diff --git a/clang/include/clang/Basic/DiagnosticDriverKinds.td b/clang/include/clang/Basic/DiagnosticDriverKinds.td --- a/clang/include/clang/Basic/DiagnosticDriverKinds.td +++ b/clang/include/clang/Basic/DiagnosticDriverKinds.td @@ -281,6 +281,7 @@ def err_drv_no_neon_modifier : Error<"[no]neon is not accepted as modifier, please use [no]simd instead">; def err_drv_invalid_omp_target : Error<"OpenMP target is invalid: '%0'">; def err_drv_debug_no_new_runtime : Error<"OpenMP target device debugging enabled with incompatible runtime">; +def err_drv_stack_size_no_new_runtime : Error<"OpenMP target stack size enabled with incompatible runtime">; def err_drv_incompatible_omp_arch : Error<"OpenMP target architecture '%0' pointer size is incompatible with host '%1'">; def err_drv_omp_host_ir_file_not_found : Error< "provided host compiler IR file '%0' is required to generate code for OpenMP " diff --git a/clang/include/clang/Basic/LangOptions.def b/clang/include/clang/Basic/LangOptions.def --- a/clang/include/clang/Basic/LangOptions.def +++ b/clang/include/clang/Basic/LangOptions.def @@ -243,6 +243,7 @@ LANGOPT(OpenMPCUDAReductionBufNum , 32, 1024, "Number of the reduction records in the intermediate reduction buffer used for the teams reductions.") LANGOPT(OpenMPTargetNewRuntime , 1, 0, "Use the new bitcode library for OpenMP offloading") LANGOPT(OpenMPTargetDebug , 32, 0, "Enable debugging in the OpenMP offloading device RTL") +LANGOPT(OpenMPTargetStackSize , 32, 2048, "Amount of shared memory to use for the data sharing stack") LANGOPT(OpenMPOptimisticCollapse , 1, 0, "Use at most 32 bits to represent the collapsed loop nest counter.") LANGOPT(RenderScript , 1, 0, "RenderScript") diff --git a/clang/include/clang/Driver/Options.td b/clang/include/clang/Driver/Options.td --- a/clang/include/clang/Driver/Options.td +++ b/clang/include/clang/Driver/Options.td @@ -2427,6 +2427,8 @@ HelpText<"Enable debugging in the OpenMP offloading device RTL">; def fno_openmp_target_debug : Flag<["-"], "fno-openmp-target-debug">, Group, Flags<[NoArgumentUnused]>; def fopenmp_target_debug_EQ : Joined<["-"], "fopenmp-target-debug=">, Group, Flags<[CC1Option, NoArgumentUnused, HelpHidden]>; +def fopenmp_target_stack_size_EQ : Joined<["-"], "fopenmp-target-stack-size=">, Group, Flags<[CC1Option, NoArgumentUnused, HelpHidden]>, + HelpText<"Amount of shared memory used for the data sharing stack in the OpenMP offloading device RTL.">; defm openmp_target_new_runtime: BoolFOption<"openmp-target-new-runtime", LangOpts<"OpenMPTargetNewRuntime">, DefaultFalse, PosFlag, diff --git a/clang/lib/CodeGen/CGOpenMPRuntimeGPU.cpp b/clang/lib/CodeGen/CGOpenMPRuntimeGPU.cpp --- a/clang/lib/CodeGen/CGOpenMPRuntimeGPU.cpp +++ b/clang/lib/CodeGen/CGOpenMPRuntimeGPU.cpp @@ -1050,7 +1050,12 @@ void CGOpenMPRuntimeGPU::emitKernelInit(CodeGenFunction &CGF, EntryFunctionState &EST, bool IsSPMD) { CGBuilderTy &Bld = CGF.Builder; - Bld.restoreIP(OMPBuilder.createTargetInit(Bld, IsSPMD, requiresFullRuntime())); + uint32_t StackSize = (CGF.CGM.getLangOpts().OpenMPTargetNewRuntime) + ? CGF.CGM.getLangOpts().OpenMPTargetStackSize + : 0; + + Bld.restoreIP(OMPBuilder.createTargetInit(Bld, IsSPMD, requiresFullRuntime(), + StackSize)); IsInTargetMasterThreadRegion = IsSPMD; if (!IsSPMD) emitGenericVarsProlog(CGF, EST.Loc); @@ -1121,6 +1126,21 @@ CGM.addCompilerUsedGlobal(GVMode); } +// Create a unique global variable to indicate the amount of dynamic shared +// memory to be used for the data sharing stack. This is to be read from the +// device by the plugin runtime library to allocate the appropriate amount of +// shared memory for the kernel. +static void setPropertyStackSize(CodeGenModule &CGM, StringRef Name, + bool Mode) { + uint32_t StackSize = CGM.getLangOpts().OpenMPTargetStackSize; + auto *GVSstackSize = new llvm::GlobalVariable( + CGM.getModule(), CGM.Int32Ty, /*isConstant=*/true, + llvm::GlobalValue::WeakAnyLinkage, + llvm::ConstantInt::get(CGM.Int32Ty, StackSize), + Twine(Name, "_stack_size")); + CGM.addCompilerUsedGlobal(GVSstackSize); +} + void CGOpenMPRuntimeGPU::createOffloadEntry(llvm::Constant *ID, llvm::Constant *Addr, uint64_t Size, int32_t, @@ -1161,6 +1181,9 @@ CodeGen); setPropertyExecutionMode(CGM, OutlinedFn->getName(), Mode); + + if (CGM.getLangOpts().OpenMPTargetNewRuntime) + setPropertyStackSize(CGM, OutlinedFn->getName(), Mode); } namespace { diff --git a/clang/lib/Driver/ToolChains/Clang.cpp b/clang/lib/Driver/ToolChains/Clang.cpp --- a/clang/lib/Driver/ToolChains/Clang.cpp +++ b/clang/lib/Driver/ToolChains/Clang.cpp @@ -5794,6 +5794,9 @@ options::OPT_fno_openmp_target_debug, /*Default=*/false)) CmdArgs.push_back("-fopenmp-target-debug"); + // When in OpenMP offloading mode, set the data sharing stack size. + Args.AddAllArgs(CmdArgs, options::OPT_fopenmp_target_stack_size_EQ); + // When in OpenMP offloading mode with NVPTX target, check if full runtime // is required. if (Args.hasFlag(options::OPT_fopenmp_cuda_force_full_runtime, diff --git a/clang/lib/Frontend/CompilerInvocation.cpp b/clang/lib/Frontend/CompilerInvocation.cpp --- a/clang/lib/Frontend/CompilerInvocation.cpp +++ b/clang/lib/Frontend/CompilerInvocation.cpp @@ -3485,6 +3485,10 @@ GenerateArg(Args, OPT_fopenmp_target_debug_EQ, Twine(Opts.OpenMPTargetDebug), SA); + if (Opts.OpenMPTargetStackSize != 2048) + GenerateArg(Args, OPT_fopenmp_target_stack_size_EQ, + Twine(Opts.OpenMPTargetStackSize), SA); + if (Opts.OpenMPCUDANumSMs != 0) GenerateArg(Args, OPT_fopenmp_cuda_number_of_sm_EQ, Twine(Opts.OpenMPCUDANumSMs), SA); @@ -3923,6 +3927,17 @@ } } + // Set the amount of shared memory to use for the data sharing stack in the + // new offloading device RTL + if (Opts.OpenMPIsDevice && Args.hasArg(OPT_fopenmp_target_stack_size_EQ)) { + if (Opts.OpenMPTargetNewRuntime) + Opts.OpenMPTargetStackSize = + getLastArgIntValue(Args, options::OPT_fopenmp_target_stack_size_EQ, + Opts.OpenMPTargetStackSize, Diags); + else + Diags.Report(diag::err_drv_stack_size_no_new_runtime); + } + // Get the OpenMP target triples if any. if (Arg *A = Args.getLastArg(options::OPT_fopenmp_targets_EQ)) { enum ArchPtrSize { Arch16Bit, Arch32Bit, Arch64Bit }; diff --git a/clang/test/OpenMP/amdgcn_target_codegen.cpp b/clang/test/OpenMP/amdgcn_target_codegen.cpp --- a/clang/test/OpenMP/amdgcn_target_codegen.cpp +++ b/clang/test/OpenMP/amdgcn_target_codegen.cpp @@ -13,7 +13,7 @@ int arr[N]; -// CHECK: call i32 @__kmpc_target_init(%struct.ident_t* addrspacecast (%struct.ident_t addrspace(1)* @1 to %struct.ident_t*), i8 1, i1 true, i1 true) +// CHECK: call i32 @__kmpc_target_init(%struct.ident_t* addrspacecast (%struct.ident_t addrspace(1)* @1 to %struct.ident_t*), i8 1, i1 true, i1 true, i32 0) #pragma omp target for (int i = 0; i < N; i++) { arr[i] = 1; @@ -27,7 +27,7 @@ int arr[N]; -// CHECK: call i32 @__kmpc_target_init(%struct.ident_t* addrspacecast (%struct.ident_t addrspace(1)* @1 to %struct.ident_t*), i8 2, i1 false, i1 false) +// CHECK: call i32 @__kmpc_target_init(%struct.ident_t* addrspacecast (%struct.ident_t addrspace(1)* @1 to %struct.ident_t*), i8 2, i1 false, i1 false, i32 0) #pragma omp target simd for (int i = 0; i < N; i++) { arr[i] = 1; diff --git a/clang/test/OpenMP/declare_target_codegen_globalization.cpp b/clang/test/OpenMP/declare_target_codegen_globalization.cpp --- a/clang/test/OpenMP/declare_target_codegen_globalization.cpp +++ b/clang/test/OpenMP/declare_target_codegen_globalization.cpp @@ -31,7 +31,7 @@ // CHECK1-NEXT: [[CAPTURED_VARS_ADDRS:%.*]] = alloca [1 x i8*], align 8 // CHECK1-NEXT: store i32* [[A]], i32** [[A_ADDR]], align 8 // CHECK1-NEXT: [[TMP0:%.*]] = load i32*, i32** [[A_ADDR]], align 8 -// CHECK1-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 true) +// CHECK1-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 true, i32 0) // CHECK1-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK1-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK1: user_code.entry: @@ -58,8 +58,8 @@ // CHECK1-NEXT: store i32* [[DOTBOUND_TID_]], i32** [[DOTBOUND_TID__ADDR]], align 8 // CHECK1-NEXT: store i32* [[A]], i32** [[A_ADDR]], align 8 // CHECK1-NEXT: [[TMP0:%.*]] = load i32*, i32** [[A_ADDR]], align 8 -// CHECK1-NEXT: [[CALL:%.*]] = call i32 @_Z3fooRi(i32* nonnull align 4 dereferenceable(4) [[B]]) #[[ATTR4:[0-9]+]] -// CHECK1-NEXT: [[CALL1:%.*]] = call i32 @_Z3barv() #[[ATTR4]] +// CHECK1-NEXT: [[CALL:%.*]] = call i32 @_Z3fooRi(i32* nonnull align 4 dereferenceable(4) [[B]]) #[[ATTR6:[0-9]+]] +// CHECK1-NEXT: [[CALL1:%.*]] = call i32 @_Z3barv() #[[ATTR6]] // CHECK1-NEXT: [[ADD:%.*]] = add nsw i32 [[CALL]], [[CALL1]] // CHECK1-NEXT: store i32 [[ADD]], i32* [[TMP0]], align 4 // CHECK1-NEXT: ret void @@ -80,7 +80,7 @@ // CHECK1-NEXT: entry: // CHECK1-NEXT: [[A:%.*]] = call i8* @__kmpc_alloc_shared(i64 4) // CHECK1-NEXT: [[A_ON_STACK:%.*]] = bitcast i8* [[A]] to i32* -// CHECK1-NEXT: [[CALL:%.*]] = call i32 @_Z3fooRi(i32* nonnull align 4 dereferenceable(4) [[A_ON_STACK]]) #[[ATTR4]] +// CHECK1-NEXT: [[CALL:%.*]] = call i32 @_Z3fooRi(i32* nonnull align 4 dereferenceable(4) [[A_ON_STACK]]) #[[ATTR6]] // CHECK1-NEXT: call void @__kmpc_free_shared(i8* [[A]], i64 4) // CHECK1-NEXT: ret i32 [[CALL]] // diff --git a/clang/test/OpenMP/nvptx_SPMD_codegen.cpp b/clang/test/OpenMP/nvptx_SPMD_codegen.cpp --- a/clang/test/OpenMP/nvptx_SPMD_codegen.cpp +++ b/clang/test/OpenMP/nvptx_SPMD_codegen.cpp @@ -21,28 +21,28 @@ // CHECK-NOT: @__omp_offloading_{{.+}}_exec_mode = weak constant i8 1 void foo() { -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false, i32 0) // CHECK-DAG: [[DISTR_LIGHT]] // CHECK-DAG: [[FOR_LIGHT]] // CHECK-DAG: [[LIGHT]] -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false, i32 0) // CHECK-DAG: [[DISTR_LIGHT]] // CHECK-DAG: [[FOR_LIGHT]] // CHECK-DAG: [[LIGHT]] -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false, i32 0) // CHECK-DAG: [[DISTR_LIGHT]] // CHECK-DAG: [[FOR_LIGHT]] // CHECK-DAG: [[LIGHT]] -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) // CHECK-DAG: [[DISTR_FULL]] // CHECK-DAG: [[FULL]] -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) // CHECK-DAG: [[DISTR_FULL]] // CHECK-DAG: [[FULL]] -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) // CHECK-DAG: [[DISTR_FULL]] // CHECK-DAG: [[FULL]] -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) // CHECK-DAG: [[DISTR_FULL]] // CHECK-DAG: [[FULL]] #pragma omp target teams distribute parallel for simd if(a) @@ -67,28 +67,28 @@ for (int i = 0; i < 10; ++i) ; int a; -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false, i32 0) // CHECK-DAG: [[DISTR_LIGHT]] // CHECK-DAG: [[FOR_LIGHT]] // CHECK-DAG: [[LIGHT]] -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false, i32 0) // CHECK-DAG: [[DISTR_LIGHT]] // CHECK-DAG: [[FOR_LIGHT]] // CHECK-DAG: [[LIGHT]] -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false, i32 0) // CHECK-DAG: [[DISTR_LIGHT]] // CHECK-DAG: [[FOR_LIGHT]] // CHECK-DAG: [[LIGHT]] -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) // CHECK-DAG: [[DISTR_FULL]] // CHECK-DAG: [[FULL]] -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) // CHECK-DAG: [[DISTR_FULL]] // CHECK-DAG: [[FULL]] -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) // CHECK-DAG: [[DISTR_FULL]] // CHECK-DAG: [[FULL]] -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) // CHECK-DAG: [[DISTR_FULL]] // CHECK-DAG: [[FULL]] #pragma omp target teams distribute parallel for lastprivate(a) @@ -112,25 +112,25 @@ #pragma omp target teams distribute parallel for schedule(guided) for (int i = 0; i < 10; ++i) ; -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false, i32 0) // CHECK-DAG: [[DISTR_LIGHT]] // CHECK-DAG: [[FOR_LIGHT]] // CHECK-DAG: [[LIGHT]] -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false, i32 0) // CHECK-DAG: [[DISTR_LIGHT]] // CHECK-DAG: [[FOR_LIGHT]] // CHECK-DAG: [[LIGHT]] // CHECK: call i32 @__kmpc_target_init( -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) // CHECK-DAG: [[DISTR_FULL]] // CHECK-DAG: [[FULL]] -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) // CHECK-DAG: [[DISTR_FULL]] // CHECK-DAG: [[FULL]] -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) // CHECK-DAG: [[DISTR_FULL]] // CHECK-DAG: [[FULL]] -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) // CHECK-DAG: [[DISTR_FULL]] // CHECK-DAG: [[FULL]] #pragma omp target teams @@ -172,28 +172,28 @@ #pragma omp distribute parallel for simd schedule(guided) for (int i = 0; i < 10; ++i) ; -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false, i32 0) // CHECK-DAG: [[DISTR_LIGHT]] // CHECK-DAG: [[FOR_LIGHT]] // CHECK-DAG: [[LIGHT]] -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false, i32 0) // CHECK-DAG: [[DISTR_LIGHT]] // CHECK-DAG: [[FOR_LIGHT]] // CHECK-DAG: [[LIGHT]] -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false, i32 0) // CHECK-DAG: [[DISTR_LIGHT]] // CHECK-DAG: [[FOR_LIGHT]] // CHECK-DAG: [[LIGHT]] -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) // CHECK-DAG: [[DISTR_FULL]] // CHECK-DAG: [[FULL]] -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) // CHECK-DAG: [[DISTR_FULL]] // CHECK-DAG: [[FULL]] -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) // CHECK-DAG: [[DISTR_FULL]] // CHECK-DAG: [[FULL]] -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) // CHECK-DAG: [[DISTR_FULL]] // CHECK-DAG: [[FULL]] #pragma omp target teams @@ -224,28 +224,28 @@ #pragma omp distribute parallel for schedule(guided) for (int i = 0; i < 10; ++i) ; -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false, i32 0) // CHECK-DAG: [[DISTR_LIGHT]] // CHECK-DAG: [[FOR_LIGHT]] // CHECK-DAG: [[LIGHT]] -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false, i32 0) // CHECK-DAG: [[DISTR_LIGHT]] // CHECK-DAG: [[FOR_LIGHT]] // CHECK-DAG: [[LIGHT]] -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false, i32 0) // CHECK-DAG: [[DISTR_LIGHT]] // CHECK-DAG: [[FOR_LIGHT]] // CHECK-DAG: [[LIGHT]] -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) // CHECK-DAG: [[DISTR_FULL]] // CHECK-DAG: [[FULL]] -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) // CHECK-DAG: [[DISTR_FULL]] // CHECK-DAG: [[FULL]] -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) // CHECK-DAG: [[DISTR_FULL]] // CHECK-DAG: [[FULL]] -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) // CHECK-DAG: [[DISTR_FULL]] // CHECK-DAG: [[FULL]] #pragma omp target @@ -283,22 +283,22 @@ #pragma omp distribute parallel for schedule(guided) for (int i = 0; i < 10; ++i) ; -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false, i32 0) // CHECK-DAG: [[FOR_LIGHT]] // CHECK-DAG: [[LIGHT]] -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false, i32 0) // CHECK-DAG: [[FOR_LIGHT]] // CHECK-DAG: [[LIGHT]] -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false, i32 0) // CHECK-DAG: [[FOR_LIGHT]] // CHECK-DAG: [[LIGHT]] -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) // CHECK-DAG: [[FULL]] -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) // CHECK-DAG: [[FULL]] -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) // CHECK-DAG: [[FULL]] -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) // CHECK-DAG: [[FULL]] #pragma omp target parallel for if(a) for (int i = 0; i < 10; ++i) @@ -321,28 +321,28 @@ #pragma omp target parallel for schedule(guided) for (int i = 0; i < 10; ++i) ; -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false, i32 0) // CHECK-DAG: [[FOR_LIGHT]] // CHECK-DAG: [[LIGHT]] // CHECK-DAG: [[BAR_LIGHT]] -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false, i32 0) // CHECK-DAG: [[FOR_LIGHT]] // CHECK-DAG: [[LIGHT]] // CHECK-DAG: [[BAR_LIGHT]] -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false, i32 0) // CHECK-DAG: [[FOR_LIGHT]] // CHECK-DAG: [[LIGHT]] // CHECK-DAG: [[BAR_LIGHT]] -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) // CHECK-DAG: [[FULL]] // CHECK-DAG: [[BAR_FULL]] -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) // CHECK-DAG: [[FULL]] // CHECK-DAG: [[BAR_FULL]] -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) // CHECK-DAG: [[FULL]] // CHECK-DAG: [[BAR_FULL]] -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) // CHECK-DAG: [[FULL]] // CHECK-DAG: [[BAR_FULL]] #pragma omp target parallel if(a) @@ -373,27 +373,27 @@ #pragma omp for simd schedule(guided) for (int i = 0; i < 10; ++i) ; -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) // CHECK-DAG: [[FULL]] // CHECK-DAG: [[BAR_FULL]] -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false, i32 0) // CHECK-DAG: [[FOR_LIGHT]] // CHECK-DAG: [[LIGHT]] // CHECK-DAG: [[BAR_LIGHT]] -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false, i32 0) // CHECK-DAG: [[FOR_LIGHT]] // CHECK-DAG: [[LIGHT]] // CHECK-DAG: [[BAR_LIGHT]] -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) // CHECK-DAG: [[FULL]] // CHECK-DAG: [[BAR_FULL]] -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) // CHECK-DAG: [[FULL]] // CHECK-DAG: [[BAR_FULL]] -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) // CHECK-DAG: [[FULL]] // CHECK-DAG: [[BAR_FULL]] -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) // CHECK-DAG: [[FULL]] // CHECK-DAG: [[BAR_FULL]] #pragma omp target @@ -431,22 +431,22 @@ #pragma omp for simd schedule(guided) for (int i = 0; i < 10; ++i) ; -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false, i32 0) // CHECK-DAG: [[FOR_LIGHT]] // CHECK-DAG: [[LIGHT]] -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false, i32 0) // CHECK-DAG: [[FOR_LIGHT]] // CHECK-DAG: [[LIGHT]] -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false, i32 0) // CHECK-DAG: [[FOR_LIGHT]] // CHECK-DAG: [[LIGHT]] -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) // CHECK-DAG: [[FULL]] -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) // CHECK-DAG: [[FULL]] -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) // CHECK-DAG: [[FULL]] -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) // CHECK-DAG: [[FULL]] #pragma omp target #pragma omp parallel for diff --git a/clang/test/OpenMP/nvptx_data_sharing.cpp b/clang/test/OpenMP/nvptx_data_sharing.cpp --- a/clang/test/OpenMP/nvptx_data_sharing.cpp +++ b/clang/test/OpenMP/nvptx_data_sharing.cpp @@ -393,7 +393,7 @@ // CHECK-NEXT: [[CAPTURED_VARS_ADDRS:%.*]] = alloca [1 x i8*], align 8 // CHECK-NEXT: [[C:%.*]] = alloca i32, align 4 // CHECK-NEXT: [[CAPTURED_VARS_ADDRS1:%.*]] = alloca [2 x i8*], align 8 -// CHECK-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true) +// CHECK-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true, i32 0) // CHECK-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 // CHECK-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK: user_code.entry: @@ -455,7 +455,7 @@ // CHECK-NEXT: [[TMP3:%.*]] = getelementptr inbounds i8*, i8** [[TMP2]], i64 0 // CHECK-NEXT: [[TMP4:%.*]] = bitcast i8** [[TMP3]] to i32** // CHECK-NEXT: [[TMP5:%.*]] = load i32*, i32** [[TMP4]], align 8 -// CHECK-NEXT: call void @__omp_outlined__(i32* [[DOTADDR1]], i32* [[DOTZERO_ADDR]], i32* [[TMP5]]) #[[ATTR1:[0-9]+]] +// CHECK-NEXT: call void @__omp_outlined__(i32* [[DOTADDR1]], i32* [[DOTZERO_ADDR]], i32* [[TMP5]]) #[[ATTR3:[0-9]+]] // CHECK-NEXT: ret void // // @@ -499,6 +499,6 @@ // CHECK-NEXT: [[TMP6:%.*]] = getelementptr inbounds i8*, i8** [[TMP2]], i64 1 // CHECK-NEXT: [[TMP7:%.*]] = bitcast i8** [[TMP6]] to i32** // CHECK-NEXT: [[TMP8:%.*]] = load i32*, i32** [[TMP7]], align 8 -// CHECK-NEXT: call void @__omp_outlined__1(i32* [[DOTADDR1]], i32* [[DOTZERO_ADDR]], i32* [[TMP5]], i32* [[TMP8]]) #[[ATTR1]] +// CHECK-NEXT: call void @__omp_outlined__1(i32* [[DOTADDR1]], i32* [[DOTZERO_ADDR]], i32* [[TMP5]], i32* [[TMP8]]) #[[ATTR3]] // CHECK-NEXT: ret void // diff --git a/clang/test/OpenMP/nvptx_distribute_parallel_generic_mode_codegen.cpp b/clang/test/OpenMP/nvptx_distribute_parallel_generic_mode_codegen.cpp --- a/clang/test/OpenMP/nvptx_distribute_parallel_generic_mode_codegen.cpp +++ b/clang/test/OpenMP/nvptx_distribute_parallel_generic_mode_codegen.cpp @@ -2995,7 +2995,7 @@ // CHECK4-NEXT: [[TMP2:%.*]] = load i32*, i32** [[A_ADDR]], align 8 // CHECK4-NEXT: [[CONV:%.*]] = bitcast i64* [[ARGC_ADDR]] to i32* // CHECK4-NEXT: [[TMP3:%.*]] = load [10 x i32]*, [10 x i32]** [[D_ADDR]], align 8 -// CHECK4-NEXT: [[TMP4:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 false) +// CHECK4-NEXT: [[TMP4:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 false, i32 0) // CHECK4-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP4]], -1 // CHECK4-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK4: user_code.entry: @@ -3005,7 +3005,7 @@ // CHECK4-NEXT: store i32 [[TMP6]], i32* [[CONV1]], align 4 // CHECK4-NEXT: [[TMP7:%.*]] = load i64, i64* [[ARGC_CASTED]], align 8 // CHECK4-NEXT: store i32 [[TMP5]], i32* [[DOTTHREADID_TEMP_]], align 4 -// CHECK4-NEXT: call void @__omp_outlined__(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], [10 x i32]* [[TMP0]], [10 x i32]* [[TMP1]], i32* [[TMP2]], i64 [[TMP7]], [10 x i32]* [[TMP3]]) #[[ATTR1:[0-9]+]] +// CHECK4-NEXT: call void @__omp_outlined__(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], [10 x i32]* [[TMP0]], [10 x i32]* [[TMP1]], i32* [[TMP2]], i64 [[TMP7]], [10 x i32]* [[TMP3]]) #[[ATTR5:[0-9]+]] // CHECK4-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 false) // CHECK4-NEXT: ret void // CHECK4: worker.exit: @@ -3259,23 +3259,23 @@ // CHECK4-NEXT: [[MUL:%.*]] = mul nsw i32 [[TMP18]], 1 // CHECK4-NEXT: [[ADD:%.*]] = add nsw i32 0, [[MUL]] // CHECK4-NEXT: store i32 [[ADD]], i32* [[I6]], align 4 -// CHECK4-NEXT: [[CALL:%.*]] = call i32 @_Z3fooPi(i32* [[I6]]) #[[ATTR5:[0-9]+]] -// CHECK4-NEXT: [[CALL9:%.*]] = call i32 @_Z3fooPi(i32* [[TMP1]]) #[[ATTR5]] +// CHECK4-NEXT: [[CALL:%.*]] = call i32 @_Z3fooPi(i32* [[I6]]) #[[ATTR7:[0-9]+]] +// CHECK4-NEXT: [[CALL9:%.*]] = call i32 @_Z3fooPi(i32* [[TMP1]]) #[[ATTR7]] // CHECK4-NEXT: [[ADD10:%.*]] = add nsw i32 [[CALL]], [[CALL9]] // CHECK4-NEXT: [[TMP19:%.*]] = load i32, i32* [[I6]], align 4 // CHECK4-NEXT: [[IDXPROM:%.*]] = sext i32 [[TMP19]] to i64 // CHECK4-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [10 x i32], [10 x i32]* [[B4]], i64 0, i64 [[IDXPROM]] -// CHECK4-NEXT: [[CALL11:%.*]] = call i32 @_Z3fooPi(i32* [[ARRAYIDX]]) #[[ATTR5]] +// CHECK4-NEXT: [[CALL11:%.*]] = call i32 @_Z3fooPi(i32* [[ARRAYIDX]]) #[[ATTR7]] // CHECK4-NEXT: [[ADD12:%.*]] = add nsw i32 [[ADD10]], [[CALL11]] // CHECK4-NEXT: [[TMP20:%.*]] = load i32, i32* [[I6]], align 4 // CHECK4-NEXT: [[IDXPROM13:%.*]] = sext i32 [[TMP20]] to i64 // CHECK4-NEXT: [[ARRAYIDX14:%.*]] = getelementptr inbounds [10 x i32], [10 x i32]* [[C5]], i64 0, i64 [[IDXPROM13]] -// CHECK4-NEXT: [[CALL15:%.*]] = call i32 @_Z3fooPi(i32* [[ARRAYIDX14]]) #[[ATTR5]] +// CHECK4-NEXT: [[CALL15:%.*]] = call i32 @_Z3fooPi(i32* [[ARRAYIDX14]]) #[[ATTR7]] // CHECK4-NEXT: [[ADD16:%.*]] = add nsw i32 [[ADD12]], [[CALL15]] // CHECK4-NEXT: [[TMP21:%.*]] = load i32, i32* [[I6]], align 4 // CHECK4-NEXT: [[IDXPROM17:%.*]] = sext i32 [[TMP21]] to i64 // CHECK4-NEXT: [[ARRAYIDX18:%.*]] = getelementptr inbounds [10 x i32], [10 x i32]* [[TMP4]], i64 0, i64 [[IDXPROM17]] -// CHECK4-NEXT: [[CALL19:%.*]] = call i32 @_Z3fooPi(i32* [[ARRAYIDX18]]) #[[ATTR5]] +// CHECK4-NEXT: [[CALL19:%.*]] = call i32 @_Z3fooPi(i32* [[ARRAYIDX18]]) #[[ATTR7]] // CHECK4-NEXT: [[ADD20:%.*]] = add nsw i32 [[ADD16]], [[CALL19]] // CHECK4-NEXT: store i32 [[ADD20]], i32* [[TMP1]], align 4 // CHECK4-NEXT: br label [[OMP_BODY_CONTINUE:%.*]] @@ -3328,7 +3328,7 @@ // CHECK5-NEXT: [[TMP1:%.*]] = load [10 x i32]*, [10 x i32]** [[C_ADDR]], align 4 // CHECK5-NEXT: [[TMP2:%.*]] = load i32*, i32** [[A_ADDR]], align 4 // CHECK5-NEXT: [[TMP3:%.*]] = load [10 x i32]*, [10 x i32]** [[D_ADDR]], align 4 -// CHECK5-NEXT: [[TMP4:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 false) +// CHECK5-NEXT: [[TMP4:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 false, i32 0) // CHECK5-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP4]], -1 // CHECK5-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK5: user_code.entry: @@ -3337,7 +3337,7 @@ // CHECK5-NEXT: store i32 [[TMP6]], i32* [[ARGC_CASTED]], align 4 // CHECK5-NEXT: [[TMP7:%.*]] = load i32, i32* [[ARGC_CASTED]], align 4 // CHECK5-NEXT: store i32 [[TMP5]], i32* [[DOTTHREADID_TEMP_]], align 4 -// CHECK5-NEXT: call void @__omp_outlined__(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], [10 x i32]* [[TMP0]], [10 x i32]* [[TMP1]], i32* [[TMP2]], i32 [[TMP7]], [10 x i32]* [[TMP3]]) #[[ATTR1:[0-9]+]] +// CHECK5-NEXT: call void @__omp_outlined__(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], [10 x i32]* [[TMP0]], [10 x i32]* [[TMP1]], i32* [[TMP2]], i32 [[TMP7]], [10 x i32]* [[TMP3]]) #[[ATTR5:[0-9]+]] // CHECK5-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 false) // CHECK5-NEXT: ret void // CHECK5: worker.exit: @@ -3585,20 +3585,20 @@ // CHECK5-NEXT: [[MUL:%.*]] = mul nsw i32 [[TMP18]], 1 // CHECK5-NEXT: [[ADD:%.*]] = add nsw i32 0, [[MUL]] // CHECK5-NEXT: store i32 [[ADD]], i32* [[I5]], align 4 -// CHECK5-NEXT: [[CALL:%.*]] = call i32 @_Z3fooPi(i32* [[I5]]) #[[ATTR5:[0-9]+]] -// CHECK5-NEXT: [[CALL7:%.*]] = call i32 @_Z3fooPi(i32* [[TMP1]]) #[[ATTR5]] +// CHECK5-NEXT: [[CALL:%.*]] = call i32 @_Z3fooPi(i32* [[I5]]) #[[ATTR7:[0-9]+]] +// CHECK5-NEXT: [[CALL7:%.*]] = call i32 @_Z3fooPi(i32* [[TMP1]]) #[[ATTR7]] // CHECK5-NEXT: [[ADD8:%.*]] = add nsw i32 [[CALL]], [[CALL7]] // CHECK5-NEXT: [[TMP19:%.*]] = load i32, i32* [[I5]], align 4 // CHECK5-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [10 x i32], [10 x i32]* [[B3]], i32 0, i32 [[TMP19]] -// CHECK5-NEXT: [[CALL9:%.*]] = call i32 @_Z3fooPi(i32* [[ARRAYIDX]]) #[[ATTR5]] +// CHECK5-NEXT: [[CALL9:%.*]] = call i32 @_Z3fooPi(i32* [[ARRAYIDX]]) #[[ATTR7]] // CHECK5-NEXT: [[ADD10:%.*]] = add nsw i32 [[ADD8]], [[CALL9]] // CHECK5-NEXT: [[TMP20:%.*]] = load i32, i32* [[I5]], align 4 // CHECK5-NEXT: [[ARRAYIDX11:%.*]] = getelementptr inbounds [10 x i32], [10 x i32]* [[C4]], i32 0, i32 [[TMP20]] -// CHECK5-NEXT: [[CALL12:%.*]] = call i32 @_Z3fooPi(i32* [[ARRAYIDX11]]) #[[ATTR5]] +// CHECK5-NEXT: [[CALL12:%.*]] = call i32 @_Z3fooPi(i32* [[ARRAYIDX11]]) #[[ATTR7]] // CHECK5-NEXT: [[ADD13:%.*]] = add nsw i32 [[ADD10]], [[CALL12]] // CHECK5-NEXT: [[TMP21:%.*]] = load i32, i32* [[I5]], align 4 // CHECK5-NEXT: [[ARRAYIDX14:%.*]] = getelementptr inbounds [10 x i32], [10 x i32]* [[TMP4]], i32 0, i32 [[TMP21]] -// CHECK5-NEXT: [[CALL15:%.*]] = call i32 @_Z3fooPi(i32* [[ARRAYIDX14]]) #[[ATTR5]] +// CHECK5-NEXT: [[CALL15:%.*]] = call i32 @_Z3fooPi(i32* [[ARRAYIDX14]]) #[[ATTR7]] // CHECK5-NEXT: [[ADD16:%.*]] = add nsw i32 [[ADD13]], [[CALL15]] // CHECK5-NEXT: store i32 [[ADD16]], i32* [[TMP1]], align 4 // CHECK5-NEXT: br label [[OMP_BODY_CONTINUE:%.*]] @@ -3651,7 +3651,7 @@ // CHECK6-NEXT: [[TMP1:%.*]] = load [10 x i32]*, [10 x i32]** [[C_ADDR]], align 4 // CHECK6-NEXT: [[TMP2:%.*]] = load i32*, i32** [[A_ADDR]], align 4 // CHECK6-NEXT: [[TMP3:%.*]] = load [10 x i32]*, [10 x i32]** [[D_ADDR]], align 4 -// CHECK6-NEXT: [[TMP4:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 false) +// CHECK6-NEXT: [[TMP4:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 false, i32 0) // CHECK6-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP4]], -1 // CHECK6-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK6: user_code.entry: @@ -3660,7 +3660,7 @@ // CHECK6-NEXT: store i32 [[TMP6]], i32* [[ARGC_CASTED]], align 4 // CHECK6-NEXT: [[TMP7:%.