CGOpenMPRegionInfo(const CapturedStmt &CS,
const CGOpenMPRegionKind RegionKind,
- const RegionCodeGenTy &CodeGen)
+ const RegionCodeGenTy &CodeGen, OpenMPDirectiveKind Kind)
: CGCapturedStmtInfo(CS, CR_OpenMP), RegionKind(RegionKind),
- CodeGen(CodeGen) {}
+ CodeGen(CodeGen), Kind(Kind) {}
CGOpenMPRegionInfo(const CGOpenMPRegionKind RegionKind,
- const RegionCodeGenTy &CodeGen)
- : CGCapturedStmtInfo(CR_OpenMP), RegionKind(RegionKind),
- CodeGen(CodeGen) {}
+ const RegionCodeGenTy &CodeGen, OpenMPDirectiveKind Kind)
+ : CGCapturedStmtInfo(CR_OpenMP), RegionKind(RegionKind), CodeGen(CodeGen),
+ Kind(Kind) {}
/// \brief Get a variable or parameter for storing global thread id
/// inside OpenMP construct.
CGOpenMPRegionKind getRegionKind() const { return RegionKind; }
+ OpenMPDirectiveKind getDirectiveKind() const { return Kind; }
+
static bool classof(const CGCapturedStmtInfo *Info) {
return Info->getKind() == CR_OpenMP;
}
protected:
CGOpenMPRegionKind RegionKind;
const RegionCodeGenTy &CodeGen;
+ OpenMPDirectiveKind Kind;
};
/// \brief API for captured statement code generation in OpenMP constructs.
class CGOpenMPOutlinedRegionInfo : public CGOpenMPRegionInfo {
public:
CGOpenMPOutlinedRegionInfo(const CapturedStmt &CS, const VarDecl *ThreadIDVar,
- const RegionCodeGenTy &CodeGen)
- : CGOpenMPRegionInfo(CS, ParallelOutlinedRegion, CodeGen),
+ const RegionCodeGenTy &CodeGen,
+ OpenMPDirectiveKind Kind)
+ : CGOpenMPRegionInfo(CS, ParallelOutlinedRegion, CodeGen, Kind),
ThreadIDVar(ThreadIDVar) {
assert(ThreadIDVar != nullptr && "No ThreadID in OpenMP region.");
}
public:
CGOpenMPTaskOutlinedRegionInfo(const CapturedStmt &CS,
const VarDecl *ThreadIDVar,
- const RegionCodeGenTy &CodeGen)
- : CGOpenMPRegionInfo(CS, TaskOutlinedRegion, CodeGen),
+ const RegionCodeGenTy &CodeGen,
+ OpenMPDirectiveKind Kind)
+ : CGOpenMPRegionInfo(CS, TaskOutlinedRegion, CodeGen, Kind),
ThreadIDVar(ThreadIDVar) {
assert(ThreadIDVar != nullptr && "No ThreadID in OpenMP region.");
}
class CGOpenMPInlinedRegionInfo : public CGOpenMPRegionInfo {
public:
CGOpenMPInlinedRegionInfo(CodeGenFunction::CGCapturedStmtInfo *OldCSI,
- const RegionCodeGenTy &CodeGen)
- : CGOpenMPRegionInfo(InlinedRegion, CodeGen), OldCSI(OldCSI),
+ const RegionCodeGenTy &CodeGen,
+ OpenMPDirectiveKind Kind)
+ : CGOpenMPRegionInfo(InlinedRegion, CodeGen, Kind), OldCSI(OldCSI),
OuterRegionInfo(dyn_cast_or_null<CGOpenMPRegionInfo>(OldCSI)) {}
// \brief Retrieve the value of the context parameter.
llvm::Value *getContextValue() const override {
/// \param CodeGen Code generation sequence for combined directives. Includes
/// a list of functions used for code generation of implicitly inlined
/// regions.
- InlinedOpenMPRegionRAII(CodeGenFunction &CGF, const RegionCodeGenTy &CodeGen)
+ InlinedOpenMPRegionRAII(CodeGenFunction &CGF, const RegionCodeGenTy &CodeGen,
+ OpenMPDirectiveKind Kind)
: CGF(CGF) {
// Start emission for the construct.
CGF.CapturedStmtInfo =
- new CGOpenMPInlinedRegionInfo(CGF.CapturedStmtInfo, CodeGen);
+ new CGOpenMPInlinedRegionInfo(CGF.CapturedStmtInfo, CodeGen, Kind);
}
~InlinedOpenMPRegionRAII() {
// Restore original CapturedStmtInfo only if we're done with code emission.
InternalVars.clear();
}
-llvm::Value *
-CGOpenMPRuntime::emitParallelOutlinedFunction(const OMPExecutableDirective &D,
- const VarDecl *ThreadIDVar,
- const RegionCodeGenTy &CodeGen) {
+llvm::Value *CGOpenMPRuntime::emitParallelOutlinedFunction(
+ const OMPExecutableDirective &D, const VarDecl *ThreadIDVar,
+ OpenMPDirectiveKind InnermostKind, const RegionCodeGenTy &CodeGen) {
assert(ThreadIDVar->getType()->isPointerType() &&
"thread id variable must be of type kmp_int32 *");
const CapturedStmt *CS = cast<CapturedStmt>(D.getAssociatedStmt());
CodeGenFunction CGF(CGM, true);
- CGOpenMPOutlinedRegionInfo CGInfo(*CS, ThreadIDVar, CodeGen);
+ CGOpenMPOutlinedRegionInfo CGInfo(*CS, ThreadIDVar, CodeGen, InnermostKind);
CodeGenFunction::CGCapturedStmtRAII CapInfoRAII(CGF, &CGInfo);
return CGF.GenerateCapturedStmtFunction(*CS);
}
-llvm::Value *
-CGOpenMPRuntime::emitTaskOutlinedFunction(const OMPExecutableDirective &D,
- const VarDecl *ThreadIDVar,
- const RegionCodeGenTy &CodeGen) {
+llvm::Value *CGOpenMPRuntime::emitTaskOutlinedFunction(
+ const OMPExecutableDirective &D, const VarDecl *ThreadIDVar,
+ OpenMPDirectiveKind InnermostKind, const RegionCodeGenTy &CodeGen) {
assert(!ThreadIDVar->getType()->isPointerType() &&
"thread id variable must be of type kmp_int32 for tasks");
auto *CS = cast<CapturedStmt>(D.getAssociatedStmt());
CodeGenFunction CGF(CGM, true);
- CGOpenMPTaskOutlinedRegionInfo CGInfo(*CS, ThreadIDVar, CodeGen);
+ CGOpenMPTaskOutlinedRegionInfo CGInfo(*CS, ThreadIDVar, CodeGen,
+ InnermostKind);
CodeGenFunction::CGCapturedStmtRAII CapInfoRAII(CGF, &CGInfo);
return CGF.GenerateCapturedStmtFunction(*CS);
}
RTLFn = CGM.CreateRuntimeFunction(FnTy, /*Name*/ "__kmpc_cancel_barrier");
break;
}
+ case OMPRTL__kmpc_barrier: {
+ // Build void __kmpc_cancel_barrier(ident_t *loc, kmp_int32 global_tid);
+ llvm::Type *TypeParams[] = {getIdentTyPointerTy(), CGM.Int32Ty};
+ llvm::FunctionType *FnTy =
+ llvm::FunctionType::get(CGM.VoidTy, TypeParams, /*isVarArg*/ false);
+ RTLFn = CGM.CreateRuntimeFunction(FnTy, /*Name*/ "__kmpc_barrier");
+ break;
+ }
case OMPRTL__kmpc_for_static_fini: {
// Build void __kmpc_for_static_fini(ident_t *loc, kmp_int32 global_tid);
llvm::Type *TypeParams[] = {getIdentTyPointerTy(), CGM.Int32Ty};
CGF.EHStack.pushCleanup<CallEndCleanup<std::extent<decltype(Args)>::value>>(
NormalAndEHCleanup, createRuntimeFunction(OMPRTL__kmpc_end_critical),
llvm::makeArrayRef(Args));
- emitInlinedDirective(CGF, CriticalOpGen);
