[clang] 423d34f - [OpenMP][Offloading] Change `bool IsSPMD` to `int8_t Mode` in `__kmpc_target_init` and `__kmpc_target_deinit`
Shilei Tian via cfe-commits
cfe-commits at lists.llvm.org
Wed Sep 22 14:16:47 PDT 2021
Author: Shilei Tian
Date: 2021-09-22T17:16:41-04:00
New Revision: 423d34f74a10ae122a67a18a76c2ead6e26924eb
URL: https://github.com/llvm/llvm-project/commit/423d34f74a10ae122a67a18a76c2ead6e26924eb
DIFF: https://github.com/llvm/llvm-project/commit/423d34f74a10ae122a67a18a76c2ead6e26924eb.diff
LOG: [OpenMP][Offloading] Change `bool IsSPMD` to `int8_t Mode` in `__kmpc_target_init` and `__kmpc_target_deinit`
This is a follow-up of D110029, which uses bitset to indicate execution mode. This patches makes the changes in the function call.
Reviewed By: jdoerfert
Differential Revision: https://reviews.llvm.org/D110279
Added:
Modified:
clang/test/OpenMP/amdgcn_target_codegen.cpp
clang/test/OpenMP/declare_target_codegen_globalization.cpp
clang/test/OpenMP/nvptx_SPMD_codegen.cpp
clang/test/OpenMP/nvptx_data_sharing.cpp
clang/test/OpenMP/nvptx_distribute_parallel_generic_mode_codegen.cpp
clang/test/OpenMP/nvptx_force_full_runtime_SPMD_codegen.cpp
clang/test/OpenMP/nvptx_lambda_capturing.cpp
clang/test/OpenMP/nvptx_multi_target_parallel_codegen.cpp
clang/test/OpenMP/nvptx_nested_parallel_codegen.cpp
clang/test/OpenMP/nvptx_parallel_codegen.cpp
clang/test/OpenMP/nvptx_parallel_for_codegen.cpp
clang/test/OpenMP/nvptx_target_codegen.cpp
clang/test/OpenMP/nvptx_target_parallel_codegen.cpp
clang/test/OpenMP/nvptx_target_parallel_num_threads_codegen.cpp
clang/test/OpenMP/nvptx_target_parallel_reduction_codegen.cpp
clang/test/OpenMP/nvptx_target_parallel_reduction_codegen_tbaa_PR46146.cpp
clang/test/OpenMP/nvptx_target_printf_codegen.c
clang/test/OpenMP/nvptx_target_simd_codegen.cpp
clang/test/OpenMP/nvptx_target_teams_codegen.cpp
clang/test/OpenMP/nvptx_target_teams_distribute_codegen.cpp
clang/test/OpenMP/nvptx_target_teams_distribute_parallel_for_codegen.cpp
clang/test/OpenMP/nvptx_target_teams_distribute_parallel_for_generic_mode_codegen.cpp
clang/test/OpenMP/nvptx_target_teams_distribute_parallel_for_simd_codegen.cpp
clang/test/OpenMP/nvptx_target_teams_distribute_simd_codegen.cpp
clang/test/OpenMP/nvptx_teams_codegen.cpp
clang/test/OpenMP/nvptx_teams_reduction_codegen.cpp
clang/test/OpenMP/target_parallel_debug_codegen.cpp
clang/test/OpenMP/target_parallel_for_debug_codegen.cpp
llvm/include/llvm/Frontend/OpenMP/OMPKinds.def
llvm/lib/Frontend/OpenMP/OMPIRBuilder.cpp
llvm/lib/Transforms/IPO/OpenMPOpt.cpp
llvm/test/Transforms/OpenMP/always_inline_device.ll
llvm/test/Transforms/OpenMP/custom_state_machines.ll
llvm/test/Transforms/OpenMP/custom_state_machines_remarks.ll
llvm/test/Transforms/OpenMP/deduplication_target.ll
llvm/test/Transforms/OpenMP/fold_generic_main_thread.ll
llvm/test/Transforms/OpenMP/get_hardware_num_threads_in_block_fold.ll
llvm/test/Transforms/OpenMP/global_constructor.ll
llvm/test/Transforms/OpenMP/gpu_state_machine_function_ptr_replacement.ll
llvm/test/Transforms/OpenMP/is_spmd_exec_mode_fold.ll
llvm/test/Transforms/OpenMP/parallel_level_fold.ll
llvm/test/Transforms/OpenMP/remove_globalization.ll
llvm/test/Transforms/OpenMP/replace_globalization.ll
llvm/test/Transforms/OpenMP/single_threaded_execution.ll
llvm/test/Transforms/OpenMP/spmdization.ll
llvm/test/Transforms/OpenMP/spmdization_assumes.ll
llvm/test/Transforms/OpenMP/spmdization_guarding.ll
llvm/test/Transforms/OpenMP/spmdization_remarks.ll
openmp/libomptarget/DeviceRTL/include/Interface.h
openmp/libomptarget/DeviceRTL/include/Types.h
openmp/libomptarget/DeviceRTL/src/Kernel.cpp
openmp/libomptarget/deviceRTLs/common/include/target.h
openmp/libomptarget/deviceRTLs/common/omptarget.h
openmp/libomptarget/deviceRTLs/common/src/omp_data.cu
openmp/libomptarget/deviceRTLs/common/src/omptarget.cu
openmp/libomptarget/deviceRTLs/common/src/support.cu
openmp/libomptarget/deviceRTLs/common/support.h
openmp/libomptarget/deviceRTLs/interface.h
Removed:
################################################################################
diff --git a/clang/test/OpenMP/amdgcn_target_codegen.cpp b/clang/test/OpenMP/amdgcn_target_codegen.cpp
index 3dbddae468eb5..9118af92fc613 100644
--- a/clang/test/OpenMP/amdgcn_target_codegen.cpp
+++ b/clang/test/OpenMP/amdgcn_target_codegen.cpp
@@ -13,7 +13,7 @@ int test_amdgcn_target_tid_threads() {
int arr[N];
-// CHECK: call i32 @__kmpc_target_init(%struct.ident_t* addrspacecast (%struct.ident_t addrspace(1)* @1 to %struct.ident_t*), i1 false, 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)
#pragma omp target
for (int i = 0; i < N; i++) {
arr[i] = 1;
@@ -27,7 +27,7 @@ int test_amdgcn_target_tid_threads_simd() {
int arr[N];
-// CHECK: call i32 @__kmpc_target_init(%struct.ident_t* addrspacecast (%struct.ident_t addrspace(1)* @1 to %struct.ident_t*), i1 true, 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)
#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
index 47610f7b84a8f..022a88c60a48b 100644
--- a/clang/test/OpenMP/declare_target_codegen_globalization.cpp
+++ b/clang/test/OpenMP/declare_target_codegen_globalization.cpp
@@ -31,7 +31,7 @@ int maini1() {
// 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]+]], i1 true, i1 false, i1 true)
+// CHECK1-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 true)
// 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:
@@ -41,7 +41,7 @@ int maini1() {
// CHECK1-NEXT: store i8* [[TMP4]], i8** [[TMP3]], align 8
// CHECK1-NEXT: [[TMP5:%.*]] = bitcast [1 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**
// CHECK1-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB2]], i32 [[TMP2]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*, i32*)* @__omp_outlined__ to i8*), i8* null, i8** [[TMP5]], i64 1)
-// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 true)
+// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 true)
// CHECK1-NEXT: ret void
// CHECK1: worker.exit:
// CHECK1-NEXT: ret void
diff --git a/clang/test/OpenMP/nvptx_SPMD_codegen.cpp b/clang/test/OpenMP/nvptx_SPMD_codegen.cpp
index bccff61fc6209..eebc41b44f849 100644
--- a/clang/test/OpenMP/nvptx_SPMD_codegen.cpp
+++ b/clang/test/OpenMP/nvptx_SPMD_codegen.cpp
@@ -21,28 +21,28 @@ int a;
// CHECK-NOT: @__omp_offloading_{{.+}}_exec_mode = weak constant i8 1
void foo() {
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 false)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false)
// CHECK-DAG: [[DISTR_LIGHT]]
// CHECK-DAG: [[FOR_LIGHT]]
// CHECK-DAG: [[LIGHT]]
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 false)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false)
// CHECK-DAG: [[DISTR_LIGHT]]
// CHECK-DAG: [[FOR_LIGHT]]
// CHECK-DAG: [[LIGHT]]
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 false)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false)
// CHECK-DAG: [[DISTR_LIGHT]]
// CHECK-DAG: [[FOR_LIGHT]]
// CHECK-DAG: [[LIGHT]]
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true)
// CHECK-DAG: [[DISTR_FULL]]
// CHECK-DAG: [[FULL]]
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true)
// CHECK-DAG: [[DISTR_FULL]]
// CHECK-DAG: [[FULL]]
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true)
// CHECK-DAG: [[DISTR_FULL]]
// CHECK-DAG: [[FULL]]
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true)
// CHECK-DAG: [[DISTR_FULL]]
// CHECK-DAG: [[FULL]]
#pragma omp target teams distribute parallel for simd if(a)
@@ -67,28 +67,28 @@ void foo() {
for (int i = 0; i < 10; ++i)
;
int a;
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 false)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false)
// CHECK-DAG: [[DISTR_LIGHT]]
// CHECK-DAG: [[FOR_LIGHT]]
// CHECK-DAG: [[LIGHT]]
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 false)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false)
// CHECK-DAG: [[DISTR_LIGHT]]
// CHECK-DAG: [[FOR_LIGHT]]
// CHECK-DAG: [[LIGHT]]
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 false)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false)
// CHECK-DAG: [[DISTR_LIGHT]]
// CHECK-DAG: [[FOR_LIGHT]]
// CHECK-DAG: [[LIGHT]]
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true)
// CHECK-DAG: [[DISTR_FULL]]
// CHECK-DAG: [[FULL]]
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true)
// CHECK-DAG: [[DISTR_FULL]]
// CHECK-DAG: [[FULL]]
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true)
// CHECK-DAG: [[DISTR_FULL]]
// CHECK-DAG: [[FULL]]
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true)
// CHECK-DAG: [[DISTR_FULL]]
// CHECK-DAG: [[FULL]]
#pragma omp target teams distribute parallel for lastprivate(a)
@@ -112,25 +112,25 @@ int a;
#pragma omp target teams distribute parallel for schedule(guided)
for (int i = 0; i < 10; ++i)
;
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 false)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false)
// CHECK-DAG: [[DISTR_LIGHT]]
// CHECK-DAG: [[FOR_LIGHT]]
// CHECK-DAG: [[LIGHT]]
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 false)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false)
// CHECK-DAG: [[DISTR_LIGHT]]
// CHECK-DAG: [[FOR_LIGHT]]
// CHECK-DAG: [[LIGHT]]
// CHECK: call i32 @__kmpc_target_init(
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true)
// CHECK-DAG: [[DISTR_FULL]]
// CHECK-DAG: [[FULL]]
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true)
// CHECK-DAG: [[DISTR_FULL]]
// CHECK-DAG: [[FULL]]
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true)
// CHECK-DAG: [[DISTR_FULL]]
// CHECK-DAG: [[FULL]]
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true)
// CHECK-DAG: [[DISTR_FULL]]
// CHECK-DAG: [[FULL]]
#pragma omp target teams
@@ -172,28 +172,28 @@ int a;
#pragma omp distribute parallel for simd schedule(guided)
for (int i = 0; i < 10; ++i)
;
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 false)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false)
// CHECK-DAG: [[DISTR_LIGHT]]
// CHECK-DAG: [[FOR_LIGHT]]
// CHECK-DAG: [[LIGHT]]
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 false)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false)
// CHECK-DAG: [[DISTR_LIGHT]]
// CHECK-DAG: [[FOR_LIGHT]]
// CHECK-DAG: [[LIGHT]]
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 false)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false)
// CHECK-DAG: [[DISTR_LIGHT]]
// CHECK-DAG: [[FOR_LIGHT]]
// CHECK-DAG: [[LIGHT]]
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true)
// CHECK-DAG: [[DISTR_FULL]]
// CHECK-DAG: [[FULL]]
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true)
// CHECK-DAG: [[DISTR_FULL]]
// CHECK-DAG: [[FULL]]
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true)
// CHECK-DAG: [[DISTR_FULL]]
// CHECK-DAG: [[FULL]]
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true)
// CHECK-DAG: [[DISTR_FULL]]
// CHECK-DAG: [[FULL]]
#pragma omp target teams
@@ -224,28 +224,28 @@ int a;
#pragma omp distribute parallel for schedule(guided)
for (int i = 0; i < 10; ++i)
;
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 false)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false)
// CHECK-DAG: [[DISTR_LIGHT]]
// CHECK-DAG: [[FOR_LIGHT]]
// CHECK-DAG: [[LIGHT]]
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 false)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false)
// CHECK-DAG: [[DISTR_LIGHT]]
// CHECK-DAG: [[FOR_LIGHT]]
// CHECK-DAG: [[LIGHT]]
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 false)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false)
// CHECK-DAG: [[DISTR_LIGHT]]
// CHECK-DAG: [[FOR_LIGHT]]
// CHECK-DAG: [[LIGHT]]
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true)
// CHECK-DAG: [[DISTR_FULL]]
// CHECK-DAG: [[FULL]]
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true)
// CHECK-DAG: [[DISTR_FULL]]
// CHECK-DAG: [[FULL]]
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true)
// CHECK-DAG: [[DISTR_FULL]]
// CHECK-DAG: [[FULL]]
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true)
// CHECK-DAG: [[DISTR_FULL]]
// CHECK-DAG: [[FULL]]
#pragma omp target
@@ -283,22 +283,22 @@ int a;
#pragma omp distribute parallel for schedule(guided)
for (int i = 0; i < 10; ++i)
;
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 false)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false)
// CHECK-DAG: [[FOR_LIGHT]]
// CHECK-DAG: [[LIGHT]]
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 false)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false)
// CHECK-DAG: [[FOR_LIGHT]]
// CHECK-DAG: [[LIGHT]]
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 false)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false)
// CHECK-DAG: [[FOR_LIGHT]]
// CHECK-DAG: [[LIGHT]]
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true)
// CHECK-DAG: [[FULL]]
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true)
// CHECK-DAG: [[FULL]]
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true)
// CHECK-DAG: [[FULL]]
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true)
// CHECK-DAG: [[FULL]]
#pragma omp target parallel for if(a)
for (int i = 0; i < 10; ++i)
@@ -321,28 +321,28 @@ int a;
#pragma omp target parallel for schedule(guided)
for (int i = 0; i < 10; ++i)
;
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 false)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false)
// CHECK-DAG: [[FOR_LIGHT]]
// CHECK-DAG: [[LIGHT]]
// CHECK-DAG: [[BAR_LIGHT]]
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 false)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false)
// CHECK-DAG: [[FOR_LIGHT]]
// CHECK-DAG: [[LIGHT]]
// CHECK-DAG: [[BAR_LIGHT]]
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 false)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false)
// CHECK-DAG: [[FOR_LIGHT]]
// CHECK-DAG: [[LIGHT]]
// CHECK-DAG: [[BAR_LIGHT]]
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true)
// CHECK-DAG: [[FULL]]
// CHECK-DAG: [[BAR_FULL]]
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true)
// CHECK-DAG: [[FULL]]
// CHECK-DAG: [[BAR_FULL]]
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true)
// CHECK-DAG: [[FULL]]
// CHECK-DAG: [[BAR_FULL]]
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true)
// CHECK-DAG: [[FULL]]
// CHECK-DAG: [[BAR_FULL]]
#pragma omp target parallel if(a)
@@ -373,27 +373,27 @@ int a;
#pragma omp for simd schedule(guided)
for (int i = 0; i < 10; ++i)
;
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true)
// CHECK-DAG: [[FULL]]
// CHECK-DAG: [[BAR_FULL]]
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 false)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false)
// CHECK-DAG: [[FOR_LIGHT]]
// CHECK-DAG: [[LIGHT]]
// CHECK-DAG: [[BAR_LIGHT]]
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 false)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false)
// CHECK-DAG: [[FOR_LIGHT]]
// CHECK-DAG: [[LIGHT]]
// CHECK-DAG: [[BAR_LIGHT]]
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true)
// CHECK-DAG: [[FULL]]
// CHECK-DAG: [[BAR_FULL]]
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true)
// CHECK-DAG: [[FULL]]
// CHECK-DAG: [[BAR_FULL]]
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true)
// CHECK-DAG: [[FULL]]
// CHECK-DAG: [[BAR_FULL]]
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true)
// CHECK-DAG: [[FULL]]
// CHECK-DAG: [[BAR_FULL]]
#pragma omp target
@@ -431,22 +431,22 @@ int a;
#pragma omp for simd schedule(guided)
for (int i = 0; i < 10; ++i)
;
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 false)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false)
// CHECK-DAG: [[FOR_LIGHT]]
// CHECK-DAG: [[LIGHT]]
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 false)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false)
// CHECK-DAG: [[FOR_LIGHT]]
// CHECK-DAG: [[LIGHT]]
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 false)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false)
// CHECK-DAG: [[FOR_LIGHT]]
// CHECK-DAG: [[LIGHT]]
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true)
// CHECK-DAG: [[FULL]]
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true)
// CHECK-DAG: [[FULL]]
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true)
// CHECK-DAG: [[FULL]]
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true)
// 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
index b71b597a1a481..d046189a4355f 100644
--- a/clang/test/OpenMP/nvptx_data_sharing.cpp
+++ b/clang/test/OpenMP/nvptx_data_sharing.cpp
@@ -393,7 +393,7 @@ void test_ds(){
// 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]+]], i1 false, i1 true, i1 true)
+// CHECK-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true)
// 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:
@@ -420,7 +420,7 @@ void test_ds(){
// CHECK-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB1]], i32 [[TMP1]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*, i32*, i32*)* @__omp_outlined__1 to i8*), i8* bitcast (void (i16, i32)* @__omp_outlined__1_wrapper to i8*), i8** [[TMP9]], i64 2)
// CHECK-NEXT: call void @__kmpc_free_shared(i8* [[B]], i64 4)
// CHECK-NEXT: call void @__kmpc_free_shared(i8* [[A]], i64 4)
-// CHECK-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK-NEXT: ret void
// CHECK: worker.exit:
// 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
index 07be3a95d8dfa..5ee867ab56a18 100644
--- 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 @@ int main(int argc, char **argv) {
// 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]+]], i1 true, i1 false, i1 false)
+// CHECK4-NEXT: [[TMP4:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 false)
// 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:
@@ -3006,7 +3006,7 @@ int main(int argc, char **argv) {
// 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 @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 false)
+// CHECK4-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 false)
// CHECK4-NEXT: ret void
// CHECK4: worker.exit:
// CHECK4-NEXT: ret void
@@ -3328,7 +3328,7 @@ int main(int argc, char **argv) {
// 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]+]], i1 true, i1 false, i1 false)
+// CHECK5-NEXT: [[TMP4:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 false)
// 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:
@@ -3338,7 +3338,7 @@ int main(int argc, char **argv) {
// 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 @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 false)
+// CHECK5-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 false)
// CHECK5-NEXT: ret void
// CHECK5: worker.exit:
// CHECK5-NEXT: ret void
@@ -3651,7 +3651,7 @@ int main(int argc, char **argv) {
// 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]+]], i1 true, i1 false, i1 false)
+// CHECK6-NEXT: [[TMP4:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 false)
// 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:
@@ -3661,7 +3661,7 @@ int main(int argc, char **argv) {
// 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 @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 false)
+// CHECK6-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 false)
// CHECK6-NEXT: ret void
// CHECK6: worker.exit:
// CHECK6-NEXT: ret void
diff --git a/clang/test/OpenMP/nvptx_force_full_runtime_SPMD_codegen.cpp b/clang/test/OpenMP/nvptx_force_full_runtime_SPMD_codegen.cpp
index 775b6794bac40..5f3838b53fc57 100644
--- 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({{.*}}, i1 true, i1 false, i1 true)
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, 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)
#pragma omp target teams distribute parallel for simd
for (int i = 0; i < 10; ++i)
;
@@ -40,13 +40,13 @@ void foo() {
for (int i = 0; i < 10; ++i)
;
int a;
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, 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)
#pragma omp target teams distribute parallel for lastprivate(a)
for (int i = 0; i < 10; ++i)
a = i;
@@ -68,13 +68,13 @@ int a;
#pragma omp target teams distribute parallel for schedule(guided)
for (int i = 0; i < 10; ++i)
;
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, 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)
#pragma omp target teams
#pragma omp distribute parallel for simd
for (int i = 0; i < 10; ++i)
@@ -103,13 +103,13 @@ int a;
#pragma omp distribute parallel for simd schedule(guided)
for (int i = 0; i < 10; ++i)
;
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, 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)
#pragma omp target teams
#pragma omp distribute parallel for
for (int i = 0; i < 10; ++i)
@@ -138,13 +138,13 @@ int a;
#pragma omp distribute parallel for schedule(guided)
for (int i = 0; i < 10; ++i)
;
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, 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)
#pragma omp target
#pragma omp teams
#pragma omp distribute parallel for
@@ -180,13 +180,13 @@ int a;
#pragma omp distribute parallel for schedule(guided)
for (int i = 0; i < 10; ++i)
;
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, 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)
#pragma omp target parallel for
for (int i = 0; i < 10; ++i)
;
@@ -208,13 +208,13 @@ int a;
#pragma omp target parallel for schedule(guided)
for (int i = 0; i < 10; ++i)
;
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, 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)
#pragma omp target parallel
#pragma omp for simd
for (int i = 0; i < 10; ++i)
@@ -243,13 +243,13 @@ int a;
#pragma omp for simd schedule(guided)
for (int i = 0; i < 10; ++i)
;
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, 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)
#pragma omp target
#pragma omp parallel
#pragma omp for simd ordered
@@ -285,13 +285,13 @@ int a;
#pragma omp for simd schedule(guided)
for (int i = 0; i < 10; ++i)
;
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, 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)
#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
index b19806df1aee9..ea59fe8b5aa48 100644
--- a/clang/test/OpenMP/nvptx_lambda_capturing.cpp
+++ b/clang/test/OpenMP/nvptx_lambda_capturing.cpp
@@ -397,7 +397,7 @@ int main(int argc, char **argv) {
//
//
// CHECK1-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}_main_l43
-// CHECK1-SAME: (i32* nonnull align 4 dereferenceable(4) [[ARGC:%.*]], i32* nonnull align 4 dereferenceable(4) [[B:%.*]], i32* nonnull align 4 dereferenceable(4) [[C:%.*]], i32* [[D:%.*]], i32* nonnull align 4 dereferenceable(4) [[A:%.*]], %class.anon* nonnull align 8 dereferenceable(40) [[L:%.*]]) #[[ATTR2:[0-9]+]] {
+// CHECK1-SAME: (i32* nonnull align 4 dereferenceable(4) [[ARGC:%.*]], i32* nonnull align 4 dereferenceable(4) [[B:%.*]], i32* nonnull align 4 dereferenceable(4) [[C:%.*]], i32* [[D:%.*]], i32* nonnull align 4 dereferenceable(4) [[A:%.*]], %class.anon* nonnull align 8 dereferenceable(40) [[L:%.*]]) #[[ATTR5:[0-9]+]] {
// CHECK1-NEXT: entry:
// CHECK1-NEXT: [[ARGC_ADDR:%.*]] = alloca i32*, align 8
// CHECK1-NEXT: [[B_ADDR:%.*]] = alloca i32*, align 8
@@ -630,7 +630,7 @@ int main(int argc, char **argv) {
//
//
// CHECK1-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZN1S3fooEv_l29
-// CHECK1-SAME: (%struct.S* [[THIS:%.*]], %class.anon.0* nonnull align 8 dereferenceable(8) [[L:%.*]]) #[[ATTR2]] {
+// CHECK1-SAME: (%struct.S* [[THIS:%.*]], %class.anon.0* nonnull align 8 dereferenceable(8) [[L:%.*]]) #[[ATTR5]] {
// CHECK1-NEXT: entry:
// CHECK1-NEXT: [[THIS_ADDR:%.*]] = alloca %struct.S*, align 8
// CHECK1-NEXT: [[L_ADDR:%.*]] = alloca %class.anon.0*, align 8
@@ -715,7 +715,7 @@ int main(int argc, char **argv) {
//
//
// CHECK1-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z3fooIZN1S3fooEvEUlvE_EiRKT__l18
-// CHECK1-SAME: (%class.anon.0* nonnull align 8 dereferenceable(8) [[T:%.*]]) #[[ATTR2]] {
+// CHECK1-SAME: (%class.anon.0* nonnull align 8 dereferenceable(8) [[T:%.*]]) #[[ATTR5]] {
// CHECK1-NEXT: entry:
// CHECK1-NEXT: [[T_ADDR:%.*]] = alloca %class.anon.0*, align 8
// CHECK1-NEXT: [[TMP:%.*]] = alloca %class.anon.0*, align 8
@@ -752,7 +752,7 @@ int main(int argc, char **argv) {
//
//
// CHECK1-LABEL: define {{[^@]+}}@.omp_offloading.requires_reg
-// CHECK1-SAME: () #[[ATTR5:[0-9]+]] {
+// CHECK1-SAME: () #[[ATTR6:[0-9]+]] {
// CHECK1-NEXT: entry:
// CHECK1-NEXT: call void @__tgt_register_requires(i64 1)
// CHECK1-NEXT: ret void
@@ -771,7 +771,7 @@ int main(int argc, char **argv) {
// 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]+]], i1 false, i1 true, i1 true)
+// CHECK2-NEXT: [[TMP2:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true)
// 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:
@@ -784,8 +784,8 @@ int main(int argc, char **argv) {
// CHECK2-NEXT: [[TMP7:%.*]] = getelementptr inbounds [[CLASS_ANON]], %class.anon* [[TMP6]], i32 0, i32 0
// CHECK2-NEXT: store %struct.S* [[TMP0]], %struct.S** [[TMP7]], align 8
// CHECK2-NEXT: [[TMP8:%.*]] = load %class.anon*, %class.anon** [[_TMP2]], align 8
-// CHECK2-NEXT: [[CALL:%.*]] = call i32 @_ZZN1S3fooEvENKUlvE_clEv(%class.anon* nonnull align 8 dereferenceable(8) [[TMP8]]) #[[ATTR4:[0-9]+]]
-// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK2-NEXT: [[CALL:%.*]] = call i32 @_ZZN1S3fooEvENKUlvE_clEv(%class.anon* nonnull align 8 dereferenceable(8) [[TMP8]]) #[[ATTR6:[0-9]+]]
+// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK2-NEXT: ret void
// CHECK2: worker.exit:
// CHECK2-NEXT: ret void
@@ -805,7 +805,7 @@ int main(int argc, char **argv) {
//
//
// CHECK2-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZN1S3fooEv_l29
-// CHECK2-SAME: (%struct.S* [[THIS:%.*]], %class.anon* nonnull align 8 dereferenceable(8) [[L:%.*]]) #[[ATTR1:[0-9]+]] {
+// CHECK2-SAME: (%struct.S* [[THIS:%.*]], %class.anon* nonnull align 8 dereferenceable(8) [[L:%.*]]) #[[ATTR3:[0-9]+]] {
// CHECK2-NEXT: entry:
// CHECK2-NEXT: [[THIS_ADDR:%.*]] = alloca %struct.S*, align 8
// CHECK2-NEXT: [[L_ADDR:%.*]] = alloca %class.anon*, align 8
@@ -816,7 +816,7 @@ int main(int argc, char **argv) {
// 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]], i1 true, i1 false, i1 true)
+// CHECK2-NEXT: [[TMP2:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true)
// 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:
@@ -830,7 +830,7 @@ int main(int argc, char **argv) {
// CHECK2-NEXT: store i8* [[TMP8]], i8** [[TMP7]], align 8
// CHECK2-NEXT: [[TMP9:%.*]] = bitcast [2 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**
// CHECK2-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB2]], i32 [[TMP3]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*, %struct.S*, %class.anon*)* @__omp_outlined__ to i8*), i8* null, i8** [[TMP9]], i64 2)
-// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 true)
+// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 true)
// CHECK2-NEXT: ret void
// CHECK2: worker.exit:
// CHECK2-NEXT: ret void
@@ -862,7 +862,7 @@ int main(int argc, char **argv) {
// CHECK2-NEXT: [[TMP6:%.*]] = getelementptr inbounds [[CLASS_ANON]], %class.anon* [[TMP5]], i32 0, i32 0
// CHECK2-NEXT: store %struct.S* [[TMP0]], %struct.S** [[TMP6]], align 8
// CHECK2-NEXT: [[TMP7:%.*]] = load %class.anon*, %class.anon** [[_TMP2]], align 8
-// CHECK2-NEXT: [[CALL:%.*]] = call i32 @_ZZN1S3fooEvENKUlvE_clEv(%class.anon* nonnull align 8 dereferenceable(8) [[TMP7]]) #[[ATTR4]]
+// CHECK2-NEXT: [[CALL:%.*]] = call i32 @_ZZN1S3fooEvENKUlvE_clEv(%class.anon* nonnull align 8 dereferenceable(8) [[TMP7]]) #[[ATTR6]]
// CHECK2-NEXT: ret void
//
//
@@ -898,7 +898,7 @@ int main(int argc, char **argv) {
// 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]], i1 false, i1 true, i1 true)
+// CHECK2-NEXT: [[TMP4:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
// 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:
@@ -929,15 +929,15 @@ int main(int argc, char **argv) {
// CHECK2-NEXT: [[TMP19:%.*]] = getelementptr inbounds [[CLASS_ANON_0]], %class.anon.0* [[TMP12]], i32 0, i32 4
// CHECK2-NEXT: store i32* [[TMP2]], i32** [[TMP19]], align 8
// CHECK2-NEXT: [[TMP20:%.*]] = load %class.anon.0*, %class.anon.0** [[_TMP4]], align 8
-// CHECK2-NEXT: [[CALL:%.*]] = call i64 @"_ZZ4mainENK3$_0clEv"(%class.anon.0* nonnull align 8 dereferenceable(40) [[TMP20]]) #[[ATTR4]]
-// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK2-NEXT: [[CALL:%.*]] = call i64 @"_ZZ4mainENK3$_0clEv"(%class.anon.0* nonnull align 8 dereferenceable(40) [[TMP20]]) #[[ATTR6]]
+// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK2-NEXT: ret void
// CHECK2: worker.exit:
// CHECK2-NEXT: ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}_main_l43
-// CHECK2-SAME: (i32* nonnull align 4 dereferenceable(4) [[ARGC:%.*]], i32* nonnull align 4 dereferenceable(4) [[B:%.*]], i32* nonnull align 4 dereferenceable(4) [[C:%.*]], i32* [[D:%.*]], i32* nonnull align 4 dereferenceable(4) [[A:%.*]], %class.anon.0* nonnull align 8 dereferenceable(40) [[L:%.*]]) #[[ATTR1]] {
+// CHECK2-SAME: (i32* nonnull align 4 dereferenceable(4) [[ARGC:%.*]], i32* nonnull align 4 dereferenceable(4) [[B:%.*]], i32* nonnull align 4 dereferenceable(4) [[C:%.*]], i32* [[D:%.*]], i32* nonnull align 4 dereferenceable(4) [[A:%.*]], %class.anon.0* nonnull align 8 dereferenceable(40) [[L:%.*]]) #[[ATTR3]] {
// CHECK2-NEXT: entry:
// CHECK2-NEXT: [[ARGC_ADDR:%.*]] = alloca i32*, align 8
// CHECK2-NEXT: [[B_ADDR:%.*]] = alloca i32*, align 8
@@ -963,7 +963,7 @@ int main(int argc, char **argv) {
// 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]], i1 true, i1 false, i1 true)
+// CHECK2-NEXT: [[TMP5:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true)
// 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:
@@ -992,7 +992,7 @@ int main(int argc, char **argv) {
// CHECK2-NEXT: store i8* [[TMP22]], i8** [[TMP21]], align 8
// CHECK2-NEXT: [[TMP23:%.*]] = bitcast [6 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**
// CHECK2-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB2]], i32 [[TMP6]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*, i32*, i32*, i32*, i32*, i32*, %class.anon.0*)* @__omp_outlined__1 to i8*), i8* null, i8** [[TMP23]], i64 6)
-// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 true)
+// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 true)
// CHECK2-NEXT: ret void
// CHECK2: worker.exit:
// CHECK2-NEXT: ret void
@@ -1067,12 +1067,12 @@ int main(int argc, char **argv) {
// CHECK2-NEXT: [[TMP21:%.*]] = getelementptr inbounds [[CLASS_ANON_0]], %class.anon.0* [[TMP14]], i32 0, i32 4
// CHECK2-NEXT: store i32* [[A10]], i32** [[TMP21]], align 8
// CHECK2-NEXT: [[TMP22:%.*]] = load %class.anon.0*, %class.anon.0** [[_TMP4]], align 8
-// CHECK2-NEXT: [[CALL:%.*]] = call i64 @"_ZZ4mainENK3$_0clEv"(%class.anon.0* nonnull align 8 dereferenceable(40) [[TMP22]]) #[[ATTR4]]
+// CHECK2-NEXT: [[CALL:%.*]] = call i64 @"_ZZ4mainENK3$_0clEv"(%class.anon.0* nonnull align 8 dereferenceable(40) [[TMP22]]) #[[ATTR6]]
// CHECK2-NEXT: ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z3fooIZN1S3fooEvEUlvE_EiRKT__l18
-// CHECK2-SAME: (%class.anon* nonnull align 8 dereferenceable(8) [[T:%.*]]) #[[ATTR1]] {
+// CHECK2-SAME: (%class.anon* nonnull align 8 dereferenceable(8) [[T:%.*]]) #[[ATTR3]] {
// CHECK2-NEXT: entry:
// CHECK2-NEXT: [[T_ADDR:%.*]] = alloca %class.anon*, align 8
// CHECK2-NEXT: [[TMP:%.*]] = alloca %class.anon*, align 8
@@ -1080,7 +1080,7 @@ int main(int argc, char **argv) {
// 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]], i1 true, i1 false, i1 true)
+// CHECK2-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true)
// 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:
@@ -1091,7 +1091,7 @@ int main(int argc, char **argv) {
// CHECK2-NEXT: store i8* [[TMP5]], i8** [[TMP4]], align 8
// CHECK2-NEXT: [[TMP6:%.*]] = bitcast [1 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**
// CHECK2-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB2]], i32 [[TMP2]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*, %class.anon*)* @__omp_outlined__2 to i8*), i8* null, i8** [[TMP6]], i64 1)
-// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 true)
+// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 true)
// CHECK2-NEXT: ret void
// CHECK2: worker.exit:
// CHECK2-NEXT: ret void
@@ -1118,7 +1118,7 @@ int main(int argc, char **argv) {
// CHECK2-NEXT: store %class.anon* [[T1]], %class.anon** [[_TMP2]], align 8
// CHECK2-NEXT: [[TMP4:%.*]] = load %class.anon*, %class.anon** [[_TMP2]], align 8
// CHECK2-NEXT: [[TMP5:%.*]] = load %class.anon*, %class.anon** [[_TMP2]], align 8
-// CHECK2-NEXT: [[CALL:%.*]] = call i32 @_ZZN1S3fooEvENKUlvE_clEv(%class.anon* nonnull align 8 dereferenceable(8) [[TMP5]]) #[[ATTR4]]
+// CHECK2-NEXT: [[CALL:%.*]] = call i32 @_ZZN1S3fooEvENKUlvE_clEv(%class.anon* nonnull align 8 dereferenceable(8) [[TMP5]]) #[[ATTR6]]
// CHECK2-NEXT: ret void
//
//
@@ -1154,7 +1154,7 @@ int main(int argc, char **argv) {
// 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]+]], i1 false, i1 true, i1 true)
+// CHECK3-NEXT: [[TMP4:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true)
// 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:
@@ -1185,15 +1185,15 @@ int main(int argc, char **argv) {
// CHECK3-NEXT: [[TMP19:%.*]] = getelementptr inbounds [[CLASS_ANON]], %class.anon* [[TMP12]], i32 0, i32 4
// CHECK3-NEXT: store i32* [[TMP2]], i32** [[TMP19]], align 8
// CHECK3-NEXT: [[TMP20:%.*]] = load %class.anon*, %class.anon** [[_TMP4]], align 8
-// CHECK3-NEXT: [[CALL:%.*]] = call i64 @"_ZZ4mainENK3$_0clEv"(%class.anon* nonnull align 8 dereferenceable(40) [[TMP20]]) #[[ATTR4:[0-9]+]]
-// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK3-NEXT: [[CALL:%.*]] = call i64 @"_ZZ4mainENK3$_0clEv"(%class.anon* nonnull align 8 dereferenceable(40) [[TMP20]]) #[[ATTR6:[0-9]+]]
+// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK3-NEXT: ret void
// CHECK3: worker.exit:
// CHECK3-NEXT: ret void
//
//
// CHECK3-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}_main_l43
-// CHECK3-SAME: (i32* nonnull align 4 dereferenceable(4) [[ARGC:%.*]], i32* nonnull align 4 dereferenceable(4) [[B:%.*]], i32* nonnull align 4 dereferenceable(4) [[C:%.*]], i32* [[D:%.*]], i32* nonnull align 4 dereferenceable(4) [[A:%.*]], %class.anon* nonnull align 8 dereferenceable(40) [[L:%.*]]) #[[ATTR1:[0-9]+]] {
+// CHECK3-SAME: (i32* nonnull align 4 dereferenceable(4) [[ARGC:%.*]], i32* nonnull align 4 dereferenceable(4) [[B:%.*]], i32* nonnull align 4 dereferenceable(4) [[C:%.*]], i32* [[D:%.*]], i32* nonnull align 4 dereferenceable(4) [[A:%.*]], %class.anon* nonnull align 8 dereferenceable(40) [[L:%.*]]) #[[ATTR3:[0-9]+]] {
// CHECK3-NEXT: entry:
// CHECK3-NEXT: [[ARGC_ADDR:%.*]] = alloca i32*, align 8
// CHECK3-NEXT: [[B_ADDR:%.*]] = alloca i32*, align 8
@@ -1219,7 +1219,7 @@ int main(int argc, char **argv) {
// 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]], i1 true, i1 false, i1 true)
+// CHECK3-NEXT: [[TMP5:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true)
// 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:
@@ -1248,7 +1248,7 @@ int main(int argc, char **argv) {
// CHECK3-NEXT: store i8* [[TMP22]], i8** [[TMP21]], align 8
// CHECK3-NEXT: [[TMP23:%.*]] = bitcast [6 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**
// CHECK3-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB2]], i32 [[TMP6]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*, i32*, i32*, i32*, i32*, i32*, %class.anon*)* @__omp_outlined__ to i8*), i8* null, i8** [[TMP23]], i64 6)
-// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 true)
+// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 true)
// CHECK3-NEXT: ret void
// CHECK3: worker.exit:
// CHECK3-NEXT: ret void
@@ -1323,7 +1323,7 @@ int main(int argc, char **argv) {
// CHECK3-NEXT: [[TMP21:%.*]] = getelementptr inbounds [[CLASS_ANON]], %class.anon* [[TMP14]], i32 0, i32 4
// CHECK3-NEXT: store i32* [[A10]], i32** [[TMP21]], align 8
// CHECK3-NEXT: [[TMP22:%.*]] = load %class.anon*, %class.anon** [[_TMP4]], align 8
-// CHECK3-NEXT: [[CALL:%.*]] = call i64 @"_ZZ4mainENK3$_0clEv"(%class.anon* nonnull align 8 dereferenceable(40) [[TMP22]]) #[[ATTR4]]
+// CHECK3-NEXT: [[CALL:%.*]] = call i64 @"_ZZ4mainENK3$_0clEv"(%class.anon* nonnull align 8 dereferenceable(40) [[TMP22]]) #[[ATTR6]]
// CHECK3-NEXT: ret void
//
//
@@ -1340,7 +1340,7 @@ int main(int argc, char **argv) {
// 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]], i1 false, i1 true, i1 true)
+// CHECK3-NEXT: [[TMP2:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
// 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:
@@ -1353,8 +1353,8 @@ int main(int argc, char **argv) {
// CHECK3-NEXT: [[TMP7:%.*]] = getelementptr inbounds [[CLASS_ANON_0]], %class.anon.0* [[TMP6]], i32 0, i32 0
// CHECK3-NEXT: store %struct.S* [[TMP0]], %struct.S** [[TMP7]], align 8
// CHECK3-NEXT: [[TMP8:%.*]] = load %class.anon.0*, %class.anon.0** [[_TMP2]], align 8
-// CHECK3-NEXT: [[CALL:%.*]] = call i32 @_ZZN1S3fooEvENKUlvE_clEv(%class.anon.0* nonnull align 8 dereferenceable(8) [[TMP8]]) #[[ATTR4]]
-// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK3-NEXT: [[CALL:%.*]] = call i32 @_ZZN1S3fooEvENKUlvE_clEv(%class.anon.0* nonnull align 8 dereferenceable(8) [[TMP8]]) #[[ATTR6]]
+// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK3-NEXT: ret void
// CHECK3: worker.exit:
// CHECK3-NEXT: ret void
@@ -1374,7 +1374,7 @@ int main(int argc, char **argv) {
//
//
// CHECK3-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZN1S3fooEv_l29
-// CHECK3-SAME: (%struct.S* [[THIS:%.*]], %class.anon.0* nonnull align 8 dereferenceable(8) [[L:%.*]]) #[[ATTR1]] {
+// CHECK3-SAME: (%struct.S* [[THIS:%.*]], %class.anon.0* nonnull align 8 dereferenceable(8) [[L:%.*]]) #[[ATTR3]] {
// CHECK3-NEXT: entry:
// CHECK3-NEXT: [[THIS_ADDR:%.*]] = alloca %struct.S*, align 8
// CHECK3-NEXT: [[L_ADDR:%.*]] = alloca %class.anon.0*, align 8
@@ -1385,7 +1385,7 @@ int main(int argc, char **argv) {
// 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]], i1 true, i1 false, i1 true)
+// CHECK3-NEXT: [[TMP2:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true)
// 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:
@@ -1399,7 +1399,7 @@ int main(int argc, char **argv) {
// CHECK3-NEXT: store i8* [[TMP8]], i8** [[TMP7]], align 8
// CHECK3-NEXT: [[TMP9:%.*]] = bitcast [2 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**
// CHECK3-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB2]], i32 [[TMP3]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*, %struct.S*, %class.anon.0*)* @__omp_outlined__1 to i8*), i8* null, i8** [[TMP9]], i64 2)
-// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 true)
+// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 true)
// CHECK3-NEXT: ret void
// CHECK3: worker.exit:
// CHECK3-NEXT: ret void
@@ -1431,12 +1431,12 @@ int main(int argc, char **argv) {
// CHECK3-NEXT: [[TMP6:%.*]] = getelementptr inbounds [[CLASS_ANON_0]], %class.anon.0* [[TMP5]], i32 0, i32 0
// CHECK3-NEXT: store %struct.S* [[TMP0]], %struct.S** [[TMP6]], align 8
// CHECK3-NEXT: [[TMP7:%.*]] = load %class.anon.0*, %class.anon.0** [[_TMP2]], align 8
-// CHECK3-NEXT: [[CALL:%.*]] = call i32 @_ZZN1S3fooEvENKUlvE_clEv(%class.anon.0* nonnull align 8 dereferenceable(8) [[TMP7]]) #[[ATTR4]]
+// CHECK3-NEXT: [[CALL:%.*]] = call i32 @_ZZN1S3fooEvENKUlvE_clEv(%class.anon.0* nonnull align 8 dereferenceable(8) [[TMP7]]) #[[ATTR6]]
// CHECK3-NEXT: ret void
//
//
// CHECK3-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z3fooIZN1S3fooEvEUlvE_EiRKT__l18
-// CHECK3-SAME: (%class.anon.0* nonnull align 8 dereferenceable(8) [[T:%.*]]) #[[ATTR1]] {
+// CHECK3-SAME: (%class.anon.0* nonnull align 8 dereferenceable(8) [[T:%.*]]) #[[ATTR3]] {
// CHECK3-NEXT: entry:
// CHECK3-NEXT: [[T_ADDR:%.*]] = alloca %class.anon.0*, align 8
// CHECK3-NEXT: [[TMP:%.*]] = alloca %class.anon.0*, align 8
@@ -1444,7 +1444,7 @@ int main(int argc, char **argv) {
// 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]], i1 true, i1 false, i1 true)
+// CHECK3-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true)
// 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:
@@ -1455,7 +1455,7 @@ int main(int argc, char **argv) {
// CHECK3-NEXT: store i8* [[TMP5]], i8** [[TMP4]], align 8
// CHECK3-NEXT: [[TMP6:%.*]] = bitcast [1 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**
// CHECK3-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB2]], i32 [[TMP2]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*, %class.anon.0*)* @__omp_outlined__2 to i8*), i8* null, i8** [[TMP6]], i64 1)
-// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 true)
+// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 true)
// CHECK3-NEXT: ret void
// CHECK3: worker.exit:
// CHECK3-NEXT: ret void
@@ -1482,7 +1482,7 @@ int main(int argc, char **argv) {
// CHECK3-NEXT: store %class.anon.0* [[T1]], %class.anon.0** [[_TMP2]], align 8
// CHECK3-NEXT: [[TMP4:%.*]] = load %class.anon.0*, %class.anon.0** [[_TMP2]], align 8
// CHECK3-NEXT: [[TMP5:%.*]] = load %class.anon.0*, %class.anon.0** [[_TMP2]], align 8
-// CHECK3-NEXT: [[CALL:%.*]] = call i32 @_ZZN1S3fooEvENKUlvE_clEv(%class.anon.0* nonnull align 8 dereferenceable(8) [[TMP5]]) #[[ATTR4]]
+// CHECK3-NEXT: [[CALL:%.*]] = call i32 @_ZZN1S3fooEvENKUlvE_clEv(%class.anon.0* nonnull align 8 dereferenceable(8) [[TMP5]]) #[[ATTR6]]
// CHECK3-NEXT: ret void
//
//
@@ -1518,7 +1518,7 @@ int main(int argc, char **argv) {
// 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]+]], i1 false, i1 true, i1 true)
+// CHECK4-NEXT: [[TMP4:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true)
// 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:
@@ -1549,15 +1549,15 @@ int main(int argc, char **argv) {
// CHECK4-NEXT: [[TMP19:%.*]] = getelementptr inbounds [[CLASS_ANON]], %class.anon* [[TMP12]], i32 0, i32 4
// CHECK4-NEXT: store i32* [[TMP2]], i32** [[TMP19]], align 8
// CHECK4-NEXT: [[TMP20:%.*]] = load %class.anon*, %class.anon** [[_TMP4]], align 8
-// CHECK4-NEXT: [[CALL:%.*]] = call i64 @"_ZZ4mainENK3$_0clEv"(%class.anon* nonnull align 8 dereferenceable(40) [[TMP20]]) #[[ATTR4:[0-9]+]]
-// CHECK4-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK4-NEXT: [[CALL:%.*]] = call i64 @"_ZZ4mainENK3$_0clEv"(%class.anon* nonnull align 8 dereferenceable(40) [[TMP20]]) #[[ATTR6:[0-9]+]]
+// CHECK4-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK4-NEXT: ret void
// CHECK4: worker.exit:
// CHECK4-NEXT: ret void
//
//
// CHECK4-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}_main_l43
-// CHECK4-SAME: (i32* nonnull align 4 dereferenceable(4) [[ARGC:%.*]], i32* nonnull align 4 dereferenceable(4) [[B:%.*]], i32* nonnull align 4 dereferenceable(4) [[C:%.*]], i32* [[D:%.*]], i32* nonnull align 4 dereferenceable(4) [[A:%.*]], %class.anon* nonnull align 8 dereferenceable(40) [[L:%.*]]) #[[ATTR1:[0-9]+]] {
+// CHECK4-SAME: (i32* nonnull align 4 dereferenceable(4) [[ARGC:%.*]], i32* nonnull align 4 dereferenceable(4) [[B:%.*]], i32* nonnull align 4 dereferenceable(4) [[C:%.*]], i32* [[D:%.*]], i32* nonnull align 4 dereferenceable(4) [[A:%.*]], %class.anon* nonnull align 8 dereferenceable(40) [[L:%.*]]) #[[ATTR3:[0-9]+]] {
// CHECK4-NEXT: entry:
// CHECK4-NEXT: [[ARGC_ADDR:%.*]] = alloca i32*, align 8
// CHECK4-NEXT: [[B_ADDR:%.*]] = alloca i32*, align 8
@@ -1583,7 +1583,7 @@ int main(int argc, char **argv) {
// 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]], i1 true, i1 false, i1 true)
+// CHECK4-NEXT: [[TMP5:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true)
// 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:
@@ -1612,7 +1612,7 @@ int main(int argc, char **argv) {
// CHECK4-NEXT: store i8* [[TMP22]], i8** [[TMP21]], align 8
// CHECK4-NEXT: [[TMP23:%.*]] = bitcast [6 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**
// CHECK4-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB2]], i32 [[TMP6]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*, i32*, i32*, i32*, i32*, i32*, %class.anon*)* @__omp_outlined__ to i8*), i8* null, i8** [[TMP23]], i64 6)
-// CHECK4-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 true)
+// CHECK4-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 true)
// CHECK4-NEXT: ret void
// CHECK4: worker.exit:
// CHECK4-NEXT: ret void
@@ -1687,7 +1687,7 @@ int main(int argc, char **argv) {
// CHECK4-NEXT: [[TMP21:%.*]] = getelementptr inbounds [[CLASS_ANON]], %class.anon* [[TMP14]], i32 0, i32 4
// CHECK4-NEXT: store i32* [[A10]], i32** [[TMP21]], align 8
// CHECK4-NEXT: [[TMP22:%.*]] = load %class.anon*, %class.anon** [[_TMP4]], align 8
-// CHECK4-NEXT: [[CALL:%.*]] = call i64 @"_ZZ4mainENK3$_0clEv"(%class.anon* nonnull align 8 dereferenceable(40) [[TMP22]]) #[[ATTR4]]
+// CHECK4-NEXT: [[CALL:%.*]] = call i64 @"_ZZ4mainENK3$_0clEv"(%class.anon* nonnull align 8 dereferenceable(40) [[TMP22]]) #[[ATTR6]]
// CHECK4-NEXT: ret void
//
//
@@ -1704,7 +1704,7 @@ int main(int argc, char **argv) {
// 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]], i1 false, i1 true, i1 true)
+// CHECK4-NEXT: [[TMP2:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
// 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:
@@ -1717,8 +1717,8 @@ int main(int argc, char **argv) {
// CHECK4-NEXT: [[TMP7:%.*]] = getelementptr inbounds [[CLASS_ANON_0]], %class.anon.0* [[TMP6]], i32 0, i32 0
// CHECK4-NEXT: store %struct.S* [[TMP0]], %struct.S** [[TMP7]], align 8
// CHECK4-NEXT: [[TMP8:%.*]] = load %class.anon.0*, %class.anon.0** [[_TMP2]], align 8
-// CHECK4-NEXT: [[CALL:%.*]] = call i32 @_ZZN1S3fooEvENKUlvE_clEv(%class.anon.0* nonnull align 8 dereferenceable(8) [[TMP8]]) #[[ATTR4]]
-// CHECK4-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK4-NEXT: [[CALL:%.*]] = call i32 @_ZZN1S3fooEvENKUlvE_clEv(%class.anon.0* nonnull align 8 dereferenceable(8) [[TMP8]]) #[[ATTR6]]
+// CHECK4-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK4-NEXT: ret void
// CHECK4: worker.exit:
// CHECK4-NEXT: ret void
@@ -1738,7 +1738,7 @@ int main(int argc, char **argv) {
//
//
// CHECK4-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__ZN1S3fooEv_l29
-// CHECK4-SAME: (%struct.S* [[THIS:%.*]], %class.anon.0* nonnull align 8 dereferenceable(8) [[L:%.*]]) #[[ATTR1]] {
+// CHECK4-SAME: (%struct.S* [[THIS:%.*]], %class.anon.0* nonnull align 8 dereferenceable(8) [[L:%.*]]) #[[ATTR3]] {
// CHECK4-NEXT: entry:
// CHECK4-NEXT: [[THIS_ADDR:%.*]] = alloca %struct.S*, align 8
// CHECK4-NEXT: [[L_ADDR:%.*]] = alloca %class.anon.0*, align 8
@@ -1749,7 +1749,7 @@ int main(int argc, char **argv) {
// 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]], i1 true, i1 false, i1 true)
+// CHECK4-NEXT: [[TMP2:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true)
// 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:
@@ -1763,7 +1763,7 @@ int main(int argc, char **argv) {
// CHECK4-NEXT: store i8* [[TMP8]], i8** [[TMP7]], align 8
// CHECK4-NEXT: [[TMP9:%.*]] = bitcast [2 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**
// CHECK4-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB2]], i32 [[TMP3]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*, %struct.S*, %class.anon.0*)* @__omp_outlined__1 to i8*), i8* null, i8** [[TMP9]], i64 2)
-// CHECK4-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 true)
+// CHECK4-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 true)
// CHECK4-NEXT: ret void
// CHECK4: worker.exit:
// CHECK4-NEXT: ret void
@@ -1795,12 +1795,12 @@ int main(int argc, char **argv) {
// CHECK4-NEXT: [[TMP6:%.*]] = getelementptr inbounds [[CLASS_ANON_0]], %class.anon.0* [[TMP5]], i32 0, i32 0
// CHECK4-NEXT: store %struct.S* [[TMP0]], %struct.S** [[TMP6]], align 8
// CHECK4-NEXT: [[TMP7:%.*]] = load %class.anon.0*, %class.anon.0** [[_TMP2]], align 8
-// CHECK4-NEXT: [[CALL:%.*]] = call i32 @_ZZN1S3fooEvENKUlvE_clEv(%class.anon.0* nonnull align 8 dereferenceable(8) [[TMP7]]) #[[ATTR4]]
+// CHECK4-NEXT: [[CALL:%.*]] = call i32 @_ZZN1S3fooEvENKUlvE_clEv(%class.anon.0* nonnull align 8 dereferenceable(8) [[TMP7]]) #[[ATTR6]]
// CHECK4-NEXT: ret void
//
//
// CHECK4-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z3fooIZN1S3fooEvEUlvE_EiRKT__l18
-// CHECK4-SAME: (%class.anon.0* nonnull align 8 dereferenceable(8) [[T:%.*]]) #[[ATTR1]] {
+// CHECK4-SAME: (%class.anon.0* nonnull align 8 dereferenceable(8) [[T:%.*]]) #[[ATTR3]] {
// CHECK4-NEXT: entry:
// CHECK4-NEXT: [[T_ADDR:%.*]] = alloca %class.anon.0*, align 8
// CHECK4-NEXT: [[TMP:%.*]] = alloca %class.anon.0*, align 8
@@ -1808,7 +1808,7 @@ int main(int argc, char **argv) {
// 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]], i1 true, i1 false, i1 true)
+// CHECK4-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true)
// 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:
@@ -1819,7 +1819,7 @@ int main(int argc, char **argv) {
// CHECK4-NEXT: store i8* [[TMP5]], i8** [[TMP4]], align 8
// CHECK4-NEXT: [[TMP6:%.*]] = bitcast [1 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**
// CHECK4-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB2]], i32 [[TMP2]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*, %class.anon.0*)* @__omp_outlined__2 to i8*), i8* null, i8** [[TMP6]], i64 1)
-// CHECK4-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 true)
+// CHECK4-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 true)
// CHECK4-NEXT: ret void
// CHECK4: worker.exit:
// CHECK4-NEXT: ret void
@@ -1846,6 +1846,6 @@ int main(int argc, char **argv) {
// CHECK4-NEXT: store %class.anon.0* [[T1]], %class.anon.0** [[_TMP2]], align 8
// CHECK4-NEXT: [[TMP4:%.*]] = load %class.anon.0*, %class.anon.0** [[_TMP2]], align 8
// CHECK4-NEXT: [[TMP5:%.*]] = load %class.anon.0*, %class.anon.0** [[_TMP2]], align 8
-// CHECK4-NEXT: [[CALL:%.*]] = call i32 @_ZZN1S3fooEvENKUlvE_clEv(%class.anon.0* nonnull align 8 dereferenceable(8) [[TMP5]]) #[[ATTR4]]
+// CHECK4-NEXT: [[CALL:%.*]] = call i32 @_ZZN1S3fooEvENKUlvE_clEv(%class.anon.0* nonnull align 8 dereferenceable(8) [[TMP5]]) #[[ATTR6]]
// CHECK4-NEXT: ret void
//
diff --git a/clang/test/OpenMP/nvptx_multi_target_parallel_codegen.cpp b/clang/test/OpenMP/nvptx_multi_target_parallel_codegen.cpp
index 232a2a33e9035..0c9c26ebea562 100644
--- a/clang/test/OpenMP/nvptx_multi_target_parallel_codegen.cpp
+++ b/clang/test/OpenMP/nvptx_multi_target_parallel_codegen.cpp
@@ -29,14 +29,14 @@ int main() {
// 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]+]], i1 true, i1 false, i1 true)
+// CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 true)
// 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: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB2:[0-9]+]])
// CHECK1-NEXT: [[TMP2:%.*]] = bitcast [0 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**
// CHECK1-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB2]], i32 [[TMP1]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*)* @__omp_outlined__ to i8*), i8* null, i8** [[TMP2]], i64 0)
-// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 true)
+// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 true)
// CHECK1-NEXT: ret void
// CHECK1: worker.exit:
// CHECK1-NEXT: ret void
@@ -66,12 +66,12 @@ int main() {
// 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]], i1 false, i1 true, i1 true)
+// CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
// 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 @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK1-NEXT: ret void
// CHECK1: worker.exit:
// CHECK1-NEXT: ret void
@@ -107,14 +107,14 @@ int main() {
// 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]+]], i1 true, i1 false, i1 true)
+// CHECK2-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 true)
// 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: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB2:[0-9]+]])
// CHECK2-NEXT: [[TMP2:%.*]] = bitcast [0 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**
// CHECK2-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB2]], i32 [[TMP1]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*)* @__omp_outlined__ to i8*), i8* null, i8** [[TMP2]], i32 0)
-// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 true)
+// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 true)
// CHECK2-NEXT: ret void
// CHECK2: worker.exit:
// CHECK2-NEXT: ret void
@@ -144,12 +144,12 @@ int main() {
// 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]], i1 false, i1 true, i1 true)
+// CHECK2-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
// 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 @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK2-NEXT: ret void
// CHECK2: worker.exit:
// CHECK2-NEXT: ret void
@@ -185,14 +185,14 @@ int main() {
// 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]+]], i1 true, i1 false, i1 true)
+// CHECK3-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 true)
// 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: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB2:[0-9]+]])
// CHECK3-NEXT: [[TMP2:%.*]] = bitcast [0 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**
// CHECK3-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB2]], i32 [[TMP1]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*)* @__omp_outlined__ to i8*), i8* null, i8** [[TMP2]], i32 0)
-// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 true)
+// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 true)
// CHECK3-NEXT: ret void
// CHECK3: worker.exit:
// CHECK3-NEXT: ret void
@@ -222,12 +222,12 @@ int main() {
// 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]], i1 false, i1 true, i1 true)
+// CHECK3-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
// 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 @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK3-NEXT: ret void
// CHECK3: worker.exit:
// CHECK3-NEXT: ret void
diff --git a/clang/test/OpenMP/nvptx_nested_parallel_codegen.cpp b/clang/test/OpenMP/nvptx_nested_parallel_codegen.cpp
index ff0bb4c518da9..8943d30c76563 100644
--- a/clang/test/OpenMP/nvptx_nested_parallel_codegen.cpp
+++ b/clang/test/OpenMP/nvptx_nested_parallel_codegen.cpp
@@ -40,7 +40,7 @@ int main() {
// 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]+]], i1 false, i1 true, i1 true)
+// CHECK1-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true)
// 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:
@@ -52,7 +52,7 @@ int main() {
// CHECK1-NEXT: store i8* [[TMP4]], i8** [[TMP3]], align 8
// CHECK1-NEXT: [[TMP5:%.*]] = bitcast [1 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**
// CHECK1-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB1]], i32 [[TMP2]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*, i32*)* @__omp_outlined__ to i8*), i8* bitcast (void (i16, i32)* @__omp_outlined___wrapper to i8*), i8** [[TMP5]], i64 1)
-// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK1-NEXT: ret void
// CHECK1: worker.exit:
// CHECK1-NEXT: ret void
@@ -173,7 +173,7 @@ int main() {