*]] = load i32, i32* [[ARGC_CASTED]], align 4 // CHECK6-NEXT: store i32 [[TMP5]], i32* [[DOTTHREADID_TEMP_]], align 4 -// CHECK6-NEXT: call void @__omp_outlined__(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], [10 x i32]* [[TMP0]], [10 x i32]* [[TMP1]], i32* [[TMP2]], i32 [[TMP7]], [10 x i32]* [[TMP3]]) #[[ATTR1:[0-9]+]] +// CHECK6-NEXT: call void @__omp_outlined__(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], [10 x i32]* [[TMP0]], [10 x i32]* [[TMP1]], i32* [[TMP2]], i32 [[TMP7]], [10 x i32]* [[TMP3]]) #[[ATTR5:[0-9]+]] // CHECK6-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 false) // CHECK6-NEXT: ret void // CHECK6: worker.exit: @@ -3908,20 +3908,20 @@ // CHECK6-NEXT: [[MUL:%.*]] = mul nsw i32 [[TMP18]], 1 // CHECK6-NEXT: [[ADD:%.*]] = add nsw i32 0, [[MUL]] // CHECK6-NEXT: store i32 [[ADD]], i32* [[I5]], align 4 -// CHECK6-NEXT: [[CALL:%.*]] = call i32 @_Z3fooPi(i32* [[I5]]) #[[ATTR5:[0-9]+]] -// CHECK6-NEXT: [[CALL7:%.*]] = call i32 @_Z3fooPi(i32* [[TMP1]]) #[[ATTR5]] +// CHECK6-NEXT: [[CALL:%.*]] = call i32 @_Z3fooPi(i32* [[I5]]) #[[ATTR7:[0-9]+]] +// CHECK6-NEXT: [[CALL7:%.*]] = call i32 @_Z3fooPi(i32* [[TMP1]]) #[[ATTR7]] // CHECK6-NEXT: [[ADD8:%.*]] = add nsw i32 [[CALL]], [[CALL7]] // CHECK6-NEXT: [[TMP19:%.*]] = load i32, i32* [[I5]], align 4 // CHECK6-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [10 x i32], [10 x i32]* [[B3]], i32 0, i32 [[TMP19]] -// CHECK6-NEXT: [[CALL9:%.*]] = call i32 @_Z3fooPi(i32* [[ARRAYIDX]]) #[[ATTR5]] +// CHECK6-NEXT: [[CALL9:%.*]] = call i32 @_Z3fooPi(i32* [[ARRAYIDX]]) #[[ATTR7]] // CHECK6-NEXT: [[ADD10:%.*]] = add nsw i32 [[ADD8]], [[CALL9]] // CHECK6-NEXT: [[TMP20:%.*]] = load i32, i32* [[I5]], align 4 // CHECK6-NEXT: [[ARRAYIDX11:%.*]] = getelementptr inbounds [10 x i32], [10 x i32]* [[C4]], i32 0, i32 [[TMP20]] -// CHECK6-NEXT: [[CALL12:%.*]] = call i32 @_Z3fooPi(i32* [[ARRAYIDX11]]) #[[ATTR5]] +// CHECK6-NEXT: [[CALL12:%.*]] = call i32 @_Z3fooPi(i32* [[ARRAYIDX11]]) #[[ATTR7]] // CHECK6-NEXT: [[ADD13:%.*]] = add nsw i32 [[ADD10]], [[CALL12]] // CHECK6-NEXT: [[TMP21:%.*]] = load i32, i32* [[I5]], align 4 // CHECK6-NEXT: [[ARRAYIDX14:%.*]] = getelementptr inbounds [10 x i32], [10 x i32]* [[TMP4]], i32 0, i32 [[TMP21]] -// CHECK6-NEXT: [[CALL15:%.*]] = call i32 @_Z3fooPi(i32* [[ARRAYIDX14]]) #[[ATTR5]] +// CHECK6-NEXT: [[CALL15:%.*]] = call i32 @_Z3fooPi(i32* [[ARRAYIDX14]]) #[[ATTR7]] // CHECK6-NEXT: [[ADD16:%.*]] = add nsw i32 [[ADD13]], [[CALL15]] // CHECK6-NEXT: store i32 [[ADD16]], i32* [[TMP1]], align 4 // CHECK6-NEXT: br label [[OMP_BODY_CONTINUE:%.*]] diff --git a/clang/test/OpenMP/nvptx_force_full_runtime_SPMD_codegen.cpp b/clang/test/OpenMP/nvptx_force_full_runtime_SPMD_codegen.cpp --- a/clang/test/OpenMP/nvptx_force_full_runtime_SPMD_codegen.cpp +++ b/clang/test/OpenMP/nvptx_force_full_runtime_SPMD_codegen.cpp @@ -11,13 +11,13 @@ // CHECK-NOT: @__omp_offloading_{{.+}}_exec_mode = weak constant i8 1 void foo() { -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) #pragma omp target teams distribute parallel for simd for (int i = 0; i < 10; ++i) ; @@ -40,13 +40,13 @@ for (int i = 0; i < 10; ++i) ; int a; -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) #pragma omp target teams distribute parallel for lastprivate(a) for (int i = 0; i < 10; ++i) a = i; @@ -68,13 +68,13 @@ #pragma omp target teams distribute parallel for schedule(guided) for (int i = 0; i < 10; ++i) ; -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) #pragma omp target teams #pragma omp distribute parallel for simd for (int i = 0; i < 10; ++i) @@ -103,13 +103,13 @@ #pragma omp distribute parallel for simd schedule(guided) for (int i = 0; i < 10; ++i) ; -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) #pragma omp target teams #pragma omp distribute parallel for for (int i = 0; i < 10; ++i) @@ -138,13 +138,13 @@ #pragma omp distribute parallel for schedule(guided) for (int i = 0; i < 10; ++i) ; -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) #pragma omp target #pragma omp teams #pragma omp distribute parallel for @@ -180,13 +180,13 @@ #pragma omp distribute parallel for schedule(guided) for (int i = 0; i < 10; ++i) ; -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) #pragma omp target parallel for for (int i = 0; i < 10; ++i) ; @@ -208,13 +208,13 @@ #pragma omp target parallel for schedule(guided) for (int i = 0; i < 10; ++i) ; -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) #pragma omp target parallel #pragma omp for simd for (int i = 0; i < 10; ++i) @@ -243,13 +243,13 @@ #pragma omp for simd schedule(guided) for (int i = 0; i < 10; ++i) ; -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) #pragma omp target #pragma omp parallel #pragma omp for simd ordered @@ -285,13 +285,13 @@ #pragma omp for simd schedule(guided) for (int i = 0; i < 10; ++i) ; -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) #pragma omp target #pragma omp parallel for for (int i = 0; i < 10; ++i) diff --git a/clang/test/OpenMP/nvptx_lambda_capturing.cpp b/clang/test/OpenMP/nvptx_lambda_capturing.cpp --- a/clang/test/OpenMP/nvptx_lambda_capturing.cpp +++ b/clang/test/OpenMP/nvptx_lambda_capturing.cpp @@ -771,7 +771,7 @@ // CHECK2-NEXT: [[TMP0:%.*]] = load %struct.S*, %struct.S** [[THIS_ADDR]], align 8 // CHECK2-NEXT: [[TMP1:%.*]] = load %class.anon*, %class.anon** [[L_ADDR]], align 8 // CHECK2-NEXT: store %class.anon* [[TMP1]], %class.anon** [[TMP]], align 8 -// CHECK2-NEXT: [[TMP2:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true) +// CHECK2-NEXT: [[TMP2:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true, i32 0) // CHECK2-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP2]], -1 // CHECK2-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK2: user_code.entry: @@ -816,7 +816,7 @@ // CHECK2-NEXT: [[TMP0:%.*]] = load %struct.S*, %struct.S** [[THIS_ADDR]], align 8 // CHECK2-NEXT: [[TMP1:%.*]] = load %class.anon*, %class.anon** [[L_ADDR]], align 8 // CHECK2-NEXT: store %class.anon* [[TMP1]], %class.anon** [[TMP]], align 8 -// CHECK2-NEXT: [[TMP2:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true) +// CHECK2-NEXT: [[TMP2:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true, i32 0) // CHECK2-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP2]], -1 // CHECK2-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK2: user_code.entry: @@ -898,7 +898,7 @@ // CHECK2-NEXT: store i32* [[TMP0]], i32** [[TMP]], align 8 // CHECK2-NEXT: store i32* [[TMP1]], i32** [[_TMP1]], align 8 // CHECK2-NEXT: store %class.anon.0* [[TMP3]], %class.anon.0** [[_TMP2]], align 8 -// CHECK2-NEXT: [[TMP4:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true) +// CHECK2-NEXT: [[TMP4:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true, i32 0) // CHECK2-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP4]], -1 // CHECK2-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK2: user_code.entry: @@ -963,7 +963,7 @@ // CHECK2-NEXT: store i32* [[TMP1]], i32** [[TMP]], align 8 // CHECK2-NEXT: store i32* [[TMP2]], i32** [[_TMP1]], align 8 // CHECK2-NEXT: store %class.anon.0* [[TMP4]], %class.anon.0** [[_TMP2]], align 8 -// CHECK2-NEXT: [[TMP5:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true) +// CHECK2-NEXT: [[TMP5:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true, i32 0) // CHECK2-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP5]], -1 // CHECK2-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK2: user_code.entry: @@ -1080,7 +1080,7 @@ // CHECK2-NEXT: store %class.anon* [[T]], %class.anon** [[T_ADDR]], align 8 // CHECK2-NEXT: [[TMP0:%.*]] = load %class.anon*, %class.anon** [[T_ADDR]], align 8 // CHECK2-NEXT: store %class.anon* [[TMP0]], %class.anon** [[TMP]], align 8 -// CHECK2-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true) +// CHECK2-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true, i32 0) // CHECK2-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK2-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK2: user_code.entry: @@ -1154,7 +1154,7 @@ // CHECK3-NEXT: store i32* [[TMP0]], i32** [[TMP]], align 8 // CHECK3-NEXT: store i32* [[TMP1]], i32** [[_TMP1]], align 8 // CHECK3-NEXT: store %class.anon* [[TMP3]], %class.anon** [[_TMP2]], align 8 -// CHECK3-NEXT: [[TMP4:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true) +// CHECK3-NEXT: [[TMP4:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true, i32 0) // CHECK3-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP4]], -1 // CHECK3-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK3: user_code.entry: @@ -1219,7 +1219,7 @@ // CHECK3-NEXT: store i32* [[TMP1]], i32** [[TMP]], align 8 // CHECK3-NEXT: store i32* [[TMP2]], i32** [[_TMP1]], align 8 // CHECK3-NEXT: store %class.anon* [[TMP4]], %class.anon** [[_TMP2]], align 8 -// CHECK3-NEXT: [[TMP5:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true) +// CHECK3-NEXT: [[TMP5:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true, i32 0) // CHECK3-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP5]], -1 // CHECK3-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK3: user_code.entry: @@ -1340,7 +1340,7 @@ // CHECK3-NEXT: [[TMP0:%.*]] = load %struct.S*, %struct.S** [[THIS_ADDR]], align 8 // CHECK3-NEXT: [[TMP1:%.*]] = load %class.anon.0*, %class.anon.0** [[L_ADDR]], align 8 // CHECK3-NEXT: store %class.anon.0* [[TMP1]], %class.anon.0** [[TMP]], align 8 -// CHECK3-NEXT: [[TMP2:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true) +// CHECK3-NEXT: [[TMP2:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true, i32 0) // CHECK3-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP2]], -1 // CHECK3-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK3: user_code.entry: @@ -1385,7 +1385,7 @@ // CHECK3-NEXT: [[TMP0:%.*]] = load %struct.S*, %struct.S** [[THIS_ADDR]], align 8 // CHECK3-NEXT: [[TMP1:%.*]] = load %class.anon.0*, %class.anon.0** [[L_ADDR]], align 8 // CHECK3-NEXT: store %class.anon.0* [[TMP1]], %class.anon.0** [[TMP]], align 8 -// CHECK3-NEXT: [[TMP2:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true) +// CHECK3-NEXT: [[TMP2:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true, i32 0) // CHECK3-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP2]], -1 // CHECK3-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK3: user_code.entry: @@ -1444,7 +1444,7 @@ // CHECK3-NEXT: store %class.anon.0* [[T]], %class.anon.0** [[T_ADDR]], align 8 // CHECK3-NEXT: [[TMP0:%.*]] = load %class.anon.0*, %class.anon.0** [[T_ADDR]], align 8 // CHECK3-NEXT: store %class.anon.0* [[TMP0]], %class.anon.0** [[TMP]], align 8 -// CHECK3-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true) +// CHECK3-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true, i32 0) // CHECK3-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK3-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK3: user_code.entry: @@ -1518,7 +1518,7 @@ // CHECK4-NEXT: store i32* [[TMP0]], i32** [[TMP]], align 8 // CHECK4-NEXT: store i32* [[TMP1]], i32** [[_TMP1]], align 8 // CHECK4-NEXT: store %class.anon* [[TMP3]], %class.anon** [[_TMP2]], align 8 -// CHECK4-NEXT: [[TMP4:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true) +// CHECK4-NEXT: [[TMP4:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true, i32 0) // CHECK4-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP4]], -1 // CHECK4-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK4: user_code.entry: @@ -1583,7 +1583,7 @@ // CHECK4-NEXT: store i32* [[TMP1]], i32** [[TMP]], align 8 // CHECK4-NEXT: store i32* [[TMP2]], i32** [[_TMP1]], align 8 // CHECK4-NEXT: store %class.anon* [[TMP4]], %class.anon** [[_TMP2]], align 8 -// CHECK4-NEXT: [[TMP5:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true) +// CHECK4-NEXT: [[TMP5:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true, i32 0) // CHECK4-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP5]], -1 // CHECK4-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK4: user_code.entry: @@ -1704,7 +1704,7 @@ // CHECK4-NEXT: [[TMP0:%.*]] = load %struct.S*, %struct.S** [[THIS_ADDR]], align 8 // CHECK4-NEXT: [[TMP1:%.*]] = load %class.anon.0*, %class.anon.0** [[L_ADDR]], align 8 // CHECK4-NEXT: store %class.anon.0* [[TMP1]], %class.anon.0** [[TMP]], align 8 -// CHECK4-NEXT: [[TMP2:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true) +// CHECK4-NEXT: [[TMP2:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true, i32 0) // CHECK4-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP2]], -1 // CHECK4-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK4: user_code.entry: @@ -1749,7 +1749,7 @@ // CHECK4-NEXT: [[TMP0:%.*]] = load %struct.S*, %struct.S** [[THIS_ADDR]], align 8 // CHECK4-NEXT: [[TMP1:%.*]] = load %class.anon.0*, %class.anon.0** [[L_ADDR]], align 8 // CHECK4-NEXT: store %class.anon.0* [[TMP1]], %class.anon.0** [[TMP]], align 8 -// CHECK4-NEXT: [[TMP2:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true) +// CHECK4-NEXT: [[TMP2:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true, i32 0) // CHECK4-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP2]], -1 // CHECK4-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK4: user_code.entry: @@ -1808,7 +1808,7 @@ // CHECK4-NEXT: store %class.anon.0* [[T]], %class.anon.0** [[T_ADDR]], align 8 // CHECK4-NEXT: [[TMP0:%.*]] = load %class.anon.0*, %class.anon.0** [[T_ADDR]], align 8 // CHECK4-NEXT: store %class.anon.0* [[TMP0]], %class.anon.0** [[TMP]], align 8 -// CHECK4-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true) +// CHECK4-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true, i32 0) // CHECK4-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK4-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK4: user_code.entry: diff --git a/clang/test/OpenMP/nvptx_multi_target_parallel_codegen.cpp b/clang/test/OpenMP/nvptx_multi_target_parallel_codegen.cpp --- a/clang/test/OpenMP/nvptx_multi_target_parallel_codegen.cpp +++ b/clang/test/OpenMP/nvptx_multi_target_parallel_codegen.cpp @@ -29,7 +29,7 @@ // CHECK1-SAME: () #[[ATTR0:[0-9]+]] { // CHECK1-NEXT: entry: // CHECK1-NEXT: [[CAPTURED_VARS_ADDRS:%.*]] = alloca [0 x i8*], align 8 -// CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 true) +// CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 true, i32 0) // CHECK1-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 // CHECK1-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK1: user_code.entry: @@ -49,7 +49,7 @@ // CHECK1-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8 // CHECK1-NEXT: store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 8 // CHECK1-NEXT: store i32* [[DOTBOUND_TID_]], i32** [[DOTBOUND_TID__ADDR]], align 8 -// CHECK1-NEXT: call void @_Z3usev() #[[ATTR6:[0-9]+]] +// CHECK1-NEXT: call void @_Z3usev() #[[ATTR7:[0-9]+]] // CHECK1-NEXT: ret void // // @@ -66,11 +66,11 @@ // CHECK1-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}_main_l23 // CHECK1-SAME: () #[[ATTR1]] { // CHECK1-NEXT: entry: -// CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true) +// CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true, i32 0) // CHECK1-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 // CHECK1-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK1: user_code.entry: -// CHECK1-NEXT: call void @_Z3usev() #[[ATTR6]] +// CHECK1-NEXT: call void @_Z3usev() #[[ATTR7]] // CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true) // CHECK1-NEXT: ret void // CHECK1: worker.exit: @@ -84,12 +84,12 @@ // CHECK1-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8 // CHECK1-NEXT: store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 8 // CHECK1-NEXT: store i32* [[DOTBOUND_TID_]], i32** [[DOTBOUND_TID__ADDR]], align 8 -// CHECK1-NEXT: call void @_Z4workv() #[[ATTR6]] +// CHECK1-NEXT: call void @_Z4workv() #[[ATTR7]] // CHECK1-NEXT: ret void // // // CHECK1-LABEL: define {{[^@]+}}@__omp_outlined__1_wrapper -// CHECK1-SAME: (i16 zeroext [[TMP0:%.*]], i32 [[TMP1:%.*]]) #[[ATTR5:[0-9]+]] { +// CHECK1-SAME: (i16 zeroext [[TMP0:%.*]], i32 [[TMP1:%.*]]) #[[ATTR6:[0-9]+]] { // CHECK1-NEXT: entry: // CHECK1-NEXT: [[DOTADDR:%.*]] = alloca i16, align 2 // CHECK1-NEXT: [[DOTADDR1:%.*]] = alloca i32, align 4 @@ -107,7 +107,7 @@ // CHECK2-SAME: () #[[ATTR0:[0-9]+]] { // CHECK2-NEXT: entry: // CHECK2-NEXT: [[CAPTURED_VARS_ADDRS:%.*]] = alloca [0 x i8*], align 4 -// CHECK2-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 true) +// CHECK2-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 true, i32 0) // CHECK2-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 // CHECK2-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK2: user_code.entry: @@ -127,7 +127,7 @@ // CHECK2-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 4 // CHECK2-NEXT: store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 4 // CHECK2-NEXT: store i32* [[DOTBOUND_TID_]], i32** [[DOTBOUND_TID__ADDR]], align 4 -// CHECK2-NEXT: call void @_Z3usev() #[[ATTR6:[0-9]+]] +// CHECK2-NEXT: call void @_Z3usev() #[[ATTR7:[0-9]+]] // CHECK2-NEXT: ret void // // @@ -144,11 +144,11 @@ // CHECK2-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}_main_l23 // CHECK2-SAME: () #[[ATTR1]] { // CHECK2-NEXT: entry: -// CHECK2-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true) +// CHECK2-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true, i32 0) // CHECK2-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 // CHECK2-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK2: user_code.entry: -// CHECK2-NEXT: call void @_Z3usev() #[[ATTR6]] +// CHECK2-NEXT: call void @_Z3usev() #[[ATTR7]] // CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true) // CHECK2-NEXT: ret void // CHECK2: worker.exit: @@ -162,12 +162,12 @@ // CHECK2-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 4 // CHECK2-NEXT: store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 4 // CHECK2-NEXT: store i32* [[DOTBOUND_TID_]], i32** [[DOTBOUND_TID__ADDR]], align 4 -// CHECK2-NEXT: call void @_Z4workv() #[[ATTR6]] +// CHECK2-NEXT: call void @_Z4workv() #[[ATTR7]] // CHECK2-NEXT: ret void // // // CHECK2-LABEL: define {{[^@]+}}@__omp_outlined__1_wrapper -// CHECK2-SAME: (i16 zeroext [[TMP0:%.*]], i32 [[TMP1:%.*]]) #[[ATTR5:[0-9]+]] { +// CHECK2-SAME: (i16 zeroext [[TMP0:%.*]], i32 [[TMP1:%.*]]) #[[ATTR6:[0-9]+]] { // CHECK2-NEXT: entry: // CHECK2-NEXT: [[DOTADDR:%.*]] = alloca i16, align 2 // CHECK2-NEXT: [[DOTADDR1:%.*]] = alloca i32, align 4 @@ -185,7 +185,7 @@ // CHECK3-SAME: () #[[ATTR0:[0-9]+]] { // CHECK3-NEXT: entry: // CHECK3-NEXT: [[CAPTURED_VARS_ADDRS:%.*]] = alloca [0 x i8*], align 4 -// CHECK3-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 true) +// CHECK3-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 true, i32 0) // CHECK3-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 // CHECK3-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK3: user_code.entry: @@ -205,7 +205,7 @@ // CHECK3-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 4 // CHECK3-NEXT: store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 4 // CHECK3-NEXT: store i32* [[DOTBOUND_TID_]], i32** [[DOTBOUND_TID__ADDR]], align 4 -// CHECK3-NEXT: call void @_Z3usev() #[[ATTR6:[0-9]+]] +// CHECK3-NEXT: call void @_Z3usev() #[[ATTR7:[0-9]+]] // CHECK3-NEXT: ret void // // @@ -222,11 +222,11 @@ // CHECK3-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}_main_l23 // CHECK3-SAME: () #[[ATTR1]] { // CHECK3-NEXT: entry: -// CHECK3-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true) +// CHECK3-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true, i32 0) // CHECK3-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 // CHECK3-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK3: user_code.entry: -// CHECK3-NEXT: call void @_Z3usev() #[[ATTR6]] +// CHECK3-NEXT: call void @_Z3usev() #[[ATTR7]] // CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true) // CHECK3-NEXT: ret void // CHECK3: worker.exit: @@ -240,12 +240,12 @@ // CHECK3-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 4 // CHECK3-NEXT: store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 4 // CHECK3-NEXT: store i32* [[DOTBOUND_TID_]], i32** [[DOTBOUND_TID__ADDR]], align 4 -// CHECK3-NEXT: call void @_Z4workv() #[[ATTR6]] +// CHECK3-NEXT: call void @_Z4workv() #[[ATTR7]] // CHECK3-NEXT: ret void // // // CHECK3-LABEL: define {{[^@]+}}@__omp_outlined__1_wrapper -// CHECK3-SAME: (i16 zeroext [[TMP0:%.*]], i32 [[TMP1:%.*]]) #[[ATTR5:[0-9]+]] { +// CHECK3-SAME: (i16 zeroext [[TMP0:%.*]], i32 [[TMP1:%.*]]) #[[ATTR6:[0-9]+]] { // CHECK3-NEXT: entry: // CHECK3-NEXT: [[DOTADDR:%.*]] = alloca i16, align 2 // CHECK3-NEXT: [[DOTADDR1:%.*]] = alloca i32, align 4 diff --git a/clang/test/OpenMP/nvptx_nested_parallel_codegen.cpp b/clang/test/OpenMP/nvptx_nested_parallel_codegen.cpp --- a/clang/test/OpenMP/nvptx_nested_parallel_codegen.cpp +++ b/clang/test/OpenMP/nvptx_nested_parallel_codegen.cpp @@ -40,12 +40,12 @@ // CHECK1-NEXT: [[CAPTURED_VARS_ADDRS:%.*]] = alloca [1 x i8*], align 8 // CHECK1-NEXT: store i32* [[C]], i32** [[C_ADDR]], align 8 // CHECK1-NEXT: [[TMP0:%.*]] = load i32*, i32** [[C_ADDR]], align 8 -// CHECK1-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true) +// CHECK1-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true, i32 0) // CHECK1-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK1-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK1: user_code.entry: // CHECK1-NEXT: [[TMP2:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) -// CHECK1-NEXT: call void @_Z3usePi(i32* [[TMP0]]) #[[ATTR5:[0-9]+]] +// CHECK1-NEXT: call void @_Z3usePi(i32* [[TMP0]]) #[[ATTR6:[0-9]+]] // CHECK1-NEXT: call void @__kmpc_push_num_threads(%struct.ident_t* @[[GLOB1]], i32 [[TMP2]], i32 2) // CHECK1-NEXT: [[TMP3:%.*]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[CAPTURED_VARS_ADDRS]], i64 0, i64 0 // CHECK1-NEXT: [[TMP4:%.*]] = bitcast i32* [[TMP0]] to i8* @@ -84,7 +84,7 @@ // CHECK1-NEXT: store i32* [[DOTBOUND_TID_]], i32** [[DOTBOUND_TID__ADDR]], align 8 // CHECK1-NEXT: store i32* [[C]], i32** [[C_ADDR]], align 8 // CHECK1-NEXT: [[TMP0:%.*]] = load i32*, i32** [[C_ADDR]], align 8 -// CHECK1-NEXT: call void @_Z3usePi(i32* [[TMP0]]) #[[ATTR5]] +// CHECK1-NEXT: call void @_Z3usePi(i32* [[TMP0]]) #[[ATTR6]] // CHECK1-NEXT: ret void // // @@ -118,7 +118,7 @@ // CHECK1-NEXT: store i32** [[C]], i32*** [[C_ADDR]], align 8 // CHECK1-NEXT: [[TMP0:%.*]] = load i32**, i32*** [[C_ADDR]], align 8 // CHECK1-NEXT: [[TMP1:%.*]] = load i32*, i32** [[TMP0]], align 8 -// CHECK1-NEXT: call void @_Z4workPi(i32* [[TMP1]]) #[[ATTR5]] +// CHECK1-NEXT: call void @_Z4workPi(i32* [[TMP1]]) #[[ATTR6]] // CHECK1-NEXT: ret void // // @@ -132,7 +132,7 @@ // CHECK1-NEXT: [[TMP0:%.*]] = load i32*, i32** [[C_ADDR]], align 8 // CHECK1-NEXT: [[TMP1:%.*]] = bitcast i32* [[TMP0]] to i8* // CHECK1-NEXT: [[TMP2:%.*]] = bitcast i32* [[ATOMIC_TEMP]] to i8* -// CHECK1-NEXT: call void @__atomic_load(i64 4, i8* [[TMP1]], i8* [[TMP2]], i32 0) #[[ATTR5]] +// CHECK1-NEXT: call void @__atomic_load(i64 4, i8* [[TMP1]], i8* [[TMP2]], i32 0) #[[ATTR6]] // CHECK1-NEXT: br label [[ATOMIC_CONT:%.*]] // CHECK1: atomic_cont: // CHECK1-NEXT: [[TMP3:%.*]] = load i32, i32* [[ATOMIC_TEMP]], align 4 @@ -141,7 +141,7 @@ // CHECK1-NEXT: [[TMP4:%.*]] = bitcast i32* [[TMP0]] to i8* // CHECK1-NEXT: [[TMP5:%.*]] = bitcast i32* [[ATOMIC_TEMP]] to i8* // CHECK1-NEXT: [[TMP6:%.*]] = bitcast i32* [[ATOMIC_TEMP1]] to i8* -// CHECK1-NEXT: [[CALL:%.*]] = call zeroext i1 @__atomic_compare_exchange(i64 4, i8* [[TMP4]], i8* [[TMP5]], i8* [[TMP6]], i32 0, i32 0) #[[ATTR5]] +// CHECK1-NEXT: [[CALL:%.*]] = call zeroext i1 @__atomic_compare_exchange(i64 4, i8* [[TMP4]], i8* [[TMP5]], i8* [[TMP6]], i32 0, i32 0) #[[ATTR6]] // CHECK1-NEXT: br i1 [[CALL]], label [[ATOMIC_EXIT:%.*]], label [[ATOMIC_CONT]] // CHECK1: atomic_exit: // CHECK1-NEXT: ret void @@ -173,12 +173,12 @@ // CHECK2-NEXT: [[CAPTURED_VARS_ADDRS:%.*]] = alloca [1 x i8*], align 4 // CHECK2-NEXT: store i32* [[C]], i32** [[C_ADDR]], align 4 // CHECK2-NEXT: [[TMP0:%.*]] = load i32*, i32** [[C_ADDR]], align 4 -// CHECK2-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true) +// CHECK2-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true, i32 0) // CHECK2-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK2-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK2: user_code.entry: // CHECK2-NEXT: [[TMP2:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) -// CHECK2-NEXT: call void @_Z3usePi(i32* [[TMP0]]) #[[ATTR5:[0-9]+]] +// CHECK2-NEXT: call void @_Z3usePi(i32* [[TMP0]]) #[[ATTR6:[0-9]+]] // CHECK2-NEXT: call void @__kmpc_push_num_threads(%struct.ident_t* @[[GLOB1]], i32 [[TMP2]], i32 2) // CHECK2-NEXT: [[TMP3:%.*]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[CAPTURED_VARS_ADDRS]], i32 0, i32 0 // CHECK2-NEXT: [[TMP4:%.*]] = bitcast i32* [[TMP0]] to i8* @@ -217,7 +217,7 @@ // CHECK2-NEXT: store i32* [[DOTBOUND_TID_]], i32** [[DOTBOUND_TID__ADDR]], align 4 // CHECK2-NEXT: store i32* [[C]], i32** [[C_ADDR]], align 4 // CHECK2-NEXT: [[TMP0:%.*]] = load i32*, i32** [[C_ADDR]], align 4 -// CHECK2-NEXT: call void @_Z3usePi(i32* [[TMP0]]) #[[ATTR5]] +// CHECK2-NEXT: call void @_Z3usePi(i32* [[TMP0]]) #[[ATTR6]] // CHECK2-NEXT: ret void // // @@ -251,7 +251,7 @@ // CHECK2-NEXT: store i32** [[C]], i32*** [[C_ADDR]], align 4 // CHECK2-NEXT: [[TMP0:%.*]] = load i32**, i32*** [[C_ADDR]], align 4 // CHECK2-NEXT: [[TMP1:%.*]] = load i32*, i32** [[TMP0]], align 4 -// CHECK2-NEXT: call void @_Z4workPi(i32* [[TMP1]]) #[[ATTR5]] +// CHECK2-NEXT: call void @_Z4workPi(i32* [[TMP1]]) #[[ATTR6]] // CHECK2-NEXT: ret void // // @@ -265,7 +265,7 @@ // CHECK2-NEXT: [[TMP0:%.*]] = load i32*, i32** [[C_ADDR]], align 4 // CHECK2-NEXT: [[TMP1:%.*]] = bitcast i32* [[TMP0]] to i8* // CHECK2-NEXT: [[TMP2:%.*]] = bitcast i32* [[ATOMIC_TEMP]] to i8* -// CHECK2-NEXT: call void @__atomic_load(i32 4, i8* [[TMP1]], i8* [[TMP2]], i32 0) #[[ATTR5]] +// CHECK2-NEXT: call void @__atomic_load(i32 4, i8* [[TMP1]], i8* [[TMP2]], i32 0) #[[ATTR6]] // CHECK2-NEXT: br label [[ATOMIC_CONT:%.*]] // CHECK2: atomic_cont: // CHECK2-NEXT: [[TMP3:%.*]] = load i32, i32* [[ATOMIC_TEMP]], align 4 @@ -274,7 +274,7 @@ // CHECK2-NEXT: [[TMP4:%.*]] = bitcast i32* [[TMP0]] to i8* // CHECK2-NEXT: [[TMP5:%.*]] = bitcast i32* [[ATOMIC_TEMP]] to i8* // CHECK2-NEXT: [[TMP6:%.*]] = bitcast i32* [[ATOMIC_TEMP1]] to i8* -// CHECK2-NEXT: [[CALL:%.*]] = call zeroext i1 @__atomic_compare_exchange(i32 4, i8* [[TMP4]], i8* [[TMP5]], i8* [[TMP6]], i32 0, i32 0) #[[ATTR5]] +// CHECK2-NEXT: [[CALL:%.*]] = call zeroext i1 @__atomic_compare_exchange(i32 4, i8* [[TMP4]], i8* [[TMP5]], i8* [[TMP6]], i32 0, i32 0) #[[ATTR6]] // CHECK2-NEXT: br i1 [[CALL]], label [[ATOMIC_EXIT:%.*]], label [[ATOMIC_CONT]] // CHECK2: atomic_exit: // CHECK2-NEXT: ret void @@ -306,12 +306,12 @@ // CHECK3-NEXT: [[CAPTURED_VARS_ADDRS:%.*]] = alloca [1 x i8*], align 4 // CHECK3-NEXT: store i32* [[C]], i32** [[C_ADDR]], align 4 // CHECK3-NEXT: [[TMP0:%.*]] = load i32*, i32** [[C_ADDR]], align 4 -// CHECK3-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true) +// CHECK3-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true, i32 0) // CHECK3-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK3-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK3: user_code.entry: // CHECK3-NEXT: [[TMP2:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) -// CHECK3-NEXT: call void @_Z3usePi(i32* [[TMP0]]) #[[ATTR5:[0-9]+]] +// CHECK3-NEXT: call void @_Z3usePi(i32* [[TMP0]]) #[[ATTR6:[0-9]+]] // CHECK3-NEXT: call void @__kmpc_push_num_threads(%struct.ident_t* @[[GLOB1]], i32 [[TMP2]], i32 2) // CHECK3-NEXT: [[TMP3:%.*]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[CAPTURED_VARS_ADDRS]], i32 0, i32 0 // CHECK3-NEXT: [[TMP4:%.*]] = bitcast i32* [[TMP0]] to i8* @@ -350,7 +350,7 @@ // CHECK3-NEXT: store i32* [[DOTBOUND_TID_]], i32** [[DOTBOUND_TID__ADDR]], align 4 // CHECK3-NEXT: store i32* [[C]], i32** [[C_ADDR]], align 4 // CHECK3-NEXT: [[TMP0:%.*]] = load i32*, i32** [[C_ADDR]], align 4 -// CHECK3-NEXT: call void @_Z3usePi(i32* [[TMP0]]) #[[ATTR5]] +// CHECK3-NEXT: call void @_Z3usePi(i32* [[TMP0]]) #[[ATTR6]] // CHECK3-NEXT: ret void // // @@ -384,7 +384,7 @@ // CHECK3-NEXT: store i32** [[C]], i32*** [[C_ADDR]], align 4 // CHECK3-NEXT: [[TMP0:%.*]] = load i32**, i32*** [[C_ADDR]], align 4 // CHECK3-NEXT: [[TMP1:%.*]] = load i32*, i32** [[TMP0]], align 4 -// CHECK3-NEXT: call void @_Z4workPi(i32* [[TMP1]]) #[[ATTR5]] +// CHECK3-NEXT: call void @_Z4workPi(i32* [[TMP1]]) #[[ATTR6]] // CHECK3-NEXT: ret void // // @@ -398,7 +398,7 @@ // CHECK3-NEXT: [[TMP0:%.*]] = load i32*, i32** [[C_ADDR]], align 4 // CHECK3-NEXT: [[TMP1:%.*]] = bitcast i32* [[TMP0]] to i8* // CHECK3-NEXT: [[TMP2:%.*]] = bitcast i32* [[ATOMIC_TEMP]] to i8* -// CHECK3-NEXT: call void @__atomic_load(i32 4, i8* [[TMP1]], i8* [[TMP2]], i32 0) #[[ATTR5]] +// CHECK3-NEXT: call void @__atomic_load(i32 4, i8* [[TMP1]], i8* [[TMP2]], i32 0) #[[ATTR6]] // CHECK3-NEXT: br label [[ATOMIC_CONT:%.*]] // CHECK3: atomic_cont: // CHECK3-NEXT: [[TMP3:%.*]] = load i32, i32* [[ATOMIC_TEMP]], align 4 @@ -407,7 +407,7 @@ // CHECK3-NEXT: [[TMP4:%.*]] = bitcast i32* [[TMP0]] to i8* // CHECK3-NEXT: [[TMP5:%.*]] = bitcast i32* [[ATOMIC_TEMP]] to i8* // CHECK3-NEXT: [[TMP6:%.*]] = bitcast i32* [[ATOMIC_TEMP1]] to i8* -// CHECK3-NEXT: [[CALL:%.*]] = call zeroext i1 @__atomic_compare_exchange(i32 4, i8* [[TMP4]], i8* [[TMP5]], i8* [[TMP6]], i32 0, i32 0) #[[ATTR5]] +// CHECK3-NEXT: [[CALL:%.*]] = call zeroext i1 @__atomic_compare_exchange(i32 4, i8* [[TMP4]], i8* [[TMP5]], i8* [[TMP6]], i32 0, i32 0) #[[ATTR6]] // CHECK3-NEXT: br i1 [[CALL]], label [[ATOMIC_EXIT:%.*]], label [[ATOMIC_CONT]] // CHECK3: atomic_exit: // CHECK3-NEXT: ret void diff --git a/clang/test/OpenMP/nvptx_parallel_codegen.cpp b/clang/test/OpenMP/nvptx_parallel_codegen.cpp --- a/clang/test/OpenMP/nvptx_parallel_codegen.cpp +++ b/clang/test/OpenMP/nvptx_parallel_codegen.cpp @@ -1446,7 +1446,7 @@ // CHECK1-NEXT: [[CAPTURED_VARS_ADDRS2:%.*]] = alloca [0 x i8*], align 8 // CHECK1-NEXT: store i64 [[A]], i64* [[A_ADDR]], align 8 // CHECK1-NEXT: [[CONV:%.*]] = bitcast i64* [[A_ADDR]] to i32* -// CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true) +// CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true, i32 0) // CHECK1-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 // CHECK1-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK1: user_code.entry: @@ -1563,7 +1563,7 @@ // CHECK1-NEXT: [[CONV1:%.*]] = bitcast i64* [[A_ADDR]] to i32* // CHECK1-NEXT: [[CONV2:%.*]] = bitcast i64* [[AA_ADDR]] to i16* // CHECK1-NEXT: [[TMP0:%.*]] = load [10 x i32]*, [10 x i32]** [[B_ADDR]], align 8 -// CHECK1-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true) +// CHECK1-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true, i32 0) // CHECK1-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK1-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK1: user_code.entry: @@ -1628,7 +1628,7 @@ // CHECK1-NEXT: [[CAPTURED_VARS_ADDRS:%.*]] = alloca [1 x i8*], align 8 // CHECK1-NEXT: store i64 [[A]], i64* [[A_ADDR]], align 8 // CHECK1-NEXT: [[CONV:%.*]] = bitcast i64* [[A_ADDR]] to i32* -// CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true) +// CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true, i32 0) // CHECK1-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 // CHECK1-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK1: user_code.entry: @@ -1721,7 +1721,7 @@ // CHECK2-NEXT: [[CAPTURED_VARS_ADDRS1:%.*]] = alloca [0 x i8*], align 4 // CHECK2-NEXT: [[CAPTURED_VARS_ADDRS2:%.*]] = alloca [0 x i8*], align 4 // CHECK2-NEXT: store i32 [[A]], i32* [[A_ADDR]], align 4 -// CHECK2-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true) +// CHECK2-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true, i32 0) // CHECK2-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 // CHECK2-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK2: user_code.entry: @@ -1836,7 +1836,7 @@ // CHECK2-NEXT: store [10 x i32]* [[B]], [10 x i32]** [[B_ADDR]], align 4 // CHECK2-NEXT: [[CONV:%.*]] = bitcast i32* [[AA_ADDR]] to i16* // CHECK2-NEXT: [[TMP0:%.*]] = load [10 x i32]*, [10 x i32]** [[B_ADDR]], align 4 -// CHECK2-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true) +// CHECK2-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true, i32 0) // CHECK2-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK2-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK2: user_code.entry: @@ -1900,7 +1900,7 @@ // CHECK2-NEXT: [[A_ADDR:%.*]] = alloca i32, align 4 // CHECK2-NEXT: [[CAPTURED_VARS_ADDRS:%.*]] = alloca [1 x i8*], align 4 // CHECK2-NEXT: store i32 [[A]], i32* [[A_ADDR]], align 4 -// CHECK2-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true) +// CHECK2-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true, i32 0) // CHECK2-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 // CHECK2-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK2: user_code.entry: diff --git a/clang/test/OpenMP/nvptx_parallel_for_codegen.cpp b/clang/test/OpenMP/nvptx_parallel_for_codegen.cpp --- a/clang/test/OpenMP/nvptx_parallel_for_codegen.cpp +++ b/clang/test/OpenMP/nvptx_parallel_for_codegen.cpp @@ -465,7 +465,7 @@ // CHECK-NEXT: store [10 x i32]* [[B]], [10 x i32]** [[B_ADDR]], align 8 // CHECK-NEXT: [[CONV:%.*]] = bitcast i64* [[N_ADDR]] to i32* // CHECK-NEXT: [[TMP0:%.*]] = load [10 x i32]*, [10 x i32]** [[B_ADDR]], align 8 -// CHECK-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true) +// CHECK-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true, i32 0) // CHECK-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK: user_code.entry: diff --git a/clang/test/OpenMP/nvptx_target_codegen.cpp b/clang/test/OpenMP/nvptx_target_codegen.cpp --- a/clang/test/OpenMP/nvptx_target_codegen.cpp +++ b/clang/test/OpenMP/nvptx_target_codegen.cpp @@ -153,7 +153,7 @@ // CHECK1-NEXT: store i32* [[PTR1]], i32** [[PTR1_ADDR]], align 8 // CHECK1-NEXT: store i32** [[PTR2]], i32*** [[PTR2_ADDR]], align 8 // CHECK1-NEXT: [[TMP0:%.*]] = load i32**, i32*** [[PTR2_ADDR]], align 8 -// CHECK1-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 true) +// CHECK1-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 true, i32 0) // CHECK1-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK1-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK1: user_code.entry: @@ -195,7 +195,7 @@ // CHECK1-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z3fooi_l39 // CHECK1-SAME: () #[[ATTR1]] { // CHECK1-NEXT: entry: -// CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true) +// CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true, i32 0) // CHECK1-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 // CHECK1-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK1: user_code.entry: @@ -211,7 +211,7 @@ // CHECK1-NEXT: [[AA_ADDR:%.*]] = alloca i64, align 8 // CHECK1-NEXT: store i64 [[AA]], i64* [[AA_ADDR]], align 8 // CHECK1-NEXT: [[CONV:%.*]] = bitcast i64* [[AA_ADDR]] to i16* -// CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true) +// CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true, i32 0) // CHECK1-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 // CHECK1-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK1: user_code.entry: @@ -261,7 +261,7 @@ // CHECK1-NEXT: [[TMP5:%.*]] = load i64, i64* [[VLA_ADDR4]], align 8 // CHECK1-NEXT: [[TMP6:%.*]] = load double*, double** [[CN_ADDR]], align 8 // CHECK1-NEXT: [[TMP7:%.*]] = load %struct.TT*, %struct.TT** [[D_ADDR]], align 8 -// CHECK1-NEXT: [[TMP8:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true) +// CHECK1-NEXT: [[TMP8:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true, i32 0) // CHECK1-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP8]], -1 // CHECK1-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK1: user_code.entry: @@ -338,7 +338,7 @@ // CHECK1-NEXT: [[CONV1:%.*]] = bitcast i64* [[AA_ADDR]] to i16* // CHECK1-NEXT: [[CONV2:%.*]] = bitcast i64* [[AAA_ADDR]] to i8* // CHECK1-NEXT: [[TMP0:%.*]] = load [10 x i32]*, [10 x i32]** [[B_ADDR]], align 8 -// CHECK1-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true) +// CHECK1-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true, i32 0) // CHECK1-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK1-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK1: user_code.entry: @@ -383,7 +383,7 @@ // CHECK1-NEXT: [[TMP1:%.*]] = load i64, i64* [[VLA_ADDR]], align 8 // CHECK1-NEXT: [[TMP2:%.*]] = load i64, i64* [[VLA_ADDR2]], align 8 // CHECK1-NEXT: [[TMP3:%.*]] = load i16*, i16** [[C_ADDR]], align 8 -// CHECK1-NEXT: [[TMP4:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true) +// CHECK1-NEXT: [[TMP4:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true, i32 0) // CHECK1-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP4]], -1 // CHECK1-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK1: user_code.entry: @@ -439,7 +439,7 @@ // CHECK1-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z16unreachable_callv_l142 // CHECK1-SAME: () #[[ATTR1]] { // CHECK1-NEXT: entry: -// CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true) +// CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true, i32 0) // CHECK1-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 // CHECK1-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK1: user_code.entry: @@ -464,7 +464,7 @@ // CHECK1-NEXT: [[CONV:%.*]] = bitcast i64* [[A_ADDR]] to i32* // CHECK1-NEXT: [[CONV1:%.*]] = bitcast i64* [[AA_ADDR]] to i16* // CHECK1-NEXT: [[TMP0:%.*]] = load [10 x i32]*, [10 x i32]** [[B_ADDR]], align 8 -// CHECK1-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true) +// CHECK1-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true, i32 0) // CHECK1-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK1-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK1: user_code.entry: @@ -540,7 +540,7 @@ // CHECK2-NEXT: store i32* [[PTR1]], i32** [[PTR1_ADDR]], align 4 // CHECK2-NEXT: store i32** [[PTR2]], i32*** [[PTR2_ADDR]], align 4 // CHECK2-NEXT: [[TMP0:%.*]] = load i32**, i32*** [[PTR2_ADDR]], align 4 -// CHECK2-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 true) +// CHECK2-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 true, i32 0) // CHECK2-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK2-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK2: user_code.entry: @@ -582,7 +582,7 @@ // CHECK2-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z3fooi_l39 // CHECK2-SAME: () #[[ATTR1]] { // CHECK2-NEXT: entry: -// CHECK2-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true) +// CHECK2-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true, i32 0) // CHECK2-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 // CHECK2-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK2: user_code.entry: @@ -598,7 +598,7 @@ // CHECK2-NEXT: [[AA_ADDR:%.*]] = alloca i32, align 4 // CHECK2-NEXT: store i32 [[AA]], i32* [[AA_ADDR]], align 4 // CHECK2-NEXT: [[CONV:%.*]] = bitcast i32* [[AA_ADDR]] to i16* -// CHECK2-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true) +// CHECK2-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true, i32 0) // CHECK2-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 // CHECK2-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK2: user_code.entry: @@ -647,7 +647,7 @@ // CHECK2-NEXT: [[TMP5:%.*]] = load i32, i32* [[VLA_ADDR4]], align 4 // CHECK2-NEXT: [[TMP6:%.*]] = load double*, double** [[CN_ADDR]], align 4 // CHECK2-NEXT: [[TMP7:%.*]] = load %struct.TT*, %struct.TT** [[D_ADDR]], align 4 -// CHECK2-NEXT: [[TMP8:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true) +// CHECK2-NEXT: [[TMP8:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true, i32 0) // CHECK2-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP8]], -1 // CHECK2-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK2: user_code.entry: @@ -723,7 +723,7 @@ // CHECK2-NEXT: [[CONV:%.*]] = bitcast i32* [[AA_ADDR]] to i16* // CHECK2-NEXT: [[CONV1:%.*]] = bitcast i32* [[AAA_ADDR]] to i8* // CHECK2-NEXT: [[TMP0:%.*]] = load [10 x i32]*, [10 x i32]** [[B_ADDR]], align 4 -// CHECK2-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true) +// CHECK2-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true, i32 0) // CHECK2-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK2-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK2: user_code.entry: @@ -767,7 +767,7 @@ // CHECK2-NEXT: [[TMP1:%.*]] = load i32, i32* [[VLA_ADDR]], align 4 // CHECK2-NEXT: [[TMP2:%.*]] = load i32, i32* [[VLA_ADDR2]], align 4 // CHECK2-NEXT: [[TMP3:%.*]] = load i16*, i16** [[C_ADDR]], align 4 -// CHECK2-NEXT: [[TMP4:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true) +// CHECK2-NEXT: [[TMP4:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true, i32 0) // CHECK2-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP4]], -1 // CHECK2-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK2: user_code.entry: @@ -823,7 +823,7 @@ // CHECK2-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z16unreachable_callv_l142 // CHECK2-SAME: () #[[ATTR1]] { // CHECK2-NEXT: entry: -// CHECK2-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true) +// CHECK2-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true, i32 0) // CHECK2-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 // CHECK2-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK2: user_code.entry: @@ -847,7 +847,7 @@ // CHECK2-NEXT: store [10 x i32]* [[B]], [10 x i32]** [[B_ADDR]], align 4 // CHECK2-NEXT: [[CONV:%.*]] = bitcast i32* [[AA_ADDR]] to i16* // CHECK2-NEXT: [[TMP0:%.*]] = load [10 x i32]*, [10 x i32]** [[B_ADDR]], align 4 -// CHECK2-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true) +// CHECK2-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true, i32 0) // CHECK2-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK2-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK2: user_code.entry: @@ -923,7 +923,7 @@ // CHECK3-NEXT: store i32* [[PTR1]], i32** [[PTR1_ADDR]], align 4 // CHECK3-NEXT: store i32** [[PTR2]], i32*** [[PTR2_ADDR]], align 4 // CHECK3-NEXT: [[TMP0:%.*]] = load i32**, i32*** [[PTR2_ADDR]], align 4 -// CHECK3-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 true) +// CHECK3-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 true, i32 0) // CHECK3-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK3-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK3: user_code.entry: @@ -965,7 +965,7 @@ // CHECK3-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z3fooi_l39 // CHECK3-SAME: () #[[ATTR1]] { // CHECK3-NEXT: entry: -// CHECK3-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true) +// CHECK3-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true, i32 0) // CHECK3-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 // CHECK3-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK3: user_code.entry: @@ -981,7 +981,7 @@ // CHECK3-NEXT: [[AA_ADDR:%.*]] = alloca i32, align 4 // CHECK3-NEXT: store i32 [[AA]], i32* [[AA_ADDR]], align 4 // CHECK3-NEXT: [[CONV:%.*]] = bitcast i32* [[AA_ADDR]] to i16* -// CHECK3-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true) +// CHECK3-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true, i32 0) // CHECK3-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 // CHECK3-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK3: user_code.entry: @@ -1030,7 +1030,7 @@ // CHECK3-NEXT: [[TMP5:%.*]] = load i32, i32* [[VLA_ADDR4]], align 4 // CHECK3-NEXT: [[TMP6:%.*]] = load double*, double** [[CN_ADDR]], align 4 // CHECK3-NEXT: [[TMP7:%.*]] = load %struct.TT*, %struct.TT** [[D_ADDR]], align 4 -// CHECK3-NEXT: [[TMP8:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true) +// CHECK3-NEXT: [[TMP8:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true, i32 0) // CHECK3-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP8]], -1 // CHECK3-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK3: user_code.entry: @@ -1106,7 +1106,7 @@ // CHECK3-NEXT: [[CONV:%.*]] = bitcast i32* [[AA_ADDR]] to i16* // CHECK3-NEXT: [[CONV1:%.*]] = bitcast i32* [[AAA_ADDR]] to i8* // CHECK3-NEXT: [[TMP0:%.*]] = load [10 x i32]*, [10 x i32]** [[B_ADDR]], align 4 -// CHECK3-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true) +// CHECK3-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true, i32 0) // CHECK3-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK3-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK3: user_code.entry: @@ -1150,7 +1150,7 @@ // CHECK3-NEXT: [[TMP1:%.*]] = load i32, i32* [[VLA_ADDR]], align 4 // CHECK3-NEXT: [[TMP2:%.*]] = load i32, i32* [[VLA_ADDR2]], align 4 // CHECK3-NEXT: [[TMP3:%.*]] = load i16*, i16** [[C_ADDR]], align 4 -// CHECK3-NEXT: [[TMP4:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true) +// CHECK3-NEXT: [[TMP4:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true, i32 0) // CHECK3-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP4]], -1 // CHECK3-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK3: user_code.entry: @@ -1206,7 +1206,7 @@ // CHECK3-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z16unreachable_callv_l142 // CHECK3-SAME: () #[[ATTR1]] { // CHECK3-NEXT: entry: -// CHECK3-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true) +// CHECK3-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true, i32 0) // CHECK3-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 // CHECK3-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK3: user_code.entry: @@ -1230,7 +1230,7 @@ // CHECK3-NEXT: store [10 x i32]* [[B]], [10 x i32]** [[B_ADDR]], align 4 // CHECK3-NEXT: [[CONV:%.*]] = bitcast i32* [[AA_ADDR]] to i16* // CHECK3-NEXT: [[TMP0:%.*]] = load [10 x i32]*, [10 x i32]** [[B_ADDR]], align 4 -// CHECK3-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true) +// CHECK3-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true, i32 0) // CHECK3-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK3-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK3: user_code.entry: diff --git a/clang/test/OpenMP/nvptx_target_parallel_codegen.cpp b/clang/test/OpenMP/nvptx_target_parallel_codegen.cpp --- a/clang/test/OpenMP/nvptx_target_parallel_codegen.cpp +++ b/clang/test/OpenMP/nvptx_target_parallel_codegen.cpp @@ -58,7 +58,7 @@ // CHECK1-NEXT: [[CAPTURED_VARS_ADDRS:%.*]] = alloca [1 x i8*], align 8 // CHECK1-NEXT: store i16* [[AA]], i16** [[AA_ADDR]], align 8 // CHECK1-NEXT: [[TMP0:%.*]] = load i16*, i16** [[AA_ADDR]], align 8 -// CHECK1-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 true) +// CHECK1-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 true, i32 0) // CHECK1-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK1-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK1: user_code.entry: @@ -105,7 +105,7 @@ // CHECK1-NEXT: [[TMP0:%.*]] = load i32*, i32** [[A_ADDR]], align 8 // CHECK1-NEXT: [[TMP1:%.*]] = load i16*, i16** [[AA_ADDR]], align 8 // CHECK1-NEXT: [[TMP2:%.*]] = load [10 x i32]*, [10 x i32]** [[B_ADDR]], align 8 -// CHECK1-NEXT: [[TMP3:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true) +// CHECK1-NEXT: [[TMP3:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true, i32 0) // CHECK1-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP3]], -1 // CHECK1-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK1: user_code.entry: @@ -165,7 +165,7 @@ // CHECK2-NEXT: [[CAPTURED_VARS_ADDRS:%.*]] = alloca [1 x i8*], align 4 // CHECK2-NEXT: store i16* [[AA]], i16** [[AA_ADDR]], align 4 // CHECK2-NEXT: [[TMP0:%.*]] = load i16*, i16** [[AA_ADDR]], align 4 -// CHECK2-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 true) +// CHECK2-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 true, i32 0) // CHECK2-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK2-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK2: user_code.entry: @@ -212,7 +212,7 @@ // CHECK2-NEXT: [[TMP0:%.*]] = load i32*, i32** [[A_ADDR]], align 4 // CHECK2-NEXT: [[TMP1:%.*]] = load i16*, i16** [[AA_ADDR]], align 4 // CHECK2-NEXT: [[TMP2:%.*]] = load [10 x i32]*, [10 x i32]** [[B_ADDR]], align 4 -// CHECK2-NEXT: [[TMP3:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true) +// CHECK2-NEXT: [[TMP3:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true, i32 0) // CHECK2-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP3]], -1 // CHECK2-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK2: user_code.entry: @@ -272,7 +272,7 @@ // CHECK3-NEXT: [[CAPTURED_VARS_ADDRS:%.*]] = alloca [1 x i8*], align 4 // CHECK3-NEXT: store i16* [[AA]], i16** [[AA_ADDR]], align 4 // CHECK3-NEXT: [[TMP0:%.*]] = load i16*, i16** [[AA_ADDR]], align 4 -// CHECK3-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 true) +// CHECK3-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 true, i32 0) // CHECK3-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK3-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK3: user_code.entry: @@ -319,7 +319,7 @@ // CHECK3-NEXT: [[TMP0:%.*]] = load i32*, i32** [[A_ADDR]], align 4 // CHECK3-NEXT: [[TMP1:%.*]] = load i16*, i16** [[AA_ADDR]], align 4 // CHECK3-NEXT: [[TMP2:%.*]] = load [10 x i32]*, [10 x i32]** [[B_ADDR]], align 4 -// CHECK3-NEXT: [[TMP3:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true) +// CHECK3-NEXT: [[TMP3:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true, i32 0) // CHECK3-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP3]], -1 // CHECK3-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK3: user_code.entry: @@ -379,7 +379,7 @@ // CHECK4-NEXT: [[CAPTURED_VARS_ADDRS:%.*]] = alloca [1 x i8*], align 8 // CHECK4-NEXT: store i16* [[AA]], i16** [[AA_ADDR]], align 8 // CHECK4-NEXT: [[TMP0:%.*]] = load i16*, i16** [[AA_ADDR]], align 8 -// CHECK4-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 true) +// CHECK4-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 true, i32 0) // CHECK4-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK4-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK4: user_code.entry: @@ -426,7 +426,7 @@ // CHECK4-NEXT: [[TMP0:%.*]] = load i32*, i32** [[A_ADDR]], align 8 // CHECK4-NEXT: [[TMP1:%.*]] = load i16*, i16** [[AA_ADDR]], align 8 // CHECK4-NEXT: [[TMP2:%.*]] = load [10 x i32]*, [10 x i32]** [[B_ADDR]], align 8 -// CHECK4-NEXT: [[TMP3:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true) +// CHECK4-NEXT: [[TMP3:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true, i32 0) // CHECK4-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP3]], -1 // CHECK4-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK4: user_code.entry: @@ -486,7 +486,7 @@ // CHECK5-NEXT: [[CAPTURED_VARS_ADDRS:%.*]] = alloca [1 x i8*], align 4 // CHECK5-NEXT: store i16* [[AA]], i16** [[AA_ADDR]], align 4 // CHECK5-NEXT: [[TMP0:%.*]] = load i16*, i16** [[AA_ADDR]], align 4 -// CHECK5-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 true) +// CHECK5-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 true, i32 0) // CHECK5-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK5-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK5: user_code.entry: @@ -533,7 +533,7 @@ // CHECK5-NEXT: [[TMP0:%.*]] = load i32*, i32** [[A_ADDR]], align 4 // CHECK5-NEXT: [[TMP1:%.*]] = load i16*, i16** [[AA_ADDR]], align 4 // CHECK5-NEXT: [[TMP2:%.*]] = load [10 x i32]*, [10 x i32]** [[B_ADDR]], align 4 -// CHECK5-NEXT: [[TMP3:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true) +// CHECK5-NEXT: [[TMP3:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true, i32 0) // CHECK5-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP3]], -1 // CHECK5-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK5: user_code.entry: @@ -593,7 +593,7 @@ // CHECK6-NEXT: [[CAPTURED_VARS_ADDRS:%.*]] = alloca [1 x i8*], align 4 // CHECK6-NEXT: store i16* [[AA]], i16** [[AA_ADDR]], align 4 // CHECK6-NEXT: [[TMP0:%.*]] = load i16*, i16** [[AA_ADDR]], align 4 -// CHECK6-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 true) +// CHECK6-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 true, i32 0) // CHECK6-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK6-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK6: user_code.entry: @@ -640,7 +640,7 @@ // CHECK6-NEXT: [[TMP0:%.*]] = load i32*, i32** [[A_ADDR]], align 4 // CHECK6-NEXT: [[TMP1:%.*]] = load i16*, i16** [[AA_ADDR]], align 4 // CHECK6-NEXT: [[TMP2:%.*]] = load [10 x i32]*, [10 x i32]** [[B_ADDR]], align 4 -// CHECK6-NEXT: [[TMP3:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true) +// CHECK6-NEXT: [[TMP3:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true, i32 0) // CHECK6-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP3]], -1 // CHECK6-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK6: user_code.entry: diff --git a/clang/test/OpenMP/nvptx_target_parallel_num_threads_codegen.cpp b/clang/test/OpenMP/nvptx_target_parallel_num_threads_codegen.cpp --- a/clang/test/OpenMP/nvptx_target_parallel_num_threads_codegen.cpp +++ b/clang/test/OpenMP/nvptx_target_parallel_num_threads_codegen.cpp @@ -53,7 +53,7 @@ // CHECK1-NEXT: [[CAPTURED_VARS_ADDRS:%.*]] = alloca [1 x i8*], align 8 // CHECK1-NEXT: store i16* [[AA]], i16** [[AA_ADDR]], align 8 // CHECK1-NEXT: [[TMP0:%.*]] = load i16*, i16** [[AA_ADDR]], align 8 -// CHECK1-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 true) +// CHECK1-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 true, i32 0) // CHECK1-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK1-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK1: user_code.entry: @@ -103,7 +103,7 @@ // CHECK1-NEXT: [[TMP1:%.*]] = load i16*, i16** [[AA_ADDR]], align 8 // CHECK1-NEXT: [[TMP2:%.*]] = load [10 x i32]*, [10 x i32]** [[B_ADDR]], align 8 // CHECK1-NEXT: [[CONV:%.*]] = bitcast i64* [[DOTCAPTURE_EXPR__ADDR]] to i32* -// CHECK1-NEXT: [[TMP3:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true) +// CHECK1-NEXT: [[TMP3:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true, i32 0) // CHECK1-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP3]], -1 // CHECK1-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK1: user_code.entry: @@ -164,7 +164,7 @@ // CHECK2-NEXT: [[CAPTURED_VARS_ADDRS:%.*]] = alloca [1 x i8*], align 4 // CHECK2-NEXT: store i16* [[AA]], i16** [[AA_ADDR]], align 4 // CHECK2-NEXT: [[TMP0:%.*]] = load i16*, i16** [[AA_ADDR]], align 4 -// CHECK2-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 true) +// CHECK2-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 true, i32 0) // CHECK2-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK2-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK2: user_code.entry: @@ -213,7 +213,7 @@ // CHECK2-NEXT: [[TMP0:%.*]] = load i32*, i32** [[A_ADDR]], align 4 // CHECK2-NEXT: [[TMP1:%.*]] = load i16*, i16** [[AA_ADDR]], align 4 // CHECK2-NEXT: [[TMP2:%.*]] = load [10 x i32]*, [10 x i32]** [[B_ADDR]], align 4 -// CHECK2-NEXT: [[TMP3:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true) +// CHECK2-NEXT: [[TMP3:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true, i32 0) // CHECK2-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP3]], -1 // CHECK2-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK2: user_code.entry: @@ -274,7 +274,7 @@ // CHECK3-NEXT: [[CAPTURED_VARS_ADDRS:%.*]] = alloca [1 x i8*], align 4 // CHECK3-NEXT: store i16* [[AA]], i16** [[AA_ADDR]], align 4 // CHECK3-NEXT: [[TMP0:%.*]] = load i16*, i16** [[AA_ADDR]], align 4 -// CHECK3-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 true) +// CHECK3-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 true, i32 0) // CHECK3-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK3-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK3: user_code.entry: @@ -323,7 +323,7 @@ // CHECK3-NEXT: [[TMP0:%.*]] = load i32*, i32** [[A_ADDR]], align 4 // CHECK3-NEXT: [[TMP1:%.*]] = load i16*, i16** [[AA_ADDR]], align 4 // CHECK3-NEXT: [[TMP2:%.*]] = load [10 x i32]*, [10 x i32]** [[B_ADDR]], align 4 -// CHECK3-NEXT: [[TMP3:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true) +// CHECK3-NEXT: [[TMP3:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true, i32 0) // CHECK3-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP3]], -1 // CHECK3-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK3: user_code.entry: @@ -384,7 +384,7 @@ // CHECK4-NEXT: [[CAPTURED_VARS_ADDRS:%.*]] = alloca [1 x i8*], align 8 // CHECK4-NEXT: store i16* [[AA]], i16** [[AA_ADDR]], align 8 // CHECK4-NEXT: [[TMP0:%.*]] = load i16*, i16** [[AA_ADDR]], align 8 -// CHECK4-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 true) +// CHECK4-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 true, i32 0) // CHECK4-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK4-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK4: user_code.entry: @@ -434,7 +434,7 @@ // CHECK4-NEXT: [[TMP1:%.*]] = load i16*, i16** [[AA_ADDR]], align 8 // CHECK4-NEXT: [[TMP2:%.*]] = load [10 x i32]*, [10 x i32]** [[B_ADDR]], align 8 // CHECK4-NEXT: [[CONV:%.*]] = bitcast i64* [[DOTCAPTURE_EXPR__ADDR]] to i32* -// CHECK4-NEXT: [[TMP3:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true) +// CHECK4-NEXT: [[TMP3:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true, i32 0) // CHECK4-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP3]], -1 // CHECK4-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK4: user_code.entry: @@ -495,7 +495,7 @@ // CHECK5-NEXT: [[CAPTURED_VARS_ADDRS:%.*]] = alloca [1 x i8*], align 4 // CHECK5-NEXT: store i16* [[AA]], i16** [[AA_ADDR]], align 4 // CHECK5-NEXT: [[TMP0:%.*]] = load i16*, i16** [[AA_ADDR]], align 4 -// CHECK5-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 true) +// CHECK5-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 true, i32 0) // CHECK5-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK5-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK5: user_code.entry: @@ -544,7 +544,7 @@ // CHECK5-NEXT: [[TMP0:%.*]] = load i32*, i32** [[A_ADDR]], align 4 // CHECK5-NEXT: [[TMP1:%.*]] = load i16*, i16** [[AA_ADDR]], align 4 // CHECK5-NEXT: [[TMP2:%.*]] = load [10 x i32]*, [10 x i32]** [[B_ADDR]], align 4 -// CHECK5-NEXT: [[TMP3:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true) +// CHECK5-NEXT: [[TMP3:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true, i32 0) // CHECK5-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP3]], -1 // CHECK5-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK5: user_code.entry: @@ -605,7 +605,7 @@ // CHECK6-NEXT: [[CAPTURED_VARS_ADDRS:%.*]] = alloca [1 x i8*], align 4 // CHECK6-NEXT: store i16* [[AA]], i16** [[AA_ADDR]], align 4 // CHECK6-NEXT: [[TMP0:%.*]] = load i16*, i16** [[AA_ADDR]], align 4 -// CHECK6-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 true) +// CHECK6-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 true, i32 0) // CHECK6-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK6-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK6: user_code.entry: @@ -654,7 +654,7 @@ // CHECK6-NEXT: [[TMP0:%.*]] = load i32*, i32** [[A_ADDR]], align 4 // CHECK6-NEXT: [[TMP1:%.*]] = load i16*, i16** [[AA_ADDR]], align 4 // CHECK6-NEXT: [[TMP2:%.*]] = load [10 x i32]*, [10 x i32]** [[B_ADDR]], align 4 -// CHECK6-NEXT: [[TMP3:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true) +// CHECK6-NEXT: [[TMP3:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true, i32 0) // CHECK6-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP3]], -1 // CHECK6-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK6: user_code.entry: diff --git a/clang/test/OpenMP/nvptx_target_parallel_reduction_codegen.cpp b/clang/test/OpenMP/nvptx_target_parallel_reduction_codegen.cpp --- a/clang/test/OpenMP/nvptx_target_parallel_reduction_codegen.cpp +++ b/clang/test/OpenMP/nvptx_target_parallel_reduction_codegen.cpp @@ -54,7 +54,7 @@ // CHECK: define {{.*}}void {{@__omp_offloading_.+template.+l27}}( // -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) // CHECK: call void @__kmpc_target_deinit({{.*}}, i8 2, i1 true) // // @@ -233,7 +233,7 @@ // CHECK: define {{.*}}void {{@__omp_offloading_.+template.+l32}}( // -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) // CHECK: call void @__kmpc_target_deinit({{.*}}, i8 2, i1 true) // // @@ -490,7 +490,7 @@ // CHECK: define {{.*}}void {{@__omp_offloading_.+template.+l38}}( // -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true, i32 0) // CHECK: call void @__kmpc_target_deinit({{.*}}, i8 2, i1 true) // // diff --git a/clang/test/OpenMP/nvptx_target_parallel_reduction_codegen_tbaa_PR46146.cpp b/clang/test/OpenMP/nvptx_target_parallel_reduction_codegen_tbaa_PR46146.cpp --- a/clang/test/OpenMP/nvptx_target_parallel_reduction_codegen_tbaa_PR46146.cpp +++ b/clang/test/OpenMP/nvptx_target_parallel_reduction_codegen_tbaa_PR46146.cpp @@ -36,7 +36,7 @@ // CHECK1-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 // CHECK1-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 // CHECK1-NEXT: store i32 0, i32* [[DOTZERO_ADDR]], align 4 -// CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true) +// CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true, i32 0) // CHECK1-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 // CHECK1-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK1: user_code.entry: @@ -603,7 +603,7 @@ // CHECK1-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 // CHECK1-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 // CHECK1-NEXT: store i32 0, i32* [[DOTZERO_ADDR]], align 4 -// CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true) +// CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true, i32 0) // CHECK1-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 // CHECK1-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK1: user_code.entry: @@ -1270,7 +1270,7 @@ // CHECK2-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 // CHECK2-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 // CHECK2-NEXT: store i32 0, i32* [[DOTZERO_ADDR]], align 4 -// CHECK2-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true) +// CHECK2-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true, i32 0) // CHECK2-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 // CHECK2-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK2: user_code.entry: @@ -1837,7 +1837,7 @@ // CHECK2-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 // CHECK2-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 // CHECK2-NEXT: store i32 0, i32* [[DOTZERO_ADDR]], align 4 -// CHECK2-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true) +// CHECK2-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true, i32 0) // CHECK2-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 // CHECK2-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK2: user_code.entry: @@ -2504,7 +2504,7 @@ // CHECK3-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 // CHECK3-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 // CHECK3-NEXT: store i32 0, i32* [[DOTZERO_ADDR]], align 4 -// CHECK3-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true) +// CHECK3-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true, i32 0) // CHECK3-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 // CHECK3-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK3: user_code.entry: @@ -3071,7 +3071,7 @@ // CHECK3-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 // CHECK3-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 // CHECK3-NEXT: store i32 0, i32* [[DOTZERO_ADDR]], align 4 -// CHECK3-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true) +// CHECK3-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true, i32 0) // CHECK3-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 // CHECK3-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK3: user_code.entry: diff --git a/clang/test/OpenMP/nvptx_target_printf_codegen.c b/clang/test/OpenMP/nvptx_target_printf_codegen.c --- a/clang/test/OpenMP/nvptx_target_printf_codegen.c +++ b/clang/test/OpenMP/nvptx_target_printf_codegen.c @@ -48,7 +48,7 @@ // CHECK-64-NEXT: entry: // CHECK-64-NEXT: [[FMT:%.*]] = alloca i8*, align 8 // CHECK-64-NEXT: [[TMP:%.*]] = alloca [[PRINTF_ARGS:%.*]], align 8 -// CHECK-64-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true) +// CHECK-64-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true, i32 0) // CHECK-64-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 // CHECK-64-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK-64: user_code.entry: @@ -71,7 +71,7 @@ // CHECK-64-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}_CheckNoArgs_l25 // CHECK-64-SAME: () #[[ATTR0]] { // CHECK-64-NEXT: entry: -// CHECK-64-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true) +// CHECK-64-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true, i32 0) // CHECK-64-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 // CHECK-64-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK-64: user_code.entry: @@ -89,7 +89,7 @@ // CHECK-64-NEXT: [[TMP:%.*]] = alloca [[PRINTF_ARGS_0:%.*]], align 8 // CHECK-64-NEXT: store i64 [[FOO]], i64* [[FOO_ADDR]], align 8 // CHECK-64-NEXT: [[CONV:%.*]] = bitcast i64* [[FOO_ADDR]] to i32* -// CHECK-64-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true) +// CHECK-64-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true, i32 0) // CHECK-64-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 // CHECK-64-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK-64: user_code.entry: @@ -117,7 +117,7 @@ // CHECK-32-NEXT: entry: // CHECK-32-NEXT: [[FMT:%.*]] = alloca i8*, align 4 // CHECK-32-NEXT: [[TMP:%.*]] = alloca [[PRINTF_ARGS:%.*]], align 8 -// CHECK-32-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true) +// CHECK-32-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true, i32 0) // CHECK-32-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 // CHECK-32-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK-32: user_code.entry: @@ -140,7 +140,7 @@ // CHECK-32-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}_CheckNoArgs_l25 // CHECK-32-SAME: () #[[ATTR0]] { // CHECK-32-NEXT: entry: -// CHECK-32-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true) +// CHECK-32-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true, i32 0) // CHECK-32-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 // CHECK-32-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK-32: user_code.entry: @@ -157,7 +157,7 @@ // CHECK-32-NEXT: [[FOO_ADDR:%.*]] = alloca i32, align 4 // CHECK-32-NEXT: [[TMP:%.*]] = alloca [[PRINTF_ARGS_0:%.*]], align 8 // CHECK-32-NEXT: store i32 [[FOO]], i32* [[FOO_ADDR]], align 4 -// CHECK-32-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true) +// CHECK-32-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true, i32 0) // CHECK-32-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 // CHECK-32-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK-32: user_code.entry: diff --git a/clang/test/OpenMP/nvptx_target_simd_codegen.cpp b/clang/test/OpenMP/nvptx_target_simd_codegen.cpp --- a/clang/test/OpenMP/nvptx_target_simd_codegen.cpp +++ b/clang/test/OpenMP/nvptx_target_simd_codegen.cpp @@ -61,28 +61,28 @@ } // CHECK-LABEL: define {{.*}}void {{@__omp_offloading_.+l32}}( -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false, i32 0) // CHECK-NOT: call void @__kmpc_for_static_init // CHECK-NOT: call void @__kmpc_for_static_fini // CHECK: call void @__kmpc_target_deinit({{.*}}, i8 2, i1 false) // CHECK: ret void // CHECK-LABEL: define {{.*}}void {{@__omp_offloading_.+l37}}( -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false, i32 0) // CHECK-NOT: call void @__kmpc_for_static_init // CHECK-NOT: call void @__kmpc_for_static_fini // CHECK: call void @__kmpc_target_deinit({{.*}}, i8 2, i1 false) // CHECK: ret void // CHECK-LABEL: define {{.*}}void {{@__omp_offloading_.+l42}}( -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false, i32 0) // CHECK-NOT: call void @__kmpc_for_static_init // CHECK-NOT: call void @__kmpc_for_static_fini // CHECK: call void @__kmpc_target_deinit({{.*}}, i8 2, i1 false) // CHECK: ret void // CHECK-LABEL: define {{.*}}void {{@__omp_offloading_.+l47}}( -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false, i32 0) // CHECK-NOT: call void @__kmpc_for_static_init // CHECK-NOT: call void @__kmpc_for_static_fini // CHECK-NOT: call void @__kmpc_nvptx_end_reduce_nowait( diff --git a/clang/test/OpenMP/nvptx_target_teams_codegen.cpp b/clang/test/OpenMP/nvptx_target_teams_codegen.cpp --- a/clang/test/OpenMP/nvptx_target_teams_codegen.cpp +++ b/clang/test/OpenMP/nvptx_target_teams_codegen.cpp @@ -59,7 +59,7 @@ // CHECK1-NEXT: store i32 0, i32* [[DOTZERO_ADDR]], align 4 // CHECK1-NEXT: store i64 [[A]], i64* [[A_ADDR]], align 8 // CHECK1-NEXT: [[CONV:%.*]] = bitcast i64* [[A_ADDR]] to i8* -// CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true) +// CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true, i32 0) // CHECK1-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 // CHECK1-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK1: user_code.entry: @@ -100,7 +100,7 @@ // CHECK1-NEXT: store i32 0, i32* [[DOTZERO_ADDR]], align 4 // CHECK1-NEXT: store i64 [[AA]], i64* [[AA_ADDR]], align 8 // CHECK1-NEXT: [[CONV:%.*]] = bitcast i64* [[AA_ADDR]] to i16* -// CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true) +// CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true, i32 0) // CHECK1-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 // CHECK1-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK1: user_code.entry: @@ -141,7 +141,7 @@ // CHECK1-NEXT: store i32 0, i32* [[DOTZERO_ADDR]], align 4 // CHECK1-NEXT: store i64 [[AA]], i64* [[AA_ADDR]], align 8 // CHECK1-NEXT: [[CONV:%.*]] = bitcast i64* [[AA_ADDR]] to i16* -// CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true) +// CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true, i32 0) // CHECK1-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 // CHECK1-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK1: user_code.entry: @@ -224,7 +224,7 @@ // CHECK2-NEXT: store i32 0, i32* [[DOTZERO_ADDR]], align 4 // CHECK2-NEXT: store i32 [[A]], i32* [[A_ADDR]], align 4 // CHECK2-NEXT: [[CONV:%.*]] = bitcast i32* [[A_ADDR]] to i8* -// CHECK2-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true) +// CHECK2-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true, i32 0) // CHECK2-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 // CHECK2-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK2: user_code.entry: @@ -265,7 +265,7 @@ // CHECK2-NEXT: store i32 0, i32* [[DOTZERO_ADDR]], align 4 // CHECK2-NEXT: store i32 [[AA]], i32* [[AA_ADDR]], align 4 // CHECK2-NEXT: [[CONV:%.*]] = bitcast i32* [[AA_ADDR]] to i16* -// CHECK2-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true) +// CHECK2-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true, i32 0) // CHECK2-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 // CHECK2-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK2: user_code.entry: @@ -306,7 +306,7 @@ // CHECK2-NEXT: store i32 0, i32* [[DOTZERO_ADDR]], align 4 // CHECK2-NEXT: store i32 [[AA]], i32* [[AA_ADDR]], align 4 // CHECK2-NEXT: [[CONV:%.*]] = bitcast i32* [[AA_ADDR]] to i16* -// CHECK2-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true) +// CHECK2-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true, i32 0) // CHECK2-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 // CHECK2-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK2: user_code.entry: @@ -389,7 +389,7 @@ // CHECK3-NEXT: store i32 0, i32* [[DOTZERO_ADDR]], align 4 // CHECK3-NEXT: store i32 [[A]], i32* [[A_ADDR]], align 4 // CHECK3-NEXT: [[CONV:%.*]] = bitcast i32* [[A_ADDR]] to i8* -// CHECK3-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true) +// CHECK3-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true, i32 0) // CHECK3-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 // CHECK3-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK3: user_code.entry: @@ -430,7 +430,7 @@ // CHECK3-NEXT: store i32 0, i32* [[DOTZERO_ADDR]], align 4 // CHECK3-NEXT: store i32 [[AA]], i32* [[AA_ADDR]], align 4 // CHECK3-NEXT: [[CONV:%.*]] = bitcast i32* [[AA_ADDR]] to i16* -// CHECK3-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true) +// CHECK3-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true, i32 0) // CHECK3-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 // CHECK3-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK3: user_code.entry: @@ -471,7 +471,7 @@ // CHECK3-NEXT: store i32 0, i32* [[DOTZERO_ADDR]], align 4 // CHECK3-NEXT: store i32 [[AA]], i32* [[AA_ADDR]], align 4 // CHECK3-NEXT: [[CONV:%.*]] = bitcast i32* [[AA_ADDR]] to i16* -// CHECK3-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true) +// CHECK3-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true, i32 0) // CHECK3-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 // CHECK3-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK3: user_code.entry: diff --git a/clang/test/OpenMP/nvptx_target_teams_distribute_codegen.cpp b/clang/test/OpenMP/nvptx_target_teams_distribute_codegen.cpp --- a/clang/test/OpenMP/nvptx_target_teams_distribute_codegen.cpp +++ b/clang/test/OpenMP/nvptx_target_teams_distribute_codegen.cpp @@ -580,7 +580,7 @@ // CHECK1-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 // CHECK1-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 // CHECK1-NEXT: store i32 0, i32* [[DOTZERO_ADDR]], align 4 -// CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true) +// CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true, i32 0) // CHECK1-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 // CHECK1-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK1: user_code.entry: @@ -702,7 +702,7 @@ // CHECK2-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 // CHECK2-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 // CHECK2-NEXT: store i32 0, i32* [[DOTZERO_ADDR]], align 4 -// CHECK2-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true) +// CHECK2-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true, i32 0) // CHECK2-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 // CHECK2-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK2: user_code.entry: @@ -824,7 +824,7 @@ // CHECK3-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 // CHECK3-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 // CHECK3-NEXT: store i32 0, i32* [[DOTZERO_ADDR]], align 4 -// CHECK3-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true) +// CHECK3-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true, i32 0) // CHECK3-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 // CHECK3-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK3: user_code.entry: diff --git a/clang/test/OpenMP/nvptx_target_teams_distribute_parallel_for_codegen.cpp b/clang/test/OpenMP/nvptx_target_teams_distribute_parallel_for_codegen.cpp --- a/clang/test/OpenMP/nvptx_target_teams_distribute_parallel_for_codegen.cpp +++ b/clang/test/OpenMP/nvptx_target_teams_distribute_parallel_for_codegen.cpp @@ -18484,7 +18484,7 @@ // CHECK1-NEXT: [[CONV:%.*]] = bitcast i64* [[N_ADDR]] to i32* // CHECK1-NEXT: [[TMP0:%.*]] = load [1000 x i32]*, [1000 x i32]** [[A_ADDR]], align 8 // CHECK1-NEXT: [[CONV1:%.*]] = bitcast i64* [[L_ADDR]] to i32* -// CHECK1-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 false) +// CHECK1-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 false, i32 0) // CHECK1-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK1-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK1: user_code.entry: @@ -18807,7 +18807,7 @@ // CHECK1-NEXT: store [1000 x i16]* [[AA]], [1000 x i16]** [[AA_ADDR]], align 8 // CHECK1-NEXT: [[CONV:%.*]] = bitcast i64* [[N_ADDR]] to i32* // CHECK1-NEXT: [[TMP0:%.*]] = load [1000 x i16]*, [1000 x i16]** [[AA_ADDR]], align 8 -// CHECK1-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false) +// CHECK1-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false, i32 0) // CHECK1-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK1-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK1: user_code.entry: @@ -19062,7 +19062,7 @@ // CHECK1-NEXT: store i32 0, i32* [[DOTZERO_ADDR]], align 4 // CHECK1-NEXT: store [10 x i32]* [[B]], [10 x i32]** [[B_ADDR]], align 8 // CHECK1-NEXT: [[TMP0:%.*]] = load [10 x i32]*, [10 x i32]** [[B_ADDR]], align 8 -// CHECK1-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false) +// CHECK1-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false, i32 0) // CHECK1-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK1-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK1: user_code.entry: @@ -19253,7 +19253,7 @@ // CHECK1-NEXT: store i64 [[F]], i64* [[F_ADDR]], align 8 // CHECK1-NEXT: [[TMP0:%.*]] = load [10 x [10 x i32]]*, [10 x [10 x i32]]** [[C_ADDR]], align 8 // CHECK1-NEXT: [[CONV:%.*]] = bitcast i64* [[F_ADDR]] to i32* -// CHECK1-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false) +// CHECK1-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false, i32 0) // CHECK1-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK1-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK1: user_code.entry: @@ -19486,7 +19486,7 @@ // CHECK1-NEXT: store [10 x [10 x i32]]* [[C]], [10 x [10 x i32]]** [[C_ADDR]], align 8 // CHECK1-NEXT: [[CONV:%.*]] = bitcast i64* [[N_ADDR]] to i32* // CHECK1-NEXT: [[TMP0:%.*]] = load [10 x [10 x i32]]*, [10 x [10 x i32]]** [[C_ADDR]], align 8 -// CHECK1-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false) +// CHECK1-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false, i32 0) // CHECK1-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK1-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK1: user_code.entry: @@ -19805,7 +19805,7 @@ // CHECK1-NEXT: store i32* [[V]], i32** [[V_ADDR]], align 8 // CHECK1-NEXT: [[CONV:%.*]] = bitcast i64* [[N_ADDR]] to i32* // CHECK1-NEXT: [[TMP0:%.*]] = load [1000 x i32]*, [1000 x i32]** [[A_ADDR]], align 8 -// CHECK1-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false) +// CHECK1-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false, i32 0) // CHECK1-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK1-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK1: user_code.entry: @@ -20078,7 +20078,7 @@ // CHECK2-NEXT: [[CONV:%.*]] = bitcast i64* [[N_ADDR]] to i32* // CHECK2-NEXT: [[TMP0:%.*]] = load [1000 x i32]*, [1000 x i32]** [[A_ADDR]], align 8 // CHECK2-NEXT: [[CONV1:%.*]] = bitcast i64* [[L_ADDR]] to i32* -// CHECK2-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 false) +// CHECK2-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 false, i32 0) // CHECK2-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK2-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK2: user_code.entry: @@ -20401,7 +20401,7 @@ // CHECK2-NEXT: store [1000 x i16]* [[AA]], [1000 x i16]** [[AA_ADDR]], align 8 // CHECK2-NEXT: [[CONV:%.*]] = bitcast i64* [[N_ADDR]] to i32* // CHECK2-NEXT: [[TMP0:%.*]] = load [1000 x i16]*, [1000 x i16]** [[AA_ADDR]], align 8 -// CHECK2-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false) +// CHECK2-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false, i32 0) // CHECK2-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK2-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK2: user_code.entry: @@ -20656,7 +20656,7 @@ // CHECK2-NEXT: store i32 0, i32* [[DOTZERO_ADDR]], align 4 // CHECK2-NEXT: store [10 x i32]* [[B]], [10 x i32]** [[B_ADDR]], align 8 // CHECK2-NEXT: [[TMP0:%.*]] = load [10 x i32]*, [10 x i32]** [[B_ADDR]], align 8 -// CHECK2-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false) +// CHECK2-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false, i32 0) // CHECK2-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK2-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK2: user_code.entry: @@ -20847,7 +20847,7 @@ // CHECK2-NEXT: store i64 [[F]], i64* [[F_ADDR]], align 8 // CHECK2-NEXT: [[TMP0:%.*]] = load [10 x [10 x i32]]*, [10 x [10 x i32]]** [[C_ADDR]], align 8 // CHECK2-NEXT: [[CONV:%.*]] = bitcast i64* [[F_ADDR]] to i32* -// CHECK2-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false) +// CHECK2-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false, i32 0) // CHECK2-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK2-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK2: user_code.entry: @@ -21080,7 +21080,7 @@ // CHECK2-NEXT: store [10 x [10 x i32]]* [[C]], [10 x [10 x i32]]** [[C_ADDR]], align 8 // CHECK2-NEXT: [[CONV:%.*]] = bitcast i64* [[N_ADDR]] to i32* // CHECK2-NEXT: [[TMP0:%.*]] = load [10 x [10 x i32]]*, [10 x [10 x i32]]** [[C_ADDR]], align 8 -// CHECK2-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false) +// CHECK2-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false, i32 0) // CHECK2-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK2-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK2: user_code.entry: @@ -21394,7 +21394,7 @@ // CHECK2-NEXT: store i32* [[V]], i32** [[V_ADDR]], align 8 // CHECK2-NEXT: [[CONV:%.*]] = bitcast i64* [[N_ADDR]] to i32* // CHECK2-NEXT: [[TMP0:%.