https://github.com/sarnex updated
https://github.com/llvm/llvm-project/pull/126956
>From 044158c176316e6c9f44ce1487ff631c8a0d15c4 Mon Sep 17 00:00:00 2001
From: "Sarnie, Nick" <nick.sar...@intel.com>
Date: Tue, 4 Mar 2025 11:16:34 -0800
Subject: [PATCH 1/4] [clang] Add isOffloadingTarget function to LangOpts
Signed-off-by: Sarnie, Nick <nick.sar...@intel.com>
---
clang/include/clang/Basic/LangOptions.h | 4 ++++
clang/lib/CodeGen/CGOpenMPRuntime.cpp | 7 +++----
clang/lib/CodeGen/CodeGenModule.cpp | 2 +-
clang/lib/CodeGen/CodeGenModule.h | 3 +--
clang/lib/Frontend/CompilerInvocation.cpp | 4 ++++
clang/lib/Sema/SemaType.cpp | 16 ++++------------
6 files changed, 17 insertions(+), 19 deletions(-)
diff --git a/clang/include/clang/Basic/LangOptions.h
b/clang/include/clang/Basic/LangOptions.h
index e925e0f3b5d85..d722ca2d926b2 100644
--- a/clang/include/clang/Basic/LangOptions.h
+++ b/clang/include/clang/Basic/LangOptions.h
@@ -636,6 +636,8 @@ class LangOptions : public LangOptionsBase {
bool AtomicFineGrainedMemory = false;
bool AtomicIgnoreDenormalMode = false;
+ bool IsOffloadingTarget = false;
+
LangOptions();
/// Set language defaults for the given input language and
@@ -838,6 +840,8 @@ class LangOptions : public LangOptionsBase {
return FPExceptionModeKind::FPE_Ignore;
return EM;
}
+
+ bool isOffloadingTarget() const { return IsOffloadingTarget; }
};
/// Floating point control options
diff --git a/clang/lib/CodeGen/CGOpenMPRuntime.cpp
b/clang/lib/CodeGen/CGOpenMPRuntime.cpp
index 06a652c146fb9..f61b46c4369a5 100644
--- a/clang/lib/CodeGen/CGOpenMPRuntime.cpp
+++ b/clang/lib/CodeGen/CGOpenMPRuntime.cpp
@@ -2613,9 +2613,8 @@ void CGOpenMPRuntime::emitDistributeStaticInit(
emitUpdateLocation(CGF, Loc, OMP_IDENT_WORK_DISTRIBUTE);
llvm::Value *ThreadId = getThreadID(CGF, Loc);
llvm::FunctionCallee StaticInitFunction;
- bool isGPUDistribute =
- CGM.getLangOpts().OpenMPIsTargetDevice &&
- (CGM.getTriple().isAMDGCN() || CGM.getTriple().isNVPTX());
+ bool isGPUDistribute = CGM.getLangOpts().OpenMPIsTargetDevice &&
+ CGM.getLangOpts().isOffloadingTarget();
StaticInitFunction = OMPBuilder.createForStaticInitFunction(
Values.IVSize, Values.IVSigned, isGPUDistribute);
@@ -2645,7 +2644,7 @@ void CGOpenMPRuntime::emitForStaticFinish(CodeGenFunction
&CGF,
auto DL = ApplyDebugLocation::CreateDefaultArtificial(CGF, Loc);
if (isOpenMPDistributeDirective(DKind) &&
CGM.getLangOpts().OpenMPIsTargetDevice &&
- (CGM.getTriple().isAMDGCN() || CGM.getTriple().isNVPTX()))
+ CGM.getLangOpts().isOffloadingTarget())
CGF.EmitRuntimeCall(
OMPBuilder.getOrCreateRuntimeFunction(
CGM.getModule(), OMPRTL___kmpc_distribute_static_fini),
diff --git a/clang/lib/CodeGen/CodeGenModule.cpp
b/clang/lib/CodeGen/CodeGenModule.cpp
index bca0a932b3495..9080b226a7b8b 100644
--- a/clang/lib/CodeGen/CodeGenModule.cpp
+++ b/clang/lib/CodeGen/CodeGenModule.cpp
@@ -843,7 +843,7 @@ static void setVisibilityFromDLLStorageClass(const
clang::LangOptions &LO,
static bool isStackProtectorOn(const LangOptions &LangOpts,
const llvm::Triple &Triple,
clang::LangOptions::StackProtectorMode Mode) {
- if (Triple.isAMDGPU() || Triple.isNVPTX())
+ if (LangOpts.isOffloadingTarget())
return false;
return LangOpts.getStackProtector() == Mode;
}
diff --git a/clang/lib/CodeGen/CodeGenModule.h
b/clang/lib/CodeGen/CodeGenModule.h
index a72b4f7338767..3448c0ed06335 100644
--- a/clang/lib/CodeGen/CodeGenModule.h
+++ b/clang/lib/CodeGen/CodeGenModule.h
@@ -1078,8 +1078,7 @@ class CodeGenModule : public CodeGenTypeCache {
bool shouldEmitRTTI(bool ForEH = false) {
return (ForEH || getLangOpts().RTTI) && !getLangOpts().CUDAIsDevice &&
!(getLangOpts().OpenMP && getLangOpts().OpenMPIsTargetDevice &&
- (getTriple().isNVPTX() || getTriple().isAMDGPU() ||
- getTriple().isSPIRV()));
+ getLangOpts().isOffloadingTarget());
}
/// Get the address of the RTTI descriptor for the given type.
