diff --git a/clang/lib/Driver/ToolChains/Linux.cpp b/clang/lib/Driver/ToolChains/Linux.cpp index 99c23e71eb0ccf..8188c972f44662 100644 --- a/clang/lib/Driver/ToolChains/Linux.cpp +++ b/clang/lib/Driver/ToolChains/Linux.cpp @@ -867,13 +867,9 @@ SanitizerMask Linux::getSupportedSanitizers() const { void Linux::addProfileRTLibs(const llvm::opt::ArgList &Args, llvm::opt::ArgStringList &CmdArgs) const { - bool Profile = needsProfileRT(Args); - if (!Profile && !needsGCovInstrumentation(Args)) - return; - // Add linker option -u__llvm_profile_runtime to cause runtime // initialization module to be linked in. - if (Profile) + if (needsProfileRT(Args)) CmdArgs.push_back(Args.MakeArgString( Twine("-u", llvm::getInstrProfRuntimeHookVarName()))); ToolChain::addProfileRTLibs(Args, CmdArgs); diff --git a/llvm/lib/Target/AMDGPU/AMDGPU.h b/llvm/lib/Target/AMDGPU/AMDGPU.h index fd18a1c2c6369c..7c86147a58e5cd 100644 --- a/llvm/lib/Target/AMDGPU/AMDGPU.h +++ b/llvm/lib/Target/AMDGPU/AMDGPU.h @@ -33,6 +33,8 @@ void initializeAMDGPUPreLegalizerCombinerPass(PassRegistry &); FunctionPass *createAMDGPUPreLegalizeCombiner(bool IsOptNone); void initializeAMDGPUPostLegalizerCombinerPass(PassRegistry &); FunctionPass *createAMDGPUPostLegalizeCombiner(bool IsOptNone); +FunctionPass *createAMDGPURegBankCombiner(bool IsOptNone); +void initializeAMDGPURegBankCombinerPass(PassRegistry &); // R600 Passes FunctionPass *createR600VectorRegMerger(); diff --git a/llvm/lib/Target/AMDGPU/AMDGPUCombine.td b/llvm/lib/Target/AMDGPU/AMDGPUCombine.td index 981aca5ab0a78f..faaf9168d0dd84 100644 --- a/llvm/lib/Target/AMDGPU/AMDGPUCombine.td +++ b/llvm/lib/Target/AMDGPU/AMDGPUCombine.td @@ -62,3 +62,8 @@ def AMDGPUPostLegalizerCombinerHelper: GICombinerHelper< uchar_to_float, cvt_f32_ubyteN]> { let DisableRuleOption = "amdgpupostlegalizercombiner-disable-rule"; } + +def AMDGPURegBankCombinerHelper : GICombinerHelper< + "AMDGPUGenRegBankCombinerHelper", []> { + let DisableRuleOption = "amdgpuregbankcombiner-disable-rule"; +} diff --git a/llvm/lib/Target/AMDGPU/AMDGPURegBankCombiner.cpp b/llvm/lib/Target/AMDGPU/AMDGPURegBankCombiner.cpp new file mode 100644 index 00000000000000..18c58c6ff5db78 --- /dev/null +++ b/llvm/lib/Target/AMDGPU/AMDGPURegBankCombiner.cpp @@ -0,0 +1,153 @@ +//=== lib/CodeGen/GlobalISel/AMDGPURegBankCombiner.cpp ---------------===// +// +// 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 pass does combining of machine instructions at the generic MI level, +// after register banks are known. +// +//===----------------------------------------------------------------------===// + +#include "AMDGPUTargetMachine.h" +#include "AMDGPULegalizerInfo.h" +#include "llvm/CodeGen/GlobalISel/Combiner.h" +#include "llvm/CodeGen/GlobalISel/CombinerHelper.h" +#include "llvm/CodeGen/GlobalISel/CombinerInfo.h" +#include "llvm/CodeGen/GlobalISel/GISelKnownBits.h" +#include "llvm/CodeGen/GlobalISel/MIPatternMatch.h" +#include "llvm/CodeGen/MachineDominators.h" +#include "llvm/CodeGen/MachineFunctionPass.h" +#include "llvm/CodeGen/TargetPassConfig.h" +#include "llvm/Support/Debug.h" +#include "MCTargetDesc/AMDGPUMCTargetDesc.h" + +#define DEBUG_TYPE "amdgpu-regbank-combiner" + +using namespace llvm; +using namespace MIPatternMatch; + + +#define AMDGPUREGBANKCOMBINERHELPER_GENCOMBINERHELPER_DEPS +#include "AMDGPUGenRegBankGICombiner.inc" +#undef AMDGPUREGBANKCOMBINERHELPER_GENCOMBINERHELPER_DEPS + +namespace { +#define AMDGPUREGBANKCOMBINERHELPER_GENCOMBINERHELPER_H +#include "AMDGPUGenRegBankGICombiner.inc" +#undef AMDGPUREGBANKCOMBINERHELPER_GENCOMBINERHELPER_H + +class AMDGPURegBankCombinerInfo : public CombinerInfo { + GISelKnownBits *KB; + MachineDominatorTree *MDT; + +public: + AMDGPUGenRegBankCombinerHelper Generated; + + AMDGPURegBankCombinerInfo(bool EnableOpt, bool OptSize, bool MinSize, + const AMDGPULegalizerInfo *LI, + GISelKnownBits *KB, MachineDominatorTree *MDT) + : CombinerInfo(/*AllowIllegalOps*/ false, /*ShouldLegalizeIllegal*/ true, + /*LegalizerInfo*/ LI, EnableOpt, OptSize, MinSize), + KB(KB), MDT(MDT) { + if (!Generated.parseCommandLineOption()) + report_fatal_error("Invalid rule identifier"); + } + + bool combine(GISelChangeObserver &Observer, MachineInstr &MI, + MachineIRBuilder &B) const override; +}; + +bool AMDGPURegBankCombinerInfo::combine(GISelChangeObserver &Observer, + MachineInstr &MI, + MachineIRBuilder &B) const { + CombinerHelper Helper(Observer, B, KB, MDT); + + if (Generated.tryCombineAll(Observer, MI, B, Helper)) + return true; + + return false; +} + +#define AMDGPUREGBANKCOMBINERHELPER_GENCOMBINERHELPER_CPP +#include "AMDGPUGenRegBankGICombiner.inc" +#undef AMDGPUREGBANKCOMBINERHELPER_GENCOMBINERHELPER_CPP + +// Pass boilerplate +// ================ + +class AMDGPURegBankCombiner : public MachineFunctionPass { +public: + static char ID; + + AMDGPURegBankCombiner(bool IsOptNone = false); + + StringRef getPassName() const override { + return "AMDGPURegBankCombiner"; + } + + bool runOnMachineFunction(MachineFunction &MF) override; + + void getAnalysisUsage(AnalysisUsage &AU) const override; +private: + bool IsOptNone; +}; +} // end anonymous namespace + +void AMDGPURegBankCombiner::getAnalysisUsage(AnalysisUsage &AU) const { + AU.addRequired(); + AU.setPreservesCFG(); + getSelectionDAGFallbackAnalysisUsage(AU); + AU.addRequired(); + AU.addPreserved(); + if (!IsOptNone) { + AU.addRequired(); + AU.addPreserved(); + } + MachineFunctionPass::getAnalysisUsage(AU); +} + +AMDGPURegBankCombiner::AMDGPURegBankCombiner(bool IsOptNone) + : MachineFunctionPass(ID), IsOptNone(IsOptNone) { + initializeAMDGPURegBankCombinerPass(*PassRegistry::getPassRegistry()); +} + +bool AMDGPURegBankCombiner::runOnMachineFunction(MachineFunction &MF) { + if (MF.getProperties().hasProperty( + MachineFunctionProperties::Property::FailedISel)) + return false; + auto *TPC = &getAnalysis(); + const Function &F = MF.getFunction(); + bool EnableOpt = + MF.getTarget().getOptLevel() != CodeGenOpt::None && !skipFunction(F); + + const GCNSubtarget &ST = MF.getSubtarget(); + const AMDGPULegalizerInfo *LI + = static_cast(ST.getLegalizerInfo()); + + GISelKnownBits *KB = &getAnalysis().get(MF); + MachineDominatorTree *MDT = + IsOptNone ? nullptr : &getAnalysis(); + AMDGPURegBankCombinerInfo PCInfo(EnableOpt, F.hasOptSize(), + F.hasMinSize(), LI, KB, MDT); + Combiner C(PCInfo, TPC); + return C.combineMachineInstrs(MF, /*CSEInfo*/ nullptr); +} + +char AMDGPURegBankCombiner::ID = 0; +INITIALIZE_PASS_BEGIN(AMDGPURegBankCombiner, DEBUG_TYPE, + "Combine AMDGPU machine instrs after regbankselect", + false, false) +INITIALIZE_PASS_DEPENDENCY(TargetPassConfig) +INITIALIZE_PASS_DEPENDENCY(GISelKnownBitsAnalysis) +INITIALIZE_PASS_END(AMDGPURegBankCombiner, DEBUG_TYPE, + "Combine AMDGPU machine instrs after regbankselect", false, + false) + +namespace llvm { +FunctionPass *createAMDGPURegBankCombiner(bool IsOptNone) { + return new AMDGPURegBankCombiner(IsOptNone); +} +} // end namespace llvm diff --git a/llvm/lib/Target/AMDGPU/CMakeLists.txt b/llvm/lib/Target/AMDGPU/CMakeLists.txt index ac7ff241a055d4..2125a47c6cc9f7 100644 --- a/llvm/lib/Target/AMDGPU/CMakeLists.txt +++ b/llvm/lib/Target/AMDGPU/CMakeLists.txt @@ -19,6 +19,8 @@ tablegen(LLVM AMDGPUGenPreLegalizeGICombiner.inc -gen-global-isel-combiner -combiners="AMDGPUPreLegalizerCombinerHelper") tablegen(LLVM AMDGPUGenPostLegalizeGICombiner.inc -gen-global-isel-combiner -combiners="AMDGPUPostLegalizerCombinerHelper") +tablegen(LLVM AMDGPUGenRegBankGICombiner.inc -gen-global-isel-combiner + -combiners="AMDGPURegBankCombinerHelper") set(LLVM_TARGET_DEFINITIONS R600.td) tablegen(LLVM R600GenAsmWriter.inc -gen-asm-writer) @@ -68,6 +70,7 @@ add_llvm_target(AMDGPUCodeGen AMDGPUPreLegalizerCombiner.cpp AMDGPUPromoteAlloca.cpp AMDGPUPropagateAttributes.cpp + AMDGPURegBankCombiner.cpp AMDGPURegisterBankInfo.cpp AMDGPURewriteOutArguments.cpp AMDGPUSubtarget.cpp