*]] = load [1000 x i32]*, [1000 x i32]** [[A_ADDR]], align 8 -// CHECK2-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false) +// CHECK2-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false, i32 0) // CHECK2-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK2-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK2: user_code.entry: @@ -21665,7 +21665,7 @@ // CHECK3-NEXT: store [1000 x i32]* [[A]], [1000 x i32]** [[A_ADDR]], align 4 // CHECK3-NEXT: store i32 [[L]], i32* [[L_ADDR]], align 4 // CHECK3-NEXT: [[TMP0:%.*]] = load [1000 x i32]*, [1000 x i32]** [[A_ADDR]], align 4 -// CHECK3-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 false) +// CHECK3-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 false, i32 0) // CHECK3-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK3-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK3: user_code.entry: @@ -21972,7 +21972,7 @@ // CHECK3-NEXT: store i32 [[N]], i32* [[N_ADDR]], align 4 // CHECK3-NEXT: store [1000 x i16]* [[AA]], [1000 x i16]** [[AA_ADDR]], align 4 // CHECK3-NEXT: [[TMP0:%.*]] = load [1000 x i16]*, [1000 x i16]** [[AA_ADDR]], align 4 -// CHECK3-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false) +// CHECK3-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false, i32 0) // CHECK3-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK3-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK3: user_code.entry: @@ -22217,7 +22217,7 @@ // CHECK3-NEXT: store i32 0, i32* [[DOTZERO_ADDR]], align 4 // CHECK3-NEXT: store [10 x i32]* [[B]], [10 x i32]** [[B_ADDR]], align 4 // CHECK3-NEXT: [[TMP0:%.*]] = load [10 x i32]*, [10 x i32]** [[B_ADDR]], align 4 -// CHECK3-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false) +// CHECK3-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false, i32 0) // CHECK3-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK3-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK3: user_code.entry: @@ -22401,7 +22401,7 @@ // CHECK3-NEXT: store [10 x [10 x i32]]* [[C]], [10 x [10 x i32]]** [[C_ADDR]], align 4 // CHECK3-NEXT: store i32 [[F]], i32* [[F_ADDR]], align 4 // CHECK3-NEXT: [[TMP0:%.*]] = load [10 x [10 x i32]]*, [10 x [10 x i32]]** [[C_ADDR]], align 4 -// CHECK3-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false) +// CHECK3-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false, i32 0) // CHECK3-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK3-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK3: user_code.entry: @@ -22622,7 +22622,7 @@ // CHECK3-NEXT: store i32 [[N]], i32* [[N_ADDR]], align 4 // CHECK3-NEXT: store [10 x [10 x i32]]* [[C]], [10 x [10 x i32]]** [[C_ADDR]], align 4 // CHECK3-NEXT: [[TMP0:%.*]] = load [10 x [10 x i32]]*, [10 x [10 x i32]]** [[C_ADDR]], align 4 -// CHECK3-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false) +// CHECK3-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false, i32 0) // CHECK3-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK3-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK3: user_code.entry: @@ -22939,7 +22939,7 @@ // CHECK3-NEXT: store [1000 x i32]* [[A]], [1000 x i32]** [[A_ADDR]], align 4 // CHECK3-NEXT: store i32* [[V]], i32** [[V_ADDR]], align 4 // CHECK3-NEXT: [[TMP0:%.*]] = load [1000 x i32]*, [1000 x i32]** [[A_ADDR]], align 4 -// CHECK3-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false) +// CHECK3-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false, i32 0) // CHECK3-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK3-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK3: user_code.entry: @@ -23199,7 +23199,7 @@ // CHECK4-NEXT: store [1000 x i32]* [[A]], [1000 x i32]** [[A_ADDR]], align 4 // CHECK4-NEXT: store i32 [[L]], i32* [[L_ADDR]], align 4 // CHECK4-NEXT: [[TMP0:%.*]] = load [1000 x i32]*, [1000 x i32]** [[A_ADDR]], align 4 -// CHECK4-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 false) +// CHECK4-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 false, i32 0) // CHECK4-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK4-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK4: user_code.entry: @@ -23506,7 +23506,7 @@ // CHECK4-NEXT: store i32 [[N]], i32* [[N_ADDR]], align 4 // CHECK4-NEXT: store [1000 x i16]* [[AA]], [1000 x i16]** [[AA_ADDR]], align 4 // CHECK4-NEXT: [[TMP0:%.*]] = load [1000 x i16]*, [1000 x i16]** [[AA_ADDR]], align 4 -// CHECK4-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false) +// CHECK4-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false, i32 0) // CHECK4-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK4-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK4: user_code.entry: @@ -23751,7 +23751,7 @@ // CHECK4-NEXT: store i32 0, i32* [[DOTZERO_ADDR]], align 4 // CHECK4-NEXT: store [10 x i32]* [[B]], [10 x i32]** [[B_ADDR]], align 4 // CHECK4-NEXT: [[TMP0:%.*]] = load [10 x i32]*, [10 x i32]** [[B_ADDR]], align 4 -// CHECK4-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false) +// CHECK4-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false, i32 0) // CHECK4-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK4-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK4: user_code.entry: @@ -23935,7 +23935,7 @@ // CHECK4-NEXT: store [10 x [10 x i32]]* [[C]], [10 x [10 x i32]]** [[C_ADDR]], align 4 // CHECK4-NEXT: store i32 [[F]], i32* [[F_ADDR]], align 4 // CHECK4-NEXT: [[TMP0:%.*]] = load [10 x [10 x i32]]*, [10 x [10 x i32]]** [[C_ADDR]], align 4 -// CHECK4-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false) +// CHECK4-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false, i32 0) // CHECK4-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK4-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK4: user_code.entry: @@ -24156,7 +24156,7 @@ // CHECK4-NEXT: store i32 [[N]], i32* [[N_ADDR]], align 4 // CHECK4-NEXT: store [10 x [10 x i32]]* [[C]], [10 x [10 x i32]]** [[C_ADDR]], align 4 // CHECK4-NEXT: [[TMP0:%.*]] = load [10 x [10 x i32]]*, [10 x [10 x i32]]** [[C_ADDR]], align 4 -// CHECK4-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false) +// CHECK4-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false, i32 0) // CHECK4-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK4-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK4: user_code.entry: @@ -24473,7 +24473,7 @@ // CHECK4-NEXT: store [1000 x i32]* [[A]], [1000 x i32]** [[A_ADDR]], align 4 // CHECK4-NEXT: store i32* [[V]], i32** [[V_ADDR]], align 4 // CHECK4-NEXT: [[TMP0:%.*]] = load [1000 x i32]*, [1000 x i32]** [[A_ADDR]], align 4 -// CHECK4-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false) +// CHECK4-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false, i32 0) // CHECK4-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK4-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK4: user_code.entry: diff --git a/clang/test/OpenMP/nvptx_target_teams_distribute_parallel_for_generic_mode_codegen.cpp b/clang/test/OpenMP/nvptx_target_teams_distribute_parallel_for_generic_mode_codegen.cpp --- a/clang/test/OpenMP/nvptx_target_teams_distribute_parallel_for_generic_mode_codegen.cpp +++ b/clang/test/OpenMP/nvptx_target_teams_distribute_parallel_for_generic_mode_codegen.cpp @@ -46,7 +46,7 @@ // CHECK1-NEXT: [[CONV:%.*]] = bitcast i64* [[ARGC_ADDR]] to i32* // CHECK1-NEXT: [[TMP0:%.*]] = load i32*, i32** [[A_ADDR]], align 8 // CHECK1-NEXT: [[CONV1:%.*]] = bitcast i64* [[DOTCAPTURE_EXPR__ADDR]] to i32* -// CHECK1-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 false) +// CHECK1-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 false, i32 0) // CHECK1-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK1-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK1: user_code.entry: @@ -354,7 +354,7 @@ // CHECK2-NEXT: store i32* [[A]], i32** [[A_ADDR]], align 4 // CHECK2-NEXT: store i32 [[DOTCAPTURE_EXPR_]], i32* [[DOTCAPTURE_EXPR__ADDR]], align 4 // CHECK2-NEXT: [[TMP0:%.*]] = load i32*, i32** [[A_ADDR]], align 4 -// CHECK2-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 false) +// CHECK2-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 false, i32 0) // CHECK2-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK2-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK2: user_code.entry: @@ -648,7 +648,7 @@ // CHECK3-NEXT: store i32* [[A]], i32** [[A_ADDR]], align 4 // CHECK3-NEXT: store i32 [[DOTCAPTURE_EXPR_]], i32* [[DOTCAPTURE_EXPR__ADDR]], align 4 // CHECK3-NEXT: [[TMP0:%.*]] = load i32*, i32** [[A_ADDR]], align 4 -// CHECK3-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 false) +// CHECK3-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 false, i32 0) // CHECK3-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK3-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK3: user_code.entry: @@ -944,7 +944,7 @@ // CHECK4-NEXT: [[CONV:%.*]] = bitcast i64* [[ARGC_ADDR]] to i32* // CHECK4-NEXT: [[TMP0:%.*]] = load i32*, i32** [[A_ADDR]], align 8 // CHECK4-NEXT: [[CONV1:%.*]] = bitcast i64* [[DOTCAPTURE_EXPR__ADDR]] to i32* -// CHECK4-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 false) +// CHECK4-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 false, i32 0) // CHECK4-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK4-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK4: user_code.entry: @@ -1252,7 +1252,7 @@ // CHECK5-NEXT: store i32* [[A]], i32** [[A_ADDR]], align 4 // CHECK5-NEXT: store i32 [[DOTCAPTURE_EXPR_]], i32* [[DOTCAPTURE_EXPR__ADDR]], align 4 // CHECK5-NEXT: [[TMP0:%.*]] = load i32*, i32** [[A_ADDR]], align 4 -// CHECK5-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 false) +// CHECK5-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 false, i32 0) // CHECK5-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK5-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK5: user_code.entry: @@ -1546,7 +1546,7 @@ // CHECK6-NEXT: store i32* [[A]], i32** [[A_ADDR]], align 4 // CHECK6-NEXT: store i32 [[DOTCAPTURE_EXPR_]], i32* [[DOTCAPTURE_EXPR__ADDR]], align 4 // CHECK6-NEXT: [[TMP0:%.*]] = load i32*, i32** [[A_ADDR]], align 4 -// CHECK6-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 false) +// CHECK6-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 false, i32 0) // CHECK6-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK6-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK6: user_code.entry: diff --git a/clang/test/OpenMP/nvptx_target_teams_distribute_parallel_for_simd_codegen.cpp b/clang/test/OpenMP/nvptx_target_teams_distribute_parallel_for_simd_codegen.cpp --- a/clang/test/OpenMP/nvptx_target_teams_distribute_parallel_for_simd_codegen.cpp +++ b/clang/test/OpenMP/nvptx_target_teams_distribute_parallel_for_simd_codegen.cpp @@ -9397,7 +9397,7 @@ // CHECK1-NEXT: [[CONV:%.*]] = bitcast i64* [[N_ADDR]] to i32* // CHECK1-NEXT: [[TMP0:%.*]] = load [1000 x i32]*, [1000 x i32]** [[A_ADDR]], align 8 // CHECK1-NEXT: [[CONV1:%.*]] = bitcast i64* [[L_ADDR]] to i32* -// CHECK1-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 false) +// CHECK1-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 false, i32 0) // CHECK1-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK1-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK1: user_code.entry: @@ -9744,7 +9744,7 @@ // CHECK1-NEXT: store [1000 x i16]* [[AA]], [1000 x i16]** [[AA_ADDR]], align 8 // CHECK1-NEXT: [[CONV:%.*]] = bitcast i64* [[N_ADDR]] to i32* // CHECK1-NEXT: [[TMP0:%.*]] = load [1000 x i16]*, [1000 x i16]** [[AA_ADDR]], align 8 -// CHECK1-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false) +// CHECK1-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false, i32 0) // CHECK1-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK1-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK1: user_code.entry: @@ -10023,7 +10023,7 @@ // CHECK1-NEXT: store i32 0, i32* [[DOTZERO_ADDR]], align 4 // CHECK1-NEXT: store [10 x i32]* [[B]], [10 x i32]** [[B_ADDR]], align 8 // CHECK1-NEXT: [[TMP0:%.*]] = load [10 x i32]*, [10 x i32]** [[B_ADDR]], align 8 -// CHECK1-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false) +// CHECK1-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false, i32 0) // CHECK1-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK1-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK1: user_code.entry: @@ -10228,7 +10228,7 @@ // CHECK1-NEXT: store i64 [[F]], i64* [[F_ADDR]], align 8 // CHECK1-NEXT: [[TMP0:%.*]] = load [10 x [10 x i32]]*, [10 x [10 x i32]]** [[C_ADDR]], align 8 // CHECK1-NEXT: [[CONV:%.*]] = bitcast i64* [[F_ADDR]] to i32* -// CHECK1-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false) +// CHECK1-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false, i32 0) // CHECK1-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK1-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK1: user_code.entry: @@ -10479,7 +10479,7 @@ // CHECK2-NEXT: store [1000 x i32]* [[A]], [1000 x i32]** [[A_ADDR]], align 4 // CHECK2-NEXT: store i32 [[L]], i32* [[L_ADDR]], align 4 // CHECK2-NEXT: [[TMP0:%.*]] = load [1000 x i32]*, [1000 x i32]** [[A_ADDR]], align 4 -// CHECK2-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 false) +// CHECK2-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 false, i32 0) // CHECK2-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK2-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK2: user_code.entry: @@ -10810,7 +10810,7 @@ // CHECK2-NEXT: store i32 [[N]], i32* [[N_ADDR]], align 4 // CHECK2-NEXT: store [1000 x i16]* [[AA]], [1000 x i16]** [[AA_ADDR]], align 4 // CHECK2-NEXT: [[TMP0:%.*]] = load [1000 x i16]*, [1000 x i16]** [[AA_ADDR]], align 4 -// CHECK2-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false) +// CHECK2-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false, i32 0) // CHECK2-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK2-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK2: user_code.entry: @@ -11079,7 +11079,7 @@ // CHECK2-NEXT: store i32 0, i32* [[DOTZERO_ADDR]], align 4 // CHECK2-NEXT: store [10 x i32]* [[B]], [10 x i32]** [[B_ADDR]], align 4 // CHECK2-NEXT: [[TMP0:%.*]] = load [10 x i32]*, [10 x i32]** [[B_ADDR]], align 4 -// CHECK2-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false) +// CHECK2-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false, i32 0) // CHECK2-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK2-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK2: user_code.entry: @@ -11277,7 +11277,7 @@ // CHECK2-NEXT: store [10 x [10 x i32]]* [[C]], [10 x [10 x i32]]** [[C_ADDR]], align 4 // CHECK2-NEXT: store i32 [[F]], i32* [[F_ADDR]], align 4 // CHECK2-NEXT: [[TMP0:%.*]] = load [10 x [10 x i32]]*, [10 x [10 x i32]]** [[C_ADDR]], align 4 -// CHECK2-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false) +// CHECK2-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false, i32 0) // CHECK2-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK2-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK2: user_code.entry: @@ -11517,7 +11517,7 @@ // CHECK3-NEXT: store [1000 x i32]* [[A]], [1000 x i32]** [[A_ADDR]], align 4 // CHECK3-NEXT: store i32 [[L]], i32* [[L_ADDR]], align 4 // CHECK3-NEXT: [[TMP0:%.*]] = load [1000 x i32]*, [1000 x i32]** [[A_ADDR]], align 4 -// CHECK3-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 false) +// CHECK3-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 false, i32 0) // CHECK3-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK3-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK3: user_code.entry: @@ -11848,7 +11848,7 @@ // CHECK3-NEXT: store i32 [[N]], i32* [[N_ADDR]], align 4 // CHECK3-NEXT: store [1000 x i16]* [[AA]], [1000 x i16]** [[AA_ADDR]], align 4 // CHECK3-NEXT: [[TMP0:%.*]] = load [1000 x i16]*, [1000 x i16]** [[AA_ADDR]], align 4 -// CHECK3-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false) +// CHECK3-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false, i32 0) // CHECK3-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK3-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK3: user_code.entry: @@ -12117,7 +12117,7 @@ // CHECK3-NEXT: store i32 0, i32* [[DOTZERO_ADDR]], align 4 // CHECK3-NEXT: store [10 x i32]* [[B]], [10 x i32]** [[B_ADDR]], align 4 // CHECK3-NEXT: [[TMP0:%.*]] = load [10 x i32]*, [10 x i32]** [[B_ADDR]], align 4 -// CHECK3-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false) +// CHECK3-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false, i32 0) // CHECK3-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK3-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK3: user_code.entry: @@ -12315,7 +12315,7 @@ // CHECK3-NEXT: store [10 x [10 x i32]]* [[C]], [10 x [10 x i32]]** [[C_ADDR]], align 4 // CHECK3-NEXT: store i32 [[F]], i32* [[F_ADDR]], align 4 // CHECK3-NEXT: [[TMP0:%.*]] = load [10 x [10 x i32]]*, [10 x [10 x i32]]** [[C_ADDR]], align 4 -// CHECK3-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false) +// CHECK3-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false, i32 0) // CHECK3-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK3-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK3: user_code.entry: diff --git a/clang/test/OpenMP/nvptx_target_teams_distribute_simd_codegen.cpp b/clang/test/OpenMP/nvptx_target_teams_distribute_simd_codegen.cpp --- a/clang/test/OpenMP/nvptx_target_teams_distribute_simd_codegen.cpp +++ b/clang/test/OpenMP/nvptx_target_teams_distribute_simd_codegen.cpp @@ -70,7 +70,7 @@ } // CHECK-LABEL: define {{.*}}void {{@__omp_offloading_.+}}_l37( -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false, i32 0) // CHECK: call void @__kmpc_target_deinit({{.*}}, i8 2, i1 false) // CHECK: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, {{.+}} 91, @@ -78,7 +78,7 @@ // CHECK: ret void // CHECK-LABEL: define {{.*}}void {{@__omp_offloading_.+}}_l43( -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false, i32 0) // CHECK: call void @__kmpc_target_deinit({{.*}}, i8 2, i1 false) // CHECK: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, {{.+}} 91, @@ -86,7 +86,7 @@ // CHECK: ret void // CHECK-LABEL: define {{.*}}void {{@__omp_offloading_.+}}_l48( -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false, i32 0) // CHECK: call void @__kmpc_target_deinit({{.*}}, i8 2, i1 false) // CHECK: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, {{.+}} 91, @@ -95,7 +95,7 @@ // CHECK: define {{.*}}void {{@__omp_offloading_.+}}_l53({{.+}}, i{{32|64}} [[F_IN:%.+]]) // CHECK: store {{.+}} [[F_IN]], {{.+}}* {{.+}}, -// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false) +// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false, i32 0) // CHECK: call void @__kmpc_target_deinit({{.*}}, i8 2, i1 false) // CHECK: store {{.+}} 99, {{.+}}* [[COMB_UB:%.+]], align diff --git a/clang/test/OpenMP/nvptx_teams_codegen.cpp b/clang/test/OpenMP/nvptx_teams_codegen.cpp --- a/clang/test/OpenMP/nvptx_teams_codegen.cpp +++ b/clang/test/OpenMP/nvptx_teams_codegen.cpp @@ -899,7 +899,7 @@ // CHECK1-NEXT: store i32 0, i32* [[DOTZERO_ADDR]], align 4 // CHECK1-NEXT: store i64 [[ARGC]], i64* [[ARGC_ADDR]], align 8 // CHECK1-NEXT: [[CONV:%.*]] = bitcast i64* [[ARGC_ADDR]] to i32* -// CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true) +// CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true, i32 0) // CHECK1-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 // CHECK1-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK1: user_code.entry: @@ -939,7 +939,7 @@ // CHECK1-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 // CHECK1-NEXT: store i32 0, i32* [[DOTZERO_ADDR]], align 4 // CHECK1-NEXT: store i8** [[ARGC]], i8*** [[ARGC_ADDR]], align 8 -// CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true) +// CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true, i32 0) // CHECK1-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 // CHECK1-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK1: user_code.entry: @@ -979,7 +979,7 @@ // CHECK2-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 // CHECK2-NEXT: store i32 0, i32* [[DOTZERO_ADDR]], align 4 // CHECK2-NEXT: store i32 [[ARGC]], i32* [[ARGC_ADDR]], align 4 -// CHECK2-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true) +// CHECK2-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true, i32 0) // CHECK2-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 // CHECK2-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK2: user_code.entry: @@ -1019,7 +1019,7 @@ // CHECK2-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 // CHECK2-NEXT: store i32 0, i32* [[DOTZERO_ADDR]], align 4 // CHECK2-NEXT: store i8** [[ARGC]], i8*** [[ARGC_ADDR]], align 4 -// CHECK2-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true) +// CHECK2-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true, i32 0) // CHECK2-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 // CHECK2-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK2: user_code.entry: @@ -1066,7 +1066,7 @@ // CHECK3-NEXT: [[CONV:%.*]] = bitcast i64* [[A_ADDR]] to i32* // CHECK3-NEXT: [[CONV1:%.*]] = bitcast i64* [[B_ADDR]] to i32* // CHECK3-NEXT: [[CONV2:%.*]] = bitcast i64* [[ARGC_ADDR]] to i32* -// CHECK3-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true) +// CHECK3-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true, i32 0) // CHECK3-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 // CHECK3-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK3: user_code.entry: @@ -1112,7 +1112,7 @@ // CHECK3-NEXT: store i8** [[ARGC]], i8*** [[ARGC_ADDR]], align 8 // CHECK3-NEXT: [[CONV:%.*]] = bitcast i64* [[A_ADDR]] to i32* // CHECK3-NEXT: [[CONV1:%.*]] = bitcast i64* [[B_ADDR]] to i32* -// CHECK3-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true) +// CHECK3-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true, i32 0) // CHECK3-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 // CHECK3-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK3: user_code.entry: @@ -1156,7 +1156,7 @@ // CHECK4-NEXT: store i32 [[A]], i32* [[A_ADDR]], align 4 // CHECK4-NEXT: store i32 [[B]], i32* [[B_ADDR]], align 4 // CHECK4-NEXT: store i32 [[ARGC]], i32* [[ARGC_ADDR]], align 4 -// CHECK4-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true) +// CHECK4-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true, i32 0) // CHECK4-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 // CHECK4-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK4: user_code.entry: @@ -1200,7 +1200,7 @@ // CHECK4-NEXT: store i32 [[A]], i32* [[A_ADDR]], align 4 // CHECK4-NEXT: store i32 [[B]], i32* [[B_ADDR]], align 4 // CHECK4-NEXT: store i8** [[ARGC]], i8*** [[ARGC_ADDR]], align 4 -// CHECK4-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true) +// CHECK4-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true, i32 0) // CHECK4-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 // CHECK4-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK4: user_code.entry: diff --git a/clang/test/OpenMP/nvptx_teams_reduction_codegen.cpp b/clang/test/OpenMP/nvptx_teams_reduction_codegen.cpp --- a/clang/test/OpenMP/nvptx_teams_reduction_codegen.cpp +++ b/clang/test/OpenMP/nvptx_teams_reduction_codegen.cpp @@ -4215,7 +4215,7 @@ // CHECK1-NEXT: store i32 0, i32* [[DOTZERO_ADDR]], align 4 // CHECK1-NEXT: store i64 [[E]], i64* [[E_ADDR]], align 8 // CHECK1-NEXT: [[CONV:%.*]] = bitcast i64* [[E_ADDR]] to double* -// CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true) +// CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true, i32 0) // CHECK1-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 // CHECK1-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK1: user_code.entry: @@ -4517,7 +4517,7 @@ // CHECK1-NEXT: store i64 [[D]], i64* [[D_ADDR]], align 8 // CHECK1-NEXT: [[CONV:%.*]] = bitcast i64* [[C_ADDR]] to i8* // CHECK1-NEXT: [[CONV1:%.*]] = bitcast i64* [[D_ADDR]] to float* -// CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true) +// CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true, i32 0) // CHECK1-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 // CHECK1-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK1: user_code.entry: @@ -4899,7 +4899,7 @@ // CHECK1-NEXT: store i64 [[B]], i64* [[B_ADDR]], align 8 // CHECK1-NEXT: [[CONV:%.*]] = bitcast i64* [[A_ADDR]] to i32* // CHECK1-NEXT: [[CONV1:%.*]] = bitcast i64* [[B_ADDR]] to i16* -// CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true) +// CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true, i32 0) // CHECK1-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 // CHECK1-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK1: user_code.entry: @@ -5537,7 +5537,7 @@ // CHECK2-NEXT: store i32 0, i32* [[DOTZERO_ADDR]], align 4 // CHECK2-NEXT: store double* [[E]], double** [[E_ADDR]], align 4 // CHECK2-NEXT: [[TMP0:%.*]] = load double*, double** [[E_ADDR]], align 4 -// CHECK2-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true) +// CHECK2-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true, i32 0) // CHECK2-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK2-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK2: user_code.entry: @@ -5836,7 +5836,7 @@ // CHECK2-NEXT: store i32 [[D]], i32* [[D_ADDR]], align 4 // CHECK2-NEXT: [[CONV:%.*]] = bitcast i32* [[C_ADDR]] to i8* // CHECK2-NEXT: [[CONV1:%.*]] = bitcast i32* [[D_ADDR]] to float* -// CHECK2-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true) +// CHECK2-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true, i32 0) // CHECK2-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 // CHECK2-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK2: user_code.entry: @@ -6217,7 +6217,7 @@ // CHECK2-NEXT: store i32 [[A]], i32* [[A_ADDR]], align 4 // CHECK2-NEXT: store i32 [[B]], i32* [[B_ADDR]], align 4 // CHECK2-NEXT: [[CONV:%.*]] = bitcast i32* [[B_ADDR]] to i16* -// CHECK2-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true) +// CHECK2-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true, i32 0) // CHECK2-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 // CHECK2-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK2: user_code.entry: @@ -6855,7 +6855,7 @@ // CHECK3-NEXT: store i32 0, i32* [[DOTZERO_ADDR]], align 4 // CHECK3-NEXT: store double* [[E]], double** [[E_ADDR]], align 4 // CHECK3-NEXT: [[TMP0:%.*]] = load double*, double** [[E_ADDR]], align 4 -// CHECK3-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true) +// CHECK3-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true, i32 0) // CHECK3-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1 // CHECK3-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK3: user_code.entry: @@ -7154,7 +7154,7 @@ // CHECK3-NEXT: store i32 [[D]], i32* [[D_ADDR]], align 4 // CHECK3-NEXT: [[CONV:%.*]] = bitcast i32* [[C_ADDR]] to i8* // CHECK3-NEXT: [[CONV1:%.*]] = bitcast i32* [[D_ADDR]] to float* -// CHECK3-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true) +// CHECK3-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true, i32 0) // CHECK3-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 // CHECK3-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK3: user_code.entry: @@ -7535,7 +7535,7 @@ // CHECK3-NEXT: store i32 [[A]], i32* [[A_ADDR]], align 4 // CHECK3-NEXT: store i32 [[B]], i32* [[B_ADDR]], align 4 // CHECK3-NEXT: [[CONV:%.*]] = bitcast i32* [[B_ADDR]] to i16* -// CHECK3-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true) +// CHECK3-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true, i32 0) // CHECK3-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 // CHECK3-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] // CHECK3: user_code.entry: diff --git a/clang/test/OpenMP/target_parallel_debug_codegen.cpp b/clang/test/OpenMP/target_parallel_debug_codegen.cpp --- a/clang/test/OpenMP/target_parallel_debug_codegen.cpp +++ b/clang/test/OpenMP/target_parallel_debug_codegen.cpp @@ -65,7 +65,7 @@ return 0; } // CHECK1-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}_main_l23_debug__ -// CHECK1-SAME: ([10 x [10 x [10 x i32]]] addrspace(1)* noalias [[C:%.*]], i32 [[A:%.*]], [10 x [10 x i32]]* noalias [[B:%.*]], i8 addrspace(1)* noalias [[BB:%.*]]) #[[ATTR0:[0-9]+]] !dbg [[DBG24:![0-9]+]] { +// CHECK1-SAME: ([10 x [10 x [10 x i32]]] addrspace(1)* noalias [[C:%.*]], i32 [[A:%.*]], [10 x [10 x i32]]* noalias [[B:%.*]], i8 addrspace(1)* noalias [[BB:%.*]]) #[[ATTR0:[0-9]+]] !dbg [[DBG23:![0-9]+]] { // CHECK1-NEXT: entry: // CHECK1-NEXT: [[C_ADDR:%.*]] = alloca [10 x [10 x [10 x i32]]] addrspace(1)*, align 8 // CHECK1-NEXT: [[A_ADDR:%.*]] = alloca i32, align 4 @@ -95,7 +95,7 @@ // CHECK1-NEXT: [[TMP6:%.*]] = addrspacecast i8 addrspace(1)* [[TMP5]] to i8*, !dbg [[DBG47]] // CHECK1-NEXT: store i8* [[TMP6]], i8** [[_TMP2]], align 8, !dbg [[DBG47]] // CHECK1-NEXT: [[TMP7:%.*]] = load i8*, i8** [[_TMP2]], align 8, !dbg [[DBG47]] -// CHECK1-NEXT: [[TMP8:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 true), !dbg [[DBG47]] +// CHECK1-NEXT: [[TMP8:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 true, i32 0), !dbg [[DBG47]] // CHECK1-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP8]], -1, !dbg [[DBG47]] // CHECK1-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]], !dbg [[DBG47]] // CHECK1: user_code.entry: @@ -256,7 +256,7 @@ // // // CHECK1-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}_main_l23 -// CHECK1-SAME: ([10 x [10 x [10 x i32]]]* nonnull align 4 dereferenceable(4000) [[C:%.*]], i64 [[A:%.*]], [10 x [10 x i32]]* nonnull align 4 dereferenceable(400) [[B:%.*]], i8* nonnull align 1 dereferenceable(1) [[BB:%.*]]) #[[ATTR4:[0-9]+]] !dbg [[DBG98:![0-9]+]] { +// CHECK1-SAME: ([10 x [10 x [10 x i32]]]* nonnull align 4 dereferenceable(4000) [[C:%.*]], i64 [[A:%.*]], [10 x [10 x i32]]* nonnull align 4 dereferenceable(400) [[B:%.*]], i8* nonnull align 1 dereferenceable(1) [[BB:%.*]]) #[[ATTR5:[0-9]+]] !dbg [[DBG115:![0-9]+]] { // CHECK1-NEXT: entry: // CHECK1-NEXT: [[C_ADDR:%.*]] = alloca [10 x [10 x [10 x i32]]]*, align 8 // CHECK1-NEXT: [[A_ADDR:%.*]] = alloca i64, align 8 @@ -316,7 +316,7 @@ // CHECK1-NEXT: [[TMP7:%.*]] = addrspacecast i8 addrspace(1)* [[TMP6]] to i8*, !dbg [[DBG137]] // CHECK1-NEXT: store i8* [[TMP7]], i8** [[_TMP2]], align 8, !dbg [[DBG137]] // CHECK1-NEXT: [[TMP8:%.*]] = load i8*, i8** [[_TMP2]], align 8, !dbg [[DBG137]] -// CHECK1-NEXT: [[TMP9:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB7:[0-9]+]], i8 2, i1 false, i1 true), !dbg [[DBG137]] +// CHECK1-NEXT: [[TMP9:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB7:[0-9]+]], i8 2, i1 false, i1 true, i32 0), !dbg [[DBG137]] // CHECK1-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP9]], -1, !dbg [[DBG137]] // CHECK1-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]], !dbg [[DBG137]] // CHECK1: user_code.entry: @@ -466,7 +466,7 @@ // // // CHECK1-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}_main_l37 -// CHECK1-SAME: ([10 x [10 x [10 x i32]]]* nonnull align 4 dereferenceable(4000) [[C:%.*]], i64 [[A:%.*]], [10 x [10 x i32]]* nonnull align 4 dereferenceable(400) [[B:%.*]], i8* nonnull align 1 dereferenceable(1) [[BB:%.*]]) #[[ATTR4]] !dbg [[DBG168:![0-9]+]] { +// CHECK1-SAME: ([10 x [10 x [10 x i32]]]* nonnull align 4 dereferenceable(4000) [[C:%.*]], i64 [[A:%.*]], [10 x [10 x i32]]* nonnull align 4 dereferenceable(400) [[B:%.*]], i8* nonnull align 1 dereferenceable(1) [[BB:%.*]]) #[[ATTR5]] !dbg [[DBG192:![0-9]+]] { // CHECK1-NEXT: entry: // CHECK1-NEXT: [[C_ADDR:%.*]] = alloca [10 x [10 x [10 x i32]]]*, align 8 // CHECK1-NEXT: [[A_ADDR:%.*]] = alloca i64, align 8 @@ -531,7 +531,7 @@ // CHECK1-NEXT: [[TMP10:%.*]] = addrspacecast i8 addrspace(1)* [[TMP9]] to i8*, !dbg [[DBG212]] // CHECK1-NEXT: store i8* [[TMP10]], i8** [[_TMP3]], align 8, !dbg [[DBG212]] // CHECK1-NEXT: [[TMP11:%.*]] = load i8*, i8** [[_TMP3]], align 8, !dbg [[DBG212]] -// CHECK1-NEXT: [[TMP12:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB13:[0-9]+]], i8 2, i1 false, i1 true), !dbg [[DBG212]] +// CHECK1-NEXT: [[TMP12:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB13:[0-9]+]], i8 2, i1 false, i1 true, i32 0), !dbg [[DBG212]] // CHECK1-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP12]], -1, !dbg [[DBG212]] // CHECK1-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]], !dbg [[DBG212]] // CHECK1: user_code.entry: @@ -687,7 +687,7 @@ // // // CHECK1-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}_main_l51 -// CHECK1-SAME: ([10 x [10 x [10 x i32]]]* nonnull align 4 dereferenceable(4000) [[C:%.*]], i32* nonnull align 4 dereferenceable(4) [[A:%.*]], [10 x [10 x i32]]* nonnull align 4 dereferenceable(400) [[B:%.*]], i8* nonnull align 1 dereferenceable(1) [[BB:%.*]]) #[[ATTR4]] !dbg [[DBG232:![0-9]+]] { +// CHECK1-SAME: ([10 x [10 x [10 x i32]]]* nonnull align 4 dereferenceable(4000) [[C:%.*]], i32* nonnull align 4 dereferenceable(4) [[A:%.*]], [10 x [10 x i32]]* nonnull align 4 dereferenceable(400) [[B:%.*]], i8* nonnull align 1 dereferenceable(1) [[BB:%.*]]) #[[ATTR5]] !dbg [[DBG270:![0-9]+]] { // CHECK1-NEXT: entry: // CHECK1-NEXT: [[C_ADDR:%.*]] = alloca [10 x [10 x [10 x i32]]]*, align 8 // CHECK1-NEXT: [[A_ADDR:%.*]] = alloca i32*, align 8 diff --git a/clang/test/OpenMP/target_parallel_for_debug_codegen.cpp b/clang/test/OpenMP/target_parallel_for_debug_codegen.cpp --- a/clang/test/OpenMP/target_parallel_for_debug_codegen.cpp +++ b/clang/test/OpenMP/target_parallel_for_debug_codegen.cpp @@ -55,7 +55,7 @@ return 0; } // CHECK1-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}_main_l13_debug__ -// CHECK1-SAME: ([10 x [10 x [10 x i32]]] addrspace(1)* noalias [[C:%.*]], i32 [[A:%.*]], [10 x [10 x i32]]* noalias [[B:%.*]], i8 addrspace(1)* noalias [[BB:%.*]], i1 zeroext [[DOTCAPTURE_EXPR_:%.*]]) #[[ATTR0:[0-9]+]] !dbg [[DBG14:![0-9]+]] { +// CHECK1-SAME: ([10 x [10 x [10 x i32]]] addrspace(1)* noalias [[C:%.*]], i32 [[A:%.*]], [10 x [10 x i32]]* noalias [[B:%.*]], i8 addrspace(1)* noalias [[BB:%.*]], i1 zeroext [[DOTCAPTURE_EXPR_:%.*]]) #[[ATTR0:[0-9]+]] !dbg [[DBG13:![0-9]+]] { // CHECK1-NEXT: entry: // CHECK1-NEXT: [[C_ADDR:%.*]] = alloca [10 x [10 x [10 x i32]]] addrspace(1)*, align 8 // CHECK1-NEXT: [[A_ADDR:%.*]] = alloca i32, align 4 @@ -89,7 +89,7 @@ // CHECK1-NEXT: [[TMP6:%.*]] = addrspacecast i8 addrspace(1)* [[TMP5]] to i8*, !dbg [[DBG41]] // CHECK1-NEXT: store i8* [[TMP6]], i8** [[_TMP2]], align 8, !dbg [[DBG41]] // CHECK1-NEXT: [[TMP7:%.*]] = load i8*, i8** [[_TMP2]], align 8, !dbg [[DBG41]] -// CHECK1-NEXT: [[TMP8:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 false), !dbg [[DBG41]] +// CHECK1-NEXT: [[TMP8:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 false, i32 0), !dbg [[DBG41]] // CHECK1-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP8]], -1, !dbg [[DBG41]] // CHECK1-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]], !dbg [[DBG41]] // CHECK1: user_code.entry: @@ -326,7 +326,7 @@ // // // CHECK1-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}_main_l13 -// CHECK1-SAME: ([10 x [10 x [10 x i32]]]* nonnull align 4 dereferenceable(4000) [[C:%.*]], i64 [[A:%.*]], [10 x [10 x i32]]* nonnull align 4 dereferenceable(400) [[B:%.*]], i8* nonnull align 1 dereferenceable(1) [[BB:%.*]], i64 [[DOTCAPTURE_EXPR_:%.*]]) #[[ATTR4:[0-9]+]] !dbg [[DBG106:![0-9]+]] { +// CHECK1-SAME: ([10 x [10 x [10 x i32]]]* nonnull align 4 dereferenceable(4000) [[C:%.*]], i64 [[A:%.*]], [10 x [10 x i32]]* nonnull align 4 dereferenceable(400) [[B:%.*]], i8* nonnull align 1 dereferenceable(1) [[BB:%.*]], i64 [[DOTCAPTURE_EXPR_:%.*]]) #[[ATTR5:[0-9]+]] !dbg [[DBG123:![0-9]+]] { // CHECK1-NEXT: entry: // CHECK1-NEXT: [[C_ADDR:%.*]] = alloca [10 x [10 x [10 x i32]]]*, align 8 // CHECK1-NEXT: [[A_ADDR:%.*]] = alloca i64, align 8 @@ -392,7 +392,7 @@ // CHECK1-NEXT: [[TMP7:%.*]] = addrspacecast i8 addrspace(1)* [[TMP6]] to i8*, !dbg [[DBG146]] // CHECK1-NEXT: store i8* [[TMP7]], i8** [[_TMP2]], align 8, !dbg [[DBG146]] // CHECK1-NEXT: [[TMP8:%.*]] = load i8*, i8** [[_TMP2]], align 8, !dbg [[DBG146]] -// CHECK1-NEXT: [[TMP9:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB10:[0-9]+]], i8 2, i1 false, i1 false), !dbg [[DBG146]] +// CHECK1-NEXT: [[TMP9:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB10:[0-9]+]], i8 2, i1 false, i1 false, i32 0), !dbg [[DBG146]] // CHECK1-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP9]], -1, !dbg [[DBG146]] // CHECK1-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]], !dbg [[DBG146]] // CHECK1: user_code.entry: @@ -615,7 +615,7 @@ // // // CHECK1-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}_main_l27 -// CHECK1-SAME: ([10 x [10 x [10 x i32]]]* nonnull align 4 dereferenceable(4000) [[C:%.*]], i64 [[A:%.*]], [10 x [10 x i32]]* nonnull align 4 dereferenceable(400) [[B:%.*]], i8* nonnull align 1 dereferenceable(1) [[BB:%.*]]) #[[ATTR4]] !dbg [[DBG190:![0-9]+]] { +// CHECK1-SAME: ([10 x [10 x [10 x i32]]]* nonnull align 4 dereferenceable(4000) [[C:%.*]], i64 [[A:%.*]], [10 x [10 x i32]]* nonnull align 4 dereferenceable(400) [[B:%.*]], i8* nonnull align 1 dereferenceable(1) [[BB:%.*]]) #[[ATTR5]] !dbg [[DBG214:![0-9]+]] { // CHECK1-NEXT: entry: // CHECK1-NEXT: [[C_ADDR:%.*]] = alloca [10 x [10 x [10 x i32]]]*, align 8 // CHECK1-NEXT: [[A_ADDR:%.*]] = alloca i64, align 8 @@ -680,7 +680,7 @@ // CHECK1-NEXT: [[TMP10:%.*]] = addrspacecast i8 addrspace(1)* [[TMP9]] to i8*, !dbg [[DBG236]] // CHECK1-NEXT: store i8* [[TMP10]], i8** [[_TMP3]], align 8, !dbg [[DBG236]] // CHECK1-NEXT: [[TMP11:%.*]] = load i8*, i8** [[_TMP3]], align 8, !dbg [[DBG236]] -// CHECK1-NEXT: [[TMP12:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB19:[0-9]+]], i8 2, i1 false, i1 false), !dbg [[DBG236]] +// CHECK1-NEXT: [[TMP12:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB19:[0-9]+]], i8 2, i1 false, i1 false, i32 0), !dbg [[DBG236]] // CHECK1-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP12]], -1, !dbg [[DBG236]] // CHECK1-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]], !dbg [[DBG236]] // CHECK1: user_code.entry: @@ -909,7 +909,7 @@ // // // CHECK1-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}_main_l41 -// CHECK1-SAME: ([10 x [10 x [10 x i32]]]* nonnull align 4 dereferenceable(4000) [[C:%.*]], i32* nonnull align 4 dereferenceable(4) [[A:%.*]], [10 x [10 x i32]]* nonnull align 4 dereferenceable(400) [[B:%.*]], i8* nonnull align 1 dereferenceable(1) [[BB:%.*]]) #[[ATTR4]] !dbg [[DBG269:![0-9]+]] { +// CHECK1-SAME: ([10 x [10 x [10 x i32]]]* nonnull align 4 dereferenceable(4000) [[C:%.*]], i32* nonnull align 4 dereferenceable(4) [[A:%.*]], [10 x [10 x i32]]* nonnull align 4 dereferenceable(400) [[B:%.*]], i8* nonnull align 1 dereferenceable(1) [[BB:%.*]]) #[[ATTR5]] !dbg [[DBG307:![0-9]+]] { // CHECK1-NEXT: entry: // CHECK1-NEXT: [[C_ADDR:%.*]] = alloca [10 x [10 x [10 x i32]]]*, align 8 // CHECK1-NEXT: [[A_ADDR:%.*]] = alloca i32*, align 8 diff --git a/clang/test/OpenMP/target_stack_size_codegen.cpp b/clang/test/OpenMP/target_stack_size_codegen.cpp new file mode 100644 --- /dev/null +++ b/clang/test/OpenMP/target_stack_size_codegen.cpp @@ -0,0 +1,20 @@ +// Test target codegen - host bc file has to be created first. +// RUN: %clang_cc1 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=nvptx64-nvidia-cuda -emit-llvm-bc %s -o %t-ppc-host.bc +// RUN: %clang_cc1 -verify -fopenmp -x c++ -triple nvptx64-unknown-unknown -fopenmp-targets=nvptx64-nvidia-cuda -emit-llvm %s -fopenmp-target-new-runtime -fopenmp-target-stack-size=8192 -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o - | FileCheck %s --check-prefix=CHECK +// RUN: %clang_cc1 -verify -fopenmp -x c++ -triple nvptx64-unknown-unknown -fopenmp-targets=nvptx64-nvidia-cuda -emit-llvm %s -fopenmp-target-new-runtime -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o - | FileCheck %s --check-prefix=CHECK-DEFAULT +// expected-no-diagnostics + +#ifndef HEADER +#define HEADER + +//. +// CHECK: @__omp_offloading_{{[0-9a-z]+_[0-9a-z]+_}}_Z3foov_{{[0-9a-z_]+}}_stack_size = weak constant i32 8192 +//. +// CHECK-DEFAULT: @__omp_offloading_{{[0-9a-z]+_[0-9a-z]+_}}_Z3foov_{{[0-9a-z_]+}}_stack_size = weak constant i32 2048 +//. +void foo() { +#pragma omp target + { } +} + +#endif diff --git a/llvm/include/llvm/Frontend/OpenMP/OMPIRBuilder.h b/llvm/include/llvm/Frontend/OpenMP/OMPIRBuilder.h --- a/llvm/include/llvm/Frontend/OpenMP/OMPIRBuilder.h +++ b/llvm/include/llvm/Frontend/OpenMP/OMPIRBuilder.h @@ -1005,8 +1005,9 @@ /// \param Loc The insert and source location description. /// \param IsSPMD Flag to indicate if the kernel is an SPMD kernel or not. /// \param RequiresFullRuntime Indicate if a full device runtime is necessary. + /// \param StackSize Amount of dynamic shared memory used for the stack. InsertPointTy createTargetInit(const LocationDescription &Loc, bool IsSPMD, - bool RequiresFullRuntime); + bool RequiresFullRuntime, uint32_t StackSize); /// Create a runtime call for kmpc_target_deinit /// diff --git a/llvm/include/llvm/Frontend/OpenMP/OMPKinds.def b/llvm/include/llvm/Frontend/OpenMP/OMPKinds.def --- a/llvm/include/llvm/Frontend/OpenMP/OMPKinds.def +++ b/llvm/include/llvm/Frontend/OpenMP/OMPKinds.def @@ -414,7 +414,7 @@ /* Int */ Int32, /* kmp_task_t */ VoidPtr) /// OpenMP Device runtime functions -__OMP_RTL(__kmpc_target_init, false, Int32, IdentPtr, Int8, Int1, Int1) +__OMP_RTL(__kmpc_target_init, false, Int32, IdentPtr, Int8, Int1, Int1, Int32) __OMP_RTL(__kmpc_target_deinit, false, Void, IdentPtr, Int8, Int1) __OMP_RTL(__kmpc_kernel_prepare_parallel, false, Void, VoidPtr) __OMP_RTL(__kmpc_parallel_51, false, Void, IdentPtr, Int32, Int32, Int32, Int32, diff --git a/llvm/lib/Frontend/OpenMP/OMPIRBuilder.cpp b/llvm/lib/Frontend/OpenMP/OMPIRBuilder.cpp --- a/llvm/lib/Frontend/OpenMP/OMPIRBuilder.cpp +++ b/llvm/lib/Frontend/OpenMP/OMPIRBuilder.cpp @@ -2755,7 +2755,8 @@ OpenMPIRBuilder::InsertPointTy OpenMPIRBuilder::createTargetInit(const LocationDescription &Loc, bool IsSPMD, - bool RequiresFullRuntime) { + bool RequiresFullRuntime, + uint32_t StackSize) { if (!updateToLocation(Loc)) return Loc.IP; @@ -2768,12 +2769,15 @@ ConstantInt::getBool(Int32->getContext(), !IsSPMD); ConstantInt *RequiresFullRuntimeVal = ConstantInt::getBool(Int32->getContext(), RequiresFullRuntime); + ConstantInt *StackSizeVal = + ConstantInt::get(Type::getInt32Ty(M.getContext()), StackSize); Function *Fn = getOrCreateRuntimeFunctionPtr( omp::RuntimeFunction::OMPRTL___kmpc_target_init); - CallInst *ThreadKind = Builder.CreateCall( - Fn, {Ident, IsSPMDVal, UseGenericStateMachine, RequiresFullRuntimeVal}); + CallInst *ThreadKind = + Builder.CreateCall(Fn, {Ident, IsSPMDVal, UseGenericStateMachine, + RequiresFullRuntimeVal, StackSizeVal}); Value *ExecUserCode = Builder.CreateICmpEQ( ThreadKind, ConstantInt::get(ThreadKind->getType(), -1), diff --git a/llvm/test/Transforms/OpenMP/always_inline_device.ll b/llvm/test/Transforms/OpenMP/always_inline_device.ll --- a/llvm/test/Transforms/OpenMP/always_inline_device.ll +++ b/llvm/test/Transforms/OpenMP/always_inline_device.ll @@ -12,7 +12,7 @@ ; CHECK: Function Attrs: convergent norecurse nounwind ; CHECK-LABEL: @__omp_offloading_fd02_c0934fc2_foo_l4( ; CHECK-NEXT: entry: -; CHECK-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 false, i1 true) +; CHECK-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 false, i1 true, i32 0) ; CHECK-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 ; CHECK-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] ; CHECK: user_code.entry: @@ -22,7 +22,7 @@ ; CHECK-NEXT: ret void ; entry: - %0 = call i32 @__kmpc_target_init(%struct.ident_t* @1, i8 1, i1 true, i1 true) + %0 = call i32 @__kmpc_target_init(%struct.ident_t* @1, i8 1, i1 true, i1 true, i32 0) %exec_user_code = icmp eq i32 %0, -1 br i1 %exec_user_code, label %user_code.entry, label %worker.exit @@ -35,7 +35,7 @@ ret void } -declare i32 @__kmpc_target_init(%struct.ident_t*, i8, i1, i1) +declare i32 @__kmpc_target_init(%struct.ident_t*, i8, i1, i1, i32) declare void @__kmpc_target_deinit(%struct.ident_t*, i8, i1) diff --git a/llvm/test/Transforms/OpenMP/custom_state_machines.ll b/llvm/test/Transforms/OpenMP/custom_state_machines.ll --- a/llvm/test/Transforms/OpenMP/custom_state_machines.ll +++ b/llvm/test/Transforms/OpenMP/custom_state_machines.ll @@ -141,7 +141,7 @@ %.zero.addr = alloca i32, align 4 %.threadid_temp. = alloca i32, align 4 store i32 0, i32* %.zero.addr, align 4 - %0 = call i32 @__kmpc_target_init(%struct.ident_t* @1, i8 1, i1 true, i1 true) + %0 = call i32 @__kmpc_target_init(%struct.ident_t* @1, i8 1, i1 true, i1 true, i32 0) %exec_user_code = icmp eq i32 %0, -1 br i1 %exec_user_code, label %user_code.entry, label %worker.exit @@ -156,7 +156,7 @@ ret void } -declare i32 @__kmpc_target_init(%struct.ident_t*, i8, i1, i1) +declare i32 @__kmpc_target_init(%struct.ident_t*, i8, i1, i1, i32) define internal void @__omp_outlined__(i32* noalias %.global_tid., i32* noalias %.bound_tid.) #0 { entry: @@ -197,7 +197,7 @@ %.zero.addr = alloca i32, align 4 %.threadid_temp. = alloca i32, align 4 store i32 0, i32* %.zero.addr, align 4 - %0 = call i32 @__kmpc_target_init(%struct.ident_t* @1, i8 1, i1 true, i1 true) + %0 = call i32 @__kmpc_target_init(%struct.ident_t* @1, i8 1, i1 true, i1 true, i32 0) %exec_user_code = icmp eq i32 %0, -1 br i1 %exec_user_code, label %user_code.entry, label %worker.exit @@ -292,7 +292,7 @@ %.zero.addr = alloca i32, align 4 %.threadid_temp. = alloca i32, align 4 store i32 0, i32* %.zero.addr, align 4 - %0 = call i32 @__kmpc_target_init(%struct.ident_t* @1, i8 1, i1 true, i1 true) + %0 = call i32 @__kmpc_target_init(%struct.ident_t* @1, i8 1, i1 true, i1 true, i32 0) %exec_user_code = icmp eq i32 %0, -1 br i1 %exec_user_code, label %user_code.entry, label %worker.exit @@ -372,7 +372,7 @@ %.zero.addr = alloca i32, align 4 %.threadid_temp. = alloca i32, align 4 store i32 0, i32* %.zero.addr, align 4 - %0 = call i32 @__kmpc_target_init(%struct.ident_t* @1, i8 1, i1 true, i1 true) + %0 = call i32 @__kmpc_target_init(%struct.ident_t* @1, i8 1, i1 true, i1 true, i32 0) %exec_user_code = icmp eq i32 %0, -1 br i1 %exec_user_code, label %user_code.entry, label %worker.exit @@ -460,7 +460,7 @@ %.zero.addr = alloca i32, align 4 %.threadid_temp. = alloca i32, align 4 store i32 0, i32* %.zero.addr, align 4 - %0 = call i32 @__kmpc_target_init(%struct.ident_t* @1, i8 1, i1 true, i1 true) + %0 = call i32 @__kmpc_target_init(%struct.ident_t* @1, i8 1, i1 true, i1 true, i32 0) %exec_user_code = icmp eq i32 %0, -1 br i1 %exec_user_code, label %user_code.entry, label %worker.exit @@ -546,7 +546,7 @@ %.zero.addr = alloca i32, align 4 %.threadid_temp. = alloca i32, align 4 store i32 0, i32* %.zero.addr, align 4 - %0 = call i32 @__kmpc_target_init(%struct.ident_t* @1, i8 1, i1 true, i1 true) + %0 = call i32 @__kmpc_target_init(%struct.ident_t* @1, i8 1, i1 true, i1 true, i32 0) %exec_user_code = icmp eq i32 %0, -1 br i1 %exec_user_code, label %user_code.entry, label %worker.exit @@ -635,7 +635,7 @@ %.zero.addr = alloca i32, align 4 %.threadid_temp. = alloca i32, align 4 store i32 0, i32* %.zero.addr, align 4 - %0 = call i32 @__kmpc_target_init(%struct.ident_t* @1, i8 1, i1 true, i1 true) + %0 = call i32 @__kmpc_target_init(%struct.ident_t* @1, i8 1, i1 true, i1 true, i32 0) %exec_user_code = icmp eq i32 %0, -1 br i1 %exec_user_code, label %user_code.entry, label %worker.exit @@ -690,7 +690,7 @@ %.zero.addr = alloca i32, align 4 %.threadid_temp. = alloca i32, align 4 store i32 0, i32* %.zero.addr, align 4 - %0 = call i32 @__kmpc_target_init(%struct.ident_t* @1, i8 1, i1 true, i1 true) + %0 = call i32 @__kmpc_target_init(%struct.ident_t* @1, i8 1, i1 true, i1 true, i32 0) %exec_user_code = icmp eq i32 %0, -1 br i1 %exec_user_code, label %user_code.entry, label %worker.exit @@ -847,7 +847,7 @@ ; AMDGPU-NEXT: entry: ; AMDGPU-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 ; AMDGPU-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 -; AMDGPU-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 false, i1 true) +; AMDGPU-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 false, i1 true, i32 0) ; AMDGPU-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 ; AMDGPU-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] ; AMDGPU: user_code.entry: @@ -911,7 +911,7 @@ ; AMDGPU-NEXT: [[WORKER_WORK_FN_ADDR:%.*]] = alloca i8*, align 8, addrspace(5) ; AMDGPU-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 ; AMDGPU-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 -; AMDGPU-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true) +; AMDGPU-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true, i32 0) ; AMDGPU-NEXT: [[THREAD_IS_WORKER:%.*]] = icmp ne i32 [[TMP0]], -1 ; AMDGPU-NEXT: br i1 [[THREAD_IS_WORKER]], label [[WORKER_STATE_MACHINE_BEGIN:%.*]], label [[THREAD_USER_CODE_CHECK:%.*]] ; AMDGPU: worker_state_machine.begin: @@ -1036,7 +1036,7 @@ ; AMDGPU-NEXT: [[WORKER_WORK_FN_ADDR:%.*]] = alloca i8*, align 8, addrspace(5) ; AMDGPU-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 ; AMDGPU-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 -; AMDGPU-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true) +; AMDGPU-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true, i32 0) ; AMDGPU-NEXT: [[THREAD_IS_WORKER:%.*]] = icmp ne i32 [[TMP0]], -1 ; AMDGPU-NEXT: br i1 [[THREAD_IS_WORKER]], label [[WORKER_STATE_MACHINE_BEGIN:%.*]], label [[THREAD_USER_CODE_CHECK:%.*]] ; AMDGPU: worker_state_machine.begin: @@ -1184,7 +1184,7 @@ ; AMDGPU-NEXT: [[WORKER_WORK_FN_ADDR:%.*]] = alloca i8*, align 8, addrspace(5) ; AMDGPU-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 ; AMDGPU-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 -; AMDGPU-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true) +; AMDGPU-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true, i32 0) ; AMDGPU-NEXT: [[THREAD_IS_WORKER:%.*]] = icmp ne i32 [[TMP0]], -1 ; AMDGPU-NEXT: br i1 [[THREAD_IS_WORKER]], label [[WORKER_STATE_MACHINE_BEGIN:%.*]], label [[THREAD_USER_CODE_CHECK:%.*]] ; AMDGPU: worker_state_machine.begin: @@ -1310,7 +1310,7 @@ ; AMDGPU-NEXT: [[WORKER_WORK_FN_ADDR:%.*]] = alloca i8*, align 8, addrspace(5) ; AMDGPU-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 ; AMDGPU-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 -; AMDGPU-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true) +; AMDGPU-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true, i32 0) ; AMDGPU-NEXT: [[THREAD_IS_WORKER:%.*]] = icmp ne i32 [[TMP0]], -1 ; AMDGPU-NEXT: br i1 [[THREAD_IS_WORKER]], label [[WORKER_STATE_MACHINE_BEGIN:%.*]], label [[THREAD_USER_CODE_CHECK:%.*]] ; AMDGPU: worker_state_machine.begin: @@ -1434,7 +1434,7 @@ ; AMDGPU-NEXT: [[WORKER_WORK_FN_ADDR:%.*]] = alloca i8*, align 8, addrspace(5) ; AMDGPU-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 ; AMDGPU-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 -; AMDGPU-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true) +; AMDGPU-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true, i32 0) ; AMDGPU-NEXT: [[THREAD_IS_WORKER:%.*]] = icmp ne i32 [[TMP0]], -1 ; AMDGPU-NEXT: br i1 [[THREAD_IS_WORKER]], label [[WORKER_STATE_MACHINE_BEGIN:%.*]], label [[THREAD_USER_CODE_CHECK:%.*]] ; AMDGPU: worker_state_machine.begin: @@ -1558,7 +1558,7 @@ ; AMDGPU-NEXT: [[WORKER_WORK_FN_ADDR:%.*]] = alloca i8*, align 8, addrspace(5) ; AMDGPU-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 ; AMDGPU-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 -; AMDGPU-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true) +; AMDGPU-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true, i32 0) ; AMDGPU-NEXT: [[THREAD_IS_WORKER:%.*]] = icmp ne i32 [[TMP0]], -1 ; AMDGPU-NEXT: br i1 [[THREAD_IS_WORKER]], label [[WORKER_STATE_MACHINE_BEGIN:%.*]], label [[THREAD_USER_CODE_CHECK:%.*]] ; AMDGPU: worker_state_machine.begin: @@ -1658,7 +1658,7 @@ ; AMDGPU-NEXT: [[WORKER_WORK_FN_ADDR:%.*]] = alloca i8*, align 8, addrspace(5) ; AMDGPU-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 ; AMDGPU-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 -; AMDGPU-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true) +; AMDGPU-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true, i32 0) ; AMDGPU-NEXT: [[THREAD_IS_WORKER:%.*]] = icmp ne i32 [[TMP0]], -1 ; AMDGPU-NEXT: br i1 [[THREAD_IS_WORKER]], label [[WORKER_STATE_MACHINE_BEGIN:%.*]], label [[THREAD_USER_CODE_CHECK:%.*]] ; AMDGPU: worker_state_machine.begin: @@ -1817,7 +1817,7 @@ ; NVPTX-NEXT: entry: ; NVPTX-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 ; NVPTX-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 -; NVPTX-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 false, i1 true) +; NVPTX-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 false, i1 true, i32 0) ; NVPTX-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 ; NVPTX-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] ; NVPTX: user_code.entry: @@ -1881,7 +1881,7 @@ ; NVPTX-NEXT: [[WORKER_WORK_FN_ADDR:%.*]] = alloca i8*, align 8 ; NVPTX-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 ; NVPTX-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 -; NVPTX-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true) +; NVPTX-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true, i32 0) ; NVPTX-NEXT: [[THREAD_IS_WORKER:%.*]] = icmp ne i32 [[TMP0]], -1 ; NVPTX-NEXT: br i1 [[THREAD_IS_WORKER]], label [[WORKER_STATE_MACHINE_BEGIN:%.*]], label [[THREAD_USER_CODE_CHECK:%.*]] ; NVPTX: worker_state_machine.begin: @@ -2005,7 +2005,7 @@ ; NVPTX-NEXT: [[WORKER_WORK_FN_ADDR:%.*]] = alloca i8*, align 8 ; NVPTX-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 ; NVPTX-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 -; NVPTX-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true) +; NVPTX-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true, i32 0) ; NVPTX-NEXT: [[THREAD_IS_WORKER:%.*]] = icmp ne i32 [[TMP0]], -1 ; NVPTX-NEXT: br i1 [[THREAD_IS_WORKER]], label [[WORKER_STATE_MACHINE_BEGIN:%.*]], label [[THREAD_USER_CODE_CHECK:%.*]] ; NVPTX: worker_state_machine.begin: @@ -2152,7 +2152,7 @@ ; NVPTX-NEXT: [[WORKER_WORK_FN_ADDR:%.*]] = alloca i8*, align 8 ; NVPTX-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 ; NVPTX-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 -; NVPTX-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true) +; NVPTX-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true, i32 0) ; NVPTX-NEXT: [[THREAD_IS_WORKER:%.*]] = icmp ne i32 [[TMP0]], -1 ; NVPTX-NEXT: br i1 [[THREAD_IS_WORKER]], label [[WORKER_STATE_MACHINE_BEGIN:%.*]], label [[THREAD_USER_CODE_CHECK:%.*]] ; NVPTX: worker_state_machine.begin: @@ -2277,7 +2277,7 @@ ; NVPTX-NEXT: [[WORKER_WORK_FN_ADDR:%.*]] = alloca i8*, align 8 ; NVPTX-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 ; NVPTX-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 -; NVPTX-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true) +; NVPTX-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true, i32 0) ; NVPTX-NEXT: [[THREAD_IS_WORKER:%.*]] = icmp ne i32 [[TMP0]], -1 ; NVPTX-NEXT: br i1 [[THREAD_IS_WORKER]], label [[WORKER_STATE_MACHINE_BEGIN:%.*]], label [[THREAD_USER_CODE_CHECK:%.*]] ; NVPTX: worker_state_machine.begin: @@ -2400,7 +2400,7 @@ ; NVPTX-NEXT: [[WORKER_WORK_FN_ADDR:%.*]] = alloca i8*, align 8 ; NVPTX-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 ; NVPTX-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 -; NVPTX-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true) +; NVPTX-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true, i32 0) ; NVPTX-NEXT: [[THREAD_IS_WORKER:%.*]] = icmp ne i32 [[TMP0]], -1 ; NVPTX-NEXT: br i1 [[THREAD_IS_WORKER]], label [[WORKER_STATE_MACHINE_BEGIN:%.*]], label [[THREAD_USER_CODE_CHECK:%.*]] ; NVPTX: worker_state_machine.begin: @@ -2523,7 +2523,7 @@ ; NVPTX-NEXT: [[WORKER_WORK_FN_ADDR:%.*]] = alloca i8*, align 8 ; NVPTX-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 ; NVPTX-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 -; NVPTX-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true) +; NVPTX-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true, i32 0) ; NVPTX-NEXT: [[THREAD_IS_WORKER:%.*]] = icmp ne i32 [[TMP0]], -1 ; NVPTX-NEXT: br i1 [[THREAD_IS_WORKER]], label [[WORKER_STATE_MACHINE_BEGIN:%.*]], label [[THREAD_USER_CODE_CHECK:%.*]] ; NVPTX: worker_state_machine.begin: @@ -2622,7 +2622,7 @@ ; NVPTX-NEXT: [[WORKER_WORK_FN_ADDR:%.*]] = alloca i8*, align 8 ; NVPTX-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 ; NVPTX-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 -; NVPTX-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true) +; NVPTX-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true, i32 0) ; NVPTX-NEXT: [[THREAD_IS_WORKER:%.*]] = icmp ne i32 [[TMP0]], -1 ; NVPTX-NEXT: br i1 [[THREAD_IS_WORKER]], label [[WORKER_STATE_MACHINE_BEGIN:%.*]], label [[THREAD_USER_CODE_CHECK:%.*]] ; NVPTX: worker_state_machine.begin: @@ -2780,7 +2780,7 @@ ; AMDGPU-DISABLED-NEXT: entry: ; AMDGPU-DISABLED-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 ; AMDGPU-DISABLED-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 -; AMDGPU-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true) +; AMDGPU-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true, i32 0) ; AMDGPU-DISABLED-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 ; AMDGPU-DISABLED-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] ; AMDGPU-DISABLED: user_code.entry: @@ -2843,7 +2843,7 @@ ; AMDGPU-DISABLED-NEXT: entry: ; AMDGPU-DISABLED-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 ; AMDGPU-DISABLED-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 -; AMDGPU-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true) +; AMDGPU-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true, i32 0) ; AMDGPU-DISABLED-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 ; AMDGPU-DISABLED-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] ; AMDGPU-DISABLED: user_code.entry: @@ -2933,7 +2933,7 @@ ; AMDGPU-DISABLED-NEXT: entry: ; AMDGPU-DISABLED-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 ; AMDGPU-DISABLED-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 -; AMDGPU-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true) +; AMDGPU-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true, i32 0) ; AMDGPU-DISABLED-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 ; AMDGPU-DISABLED-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] ; AMDGPU-DISABLED: user_code.entry: @@ -3040,7 +3040,7 @@ ; AMDGPU-DISABLED-NEXT: entry: ; AMDGPU-DISABLED-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 ; AMDGPU-DISABLED-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 -; AMDGPU-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true) +; AMDGPU-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true, i32 0) ; AMDGPU-DISABLED-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 ; AMDGPU-DISABLED-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] ; AMDGPU-DISABLED: user_code.entry: @@ -3129,7 +3129,7 @@ ; AMDGPU-DISABLED-NEXT: entry: ; AMDGPU-DISABLED-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 ; AMDGPU-DISABLED-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 -; AMDGPU-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true) +; AMDGPU-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true, i32 0) ; AMDGPU-DISABLED-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 ; AMDGPU-DISABLED-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] ; AMDGPU-DISABLED: user_code.entry: @@ -3218,7 +3218,7 @@ ; AMDGPU-DISABLED-NEXT: entry: ; AMDGPU-DISABLED-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 ; AMDGPU-DISABLED-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 -; AMDGPU-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true) +; AMDGPU-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true, i32 0) ; AMDGPU-DISABLED-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 ; AMDGPU-DISABLED-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] ; AMDGPU-DISABLED: user_code.entry: @@ -3307,7 +3307,7 @@ ; AMDGPU-DISABLED-NEXT: entry: ; AMDGPU-DISABLED-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 ; AMDGPU-DISABLED-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 -; AMDGPU-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true) +; AMDGPU-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true, i32 0) ; AMDGPU-DISABLED-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 ; AMDGPU-DISABLED-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] ; AMDGPU-DISABLED: user_code.entry: @@ -3376,7 +3376,7 @@ ; AMDGPU-DISABLED-NEXT: entry: ; AMDGPU-DISABLED-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 ; AMDGPU-DISABLED-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 -; AMDGPU-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true) +; AMDGPU-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true, i32 0) ; AMDGPU-DISABLED-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 ; AMDGPU-DISABLED-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] ; AMDGPU-DISABLED: user_code.entry: @@ -3511,7 +3511,7 @@ ; NVPTX-DISABLED-NEXT: entry: ; NVPTX-DISABLED-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 ; NVPTX-DISABLED-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 -; NVPTX-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true) +; NVPTX-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true, i32 0) ; NVPTX-DISABLED-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 ; NVPTX-DISABLED-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] ; NVPTX-DISABLED: user_code.entry: @@ -3574,7 +3574,7 @@ ; NVPTX-DISABLED-NEXT: entry: ; NVPTX-DISABLED-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 ; NVPTX-DISABLED-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 -; NVPTX-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true) +; NVPTX-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true, i32 0) ; NVPTX-DISABLED-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 ; NVPTX-DISABLED-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] ; NVPTX-DISABLED: user_code.entry: @@ -3664,7 +3664,7 @@ ; NVPTX-DISABLED-NEXT: entry: ; NVPTX-DISABLED-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 ; NVPTX-DISABLED-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 -; NVPTX-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true) +; NVPTX-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true, i32 0) ; NVPTX-DISABLED-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 ; NVPTX-DISABLED-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] ; NVPTX-DISABLED: user_code.entry: @@ -3771,7 +3771,7 @@ ; NVPTX-DISABLED-NEXT: entry: ; NVPTX-DISABLED-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 ; NVPTX-DISABLED-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 -; NVPTX-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true) +; NVPTX-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true, i32 0) ; NVPTX-DISABLED-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 ; NVPTX-DISABLED-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] ; NVPTX-DISABLED: user_code.entry: @@ -3860,7 +3860,7 @@ ; NVPTX-DISABLED-NEXT: entry: ; NVPTX-DISABLED-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 ; NVPTX-DISABLED-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 -; NVPTX-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true) +; NVPTX-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true, i32 0) ; NVPTX-DISABLED-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 ; NVPTX-DISABLED-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] ; NVPTX-DISABLED: user_code.entry: @@ -3949,7 +3949,7 @@ ; NVPTX-DISABLED-NEXT: entry: ; NVPTX-DISABLED-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 ; NVPTX-DISABLED-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 -; NVPTX-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true) +; NVPTX-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true, i32 0) ; NVPTX-DISABLED-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 ; NVPTX-DISABLED-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] ; NVPTX-DISABLED: user_code.entry: @@ -4038,7 +4038,7 @@ ; NVPTX-DISABLED-NEXT: entry: ; NVPTX-DISABLED-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 ; NVPTX-DISABLED-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 -; NVPTX-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true) +; NVPTX-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true, i32 0) ; NVPTX-DISABLED-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 ; NVPTX-DISABLED-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] ; NVPTX-DISABLED: user_code.entry: @@ -4107,7 +4107,7 @@ ; NVPTX-DISABLED-NEXT: entry: ; NVPTX-DISABLED-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 ; NVPTX-DISABLED-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 -; NVPTX-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true) +; NVPTX-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true, i32 0) ; NVPTX-DISABLED-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 ; NVPTX-DISABLED-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] ; NVPTX-DISABLED: user_code.entry: diff --git a/llvm/test/Transforms/OpenMP/custom_state_machines_remarks.ll b/llvm/test/Transforms/OpenMP/custom_state_machines_remarks.ll --- a/llvm/test/Transforms/OpenMP/custom_state_machines_remarks.ll +++ b/llvm/test/Transforms/OpenMP/custom_state_machines_remarks.ll @@ -60,7 +60,7 @@ define weak void @__omp_offloading_2a_d80d3d_test_fallback_l11() local_unnamed_addr #0 !dbg !15 { entry: %captured_vars_addrs.i.i = alloca [0 x i8*], align 8 - %0 = call i32 @__kmpc_target_init(%struct.ident_t* nonnull @1, i8 1, i1 true, i1 true) #3, !dbg !18 + %0 = call i32 @__kmpc_target_init(%struct.ident_t* nonnull @1, i8 1, i1 true, i1 true, i32 0) #3, !dbg !18 %exec_user_code = icmp eq i32 %0, -1, !dbg !18 br i1 %exec_user_code, label %user_code.entry, label %common.ret, !dbg !18 @@ -81,7 +81,7 @@ br label %common.ret } -declare i32 @__kmpc_target_init(%struct.ident_t*, i8, i1, i1) local_unnamed_addr +declare i32 @__kmpc_target_init(%struct.ident_t*, i8, i1, i1, i32) local_unnamed_addr ; Function Attrs: convergent declare void @unknown() local_unnamed_addr #1 @@ -105,7 +105,7 @@ define weak void @__omp_offloading_2a_d80d3d_test_no_fallback_l20() local_unnamed_addr #4 !dbg !32 { entry: %captured_vars_addrs.i2.i = alloca [0 x i8*], align 8 - %0 = call i32 @__kmpc_target_init(%struct.ident_t* nonnull @7, i8 1, i1 true, i1 true) #3, !dbg !33 + %0 = call i32 @__kmpc_target_init(%struct.ident_t* nonnull @7, i8 1, i1 true, i1 true, i32 0) #3, !dbg !33 %exec_user_code = icmp eq i32 %0, -1, !dbg !33 br i1 %exec_user_code, label %user_code.entry, label %common.ret, !dbg !33 diff --git a/llvm/test/Transforms/OpenMP/deduplication_target.ll b/llvm/test/Transforms/OpenMP/deduplication_target.ll --- a/llvm/test/Transforms/OpenMP/deduplication_target.ll +++ b/llvm/test/Transforms/OpenMP/deduplication_target.ll @@ -20,7 +20,7 @@ ; CHECK-SAME: () #[[ATTR0:[0-9]+]] { ; CHECK-NEXT: entry: ; CHECK-NEXT: [[CAPTURED_VARS_ADDRS:%.*]] = alloca [0 x i8*], align 8 -; CHECK-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 true) +; CHECK-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 true, i32 0) ; CHECK-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB2:[0-9]+]]) ; CHECK-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 ; CHECK-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] @@ -32,7 +32,7 @@ ; entry: %captured_vars_addrs = alloca [0 x i8*], align 8 - %0 = call i32 @__kmpc_target_init(%struct.ident_t* @1, i8 2, i1 false, i1 true) + %0 = call i32 @__kmpc_target_init(%struct.ident_t* @1, i8 2, i1 false, i1 true, i32 0) %exec_user_code = icmp eq i32 %0, -1 br i1 %exec_user_code, label %user_code.entry, label %worker.exit @@ -46,7 +46,7 @@ ret void } -declare i32 @__kmpc_target_init(%struct.ident_t*, i8, i1, i1) +declare i32 @__kmpc_target_init(%struct.ident_t*, i8, i1, i1, i32) declare i32 @__kmpc_global_thread_num(%struct.ident_t*) #1 diff --git a/llvm/test/Transforms/OpenMP/fold_generic_main_thread.ll b/llvm/test/Transforms/OpenMP/fold_generic_main_thread.ll --- a/llvm/test/Transforms/OpenMP/fold_generic_main_thread.ll +++ b/llvm/test/Transforms/OpenMP/fold_generic_main_thread.ll @@ -10,7 +10,7 @@ define void @kernel() { ; CHECK-LABEL: define {{[^@]+}}@kernel() { -; CHECK-NEXT: [[CALL:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* nonnull @[[GLOB1:[0-9]+]], i8 1, i1 false, i1 false) +; CHECK-NEXT: [[CALL:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* nonnull @[[GLOB1:[0-9]+]], i8 1, i1 false, i1 false, i32 0) ; CHECK-NEXT: [[CMP:%.*]] = icmp eq i32 [[CALL]], -1 ; CHECK-NEXT: br i1 [[CMP]], label [[IF_THEN:%.*]], label [[IF_ELSE:%.*]] ; CHECK: if.then: @@ -24,7 +24,7 @@ ; CHECK-NEXT: ret void ; ; CHECK-DISABLED-LABEL: define {{[^@]+}}@kernel() { -; CHECK-DISABLED-NEXT: [[CALL:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* nonnull @[[GLOB1:[0-9]+]], i8 1, i1 false, i1 false) +; CHECK-DISABLED-NEXT: [[CALL:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* nonnull @[[GLOB1:[0-9]+]], i8 1, i1 false, i1 false, i32 0) ; CHECK-DISABLED-NEXT: [[CMP:%.*]] = icmp eq i32 [[CALL]], -1 ; CHECK-DISABLED-NEXT: br i1 [[CMP]], label [[IF_THEN:%.*]], label [[IF_ELSE:%.*]] ; CHECK-DISABLED: if.then: @@ -37,7 +37,7 @@ ; CHECK-DISABLED-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* null, i8 1, i1 true) ; CHECK-DISABLED-NEXT: ret void ; - %call = call i32 @__kmpc_target_init(%struct.ident_t* nonnull @1, i8 1, i1 false, i1 false) + %call = call i32 @__kmpc_target_init(%struct.ident_t* nonnull @1, i8 1, i1 false, i1 false, i32 0) %cmp = icmp eq i32 %call, -1 br i1 %cmp, label %if.then, label %if.else if.then: @@ -135,7 +135,7 @@ declare i32 @__kmpc_get_hardware_thread_id() -declare i32 @__kmpc_target_init(%struct.ident_t*, i8, i1, i1) +declare i32 @__kmpc_target_init(%struct.ident_t*, i8, i1, i1, i32) declare void @__kmpc_target_deinit(%struct.ident_t*, i8, i1) diff --git a/llvm/test/Transforms/OpenMP/get_hardware_num_threads_in_block_fold.ll b/llvm/test/Transforms/OpenMP/get_hardware_num_threads_in_block_fold.ll --- a/llvm/test/Transforms/OpenMP/get_hardware_num_threads_in_block_fold.ll +++ b/llvm/test/Transforms/OpenMP/get_hardware_num_threads_in_block_fold.ll @@ -18,14 +18,14 @@ define weak void @kernel0() #0 { ; CHECK-LABEL: define {{[^@]+}}@kernel0 ; CHECK-SAME: () #[[ATTR0:[0-9]+]] { -; CHECK-NEXT: [[I:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* null, i8 2, i1 false, i1 false) +; CHECK-NEXT: [[I:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* null, i8 2, i1 false, i1 false, i32 0) ; CHECK-NEXT: call void @helper0() #[[ATTR1:[0-9]+]] ; CHECK-NEXT: call void @helper1() #[[ATTR1]] ; CHECK-NEXT: call void @helper2() #[[ATTR1]] ; CHECK-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* null, i8 2, i1 false) ; CHECK-NEXT: ret void ; - %i = call i32 @__kmpc_target_init(%struct.ident_t* null, i8 2, i1 false, i1 false) + %i = call i32 @__kmpc_target_init(%struct.ident_t* null, i8 2, i1 false, i1 false, i32 0) call void @helper0() call void @helper1() call void @helper2() @@ -38,12 +38,12 @@ define weak void @kernel1() #0 { ; CHECK-LABEL: define {{[^@]+}}@kernel1 ; CHECK-SAME: () #[[ATTR0]] { -; CHECK-NEXT: [[I:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* null, i8 2, i1 false, i1 false) +; CHECK-NEXT: [[I:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* null, i8 2, i1 false, i1 false, i32 0) ; CHECK-NEXT: call void @helper1() #[[ATTR1]] ; CHECK-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* null, i8 2, i1 false) ; CHECK-NEXT: ret void ; - %i = call i32 @__kmpc_target_init(%struct.ident_t* null, i8 2, i1 false, i1 false) + %i = call i32 @__kmpc_target_init(%struct.ident_t* null, i8 2, i1 false, i1 false, i32 0) call void @helper1() call void @__kmpc_target_deinit(%struct.ident_t* null, i8 2, i1 false) ret void @@ -56,7 +56,7 @@ ; CHECK-SAME: () #[[ATTR0]] { ; CHECK-NEXT: entry: ; CHECK-NEXT: [[CAPTURED_VARS_ADDRS:%.*]] = alloca [0 x i8*], align 8 -; CHECK-NEXT: [[I:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* null, i8 2, i1 false, i1 false) +; CHECK-NEXT: [[I:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* null, i8 2, i1 false, i1 false, i32 0) ; CHECK-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[I]], -1 ; CHECK-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[COMMON_RET:%.*]] ; CHECK: common.ret: @@ -73,7 +73,7 @@ ; entry: %captured_vars_addrs = alloca [0 x i8*], align 8 - %i = call i32 @__kmpc_target_init(%struct.ident_t* null, i8 1, i1 true, i1 true) + %i = call i32 @__kmpc_target_init(%struct.ident_t* null, i8 1, i1 true, i1 true, i32 0) %exec_user_code = icmp eq i32 %i, -1 br i1 %exec_user_code, label %user_code.entry, label %common.ret @@ -179,7 +179,7 @@ } declare i32 @__kmpc_get_hardware_num_threads_in_block() -declare i32 @__kmpc_target_init(%struct.ident_t*, i8, i1 zeroext, i1 zeroext) #1 +declare i32 @__kmpc_target_init(%struct.ident_t*, i8, i1 zeroext, i1 zeroext, i32) #1 declare void @__kmpc_target_deinit(%struct.ident_t* nocapture readnone, i8, i1 zeroext) #1 declare void @__kmpc_parallel_51(%struct.ident_t*, i32, i32, i32, i32, i8*, i8*, i8**, i64) declare i32 @__kmpc_global_thread_num(%struct.ident_t*) diff --git a/llvm/test/Transforms/OpenMP/global_constructor.ll b/llvm/test/Transforms/OpenMP/global_constructor.ll --- a/llvm/test/Transforms/OpenMP/global_constructor.ll +++ b/llvm/test/Transforms/OpenMP/global_constructor.ll @@ -10,7 +10,7 @@ define weak void @__omp_offloading_fd02_85283c04_main_l11(double* nonnull align 8 dereferenceable(8) %X) local_unnamed_addr { entry: - %0 = tail call i32 @__kmpc_target_init(%struct.ident_t* nonnull @1, i8 2, i1 false, i1 false) #0 + %0 = tail call i32 @__kmpc_target_init(%struct.ident_t* nonnull @1, i8 2, i1 false, i1 false, i32 0) #0 %exec_user_code = icmp eq i32 %0, -1 br i1 %exec_user_code, label %user_code.entry, label %common.ret @@ -33,7 +33,7 @@ br label %common.ret } -declare i32 @__kmpc_target_init(%struct.ident_t*, i8, i1, i1) local_unnamed_addr +declare i32 @__kmpc_target_init(%struct.ident_t*, i8, i1, i1, i32) local_unnamed_addr declare void @__kmpc_target_deinit(%struct.ident_t*, i8, i1) local_unnamed_addr @@ -78,7 +78,7 @@ ; CHECK-LABEL: define {{[^@]+}}@__omp_offloading_fd02_85283c04_main_l11 ; CHECK-SAME: (double* nonnull align 8 dereferenceable(8) [[X:%.*]]) local_unnamed_addr { ; CHECK-NEXT: entry: -; CHECK-NEXT: [[TMP0:%.*]] = tail call i32 @__kmpc_target_init(%struct.ident_t* nonnull @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 false) #[[ATTR1:[0-9]+]] +; CHECK-NEXT: [[TMP0:%.*]] = tail call i32 @__kmpc_target_init(%struct.ident_t* nonnull @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 false, i32 0) #[[ATTR1:[0-9]+]] ; CHECK-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 ; CHECK-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[COMMON_RET:%.*]] ; CHECK: common.ret: diff --git a/llvm/test/Transforms/OpenMP/gpu_state_machine_function_ptr_replacement.ll b/llvm/test/Transforms/OpenMP/gpu_state_machine_function_ptr_replacement.ll --- a/llvm/test/Transforms/OpenMP/gpu_state_machine_function_ptr_replacement.ll +++ b/llvm/test/Transforms/OpenMP/gpu_state_machine_function_ptr_replacement.ll @@ -48,7 +48,7 @@ %.zero.addr = alloca i32, align 4 %.threadid_temp. = alloca i32, align 4 store i32 0, i32* %.zero.addr, align 4 - %0 = call i32 @__kmpc_target_init(%struct.ident_t* @1, i8 1, i1 true, i1 true) + %0 = call i32 @__kmpc_target_init(%struct.ident_t* @1, i8 1, i1 true, i1 true, i32 0) %exec_user_code = icmp eq i32 %0, -1 br i1 %exec_user_code, label %user_code.entry, label %worker.exit @@ -63,7 +63,7 @@ ret void } -declare i32 @__kmpc_target_init(%struct.ident_t*, i8, i1, i1) +declare i32 @__kmpc_target_init(%struct.ident_t*, i8, i1, i1, i32) declare void @unknown() define internal void @__omp_outlined__(i32* noalias %.global_tid., i32* noalias %.bound_tid.) { diff --git a/llvm/test/Transforms/OpenMP/is_spmd_exec_mode_fold.ll b/llvm/test/Transforms/OpenMP/is_spmd_exec_mode_fold.ll --- a/llvm/test/Transforms/OpenMP/is_spmd_exec_mode_fold.ll +++ b/llvm/test/Transforms/OpenMP/is_spmd_exec_mode_fold.ll @@ -21,14 +21,14 @@ ;. define weak void @is_spmd() { ; CHECK-LABEL: define {{[^@]+}}@is_spmd() { -; CHECK-NEXT: [[I:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* null, i8 2, i1 false, i1 false) +; CHECK-NEXT: [[I:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* null, i8 2, i1 false, i1 false, i32 0) ; CHECK-NEXT: call void @is_spmd_helper1() ; CHECK-NEXT: call void @is_spmd_helper2() ; CHECK-NEXT: call void @is_mixed_helper() ; CHECK-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* null, i8 2, i1 false) ; CHECK-NEXT: ret void ; - %i = call i32 @__kmpc_target_init(%struct.ident_t* null, i8 2, i1 false, i1 false) + %i = call i32 @__kmpc_target_init(%struct.ident_t* null, i8 2, i1 false, i1 false, i32 0) call void @is_spmd_helper1() call void @is_spmd_helper2() call void @is_mixed_helper() @@ -40,7 +40,7 @@ ; CHECK-LABEL: define {{[^@]+}}@will_be_spmd() { ; CHECK-NEXT: entry: ; CHECK-NEXT: [[CAPTURED_VARS_ADDRS:%.*]] = alloca [0 x i8*], align 8 -; CHECK-NEXT: [[I:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* null, i8 2, i1 false, i1 false) +; CHECK-NEXT: [[I:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* null, i8 2, i1 false, i1 false, i32 0) ; CHECK-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[I]], -1 ; CHECK-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[COMMON_RET:%.*]] ; CHECK: common.ret: @@ -55,7 +55,7 @@ ; entry: %captured_vars_addrs = alloca [0 x i8*], align 8 - %i = call i32 @__kmpc_target_init(%struct.ident_t* null, i8 1, i1 true, i1 true) + %i = call i32 @__kmpc_target_init(%struct.ident_t* null, i8 1, i1 true, i1 true, i32 0) %exec_user_code = icmp eq i32 %i, -1 br i1 %exec_user_code, label %user_code.entry, label %common.ret @@ -73,14 +73,14 @@ define weak void @non_spmd() { ; CHECK-LABEL: define {{[^@]+}}@non_spmd() { -; CHECK-NEXT: [[I:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* null, i8 1, i1 false, i1 false) +; CHECK-NEXT: [[I:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* null, i8 1, i1 false, i1 false, i32 0) ; CHECK-NEXT: call void @is_generic_helper1() ; CHECK-NEXT: call void @is_generic_helper2() ; CHECK-NEXT: call void @is_mixed_helper() ; CHECK-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* null, i8 1, i1 false) ; CHECK-NEXT: ret void ; - %i = call i32 @__kmpc_target_init(%struct.ident_t* null, i8 1, i1 false, i1 false) + %i = call i32 @__kmpc_target_init(%struct.ident_t* null, i8 1, i1 false, i1 false, i32 0) call void @is_generic_helper1() call void @is_generic_helper2() call void @is_mixed_helper() @@ -90,14 +90,14 @@ define weak void @will_not_be_spmd() { ; CHECK-LABEL: define {{[^@]+}}@will_not_be_spmd() { -; CHECK-NEXT: [[I:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* null, i8 1, i1 false, i1 false) +; CHECK-NEXT: [[I:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* null, i8 1, i1 false, i1 false, i32 0) ; CHECK-NEXT: call void @is_generic_helper1() ; CHECK-NEXT: call void @is_generic_helper2() ; CHECK-NEXT: call void @is_mixed_helper() ; CHECK-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* null, i8 1, i1 false) ; CHECK-NEXT: ret void ; - %i = call i32 @__kmpc_target_init(%struct.ident_t* null, i8 1, i1 false, i1 false) + %i = call i32 @__kmpc_target_init(%struct.ident_t* null, i8 1, i1 false, i1 false, i32 0) call void @is_generic_helper1() call void @is_generic_helper2() call void @is_mixed_helper() @@ -197,7 +197,7 @@ declare void @spmd_compatible() "llvm.assume"="ompx_spmd_amenable" declare i8 @__kmpc_is_spmd_exec_mode() -declare i32 @__kmpc_target_init(%struct.ident_t*, i8, i1 zeroext, i1 zeroext) +declare i32 @__kmpc_target_init(%struct.ident_t*, i8, i1 zeroext, i1 zeroext, i32) declare void @__kmpc_target_deinit(%struct.ident_t* nocapture readnone, i8, i1 zeroext) declare void @__kmpc_parallel_51(%struct.ident_t*, i32, i32, i32, i32, i8*, i8*, i8**, i64) declare i32 @__kmpc_global_thread_num(%struct.ident_t*) diff --git a/llvm/test/Transforms/OpenMP/parallel_level_fold.ll b/llvm/test/Transforms/OpenMP/parallel_level_fold.ll --- a/llvm/test/Transforms/OpenMP/parallel_level_fold.ll +++ b/llvm/test/Transforms/OpenMP/parallel_level_fold.ll @@ -19,13 +19,13 @@ ;. define weak void @none_spmd() { ; CHECK-LABEL: define {{[^@]+}}@none_spmd() { -; CHECK-NEXT: [[I:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* null, i8 1, i1 false, i1 false) +; CHECK-NEXT: [[I:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* null, i8 1, i1 false, i1 false, i32 0) ; CHECK-NEXT: call void @none_spmd_helper() ; CHECK-NEXT: call void @mixed_helper() ; CHECK-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* null, i8 1, i1 false) ; CHECK-NEXT: ret void ; - %i = call i32 @__kmpc_target_init(%struct.ident_t* null, i8 1, i1 false, i1 false) + %i = call i32 @__kmpc_target_init(%struct.ident_t* null, i8 1, i1 false, i1 false, i32 0) call void @none_spmd_helper() call void @mixed_helper() call void @__kmpc_target_deinit(%struct.ident_t* null, i8 1, i1 false) @@ -34,13 +34,13 @@ define weak void @spmd() { ; CHECK-LABEL: define {{[^@]+}}@spmd() { -; CHECK-NEXT: [[I:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* null, i8 2, i1 false, i1 false) +; CHECK-NEXT: [[I:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* null, i8 2, i1 false, i1 false, i32 0) ; CHECK-NEXT: call void @spmd_helper() ; CHECK-NEXT: call void @mixed_helper() ; CHECK-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* null, i8 2, i1 false) ; CHECK-NEXT: ret void ; - %i = call i32 @__kmpc_target_init(%struct.ident_t* null, i8 2, i1 false, i1 false) + %i = call i32 @__kmpc_target_init(%struct.ident_t* null, i8 2, i1 false, i1 false, i32 0) call void @spmd_helper() call void @mixed_helper() call void @__kmpc_target_deinit(%struct.ident_t* null, i8 2, i1 false) @@ -49,13 +49,13 @@ define weak void @parallel() { ; CHECK-LABEL: define {{[^@]+}}@parallel() { -; CHECK-NEXT: [[I:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* align 1073741824 null, i8 2, i1 false, i1 false) +; CHECK-NEXT: [[I:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* align 1073741824 null, i8 2, i1 false, i1 false, i32 0) ; CHECK-NEXT: call void @spmd_helper() ; CHECK-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* noalias noundef align 1073741824 null, i32 noundef 0, i32 noundef 0, i32 noundef 0, i32 noundef 0, i8* noalias noundef align 1073741824 null, i8* noalias noundef align 1073741824 null, i8** noalias noundef align 1073741824 null, i64 noundef 0) ; CHECK-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* null, i8 2, i1 false) ; CHECK-NEXT: ret void ; - %i = call i32 @__kmpc_target_init(%struct.ident_t* null, i8 2, i1 false, i1 false) + %i = call i32 @__kmpc_target_init(%struct.ident_t* null, i8 2, i1 false, i1 false, i32 0) call void @spmd_helper() call void @__kmpc_parallel_51(%struct.ident_t* null, i32 0, i32 0, i32 0, i32 0, i8* null, i8* null, i8** null, i64 0) call void @__kmpc_target_deinit(%struct.ident_t* null, i8 2, i1 false) @@ -130,7 +130,7 @@ declare void @foo() declare void @bar() declare i8 @__kmpc_parallel_level() -declare i32 @__kmpc_target_init(%struct.ident_t*, i8 zeroext, i1 zeroext, i1 zeroext) #1 +declare i32 @__kmpc_target_init(%struct.ident_t*, i8 zeroext, i1 zeroext, i1 zeroext, i32) #1 declare void @__kmpc_target_deinit(%struct.ident_t* nocapture readnone, i8 zeroext, i1 zeroext) #1 !llvm.module.flags = !{!0, !1} diff --git a/llvm/test/Transforms/OpenMP/remove_globalization.ll b/llvm/test/Transforms/OpenMP/remove_globalization.ll --- a/llvm/test/Transforms/OpenMP/remove_globalization.ll +++ b/llvm/test/Transforms/OpenMP/remove_globalization.ll @@ -14,7 +14,7 @@ %struct.ident_t = type { i32, i32, i32, i32, i8* } -declare i32 @__kmpc_target_init(%struct.ident_t*, i8, i1, i1) +declare i32 @__kmpc_target_init(%struct.ident_t*, i8, i1, i1, i32) declare void @__kmpc_target_deinit(%struct.ident_t*, i8, i1) ;. @@ -25,7 +25,7 @@ define void @kernel() { ; CHECK-LABEL: define {{[^@]+}}@kernel() { ; CHECK-NEXT: entry: -; CHECK-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* nonnull null, i8 1, i1 false, i1 true) +; CHECK-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* nonnull null, i8 1, i1 false, i1 true, i32 0) ; CHECK-NEXT: call void @foo() #[[ATTR4:[0-9]+]] ; CHECK-NEXT: call void @bar() #[[ATTR4]] ; CHECK-NEXT: call void @unknown_no_openmp() @@ -34,7 +34,7 @@ ; ; CHECK-DISABLED-LABEL: define {{[^@]+}}@kernel() { ; CHECK-DISABLED-NEXT: entry: -; CHECK-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* nonnull null, i8 1, i1 false, i1 true) +; CHECK-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* nonnull null, i8 1, i1 false, i1 true, i32 0) ; CHECK-DISABLED-NEXT: call void @foo() #[[ATTR4:[0-9]+]] ; CHECK-DISABLED-NEXT: call void @bar() #[[ATTR4]] ; CHECK-DISABLED-NEXT: call void @unknown_no_openmp() @@ -42,7 +42,7 @@ ; CHECK-DISABLED-NEXT: ret void ; entry: - %0 = call i32 @__kmpc_target_init(%struct.ident_t* nonnull null, i8 1, i1 true, i1 true) + %0 = call i32 @__kmpc_target_init(%struct.ident_t* nonnull null, i8 1, i1 true, i1 true, i32 0) call void @foo() call void @bar() call void @unknown_no_openmp() diff --git a/llvm/test/Transforms/OpenMP/replace_globalization.ll b/llvm/test/Transforms/OpenMP/replace_globalization.ll --- a/llvm/test/Transforms/OpenMP/replace_globalization.ll +++ b/llvm/test/Transforms/OpenMP/replace_globalization.ll @@ -19,7 +19,7 @@ define dso_local void @foo() { entry: - %c = call i32 @__kmpc_target_init(%struct.ident_t* @1, i8 1, i1 true, i1 true) + %c = call i32 @__kmpc_target_init(%struct.ident_t* @1, i8 1, i1 true, i1 true, i32 0) %x = call i8* @__kmpc_alloc_shared(i64 4) call void @unknown_no_openmp() %x_on_stack = bitcast i8* %x to i32* @@ -31,7 +31,7 @@ } define void @bar() { - %c = call i32 @__kmpc_target_init(%struct.ident_t* @1, i8 1, i1 true, i1 true) + %c = call i32 @__kmpc_target_init(%struct.ident_t* @1, i8 1, i1 true, i1 true, i32 0) call void @unknown_no_openmp() %cmp = icmp eq i32 %c, -1 br i1 %cmp, label %master1, label %exit @@ -59,7 +59,7 @@ } define void @baz_spmd() { - %c = call i32 @__kmpc_target_init(%struct.ident_t* @1, i8 2, i1 true, i1 true) + %c = call i32 @__kmpc_target_init(%struct.ident_t* @1, i8 2, i1 true, i1 true, i32 0) call void @unknown_no_openmp() %c0 = icmp eq i32 %c, -1 br i1 %c0, label %master3, label %exit @@ -91,7 +91,7 @@ declare i32 @llvm.nvvm.read.ptx.sreg.warpsize() -declare i32 @__kmpc_target_init(%struct.ident_t*, i8, i1, i1) +declare i32 @__kmpc_target_init(%struct.ident_t*, i8, i1, i1, i32) declare void @__kmpc_target_deinit(%struct.ident_t*, i8, i1) @@ -124,7 +124,7 @@ ;. ; CHECK-LABEL: define {{[^@]+}}@foo() { ; CHECK-NEXT: entry: -; CHECK-NEXT: [[C:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true) +; CHECK-NEXT: [[C:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true, i32 0) ; CHECK-NEXT: [[X:%.*]] = call i8* @__kmpc_alloc_shared(i64 4) #[[ATTR4:[0-9]+]] ; CHECK-NEXT: call void @unknown_no_openmp() ; CHECK-NEXT: call void @use.internalized(i8* nofree writeonly [[X]]) #[[ATTR5:[0-9]+]] @@ -134,7 +134,7 @@ ; ; ; CHECK-LABEL: define {{[^@]+}}@bar() { -; CHECK-NEXT: [[C:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true) +; CHECK-NEXT: [[C:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true, i32 0) ; CHECK-NEXT: call void @unknown_no_openmp() ; CHECK-NEXT: [[CMP:%.*]] = icmp eq i32 [[C]], -1 ; CHECK-NEXT: br i1 [[CMP]], label [[MASTER1:%.*]], label [[EXIT:%.*]] @@ -153,7 +153,7 @@ ; ; ; CHECK-LABEL: define {{[^@]+}}@baz_spmd() { -; CHECK-NEXT: [[C:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 true, i1 true) +; CHECK-NEXT: [[C:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 true, i1 true, i32 0) ; CHECK-NEXT: call void @unknown_no_openmp() ; CHECK-NEXT: [[C0:%.*]] = icmp eq i32 [[C]], -1 ; CHECK-NEXT: br i1 [[C0]], label [[MASTER3:%.*]], label [[EXIT:%.*]] diff --git a/llvm/test/Transforms/OpenMP/single_threaded_execution.ll b/llvm/test/Transforms/OpenMP/single_threaded_execution.ll --- a/llvm/test/Transforms/OpenMP/single_threaded_execution.ll +++ b/llvm/test/Transforms/OpenMP/single_threaded_execution.ll @@ -14,7 +14,7 @@ ; CHECK-NOT: [openmp-opt] Basic block @kernel if.else is executed by a single thread. ; CHECK-NOT: [openmp-opt] Basic block @kernel if.end is executed by a single thread. define void @kernel() { - %call = call i32 @__kmpc_target_init(%struct.ident_t* nonnull @1, i8 1, i1 false, i1 false) + %call = call i32 @__kmpc_target_init(%struct.ident_t* nonnull @1, i8 1, i1 false, i1 false, i32 0) %cmp = icmp eq i32 %call, -1 br i1 %cmp, label %if.then, label %if.else if.then: @@ -105,7 +105,7 @@ declare void @__kmpc_kernel_prepare_parallel(i8*) -declare i32 @__kmpc_target_init(%struct.ident_t*, i8, i1, i1) +declare i32 @__kmpc_target_init(%struct.ident_t*, i8, i1, i1, i32) declare void @__kmpc_target_deinit(%struct.ident_t*, i8, i1) diff --git a/llvm/test/Transforms/OpenMP/spmdization.ll b/llvm/test/Transforms/OpenMP/spmdization.ll --- a/llvm/test/Transforms/OpenMP/spmdization.ll +++ b/llvm/test/Transforms/OpenMP/spmdization.ll @@ -147,7 +147,7 @@ ; AMDGPU-NEXT: entry: ; AMDGPU-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 ; AMDGPU-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 -; AMDGPU-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false) +; AMDGPU-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false, i32 0) ; AMDGPU-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 ; AMDGPU-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] ; AMDGPU: user_code.entry: @@ -164,7 +164,7 @@ ; NVPTX-NEXT: entry: ; NVPTX-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 ; NVPTX-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 -; NVPTX-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false) +; NVPTX-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false, i32 0) ; NVPTX-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 ; NVPTX-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] ; NVPTX: user_code.entry: @@ -183,7 +183,7 @@ ; AMDGPU-DISABLED-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 ; AMDGPU-DISABLED-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 ; AMDGPU-DISABLED-NEXT: store i32 0, i32* [[DOTZERO_ADDR]], align 4 -; AMDGPU-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true) +; AMDGPU-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true, i32 0) ; AMDGPU-DISABLED-NEXT: [[THREAD_IS_WORKER:%.*]] = icmp ne i32 [[TMP0]], -1 ; AMDGPU-DISABLED-NEXT: br i1 [[THREAD_IS_WORKER]], label [[WORKER_STATE_MACHINE_BEGIN:%.*]], label [[THREAD_USER_CODE_CHECK:%.*]] ; AMDGPU-DISABLED: worker_state_machine.begin: @@ -232,7 +232,7 @@ ; NVPTX-DISABLED-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 ; NVPTX-DISABLED-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 ; NVPTX-DISABLED-NEXT: store i32 0, i32* [[DOTZERO_ADDR]], align 4 -; NVPTX-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true) +; NVPTX-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true, i32 0) ; NVPTX-DISABLED-NEXT: [[THREAD_IS_WORKER:%.*]] = icmp ne i32 [[TMP0]], -1 ; NVPTX-DISABLED-NEXT: br i1 [[THREAD_IS_WORKER]], label [[WORKER_STATE_MACHINE_BEGIN:%.*]], label [[THREAD_USER_CODE_CHECK:%.*]] ; NVPTX-DISABLED: worker_state_machine.begin: @@ -277,7 +277,7 @@ %.zero.addr = alloca i32, align 4 %.threadid_temp. = alloca i32, align 4 store i32 0, i32* %.zero.addr, align 4 - %0 = call i32 @__kmpc_target_init(%struct.ident_t* @1, i8 1, i1 true, i1 true) + %0 = call i32 @__kmpc_target_init(%struct.ident_t* @1, i8 1, i1 true, i1 true, i32 0) %exec_user_code = icmp eq i32 %0, -1 br i1 %exec_user_code, label %user_code.entry, label %worker.exit @@ -292,7 +292,7 @@ ret void } -declare i32 @__kmpc_target_init(%struct.ident_t*, i8, i1, i1) +declare i32 @__kmpc_target_init(%struct.ident_t*, i8, i1, i1, i32) define internal void @__omp_outlined__(i32* noalias %.global_tid., i32* noalias %.bound_tid.) #0 { ; @@ -578,7 +578,7 @@ ; AMDGPU-NEXT: entry: ; AMDGPU-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 ; AMDGPU-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 -; AMDGPU-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false) +; AMDGPU-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false, i32 0) ; AMDGPU-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 ; AMDGPU-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] ; AMDGPU: user_code.entry: @@ -595,7 +595,7 @@ ; NVPTX-NEXT: entry: ; NVPTX-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 ; NVPTX-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 -; NVPTX-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false) +; NVPTX-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false, i32 0) ; NVPTX-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 ; NVPTX-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] ; NVPTX: user_code.entry: @@ -614,7 +614,7 @@ ; AMDGPU-DISABLED-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 ; AMDGPU-DISABLED-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 ; AMDGPU-DISABLED-NEXT: store i32 0, i32* [[DOTZERO_ADDR]], align 4 -; AMDGPU-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true) +; AMDGPU-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true, i32 0) ; AMDGPU-DISABLED-NEXT: [[THREAD_IS_WORKER:%.*]] = icmp ne i32 [[TMP0]], -1 ; AMDGPU-DISABLED-NEXT: br i1 [[THREAD_IS_WORKER]], label [[WORKER_STATE_MACHINE_BEGIN:%.*]], label [[THREAD_USER_CODE_CHECK:%.*]] ; AMDGPU-DISABLED: worker_state_machine.begin: @@ -663,7 +663,7 @@ ; NVPTX-DISABLED-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 ; NVPTX-DISABLED-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 ; NVPTX-DISABLED-NEXT: store i32 0, i32* [[DOTZERO_ADDR]], align 4 -; NVPTX-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true) +; NVPTX-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true, i32 0) ; NVPTX-DISABLED-NEXT: [[THREAD_IS_WORKER:%.*]] = icmp ne i32 [[TMP0]], -1 ; NVPTX-DISABLED-NEXT: br i1 [[THREAD_IS_WORKER]], label [[WORKER_STATE_MACHINE_BEGIN:%.*]], label [[THREAD_USER_CODE_CHECK:%.*]] ; NVPTX-DISABLED: worker_state_machine.begin: @@ -708,7 +708,7 @@ %.zero.addr = alloca i32, align 4 %.threadid_temp. = alloca i32, align 4 store i32 0, i32* %.zero.addr, align 4 - %0 = call i32 @__kmpc_target_init(%struct.ident_t* @1, i8 1, i1 true, i1 true) + %0 = call i32 @__kmpc_target_init(%struct.ident_t* @1, i8 1, i1 true, i1 true, i32 0) %exec_user_code = icmp eq i32 %0, -1 br i1 %exec_user_code, label %user_code.entry, label %worker.exit @@ -1015,7 +1015,7 @@ ; AMDGPU-NEXT: entry: ; AMDGPU-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 ; AMDGPU-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 -; AMDGPU-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false) +; AMDGPU-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false, i32 0) ; AMDGPU-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 ; AMDGPU-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] ; AMDGPU: user_code.entry: @@ -1032,7 +1032,7 @@ ; NVPTX-NEXT: entry: ; NVPTX-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 ; NVPTX-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 -; NVPTX-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false) +; NVPTX-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false, i32 0) ; NVPTX-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 ; NVPTX-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] ; NVPTX: user_code.entry: @@ -1051,7 +1051,7 @@ ; AMDGPU-DISABLED-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 ; AMDGPU-DISABLED-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 ; AMDGPU-DISABLED-NEXT: store i32 0, i32* [[DOTZERO_ADDR]], align 4 -; AMDGPU-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true) +; AMDGPU-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true, i32 0) ; AMDGPU-DISABLED-NEXT: [[THREAD_IS_WORKER:%.*]] = icmp ne i32 [[TMP0]], -1 ; AMDGPU-DISABLED-NEXT: br i1 [[THREAD_IS_WORKER]], label [[WORKER_STATE_MACHINE_BEGIN:%.*]], label [[THREAD_USER_CODE_CHECK:%.*]] ; AMDGPU-DISABLED: worker_state_machine.begin: @@ -1100,7 +1100,7 @@ ; NVPTX-DISABLED-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 ; NVPTX-DISABLED-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 ; NVPTX-DISABLED-NEXT: store i32 0, i32* [[DOTZERO_ADDR]], align 4 -; NVPTX-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true) +; NVPTX-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true, i32 0) ; NVPTX-DISABLED-NEXT: [[THREAD_IS_WORKER:%.*]] = icmp ne i32 [[TMP0]], -1 ; NVPTX-DISABLED-NEXT: br i1 [[THREAD_IS_WORKER]], label [[WORKER_STATE_MACHINE_BEGIN:%.*]], label [[THREAD_USER_CODE_CHECK:%.*]] ; NVPTX-DISABLED: worker_state_machine.begin: @@ -1145,7 +1145,7 @@ %.zero.addr = alloca i32, align 4 %.threadid_temp. = alloca i32, align 4 store i32 0, i32* %.zero.addr, align 4 - %0 = call i32 @__kmpc_target_init(%struct.ident_t* @1, i8 1, i1 true, i1 true) + %0 = call i32 @__kmpc_target_init(%struct.ident_t* @1, i8 1, i1 true, i1 true, i32 0) %exec_user_code = icmp eq i32 %0, -1 br i1 %exec_user_code, label %user_code.entry, label %worker.exit @@ -1492,7 +1492,7 @@ ; AMDGPU-NEXT: entry: ; AMDGPU-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 ; AMDGPU-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 -; AMDGPU-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false) +; AMDGPU-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false, i32 0) ; AMDGPU-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 ; AMDGPU-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] ; AMDGPU: user_code.entry: @@ -1509,7 +1509,7 @@ ; NVPTX-NEXT: entry: ; NVPTX-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 ; NVPTX-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 -; NVPTX-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false) +; NVPTX-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false, i32 0) ; NVPTX-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 ; NVPTX-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] ; NVPTX: user_code.entry: @@ -1528,7 +1528,7 @@ ; AMDGPU-DISABLED-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 ; AMDGPU-DISABLED-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 ; AMDGPU-DISABLED-NEXT: store i32 0, i32* [[DOTZERO_ADDR]], align 4 -; AMDGPU-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true) +; AMDGPU-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true, i32 0) ; AMDGPU-DISABLED-NEXT: [[THREAD_IS_WORKER:%.*]] = icmp ne i32 [[TMP0]], -1 ; AMDGPU-DISABLED-NEXT: br i1 [[THREAD_IS_WORKER]], label [[WORKER_STATE_MACHINE_BEGIN:%.*]], label [[THREAD_USER_CODE_CHECK:%.*]] ; AMDGPU-DISABLED: worker_state_machine.begin: @@ -1577,7 +1577,7 @@ ; NVPTX-DISABLED-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 ; NVPTX-DISABLED-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 ; NVPTX-DISABLED-NEXT: store i32 0, i32* [[DOTZERO_ADDR]], align 4 -; NVPTX-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true) +; NVPTX-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true, i32 0) ; NVPTX-DISABLED-NEXT: [[THREAD_IS_WORKER:%.*]] = icmp ne i32 [[TMP0]], -1 ; NVPTX-DISABLED-NEXT: br i1 [[THREAD_IS_WORKER]], label [[WORKER_STATE_MACHINE_BEGIN:%.*]], label [[THREAD_USER_CODE_CHECK:%.*]] ; NVPTX-DISABLED: worker_state_machine.begin: @@ -1622,7 +1622,7 @@ %.zero.addr = alloca i32, align 4 %.threadid_temp. = alloca i32, align 4 store i32 0, i32* %.zero.addr, align 4 - %0 = call i32 @__kmpc_target_init(%struct.ident_t* @1, i8 1, i1 true, i1 true) + %0 = call i32 @__kmpc_target_init(%struct.ident_t* @1, i8 1, i1 true, i1 true, i32 0) %exec_user_code = icmp eq i32 %0, -1 br i1 %exec_user_code, label %user_code.entry, label %worker.exit @@ -2005,7 +2005,7 @@ ; AMDGPU-NEXT: [[WORKER_WORK_FN_ADDR:%.*]] = alloca i8*, align 8, addrspace(5) ; AMDGPU-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 ; AMDGPU-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 -; AMDGPU-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true) +; AMDGPU-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true, i32 0) ; AMDGPU-NEXT: [[THREAD_IS_WORKER:%.*]] = icmp ne i32 [[TMP0]], -1 ; AMDGPU-NEXT: br i1 [[THREAD_IS_WORKER]], label [[WORKER_STATE_MACHINE_BEGIN:%.*]], label [[THREAD_USER_CODE_CHECK:%.*]] ; AMDGPU: worker_state_machine.begin: @@ -2046,7 +2046,7 @@ ; NVPTX-NEXT: [[WORKER_WORK_FN_ADDR:%.*]] = alloca i8*, align 8 ; NVPTX-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 ; NVPTX-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 -; NVPTX-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true) +; NVPTX-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true, i32 0) ; NVPTX-NEXT: [[THREAD_IS_WORKER:%.*]] = icmp ne i32 [[TMP0]], -1 ; NVPTX-NEXT: br i1 [[THREAD_IS_WORKER]], label [[WORKER_STATE_MACHINE_BEGIN:%.*]], label [[THREAD_USER_CODE_CHECK:%.*]] ; NVPTX: worker_state_machine.begin: @@ -2087,7 +2087,7 @@ ; AMDGPU-DISABLED-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 ; AMDGPU-DISABLED-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 ; AMDGPU-DISABLED-NEXT: store i32 0, i32* [[DOTZERO_ADDR]], align 4 -; AMDGPU-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true) +; AMDGPU-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true, i32 0) ; AMDGPU-DISABLED-NEXT: [[THREAD_IS_WORKER:%.*]] = icmp ne i32 [[TMP0]], -1 ; AMDGPU-DISABLED-NEXT: br i1 [[THREAD_IS_WORKER]], label [[WORKER_STATE_MACHINE_BEGIN:%.*]], label [[THREAD_USER_CODE_CHECK:%.*]] ; AMDGPU-DISABLED: worker_state_machine.begin: @@ -2129,7 +2129,7 @@ ; NVPTX-DISABLED-NEXT: [[DOTZERO_ADDR:%.*]] = alloca i32, align 4 ; NVPTX-DISABLED-NEXT: [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4 ; NVPTX-DISABLED-NEXT: store i32 0, i32* [[DOTZERO_ADDR]], align 4 -; NVPTX-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true) +; NVPTX-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true, i32 0) ; NVPTX-DISABLED-NEXT: [[THREAD_IS_WORKER:%.*]] = icmp ne i32 [[TMP0]], -1 ; NVPTX-DISABLED-NEXT: br i1 [[THREAD_IS_WORKER]], label [[WORKER_STATE_MACHINE_BEGIN:%.*]], label [[THREAD_USER_CODE_CHECK:%.*]] ; NVPTX-DISABLED: worker_state_machine.begin: @@ -2167,7 +2167,7 @@ %.zero.addr = alloca i32, align 4 %.threadid_temp. = alloca i32, align 4 store i32 0, i32* %.zero.addr, align 4 - %0 = call i32 @__kmpc_target_init(%struct.ident_t* @1, i8 1, i1 true, i1 true) + %0 = call i32 @__kmpc_target_init(%struct.ident_t* @1, i8 1, i1 true, i1 true, i32 0) %exec_user_code = icmp eq i32 %0, -1 br i1 %exec_user_code, label %user_code.entry, label %worker.exit diff --git a/llvm/test/Transforms/OpenMP/spmdization_assumes.ll b/llvm/test/Transforms/OpenMP/spmdization_assumes.ll --- a/llvm/test/Transforms/OpenMP/spmdization_assumes.ll +++ b/llvm/test/Transforms/OpenMP/spmdization_assumes.ll @@ -31,7 +31,7 @@ ; CHECK-SAME: (double* nonnull align 8 dereferenceable(8) [[X:%.*]]) local_unnamed_addr #[[ATTR0:[0-9]+]] { ; CHECK-NEXT: entry: ; CHECK-NEXT: [[CAPTURED_VARS_ADDRS:%.*]] = alloca [0 x i8*], align 8 -; CHECK-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* nonnull @[[GLOB1]], i8 2, i1 false, i1 false) #[[ATTR3:[0-9]+]] +; CHECK-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* nonnull @[[GLOB1]], i8 2, i1 false, i1 false, i32 0) #[[ATTR3:[0-9]+]] ; CHECK-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 ; CHECK-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[COMMON_RET:%.*]] ; CHECK: common.ret: @@ -60,7 +60,7 @@ ; entry: %captured_vars_addrs = alloca [0 x i8*], align 8 - %0 = call i32 @__kmpc_target_init(%struct.ident_t* nonnull @1, i8 1, i1 true, i1 true) #3 + %0 = call i32 @__kmpc_target_init(%struct.ident_t* nonnull @1, i8 1, i1 true, i1 true, i32 0) #3 %exec_user_code = icmp eq i32 %0, -1 br i1 %exec_user_code, label %user_code.entry, label %common.ret @@ -77,7 +77,7 @@ br label %common.ret } -declare i32 @__kmpc_target_init(%struct.ident_t*, i8, i1, i1) local_unnamed_addr +declare i32 @__kmpc_target_init(%struct.ident_t*, i8, i1, i1, i32) local_unnamed_addr ; Function Attrs: alwaysinline mustprogress nofree norecurse nosync nounwind readnone willreturn define internal void @__omp_outlined__(i32* noalias nocapture %.global_tid., i32* noalias nocapture %.bound_tid.) #1 { diff --git a/llvm/test/Transforms/OpenMP/spmdization_guarding.ll b/llvm/test/Transforms/OpenMP/spmdization_guarding.ll --- a/llvm/test/Transforms/OpenMP/spmdization_guarding.ll +++ b/llvm/test/Transforms/OpenMP/spmdization_guarding.ll @@ -59,7 +59,7 @@ ; CHECK-SAME: (i32* [[X:%.*]], i64 [[N:%.*]]) #[[ATTR0:[0-9]+]] { ; CHECK-NEXT: entry: ; CHECK-NEXT: [[N_ADDR_SROA_0_0_EXTRACT_TRUNC:%.*]] = trunc i64 [[N]] to i32 -; CHECK-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* nonnull @[[GLOB1]], i8 2, i1 false, i1 false) #[[ATTR4:[0-9]+]] +; CHECK-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* nonnull @[[GLOB1]], i8 2, i1 false, i1 false, i32 0) #[[ATTR4:[0-9]+]] ; CHECK-NEXT: [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1 ; CHECK-NEXT: br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]] ; CHECK: user_code.entry: @@ -176,7 +176,7 @@ ; CHECK-DISABLED-NEXT: entry: ; CHECK-DISABLED-NEXT: [[WORKER_WORK_FN_ADDR:%.*]] = alloca i8*, align 8 ; CHECK-DISABLED-NEXT: [[N_ADDR_SROA_0_0_EXTRACT_TRUNC:%.*]] = trunc i64 [[N]] to i32 -; CHECK-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* nonnull @[[GLOB1]], i8 1, i1 false, i1 true) #[[ATTR4:[0-9]+]] +; CHECK-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* nonnull @[[GLOB1]], i8 1, i1 false, i1 true, i32 0) #[[ATTR4:[0-9]+]] ; CHECK-DISABLED-NEXT: [[THREAD_IS_WORKER:%.*]] = icmp ne i32 [[TMP0]], -1 ; CHECK-DISABLED-NEXT: br i1 [[THREAD_IS_WORKER]], label [[WORKER_STATE_MACHINE_BEGIN:%.*]], label [[THREAD_USER_CODE_CHECK:%.*]] ; CHECK-DISABLED: worker_state_machine.begin: @@ -252,7 +252,7 @@ ; entry: %N.addr.sroa.0.0.extract.trunc = trunc i64 %N to i32 - %0 = call i32 @__kmpc_target_init(%struct.ident_t* nonnull @1, i8 1, i1 true, i1 true) #3 + %0 = call i32 @__kmpc_target_init(%struct.ident_t* nonnull @1, i8 1, i1 true, i1 true, i32 0) #3 %exec_user_code = icmp eq i32 %0, -1 br i1 %exec_user_code, label %user_code.entry, label %worker.exit @@ -330,7 +330,7 @@ declare void @__kmpc_parallel_51(%struct.ident_t*, i32, i32, i32, i32, i8*, i8*, i8**, i64) -declare i32 @__kmpc_target_init(%struct.ident_t*, i8, i1, i1) +declare i32 @__kmpc_target_init(%struct.ident_t*, i8, i1, i1, i32) ; Function Attrs: convergent declare i32 @no_openmp(i32*) #1 diff --git a/llvm/test/Transforms/OpenMP/spmdization_remarks.ll b/llvm/test/Transforms/OpenMP/spmdization_remarks.ll --- a/llvm/test/Transforms/OpenMP/spmdization_remarks.ll +++ b/llvm/test/Transforms/OpenMP/spmdization_remarks.ll @@ -8,7 +8,6 @@ ; CHECK: remark: llvm/test/Transforms/OpenMP/spmdization_remarks.c:15:5: Call may contain unknown parallel regions. Use `__attribute__((assume("omp_no_parallelism")))` to override. ; CHECK: remark: llvm/test/Transforms/OpenMP/spmdization_remarks.c:20:1: Transformed generic-mode kernel to SPMD-mode. - ;; void unknown(void); ;; void known(void) { ;; #pragma omp parallel @@ -62,7 +61,7 @@ define weak void @__omp_offloading_2a_d80d3d_test_fallback_l11() local_unnamed_addr #0 !dbg !15 { entry: %captured_vars_addrs.i.i = alloca [0 x i8*], align 8 - %0 = call i32 @__kmpc_target_init(%struct.ident_t* nonnull @1, i8 1, i1 true, i1 true) #3, !dbg !18 + %0 = call i32 @__kmpc_target_init(%struct.ident_t* nonnull @1, i8 1, i1 true, i1 true, i32 0) #3, !dbg !18 %exec_user_code = icmp eq i32 %0, -1, !dbg !18 br i1 %exec_user_code, label %user_code.entry, label %common.ret, !dbg !18 @@ -83,7 +82,7 @@ br label %common.ret } -declare i32 @__kmpc_target_init(%struct.ident_t*, i8, i1, i1) local_unnamed_addr +declare i32 @__kmpc_target_init(%struct.ident_t*, i8, i1, i1, i32) local_unnamed_addr ; Function Attrs: convergent declare void @unknown() local_unnamed_addr #1 @@ -107,7 +106,7 @@ define weak void @__omp_offloading_2a_d80d3d_test_no_fallback_l20() local_unnamed_addr #4 !dbg !32 { entry: %captured_vars_addrs.i2.i = alloca [0 x i8*], align 8 - %0 = call i32 @__kmpc_target_init(%struct.ident_t* nonnull @7, i8 1, i1 true, i1 true) #3, !dbg !33 + %0 = call i32 @__kmpc_target_init(%struct.ident_t* nonnull @7, i8 1, i1 true, i1 true, i32 0) #3, !dbg !33 %exec_user_code = icmp eq i32 %0, -1, !dbg !33 br i1 %exec_user_code, label %user_code.entry, label %common.ret, !dbg !33