diff --git a/clang/lib/Frontend/CompilerInvocation.cpp
b/clang/lib/Frontend/CompilerInvocation.cpp
index 4eb743acf327f..33d26ee7f9f7d 100644
--- a/clang/lib/Frontend/CompilerInvocation.cpp
+++ b/clang/lib/Frontend/CompilerInvocation.cpp
@@ -4357,6 +4357,10 @@ bool CompilerInvocation::ParseLangArgs(LangOptions
&Opts, ArgList &Args,
Opts.OpenACCMacroOverride = A->getValue();
}
+ Opts.IsOffloadingTarget =
+ (Opts.OpenMPIsTargetDevice || Opts.SYCLIsDevice || Opts.CUDAIsDevice) &&
+ (T.isNVPTX() || T.isAMDGCN() || T.isSPIROrSPIRV());
+
// FIXME: Eliminate this dependency.
unsigned Opt = getOptimizationLevel(Args, IK, Diags),
OptSize = getOptimizationLevelSize(Args);
diff --git a/clang/lib/Sema/SemaType.cpp b/clang/lib/Sema/SemaType.cpp
index 11943c0b53591..0f115a6f04509 100644
--- a/clang/lib/Sema/SemaType.cpp
+++ b/clang/lib/Sema/SemaType.cpp
@@ -1122,8 +1122,7 @@ static QualType ConvertDeclSpecToType(TypeProcessingState
&state) {
}
case DeclSpec::TST_int128:
if (!S.Context.getTargetInfo().hasInt128Type() &&
- !(S.getLangOpts().SYCLIsDevice || S.getLangOpts().CUDAIsDevice ||
- (S.getLangOpts().OpenMP && S.getLangOpts().OpenMPIsTargetDevice)))
+ !(S.getLangOpts().isOffloadingTarget()))
S.Diag(DS.getTypeSpecTypeLoc(), diag::err_type_unsupported)
<< "__int128";
if (DS.getTypeSpecSign() == TypeSpecifierSign::Unsigned)
@@ -1168,8 +1167,7 @@ static QualType ConvertDeclSpecToType(TypeProcessingState
&state) {
break;
case DeclSpec::TST_float128:
if (!S.Context.getTargetInfo().hasFloat128Type() &&
- !S.getLangOpts().SYCLIsDevice && !S.getLangOpts().CUDAIsDevice &&
- !(S.getLangOpts().OpenMP && S.getLangOpts().OpenMPIsTargetDevice))
+ !S.getLangOpts().isOffloadingTarget())
S.Diag(DS.getTypeSpecTypeLoc(), diag::err_type_unsupported)
<< "__float128";
Result = Context.Float128Ty;
@@ -8362,12 +8360,7 @@ static bool verifyValidIntegerConstantExpr(Sema &S,
const ParsedAttr &Attr,
/// match one of the standard Neon vector types.
static void HandleNeonVectorTypeAttr(QualType &CurType, const ParsedAttr &Attr,
Sema &S, VectorKind VecKind) {
- bool IsTargetCUDAAndHostARM = false;
- if (S.getLangOpts().CUDAIsDevice) {
- const TargetInfo *AuxTI = S.getASTContext().getAuxTargetInfo();
- IsTargetCUDAAndHostARM =
- AuxTI && (AuxTI->getTriple().isAArch64() ||
AuxTI->getTriple().isARM());
- }
+ bool IsTargetOffloading = S.getLangOpts().isOffloadingTarget();
// Target must have NEON (or MVE, whose vectors are similar enough
// not to need a separate attribute)
@@ -8401,8 +8394,7 @@ static void HandleNeonVectorTypeAttr(QualType &CurType,
const ParsedAttr &Attr,
return;
// Only certain element types are supported for Neon vectors.
- if (!isPermittedNeonBaseType(CurType, VecKind, S) &&
- !IsTargetCUDAAndHostARM) {
+ if (!isPermittedNeonBaseType(CurType, VecKind, S) && !IsTargetOffloading) {
S.Diag(Attr.getLoc(), diag::err_attribute_invalid_vector_type) << CurType;
Attr.setInvalid();
return;
>From 417a6b092c1cfe6aed78167112a2855daec00182 Mon Sep 17 00:00:00 2001
From: "Sarnie, Nick" <nick.sar...@intel.com>
Date: Fri, 21 Mar 2025 08:05:58 -0700
Subject: [PATCH 2/4] add two flags
Signed-off-by: Sarnie, Nick <nick.sar...@intel.com>
---
clang/include/clang/Basic/LangOptions.h | 7 ++++---
clang/lib/AST/Decl.cpp | 3 +--
clang/lib/CodeGen/CGOpenMPRuntime.cpp | 7 +++----
clang/lib/CodeGen/CodeGenModule.cpp | 2 +-
clang/lib/CodeGen/CodeGenModule.h | 5 ++---
clang/lib/Driver/ToolChains/Clang.cpp | 10 ++++------
clang/lib/Frontend/CompilerInstance.cpp | 3 +--
clang/lib/Frontend/CompilerInvocation.cpp | 7 +------
clang/lib/Frontend/InitPreprocessor.cpp | 4 +---
clang/lib/Sema/SemaDecl.cpp | 11 ++++------
clang/lib/Sema/SemaType.cpp | 6 +++---
.../ClangLinkerWrapper.cpp | 2 +-
flang/lib/Frontend/CompilerInvocation.cpp | 20 ++++++-------------
llvm/include/llvm/TargetParser/Triple.h | 3 +++
llvm/include/llvm/Transforms/IPO/Attributor.h | 4 +---
llvm/lib/ProfileData/InstrProf.cpp | 2 +-
llvm/lib/Transforms/IPO/Attributor.cpp | 4 ++--
.../SanitizerBinaryMetadata.cpp | 3 +--
18 files changed, 40 insertions(+), 63 deletions(-)
diff --git a/clang/include/clang/Basic/LangOptions.h
b/clang/include/clang/Basic/LangOptions.h
index d722ca2d926b2..bbebf7af9ede3 100644
--- a/clang/include/clang/Basic/LangOptions.h
+++ b/clang/include/clang/Basic/LangOptions.h
@@ -636,8 +636,6 @@ class LangOptions : public LangOptionsBase {
bool AtomicFineGrainedMemory = false;
bool AtomicIgnoreDenormalMode = false;
- bool IsOffloadingTarget = false;
-
LangOptions();
/// Set language defaults for the given input language and
@@ -841,7 +839,10 @@ class LangOptions : public LangOptionsBase {
return EM;
}
- bool isOffloadingTarget() const { return IsOffloadingTarget; }
+ /// True when compiling for an offloading target device.
+ bool isTargetDevice() const {
+ return OpenMPIsTargetDevice || CUDAIsDevice || SYCLIsDevice;
+ }
};
/// Floating point control options
diff --git a/clang/lib/AST/Decl.cpp b/clang/lib/AST/Decl.cpp
index e8aeacf24374f..5f5568518e0b9 100644
--- a/clang/lib/AST/Decl.cpp
+++ b/clang/lib/AST/Decl.cpp
@@ -810,8 +810,7 @@ LinkageComputer::getLVForNamespaceScopeDecl(const NamedDecl
*D,
// functions as the host-callable kernel functions are emitted at codegen.
if (Context.getLangOpts().OpenMP &&
Context.getLangOpts().OpenMPIsTargetDevice &&
- ((Context.getTargetInfo().getTriple().isAMDGPU() ||
- Context.getTargetInfo().getTriple().isNVPTX()) ||
+ (Context.getTargetInfo().getTriple().isGPU() ||
OMPDeclareTargetDeclAttr::isDeclareTargetDeclaration(Function)))
LV.mergeVisibility(HiddenVisibility, /*newExplicit=*/false);
diff --git a/clang/lib/CodeGen/CGOpenMPRuntime.cpp
b/clang/lib/CodeGen/CGOpenMPRuntime.cpp
index f61b46c4369a5..b706fa3759c0d 100644
--- a/clang/lib/CodeGen/CGOpenMPRuntime.cpp
+++ b/clang/lib/CodeGen/CGOpenMPRuntime.cpp
@@ -2613,8 +2613,8 @@ void CGOpenMPRuntime::emitDistributeStaticInit(
emitUpdateLocation(CGF, Loc, OMP_IDENT_WORK_DISTRIBUTE);
llvm::Value *ThreadId = getThreadID(CGF, Loc);
llvm::FunctionCallee StaticInitFunction;
- bool isGPUDistribute = CGM.getLangOpts().OpenMPIsTargetDevice &&
- CGM.getLangOpts().isOffloadingTarget();
+ bool isGPUDistribute =
+ CGM.getLangOpts().OpenMPIsTargetDevice && CGM.getTriple().isGPU();
StaticInitFunction = OMPBuilder.createForStaticInitFunction(
Values.IVSize, Values.IVSigned, isGPUDistribute);
@@ -2643,8 +2643,7 @@ void CGOpenMPRuntime::emitForStaticFinish(CodeGenFunction
&CGF,
getThreadID(CGF, Loc)};
auto DL = ApplyDebugLocation::CreateDefaultArtificial(CGF, Loc);
if (isOpenMPDistributeDirective(DKind) &&
- CGM.getLangOpts().OpenMPIsTargetDevice &&
- CGM.getLangOpts().isOffloadingTarget())
+ CGM.getLangOpts().OpenMPIsTargetDevice && CGM.getTriple().isGPU())
CGF.EmitRuntimeCall(
OMPBuilder.getOrCreateRuntimeFunction(
CGM.getModule(), OMPRTL___kmpc_distribute_static_fini),
diff --git a/clang/lib/CodeGen/CodeGenModule.cpp
b/clang/lib/CodeGen/CodeGenModule.cpp
index 9080b226a7b8b..78c709a13982d 100644
--- a/clang/lib/CodeGen/CodeGenModule.cpp
+++ b/clang/lib/CodeGen/CodeGenModule.cpp
@@ -843,7 +843,7 @@ static void setVisibilityFromDLLStorageClass(const
clang::LangOptions &LO,
static bool isStackProtectorOn(const LangOptions &LangOpts,
const llvm::Triple &Triple,
clang::LangOptions::StackProtectorMode Mode) {
- if (LangOpts.isOffloadingTarget())
+ if (Triple.isGPU())
return false;
return LangOpts.getStackProtector() == Mode;
}
diff --git a/clang/lib/CodeGen/CodeGenModule.h
b/clang/lib/CodeGen/CodeGenModule.h
index 3448c0ed06335..99515cb18dd64 100644
--- a/clang/lib/CodeGen/CodeGenModule.h
+++ b/clang/lib/CodeGen/CodeGenModule.h
@@ -1076,9 +1076,8 @@ class CodeGenModule : public CodeGenTypeCache {
// Return whether RTTI information should be emitted for this target.
bool shouldEmitRTTI(bool ForEH = false) {
- return (ForEH || getLangOpts().RTTI) && !getLangOpts().CUDAIsDevice &&
- !(getLangOpts().OpenMP && getLangOpts().OpenMPIsTargetDevice &&
- getLangOpts().isOffloadingTarget());
+ return (ForEH || getLangOpts().RTTI) &&
+ (!getLangOpts().isTargetDevice() || !getTriple().isGPU());
}
/// Get the address of the RTTI descriptor for the given type.
diff --git a/clang/lib/Driver/ToolChains/Clang.cpp
b/clang/lib/Driver/ToolChains/Clang.cpp
index 4ebbd241d2f0b..4f2d898ec540b 100644
--- a/clang/lib/Driver/ToolChains/Clang.cpp
+++ b/clang/lib/Driver/ToolChains/Clang.cpp
@@ -1101,8 +1101,7 @@ void Clang::AddPreprocessingOptions(Compilation &C, const
JobAction &JA,
if (JA.isDeviceOffloading(Action::OFK_OpenMP) &&
!Args.hasArg(options::OPT_nostdinc) &&
!Args.hasArg(options::OPT_nogpuinc) &&
- (getToolChain().getTriple().isNVPTX() ||
- getToolChain().getTriple().isAMDGCN())) {
+ getToolChain().getTriple().isGPU()) {
if (!Args.hasArg(options::OPT_nobuiltininc)) {
// Add openmp_wrappers/* to our system include path. This lets us wrap
// standard library headers.
@@ -1289,8 +1288,7 @@ void Clang::AddPreprocessingOptions(Compilation &C, const
JobAction &JA,
// Without an offloading language we will include these headers directly.
// Offloading languages will instead only use the declarations stored in
// the resource directory at clang/lib/Headers/llvm_libc_wrappers.
- if ((getToolChain().getTriple().isNVPTX() ||
- getToolChain().getTriple().isAMDGCN()) &&
+ if (getToolChain().getTriple().isGPU() &&
C.getActiveOffloadKinds() == Action::OFK_None) {
SmallString<128> P(llvm::sys::path::parent_path(D.Dir));
llvm::sys::path::append(P, "include");
@@ -6388,10 +6386,10 @@ void Clang::ConstructJob(Compilation &C, const
JobAction &JA,
Args.AddLastArg(CmdArgs, options::OPT_fconvergent_functions,
options::OPT_fno_convergent_functions);
- // NVPTX/AMDGCN doesn't support PGO or coverage. There's no runtime support
+ // GPUs don't support PGO or coverage. There's no runtime support
// for sampling, overhead of call arc collection is way too high and there's
// no way to collect the output.
- if (!Triple.isNVPTX() && !Triple.isAMDGCN())
+ if (!Triple.isGPU())
addPGOAndCoverageFlags(TC, C, JA, Output, Args, SanitizeArgs, CmdArgs);
Args.AddLastArg(CmdArgs, options::OPT_fclang_abi_compat_EQ);
diff --git a/clang/lib/Frontend/CompilerInstance.cpp
b/clang/lib/Frontend/CompilerInstance.cpp
index c11c857ea0606..d23eaf1015b45 100644
--- a/clang/lib/Frontend/CompilerInstance.cpp
+++ b/clang/lib/Frontend/CompilerInstance.cpp
@@ -116,8 +116,7 @@ bool CompilerInstance::createTarget() {
// Check whether AuxTarget exists, if not, then create TargetInfo for the
// other side of CUDA/OpenMP/SYCL compilation.
if (!getAuxTarget() &&
- (getLangOpts().CUDA || getLangOpts().OpenMPIsTargetDevice ||
- getLangOpts().SYCLIsDevice) &&
+ (getLangOpts().CUDA || getLangOpts().isTargetDevice()) &&
!getFrontendOpts().AuxTriple.empty()) {
auto TO = std::make_shared<TargetOptions>();
TO->Triple = llvm::Triple::normalize(getFrontendOpts().AuxTriple);
diff --git a/clang/lib/Frontend/CompilerInvocation.cpp
b/clang/lib/Frontend/CompilerInvocation.cpp
index 33d26ee7f9f7d..dae5ee38a3c4d 100644
--- a/clang/lib/Frontend/CompilerInvocation.cpp
+++ b/clang/lib/Frontend/CompilerInvocation.cpp
@@ -4183,8 +4183,7 @@ bool CompilerInvocation::ParseLangArgs(LangOptions &Opts,
ArgList &Args,
Opts.Blocks = Args.hasArg(OPT_fblocks) || (Opts.OpenCL
&& Opts.OpenCLVersion == 200);
- bool HasConvergentOperations = Opts.OpenMPIsTargetDevice || Opts.OpenCL ||
- Opts.CUDAIsDevice || Opts.SYCLIsDevice ||
+ bool HasConvergentOperations = Opts.isTargetDevice() || Opts.OpenCL ||
Opts.HLSL || T.isAMDGPU() || T.isNVPTX();
Opts.ConvergentFunctions =
Args.hasFlag(OPT_fconvergent_functions, OPT_fno_convergent_functions,
@@ -4357,10 +4356,6 @@ bool CompilerInvocation::ParseLangArgs(LangOptions
&Opts, ArgList &Args,
Opts.OpenACCMacroOverride = A->getValue();
}
- Opts.IsOffloadingTarget =
- (Opts.OpenMPIsTargetDevice || Opts.SYCLIsDevice || Opts.CUDAIsDevice) &&
- (T.isNVPTX() || T.isAMDGCN() || T.isSPIROrSPIRV());
-
// FIXME: Eliminate this dependency.
unsigned Opt = getOptimizationLevel(Args, IK, Diags),
OptSize = getOptimizationLevelSize(Args);
diff --git a/clang/lib/Frontend/InitPreprocessor.cpp
b/clang/lib/Frontend/InitPreprocessor.cpp
index 1a816cb6269d4..863b9c14b6d4f 100644
--- a/clang/lib/Frontend/InitPreprocessor.cpp
+++ b/clang/lib/Frontend/InitPreprocessor.cpp
@@ -1564,9 +1564,7 @@ void clang::InitializePreprocessor(Preprocessor &PP,
if (InitOpts.UsePredefines) {
// FIXME: This will create multiple definitions for most of the predefined
// macros. This is not the right way to handle this.
- if ((LangOpts.CUDA || LangOpts.OpenMPIsTargetDevice ||
- LangOpts.SYCLIsDevice) &&
- PP.getAuxTargetInfo())
+ if ((LangOpts.CUDA || LangOpts.isTargetDevice()) && PP.getAuxTargetInfo())
InitializePredefinedMacros(*PP.getAuxTargetInfo(), LangOpts, FEOpts,
PP.getPreprocessorOpts(), Builder);
diff --git a/clang/lib/Sema/SemaDecl.cpp b/clang/lib/Sema/SemaDecl.cpp
index 36de02d91e151..20e1b478761e5 100644
--- a/clang/lib/Sema/SemaDecl.cpp
+++ b/clang/lib/Sema/SemaDecl.cpp
@@ -7854,8 +7854,7 @@ NamedDecl *Sema::ActOnVariableDeclarator(
diag::err_thread_non_global)
<< DeclSpec::getSpecifierName(TSCS);
else if (!Context.getTargetInfo().isTLSSupported()) {
- if (getLangOpts().CUDA || getLangOpts().OpenMPIsTargetDevice ||
- getLangOpts().SYCLIsDevice) {
+ if (getLangOpts().CUDA || getLangOpts().isTargetDevice()) {
// Postpone error emission until we've collected attributes required to
// figure out whether it's a host or device variable and whether the
// error should be ignored.
@@ -7986,8 +7985,7 @@ NamedDecl *Sema::ActOnVariableDeclarator(
if (const auto *TT = R->getAs<TypedefType>())
copyAttrFromTypedefToDecl<AllocSizeAttr>(*this, NewVD, TT);
- if (getLangOpts().CUDA || getLangOpts().OpenMPIsTargetDevice ||
- getLangOpts().SYCLIsDevice) {
+ if (getLangOpts().CUDA || getLangOpts().isTargetDevice()) {
if (EmitTLSUnsupportedError &&
((getLangOpts().CUDA && DeclAttrsMatchCUDAMode(getLangOpts(), NewVD))
||
(getLangOpts().OpenMPIsTargetDevice &&
@@ -16554,9 +16552,8 @@ Decl *Sema::ActOnFinishFunctionBody(Decl *dcl, Stmt
*Body,
DiscardCleanupsInEvaluationContext();
}
- if (FD && ((LangOpts.OpenMP && (LangOpts.OpenMPIsTargetDevice ||
- !LangOpts.OMPTargetTriples.empty())) ||
- LangOpts.CUDA || LangOpts.SYCLIsDevice)) {
+ if (FD && (LangOpts.isTargetDevice() || LangOpts.CUDA ||
+ (LangOpts.OpenMP && !LangOpts.OMPTargetTriples.empty()))) {
auto ES = getEmissionStatus(FD);
if (ES == Sema::FunctionEmissionStatus::Emitted ||
ES == Sema::FunctionEmissionStatus::Unknown)
diff --git a/clang/lib/Sema/SemaType.cpp b/clang/lib/Sema/SemaType.cpp
index 0f115a6f04509..ad287f2989fce 100644
--- a/clang/lib/Sema/SemaType.cpp
+++ b/clang/lib/Sema/SemaType.cpp
@@ -1122,7 +1122,7 @@ static QualType ConvertDeclSpecToType(TypeProcessingState
&state) {
}
case DeclSpec::TST_int128:
if (!S.Context.getTargetInfo().hasInt128Type() &&
- !(S.getLangOpts().isOffloadingTarget()))
+ !(S.getLangOpts().isTargetDevice()))
S.Diag(DS.getTypeSpecTypeLoc(), diag::err_type_unsupported)
<< "__int128";
if (DS.getTypeSpecSign() == TypeSpecifierSign::Unsigned)
@@ -1167,7 +1167,7 @@ static QualType ConvertDeclSpecToType(TypeProcessingState
&state) {
break;
case DeclSpec::TST_float128:
if (!S.Context.getTargetInfo().hasFloat128Type() &&
- !S.getLangOpts().isOffloadingTarget())
+ !S.getLangOpts().isTargetDevice())
S.Diag(DS.getTypeSpecTypeLoc(), diag::err_type_unsupported)
<< "__float128";
Result = Context.Float128Ty;
@@ -8360,7 +8360,7 @@ static bool verifyValidIntegerConstantExpr(Sema &S, const
ParsedAttr &Attr,
/// match one of the standard Neon vector types.
static void HandleNeonVectorTypeAttr(QualType &CurType, const ParsedAttr &Attr,
Sema &S, VectorKind VecKind) {
- bool IsTargetOffloading = S.getLangOpts().isOffloadingTarget();
+ bool IsTargetOffloading = S.getLangOpts().isTargetDevice();
// Target must have NEON (or MVE, whose vectors are similar enough
// not to need a separate attribute)
diff --git a/clang/tools/clang-linker-wrapper/ClangLinkerWrapper.cpp
b/clang/tools/clang-linker-wrapper/ClangLinkerWrapper.cpp
index 7db8f3e27d704..b38aca051dac2 100644
--- a/clang/tools/clang-linker-wrapper/ClangLinkerWrapper.cpp
+++ b/clang/tools/clang-linker-wrapper/ClangLinkerWrapper.cpp
@@ -508,7 +508,7 @@ Expected<StringRef> clang(ArrayRef<StringRef> InputFiles,
const ArgList &Args) {
CmdArgs.push_back(InputFile);
// If this is CPU offloading we copy the input libraries.