+ emitInlinedDirective(CGF, OMPD_critical, CriticalOpGen);
}
}
static void emitIfStmt(CodeGenFunction &CGF, llvm::Value *IfCond,
+ OpenMPDirectiveKind Kind,
const RegionCodeGenTy &BodyOpGen) {
llvm::Value *CallBool = CGF.EmitScalarConversion(
IfCond,
// Generate the branch (If-stmt)
CGF.Builder.CreateCondBr(CallBool, ThenBlock, ContBlock);
CGF.EmitBlock(ThenBlock);
- CGF.CGM.getOpenMPRuntime().emitInlinedDirective(CGF, BodyOpGen);
+ CGF.CGM.getOpenMPRuntime().emitInlinedDirective(CGF, Kind, BodyOpGen);
// Emit the rest of bblocks/branches
CGF.EmitBranch(ContBlock);
CGF.EmitBlock(ContBlock, true);
CGF.EmitRuntimeCall(createRuntimeFunction(OMPRTL__kmpc_master), Args);
typedef CallEndCleanup<std::extent<decltype(Args)>::value>
MasterCallEndCleanup;
- emitIfStmt(CGF, IsMaster, [&](CodeGenFunction &CGF) -> void {
+ emitIfStmt(CGF, IsMaster, OMPD_master, [&](CodeGenFunction &CGF) -> void {
CodeGenFunction::RunCleanupsScope Scope(CGF);
CGF.EHStack.pushCleanup<MasterCallEndCleanup>(
NormalAndEHCleanup, createRuntimeFunction(OMPRTL__kmpc_end_master),
CGF.EHStack.pushCleanup<CallEndCleanup<std::extent<decltype(Args)>::value>>(
NormalAndEHCleanup, createRuntimeFunction(OMPRTL__kmpc_end_taskgroup),
llvm::makeArrayRef(Args));
- emitInlinedDirective(CGF, TaskgroupOpGen);
+ emitInlinedDirective(CGF, OMPD_taskgroup, TaskgroupOpGen);
}
}
CGF.EmitRuntimeCall(createRuntimeFunction(OMPRTL__kmpc_single), Args);
typedef CallEndCleanup<std::extent<decltype(Args)>::value>
SingleCallEndCleanup;
- emitIfStmt(CGF, IsSingle, [&](CodeGenFunction &CGF) -> void {
+ emitIfStmt(CGF, IsSingle, OMPD_single, [&](CodeGenFunction &CGF) -> void {
CodeGenFunction::RunCleanupsScope Scope(CGF);
CGF.EHStack.pushCleanup<SingleCallEndCleanup>(
NormalAndEHCleanup, createRuntimeFunction(OMPRTL__kmpc_end_single),
CGF.EHStack.pushCleanup<CallEndCleanup<std::extent<decltype(Args)>::value>>(
NormalAndEHCleanup, createRuntimeFunction(OMPRTL__kmpc_end_ordered),
llvm::makeArrayRef(Args));
- emitInlinedDirective(CGF, OrderedOpGen);
+ emitInlinedDirective(CGF, OMPD_ordered, OrderedOpGen);
}
}
void CGOpenMPRuntime::emitBarrierCall(CodeGenFunction &CGF, SourceLocation Loc,
- OpenMPDirectiveKind Kind) {
+ OpenMPDirectiveKind Kind,
+ bool CheckForCancel) {
// Build call __kmpc_cancel_barrier(loc, thread_id);
+ // Build call __kmpc_barrier(loc, thread_id);
OpenMPLocationFlags Flags = OMP_IDENT_KMPC;
if (Kind == OMPD_for) {
Flags =
} else {
Flags = static_cast<OpenMPLocationFlags>(Flags | OMP_IDENT_BARRIER_IMPL);
}
- // Build call __kmpc_cancel_barrier(loc, thread_id);
- // Replace __kmpc_barrier() function by __kmpc_cancel_barrier() because this
- // one provides the same functionality and adds initial support for
- // cancellation constructs introduced in OpenMP 4.0. __kmpc_cancel_barrier()
- // is provided default by the runtime library so it safe to make such
- // replacement.
+ // Build call __kmpc_cancel_barrier(loc, thread_id) or __kmpc_barrier(loc,
+ // thread_id);
llvm::Value *Args[] = {emitUpdateLocation(CGF, Loc, Flags),
getThreadID(CGF, Loc)};
- CGF.EmitRuntimeCall(createRuntimeFunction(OMPRTL__kmpc_cancel_barrier), Args);
+ if (auto *OMPRegionInfo =
+ dyn_cast_or_null<CGOpenMPRegionInfo>(CGF.CapturedStmtInfo)) {
+ auto CancelDestination =
+ CGF.getOMPCancelDestination(OMPRegionInfo->getDirectiveKind());
+ if (CancelDestination.isValid()) {
+ auto *Result = CGF.EmitRuntimeCall(
+ createRuntimeFunction(OMPRTL__kmpc_cancel_barrier), Args);
+ if (CheckForCancel) {
+ // if (__kmpc_cancel_barrier()) {
+ // exit from construct;
+ // }
+ auto *ExitBB = CGF.createBasicBlock(".cancel.exit");
+ auto *ContBB = CGF.createBasicBlock(".cancel.continue");
+ auto *Cmp = CGF.Builder.CreateIsNotNull(Result);
+ CGF.Builder.CreateCondBr(Cmp, ExitBB, ContBB);
+ CGF.EmitBlock(ExitBB);
+ // exit from construct;
+ CGF.EmitBranchThroughCleanup(CancelDestination);
+ CGF.EmitBlock(ContBB, /*IsFinished=*/true);
+ }
+ return;
+ }
+ }
+ CGF.EmitRuntimeCall(createRuntimeFunction(OMPRTL__kmpc_barrier), Args);
}
/// \brief Schedule types for 'omp for' loops (these enumerators are taken from
}
void CGOpenMPRuntime::emitInlinedDirective(CodeGenFunction &CGF,
+ OpenMPDirectiveKind InnerKind,
const RegionCodeGenTy &CodeGen) {
- InlinedOpenMPRegionRAII Region(CGF, CodeGen);
+ InlinedOpenMPRegionRAII Region(CGF, CodeGen, InnerKind);
CGF.CapturedStmtInfo->EmitBody(CGF, /*S=*/nullptr);
}
assert(CancelRegion == OMPD_taskgroup);
CancelKind = CancelTaskgroup;
}
- llvm::Value *Args[] = {emitUpdateLocation(CGF, Loc), getThreadID(CGF, Loc),
- CGF.Builder.getInt32(CancelKind)};
- // Ignore return result until untied tasks are supported.
- auto *Result = CGF.EmitRuntimeCall(
- createRuntimeFunction(OMPRTL__kmpc_cancellationpoint), Args);
- // if (__kmpc_cancellationpoint())
- // exit from construct;
- auto *ExitBB = CGF.createBasicBlock(".cancel.exit");
- auto *ContBB = CGF.createBasicBlock(".cancel.continue");
- auto *Cmp = CGF.Builder.CreateIsNotNull(Result);
- CGF.Builder.CreateCondBr(Cmp, ExitBB, ContBB);
- CGF.EmitBlock(ExitBB);
- if (CancelRegion == OMPD_parallel || CancelRegion == OMPD_taskgroup) {
- CGF.EmitBranchThroughCleanup(CGF.ReturnBlock);
- } else {
- assert(CancelRegion == OMPD_for || CancelRegion == OMPD_sections);
- BreakStmt PseudoBrStmt(Loc);
- CGF.EmitBreakStmt(PseudoBrStmt);
+ if (auto *OMPRegionInfo =
+ dyn_cast_or_null<CGOpenMPRegionInfo>(CGF.CapturedStmtInfo)) {
+ auto CancelDest =
+ CGF.getOMPCancelDestination(OMPRegionInfo->getDirectiveKind());
+ if (CancelDest.isValid()) {
+ llvm::Value *Args[] = {emitUpdateLocation(CGF, Loc),
+ getThreadID(CGF, Loc),
+ CGF.Builder.getInt32(CancelKind)};
+ // Ignore return result until untied tasks are supported.
