[OPENMP50]Add support for 'parallel master taskloop' construct.
Added parsing/sema/codegen support for 'parallel master taskloop'
constructs. Some of the clauses, like 'grainsize', 'num_tasks', 'final'
and 'priority' are not supported in full, only constant expressions can
be used currently in these clauses.
llvm-svn: 374791
diff --git a/clang/lib/Sema/SemaOpenMP.cpp b/clang/lib/Sema/SemaOpenMP.cpp
index fe69b70..f717dc7 100644
--- a/clang/lib/Sema/SemaOpenMP.cpp
+++ b/clang/lib/Sema/SemaOpenMP.cpp
@@ -966,7 +966,8 @@
// In a parallel construct, if no default clause is present, these
// variables are shared.
DVar.ImplicitDSALoc = Iter->DefaultAttrLoc;
- if (isOpenMPParallelDirective(DVar.DKind) ||
+ if ((isOpenMPParallelDirective(DVar.DKind) &&
+ !isOpenMPTaskLoopDirective(DVar.DKind)) ||
isOpenMPTeamsDirective(DVar.DKind)) {
DVar.CKind = OMPC_shared;
return DVar;
@@ -3280,6 +3281,56 @@
AlwaysInlineAttr::Keyword_forceinline));
break;
}
+ case OMPD_parallel_master_taskloop: {
+ QualType KmpInt32Ty =
+ Context.getIntTypeForBitwidth(/*DestWidth=*/32, /*Signed=*/1)
+ .withConst();
+ QualType KmpUInt64Ty =
+ Context.getIntTypeForBitwidth(/*DestWidth=*/64, /*Signed=*/0)
+ .withConst();
+ QualType KmpInt64Ty =
+ Context.getIntTypeForBitwidth(/*DestWidth=*/64, /*Signed=*/1)
+ .withConst();
+ QualType VoidPtrTy = Context.VoidPtrTy.withConst().withRestrict();
+ QualType KmpInt32PtrTy =
+ Context.getPointerType(KmpInt32Ty).withConst().withRestrict();
+ Sema::CapturedParamNameType ParamsParallel[] = {
+ std::make_pair(".global_tid.", KmpInt32PtrTy),
+ std::make_pair(".bound_tid.", KmpInt32PtrTy),
+ std::make_pair(StringRef(), QualType()) // __context with shared vars
+ };
+ // Start a captured region for 'parallel'.
+ ActOnCapturedRegionStart(DSAStack->getConstructLoc(), CurScope, CR_OpenMP,
+ ParamsParallel, /*OpenMPCaptureLevel=*/1);
+ QualType Args[] = {VoidPtrTy};
+ FunctionProtoType::ExtProtoInfo EPI;
+ EPI.Variadic = true;
+ QualType CopyFnType = Context.getFunctionType(Context.VoidTy, Args, EPI);
+ Sema::CapturedParamNameType Params[] = {
+ std::make_pair(".global_tid.", KmpInt32Ty),
+ std::make_pair(".part_id.", KmpInt32PtrTy),
+ std::make_pair(".privates.", VoidPtrTy),
+ std::make_pair(
+ ".copy_fn.",
+ Context.getPointerType(CopyFnType).withConst().withRestrict()),
+ std::make_pair(".task_t.", Context.VoidPtrTy.withConst()),
+ std::make_pair(".lb.", KmpUInt64Ty),
+ std::make_pair(".ub.", KmpUInt64Ty),
+ std::make_pair(".st.", KmpInt64Ty),
+ std::make_pair(".liter.", KmpInt32Ty),
+ std::make_pair(".reductions.", VoidPtrTy),
+ std::make_pair(StringRef(), QualType()) // __context with shared vars
+ };
+ ActOnCapturedRegionStart(DSAStack->getConstructLoc(), CurScope, CR_OpenMP,
+ Params, /*OpenMPCaptureLevel=*/2);
+ // Mark this captured region as inlined, because we don't use outlined
+ // function directly.
+ getCurCapturedRegion()->TheCapturedDecl->addAttr(
+ AlwaysInlineAttr::CreateImplicit(
+ Context, {}, AttributeCommonInfo::AS_Keyword,
+ AlwaysInlineAttr::Keyword_forceinline));
+ break;
+ }
case OMPD_distribute_parallel_for_simd:
case OMPD_distribute_parallel_for: {
QualType KmpInt32Ty = Context.getIntTypeForBitwidth(32, 1).withConst();
@@ -4420,6 +4471,12 @@
ClausesWithImplicit, AStmt, StartLoc, EndLoc, VarsWithInheritedDSA);
AllowedNameModifiers.push_back(OMPD_taskloop);
break;
+ case OMPD_parallel_master_taskloop:
+ Res = ActOnOpenMPParallelMasterTaskLoopDirective(
+ ClausesWithImplicit, AStmt, StartLoc, EndLoc, VarsWithInheritedDSA);
+ AllowedNameModifiers.push_back(OMPD_taskloop);
+ AllowedNameModifiers.push_back(OMPD_parallel);
+ break;
case OMPD_distribute:
Res = ActOnOpenMPDistributeDirective(ClausesWithImplicit, AStmt, StartLoc,
EndLoc, VarsWithInheritedDSA);
@@ -6445,6 +6502,7 @@
DVar.CKind != OMPC_private))) ||
((isOpenMPWorksharingDirective(DKind) || DKind == OMPD_taskloop ||
DKind == OMPD_master_taskloop ||
+ DKind == OMPD_parallel_master_taskloop ||
isOpenMPDistributeDirective(DKind)) &&
!isOpenMPSimdDirective(DKind) && DVar.CKind != OMPC_unknown &&
DVar.CKind != OMPC_private && DVar.CKind != OMPC_lastprivate)) &&
@@ -9295,6 +9353,61 @@
NestedLoopCount, Clauses, AStmt, B);
}
+StmtResult Sema::ActOnOpenMPParallelMasterTaskLoopDirective(
+ ArrayRef<OMPClause *> Clauses, Stmt *AStmt, SourceLocation StartLoc,
+ SourceLocation EndLoc, VarsWithInheritedDSAType &VarsWithImplicitDSA) {
+ if (!AStmt)
+ return StmtError();
+
+ assert(isa<CapturedStmt>(AStmt) && "Captured statement expected");
+ auto *CS = cast<CapturedStmt>(AStmt);
+ // 1.2.2 OpenMP Language Terminology
+ // Structured block - An executable statement with a single entry at the
+ // top and a single exit at the bottom.
