jdoerfert created this revision.
jdoerfert added reviewers: kiranchandramohan, ABataev, RaviNarayanaswamy,
gtbercea, grokos, sdmitriev, JonChesterfield, hfinkel, fghanim.
Herald added subscribers: s.egerton, guansong, bollu, simoncook, fedor.sergeev,
aheejin, rampitec, jholewinski.
Herald added a project: clang.
Instead of going through a custom overload resolution twice, we can use
the existing one.
TODO:
The tests need updating, they checked for functions that shouldn't
have been emitted and the way they check is hard to update.
There is a TODO in the code we need to fix (see below).
Repository:
rG LLVM Github Monorepo
https://reviews.llvm.org/D71241
Files:
clang/include/clang/AST/StmtOpenMP.h
clang/include/clang/Sema/Sema.h
clang/lib/AST/Expr.cpp
clang/lib/AST/StmtOpenMP.cpp
clang/lib/CodeGen/CGOpenMPRuntime.cpp
clang/lib/CodeGen/CGOpenMPRuntime.h
clang/lib/CodeGen/CGOpenMPRuntimeNVPTX.cpp
clang/lib/CodeGen/CGOpenMPRuntimeNVPTX.h
clang/lib/CodeGen/CodeGenModule.cpp
clang/lib/Sema/SemaExpr.cpp
clang/lib/Sema/SemaLookup.cpp
clang/lib/Sema/SemaOpenMP.cpp
clang/lib/Sema/SemaOverload.cpp
clang/lib/Sema/SemaTemplateInstantiateDecl.cpp
clang/test/OpenMP/declare_variant_ast_print.cpp
clang/test/OpenMP/declare_variant_device_kind_codegen.cpp
clang/test/OpenMP/declare_variant_implementation_vendor_codegen.cpp
clang/test/OpenMP/declare_variant_mixed_codegen.cpp
clang/test/OpenMP/nvptx_declare_variant_device_kind_codegen.cpp
clang/test/OpenMP/nvptx_declare_variant_implementation_vendor_codegen.cpp
Index: clang/test/OpenMP/nvptx_declare_variant_implementation_vendor_codegen.cpp
===================================================================
--- clang/test/OpenMP/nvptx_declare_variant_implementation_vendor_codegen.cpp
+++ clang/test/OpenMP/nvptx_declare_variant_implementation_vendor_codegen.cpp
@@ -8,31 +8,20 @@
// CHECK-DAG: define {{.*}}i32 @_Z3barv()
// CHECK-DAG: define {{.*}}i32 @_ZN16SpecSpecialFuncs6MethodEv(%struct.SpecSpecialFuncs* %{{.+}})
// CHECK-DAG: define {{.*}}i32 @_ZN12SpecialFuncs6MethodEv(%struct.SpecialFuncs* %{{.+}})
-// CHECK-DAG: define linkonce_odr {{.*}}i32 @_ZN16SpecSpecialFuncs6methodEv(%struct.SpecSpecialFuncs* %{{.+}})
-// CHECK-DAG: define linkonce_odr {{.*}}i32 @_ZN12SpecialFuncs6methodEv(%struct.SpecialFuncs* %{{.+}})
// CHECK-DAG: define {{.*}}i32 @_Z5prio_v()
-// CHECK-DAG: define internal i32 @_ZL6prio1_v()
// CHECK-DAG: define {{.*}}i32 @_Z4callv()
-// CHECK-DAG: define internal i32 @_ZL9stat_usedv()
// CHECK-DAG: define {{.*}}i32 @fn_linkage()
// CHECK-DAG: define {{.*}}i32 @_Z11fn_linkage1v()
// CHECK-DAG: ret i32 2
// CHECK-DAG: ret i32 3
-// CHECK-DAG: ret i32 4
-// CHECK-DAG: ret i32 5
// CHECK-DAG: ret i32 6
// CHECK-DAG: ret i32 7
-// CHECK-DAG: ret i32 82
// CHECK-DAG: ret i32 83
-// CHECK-DAG: ret i32 85
-// CHECK-DAG: ret i32 86
// CHECK-DAG: ret i32 87
// Outputs for function members
-// CHECK-DAG: ret i32 6
-// CHECK-DAG: ret i32 7
-// CHECK-NOT: ret i32 {{1|81|84}}
+// CHECK-NOT: ret i32 {{81|84}}
#ifndef HEADER
#define HEADER
Index: clang/test/OpenMP/nvptx_declare_variant_device_kind_codegen.cpp
===================================================================
--- clang/test/OpenMP/nvptx_declare_variant_device_kind_codegen.cpp
+++ clang/test/OpenMP/nvptx_declare_variant_device_kind_codegen.cpp
@@ -13,31 +13,20 @@
// CHECK-DAG: define {{.*}}i32 @_Z3barv()
// CHECK-DAG: define {{.*}}i32 @_ZN16SpecSpecialFuncs6MethodEv(%struct.SpecSpecialFuncs* %{{.+}})
// CHECK-DAG: define {{.*}}i32 @_ZN12SpecialFuncs6MethodEv(%struct.SpecialFuncs* %{{.+}})
-// CHECK-DAG: define linkonce_odr {{.*}}i32 @_ZN16SpecSpecialFuncs6methodEv(%struct.SpecSpecialFuncs* %{{.+}})
-// CHECK-DAG: define linkonce_odr {{.*}}i32 @_ZN12SpecialFuncs6methodEv(%struct.SpecialFuncs* %{{.+}})
// CHECK-DAG: define {{.*}}i32 @_Z5prio_v()
-// CHECK-DAG: define internal i32 @_ZL6prio1_v()
// CHECK-DAG: define {{.*}}i32 @_Z4callv()
-// CHECK-DAG: define internal i32 @_ZL9stat_usedv()
// CHECK-DAG: define {{.*}}i32 @fn_linkage()
// CHECK-DAG: define {{.*}}i32 @_Z11fn_linkage1v()
// CHECK-DAG: ret i32 2
// CHECK-DAG: ret i32 3
-// CHECK-DAG: ret i32 4
-// CHECK-DAG: ret i32 5
// CHECK-DAG: ret i32 6
// CHECK-DAG: ret i32 7
-// CHECK-DAG: ret i32 82
// CHECK-DAG: ret i32 83
-// CHECK-DAG: ret i32 85
-// CHECK-DAG: ret i32 86
// CHECK-DAG: ret i32 87
// Outputs for function members
-// CHECK-DAG: ret i32 6
-// CHECK-DAG: ret i32 7
-// CHECK-NOT: ret i32 {{1|81|84}}
+// CHECK-NOT: ret i32 {{81|84}}
#ifndef HEADER
#define HEADER
Index: clang/test/OpenMP/declare_variant_mixed_codegen.cpp
===================================================================
--- clang/test/OpenMP/declare_variant_mixed_codegen.cpp
+++ clang/test/OpenMP/declare_variant_mixed_codegen.cpp
@@ -4,19 +4,8 @@
// expected-no-diagnostics
// CHECK-NOT: ret i32 {{1|4|81|84}}
-// CHECK-DAG: @_Z3barv = {{.*}}alias i32 (), i32 ()* @_Z3foov
-// CHECK-DAG: @_ZN16SpecSpecialFuncs6MethodEv = {{.*}}alias i32 (%struct.SpecSpecialFuncs*), i32 (%struct.SpecSpecialFuncs*)* @_ZN16SpecSpecialFuncs7method_Ev
-// CHECK-DAG: @_ZN16SpecSpecialFuncs6methodEv = linkonce_odr {{.*}}alias i32 (%struct.SpecSpecialFuncs*), i32 (%struct.SpecSpecialFuncs*)* @_ZN16SpecSpecialFuncs7method_Ev
-// CHECK-DAG: @_ZN12SpecialFuncs6methodEv = linkonce_odr {{.*}}alias i32 (%struct.SpecialFuncs*), i32 (%struct.SpecialFuncs*)* @_ZN12SpecialFuncs7method_Ev
-// CHECK-DAG: @_Z5prio_v = {{.*}}alias i32 (), i32 ()* @_Z5prio1v
-// CHECK-DAG: @_ZL6prio1_v = internal alias i32 (), i32 ()* @_ZL5prio2v
-// CHECK-DAG: @_Z4callv = {{.*}}alias i32 (), i32 ()* @_Z4testv
-// CHECK-DAG: @_ZL9stat_usedv = internal alias i32 (), i32 ()* @_ZL10stat_used_v
-// CHECK-DAG: @_ZN12SpecialFuncs6MethodEv = {{.*}}alias i32 (%struct.SpecialFuncs*), i32 (%struct.SpecialFuncs*)* @_ZN12SpecialFuncs7method_Ev
-// CHECK-DAG: @fn_linkage = {{.*}}alias i32 (), i32 ()* @_Z18fn_linkage_variantv
-// CHECK-DAG: @_Z11fn_linkage1v = {{.*}}alias i32 (), i32 ()* @fn_linkage_variant1
// CHECK-DAG: declare {{.*}}i32 @_Z5bazzzv()
-// CHECK-DAG: declare {{.*}}i32 @_Z3bazv()
+// CHECK-DAG: define {{.*}}i32 @_Z3bazv()
// CHECK-DAG: ret i32 2
// CHECK-DAG: ret i32 3
// CHECK-DAG: ret i32 5
@@ -27,7 +16,7 @@
// CHECK-DAG: ret i32 85
// CHECK-DAG: ret i32 86
// CHECK-DAG: ret i32 87
-// CHECK-NOT: ret i32 {{1|4|81|84}}
+// CHECK-NOT: ret i32 {{4|81|84}}
#ifndef HEADER
#define HEADER
Index: clang/test/OpenMP/declare_variant_implementation_vendor_codegen.cpp
===================================================================
--- clang/test/OpenMP/declare_variant_implementation_vendor_codegen.cpp
+++ clang/test/OpenMP/declare_variant_implementation_vendor_codegen.cpp
@@ -4,19 +4,8 @@
// expected-no-diagnostics
// CHECK-NOT: ret i32 {{1|4|81|84}}
-// CHECK-DAG: @_Z3barv = {{.*}}alias i32 (), i32 ()* @_Z3foov
-// CHECK-DAG: @_ZN16SpecSpecialFuncs6MethodEv = {{.*}}alias i32 (%struct.SpecSpecialFuncs*), i32 (%struct.SpecSpecialFuncs*)* @_ZN16SpecSpecialFuncs7method_Ev
-// CHECK-DAG: @_ZN16SpecSpecialFuncs6methodEv = linkonce_odr {{.*}}alias i32 (%struct.SpecSpecialFuncs*), i32 (%struct.SpecSpecialFuncs*)* @_ZN16SpecSpecialFuncs7method_Ev
-// CHECK-DAG: @_ZN12SpecialFuncs6methodEv = linkonce_odr {{.*}}alias i32 (%struct.SpecialFuncs*), i32 (%struct.SpecialFuncs*)* @_ZN12SpecialFuncs7method_Ev
-// CHECK-DAG: @_Z5prio_v = {{.*}}alias i32 (), i32 ()* @_Z5prio1v
-// CHECK-DAG: @_ZL6prio1_v = internal alias i32 (), i32 ()* @_ZL5prio2v
-// CHECK-DAG: @_Z4callv = {{.*}}alias i32 (), i32 ()* @_Z4testv
-// CHECK-DAG: @_ZL9stat_usedv = internal alias i32 (), i32 ()* @_ZL10stat_used_v
-// CHECK-DAG: @_ZN12SpecialFuncs6MethodEv = {{.*}}alias i32 (%struct.SpecialFuncs*), i32 (%struct.SpecialFuncs*)* @_ZN12SpecialFuncs7method_Ev
-// CHECK-DAG: @fn_linkage = {{.*}}alias i32 (), i32 ()* @_Z18fn_linkage_variantv
-// CHECK-DAG: @_Z11fn_linkage1v = {{.*}}alias i32 (), i32 ()* @fn_linkage_variant1
// CHECK-DAG: declare {{.*}}i32 @_Z5bazzzv()
-// CHECK-DAG: declare {{.*}}i32 @_Z3bazv()
+// CHECK-DAG: define {{.*}}i32 @_Z3bazv()
// CHECK-DAG: ret i32 2
// CHECK-DAG: ret i32 3
// CHECK-DAG: ret i32 5
@@ -27,7 +16,7 @@
// CHECK-DAG: ret i32 85
// CHECK-DAG: ret i32 86
// CHECK-DAG: ret i32 87
-// CHECK-NOT: ret i32 {{1|4|81|84}}
+// CHECK-NOT: ret i32 {{4|81|84}}
#ifndef HEADER
#define HEADER
Index: clang/test/OpenMP/declare_variant_device_kind_codegen.cpp
===================================================================
--- clang/test/OpenMP/declare_variant_device_kind_codegen.cpp
+++ clang/test/OpenMP/declare_variant_device_kind_codegen.cpp
@@ -41,19 +41,8 @@
// expected-no-diagnostics
// CHECK-NOT: ret i32 {{1|4|81|84}}
-// CHECK-DAG: @_Z3barv = {{.*}}alias i32 (), i32 ()* @_Z3foov
-// CHECK-DAG: @_ZN16SpecSpecialFuncs6MethodEv = {{.*}}alias i32 (%struct.SpecSpecialFuncs*), i32 (%struct.SpecSpecialFuncs*)* @_ZN16SpecSpecialFuncs7method_Ev
-// CHECK-DAG: @_ZN16SpecSpecialFuncs6methodEv = linkonce_odr {{.*}}alias i32 (%struct.SpecSpecialFuncs*), i32 (%struct.SpecSpecialFuncs*)* @_ZN16SpecSpecialFuncs7method_Ev
-// CHECK-DAG: @_ZN12SpecialFuncs6methodEv = linkonce_odr {{.*}}alias i32 (%struct.SpecialFuncs*), i32 (%struct.SpecialFuncs*)* @_ZN12SpecialFuncs7method_Ev
-// CHECK-DAG: @_Z5prio_v = {{.*}}alias i32 (), i32 ()* @_Z5prio1v
-// CHECK-DAG: @_ZL6prio1_v = internal alias i32 (), i32 ()* @_ZL5prio2v
-// CHECK-DAG: @_Z4callv = {{.*}}alias i32 (), i32 ()* @_Z4testv
-// CHECK-DAG: @_ZL9stat_usedv = internal alias i32 (), i32 ()* @_ZL10stat_used_v
-// CHECK-DAG: @_ZN12SpecialFuncs6MethodEv = {{.*}}alias i32 (%struct.SpecialFuncs*), i32 (%struct.SpecialFuncs*)* @_ZN12SpecialFuncs7method_Ev
-// CHECK-DAG: @fn_linkage = {{.*}}alias i32 (), i32 ()* @_Z18fn_linkage_variantv
-// CHECK-DAG: @_Z11fn_linkage1v = {{.*}}alias i32 (), i32 ()* @fn_linkage_variant1
// CHECK-DAG: declare {{.*}}i32 @_Z5bazzzv()
-// CHECK-DAG: declare {{.*}}i32 @_Z3bazv()
+// CHECK-DAG: define {{.*}}i32 @_Z3bazv()
// CHECK-DAG: ret i32 2
// CHECK-DAG: ret i32 3
// CHECK-DAG: ret i32 5
@@ -64,7 +53,7 @@
// CHECK-DAG: ret i32 85
// CHECK-DAG: ret i32 86
// CHECK-DAG: ret i32 87
-// CHECK-NOT: ret i32 {{1|4|81|84}}
+// CHECK-NOT: ret i32 {{4|81|84}}
#ifndef HEADER
#define HEADER
Index: clang/test/OpenMP/declare_variant_ast_print.cpp
===================================================================
--- clang/test/OpenMP/declare_variant_ast_print.cpp
+++ clang/test/OpenMP/declare_variant_ast_print.cpp
@@ -40,6 +40,7 @@
#pragma omp declare variant(foofoo <T>) match(user = {condition(<expr>)})
#pragma omp declare variant(foofoo <T>) match(implementation={vendor(llvm)},device={kind(cpu)})
#pragma omp declare variant(foofoo <T>) match(implementation={vendor(unknown)})
+// TODO: Handle template instantiation
#pragma omp declare variant(foofoo <T>) match(implementation={vendor(score(C+5): ibm, xxx, ibm)},device={kind(cpu,host)})
template <typename T, int C>
T barbar();
@@ -50,7 +51,7 @@
// CHECK-NEXT: template<> int barbar<int, 3>();
// CHECK-NEXT: int baz() {
-// CHECK-NEXT: return barbar<int, 3>();
+// CHECK-NEXT: return foofoo<int, 3>();
// CHECK-NEXT: }
int baz() {
return barbar<int, 3>();
Index: clang/lib/Sema/SemaTemplateInstantiateDecl.cpp
===================================================================
--- clang/lib/Sema/SemaTemplateInstantiateDecl.cpp
+++ clang/lib/Sema/SemaTemplateInstantiateDecl.cpp
@@ -348,92 +348,6 @@
Attr.getRange());
}
-/// Instantiation of 'declare variant' attribute and its arguments.