// 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]+]], i1 false, i1 true, i1 true)
+// CHECK2-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true)
// 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:
@@ -185,7 +185,7 @@ int main() {
// CHECK2-NEXT: store i8* [[TMP4]], i8** [[TMP3]], align 4
// CHECK2-NEXT: [[TMP5:%.*]] = bitcast [1 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**
// CHECK2-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB1]], i32 [[TMP2]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*, i32*)* @__omp_outlined__ to i8*), i8* bitcast (void (i16, i32)* @__omp_outlined___wrapper to i8*), i8** [[TMP5]], i32 1)
-// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK2-NEXT: ret void
// CHECK2: worker.exit:
// CHECK2-NEXT: ret void
@@ -306,7 +306,7 @@ int main() {
// 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]+]], i1 false, i1 true, i1 true)
+// CHECK3-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true)
// 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:
@@ -318,7 +318,7 @@ int main() {
// CHECK3-NEXT: store i8* [[TMP4]], i8** [[TMP3]], align 4
// CHECK3-NEXT: [[TMP5:%.*]] = bitcast [1 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**
// CHECK3-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB1]], i32 [[TMP2]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*, i32*)* @__omp_outlined__ to i8*), i8* bitcast (void (i16, i32)* @__omp_outlined___wrapper to i8*), i8** [[TMP5]], i32 1)
-// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK3-NEXT: ret void
// CHECK3: worker.exit:
// CHECK3-NEXT: ret void
diff --git a/clang/test/OpenMP/nvptx_parallel_codegen.cpp b/clang/test/OpenMP/nvptx_parallel_codegen.cpp
index 712c5a41c573d..35d6813d25267 100644
--- a/clang/test/OpenMP/nvptx_parallel_codegen.cpp
+++ b/clang/test/OpenMP/nvptx_parallel_codegen.cpp
@@ -1446,7 +1446,7 @@ int bar(int n){
// 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]+]], i1 false, i1 true, i1 true)
+// CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true)
// 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:
@@ -1460,7 +1460,7 @@ int bar(int n){
// CHECK1-NEXT: [[TMP5:%.*]] = load i32, i32* [[CONV]], align 8
// CHECK1-NEXT: [[ADD:%.*]] = add nsw i32 [[TMP5]], 1
// CHECK1-NEXT: store i32 [[ADD]], i32* [[CONV]], align 8
-// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK1-NEXT: ret void
// CHECK1: worker.exit:
// CHECK1-NEXT: ret void
@@ -1563,7 +1563,7 @@ int bar(int n){
// 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]], i1 false, i1 true, i1 true)
+// CHECK1-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
// 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:
@@ -1585,7 +1585,7 @@ int bar(int n){
// CHECK1-NEXT: [[TMP8:%.*]] = load i32, i32* [[ARRAYIDX]], align 4
// CHECK1-NEXT: [[ADD6:%.*]] = add nsw i32 [[TMP8]], 1
// CHECK1-NEXT: store i32 [[ADD6]], i32* [[ARRAYIDX]], align 4
-// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK1-NEXT: ret void
// CHECK1: worker.exit:
// CHECK1-NEXT: ret void
@@ -1628,7 +1628,7 @@ int bar(int n){
// 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]], i1 false, i1 true, i1 true)
+// CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
// 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:
@@ -1646,7 +1646,7 @@ int bar(int n){
// CHECK1-NEXT: [[INC:%.*]] = add nsw i32 [[TMP6]], 1
// CHECK1-NEXT: store i32 [[INC]], i32* [[A_ON_STACK]], align 4
// CHECK1-NEXT: call void @__kmpc_free_shared(i8* [[A1]], i64 4)
-// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK1-NEXT: ret void
// CHECK1: worker.exit:
// CHECK1-NEXT: ret void
@@ -1721,7 +1721,7 @@ int bar(int n){
// 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]+]], i1 false, i1 true, i1 true)
+// CHECK2-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true)
// 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:
@@ -1735,7 +1735,7 @@ int bar(int n){
// CHECK2-NEXT: [[TMP5:%.*]] = load i32, i32* [[A_ADDR]], align 4
// CHECK2-NEXT: [[ADD:%.*]] = add nsw i32 [[TMP5]], 1
// CHECK2-NEXT: store i32 [[ADD]], i32* [[A_ADDR]], align 4
-// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK2-NEXT: ret void
// CHECK2: worker.exit:
// CHECK2-NEXT: ret void
@@ -1836,7 +1836,7 @@ int bar(int n){
// 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]], i1 false, i1 true, i1 true)
+// CHECK2-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
// 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:
@@ -1858,7 +1858,7 @@ int bar(int n){
// CHECK2-NEXT: [[TMP8:%.*]] = load i32, i32* [[ARRAYIDX]], align 4
// CHECK2-NEXT: [[ADD4:%.*]] = add nsw i32 [[TMP8]], 1
// CHECK2-NEXT: store i32 [[ADD4]], i32* [[ARRAYIDX]], align 4
-// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK2-NEXT: ret void
// CHECK2: worker.exit:
// CHECK2-NEXT: ret void
@@ -1900,7 +1900,7 @@ int bar(int n){
// 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]], i1 false, i1 true, i1 true)
+// CHECK2-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
// 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:
@@ -1918,7 +1918,7 @@ int bar(int n){
// CHECK2-NEXT: [[INC:%.*]] = add nsw i32 [[TMP6]], 1
// CHECK2-NEXT: store i32 [[INC]], i32* [[A_ON_STACK]], align 4
// CHECK2-NEXT: call void @__kmpc_free_shared(i8* [[A1]], i32 4)
-// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK2-NEXT: ret void
// CHECK2: worker.exit:
// CHECK2-NEXT: ret void
diff --git a/clang/test/OpenMP/nvptx_parallel_for_codegen.cpp b/clang/test/OpenMP/nvptx_parallel_for_codegen.cpp
index 5f9da8346cc6f..1996299071957 100644
--- a/clang/test/OpenMP/nvptx_parallel_for_codegen.cpp
+++ b/clang/test/OpenMP/nvptx_parallel_for_codegen.cpp
@@ -465,7 +465,7 @@ int bar(int n){
// 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]+]], i1 false, i1 true, i1 true)
+// CHECK-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true)
// 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:
@@ -487,7 +487,7 @@ int bar(int n){
// CHECK-NEXT: [[ADD:%.*]] = add nsw i32 [[TMP9]], 1
// CHECK-NEXT: store i32 [[ADD]], i32* [[ARRAYIDX]], align 4
// CHECK-NEXT: call void @__kmpc_free_shared(i8* [[D]], i64 4)
-// CHECK-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK-NEXT: ret void
// CHECK: worker.exit:
// CHECK-NEXT: ret void
@@ -584,7 +584,7 @@ int bar(int n){
//
//
// CHECK-LABEL: define {{[^@]+}}@__omp_outlined___wrapper
-// CHECK-SAME: (i16 zeroext [[TMP0:%.*]], i32 [[TMP1:%.*]]) #[[ATTR2:[0-9]+]] {
+// CHECK-SAME: (i16 zeroext [[TMP0:%.*]], i32 [[TMP1:%.*]]) #[[ATTR3:[0-9]+]] {
// CHECK-NEXT: entry:
// CHECK-NEXT: [[DOTADDR:%.*]] = alloca i16, align 2
// CHECK-NEXT: [[DOTADDR1:%.*]] = alloca i32, align 4
@@ -601,6 +601,6 @@ int bar(int n){
// 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__(i32* [[DOTADDR1]], i32* [[DOTZERO_ADDR]], [10 x i32]* [[TMP5]], i32* [[TMP8]]) #[[ATTR1:[0-9]+]]
+// CHECK-NEXT: call void @__omp_outlined__(i32* [[DOTADDR1]], i32* [[DOTZERO_ADDR]], [10 x i32]* [[TMP5]], i32* [[TMP8]]) #[[ATTR2:[0-9]+]]
// CHECK-NEXT: ret void
//
diff --git a/clang/test/OpenMP/nvptx_target_codegen.cpp b/clang/test/OpenMP/nvptx_target_codegen.cpp
index 5bf54060b85a7..63560b26aa644 100644
--- a/clang/test/OpenMP/nvptx_target_codegen.cpp
+++ b/clang/test/OpenMP/nvptx_target_codegen.cpp
@@ -153,7 +153,7 @@ void unreachable_call() {
// 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]+]], i1 true, i1 false, i1 true)
+// CHECK1-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 true)
// 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:
@@ -166,7 +166,7 @@ void unreachable_call() {
// CHECK1-NEXT: store i8* [[TMP6]], i8** [[TMP5]], align 8
// CHECK1-NEXT: [[TMP7:%.*]] = bitcast [2 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**
// CHECK1-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB2]], i32 [[TMP2]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*, i32**, i32**)* @__omp_outlined__ to i8*), i8* null, i8** [[TMP7]], i64 2)
-// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 true)
+// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 true)
// CHECK1-NEXT: ret void
// CHECK1: worker.exit:
// CHECK1-NEXT: ret void
@@ -195,11 +195,11 @@ void unreachable_call() {
// 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]], i1 false, i1 true, i1 true)
+// CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
// 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 @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK1-NEXT: ret void
// CHECK1: worker.exit:
// CHECK1-NEXT: ret void
@@ -211,7 +211,7 @@ void unreachable_call() {
// 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]], i1 false, i1 true, i1 true)
+// CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
// 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:
@@ -225,7 +225,7 @@ void unreachable_call() {
// CHECK1-NEXT: [[ADD4:%.*]] = add nsw i32 [[CONV3]], 2
// CHECK1-NEXT: [[CONV5:%.*]] = trunc i32 [[ADD4]] to i16
// CHECK1-NEXT: store i16 [[CONV5]], i16* [[CONV]], align 8
-// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK1-NEXT: ret void
// CHECK1: worker.exit:
// CHECK1-NEXT: ret void
@@ -261,7 +261,7 @@ void unreachable_call() {
// 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]], i1 false, i1 true, i1 true)
+// CHECK1-NEXT: [[TMP8:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
// 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:
@@ -305,7 +305,7 @@ void unreachable_call() {
// CHECK1-NEXT: [[TMP17:%.*]] = load i64, i64* [[CALL]], align 8
// CHECK1-NEXT: [[ADD22:%.*]] = add nsw i64 [[TMP17]], 1
// CHECK1-NEXT: store i64 [[ADD22]], i64* [[CALL]], align 8
-// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK1-NEXT: ret void
// CHECK1: worker.exit:
// CHECK1-NEXT: ret void
@@ -338,7 +338,7 @@ void unreachable_call() {
// 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]], i1 false, i1 true, i1 true)
+// CHECK1-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
// 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:
@@ -359,7 +359,7 @@ void unreachable_call() {
// CHECK1-NEXT: [[TMP5:%.*]] = load i32, i32* [[ARRAYIDX]], align 4
// CHECK1-NEXT: [[ADD9:%.*]] = add nsw i32 [[TMP5]], 1
// CHECK1-NEXT: store i32 [[ADD9]], i32* [[ARRAYIDX]], align 4
-// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK1-NEXT: ret void
// CHECK1: worker.exit:
// CHECK1-NEXT: ret void
@@ -383,7 +383,7 @@ void unreachable_call() {
// 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]], i1 false, i1 true, i1 true)
+// CHECK1-NEXT: [[TMP4:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
// 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:
@@ -406,7 +406,7 @@ void unreachable_call() {
// CHECK1-NEXT: [[CONV8:%.*]] = fptosi double [[TMP8]] to i32
// CHECK1-NEXT: [[A9:%.*]] = getelementptr inbounds [[STRUCT_S1]], %struct.S1* [[TMP0]], i32 0, i32 0
// CHECK1-NEXT: [[CALL:%.*]] = call i32 @_Z3baziRd(i32 [[CONV8]], double* nonnull align 8 dereferenceable(8) [[A9]]) #[[ATTR6]]
-// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK1-NEXT: ret void
// CHECK1: worker.exit:
// CHECK1-NEXT: ret void
@@ -439,7 +439,7 @@ void unreachable_call() {
// 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]], i1 false, i1 true, i1 true)
+// CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
// 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:
@@ -448,7 +448,7 @@ void unreachable_call() {
// CHECK1: worker.exit:
// CHECK1-NEXT: ret void
// CHECK1: 1:
-// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK1-NEXT: ret void
//
//
@@ -464,7 +464,7 @@ void unreachable_call() {
// 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]], i1 false, i1 true, i1 true)
+// CHECK1-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
// 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:
@@ -480,7 +480,7 @@ void unreachable_call() {
// CHECK1-NEXT: [[TMP4:%.*]] = load i32, i32* [[ARRAYIDX]], align 4
// CHECK1-NEXT: [[ADD5:%.*]] = add nsw i32 [[TMP4]], 1
// CHECK1-NEXT: store i32 [[ADD5]], i32* [[ARRAYIDX]], align 4
-// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK1-NEXT: ret void
// CHECK1: worker.exit:
// CHECK1-NEXT: ret void
@@ -540,7 +540,7 @@ void unreachable_call() {
// 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]+]], i1 true, i1 false, i1 true)
+// CHECK2-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 true)
// 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:
@@ -553,7 +553,7 @@ void unreachable_call() {
// CHECK2-NEXT: store i8* [[TMP6]], i8** [[TMP5]], align 4
// CHECK2-NEXT: [[TMP7:%.*]] = bitcast [2 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**
// CHECK2-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB2]], i32 [[TMP2]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*, i32**, i32**)* @__omp_outlined__ to i8*), i8* null, i8** [[TMP7]], i32 2)
-// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 true)
+// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 true)
// CHECK2-NEXT: ret void
// CHECK2: worker.exit:
// CHECK2-NEXT: ret void
@@ -582,11 +582,11 @@ void unreachable_call() {
// 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]], i1 false, i1 true, i1 true)
+// CHECK2-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
// 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 @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK2-NEXT: ret void
// CHECK2: worker.exit:
// CHECK2-NEXT: ret void
@@ -598,7 +598,7 @@ void unreachable_call() {
// 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]], i1 false, i1 true, i1 true)
+// CHECK2-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
// 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:
@@ -612,7 +612,7 @@ void unreachable_call() {
// CHECK2-NEXT: [[ADD4:%.*]] = add nsw i32 [[CONV3]], 2
// CHECK2-NEXT: [[CONV5:%.*]] = trunc i32 [[ADD4]] to i16
// CHECK2-NEXT: store i16 [[CONV5]], i16* [[CONV]], align 4
-// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK2-NEXT: ret void
// CHECK2: worker.exit:
// CHECK2-NEXT: ret void
@@ -647,7 +647,7 @@ void unreachable_call() {
// 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]], i1 false, i1 true, i1 true)
+// CHECK2-NEXT: [[TMP8:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
// 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:
@@ -691,7 +691,7 @@ void unreachable_call() {
// CHECK2-NEXT: [[TMP17:%.*]] = load i64, i64* [[CALL]], align 8
// CHECK2-NEXT: [[ADD21:%.*]] = add nsw i64 [[TMP17]], 1
// CHECK2-NEXT: store i64 [[ADD21]], i64* [[CALL]], align 8
-// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK2-NEXT: ret void
// CHECK2: worker.exit:
// CHECK2-NEXT: ret void
@@ -723,7 +723,7 @@ void unreachable_call() {
// 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]], i1 false, i1 true, i1 true)
+// CHECK2-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
// 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:
@@ -744,7 +744,7 @@ void unreachable_call() {
// CHECK2-NEXT: [[TMP5:%.*]] = load i32, i32* [[ARRAYIDX]], align 4
// CHECK2-NEXT: [[ADD8:%.*]] = add nsw i32 [[TMP5]], 1
// CHECK2-NEXT: store i32 [[ADD8]], i32* [[ARRAYIDX]], align 4
-// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK2-NEXT: ret void
// CHECK2: worker.exit:
// CHECK2-NEXT: ret void
@@ -767,7 +767,7 @@ void unreachable_call() {
// 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]], i1 false, i1 true, i1 true)
+// CHECK2-NEXT: [[TMP4:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
// 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:
@@ -790,7 +790,7 @@ void unreachable_call() {
// CHECK2-NEXT: [[CONV7:%.*]] = fptosi double [[TMP8]] to i32
// CHECK2-NEXT: [[A8:%.*]] = getelementptr inbounds [[STRUCT_S1]], %struct.S1* [[TMP0]], i32 0, i32 0
// CHECK2-NEXT: [[CALL:%.*]] = call i32 @_Z3baziRd(i32 [[CONV7]], double* nonnull align 8 dereferenceable(8) [[A8]]) #[[ATTR6]]
-// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK2-NEXT: ret void
// CHECK2: worker.exit:
// CHECK2-NEXT: ret void
@@ -823,7 +823,7 @@ void unreachable_call() {
// 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]], i1 false, i1 true, i1 true)
+// CHECK2-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
// 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:
@@ -832,7 +832,7 @@ void unreachable_call() {
// CHECK2: worker.exit:
// CHECK2-NEXT: ret void
// CHECK2: 1:
-// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK2-NEXT: ret void
//
//
@@ -847,7 +847,7 @@ void unreachable_call() {
// 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]], i1 false, i1 true, i1 true)
+// CHECK2-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
// 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:
@@ -863,7 +863,7 @@ void unreachable_call() {
// CHECK2-NEXT: [[TMP4:%.*]] = load i32, i32* [[ARRAYIDX]], align 4
// CHECK2-NEXT: [[ADD4:%.*]] = add nsw i32 [[TMP4]], 1
// CHECK2-NEXT: store i32 [[ADD4]], i32* [[ARRAYIDX]], align 4
-// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK2-NEXT: ret void
// CHECK2: worker.exit:
// CHECK2-NEXT: ret void
@@ -923,7 +923,7 @@ void unreachable_call() {
// 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]+]], i1 true, i1 false, i1 true)
+// CHECK3-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 true)
// 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:
@@ -936,7 +936,7 @@ void unreachable_call() {
// CHECK3-NEXT: store i8* [[TMP6]], i8** [[TMP5]], align 4
// CHECK3-NEXT: [[TMP7:%.*]] = bitcast [2 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**
// CHECK3-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB2]], i32 [[TMP2]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*, i32**, i32**)* @__omp_outlined__ to i8*), i8* null, i8** [[TMP7]], i32 2)
-// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 true)
+// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 true)
// CHECK3-NEXT: ret void
// CHECK3: worker.exit:
// CHECK3-NEXT: ret void
@@ -965,11 +965,11 @@ void unreachable_call() {
// 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]], i1 false, i1 true, i1 true)
+// CHECK3-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
// 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 @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK3-NEXT: ret void
// CHECK3: worker.exit:
// CHECK3-NEXT: ret void
@@ -981,7 +981,7 @@ void unreachable_call() {
// 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]], i1 false, i1 true, i1 true)
+// CHECK3-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
// 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:
@@ -995,7 +995,7 @@ void unreachable_call() {
// CHECK3-NEXT: [[ADD4:%.*]] = add nsw i32 [[CONV3]], 2
// CHECK3-NEXT: [[CONV5:%.*]] = trunc i32 [[ADD4]] to i16
// CHECK3-NEXT: store i16 [[CONV5]], i16* [[CONV]], align 4
-// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK3-NEXT: ret void
// CHECK3: worker.exit:
// CHECK3-NEXT: ret void
@@ -1030,7 +1030,7 @@ void unreachable_call() {
// 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]], i1 false, i1 true, i1 true)
+// CHECK3-NEXT: [[TMP8:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
// 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:
@@ -1074,7 +1074,7 @@ void unreachable_call() {
// CHECK3-NEXT: [[TMP17:%.*]] = load i64, i64* [[CALL]], align 8
// CHECK3-NEXT: [[ADD21:%.*]] = add nsw i64 [[TMP17]], 1
// CHECK3-NEXT: store i64 [[ADD21]], i64* [[CALL]], align 8
-// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK3-NEXT: ret void
// CHECK3: worker.exit:
// CHECK3-NEXT: ret void
@@ -1106,7 +1106,7 @@ void unreachable_call() {
// 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]], i1 false, i1 true, i1 true)
+// CHECK3-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
// 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:
@@ -1127,7 +1127,7 @@ void unreachable_call() {
// CHECK3-NEXT: [[TMP5:%.*]] = load i32, i32* [[ARRAYIDX]], align 4
// CHECK3-NEXT: [[ADD8:%.*]] = add nsw i32 [[TMP5]], 1
// CHECK3-NEXT: store i32 [[ADD8]], i32* [[ARRAYIDX]], align 4
-// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK3-NEXT: ret void
// CHECK3: worker.exit:
// CHECK3-NEXT: ret void
@@ -1150,7 +1150,7 @@ void unreachable_call() {
// 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]], i1 false, i1 true, i1 true)
+// CHECK3-NEXT: [[TMP4:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
// 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:
@@ -1173,7 +1173,7 @@ void unreachable_call() {
// CHECK3-NEXT: [[CONV7:%.*]] = fptosi double [[TMP8]] to i32
// CHECK3-NEXT: [[A8:%.*]] = getelementptr inbounds [[STRUCT_S1]], %struct.S1* [[TMP0]], i32 0, i32 0
// CHECK3-NEXT: [[CALL:%.*]] = call i32 @_Z3baziRd(i32 [[CONV7]], double* nonnull align 8 dereferenceable(8) [[A8]]) #[[ATTR6]]
-// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK3-NEXT: ret void
// CHECK3: worker.exit:
// CHECK3-NEXT: ret void
@@ -1206,7 +1206,7 @@ void unreachable_call() {
// 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]], i1 false, i1 true, i1 true)
+// CHECK3-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
// 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:
@@ -1215,7 +1215,7 @@ void unreachable_call() {
// CHECK3: worker.exit:
// CHECK3-NEXT: ret void
// CHECK3: 1:
-// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK3-NEXT: ret void
//
//
@@ -1230,7 +1230,7 @@ void unreachable_call() {
// 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]], i1 false, i1 true, i1 true)
+// CHECK3-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
// 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:
@@ -1246,7 +1246,7 @@ void unreachable_call() {
// CHECK3-NEXT: [[TMP4:%.*]] = load i32, i32* [[ARRAYIDX]], align 4
// CHECK3-NEXT: [[ADD4:%.*]] = add nsw i32 [[TMP4]], 1
// CHECK3-NEXT: store i32 [[ADD4]], i32* [[ARRAYIDX]], align 4
-// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK3-NEXT: ret void
// CHECK3: worker.exit:
// CHECK3-NEXT: ret void