+ // The point of exit cannot be a branch out of the structured block.
+ // longjmp() and throw() must not violate the entry/exit criteria.
+ CS->getCapturedDecl()->setNothrow();
+ for (int ThisCaptureLevel =
+ getOpenMPCaptureLevels(OMPD_parallel_master_taskloop);
+ ThisCaptureLevel > 1; --ThisCaptureLevel) {
+ CS = cast<CapturedStmt>(CS->getCapturedStmt());
+ // 1.2.2 OpenMP Language Terminology
+ // Structured block - An executable statement with a single entry at the
+ // top and a single exit at the bottom.
+ // The point of exit cannot be a branch out of the structured block.
+ // longjmp() and throw() must not violate the entry/exit criteria.
+ CS->getCapturedDecl()->setNothrow();
+ }
+
+ OMPLoopDirective::HelperExprs B;
+ // In presence of clause 'collapse' or 'ordered' with number of loops, it will
+ // define the nested loops number.
+ unsigned NestedLoopCount = checkOpenMPLoop(
+ OMPD_parallel_master_taskloop, getCollapseNumberExpr(Clauses),
+ /*OrderedLoopCountExpr=*/nullptr, CS, *this, *DSAStack,
+ VarsWithImplicitDSA, B);
+ if (NestedLoopCount == 0)
+ return StmtError();
+
+ assert((CurContext->isDependentContext() || B.builtAll()) &&
+ "omp for loop exprs were not built");
+
+ // OpenMP, [2.9.2 taskloop Construct, Restrictions]
+ // The grainsize clause and num_tasks clause are mutually exclusive and may
+ // not appear on the same taskloop directive.
+ if (checkGrainsizeNumTasksClauses(*this, Clauses))
+ return StmtError();
+ // OpenMP, [2.9.2 taskloop Construct, Restrictions]
+ // If a reduction clause is present on the taskloop directive, the nogroup
+ // clause must not be specified.
+ if (checkReductionClauseWithNogroup(*this, Clauses))
+ return StmtError();
+
+ setFunctionHasBranchProtectedScope();
+ return OMPParallelMasterTaskLoopDirective::Create(
+ Context, StartLoc, EndLoc, NestedLoopCount, Clauses, AStmt, B);
+}
+
StmtResult Sema::ActOnOpenMPDistributeDirective(
ArrayRef<OMPClause *> Clauses, Stmt *AStmt, SourceLocation StartLoc,
SourceLocation EndLoc, VarsWithInheritedDSAType &VarsWithImplicitDSA) {
@@ -10186,6 +10299,10 @@
case OMPD_target_exit_data:
CaptureRegion = OMPD_task;
break;
+ case OMPD_parallel_master_taskloop:
+ if (NameModifier == OMPD_unknown || NameModifier == OMPD_taskloop)
+ CaptureRegion = OMPD_parallel;
+ break;
case OMPD_cancel:
case OMPD_parallel:
case OMPD_parallel_sections:
@@ -10259,6 +10376,7 @@
case OMPD_parallel_for_simd:
case OMPD_distribute_parallel_for:
case OMPD_distribute_parallel_for_simd:
+ case OMPD_parallel_master_taskloop:
// Do not capture num_threads-clause expressions.
break;
case OMPD_target_data:
@@ -10332,6 +10450,7 @@
case OMPD_taskloop:
case OMPD_taskloop_simd:
case OMPD_master_taskloop:
+ case OMPD_parallel_master_taskloop:
case OMPD_target_data:
case OMPD_target_enter_data:
case OMPD_target_exit_data:
@@ -10400,6 +10519,7 @@
case OMPD_taskloop:
case OMPD_taskloop_simd:
case OMPD_master_taskloop:
+ case OMPD_parallel_master_taskloop:
case OMPD_target_data:
case OMPD_target_enter_data:
case OMPD_target_exit_data:
@@ -10468,6 +10588,7 @@
case OMPD_taskloop:
case OMPD_taskloop_simd:
case OMPD_master_taskloop:
+ case OMPD_parallel_master_taskloop:
case OMPD_target_data:
case OMPD_target_enter_data:
case OMPD_target_exit_data:
@@ -10540,6 +10661,7 @@
case OMPD_taskloop:
case OMPD_taskloop_simd:
case OMPD_master_taskloop:
+ case OMPD_parallel_master_taskloop:
case OMPD_target_data:
case OMPD_target_enter_data:
case OMPD_target_exit_data:
@@ -10613,6 +10735,7 @@
case OMPD_taskloop:
case OMPD_taskloop_simd:
case OMPD_master_taskloop:
+ case OMPD_parallel_master_taskloop:
case OMPD_cancel:
case OMPD_parallel:
case OMPD_parallel_sections: