineFunctionInfo.h @@ -311,7 +311,7 @@ struct SIMachineFunctionInfo final : public yaml::MachineFunctionInfo { const llvm::MachineFunction &MF); void mappingImpl(yaml::IO &YamlIO) override; - ~SIMachineFunctionInfo() = default; + ~SIMachineFunctionInfo() override = default; }; template <> struct MappingTraits<SIMachineFunctionInfo> { diff --git a/llvm/lib/Target/AMDGPU/SIMemoryLegalizer.cpp b/llvm/lib/Target/AMDGPU/SIMemoryLegalizer.cpp index 07264d9..a177a42 100644 --- a/llvm/lib/Target/AMDGPU/SIMemoryLegalizer.cpp +++ b/llvm/lib/Target/AMDGPU/SIMemoryLegalizer.cpp @@ -640,7 +640,7 @@ public: bool finalizeStore(MachineInstr &MI, bool Atomic) const override; - virtual bool handleCooperativeAtomic(MachineInstr &MI) const override; + bool handleCooperativeAtomic(MachineInstr &MI) const override; bool insertRelease(MachineBasicBlock::iterator &MI, SIAtomicScope Scope, SIAtomicAddrSpace AddrSpace, bool IsCrossAddrSpaceOrdering, diff --git a/llvm/lib/Target/ARM/ARMConstantPoolValue.h b/llvm/lib/Target/ARM/ARMConstantPoolValue.h index 261070a..e21b2c9 100644 --- a/llvm/lib/Target/ARM/ARMConstantPoolValue.h +++ b/llvm/lib/Target/ARM/ARMConstantPoolValue.h @@ -176,9 +176,7 @@ public: using promoted_iterator = SmallPtrSet<const GlobalVariable *, 1>::iterator; - iterator_range<promoted_iterator> promotedGlobals() { - return iterator_range<promoted_iterator>(GVars.begin(), GVars.end()); - } + iterator_range<promoted_iterator> promotedGlobals() { return GVars; } const Constant *getPromotedGlobalInit() const { return CVal; diff --git a/llvm/lib/Target/ARM/ARMMachineFunctionInfo.h b/llvm/lib/Target/ARM/ARMMachineFunctionInfo.h index a9c4b53..72eb3d0 100644 --- a/llvm/lib/Target/ARM/ARMMachineFunctionInfo.h +++ b/llvm/lib/Target/ARM/ARMMachineFunctionInfo.h @@ -312,7 +312,7 @@ struct ARMFunctionInfo final : public yaml::MachineFunctionInfo { ARMFunctionInfo(const llvm::ARMFunctionInfo &MFI); void mappingImpl(yaml::IO &YamlIO) override; - ~ARMFunctionInfo() = default; + ~ARMFunctionInfo() override = default; }; template <> struct MappingTraits<ARMFunctionInfo> { diff --git a/llvm/lib/Target/AVR/Disassembler/AVRDisassembler.cpp b/llvm/lib/Target/AVR/Disassembler/AVRDisassembler.cpp index 3a840a3..5548ad1 100644 --- a/llvm/lib/Target/AVR/Disassembler/AVRDisassembler.cpp +++ b/llvm/lib/Target/AVR/Disassembler/AVRDisassembler.cpp @@ -40,7 +40,7 @@ class AVRDisassembler : public MCDisassembler { public: AVRDisassembler(const MCSubtargetInfo &STI, MCContext &Ctx) : MCDisassembler(STI, Ctx) {} - virtual ~AVRDisassembler() = default; + ~AVRDisassembler() override = default; DecodeStatus getInstruction(MCInst &Instr, uint64_t &Size, ArrayRef<uint8_t> Bytes, uint64_t Address, diff --git a/llvm/lib/Target/AVR/MCTargetDesc/AVRELFObjectWriter.cpp b/llvm/lib/Target/AVR/MCTargetDesc/AVRELFObjectWriter.cpp index 619efb3..03c60e8 100644 --- a/llvm/lib/Target/AVR/MCTargetDesc/AVRELFObjectWriter.cpp +++ b/llvm/lib/Target/AVR/MCTargetDesc/AVRELFObjectWriter.cpp @@ -24,7 +24,7 @@ class AVRELFObjectWriter : public MCELFObjectTargetWriter { public: AVRELFObjectWriter(uint8_t OSABI); - virtual ~AVRELFObjectWriter() = default; + ~AVRELFObjectWriter() override = default; unsigned getRelocType(const MCFixup &, const MCValue &, bool IsPCRel) const override; diff --git a/llvm/lib/Target/BPF/BPFAsmPrinter.h b/llvm/lib/Target/BPF/BPFAsmPrinter.h index 0cfb283..90ef207 100644 --- a/llvm/lib/Target/BPF/BPFAsmPrinter.h +++ b/llvm/lib/Target/BPF/BPFAsmPrinter.h @@ -32,7 +32,7 @@ public: void emitInstruction(const MachineInstr *MI) override; MCSymbol *getJTPublicSymbol(unsigned JTI); - virtual void emitJumpTableInfo() override; + void emitJumpTableInfo() override; static char ID; diff --git a/llvm/lib/Target/BPF/BPFCheckAndAdjustIR.cpp b/llvm/lib/Target/BPF/BPFCheckAndAdjustIR.cpp index e3c39a1..b12985d 100644 --- a/llvm/lib/Target/BPF/BPFCheckAndAdjustIR.cpp +++ b/llvm/lib/Target/BPF/BPFCheckAndAdjustIR.cpp @@ -46,7 +46,7 @@ class BPFCheckAndAdjustIR final : public ModulePass { public: static char ID; BPFCheckAndAdjustIR() : ModulePass(ID) {} - virtual void getAnalysisUsage(AnalysisUsage &AU) const override; + void getAnalysisUsage(AnalysisUsage &AU) const override; private: void checkIR(Module &M); diff --git a/llvm/lib/Target/BPF/BPFTargetLoweringObjectFile.h b/llvm/lib/Target/BPF/BPFTargetLoweringObjectFile.h index f3064c0c..af3542e 100644 --- a/llvm/lib/Target/BPF/BPFTargetLoweringObjectFile.h +++ b/llvm/lib/Target/BPF/BPFTargetLoweringObjectFile.h @@ -16,7 +16,7 @@ namespace llvm { class BPFTargetLoweringObjectFileELF : public TargetLoweringObjectFileELF { public: - virtual MCSection * + MCSection * getSectionForJumpTable(const Function &F, const TargetMachine &TM, const MachineJumpTableEntry *JTE) const override; }; diff --git a/llvm/lib/Target/LoongArch/LoongArchTargetMachine.cpp b/llvm/lib/Target/LoongArch/LoongArchTargetMachine.cpp index c5e26c1..9de4c9d 100644 --- a/llvm/lib/Target/LoongArch/LoongArchTargetMachine.cpp +++ b/llvm/lib/Target/LoongArch/LoongArchTargetMachine.cpp @@ -62,8 +62,7 @@ static cl::opt<bool> cl::desc("Enable the merge base offset pass"), cl::init(true), cl::Hidden); -static Reloc::Model getEffectiveRelocModel(const Triple &TT, - std::optional<Reloc::Model> RM) { +static Reloc::Model getEffectiveRelocModel(std::optional<Reloc::Model> RM) { return RM.value_or(Reloc::Static); } @@ -92,7 +91,7 @@ LoongArchTargetMachine::LoongArchTargetMachine( const TargetOptions &Options, std::optional<Reloc::Model> RM, std::optional<CodeModel::Model> CM, CodeGenOptLevel OL, bool JIT) : CodeGenTargetMachineImpl(T, TT.computeDataLayout(), TT, CPU, FS, Options, - getEffectiveRelocModel(TT, RM), + getEffectiveRelocModel(RM), getEffectiveLoongArchCodeModel(TT, CM), OL), TLOF(std::make_unique<TargetLoweringObjectFileELF>()) { initAsmInfo(); diff --git a/llvm/lib/Target/M68k/M68kTargetMachine.cpp b/llvm/lib/Target/M68k/M68kTargetMachine.cpp index 847c27ba..f525d43 100644 --- a/llvm/lib/Target/M68k/M68kTargetMachine.cpp +++ b/llvm/lib/Target/M68k/M68kTargetMachine.cpp @@ -46,13 +46,9 @@ extern "C" LLVM_EXTERNAL_VISIBILITY void LLVMInitializeM68kTarget() { namespace { -Reloc::Model getEffectiveRelocModel(const Triple &TT, - std::optional<Reloc::Model> RM) { +Reloc::Model getEffectiveRelocModel(std::optional<Reloc::Model> RM) { // If not defined we default to static - if (!RM.has_value()) - return Reloc::Static; - - return *RM; + return RM.value_or(Reloc::Static); } CodeModel::Model getEffectiveCodeModel(std::optional<CodeModel::Model> CM, @@ -73,7 +69,7 @@ M68kTargetMachine::M68kTargetMachine(const Target &T, const Triple &TT, std::optional<CodeModel::Model> CM, CodeGenOptLevel OL, bool JIT) : CodeGenTargetMachineImpl(T, TT.computeDataLayout(), TT, CPU, FS, Options, - getEffectiveRelocModel(TT, RM), + getEffectiveRelocModel(RM), ::getEffectiveCodeModel(CM, JIT), OL), TLOF(std::make_unique<M68kELFTargetObjectFile>()), Subtarget(TT, CPU, FS, *this) { diff --git a/llvm/lib/Target/Mips/MipsRegisterBankInfo.cpp b/llvm/lib/Target/Mips/MipsRegisterBankInfo.cpp index 4aecaf1..8e791e6 100644 --- a/llvm/lib/Target/Mips/MipsRegisterBankInfo.cpp +++ b/llvm/lib/Target/Mips/MipsRegisterBankInfo.cpp @@ -635,7 +635,7 @@ public: B.setChangeObserver(*this); } - ~InstManager() { B.stopObservingChanges(); } + ~InstManager() override { B.stopObservingChanges(); } void createdInstr(MachineInstr &MI) override { InstList.insert(&MI); } void erasingInstr(MachineInstr &MI) override {} diff --git a/llvm/lib/Target/RISCV/MCA/RISCVCustomBehaviour.h b/llvm/lib/Target/RISCV/MCA/RISCVCustomBehaviour.h index 34efa0b..4ccd3cf 100644 --- a/llvm/lib/Target/RISCV/MCA/RISCVCustomBehaviour.h +++ b/llvm/lib/Target/RISCV/MCA/RISCVCustomBehaviour.h @@ -33,7 +33,7 @@ public: explicit RISCVLMULInstrument(StringRef Data) : Instrument(DESC_NAME, Data) {} - ~RISCVLMULInstrument() = default; + ~RISCVLMULInstrument() override = default; uint8_t getLMUL() const; }; @@ -45,7 +45,7 @@ public: explicit RISCVSEWInstrument(StringRef Data) : Instrument(DESC_NAME, Data) {} - ~RISCVSEWInstrument() = default; + ~RISCVSEWInstrument() override = default; uint8_t getSEW() const; }; diff --git a/llvm/lib/Target/RISCV/RISCVConstantPoolValue.h b/llvm/lib/Target/RISCV/RISCVConstantPoolValue.h index b69904d..b2ce57a 100644 --- a/llvm/lib/Target/RISCV/RISCVConstantPoolValue.h +++ b/llvm/lib/Target/RISCV/RISCVConstantPoolValue.h @@ -37,7 +37,7 @@ private: RISCVCPKind Kind; public: - ~RISCVConstantPoolValue() = default; + ~RISCVConstantPoolValue() override = default; static RISCVConstantPoolValue *Create(const GlobalValue *GV); static RISCVConstantPoolValue *Create(LLVMContext &C, StringRef S); diff --git a/llvm/lib/Target/RISCV/RISCVMachineFunctionInfo.h b/llvm/lib/Target/RISCV/RISCVMachineFunctionInfo.h index 4fa93f1..f9be80f 100644 --- a/llvm/lib/Target/RISCV/RISCVMachineFunctionInfo.h +++ b/llvm/lib/Target/RISCV/RISCVMachineFunctionInfo.h @@ -31,7 +31,7 @@ struct RISCVMachineFunctionInfo final : public yaml::MachineFunctionInfo { RISCVMachineFunctionInfo(const llvm::RISCVMachineFunctionInfo &MFI); void mappingImpl(yaml::IO &YamlIO) override; - ~RISCVMachineFunctionInfo() = default; + ~RISCVMachineFunctionInfo() override = default; }; template <> struct MappingTraits<RISCVMachineFunctionInfo> { diff --git a/llvm/lib/Target/RISCV/RISCVTargetMachine.cpp b/llvm/lib/Target/RISCV/RISCVTargetMachine.cpp index f81b1e12..ae54ff1 100644 --- a/llvm/lib/Target/RISCV/RISCVTargetMachine.cpp +++ b/llvm/lib/Target/RISCV/RISCVTargetMachine.cpp @@ -141,8 +141,7 @@ extern "C" LLVM_ABI LLVM_EXTERNAL_VISIBILITY void LLVMInitializeRISCVTarget() { initializeRISCVAsmPrinterPass(*PR); } -static Reloc::Model getEffectiveRelocModel(const Triple &TT, - std::optional<Reloc::Model> RM) { +static Reloc::Model getEffectiveRelocModel(std::optional<Reloc::Model> RM) { return RM.value_or(Reloc::Static); } @@ -154,7 +153,7 @@ RISCVTargetMachine::RISCVTargetMachine(const Target &T, const Triple &TT, CodeGenOptLevel OL, bool JIT) : CodeGenTargetMachineImpl( T, TT.computeDataLayout(Options.MCOptions.getABIName()), TT, CPU, FS, - Options, getEffectiveRelocModel(TT, RM), + Options, getEffectiveRelocModel(RM), getEffectiveCodeModel(CM, CodeModel::Small), OL), TLOF(std::make_unique<RISCVELFTargetObjectFile>()) { initAsmInfo(); diff --git a/llvm/lib/Target/SPIRV/SPIRVLegalizePointerCast.cpp b/llvm/lib/Target/SPIRV/SPIRVLegalizePointerCast.cpp index 28a1690..6e444c9 100644 --- a/llvm/lib/Target/SPIRV/SPIRVLegalizePointerCast.cpp +++ b/llvm/lib/Target/SPIRV/SPIRVLegalizePointerCast.cpp @@ -347,7 +347,7 @@ class SPIRVLegalizePointerCast : public FunctionPass { public: SPIRVLegalizePointerCast(SPIRVTargetMachine *TM) : FunctionPass(ID), TM(TM) {} - virtual bool runOnFunction(Function &F) override { + bool runOnFunction(Function &F) override { const SPIRVSubtarget &ST = TM->getSubtarget<SPIRVSubtarget>(F); GR = ST.getSPIRVGlobalRegistry(); DeadInstructions.clear(); diff --git a/llvm/lib/Target/SPIRV/SPIRVMergeRegionExitTargets.cpp b/llvm/lib/Target/SPIRV/SPIRVMergeRegionExitTargets.cpp index 60d39c9..aba9cf7 100644 --- a/llvm/lib/Target/SPIRV/SPIRVMergeRegionExitTargets.cpp +++ b/llvm/lib/Target/SPIRV/SPIRVMergeRegionExitTargets.cpp @@ -234,7 +234,7 @@ public: } #endif - virtual bool runOnFunction(Function &F) override { + bool runOnFunction(Function &F) override { LoopInfo &LI = getAnalysis<LoopInfoWrapperPass>().getLoopInfo(); auto *TopLevelRegion = getAnalysis<SPIRVConvergenceRegionAnalysisWrapperPass>() diff --git a/llvm/lib/Target/SPIRV/SPIRVStripConvergentIntrinsics.cpp b/llvm/lib/Target/SPIRV/SPIRVStripConvergentIntrinsics.cpp index e621bcd44..b1a8d1a 100644 --- a/llvm/lib/Target/SPIRV/SPIRVStripConvergentIntrinsics.cpp +++ b/llvm/lib/Target/SPIRV/SPIRVStripConvergentIntrinsics.cpp @@ -28,7 +28,7 @@ public: SPIRVStripConvergentIntrinsics() : FunctionPass(ID) {} - virtual bool runOnFunction(Function &F) override { + bool runOnFunction(Function &F) override { DenseSet<Instruction *> ToRemove; // Is the instruction is a convergent intrinsic, add it to kill-list and diff --git a/llvm/lib/Target/SPIRV/SPIRVStructurizer.cpp b/llvm/lib/Target/SPIRV/SPIRVStructurizer.cpp index 5b149f8..ea634fb 100644 --- a/llvm/lib/Target/SPIRV/SPIRVStructurizer.cpp +++ b/llvm/lib/Target/SPIRV/SPIRVStructurizer.cpp @@ -1113,7 +1113,7 @@ public: SPIRVStructurizer() : FunctionPass(ID) {} - virtual bool runOnFunction(Function &F) override { + bool runOnFunction(Function &F) override { bool Modified = false; // In LLVM, Switches are allowed to have several cases branching to the same diff --git a/llvm/lib/Target/Sparc/Disassembler/SparcDisassembler.cpp b/llvm/lib/Target/Sparc/Disassembler/SparcDisassembler.cpp index e585e5a..b4dadaa 100644 --- a/llvm/lib/Target/Sparc/Disassembler/SparcDisassembler.cpp +++ b/llvm/lib/Target/Sparc/Disassembler/SparcDisassembler.cpp @@ -36,7 +36,7 @@ class SparcDisassembler : public MCDisassembler { public: SparcDisassembler(const MCSubtargetInfo &STI, MCContext &Ctx) : MCDisassembler(STI, Ctx) {} - virtual ~SparcDisassembler() = default; + ~SparcDisassembler() override = default; DecodeStatus getInstruction(MCInst &Instr, uint64_t &Size, ArrayRef<uint8_t> Bytes, uint64_t Address, diff --git a/llvm/lib/Target/SystemZ/SystemZMachineScheduler.h b/llvm/lib/Target/SystemZ/SystemZMachineScheduler.h index 1ff6cc8..ba325b5 100644 --- a/llvm/lib/Target/SystemZ/SystemZMachineScheduler.h +++ b/llvm/lib/Target/SystemZ/SystemZMachineScheduler.h @@ -111,7 +111,7 @@ class SystemZPostRASchedStrategy : public MachineSchedStrategy { public: SystemZPostRASchedStrategy(const MachineSchedContext *C); - virtual ~SystemZPostRASchedStrategy(); + ~SystemZPostRASchedStrategy() override; /// Called for a region before scheduling. void initPolicy(MachineBasicBlock::iterator Begin, diff --git a/llvm/lib/Target/SystemZ/SystemZRegisterInfo.h b/llvm/lib/Target/SystemZ/SystemZRegisterInfo.h index b1de145..bea8b9f 100644 --- a/llvm/lib/Target/SystemZ/SystemZRegisterInfo.h +++ b/llvm/lib/Target/SystemZ/SystemZRegisterInfo.h @@ -101,7 +101,7 @@ public: int getStackPointerBias() final { return 2048; } /// Destroys the object. Bogus destructor overriding base class destructor - ~SystemZXPLINK64Registers() = default; + ~SystemZXPLINK64Registers() override = default; }; /// ELF calling convention specific use registers @@ -124,7 +124,7 @@ public: int getStackPointerBias() final { return 0; } /// Destroys the object. Bogus destructor overriding base class destructor - ~SystemZELFRegisters() = default; + ~SystemZELFRegisters() override = default; }; struct SystemZRegisterInfo : public SystemZGenRegisterInfo { diff --git a/llvm/lib/Target/VE/Disassembler/VEDisassembler.cpp b/llvm/lib/Target/VE/Disassembler/VEDisassembler.cpp index aad826b..465e074 100644 --- a/llvm/lib/Target/VE/Disassembler/VEDisassembler.cpp +++ b/llvm/lib/Target/VE/Disassembler/VEDisassembler.cpp @@ -36,7 +36,7 @@ class VEDisassembler : public MCDisassembler { public: VEDisassembler(const MCSubtargetInfo &STI, MCContext &Ctx) : MCDisassembler(STI, Ctx) {} - virtual ~VEDisassembler() = default; + ~VEDisassembler() override = default; DecodeStatus getInstruction(MCInst &Instr, uint64_t &Size, ArrayRef<uint8_t> Bytes, uint64_t Address, diff --git a/llvm/lib/Target/WebAssembly/AsmParser/WebAssemblyAsmParser.cpp b/llvm/lib/Target/WebAssembly/AsmParser/WebAssemblyAsmParser.cpp index 45bbf12..9175b27 100644 --- a/llvm/lib/Target/WebAssembly/AsmParser/WebAssemblyAsmParser.cpp +++ b/llvm/lib/Target/WebAssembly/AsmParser/WebAssemblyAsmParser.cpp @@ -102,7 +102,7 @@ struct WebAssemblyOperand : public MCParsedAsmOperand { WebAssemblyOperand(SMLoc Start, SMLoc End, CaLOp C) : Kind(CatchList), StartLoc(Start), EndLoc(End), CaL(C) {} - ~WebAssemblyOperand() { + ~WebAssemblyOperand() override { if (isBrList()) BrL.~BrLOp(); if (isCatchList()) diff --git a/llvm/lib/Target/WebAssembly/WebAssemblyMachineFunctionInfo.h b/llvm/lib/Target/WebAssembly/WebAssemblyMachineFunctionInfo.h index 40ae4ae..ff4d6469 100644 --- a/llvm/lib/Target/WebAssembly/WebAssemblyMachineFunctionInfo.h +++ b/llvm/lib/Target/WebAssembly/WebAssemblyMachineFunctionInfo.h @@ -192,7 +192,7 @@ struct WebAssemblyFunctionInfo final : public yaml::MachineFunctionInfo { const llvm::WebAssemblyFunctionInfo &MFI); void mappingImpl(yaml::IO &YamlIO) override; - ~WebAssemblyFunctionInfo() = default; + ~WebAssemblyFunctionInfo() override = default; }; template <> struct MappingTraits<WebAssemblyFunctionInfo> { diff --git a/llvm/lib/Target/WebAssembly/WebAssemblyTargetMachine.cpp b/llvm/lib/Target/WebAssembly/WebAssemblyTargetMachine.cpp index a9c638c..621640c 100644 --- a/llvm/lib/Target/WebAssembly/WebAssemblyTargetMachine.cpp +++ b/llvm/lib/Target/WebAssembly/WebAssemblyTargetMachine.cpp @@ -127,16 +127,11 @@ LLVMInitializeWebAssemblyTarget() { // WebAssembly Lowering public interface. //===----------------------------------------------------------------------===// -static Reloc::Model getEffectiveRelocModel(std::optional<Reloc::Model> RM, - const Triple &TT) { - if (!RM) { - // Default to static relocation model. This should always be more optimial - // than PIC since the static linker can determine all global addresses and - // assume direct function calls. - return Reloc::Static; - } - - return *RM; +static Reloc::Model getEffectiveRelocModel(std::optional<Reloc::Model> RM) { + // Default to static relocation model. This should always be more optimial + // than PIC since the static linker can determine all global addresses and + // assume direct function calls. + return RM.value_or(Reloc::Static); } using WebAssembly::WasmEnableEH; @@ -197,7 +192,7 @@ WebAssemblyTargetMachine::WebAssemblyTargetMachine( const TargetOptions &Options, std::optional<Reloc::Model> RM, std::optional<CodeModel::Model> CM, CodeGenOptLevel OL, bool JIT) : CodeGenTargetMachineImpl(T, TT.computeDataLayout(), TT, CPU, FS, Options, - getEffectiveRelocModel(RM, TT), + getEffectiveRelocModel(RM), getEffectiveCodeModel(CM, CodeModel::Large), OL), TLOF(new WebAssemblyTargetObjectFile()), UsesMultivalueABI(Options.MCOptions.getABIName() == "experimental-mv") { diff --git a/llvm/lib/Target/X86/AsmParser/X86AsmParser.cpp b/llvm/lib/Target/X86/AsmParser/X86AsmParser.cpp index ac251fd..127ee67 100644 --- a/llvm/lib/Target/X86/AsmParser/X86AsmParser.cpp +++ b/llvm/lib/Target/X86/AsmParser/X86AsmParser.cpp @@ -3533,10 +3533,10 @@ bool X86AsmParser::parseInstruction(ParseInstructionInfo &Info, StringRef Name, while (isLockRepeatNtPrefix(Name.lower())) { unsigned Prefix = StringSwitch<unsigned>(Name) - .Cases("lock", "lock", X86::IP_HAS_LOCK) - .Cases("rep", "repe", "repz", X86::IP_HAS_REPEAT) - .Cases("repne", "repnz", X86::IP_HAS_REPEAT_NE) - .Cases("notrack", "notrack", X86::IP_HAS_NOTRACK) + .Case("lock", X86::IP_HAS_LOCK) + .Cases({"rep", "repe", "repz"}, X86::IP_HAS_REPEAT) + .Cases({"repne", "repnz"}, X86::IP_HAS_REPEAT_NE) + .Case("notrack", X86::IP_HAS_NOTRACK) .Default(X86::IP_NO_PREFIX); // Invalid prefix (impossible) Flags |= Prefix; if (getLexer().is(AsmToken::EndOfStatement)) { diff --git a/llvm/lib/Target/X86/MCA/X86CustomBehaviour.h b/llvm/lib/Target/X86/MCA/X86CustomBehaviour.h index d6197f3..05a1c22 100644 --- a/llvm/lib/Target/X86/MCA/X86CustomBehaviour.h +++ b/llvm/lib/Target/X86/MCA/X86CustomBehaviour.h @@ -37,7 +37,7 @@ public: X86InstrPostProcess(const MCSubtargetInfo &STI, const MCInstrInfo &MCII) : InstrPostProcess(STI, MCII) {} - ~X86InstrPostProcess() = default; + ~X86InstrPostProcess() override = default; void postProcessInstruction(Instruction &Inst, const MCInst &MCI) override; }; diff --git a/llvm/lib/Target/X86/MCTargetDesc/X86MCTargetDesc.cpp b/llvm/lib/Target/X86/MCTargetDesc/X86MCTargetDesc.cpp index 1c06dc4..af5a698 100644 --- a/llvm/lib/Target/X86/MCTargetDesc/X86MCTargetDesc.cpp +++ b/llvm/lib/Target/X86/MCTargetDesc/X86MCTargetDesc.cpp @@ -491,7 +491,7 @@ namespace X86_MC { class X86MCInstrAnalysis : public MCInstrAnalysis { X86MCInstrAnalysis(const X86MCInstrAnalysis &) = delete; X86MCInstrAnalysis &operator=(const X86MCInstrAnalysis &) = delete; - virtual ~X86MCInstrAnalysis() = default; + ~X86MCInstrAnalysis() override = default; public: X86MCInstrAnalysis(const MCInstrInfo *MCII) : MCInstrAnalysis(MCII) {} diff --git a/llvm/lib/Target/X86/X86DomainReassignment.cpp b/llvm/lib/Target/X86/X86DomainReassignment.cpp index 339e2f3..5d19011 100644 --- a/llvm/lib/Target/X86/X86DomainReassignment.cpp +++ b/llvm/lib/Target/X86/X86DomainReassignment.cpp @@ -324,9 +324,7 @@ public: bool insertEdge(Register Reg) { return Edges.insert(Reg).second; } using const_edge_iterator = DenseSet<Register>::const_iterator; - iterator_range<const_edge_iterator> edges() const { - return iterator_range<const_edge_iterator>(Edges.begin(), Edges.end()); - } + iterator_range<const_edge_iterator> edges() const { return Edges; } void addInstruction(MachineInstr *I) { Instrs.push_back(I); diff --git a/llvm/lib/Target/X86/X86MachineFunctionInfo.h b/llvm/lib/Target/X86/X86MachineFunctionInfo.h index 5f974e5..1bda505 100644 --- a/llvm/lib/Target/X86/X86MachineFunctionInfo.h +++ b/llvm/lib/Target/X86/X86MachineFunctionInfo.h @@ -43,7 +43,7 @@ struct X86MachineFunctionInfo final : public yaml::MachineFunctionInfo { X86MachineFunctionInfo(const llvm::X86MachineFunctionInfo &MFI); void mappingImpl(yaml::IO &YamlIO) override; - ~X86MachineFunctionInfo() = default; + ~X86MachineFunctionInfo() override = default; }; template <> struct MappingTraits<X86MachineFunctionInfo> { diff --git a/llvm/lib/Transforms/IPO/Attributor.cpp b/llvm/lib/Transforms/IPO/Attributor.cpp index 077d29f..3b59ebb 100644 --- a/llvm/lib/Transforms/IPO/Attributor.cpp +++ b/llvm/lib/Transforms/IPO/Attributor.cpp @@ -272,6 +272,9 @@ AA::getInitialValueForObj(Attributor &A, const AbstractAttribute &QueryingAA, } if (RangePtr && !RangePtr->offsetOrSizeAreUnknown()) { + int64_t StorageSize = DL.getTypeStoreSize(&Ty); + if (StorageSize != RangePtr->Size) + return nullptr; APInt Offset = APInt(64, RangePtr->Offset); return ConstantFoldLoadFromConst(Initializer, &Ty, Offset, DL); } diff --git a/llvm/lib/Transforms/Vectorize/VPlanUtils.cpp b/llvm/lib/Transforms/Vectorize/VPlanUtils.cpp index fe66f13..4db92e7 100644 --- a/llvm/lib/Transforms/Vectorize/VPlanUtils.cpp +++ b/llvm/lib/Transforms/Vectorize/VPlanUtils.cpp @@ -89,6 +89,60 @@ const SCEV *vputils::getSCEVExprForVPValue(VPValue *V, ScalarEvolution &SE) { .Default([&SE](const VPRecipeBase *) { return SE.getCouldNotCompute(); }); } +bool vputils::isSingleScalar(const VPValue *VPV) { + auto PreservesUniformity = [](unsigned Opcode) -> bool { + if (Instruction::isBinaryOp(Opcode) || Instruction::isCast(Opcode)) + return true; + switch (Opcode) { + case Instruction::GetElementPtr: + case Instruction::ICmp: + case Instruction::FCmp: + case Instruction::Select: + case VPInstruction::Not: + case VPInstruction::Broadcast: + case VPInstruction::PtrAdd: + return true; + default: + return false; + } + }; + + // A live-in must be uniform across the scope of VPlan. + if (VPV->isLiveIn()) + return true; + + if (auto *Rep = dyn_cast<VPReplicateRecipe>(VPV)) { + const VPRegionBlock *RegionOfR = Rep->getRegion(); + // Don't consider recipes in replicate regions as uniform yet; their first + // lane cannot be accessed when executing the replicate region for other + // lanes. + if (RegionOfR && RegionOfR->isReplicator()) + return false; + return Rep->isSingleScalar() || (PreservesUniformity(Rep->getOpcode()) && + all_of(Rep->operands(), isSingleScalar)); + } + if (isa<VPWidenGEPRecipe, VPDerivedIVRecipe, VPBlendRecipe, + VPWidenSelectRecipe>(VPV)) + return all_of(VPV->getDefiningRecipe()->operands(), isSingleScalar); + if (auto *WidenR = dyn_cast<VPWidenRecipe>(VPV)) { + return PreservesUniformity(WidenR->getOpcode()) && + all_of(WidenR->operands(), isSingleScalar); + } + if (auto *VPI = dyn_cast<VPInstruction>(VPV)) + return VPI->isSingleScalar() || VPI->isVectorToScalar() || + (PreservesUniformity(VPI->getOpcode()) && + all_of(VPI->operands(), isSingleScalar)); + if (isa<VPPartialReductionRecipe>(VPV)) + return false; + if (isa<VPReductionRecipe>(VPV)) + return true; + if (auto *Expr = dyn_cast<VPExpressionRecipe>(VPV)) + return Expr->isSingleScalar(); + + // VPExpandSCEVRecipes must be placed in the entry and are always uniform. + return isa<VPExpandSCEVRecipe>(VPV); +} + bool vputils::isUniformAcrossVFsAndUFs(VPValue *V) { // Live-ins are uniform. if (V->isLiveIn()) diff --git a/llvm/lib/Transforms/Vectorize/VPlanUtils.h b/llvm/lib/Transforms/Vectorize/VPlanUtils.h index 840a5b9..37cd413 100644 --- a/llvm/lib/Transforms/Vectorize/VPlanUtils.h +++ b/llvm/lib/Transforms/Vectorize/VPlanUtils.h @@ -41,59 +41,7 @@ const SCEV *getSCEVExprForVPValue(VPValue *V, ScalarEvolution &SE); /// Returns true if \p VPV is a single scalar, either because it produces the /// same value for all lanes or only has its first lane used. -inline bool isSingleScalar(const VPValue *VPV) { - auto PreservesUniformity = [](unsigned Opcode) -> bool { - if (Instruction::isBinaryOp(Opcode) || Instruction::isCast(Opcode)) - return true; - switch (Opcode) { - case Instruction::GetElementPtr: - case Instruction::ICmp: - case Instruction::FCmp: - case Instruction::Select: - case VPInstruction::Not: - case VPInstruction::Broadcast: - case VPInstruction::PtrAdd: - return true; - default: - return false; - } - }; - - // A live-in must be uniform across the scope of VPlan. - if (VPV->isLiveIn()) - return true; - - if (auto *Rep = dyn_cast<VPReplicateRecipe>(VPV)) { - const VPRegionBlock *RegionOfR = Rep->getRegion(); - // Don't consider recipes in replicate regions as uniform yet; their first - // lane cannot be accessed when executing the replicate region for other - // lanes. - if (RegionOfR && RegionOfR->isReplicator()) - return false; - return Rep->isSingleScalar() || (PreservesUniformity(Rep->getOpcode()) && - all_of(Rep->operands(), isSingleScalar)); - } - if (isa<VPWidenGEPRecipe, VPDerivedIVRecipe, VPBlendRecipe, - VPWidenSelectRecipe>(VPV)) - return all_of(VPV->getDefiningRecipe()->operands(), isSingleScalar); - if (auto *WidenR = dyn_cast<VPWidenRecipe>(VPV)) { - return PreservesUniformity(WidenR->getOpcode()) && - all_of(WidenR->operands(), isSingleScalar); - } - if (auto *VPI = dyn_cast<VPInstruction>(VPV)) - return VPI->isSingleScalar() || VPI->isVectorToScalar() || - (PreservesUniformity(VPI->getOpcode()) && - all_of(VPI->operands(), isSingleScalar)); - if (isa<VPPartialReductionRecipe>(VPV)) - return false; - if (isa<VPReductionRecipe>(VPV)) - return true; - if (auto *Expr = dyn_cast<VPExpressionRecipe>(VPV)) - return Expr->isSingleScalar(); - - // VPExpandSCEVRecipes must be placed in the entry and are alway uniform. - return isa<VPExpandSCEVRecipe>(VPV); -} +bool isSingleScalar(const VPValue *VPV); /// Return true if \p V is a header mask in \p Plan. bool isHeaderMask(const VPValue *V, const VPlan &Plan); |