- if (!Triple.isAMDGPU() && !Triple.isNVPTX() && !Triple.isSPIRV()) {
+ if (!Triple.isGPU()) {
CmdArgs.push_back("-Wl,-Bsymbolic");
CmdArgs.push_back("-shared");
ArgStringList LinkerArgs;
diff --git a/flang/lib/Frontend/CompilerInvocation.cpp
b/flang/lib/Frontend/CompilerInvocation.cpp
index 8b07a50824899..aea975a427f0b 100644
--- a/flang/lib/Frontend/CompilerInvocation.cpp
+++ b/flang/lib/Frontend/CompilerInvocation.cpp
@@ -1126,22 +1126,17 @@ static bool parseOpenMPArgs(CompilerInvocation &res,
llvm::opt::ArgList &args,
if (args.hasArg(clang::driver::options::OPT_no_offloadlib))
res.getLangOpts().NoGPULib = 1;
}
-
- switch (llvm::Triple(res.getTargetOpts().triple).getArch()) {
- case llvm::Triple::nvptx:
- case llvm::Triple::nvptx64:
- case llvm::Triple::amdgcn:
+ if (llvm::Triple(res.getTargetOpts().triple).isGPU()) {
if (!res.getLangOpts().OpenMPIsTargetDevice) {
const unsigned diagID = diags.getCustomDiagID(
clang::DiagnosticsEngine::Error,
- "OpenMP AMDGPU/NVPTX is only prepared to deal with device code.");
+ "OpenMP GPU is only prepared to deal with device code.");
diags.Report(diagID);
}
res.getLangOpts().OpenMPIsGPU = 1;
- break;
- default:
- res.getLangOpts().OpenMPIsGPU = 0;
- break;
+ else {
+ res.getLangOpts().OpenMPIsGPU = 0;
+ }
}
// Get the OpenMP target triples if any.
@@ -1163,11 +1158,8 @@ static bool parseOpenMPArgs(CompilerInvocation &res,
llvm::opt::ArgList &args,
if (tt.getArch() == llvm::Triple::UnknownArch ||
!(tt.getArch() == llvm::Triple::aarch64 || tt.isPPC() ||
tt.getArch() == llvm::Triple::systemz ||
- tt.getArch() == llvm::Triple::nvptx ||
- tt.getArch() == llvm::Triple::nvptx64 ||
- tt.getArch() == llvm::Triple::amdgcn ||
tt.getArch() == llvm::Triple::x86 ||
- tt.getArch() == llvm::Triple::x86_64))
+ tt.getArch() == llvm::Triple::x86_64 || tt.isGPU()))
diags.Report(clang::diag::err_drv_invalid_omp_target)
<< arg->getValue(i);
else if (getArchPtrSize(t) != getArchPtrSize(tt))
diff --git a/llvm/include/llvm/TargetParser/Triple.h
b/llvm/include/llvm/TargetParser/Triple.h
index 09c0d223d9b4d..a68bd032fe846 100644
--- a/llvm/include/llvm/TargetParser/Triple.h
+++ b/llvm/include/llvm/TargetParser/Triple.h
@@ -1218,6 +1218,9 @@ class Triple {
/// Test whether target triples are compatible.
bool isCompatibleWith(const Triple &Other) const;
+ /// Test whether the target triple is for a GPU.
+ bool isGPU() const { return isSPIRV() || isNVPTX() || isAMDGPU(); }
+
/// Merge target triples.
std::string merge(const Triple &Other) const;
diff --git a/llvm/include/llvm/Transforms/IPO/Attributor.h
b/llvm/include/llvm/Transforms/IPO/Attributor.h
index 05951f87b5062..543926fe76f61 100644
--- a/llvm/include/llvm/Transforms/IPO/Attributor.h
+++ b/llvm/include/llvm/Transforms/IPO/Attributor.h
@@ -1339,9 +1339,7 @@ struct InformationCache {
bool stackIsAccessibleByOtherThreads() { return !targetIsGPU(); }
/// Return true if the target is a GPU.
- bool targetIsGPU() {
- return TargetTriple.isAMDGPU() || TargetTriple.isNVPTX();
- }
+ bool targetIsGPU() { return TargetTriple.isGPU(); }
/// Return all functions that might be called indirectly, only valid for
/// closed world modules (see isClosedWorldModule).
diff --git a/llvm/lib/ProfileData/InstrProf.cpp
b/llvm/lib/ProfileData/InstrProf.cpp
index 47eb751328b51..b39c34dbc128e 100644
--- a/llvm/lib/ProfileData/InstrProf.cpp
+++ b/llvm/lib/ProfileData/InstrProf.cpp
@@ -439,7 +439,7 @@ std::string getPGOFuncNameVarName(StringRef FuncName,
bool isGPUProfTarget(const Module &M) {
const auto &T = Triple(M.getTargetTriple());
- return T.isAMDGPU() || T.isNVPTX();
+ return T.isGPU();
}
void setPGOFuncVisibility(Module &M, GlobalVariable *FuncNameVar) {
diff --git a/llvm/lib/Transforms/IPO/Attributor.cpp
b/llvm/lib/Transforms/IPO/Attributor.cpp
index c29a5e53b82ea..dd8328ff805d6 100644
--- a/llvm/lib/Transforms/IPO/Attributor.cpp
+++ b/llvm/lib/Transforms/IPO/Attributor.cpp
@@ -199,7 +199,7 @@ ChangeStatus &llvm::operator&=(ChangeStatus &L,
ChangeStatus R) {
bool AA::isGPU(const Module &M) {
Triple T(M.getTargetTriple());
- return T.isAMDGPU() || T.isNVPTX();
+ return T.isGPU();
}
bool AA::isNoSyncInst(Attributor &A, const Instruction &I,
@@ -3301,7 +3301,7 @@
InformationCache::getIndirectlyCallableFunctions(Attributor &A) const {
}
std::optional<unsigned> InformationCache::getFlatAddressSpace() const {
- if (TargetTriple.isAMDGPU() || TargetTriple.isNVPTX())
+ if (TargetTriple.isGPU())
return 0;
return std::nullopt;
}
diff --git a/llvm/lib/Transforms/Instrumentation/SanitizerBinaryMetadata.cpp
b/llvm/lib/Transforms/Instrumentation/SanitizerBinaryMetadata.cpp
index 9ebc235f58142..270ba309e14dd 100644
--- a/llvm/lib/Transforms/Instrumentation/SanitizerBinaryMetadata.cpp
+++ b/llvm/lib/Transforms/Instrumentation/SanitizerBinaryMetadata.cpp
@@ -133,8 +133,7 @@ class SanitizerBinaryMetadata {
VersionStr(utostr(getVersion())), IRB(M.getContext()) {
// FIXME: Make it work with other formats.
assert(TargetTriple.isOSBinFormatELF() && "ELF only");
- assert(!(TargetTriple.isNVPTX() || TargetTriple.isAMDGPU()) &&
- "Device targets are not supported");
+ assert(!TargetTriple.isGPU() && "Device targets are not supported");
}
bool run();
>From 3167b3f644a4518e8eb779054f1d2404080503bd Mon Sep 17 00:00:00 2001
From: "Sarnie, Nick" <nick.sar...@intel.com>
Date: Fri, 21 Mar 2025 12:09:38 -0700
Subject: [PATCH 3/4] remove newline
Signed-off-by: Sarnie, Nick <nick.sar...@intel.com>
---
clang/lib/Driver/ToolChains/Clang.cpp | 1 -
1 file changed, 1 deletion(-)
diff --git a/clang/lib/Driver/ToolChains/Clang.cpp
b/clang/lib/Driver/ToolChains/Clang.cpp
index d603dab7076d9..cf95cf0af8f5b 100644
--- a/clang/lib/Driver/ToolChains/Clang.cpp
+++ b/clang/lib/Driver/ToolChains/Clang.cpp
@@ -6388,7 +6388,6 @@ void Clang::ConstructJob(Compilation &C, const JobAction
&JA,
addPGOAndCoverageFlags(TC, C, JA, Output, Args, SanitizeArgs, CmdArgs);
-
Args.AddLastArg(CmdArgs, options::OPT_fclang_abi_compat_EQ);
if (getLastProfileSampleUseArg(Args) &&
>From 3224283266e64d0913873c5d358326365d6020a4 Mon Sep 17 00:00:00 2001
From: "Sarnie, Nick" <nick.sar...@intel.com>
Date: Mon, 24 Mar 2025 08:29:20 -0700
Subject: [PATCH 4/4] flang fixes
Signed-off-by: Sarnie, Nick <nick.sar...@intel.com>
---
flang/lib/Frontend/CompilerInvocation.cpp | 5 ++---
flang/test/Lower/OpenMP/omp-is-gpu.f90 | 2 +-
2 files changed, 3 insertions(+), 4 deletions(-)
diff --git a/flang/lib/Frontend/CompilerInvocation.cpp
b/flang/lib/Frontend/CompilerInvocation.cpp
index 7d902fc92c226..376ae0c34c399 100644
--- a/flang/lib/Frontend/CompilerInvocation.cpp
+++ b/flang/lib/Frontend/CompilerInvocation.cpp
@@ -1144,9 +1144,8 @@ static bool parseOpenMPArgs(CompilerInvocation &res,
llvm::opt::ArgList &args,
diags.Report(diagID);
}
res.getLangOpts().OpenMPIsGPU = 1;
- else {
- res.getLangOpts().OpenMPIsGPU = 0;
- }
+ } else {
+ res.getLangOpts().OpenMPIsGPU = 0;
}
// Get the OpenMP target triples if any.
diff --git a/flang/test/Lower/OpenMP/omp-is-gpu.f90
b/flang/test/Lower/OpenMP/omp-is-gpu.f90
index 3e6daeb522d77..77d347c9a9fde 100644
--- a/flang/test/Lower/OpenMP/omp-is-gpu.f90
+++ b/flang/test/Lower/OpenMP/omp-is-gpu.f90
@@ -12,5 +12,5 @@
subroutine omp_subroutine()
end subroutine omp_subroutine
-!FLANG-ERROR: error: OpenMP AMDGPU/NVPTX is only prepared to deal with device
code.
+!FLANG-ERROR: error: OpenMP GPU is only prepared to deal with device code.
!BBC-ERROR: FATAL: -fopenmp-is-gpu can only be set if
-fopenmp-is-target-device is also set
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