+ auto *Result = CGF.EmitRuntimeCall(
+ createRuntimeFunction(OMPRTL__kmpc_cancellationpoint), Args);
+ // if (__kmpc_cancellationpoint()) {
+ // __kmpc_cancel_barrier();
+ // exit from construct;
+ // }
+ auto *ExitBB = CGF.createBasicBlock(".cancel.exit");
+ auto *ContBB = CGF.createBasicBlock(".cancel.continue");
+ auto *Cmp = CGF.Builder.CreateIsNotNull(Result);
+ CGF.Builder.CreateCondBr(Cmp, ExitBB, ContBB);
+ CGF.EmitBlock(ExitBB);
+ // __kmpc_cancel_barrier();
+ emitBarrierCall(CGF, Loc, OMPD_unknown, /*CheckForCancel=*/false);
+ // exit from construct;
+ CGF.EmitBranchThroughCleanup(CancelDest);
+ CGF.EmitBlock(ContBB, /*IsFinished=*/true);
+ }
}
- CGF.EmitBlock(ContBB, /*IsFinished=*/true);
}
// Call to kmp_int32 __kmpc_cancel_barrier(ident_t *loc, kmp_int32
// global_tid);
OMPRTL__kmpc_cancel_barrier,
+ // Call to void __kmpc_barrier(ident_t *loc, kmp_int32 global_tid);
+ OMPRTL__kmpc_barrier,
// Call to void __kmpc_for_static_fini(ident_t *loc, kmp_int32 global_tid);
OMPRTL__kmpc_for_static_fini,
// Call to void __kmpc_serialized_parallel(ident_t *loc, kmp_int32
/// kmp_int32 BoundID, struct context_vars*).
/// \param D OpenMP directive.
/// \param ThreadIDVar Variable for thread id in the current OpenMP region.
+ /// \param InnermostKind Kind of innermost directive (for simple directives it
+ /// is a directive itself, for combined - its innermost directive).
/// \param CodeGen Code generation sequence for the \a D directive.
- virtual llvm::Value *
- emitParallelOutlinedFunction(const OMPExecutableDirective &D,
- const VarDecl *ThreadIDVar,
- const RegionCodeGenTy &CodeGen);
+ virtual llvm::Value *emitParallelOutlinedFunction(
+ const OMPExecutableDirective &D, const VarDecl *ThreadIDVar,
+ OpenMPDirectiveKind InnermostKind, const RegionCodeGenTy &CodeGen);
/// \brief Emits outlined function for the OpenMP task directive \a D. This
/// outlined function has type void(*)(kmp_int32 ThreadID, kmp_int32
/// PartID, struct context_vars*).
/// \param D OpenMP directive.
/// \param ThreadIDVar Variable for thread id in the current OpenMP region.
+ /// \param InnermostKind Kind of innermost directive (for simple directives it
+ /// is a directive itself, for combined - its innermost directive).
/// \param CodeGen Code generation sequence for the \a D directive.
///
- virtual llvm::Value *emitTaskOutlinedFunction(const OMPExecutableDirective &D,
- const VarDecl *ThreadIDVar,
- const RegionCodeGenTy &CodeGen);
+ virtual llvm::Value *emitTaskOutlinedFunction(
+ const OMPExecutableDirective &D, const VarDecl *ThreadIDVar,
+ OpenMPDirectiveKind InnermostKind, const RegionCodeGenTy &CodeGen);
/// \brief Cleans up references to the objects in finished function.
///
/// \brief Emit an implicit/explicit barrier for OpenMP threads.
/// \param Kind Directive for which this implicit barrier call must be
/// generated. Must be OMPD_barrier for explicit barrier generation.
+ /// \param CheckForCancel true if check for possible cancellation must be
+ /// performed, false otherwise.
///
virtual void emitBarrierCall(CodeGenFunction &CGF, SourceLocation Loc,
- OpenMPDirectiveKind Kind);
+ OpenMPDirectiveKind Kind,
+ bool CheckForCancel = true);
/// \brief Check if the specified \a ScheduleKind is static non-chunked.
/// This kind of worksharing directive is emitted without outer loop.
/// \brief Emit code for the directive that does not require outlining.
///
+ /// \param InnermostKind Kind of innermost directive (for simple directives it
+ /// is a directive itself, for combined - its innermost directive).
/// \param CodeGen Code generation sequence for the \a D directive.
virtual void emitInlinedDirective(CodeGenFunction &CGF,
+ OpenMPDirectiveKind InnermostKind,
const RegionCodeGenTy &CodeGen);
/// \brief Emit a code for reduction clause. Next code should be emitted for
/// reduction:
static void emitCommonOMPParallelDirective(CodeGenFunction &CGF,
const OMPExecutableDirective &S,
+ OpenMPDirectiveKind InnermostKind,
const RegionCodeGenTy &CodeGen) {
auto CS = cast<CapturedStmt>(S.getAssociatedStmt());
auto CapturedStruct = CGF.GenerateCapturedStmtArgument(*CS);
auto OutlinedFn = CGF.CGM.getOpenMPRuntime().emitParallelOutlinedFunction(
- S, *CS->getCapturedDecl()->param_begin(), CodeGen);
+ S, *CS->getCapturedDecl()->param_begin(), InnermostKind, CodeGen);
if (auto C = S.getSingleClause(OMPC_num_threads)) {
CodeGenFunction::RunCleanupsScope NumThreadsScope(CGF);
auto NumThreadsClause = cast<OMPNumThreadsClause>(C);
CGF.CGM.getOpenMPRuntime().emitBarrierCall(CGF, S.getLocStart(),
OMPD_unknown);
};
- emitCommonOMPParallelDirective(*this, S, CodeGen);
+ emitCommonOMPParallelDirective(*this, S, OMPD_parallel, CodeGen);
}
void CodeGenFunction::EmitOMPLoopBody(const OMPLoopDirective &D,
CGF.EmitBlock(ContBlock, true);
}
};
- CGM.getOpenMPRuntime().emitInlinedDirective(*this, CodeGen);
+ CGM.getOpenMPRuntime().emitInlinedDirective(*this, OMPD_simd, CodeGen);
}
void CodeGenFunction::EmitOMPForOuterLoop(OpenMPScheduleClauseKind ScheduleKind,
auto &&CodeGen = [&S, &HasLastprivates](CodeGenFunction &CGF) {
HasLastprivates = CGF.EmitOMPWorksharingLoop(S);
};
- CGM.getOpenMPRuntime().emitInlinedDirective(*this, CodeGen);
+ CGM.getOpenMPRuntime().emitInlinedDirective(*this, OMPD_for, CodeGen);
// Emit an implicit barrier at the end.
if (!S.getSingleClause(OMPC_nowait) || HasLastprivates) {
auto &&CodeGen = [&S, &HasLastprivates](CodeGenFunction &CGF) {
HasLastprivates = CGF.EmitOMPWorksharingLoop(S);
};
- CGM.getOpenMPRuntime().emitInlinedDirective(*this, CodeGen);
+ CGM.getOpenMPRuntime().emitInlinedDirective(*this, OMPD_simd, CodeGen);
// Emit an implicit barrier at the end.
if (!S.getSingleClause(OMPC_nowait) || HasLastprivates) {
CGF.EmitLoadOfScalar(IL, S.getLocStart())));
};
- CGM.getOpenMPRuntime().emitInlinedDirective(*this, CodeGen);
+ CGM.getOpenMPRuntime().emitInlinedDirective(*this, OMPD_sections, CodeGen);
// Emit barrier for lastprivates only if 'sections' directive has 'nowait'
// clause. Otherwise the barrier will be generated by the codegen for the
// directive.
CGF.EmitOMPPrivateClause(S, SingleScope);
(void)SingleScope.Privatize();
- CGF.BreakContinueStack.push_back(
- BreakContinue(CGF.getJumpDestInCurrentScope(
- CGF.createBasicBlock("omp.sections.exit")),
- JumpDest()));
CGF.EmitStmt(Stmt);
- CGF.EmitBlock(CGF.BreakContinueStack.back().BreakBlock.getBlock());
- CGF.BreakContinueStack.pop_back();
};
CGM.getOpenMPRuntime().emitSingleRegion(*this, CodeGen, S.getLocStart(),
llvm::None, llvm::None, llvm::None,
S.getSingleClause(OMPC_nowait)) {
// Emit implicit barrier to synchronize threads and avoid data races on
// initialization of firstprivate variables.
- CGM.getOpenMPRuntime().emitBarrierCall(*this, S.getLocStart(),
- OMPD_unknown);
+ CGM.getOpenMPRuntime().emitBarrierCall(*this, S.getLocStart(), OMPD_unknown);
}
return OMPD_single;
}
CGF.EmitStmt(cast<CapturedStmt>(S.getAssociatedStmt())->getCapturedStmt());
CGF.EnsureInsertPoint();
};
- CGM.getOpenMPRuntime().emitInlinedDirective(*this, CodeGen);
+ CGM.getOpenMPRuntime().emitInlinedDirective(*this, OMPD_section, CodeGen);
}
void CodeGenFunction::EmitOMPSingleDirective(const OMPSingleDirective &S) {
CGF.CGM.getOpenMPRuntime().emitBarrierCall(CGF, S.getLocStart(),
OMPD_parallel);
};
- emitCommonOMPParallelDirective(*this, S, CodeGen);
+ emitCommonOMPParallelDirective(*this, S, OMPD_for, CodeGen);
}
void CodeGenFunction::EmitOMPParallelForSimdDirective(
CGF.CGM.getOpenMPRuntime().emitBarrierCall(CGF, S.getLocStart(),
OMPD_parallel);
};
- emitCommonOMPParallelDirective(*this, S, CodeGen);
+ emitCommonOMPParallelDirective(*this, S, OMPD_simd, CodeGen);
}
void CodeGenFunction::EmitOMPParallelSectionsDirective(
CGF.CGM.getOpenMPRuntime().emitBarrierCall(CGF, S.getLocStart(),
OMPD_parallel);
};
- emitCommonOMPParallelDirective(*this, S, CodeGen);
+ emitCommonOMPParallelDirective(*this, S, OMPD_sections, CodeGen);
}
void CodeGenFunction::EmitOMPTaskDirective(const OMPTaskDirective &S) {
}
CGF.EmitStmt(CS->getCapturedStmt());
};
- auto OutlinedFn =
- CGM.getOpenMPRuntime().emitTaskOutlinedFunction(S, *I, CodeGen);
+ auto OutlinedFn = CGM.getOpenMPRuntime().emitTaskOutlinedFunction(
+ S, *I, OMPD_task, CodeGen);
// Check if we should emit tied or untied task.
bool Tied = !S.getSingleClause(OMPC_untied);
// Check if the task is final
S.getV(), S.getExpr(), S.getUpdateExpr(),
S.isXLHSInRHSPart(), S.getLocStart());
};
- CGM.getOpenMPRuntime().emitInlinedDirective(*this, CodeGen);
+ CGM.getOpenMPRuntime().emitInlinedDirective(*this, OMPD_atomic, CodeGen);
}
void CodeGenFunction::EmitOMPTargetDirective(const OMPTargetDirective &) {
llvm_unreachable("CodeGen for 'omp cancel' is not supported yet.");
}
+CodeGenFunction::JumpDest
+CodeGenFunction::getOMPCancelDestination(OpenMPDirectiveKind Kind) {
+ if (Kind == OMPD_parallel || Kind == OMPD_task)
+ return ReturnBlock;
+ else if (Kind == OMPD_for || Kind == OMPD_section || Kind == OMPD_sections)
+ return BreakContinueStack.empty() ? JumpDest()
+ : BreakContinueStack.back().BreakBlock;
+ return JumpDest();
+}
const llvm::function_ref<void(CodeGenFunction &)> &BodyGen,
const llvm::function_ref<void(CodeGenFunction &)> &PostIncGen);
+ JumpDest getOMPCancelDestination(OpenMPDirectiveKind Kind);
+
private:
/// Helpers for the OpenMP loop directives.
static int a;
#pragma omp barrier
// CHECK: [[GTID:%.+]] = call i32 @__kmpc_global_thread_num([[IDENT_T]]* [[LOC]])
- // CHECK: call i32 @__kmpc_cancel_barrier([[IDENT_T]]* [[EXPLICIT_BARRIER_LOC]], i32 [[GTID]])
+ // CHECK: call void @__kmpc_barrier([[IDENT_T]]* [[EXPLICIT_BARRIER_LOC]], i32 [[GTID]])
// CHECK: call {{.+}} [[TMAIN_INT:@.+]](i{{[0-9][0-9]}}
// CHECK: call {{.+}} [[TMAIN_CHAR:@.+]](i{{[0-9]}}
return tmain(argc) + tmain(argv[0][0]) + a;
// CHECK: define {{.+}} [[TMAIN_INT]](
// CHECK: [[GTID:%.+]] = call i32 @__kmpc_global_thread_num([[IDENT_T]]* [[LOC]])
-// CHECK: call i32 @__kmpc_cancel_barrier([[IDENT_T]]* [[EXPLICIT_BARRIER_LOC]], i32 [[GTID]])
+// CHECK: call void @__kmpc_barrier([[IDENT_T]]* [[EXPLICIT_BARRIER_LOC]], i32 [[GTID]])
// CHECK: define {{.+}} [[TMAIN_CHAR]](
// CHECK: [[GTID:%.+]] = call i32 @__kmpc_global_thread_num([[IDENT_T]]* [[LOC]])
-// CHECK: call i32 @__kmpc_cancel_barrier([[IDENT_T]]* [[EXPLICIT_BARRIER_LOC]], i32 [[GTID]])
+// CHECK: call void @__kmpc_barrier([[IDENT_T]]* [[EXPLICIT_BARRIER_LOC]], i32 [[GTID]])
#endif
#pragma omp cancellation point sections
}
// CHECK: call i32 @__kmpc_single(
+// CHECK-NOT: @__kmpc_cancellationpoint
+// CHECK: call void @__kmpc_end_single(
+// CHECK: call void @__kmpc_barrier(%ident_t*
+#pragma omp sections
+{
+#pragma omp cancellation point sections
+#pragma omp section
+ {
+#pragma omp cancellation point sections
+ }
+}
+// CHECK: call void @__kmpc_for_static_init_4(
// CHECK: [[RES:%.+]] = call i32 @__kmpc_cancellationpoint(%ident_t* {{[^,]+}}, i32 [[GTID]], i32 3)
// CHECK: [[CMP:%.+]] = icmp ne i32 [[RES]], 0
// CHECK: br i1 [[CMP]], label %[[EXIT:[^,].+]], label %[[CONTINUE:.+]]
// CHECK: [[EXIT]]
+// CHECK: call i32 @__kmpc_cancel_barrier(%ident_t*
// CHECK: br label
// CHECK: [[CONTINUE]]
// CHECK: br label
-// CHECK: call void @__kmpc_end_single(
+// CHECK: [[RES:%.+]] = call i32 @__kmpc_cancellationpoint(%ident_t* {{[^,]+}}, i32 [[GTID]], i32 3)
+// CHECK: [[CMP:%.+]] = icmp ne i32 [[RES]], 0
+// CHECK: br i1 [[CMP]], label %[[EXIT:[^,].+]], label %[[CONTINUE:.+]]
+// CHECK: [[EXIT]]
+// CHECK: call i32 @__kmpc_cancel_barrier(%ident_t*
+// CHECK: br label
+// CHECK: [[CONTINUE]]
+// CHECK: br label
+// CHECK: call void @__kmpc_for_static_fini(
#pragma omp for
for (int i = 0; i < argc; ++i) {
#pragma omp cancellation point for
// CHECK: [[CMP:%.+]] = icmp ne i32 [[RES]], 0
// CHECK: br i1 [[CMP]], label %[[EXIT:[^,].+]], label %[[CONTINUE:.+]]
// CHECK: [[EXIT]]
+// CHECK: call i32 @__kmpc_cancel_barrier(%ident_t*
// CHECK: br label
// CHECK: [[CONTINUE]]
// CHECK: br label
// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: call void @__kmpc_barrier(%ident_t*
#pragma omp task
{
#pragma omp cancellation point taskgroup
// CHECK: [[CMP:%.+]] = icmp ne i32 [[RES]], 0
// CHECK: br i1 [[CMP]], label %[[EXIT:[^,]+]],
// CHECK: [[EXIT]]
+// CHECK: call i32 @__kmpc_cancel_barrier(%ident_t*
// CHECK: br label %[[RETURN:.+]]
// CHECK: [[RETURN]]
// CHECK: ret void
// CHECK: [[CMP:%.+]] = icmp ne i32 [[RES]], 0
// CHECK: br i1 [[CMP]], label %[[EXIT:[^,]+]],
// CHECK: [[EXIT]]
+// CHECK: call i32 @__kmpc_cancel_barrier(%ident_t*
// CHECK: br label %[[RETURN:.+]]
// CHECK: [[RETURN]]
// CHECK: ret i32 0
}
// CHECK: [[LOOP1_END]]
// CHECK: call void @__kmpc_for_static_fini([[IDENT_T_TY]]* [[DEFAULT_LOC]], i32 [[GTID]])
-// CHECK-NOT: __kmpc_cancel_barrier
+// CHECK-NOT: __kmpc_barrier
// CHECK: ret void
}
}
// CHECK: [[LOOP1_END]]
// CHECK: call void @__kmpc_for_static_fini([[IDENT_T_TY]]* [[DEFAULT_LOC]], i32 [[GTID]])
-// CHECK: call {{.+}} @__kmpc_cancel_barrier([[IDENT_T_TY]]* [[IMPLICIT_BARRIER_LOC]], i32 [[GTID]])
+// CHECK: call {{.+}} @__kmpc_barrier([[IDENT_T_TY]]* [[IMPLICIT_BARRIER_LOC]], i32 [[GTID]])
// CHECK: ret void
}
// CHECK: [[O_LOOP1_END]]
// CHECK: call void @__kmpc_for_static_fini([[IDENT_T_TY]]* [[DEFAULT_LOC]], i32 [[GTID]])
-// CHECK: call {{.+}} @__kmpc_cancel_barrier([[IDENT_T_TY]]* [[IMPLICIT_BARRIER_LOC]], i32 [[GTID]])
+// CHECK: call {{.+}} @__kmpc_barrier([[IDENT_T_TY]]* [[IMPLICIT_BARRIER_LOC]], i32 [[GTID]])
// CHECK: ret void
}
}
// CHECK: [[LOOP1_END]]
// CHECK: [[O_LOOP1_END]]
-// CHECK: call {{.+}} @__kmpc_cancel_barrier([[IDENT_T_TY]]* [[IMPLICIT_BARRIER_LOC]], i32 [[GTID]])
+// CHECK: call {{.+}} @__kmpc_barrier([[IDENT_T_TY]]* [[IMPLICIT_BARRIER_LOC]], i32 [[GTID]])
// CHECK: ret void
}
}
// CHECK: [[LOOP1_END]]
// CHECK: [[O_LOOP1_END]]
-// CHECK: call {{.+}} @__kmpc_cancel_barrier([[IDENT_T_TY]]* [[IMPLICIT_BARRIER_LOC]], i32 [[GTID]])
+// CHECK: call {{.+}} @__kmpc_barrier([[IDENT_T_TY]]* [[IMPLICIT_BARRIER_LOC]], i32 [[GTID]])
// CHECK: ret void
}
}
// CHECK: [[LOOP1_END]]
// CHECK: [[O_LOOP1_END]]
-// CHECK: call {{.+}} @__kmpc_cancel_barrier([[IDENT_T_TY]]* [[IMPLICIT_BARRIER_LOC]], i32 [[GTID]])
+// CHECK: call {{.+}} @__kmpc_barrier([[IDENT_T_TY]]* [[IMPLICIT_BARRIER_LOC]], i32 [[GTID]])
// CHECK: ret void
}
}
// CHECK: [[LOOP1_END]]
// CHECK: [[O_LOOP1_END]]
-// CHECK: call {{.+}} @__kmpc_cancel_barrier([[IDENT_T_TY]]* [[IMPLICIT_BARRIER_LOC]], i32 [[GTID]])
+// CHECK: call {{.+}} @__kmpc_barrier([[IDENT_T_TY]]* [[IMPLICIT_BARRIER_LOC]], i32 [[GTID]])
// CHECK: ret void
}
// LAMBDA: [[G_PRIVATE_ADDR:%.+]] = alloca i{{[0-9]+}},
// LAMBDA: [[G_VAL:%.+]] = load volatile i{{[0-9]+}}, i{{[0-9]+}}* [[G]]
// LAMBDA: store i{{[0-9]+}} [[G_VAL]], i{{[0-9]+}}* [[G_PRIVATE_ADDR]]
- // LAMBDA: call i32 @__kmpc_cancel_barrier(
+ // LAMBDA: call void @__kmpc_barrier(
g = 1;
// LAMBDA: call void @__kmpc_for_static_init_4(
// LAMBDA: store volatile i{{[0-9]+}} 1, i{{[0-9]+}}* [[G_PRIVATE_ADDR]],
// BLOCKS: [[G_PRIVATE_ADDR:%.+]] = alloca i{{[0-9]+}},
// BLOCKS: [[G_VAL:%.+]] = load volatile i{{[0-9]+}}, i{{[0-9]+}}* [[G]]
// BLOCKS: store i{{[0-9]+}} [[G_VAL]], i{{[0-9]+}}* [[G_PRIVATE_ADDR]]
- // BLOCKS: call i32 @__kmpc_cancel_barrier(
+ // BLOCKS: call void @__kmpc_barrier(
g = 1;
// BLOCKS: call void @__kmpc_for_static_init_4(
// BLOCKS: store volatile i{{[0-9]+}} 1, i{{[0-9]+}}* [[G_PRIVATE_ADDR]],
// CHECK: call {{.*}} [[ST_TY_DESTR]]([[ST_TY]]* [[ST_TY_TEMP]])
// Synchronization for initialization.
-// CHECK: call i32 @__kmpc_cancel_barrier(%{{.+}}* [[IMPLICIT_BARRIER_LOC]], i{{[0-9]+}} [[GTID]])
+// CHECK: call void @__kmpc_barrier(%{{.+}}* [[IMPLICIT_BARRIER_LOC]], i{{[0-9]+}} [[GTID]])
// CHECK: call void @__kmpc_for_static_init_4(
// CHECK: call void @__kmpc_for_static_fini(
// ~(firstprivate var), ~(firstprivate s_arr)
// CHECK-DAG: call {{.*}} [[S_FLOAT_TY_DESTR]]([[S_FLOAT_TY]]* [[VAR_PRIV]])
// CHECK-DAG: call {{.*}} [[S_FLOAT_TY_DESTR]]([[S_FLOAT_TY]]*
-// CHECK: call i32 @__kmpc_cancel_barrier(%{{.+}}* [[IMPLICIT_BARRIER_LOC]], i{{[0-9]+}} [[GTID]])
+// CHECK: call void @__kmpc_barrier(%{{.+}}* [[IMPLICIT_BARRIER_LOC]], i{{[0-9]+}} [[GTID]])
// CHECK: = call {{.*}}i{{.+}} [[TMAIN_INT:@.+]]()
// Synchronization for initialization.
// CHECK: [[GTID_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[GTID_ADDR_ADDR]]
// CHECK: [[GTID:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[GTID_REF]]
-// CHECK: call i32 @__kmpc_cancel_barrier(%{{.+}}* [[IMPLICIT_BARRIER_LOC]], i{{[0-9]+}} [[GTID]])
+// CHECK: call void @__kmpc_barrier(%{{.+}}* [[IMPLICIT_BARRIER_LOC]], i{{[0-9]+}} [[GTID]])
// CHECK: call void @__kmpc_for_static_init_4(
// CHECK: call void @__kmpc_for_static_fini(
}
// CHECK: [[SIMPLE_LOOP1_END]]
// CHECK: call void @__kmpc_for_static_fini(%ident_t* {{.+}}, i32 %{{.+}})
-// CHECK: call i32 @__kmpc_cancel_barrier(%ident_t* {{.+}}, i32 %{{.+}})
+// CHECK: call void @__kmpc_barrier(%ident_t* {{.+}}, i32 %{{.+}})
long long k = get_val();
// CHECK: [[LIN0_2:%.+]] = load i64, i64* [[LIN0]]
// CHECK-NEXT: [[LIN_ADD2:%.+]] = add nsw i64 [[LIN0_2]], 27
// CHECK-NEXT: store i64 [[LIN_ADD2]], i64* [[K_VAR]]
-// CHECK: call i32 @__kmpc_cancel_barrier(%ident_t* {{.+}}, i32 %{{.+}})
+// CHECK: call void @__kmpc_barrier(%ident_t* {{.+}}, i32 %{{.+}})
int lin = 12;
#pragma omp for simd linear(lin : get_val()), linear(g_ptr)
// CHECK: store i32 {{.+}}, i32* [[LIN_VAR]],
// CHECK: [[GLINSTART:.+]] = load double*, double** [[GLIN_START]]
// CHECK: store double* {{.*}}[[GLIN_VAR]]
-// CHECK: call i32 @__kmpc_cancel_barrier(%ident_t* {{.+}}, i32 %{{.+}})
+// CHECK: call void @__kmpc_barrier(%ident_t* {{.+}}, i32 %{{.+}})
#pragma omp for simd
// CHECK: call void @__kmpc_for_static_init_4(%ident_t* {{[^,]+}}, i32 %{{[^,]+}}, i32 34, i32* %{{[^,]+}}, i32* [[LB:%[^,]+]], i32* [[UB:%[^,]+]], i32* [[STRIDE:%[^,]+]], i32 1, i32 1)
}
// CHECK: [[SIMPLE_LOOP4_END]]
// CHECK: call void @__kmpc_for_static_fini(%ident_t* {{.+}}, i32 %{{.+}})
-// CHECK: call i32 @__kmpc_cancel_barrier(%ident_t* {{.+}}, i32 %{{.+}})
+// CHECK: call void @__kmpc_barrier(%ident_t* {{.+}}, i32 %{{.+}})
#pragma omp for simd
// CHECK: call void @__kmpc_for_static_init_4(%ident_t* {{[^,]+}}, i32 %{{[^,]+}}, i32 34, i32* %{{[^,]+}}, i32* [[LB:%[^,]+]], i32* [[UB:%[^,]+]], i32* [[STRIDE:%[^,]+]], i32 1, i32 1)
}
// CHECK: [[SIMPLE_LOOP5_END]]
// CHECK: call void @__kmpc_for_static_fini(%ident_t* {{.+}}, i32 %{{.+}})
-// CHECK: call i32 @__kmpc_cancel_barrier(%ident_t* {{.+}}, i32 %{{.+}})
+// CHECK: call void @__kmpc_barrier(%ident_t* {{.+}}, i32 %{{.+}})
// CHECK-NOT: mul i32 %{{.+}}, 10
#pragma omp for simd
// CHECK-NEXT: br label {{%.+}}
// CHECK: [[T1_END]]
// CHECK: call void @__kmpc_for_static_fini(%ident_t* {{.+}}, i32 %{{.+}})
-// CHECK: call i32 @__kmpc_cancel_barrier(%ident_t* {{.+}}, i32 %{{.+}})
+// CHECK: call void @__kmpc_barrier(%ident_t* {{.+}}, i32 %{{.+}})
// CHECK: ret i32 0
//
void inst_templ1() {
}
// CHECK: [[IT_END]]
// CHECK: call void @__kmpc_for_static_fini(%ident_t* {{.+}}, i32 %{{.+}})
-// CHECK: call i32 @__kmpc_cancel_barrier(%ident_t* {{.+}}, i32 %{{.+}})
+// CHECK: call void @__kmpc_barrier(%ident_t* {{.+}}, i32 %{{.+}})
// CHECK: ret void
}
// CHECK-NEXT: store i32 3, i32* [[I:%[^,]+]]
// CHECK-NEXT: store i32 5, i32* [[I:%[^,]+]]
// CHECK-NEXT: store i16 9, i16* [[I:%[^,]+]]
-// CHECK: call i32 @__kmpc_cancel_barrier(%ident_t* {{.+}}, i32 %{{.+}})
+// CHECK: call void @__kmpc_barrier(%ident_t* {{.+}}, i32 %{{.+}})
// CHECK: ret void
}
}
// CHECK: [[LOOP1_END]]
// CHECK: [[O_LOOP1_END]]
-// CHECK: call {{.+}} @__kmpc_cancel_barrier([[IDENT_T_TY]]* [[IMPLICIT_BARRIER_LOC]], i32 [[GTID]])
+// CHECK: call {{.+}} @__kmpc_barrier([[IDENT_T_TY]]* [[IMPLICIT_BARRIER_LOC]], i32 [[GTID]])
// CHECK: ret void
}
}
// CHECK: [[LOOP1_END]]
// CHECK: [[O_LOOP1_END]]
-// CHECK: call {{.+}} @__kmpc_cancel_barrier([[IDENT_T_TY]]* [[IMPLICIT_BARRIER_LOC]], i32 [[GTID]])
+// CHECK: call {{.+}} @__kmpc_barrier([[IDENT_T_TY]]* [[IMPLICIT_BARRIER_LOC]], i32 [[GTID]])
// CHECK: ret void
}
}
// CHECK: [[LOOP1_END]]
// CHECK: [[O_LOOP1_END]]
-// CHECK: call {{.+}} @__kmpc_cancel_barrier([[IDENT_T_TY]]* [[IMPLICIT_BARRIER_LOC]], i32 [[GTID]])
+// CHECK: call {{.+}} @__kmpc_barrier([[IDENT_T_TY]]* [[IMPLICIT_BARRIER_LOC]], i32 [[GTID]])
// CHECK: ret void
}
}
// CHECK: [[LOOP1_END]]
// CHECK: [[O_LOOP1_END]]
-// CHECK: call {{.+}} @__kmpc_cancel_barrier([[IDENT_T_TY]]* [[IMPLICIT_BARRIER_LOC]], i32 [[GTID]])
+// CHECK: call {{.+}} @__kmpc_barrier([[IDENT_T_TY]]* [[IMPLICIT_BARRIER_LOC]], i32 [[GTID]])
// CHECK: ret void
}
// CHECK: [[CHUNK_SIZE:%.+]] = sext i8 [[CHUNK_VAL]] to i64
// CHECK: call void @__kmpc_for_static_init_8u([[IDENT_T_TY]]* [[DEFAULT_LOC:@[^,]+]], i32 [[GTID:%[^,]+]], i32 33, i32* [[IS_LAST:%[^,]+]], i64* [[OMP_LB:%[^,]+]], i64* [[OMP_UB:%[^,]+]], i64* [[OMP_ST:%[^,]+]], i64 1, i64 [[CHUNK_SIZE]])
// CHECK: call void @__kmpc_for_static_fini([[IDENT_T_TY]]* [[DEFAULT_LOC]], i32 [[GTID]])
-// CHECK: __kmpc_cancel_barrier
+// CHECK: __kmpc_barrier
#pragma omp parallel for schedule(static, char(a))
for (unsigned long long i = 1; i < 2; ++i) {
}
}
// CHECK: [[LOOP1_END]]
// CHECK: call void @__kmpc_for_static_fini([[IDENT_T_TY]]* [[DEFAULT_LOC]], i32 [[GTID]])
-// CHECK: call {{.+}} @__kmpc_cancel_barrier([[IDENT_T_TY]]* [[DEFAULT_LOC_BARRIER:[@%].+]], i32 [[GTID]])
+// CHECK: call {{.+}} @__kmpc_barrier([[IDENT_T_TY]]* [[DEFAULT_LOC_BARRIER:[@%].+]], i32 [[GTID]])
// CHECK: ret void
}
}
// CHECK: [[LOOP1_END]]
// CHECK: call void @__kmpc_for_static_fini([[IDENT_T_TY]]* [[DEFAULT_LOC]], i32 [[GTID]])
-// CHECK: call {{.+}} @__kmpc_cancel_barrier([[IDENT_T_TY]]* [[DEFAULT_LOC_BARRIER:[@%].+]], i32 [[GTID]])
+// CHECK: call {{.+}} @__kmpc_barrier([[IDENT_T_TY]]* [[DEFAULT_LOC_BARRIER:[@%].+]], i32 [[GTID]])
// CHECK: ret void
}
// CHECK: [[O_LOOP1_END]]
// CHECK: call void @__kmpc_for_static_fini([[IDENT_T_TY]]* [[DEFAULT_LOC]], i32 [[GTID]])
-// CHECK: call {{.+}} @__kmpc_cancel_barrier([[IDENT_T_TY]]* [[DEFAULT_LOC_BARRIER:[@%].+]], i32 [[GTID]])
+// CHECK: call {{.+}} @__kmpc_barrier([[IDENT_T_TY]]* [[DEFAULT_LOC_BARRIER:[@%].+]], i32 [[GTID]])
// CHECK: ret void
}
}
// CHECK: [[LOOP1_END]]
// CHECK: [[O_LOOP1_END]]
-// CHECK: call {{.+}} @__kmpc_cancel_barrier([[IDENT_T_TY]]* [[DEFAULT_LOC_BARRIER:[@%].+]], i32 [[GTID]])
+// CHECK: call {{.+}} @__kmpc_barrier([[IDENT_T_TY]]* [[DEFAULT_LOC_BARRIER:[@%].+]], i32 [[GTID]])
// CHECK: ret void
}
}
// CHECK: [[LOOP1_END]]
// CHECK: [[O_LOOP1_END]]
-// CHECK: call {{.+}} @__kmpc_cancel_barrier([[IDENT_T_TY]]* [[DEFAULT_LOC_BARRIER:[@%].+]], i32 [[GTID]])
+// CHECK: call {{.+}} @__kmpc_barrier([[IDENT_T_TY]]* [[DEFAULT_LOC_BARRIER:[@%].+]], i32 [[GTID]])
// CHECK: ret void
}
// CHECK: [[O_LOOP1_END]]
// CHECK: [[GTID_REF:%.+]] = load i32*, i32** [[GTID_REF_ADDR]],
// CHECK: [[GTID:%.+]] = load i32, i32* [[GTID_REF]],
-// CHECK: call {{.+}} @__kmpc_cancel_barrier([[IDENT_T_TY]]* [[DEFAULT_LOC_BARRIER:[@%].+]], i32 [[GTID]])
+// CHECK: call {{.+}} @__kmpc_barrier([[IDENT_T_TY]]* [[DEFAULT_LOC_BARRIER:[@%].+]], i32 [[GTID]])
// CHECK: ret void
}
// CHECK: [[O_LOOP1_END]]
// CHECK: [[GTID_REF:%.+]] = load i32*, i32** [[GTID_REF_ADDR]],
// CHECK: [[GTID:%.+]] = load i32, i32* [[GTID_REF]],
-// CHECK: call {{.+}} @__kmpc_cancel_barrier([[IDENT_T_TY]]* [[DEFAULT_LOC_BARRIER:[@%].+]], i32 [[GTID]])
+// CHECK: call {{.+}} @__kmpc_barrier([[IDENT_T_TY]]* [[DEFAULT_LOC_BARRIER:[@%].+]], i32 [[GTID]])
// CHECK: ret void
}
// TERM_DEBUG: unwind label %[[TERM_LPAD:.+]],
// TERM_DEBUG-NOT: __kmpc_global_thread_num
// TERM_DEBUG: call void @__kmpc_for_static_fini({{.+}}), !dbg [[DBG_LOC_END:![0-9]+]]
- // TERM_DEBUG: call {{.+}} @__kmpc_cancel_barrier({{.+}}), !dbg [[DBG_LOC_CANCEL:![0-9]+]]
+ // TERM_DEBUG: call {{.+}} @__kmpc_barrier({{.+}}), !dbg [[DBG_LOC_CANCEL:![0-9]+]]
// TERM_DEBUG: [[TERM_LPAD]]
// TERM_DEBUG: call void @__clang_call_terminate
// TERM_DEBUG: unreachable
// CLEANUP-NOT: __kmpc_global_thread_num
// CLEANUP: call void @__kmpc_for_static_init_4u({{.+}})
// CLEANUP: call void @__kmpc_for_static_fini({{.+}})
- // CLEANUP: call {{.+}} @__kmpc_cancel_barrier({{.+}})
+ // CLEANUP: call {{.+}} @__kmpc_barrier({{.+}})
for (unsigned i = 131071; i <= 2147483647; i += 127)
a[i] += foo();
}
}
// CHECK: [[SIMPLE_LOOP1_END]]
// CHECK: call void @__kmpc_for_static_fini(%ident_t* {{.+}}, i32 %{{.+}})
-// CHECK: call i32 @__kmpc_cancel_barrier(%ident_t* {{.+}}, i32 %{{.+}})
+// CHECK: call void @__kmpc_barrier(%ident_t* {{.+}}, i32 %{{.+}})
long long k = get_val();
// CHECK: [[LIN0_2:%.+]] = load i64, i64* [[LIN0]]
// CHECK-NEXT: [[LIN_ADD2:%.+]] = add nsw i64 [[LIN0_2]], 27
// CHECK-NEXT: store i64 [[LIN_ADD2]], i64* %{{.+}}
-// CHECK: call i32 @__kmpc_cancel_barrier(%ident_t* {{.+}}, i32 %{{.+}})
+// CHECK: call void @__kmpc_barrier(%ident_t* {{.+}}, i32 %{{.+}})
int lin = 12;
#pragma omp parallel for simd linear(lin : get_val()), linear(g_ptr)
// CHECK: [[GLIN_VAR:%.+]] = load double**, double*** %
// CHECK: [[GLINSTART:.+]] = load double*, double** [[GLIN_START]]
// CHECK: store double* {{.*}}[[GLIN_VAR]]
-// CHECK: call i32 @__kmpc_cancel_barrier(%ident_t* {{.+}}, i32 %{{.+}})
+// CHECK: call void @__kmpc_barrier(%ident_t* {{.+}}, i32 %{{.+}})
#pragma omp parallel for simd
// CHECK: call void @__kmpc_for_static_init_4(%ident_t* {{[^,]+}}, i32 %{{[^,]+}}, i32 34, i32* %{{[^,]+}}, i32* [[LB:%[^,]+]], i32* [[UB:%[^,]+]], i32* [[STRIDE:%[^,]+]], i32 1, i32 1)
}
// CHECK: [[SIMPLE_LOOP4_END]]
// CHECK: call void @__kmpc_for_static_fini(%ident_t* {{.+}}, i32 %{{.+}})
-// CHECK: call i32 @__kmpc_cancel_barrier(%ident_t* {{.+}}, i32 %{{.+}})
+// CHECK: call void @__kmpc_barrier(%ident_t* {{.+}}, i32 %{{.+}})
#pragma omp parallel for simd
// CHECK: call void @__kmpc_for_static_init_4(%ident_t* {{[^,]+}}, i32 %{{[^,]+}}, i32 34, i32* %{{[^,]+}}, i32* [[LB:%[^,]+]], i32* [[UB:%[^,]+]], i32* [[STRIDE:%[^,]+]], i32 1, i32 1)
}
// CHECK: [[SIMPLE_LOOP5_END]]
// CHECK: call void @__kmpc_for_static_fini(%ident_t* {{.+}}, i32 %{{.+}})
-// CHECK: call i32 @__kmpc_cancel_barrier(%ident_t* {{.+}}, i32 %{{.+}})
+// CHECK: call void @__kmpc_barrier(%ident_t* {{.+}}, i32 %{{.+}})
// CHECK-NOT: mul i32 %{{.+}}, 10
#pragma omp parallel for simd
// CHECK: [[A_PRIV_VAL:%.+]] = load i32, i32* [[A_PRIV]],
// CHECK-NEXT: store i32 [[A_PRIV_VAL]], i32* %{{.+}},
// CHECK-NEXT: br label
-// CHECK: call i32 @__kmpc_cancel_barrier(%ident_t* {{.+}}, i32 %{{.+}})
+// CHECK: call void @__kmpc_barrier(%ident_t* {{.+}}, i32 %{{.+}})
}
int R;
{
// CHECK: [[RED:%.+]] = mul nsw i32 %{{.+}}, [[R_PRIV_VAL]]
// CHECK-NEXT: store i32 [[RED]], i32* %{{.+}},
// CHECK-NEXT: call void @__kmpc_end_reduce_nowait(
-// CHECK: call i32 @__kmpc_cancel_barrier(%ident_t* {{.+}}, i32 %{{.+}})
+// CHECK: call void @__kmpc_barrier(%ident_t* {{.+}}, i32 %{{.+}})
}
}
}
// CHECK: [[IT_END]]
// CHECK: call void @__kmpc_for_static_fini(%ident_t* {{.+}}, i32 %{{.+}})
-// CHECK: call i32 @__kmpc_cancel_barrier(%ident_t* {{.+}}, i32 %{{.+}})
+// CHECK: call void @__kmpc_barrier(%ident_t* {{.+}}, i32 %{{.+}})
// CHECK: ret void
}
// CHECK: store i32 3, i32* [[I:%[^,]+]]
// CHECK: store i32 5, i32* [[I:%[^,]+]]
// CHECK: store i16 9, i16* [[I:%[^,]+]]
-// CHECK: call i32 @__kmpc_cancel_barrier(%ident_t* {{.+}}, i32 %{{.+}})
+// CHECK: call void @__kmpc_barrier(%ident_t* {{.+}}, i32 %{{.+}})
// CHECK: ret void
}
// CHECK-NEXT: br label {{%.+}}
// CHECK: [[T1_END]]
// CHECK: call void @__kmpc_for_static_fini(%ident_t* {{.+}}, i32 %{{.+}})
-// CHECK: call i32 @__kmpc_cancel_barrier(%ident_t* {{.+}}, i32 %{{.+}})
+// CHECK: call void @__kmpc_barrier(%ident_t* {{.+}}, i32 %{{.+}})
// CHECK: ret void
//
// TERM_DEBUG-LABEL: bar
// CHECK: [[INNER_LOOP_END]]
}
// CHECK: call void @__kmpc_for_static_fini(%{{.+}}* @{{.+}}, i32 [[GTID]])
-// CHECK: call i32 @__kmpc_cancel_barrier(%{{.+}}* [[IMPLICIT_BARRIER_LOC]],
+// CHECK: call void @__kmpc_barrier(%{{.+}}* [[IMPLICIT_BARRIER_LOC]],
return tmain<int>();
}
// CHECK: call void @__kmpc_end_single(
// CHECK-NEXT: br label %[[END]]
// CHECK: [[END]]
-// CHECK-NEXT: call i32 @__kmpc_cancel_barrier(%{{.+}}* [[IMPLICIT_BARRIER_LOC]],
+// CHECK-NEXT: call void @__kmpc_barrier(%{{.+}}* [[IMPLICIT_BARRIER_LOC]],
// CHECK-NEXT: ret
// CHECK: [[TERM_LPAD]]
// CHECK: call void @__clang_call_terminate(i8*
// CHECK: [[INNER_LOOP_END]]
}
// CHECK: call void @__kmpc_for_static_fini(%{{.+}}* @{{.+}}, i32 [[GTID]])
-// CHECK: call i32 @__kmpc_cancel_barrier(%{{.+}}* [[IMPLICIT_BARRIER_SECTIONS_LOC]],
+// CHECK: call void @__kmpc_barrier(%{{.+}}* [[IMPLICIT_BARRIER_SECTIONS_LOC]],
#pragma omp sections nowait
{
foo();
// CHECK-NEXT: br label %[[END]]
// CHECK: [[END]]
// CHECK-NEXT: call i32 @__kmpc_cancel_barrier(%{{.+}}* [[IMPLICIT_BARRIER_SINGLE_LOC]],
-// CHECK-NEXT: call i32 @__kmpc_cancel_barrier(
-// CHECK-NEXT: ret
+// CHECK: call i32 @__kmpc_cancel_barrier(
+// CHECK: ret
// CHECK: [[TERM_LPAD]]
// CHECK: call void @__clang_call_terminate(i8*
// CHECK-NEXT: unreachable
// LAMBDA: [[G_PRIVATE_ADDR:%.+]] = alloca i{{[0-9]+}},
// LAMBDA: [[G_VAL:%.+]] = load volatile i{{[0-9]+}}, i{{[0-9]+}}* [[G]]
// LAMBDA: store i{{[0-9]+}} [[G_VAL]], i{{[0-9]+}}* [[G_PRIVATE_ADDR]]
- // LAMBDA: call i32 @__kmpc_cancel_barrier(
+ // LAMBDA: call void @__kmpc_barrier(
g = 1;
// LAMBDA: call void @__kmpc_for_static_init_4(
// LAMBDA: store volatile i{{[0-9]+}} 1, i{{[0-9]+}}* [[G_PRIVATE_ADDR]],
// BLOCKS: [[G_PRIVATE_ADDR:%.+]] = alloca i{{[0-9]+}},
// BLOCKS: [[G_VAL:%.+]] = load volatile i{{[0-9]+}}, i{{[0-9]+}}* [[G]]
// BLOCKS: store i{{[0-9]+}} [[G_VAL]], i{{[0-9]+}}* [[G_PRIVATE_ADDR]]
- // BLOCKS: call i32 @__kmpc_cancel_barrier(
+ // BLOCKS: call void @__kmpc_barrier(
g = 1;
// BLOCKS: call void @__kmpc_for_static_init_4(
// BLOCKS: store volatile i{{[0-9]+}} 1, i{{[0-9]+}}* [[G_PRIVATE_ADDR]],
// CHECK-DAG: call {{.*}} [[S_FLOAT_TY_DESTR]]([[S_FLOAT_TY]]*
// CHECK: call void @__kmpc_end_single(
-// CHECK: call i32 @__kmpc_cancel_barrier(%{{.+}}* [[IMPLICIT_BARRIER_LOC]], i{{[0-9]+}} [[GTID]])
+// CHECK: call void @__kmpc_barrier(%{{.+}}* [[IMPLICIT_BARRIER_LOC]], i{{[0-9]+}} [[GTID]])
// CHECK: = call {{.*}}i{{.+}} [[TMAIN_INT:@.+]]()
// Synchronization for initialization.
// CHECK: [[GTID_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[GTID_ADDR_ADDR]]
// CHECK: [[GTID:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[GTID_REF]]
-// CHECK: call i32 @__kmpc_cancel_barrier(%{{.+}}* [[IMPLICIT_BARRIER_LOC]], i{{[0-9]+}} [[GTID]])
+// CHECK: call void @__kmpc_barrier(%{{.+}}* [[IMPLICIT_BARRIER_LOC]], i{{[0-9]+}} [[GTID]])
// CHECK: call void @__kmpc_for_static_init_4(
// CHECK: call void @__kmpc_for_static_fini(
// CHECK-NEXT: call void @__kmpc_end_single([[IDENT_T_TY]]* [[DEFAULT_LOC]], i32 [[GTID]])
// CHECK-NEXT: br label {{%?}}[[EXIT]]
// CHECK: [[EXIT]]
-// CHECK: call{{.*}} @__kmpc_cancel_barrier([[IDENT_T_TY]]* [[IMPLICIT_BARRIER_SINGLE_LOC]], i32 [[GTID]])
+// CHECK: call{{.*}} @__kmpc_barrier([[IDENT_T_TY]]* [[IMPLICIT_BARRIER_SINGLE_LOC]], i32 [[GTID]])
#pragma omp single
a = 2;
// CHECK: store i32 0, i32* [[DID_IT]]
// CHECK-DAG: call {{.*}} [[S_FLOAT_TY_DESTR]]([[S_FLOAT_TY]]*
// CHECK: call void @__kmpc_end_single(
-// CHECK: call i32 @__kmpc_cancel_barrier(%{{.+}}* [[IMPLICIT_BARRIER_LOC]], i{{[0-9]+}} [[GTID]])
+// CHECK: call void @__kmpc_barrier(%{{.+}}* [[IMPLICIT_BARRIER_LOC]], i{{[0-9]+}} [[GTID]])
// CHECK: = call {{.*}}i{{.+}} [[TMAIN_INT:@.+]]()