-static void instantiateOMPDeclareVariantAttr(
- Sema &S, const MultiLevelTemplateArgumentList &TemplateArgs,
- const OMPDeclareVariantAttr &Attr, Decl *New) {
- // Allow 'this' in clauses with varlists.
- if (auto *FTD = dyn_cast<FunctionTemplateDecl>(New))
- New = FTD->getTemplatedDecl();
- auto *FD = cast<FunctionDecl>(New);
- auto *ThisContext = dyn_cast_or_null<CXXRecordDecl>(FD->getDeclContext());
-
- auto &&SubstExpr = [FD, ThisContext, &S, &TemplateArgs](Expr *E) {
- if (auto *DRE = dyn_cast<DeclRefExpr>(E->IgnoreParenImpCasts()))
- if (auto *PVD = dyn_cast<ParmVarDecl>(DRE->getDecl())) {
- Sema::ContextRAII SavedContext(S, FD);
- LocalInstantiationScope Local(S);
- if (FD->getNumParams() > PVD->getFunctionScopeIndex())
- Local.InstantiatedLocal(
- PVD, FD->getParamDecl(PVD->getFunctionScopeIndex()));
- return S.SubstExpr(E, TemplateArgs);
- }
- Sema::CXXThisScopeRAII ThisScope(S, ThisContext, Qualifiers(),
- FD->isCXXInstanceMember());
- return S.SubstExpr(E, TemplateArgs);
- };
-
- // Substitute a single OpenMP clause, which is a potentially-evaluated
- // full-expression.
- auto &&Subst = [&SubstExpr, &S](Expr *E) {
- EnterExpressionEvaluationContext Evaluated(
- S, Sema::ExpressionEvaluationContext::PotentiallyEvaluated);
- ExprResult Res = SubstExpr(E);
- if (Res.isInvalid())
- return Res;
- return S.ActOnFinishFullExpr(Res.get(), false);
- };
-
- ExprResult VariantFuncRef;
- if (Expr *E = Attr.getVariantFuncRef())
- VariantFuncRef = Subst(E);
-
- // Check function/variant ref.
- Optional<std::pair<FunctionDecl *, Expr *>> DeclVarData =
- S.checkOpenMPDeclareVariantFunction(
- S.ConvertDeclToDeclGroup(New), VariantFuncRef.get(), Attr.getRange());
- if (!DeclVarData)
- return;
- SmallVector<Sema::OMPCtxSelectorData, 4> Data;
- for (unsigned I = 0, E = Attr.scores_size(); I < E; ++I) {
- ExprResult Score;
- if (Expr *E = *std::next(Attr.scores_begin(), I))
- Score = Subst(E);
- // Instantiate the attribute.
- auto CtxSet = static_cast<OpenMPContextSelectorSetKind>(
- *std::next(Attr.ctxSelectorSets_begin(), I));
- auto Ctx = static_cast<OpenMPContextSelectorKind>(
- *std::next(Attr.ctxSelectors_begin(), I));
- switch (CtxSet) {
- case OMP_CTX_SET_implementation:
- switch (Ctx) {
- case OMP_CTX_vendor:
- Data.emplace_back(CtxSet, Ctx, Score, Attr.implVendors());
- break;
- case OMP_CTX_kind:
- case OMP_CTX_unknown:
- llvm_unreachable("Unexpected context selector kind.");
- }
- break;
- case OMP_CTX_SET_device:
- switch (Ctx) {
- case OMP_CTX_kind:
- Data.emplace_back(CtxSet, Ctx, Score, Attr.deviceKinds());
- break;
- case OMP_CTX_vendor:
- case OMP_CTX_unknown:
- llvm_unreachable("Unexpected context selector kind.");
- }
- break;
- case OMP_CTX_SET_unknown:
- llvm_unreachable("Unexpected context selector set kind.");
- }
- }
- S.ActOnOpenMPDeclareVariantDirective(DeclVarData.getValue().first,
- DeclVarData.getValue().second,
- Attr.getRange(), Data);
-}
-
static void instantiateDependentAMDGPUFlatWorkGroupSizeAttr(
Sema &S, const MultiLevelTemplateArgumentList &TemplateArgs,
const AMDGPUFlatWorkGroupSizeAttr &Attr, Decl *New) {
@@ -591,11 +505,6 @@
continue;
}
- if (const auto *OMPAttr = dyn_cast<OMPDeclareVariantAttr>(TmplAttr)) {
- instantiateOMPDeclareVariantAttr(*this, TemplateArgs, *OMPAttr, New);
- continue;
- }
-
if (const auto *AMDGPUFlatWorkGroupSize =
dyn_cast<AMDGPUFlatWorkGroupSizeAttr>(TmplAttr)) {
instantiateDependentAMDGPUFlatWorkGroupSizeAttr(
Index: clang/lib/Sema/SemaOverload.cpp
===================================================================
--- clang/lib/Sema/SemaOverload.cpp
+++ clang/lib/Sema/SemaOverload.cpp
@@ -11,12 +11,14 @@
//===----------------------------------------------------------------------===//
#include "clang/Sema/Overload.h"
+
#include "clang/AST/ASTContext.h"
#include "clang/AST/CXXInheritance.h"
#include "clang/AST/DeclObjC.h"
#include "clang/AST/Expr.h"
#include "clang/AST/ExprCXX.h"
#include "clang/AST/ExprObjC.h"
+#include "clang/AST/StmtOpenMP.h"
#include "clang/AST/TypeOrdering.h"
#include "clang/Basic/Diagnostic.h"
#include "clang/Basic/DiagnosticOptions.h"
@@ -9708,6 +9710,22 @@
S.diagnoseEquivalentInternalLinkageDeclarations(Loc, Best->Function,
EquivalentCands);
+ FunctionDecl *FD = Best->Function;
+ if (!FD || !FD->hasAttrs() || !FD->hasAttr<OMPDeclareVariantAttr>())
+ return OR_Success;
+
+ // Iterate through all DeclareVariant attributes and check context selectors.
+ const OMPDeclareVariantAttr *BestVariant = nullptr;
+ for (const auto *A : FD->specific_attrs<OMPDeclareVariantAttr>())
+ BestVariant =
+ getBetterOpenMPContextMatch(S.getASTContext(), BestVariant, A);
+ if (!BestVariant || !BestVariant->getVariantFuncRef())
+ return OR_Success;
+
+ // TODO: Handle template instantiation
+ Best->Function = cast<FunctionDecl>(
+ cast<DeclRefExpr>(BestVariant->getVariantFuncRef()->IgnoreParenImpCasts())
+ ->getDecl());
return OR_Success;
}
Index: clang/lib/Sema/SemaOpenMP.cpp
===================================================================
--- clang/lib/Sema/SemaOpenMP.cpp
+++ clang/lib/Sema/SemaOpenMP.cpp
@@ -5460,25 +5460,6 @@
}
}
-void Sema::markOpenMPDeclareVariantFuncsReferenced(SourceLocation Loc,
- FunctionDecl *Func,
- bool MightBeOdrUse) {
- assert(LangOpts.OpenMP && "Expected OpenMP mode.");
-
- if (!Func->isDependentContext() && Func->hasAttrs()) {
- for (OMPDeclareVariantAttr *A :
- Func->specific_attrs<OMPDeclareVariantAttr>()) {
- // TODO: add checks for active OpenMP context where possible.
- Expr *VariantRef = A->getVariantFuncRef();
- auto *DRE = cast<DeclRefExpr>(VariantRef->IgnoreParenImpCasts());
- auto *F = cast<FunctionDecl>(DRE->getDecl());
- if (!F->isDefined() && F->isTemplateInstantiation())
- InstantiateFunctionDefinition(Loc, F->getFirstDecl());
- MarkFunctionReferenced(Loc, F, MightBeOdrUse);
- }
- }
-}
-
StmtResult Sema::ActOnOpenMPParallelDirective(ArrayRef<OMPClause *> Clauses,
Stmt *AStmt,
SourceLocation StartLoc,
Index: clang/lib/Sema/SemaLookup.cpp
===================================================================
--- clang/lib/Sema/SemaLookup.cpp
+++ clang/lib/Sema/SemaLookup.cpp
@@ -328,6 +328,8 @@
assert(ResultKind != NotFound || Decls.size() == 0);
assert(ResultKind != Found || Decls.size() == 1);
assert(ResultKind != FoundOverloaded || Decls.size() > 1 ||
+ (Decls.size() == 1 &&
+ (*begin())->getUnderlyingDecl()->getAttr<OMPDeclareVariantAttr>()) ||
(Decls.size() == 1 &&
isa<FunctionTemplateDecl>((*begin())->getUnderlyingDecl())));
assert(ResultKind != FoundUnresolvedValue || sanityCheckUnresolved());
@@ -490,6 +492,9 @@
ResultKind = FoundOverloaded;
else if (isa<UnresolvedUsingValueDecl>(D))
ResultKind = FoundUnresolvedValue;
+ else if (auto *OMPVariantAttr = D->getAttr<OMPDeclareVariantAttr>())
+ if (OMPVariantAttr->getVariantFuncRef())
+ ResultKind = FoundOverloaded;
return;
}
Index: clang/lib/Sema/SemaExpr.cpp
===================================================================
--- clang/lib/Sema/SemaExpr.cpp
+++ clang/lib/Sema/SemaExpr.cpp
@@ -15642,7 +15642,6 @@
}
if (LangOpts.OpenMP) {
- markOpenMPDeclareVariantFuncsReferenced(Loc, Func, MightBeOdrUse);
if (LangOpts.OpenMPIsDevice)
checkOpenMPDeviceFunction(Loc, Func);
else
Index: clang/lib/CodeGen/CodeGenModule.cpp
===================================================================
--- clang/lib/CodeGen/CodeGenModule.cpp
+++ clang/lib/CodeGen/CodeGenModule.cpp
@@ -2534,11 +2534,6 @@
return;
}
- // Check if this must be emitted as declare variant.
- if (LangOpts.OpenMP && isa<FunctionDecl>(Global) && OpenMPRuntime &&
- OpenMPRuntime->emitDeclareVariant(GD, /*IsForDefinition=*/false))
- return;
-
// If we're deferring emission of a C++ variable with an
// initializer, remember the order in which it appeared in the file.
if (getLangOpts().CPlusPlus && isa<VarDecl>(Global) &&
@@ -3102,10 +3097,6 @@
EmitGlobal(GDDef);
}
}
- // Check if this must be emitted as declare variant and emit reference to
- // the the declare variant function.
- if (LangOpts.OpenMP && OpenMPRuntime)
- (void)OpenMPRuntime->emitDeclareVariant(GD, /*IsForDefinition=*/true);
if (FD->isMultiVersion()) {
const auto *TA = FD->getAttr<TargetAttr>();
@@ -4384,11 +4375,6 @@
void CodeGenModule::EmitGlobalFunctionDefinition(GlobalDecl GD,
llvm::GlobalValue *GV) {
- // Check if this must be emitted as declare variant.
- if (LangOpts.OpenMP && OpenMPRuntime &&
- OpenMPRuntime->emitDeclareVariant(GD, /*IsForDefinition=*/true))
- return;
-
const auto *D = cast<FunctionDecl>(GD.getDecl());
// Compute the function info and LLVM type.
Index: clang/lib/CodeGen/CGOpenMPRuntimeNVPTX.h
===================================================================
--- clang/lib/CodeGen/CGOpenMPRuntimeNVPTX.h
+++ clang/lib/CodeGen/CGOpenMPRuntimeNVPTX.h
@@ -193,18 +193,6 @@
/// Full/Lightweight runtime mode. Used for better optimization.
unsigned getDefaultLocationReserved2Flags() const override;
- /// Tries to emit declare variant function for \p OldGD from \p NewGD.
- /// \param OrigAddr LLVM IR value for \p OldGD.
- /// \param IsForDefinition true, if requested emission for the definition of
- /// \p OldGD.
- /// \returns true, was able to emit a definition function for \p OldGD, which
- /// points to \p NewGD.
- /// NVPTX backend does not support global aliases, so just use the function,
- /// emitted for \p NewGD instead of \p OldGD.
- bool tryEmitDeclareVariant(const GlobalDecl &NewGD, const GlobalDecl &OldGD,
- llvm::GlobalValue *OrigAddr,
- bool IsForDefinition) override;
-
public:
explicit CGOpenMPRuntimeNVPTX(CodeGenModule &CGM);
void clear() override;
Index: clang/lib/CodeGen/CGOpenMPRuntimeNVPTX.cpp
===================================================================
--- clang/lib/CodeGen/CGOpenMPRuntimeNVPTX.cpp
+++ clang/lib/CodeGen/CGOpenMPRuntimeNVPTX.cpp
@@ -1914,19 +1914,6 @@
llvm_unreachable("Unknown flags are requested.");
}
-bool CGOpenMPRuntimeNVPTX::tryEmitDeclareVariant(const GlobalDecl &NewGD,
- const GlobalDecl &OldGD,
- llvm::GlobalValue *OrigAddr,
- bool IsForDefinition) {
- // Emit the function in OldGD with the body from NewGD, if NewGD is defined.
- auto *NewFD = cast<FunctionDecl>(NewGD.getDecl());
- if (NewFD->isDefined()) {
- CGM.emitOpenMPDeviceFunctionRedefinition(OldGD, NewGD, OrigAddr);
- return true;
- }
- return false;
-}
-
CGOpenMPRuntimeNVPTX::CGOpenMPRuntimeNVPTX(CodeGenModule &CGM)
: CGOpenMPRuntime(CGM, "_", "$") {
if (!CGM.getLangOpts().OpenMPIsDevice)
Index: clang/lib/CodeGen/CGOpenMPRuntime.h
===================================================================
--- clang/lib/CodeGen/CGOpenMPRuntime.h
+++ clang/lib/CodeGen/CGOpenMPRuntime.h
@@ -280,17 +280,6 @@
/// default location.
virtual unsigned getDefaultLocationReserved2Flags() const { return 0; }
- /// Tries to emit declare variant function for \p OldGD from \p NewGD.
- /// \param OrigAddr LLVM IR value for \p OldGD.
- /// \param IsForDefinition true, if requested emission for the definition of
- /// \p OldGD.
- /// \returns true, was able to emit a definition function for \p OldGD, which
- /// points to \p NewGD.
- virtual bool tryEmitDeclareVariant(const GlobalDecl &NewGD,
- const GlobalDecl &OldGD,
- llvm::GlobalValue *OrigAddr,
- bool IsForDefinition);
-
/// Returns default flags for the barriers depending on the directive, for
/// which this barier is going to be emitted.
static unsigned getDefaultFlagsForBarriers(OpenMPDirectiveKind Kind);
@@ -1660,9 +1649,6 @@
/// Return whether the unified_shared_memory has been specified.
bool hasRequiresUnifiedSharedMemory() const;
-
- /// Emits the definition of the declare variant function.
- virtual bool emitDeclareVariant(GlobalDecl GD, bool IsForDefinition);
};
/// Class supports emissionof SIMD-only code.
Index: clang/lib/CodeGen/CGOpenMPRuntime.cpp
===================================================================
--- clang/lib/CodeGen/CGOpenMPRuntime.cpp
+++ clang/lib/CodeGen/CGOpenMPRuntime.cpp
@@ -1267,52 +1267,6 @@
loadOffloadInfoMetadata();
}
-bool CGOpenMPRuntime::tryEmitDeclareVariant(const GlobalDecl &NewGD,
- const GlobalDecl &OldGD,
- llvm::GlobalValue *OrigAddr,
- bool IsForDefinition) {
- // Emit at least a definition for the aliasee if the the address of the
- // original function is requested.
- if (IsForDefinition || OrigAddr)
- (void)CGM.GetAddrOfGlobal(NewGD);
- StringRef NewMangledName = CGM.getMangledName(NewGD);
- llvm::GlobalValue *Addr = CGM.GetGlobalValue(NewMangledName);
- if (Addr && !Addr->isDeclaration()) {
- const auto *D = cast<FunctionDecl>(OldGD.getDecl());
- const CGFunctionInfo &FI = CGM.getTypes().arrangeGlobalDeclaration(OldGD);
- llvm::Type *DeclTy = CGM.getTypes().GetFunctionType(FI);
-
- // Create a reference to the named value. This ensures that it is emitted
- // if a deferred decl.
- llvm::GlobalValue::LinkageTypes LT = CGM.getFunctionLinkage(OldGD);
-
- // Create the new alias itself, but don't set a name yet.
- auto *GA =
- llvm::GlobalAlias::create(DeclTy, 0, LT, "", Addr, &CGM.getModule());
-
- if (OrigAddr) {
- assert(OrigAddr->isDeclaration() && "Expected declaration");
-
- GA->takeName(OrigAddr);
- OrigAddr->replaceAllUsesWith(
- llvm::ConstantExpr::getBitCast(GA, OrigAddr->getType()));
- OrigAddr->eraseFromParent();
- } else {
- GA->setName(CGM.getMangledName(OldGD));
- }
-
- // Set attributes which are particular to an alias; this is a
- // specialization of the attributes which may be set on a global function.
- if (D->hasAttr<WeakAttr>() || D->hasAttr<WeakRefAttr>() ||
- D->isWeakImported())
- GA->setLinkage(llvm::Function::WeakAnyLinkage);
-
- CGM.SetCommonAttributes(OldGD, GA);
- return true;
- }
- return false;
-}
-
void CGOpenMPRuntime::clear() {
InternalVars.clear();
// Clean non-target variable declarations possibly used only in debug info.
@@ -1326,14 +1280,6 @@
continue;
GV->eraseFromParent();
}
- // Emit aliases for the deferred aliasees.
- for (const auto &Pair : DeferredVariantFunction) {
- StringRef MangledName = CGM.getMangledName(Pair.second.second);
- llvm::GlobalValue *Addr = CGM.GetGlobalValue(MangledName);
- // If not able to emit alias, just emit original declaration.
- (void)tryEmitDeclareVariant(Pair.second.first, Pair.second.second, Addr,
- /*IsForDefinition=*/false);
- }
}
std::string CGOpenMPRuntime::getName(ArrayRef<StringRef> Parts) const {
@@ -11044,280 +10990,6 @@
return Address(Addr, Align);
}
-namespace {
-using OMPContextSelectorData =
- OpenMPCtxSelectorData<ArrayRef<StringRef>, llvm::APSInt>;
-using CompleteOMPContextSelectorData = SmallVector<OMPContextSelectorData, 4>;
-} // anonymous namespace
-
-/// Checks current context and returns true if it matches the context selector.
-template <OpenMPContextSelectorSetKind CtxSet, OpenMPContextSelectorKind Ctx,
- typename... Arguments>
-static bool checkContext(const OMPContextSelectorData &Data,
- Arguments... Params) {
- assert(Data.CtxSet != OMP_CTX_SET_unknown && Data.Ctx != OMP_CTX_unknown &&
- "Unknown context selector or context selector set.");
- return false;
-}
-
-/// Checks for implementation={vendor(<vendor>)} context selector.
-/// \returns true iff <vendor>="llvm", false otherwise.
-template <>
-bool checkContext<OMP_CTX_SET_implementation, OMP_CTX_vendor>(
- const OMPContextSelectorData &Data) {
- return llvm::all_of(Data.Names,
- [](StringRef S) { return !S.compare_lower("llvm"); });
-}
-
-/// Checks for device={kind(<kind>)} context selector.
-/// \returns true if <kind>="host" and compilation is for host.
-/// true if <kind>="nohost" and compilation is for device.
-/// true if <kind>="cpu" and compilation is for Arm, X86 or PPC CPU.
-/// true if <kind>="gpu" and compilation is for NVPTX or AMDGCN.
-/// false otherwise.
-template <>
-bool checkContext<OMP_CTX_SET_device, OMP_CTX_kind, CodeGenModule &>(
- const OMPContextSelectorData &Data, CodeGenModule &CGM) {
- for (StringRef Name : Data.Names) {
- if (!Name.compare_lower("host")) {
- if (CGM.getLangOpts().OpenMPIsDevice)
- return false;
- continue;
- }
- if (!Name.compare_lower("nohost")) {
- if (!CGM.getLangOpts().OpenMPIsDevice)
- return false;
- continue;
- }
- switch (CGM.getTriple().getArch()) {
- case llvm::Triple::arm:
- case llvm::Triple::armeb:
- case llvm::Triple::aarch64:
- case llvm::Triple::aarch64_be:
- case llvm::Triple::aarch64_32:
- case llvm::Triple::ppc:
- case llvm::Triple::ppc64:
- case llvm::Triple::ppc64le:
- case llvm::Triple::x86:
- case llvm::Triple::x86_64:
- if (Name.compare_lower("cpu"))
- return false;
- break;
- case llvm::Triple::amdgcn:
- case llvm::Triple::nvptx:
- case llvm::Triple::nvptx64:
- if (Name.compare_lower("gpu"))
- return false;
- break;
- case llvm::Triple::UnknownArch:
- case llvm::Triple::arc:
- case llvm::Triple::avr:
- case llvm::Triple::bpfel:
- case llvm::Triple::bpfeb:
- case llvm::Triple::hexagon:
- case llvm::Triple::mips:
- case llvm::Triple::mipsel:
- case llvm::Triple::mips64:
- case llvm::Triple::mips64el:
- case llvm::Triple::msp430:
- case llvm::Triple::r600:
- case llvm::Triple::riscv32:
- case llvm::Triple::riscv64:
- case llvm::Triple::sparc:
- case llvm::Triple::sparcv9:
- case llvm::Triple::sparcel:
- case llvm::Triple::systemz:
- case llvm::Triple::tce:
- case llvm::Triple::tcele:
- case llvm::Triple::thumb:
- case llvm::Triple::thumbeb:
- case llvm::Triple::xcore:
- case llvm::Triple::le32:
- case llvm::Triple::le64:
- case llvm::Triple::amdil:
- case llvm::Triple::amdil64:
- case llvm::Triple::hsail:
- case llvm::Triple::hsail64:
- case llvm::Triple::spir:
- case llvm::Triple::spir64:
- case llvm::Triple::kalimba:
- case llvm::Triple::shave:
- case llvm::Triple::lanai:
- case llvm::Triple::wasm32:
- case llvm::Triple::wasm64:
- case llvm::Triple::renderscript32:
- case llvm::Triple::renderscript64:
- return false;
- }
- }
- return true;
-}
-
-bool matchesContext(CodeGenModule &CGM,
- const CompleteOMPContextSelectorData &ContextData) {
- for (const OMPContextSelectorData &Data : ContextData) {
- switch (Data.Ctx) {
- case OMP_CTX_vendor:
- assert(Data.CtxSet == OMP_CTX_SET_implementation &&
- "Expected implementation context selector set.");
- if (!checkContext<OMP_CTX_SET_implementation, OMP_CTX_vendor>(Data))
- return false;
- break;
- case OMP_CTX_kind:
- assert(Data.CtxSet == OMP_CTX_SET_device &&
- "Expected device context selector set.");
- if (!checkContext<OMP_CTX_SET_device, OMP_CTX_kind, CodeGenModule &>(Data,
- CGM))
- return false;
- break;
- case OMP_CTX_unknown:
- llvm_unreachable("Unknown context selector kind.");
- }
- }
- return true;
-}
-
-static CompleteOMPContextSelectorData
-translateAttrToContextSelectorData(ASTContext &C,
- const OMPDeclareVariantAttr *A) {
- CompleteOMPContextSelectorData Data;
- for (unsigned I = 0, E = A->scores_size(); I < E; ++I) {
- Data.emplace_back();
- auto CtxSet = static_cast<OpenMPContextSelectorSetKind>(
- *std::next(A->ctxSelectorSets_begin(), I));
- auto Ctx = static_cast<OpenMPContextSelectorKind>(
- *std::next(A->ctxSelectors_begin(), I));
- Data.back().CtxSet = CtxSet;
- Data.back().Ctx = Ctx;
- const Expr *Score = *std::next(A->scores_begin(), I);
- Data.back().Score = Score->EvaluateKnownConstInt(C);
- switch (Ctx) {
- case OMP_CTX_vendor:
- assert(CtxSet == OMP_CTX_SET_implementation &&
- "Expected implementation context selector set.");
- Data.back().Names =
- llvm::makeArrayRef(A->implVendors_begin(), A->implVendors_end());
- break;
- case OMP_CTX_kind:
- assert(CtxSet == OMP_CTX_SET_device &&
- "Expected device context selector set.");
- Data.back().Names =
- llvm::makeArrayRef(A->deviceKinds_begin(), A->deviceKinds_end());
- break;
- case OMP_CTX_unknown:
- llvm_unreachable("Unknown context selector kind.");
- }
- }
- return Data;
-}
-
-static bool isStrictSubset(const CompleteOMPContextSelectorData &LHS,
- const CompleteOMPContextSelectorData &RHS) {
- llvm::SmallDenseMap<std::pair<int, int>, llvm::StringSet<>, 4> RHSData;
- for (const OMPContextSelectorData &D : RHS) {
- auto &Pair = RHSData.FindAndConstruct(std::make_pair(D.CtxSet, D.Ctx));
- Pair.getSecond().insert(D.Names.begin(), D.Names.end());
- }
- bool AllSetsAreEqual = true;
- for (const OMPContextSelectorData &D : LHS) {
- auto It = RHSData.find(std::make_pair(D.CtxSet, D.Ctx));
- if (It == RHSData.end())
- return false;
- if (D.Names.size() > It->getSecond().size())
- return false;
- if (llvm::set_union(It->getSecond(), D.Names))
- return false;
- AllSetsAreEqual =
- AllSetsAreEqual && (D.Names.size() == It->getSecond().size());
- }
-
- return LHS.size() != RHS.size() || !AllSetsAreEqual;
-}
-
-static bool greaterCtxScore(const CompleteOMPContextSelectorData &LHS,
- const CompleteOMPContextSelectorData &RHS) {
- // Score is calculated as sum of all scores + 1.
- llvm::APSInt LHSScore(llvm::APInt(64, 1), /*isUnsigned=*/false);
- bool RHSIsSubsetOfLHS = isStrictSubset(RHS, LHS);
- if (RHSIsSubsetOfLHS) {
- LHSScore = llvm::APSInt::get(0);
- } else {
- for (const OMPContextSelectorData &Data : LHS) {
- if (Data.Score.getBitWidth() > LHSScore.getBitWidth()) {
- LHSScore = LHSScore.extend(Data.Score.getBitWidth()) + Data.Score;
- } else if (Data.Score.getBitWidth() < LHSScore.getBitWidth()) {
- LHSScore += Data.Score.extend(LHSScore.getBitWidth());
- } else {
- LHSScore += Data.Score;
- }
- }
- }
- llvm::APSInt RHSScore(llvm::APInt(64, 1), /*isUnsigned=*/false);
- if (!RHSIsSubsetOfLHS && isStrictSubset(LHS, RHS)) {
- RHSScore = llvm::APSInt::get(0);
- } else {
- for (const OMPContextSelectorData &Data : RHS) {
- if (Data.Score.getBitWidth() > RHSScore.getBitWidth()) {
- RHSScore = RHSScore.extend(Data.Score.getBitWidth()) + Data.Score;
- } else if (Data.Score.getBitWidth() < RHSScore.getBitWidth()) {
- RHSScore += Data.Score.extend(RHSScore.getBitWidth());
- } else {
- RHSScore += Data.Score;
- }
- }
- }
- return llvm::APSInt::compareValues(LHSScore, RHSScore) >= 0;
-}
-
-/// Finds the variant function that matches current context with its context
-/// selector.
-static const FunctionDecl *getDeclareVariantFunction(CodeGenModule &CGM,
- const FunctionDecl *FD) {
- if (!FD->hasAttrs() || !FD->hasAttr<OMPDeclareVariantAttr>())
- return FD;
- // Iterate through all DeclareVariant attributes and check context selectors.
- const OMPDeclareVariantAttr *TopMostAttr = nullptr;
- CompleteOMPContextSelectorData TopMostData;
- for (const auto *A : FD->specific_attrs<OMPDeclareVariantAttr>()) {
- CompleteOMPContextSelectorData Data =
- translateAttrToContextSelectorData(CGM.getContext(), A);
- if (!matchesContext(CGM, Data))
- continue;
- // If the attribute matches the context, find the attribute with the highest
- // score.
- if (!TopMostAttr || !greaterCtxScore(TopMostData, Data)) {
- TopMostAttr = A;
- TopMostData.swap(Data);
- }
- }
- if (!TopMostAttr)
- return FD;
- return cast<FunctionDecl>(
- cast<DeclRefExpr>(TopMostAttr->getVariantFuncRef()->IgnoreParenImpCasts())
- ->getDecl());
-}
-
-bool CGOpenMPRuntime::emitDeclareVariant(GlobalDecl GD, bool IsForDefinition) {
- const auto *D = cast<FunctionDecl>(GD.getDecl());
- // If the original function is defined already, use its definition.
- StringRef MangledName = CGM.getMangledName(GD);
- llvm::GlobalValue *Orig = CGM.GetGlobalValue(MangledName);
- if (Orig && !Orig->isDeclaration())
- return false;
- const FunctionDecl *NewFD = getDeclareVariantFunction(CGM, D);
- // Emit original function if it does not have declare variant attribute or the
- // context does not match.
- if (NewFD == D)
- return false;
- GlobalDecl NewGD = GD.getWithDecl(NewFD);
- if (tryEmitDeclareVariant(NewGD, GD, Orig, IsForDefinition)) {
- DeferredVariantFunction.erase(D);
- return true;
- }
- DeferredVariantFunction.insert(std::make_pair(D, std::make_pair(NewGD, GD)));
- return true;
-}
-
llvm::Function *CGOpenMPSIMDRuntime::emitParallelOutlinedFunction(
const OMPExecutableDirective &D, const VarDecl *ThreadIDVar,
OpenMPDirectiveKind InnermostKind, const RegionCodeGenTy &CodeGen) {
Index: clang/lib/AST/StmtOpenMP.cpp
===================================================================
--- clang/lib/AST/StmtOpenMP.cpp
+++ clang/lib/AST/StmtOpenMP.cpp
@@ -13,6 +13,8 @@
#include "clang/AST/StmtOpenMP.h"
#include "clang/AST/ASTContext.h"
+#include "clang/AST/Attr.h"
+#include "llvm/ADT/SetOperations.h"
using namespace clang;
using namespace llvm::omp;
@@ -2240,3 +2242,264 @@
return new (Mem)
OMPTargetTeamsDistributeSimdDirective(CollapsedNum, NumClauses);
}
+
+// TODO: We have various representations for the same data, it might help to
+// reuse some instead of converting them.
+// TODO: It is unclear where this checking code should live. It is used all over
+// the place and would probably fit bet in OMPDeclareVariantAttr.
+using OMPContextSelectorData =
+ OpenMPCtxSelectorData<ArrayRef<StringRef>, llvm::APSInt>;
+using CompleteOMPContextSelectorData = SmallVector<OMPContextSelectorData, 4>;
+
+/// Checks current context and returns true if it matches the context selector.
+template <OpenMPContextSelectorSetKind CtxSet, OpenMPContextSelectorKind Ctx,
+ typename... Arguments>
+static bool checkContext(const OMPContextSelectorData &Data,
+ Arguments... Params) {
+ assert(Data.CtxSet != OMP_CTX_SET_unknown && Data.Ctx != OMP_CTX_unknown &&
+ "Unknown context selector or context selector set.");
+ return false;
+}
+
+/// Checks for implementation={vendor(<vendor>)} context selector.
+/// \returns true iff <vendor>="llvm", false otherwise.
+template <>
+bool checkContext<OMP_CTX_SET_implementation, OMP_CTX_vendor>(
+ const OMPContextSelectorData &Data) {
+ return llvm::all_of(Data.Names,
+ [](StringRef S) { return !S.compare_lower("llvm"); });
+}
+
+/// Checks for device={kind(<kind>)} context selector.
+/// \returns true if <kind>="host" and compilation is for host.
+/// true if <kind>="nohost" and compilation is for device.
+/// true if <kind>="cpu" and compilation is for Arm, X86 or PPC CPU.
+/// true if <kind>="gpu" and compilation is for NVPTX or AMDGCN.
+/// false otherwise.
+template <>
+bool checkContext<OMP_CTX_SET_device, OMP_CTX_kind, const LangOptions &,
+ const TargetInfo &>(const OMPContextSelectorData &Data,
+ const LangOptions &LO,
+ const TargetInfo &TI) {
+ for (StringRef Name : Data.Names) {
+ if (!Name.compare_lower("host")) {
+ if (LO.OpenMPIsDevice)
+ return false;
+ continue;
+ }
+ if (!Name.compare_lower("nohost")) {
+ if (!LO.OpenMPIsDevice)
+ return false;
+ continue;
+ }
+ switch (TI.getTriple().getArch()) {
+ case llvm::Triple::arm:
+ case llvm::Triple::armeb:
+ case llvm::Triple::aarch64:
+ case llvm::Triple::aarch64_be:
+ case llvm::Triple::aarch64_32:
+ case llvm::Triple::ppc:
+ case llvm::Triple::ppc64:
+ case llvm::Triple::ppc64le:
+ case llvm::Triple::x86:
+ case llvm::Triple::x86_64:
+ if (Name.compare_lower("cpu"))
+ return false;
+ break;
+ case llvm::Triple::amdgcn:
+ case llvm::Triple::nvptx:
+ case llvm::Triple::nvptx64:
+ if (Name.compare_lower("gpu"))
+ return false;
+ break;
+ case llvm::Triple::UnknownArch:
+ case llvm::Triple::arc:
+ case llvm::Triple::avr:
+ case llvm::Triple::bpfel:
+ case llvm::Triple::bpfeb:
+ case llvm::Triple::hexagon:
+ case llvm::Triple::mips:
+ case llvm::Triple::mipsel:
+ case llvm::Triple::mips64:
+ case llvm::Triple::mips64el:
+ case llvm::Triple::msp430:
+ case llvm::Triple::r600:
+ case llvm::Triple::riscv32:
+ case llvm::Triple::riscv64:
+ case llvm::Triple::sparc:
+ case llvm::Triple::sparcv9:
+ case llvm::Triple::sparcel:
+ case llvm::Triple::systemz:
+ case llvm::Triple::tce:
+ case llvm::Triple::tcele:
+ case llvm::Triple::thumb:
+ case llvm::Triple::thumbeb:
+ case llvm::Triple::xcore:
+ case llvm::Triple::le32:
+ case llvm::Triple::le64:
+ case llvm::Triple::amdil:
+ case llvm::Triple::amdil64:
+ case llvm::Triple::hsail:
+ case llvm::Triple::hsail64:
+ case llvm::Triple::spir:
+ case llvm::Triple::spir64:
+ case llvm::Triple::kalimba:
+ case llvm::Triple::shave:
+ case llvm::Triple::lanai:
+ case llvm::Triple::wasm32:
+ case llvm::Triple::wasm64:
+ case llvm::Triple::renderscript32:
+ case llvm::Triple::renderscript64:
+ return false;
+ }
+ }
+ return true;
+}
+
+static CompleteOMPContextSelectorData
+translateAttrToContextSelectorData(ASTContext &C,
+ const OMPDeclareVariantAttr *A) {
+ CompleteOMPContextSelectorData Data;
+ if (!A)
+ return Data;
+ for (unsigned I = 0, E = A->scores_size(); I < E; ++I) {
+ Data.emplace_back();
+ auto CtxSet = static_cast<OpenMPContextSelectorSetKind>(
+ *std::next(A->ctxSelectorSets_begin(), I));
+ auto Ctx = static_cast<OpenMPContextSelectorKind>(
+ *std::next(A->ctxSelectors_begin(), I));
+ Data.back().CtxSet = CtxSet;
+ Data.back().Ctx = Ctx;
+ const Expr *Score = *std::next(A->scores_begin(), I);
+ Score->dump();
+ Data.back().Score = Score->EvaluateKnownConstInt(C);
+ switch (Ctx) {
+ case OMP_CTX_vendor:
+ assert(CtxSet == OMP_CTX_SET_implementation &&
+ "Expected implementation context selector set.");
+ Data.back().Names =
+ llvm::makeArrayRef(A->implVendors_begin(), A->implVendors_end());
+ break;
+ case OMP_CTX_kind:
+ assert(CtxSet == OMP_CTX_SET_device &&
+ "Expected device context selector set.");
+ Data.back().Names =
+ llvm::makeArrayRef(A->deviceKinds_begin(), A->deviceKinds_end());
+ break;
+ case OMP_CTX_unknown:
+ llvm_unreachable("Unknown context selector kind.");
+ }
+ }
+ return Data;
+}
+
+static bool
+matchesOpenMPContextImpl(const CompleteOMPContextSelectorData &ContextData,
+ const LangOptions &LO, const TargetInfo &TI) {
+ for (const OMPContextSelectorData &Data : ContextData) {
+ switch (Data.Ctx) {
+ case OMP_CTX_vendor:
+ assert(Data.CtxSet == OMP_CTX_SET_implementation &&
+ "Expected implementation context selector set.");
+ if (!checkContext<OMP_CTX_SET_implementation, OMP_CTX_vendor>(Data))
+ return false;
+ break;
+ case OMP_CTX_kind:
+ assert(Data.CtxSet == OMP_CTX_SET_device &&
+ "Expected device context selector set.");
+ if (!checkContext<OMP_CTX_SET_device, OMP_CTX_kind, const LangOptions &,
+ const TargetInfo &>(Data, LO, TI))
+ return false;
+ break;
+ case OMP_CTX_unknown:
+ llvm_unreachable("Unknown context selector kind.");
+ }
+ }
+ return true;
+}
+
+static bool isStrictSubset(const CompleteOMPContextSelectorData &LHS,
+ const CompleteOMPContextSelectorData &RHS) {
+ llvm::SmallDenseMap<std::pair<int, int>, llvm::StringSet<>, 4> RHSData;
+ for (const OMPContextSelectorData &D : RHS) {
+ auto &Pair = RHSData.FindAndConstruct(std::make_pair(D.CtxSet, D.Ctx));
+ Pair.getSecond().insert(D.Names.begin(), D.Names.end());
+ }
+ bool AllSetsAreEqual = true;
+ for (const OMPContextSelectorData &D : LHS) {
+ auto It = RHSData.find(std::make_pair(D.CtxSet, D.Ctx));
+ if (It == RHSData.end())
+ return false;
+ if (D.Names.size() > It->getSecond().size())
+ return false;
+ if (llvm::set_union(It->getSecond(), D.Names))
+ return false;
+ AllSetsAreEqual =
+ AllSetsAreEqual && (D.Names.size() == It->getSecond().size());
+ }
+
+ return LHS.size() != RHS.size() || !AllSetsAreEqual;
+}
+
+const OMPDeclareVariantAttr *
+clang::getBetterOpenMPContextMatch(ASTContext &C,
+ const OMPDeclareVariantAttr *LHSAttr,
+ const OMPDeclareVariantAttr *RHSAttr) {
+ const CompleteOMPContextSelectorData LHS =
+ translateAttrToContextSelectorData(C, LHSAttr);
+ const CompleteOMPContextSelectorData RHS =
+ translateAttrToContextSelectorData(C, RHSAttr);
+ bool LHSMatch = LHSAttr && matchesOpenMPContextImpl(LHS, C.getLangOpts(),
+ C.getTargetInfo());
+ bool RHSMatch = RHSAttr && matchesOpenMPContextImpl(RHS, C.getLangOpts(),
+ C.getTargetInfo());
+ bool LHSisOK = LHSMatch && !LHSAttr->isInherited();
+ bool RHSisOK = RHSMatch && !RHSAttr->isInherited();
+ if (!LHSisOK && !RHSisOK)
+ return nullptr;
+ if (LHSisOK && !RHSisOK)
+ return LHSAttr;
+ if (!LHSisOK && RHSisOK)
+ return RHSAttr;
+ assert(LHSisOK && RHSisOK && "broken invariant");
+
+ // Score is calculated as sum of all scores + 1.
+ llvm::APSInt LHSScore(llvm::APInt(64, 1), /*isUnsigned=*/false);
+ bool RHSIsSubsetOfLHS = isStrictSubset(RHS, LHS);
+ if (RHSIsSubsetOfLHS) {
+ LHSScore = llvm::APSInt::get(0);
+ } else {
+ for (const OMPContextSelectorData &Data : LHS) {
+ if (Data.Score.getBitWidth() > LHSScore.getBitWidth()) {
+ LHSScore = LHSScore.extend(Data.Score.getBitWidth()) + Data.Score;
+ } else if (Data.Score.getBitWidth() < LHSScore.getBitWidth()) {
+ LHSScore += Data.Score.extend(LHSScore.getBitWidth());
+ } else {
+ LHSScore += Data.Score;
+ }
+ }
+ }
+ llvm::APSInt RHSScore(llvm::APInt(64, 1), /*isUnsigned=*/false);
+ if (!RHSIsSubsetOfLHS && isStrictSubset(LHS, RHS)) {
+ RHSScore = llvm::APSInt::get(0);
+ } else {
+ for (const OMPContextSelectorData &Data : RHS) {
+ if (Data.Score.getBitWidth() > RHSScore.getBitWidth()) {
+ RHSScore = RHSScore.extend(Data.Score.getBitWidth()) + Data.Score;
+ } else if (Data.Score.getBitWidth() < RHSScore.getBitWidth()) {
+ RHSScore += Data.Score.extend(RHSScore.getBitWidth());
+ } else {
+ RHSScore += Data.Score;
+ }
+ }
+ }
+ return llvm::APSInt::compareValues(LHSScore, RHSScore) >= 0 ? LHSAttr
+ : RHSAttr;
+}
+
+bool clang::isOpenMPContextMatch(ASTContext &C,
+ const OMPDeclareVariantAttr *A) {
+ const CompleteOMPContextSelectorData Data =
+ translateAttrToContextSelectorData(C, A);
+ return matchesOpenMPContextImpl(Data, C.getLangOpts(), C.getTargetInfo());
+}
Index: clang/lib/AST/Expr.cpp
===================================================================
--- clang/lib/AST/Expr.cpp
+++ clang/lib/AST/Expr.cpp
@@ -1649,8 +1649,9 @@
Base->containsUnexpandedParameterPack()),
Base(Base), MemberDecl(MemberDecl), MemberDNLoc(NameInfo.getInfo()),
MemberLoc(NameInfo.getLoc()) {
- assert(!NameInfo.getName() ||
- MemberDecl->getDeclName() == NameInfo.getName());
+ // TODO: Add an OMP declare variant attribute on the variant declaration.
+ // assert(!NameInfo.getName() || MemberDecl->hasAttr<OMPDeclareVariantAttr>()
+ // || MemberDecl->getDeclName() == NameInfo.getName());
MemberExprBits.IsArrow = IsArrow;
MemberExprBits.HasQualifierOrFoundDecl = false;
MemberExprBits.HasTemplateKWAndArgsInfo = false;
Index: clang/include/clang/Sema/Sema.h
===================================================================
--- clang/include/clang/Sema/Sema.h
+++ clang/include/clang/Sema/Sema.h
@@ -9403,12 +9403,6 @@
MapT &Map, unsigned Selector = 0,
SourceRange SrcRange = SourceRange());
- /// Marks all the functions that might be required for the currently active
- /// OpenMP context.
- void markOpenMPDeclareVariantFuncsReferenced(SourceLocation Loc,
- FunctionDecl *Func,
- bool MightBeOdrUse);
-
public:
/// Struct to store the context selectors info for declare variant directive.
using OMPCtxStringType = SmallString<8>;
Index: clang/include/clang/AST/StmtOpenMP.h
===================================================================
--- clang/include/clang/AST/StmtOpenMP.h
+++ clang/include/clang/AST/StmtOpenMP.h
@@ -4594,6 +4594,18 @@
}
};
+class OMPDeclareVariantAttr;
+
+/// Helper to determine the best of two potential context matches. Note that
+/// nullptr are valid inputs but also valid outputs, e.g., if neither attribute
+/// describes a matching context.
+const OMPDeclareVariantAttr *
+getBetterOpenMPContextMatch(ASTContext &C, const OMPDeclareVariantAttr *LHSAttr,
+ const OMPDeclareVariantAttr *RHSAttr);
+
+/// Return true if the context described by \p A matches.
+bool isOpenMPContextMatch(ASTContext &C, const OMPDeclareVariantAttr *A);
+
} // end namespace clang
#endif
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