diff --git a/clang/test/OpenMP/nvptx_target_parallel_codegen.cpp b/clang/test/OpenMP/nvptx_target_parallel_codegen.cpp
index 11f18c07f38a3..cbf44c5060336 100644
--- a/clang/test/OpenMP/nvptx_target_parallel_codegen.cpp
+++ b/clang/test/OpenMP/nvptx_target_parallel_codegen.cpp
@@ -58,7 +58,7 @@ int bar(int n){
// 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]+]], i1 true, i1 false, i1 true)
+// CHECK1-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 true)
// 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:
@@ -68,7 +68,7 @@ int bar(int n){
// CHECK1-NEXT: store i8* [[TMP4]], i8** [[TMP3]], align 8
// CHECK1-NEXT: [[TMP5:%.*]] = bitcast [1 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**
// CHECK1-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB2]], i32 [[TMP2]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*, i16*)* @__omp_outlined__ to i8*), i8* null, i8** [[TMP5]], i64 1)
-// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 true)
+// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 true)
// CHECK1-NEXT: ret void
// CHECK1: worker.exit:
// CHECK1-NEXT: ret void
@@ -105,7 +105,7 @@ int bar(int n){
// 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]], i1 true, i1 false, i1 true)
+// CHECK1-NEXT: [[TMP3:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true)
// 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:
@@ -121,7 +121,7 @@ int bar(int n){
// CHECK1-NEXT: store i8* [[TMP10]], i8** [[TMP9]], align 8
// CHECK1-NEXT: [[TMP11:%.*]] = bitcast [3 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**
// CHECK1-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB2]], i32 [[TMP4]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*, i32*, i16*, [10 x i32]*)* @__omp_outlined__1 to i8*), i8* null, i8** [[TMP11]], i64 3)
-// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 true)
+// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 true)
// CHECK1-NEXT: ret void
// CHECK1: worker.exit:
// CHECK1-NEXT: ret void
@@ -165,7 +165,7 @@ int bar(int n){
// 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]+]], i1 true, i1 false, i1 true)
+// CHECK2-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 true)
// 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:
@@ -175,7 +175,7 @@ int bar(int n){
// CHECK2-NEXT: store i8* [[TMP4]], i8** [[TMP3]], align 4
// CHECK2-NEXT: [[TMP5:%.*]] = bitcast [1 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**
// CHECK2-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB2]], i32 [[TMP2]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*, i16*)* @__omp_outlined__ to i8*), i8* null, i8** [[TMP5]], i32 1)
-// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 true)
+// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 true)
// CHECK2-NEXT: ret void
// CHECK2: worker.exit:
// CHECK2-NEXT: ret void
@@ -212,7 +212,7 @@ int bar(int n){
// 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]], i1 true, i1 false, i1 true)
+// CHECK2-NEXT: [[TMP3:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true)
// 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:
@@ -228,7 +228,7 @@ int bar(int n){
// CHECK2-NEXT: store i8* [[TMP10]], i8** [[TMP9]], align 4
// CHECK2-NEXT: [[TMP11:%.*]] = bitcast [3 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**
// CHECK2-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB2]], i32 [[TMP4]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*, i32*, i16*, [10 x i32]*)* @__omp_outlined__1 to i8*), i8* null, i8** [[TMP11]], i32 3)
-// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 true)
+// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 true)
// CHECK2-NEXT: ret void
// CHECK2: worker.exit:
// CHECK2-NEXT: ret void
@@ -272,7 +272,7 @@ int bar(int n){
// 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]+]], i1 true, i1 false, i1 true)
+// CHECK3-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 true)
// 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:
@@ -282,7 +282,7 @@ int bar(int n){
// CHECK3-NEXT: store i8* [[TMP4]], i8** [[TMP3]], align 4
// CHECK3-NEXT: [[TMP5:%.*]] = bitcast [1 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**
// CHECK3-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB2]], i32 [[TMP2]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*, i16*)* @__omp_outlined__ to i8*), i8* null, i8** [[TMP5]], i32 1)
-// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 true)
+// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 true)
// CHECK3-NEXT: ret void
// CHECK3: worker.exit:
// CHECK3-NEXT: ret void
@@ -319,7 +319,7 @@ int bar(int n){
// 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]], i1 true, i1 false, i1 true)
+// CHECK3-NEXT: [[TMP3:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true)
// 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:
@@ -335,7 +335,7 @@ int bar(int n){
// CHECK3-NEXT: store i8* [[TMP10]], i8** [[TMP9]], align 4
// CHECK3-NEXT: [[TMP11:%.*]] = bitcast [3 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**
// CHECK3-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB2]], i32 [[TMP4]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*, i32*, i16*, [10 x i32]*)* @__omp_outlined__1 to i8*), i8* null, i8** [[TMP11]], i32 3)
-// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 true)
+// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 true)
// CHECK3-NEXT: ret void
// CHECK3: worker.exit:
// CHECK3-NEXT: ret void
@@ -379,7 +379,7 @@ int bar(int n){
// 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]+]], i1 true, i1 false, i1 true)
+// CHECK4-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 true)
// 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:
@@ -389,7 +389,7 @@ int bar(int n){
// CHECK4-NEXT: store i8* [[TMP4]], i8** [[TMP3]], align 8
// CHECK4-NEXT: [[TMP5:%.*]] = bitcast [1 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**
// CHECK4-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB2]], i32 [[TMP2]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*, i16*)* @__omp_outlined__ to i8*), i8* null, i8** [[TMP5]], i64 1)
-// CHECK4-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 true)
+// CHECK4-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 true)
// CHECK4-NEXT: ret void
// CHECK4: worker.exit:
// CHECK4-NEXT: ret void
@@ -426,7 +426,7 @@ int bar(int n){
// 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]], i1 true, i1 false, i1 true)
+// CHECK4-NEXT: [[TMP3:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true)
// 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:
@@ -442,7 +442,7 @@ int bar(int n){
// CHECK4-NEXT: store i8* [[TMP10]], i8** [[TMP9]], align 8
// CHECK4-NEXT: [[TMP11:%.*]] = bitcast [3 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**
// CHECK4-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB2]], i32 [[TMP4]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*, i32*, i16*, [10 x i32]*)* @__omp_outlined__1 to i8*), i8* null, i8** [[TMP11]], i64 3)
-// CHECK4-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 true)
+// CHECK4-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 true)
// CHECK4-NEXT: ret void
// CHECK4: worker.exit:
// CHECK4-NEXT: ret void
@@ -486,7 +486,7 @@ int bar(int n){
// 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]+]], i1 true, i1 false, i1 true)
+// CHECK5-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 true)
// 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:
@@ -496,7 +496,7 @@ int bar(int n){
// CHECK5-NEXT: store i8* [[TMP4]], i8** [[TMP3]], align 4
// CHECK5-NEXT: [[TMP5:%.*]] = bitcast [1 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**
// CHECK5-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB2]], i32 [[TMP2]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*, i16*)* @__omp_outlined__ to i8*), i8* null, i8** [[TMP5]], i32 1)
-// CHECK5-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 true)
+// CHECK5-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 true)
// CHECK5-NEXT: ret void
// CHECK5: worker.exit:
// CHECK5-NEXT: ret void
@@ -533,7 +533,7 @@ int bar(int n){
// 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]], i1 true, i1 false, i1 true)
+// CHECK5-NEXT: [[TMP3:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true)
// 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:
@@ -549,7 +549,7 @@ int bar(int n){
// CHECK5-NEXT: store i8* [[TMP10]], i8** [[TMP9]], align 4
// CHECK5-NEXT: [[TMP11:%.*]] = bitcast [3 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**
// CHECK5-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB2]], i32 [[TMP4]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*, i32*, i16*, [10 x i32]*)* @__omp_outlined__1 to i8*), i8* null, i8** [[TMP11]], i32 3)
-// CHECK5-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 true)
+// CHECK5-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 true)
// CHECK5-NEXT: ret void
// CHECK5: worker.exit:
// CHECK5-NEXT: ret void
@@ -593,7 +593,7 @@ int bar(int n){
// 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]+]], i1 true, i1 false, i1 true)
+// CHECK6-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 true)
// 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:
@@ -603,7 +603,7 @@ int bar(int n){
// CHECK6-NEXT: store i8* [[TMP4]], i8** [[TMP3]], align 4
// CHECK6-NEXT: [[TMP5:%.*]] = bitcast [1 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**
// CHECK6-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB2]], i32 [[TMP2]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*, i16*)* @__omp_outlined__ to i8*), i8* null, i8** [[TMP5]], i32 1)
-// CHECK6-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 true)
+// CHECK6-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 true)
// CHECK6-NEXT: ret void
// CHECK6: worker.exit:
// CHECK6-NEXT: ret void
@@ -640,7 +640,7 @@ int bar(int n){
// 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]], i1 true, i1 false, i1 true)
+// CHECK6-NEXT: [[TMP3:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true)
// 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:
@@ -656,7 +656,7 @@ int bar(int n){
// CHECK6-NEXT: store i8* [[TMP10]], i8** [[TMP9]], align 4
// CHECK6-NEXT: [[TMP11:%.*]] = bitcast [3 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**
// CHECK6-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB2]], i32 [[TMP4]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*, i32*, i16*, [10 x i32]*)* @__omp_outlined__1 to i8*), i8* null, i8** [[TMP11]], i32 3)
-// CHECK6-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 true)
+// CHECK6-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 true)
// CHECK6-NEXT: ret void
// CHECK6: worker.exit:
// CHECK6-NEXT: ret void
diff --git a/clang/test/OpenMP/nvptx_target_parallel_num_threads_codegen.cpp b/clang/test/OpenMP/nvptx_target_parallel_num_threads_codegen.cpp
index 84d5e70816c1f..fb1e427cfed72 100644
--- 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 @@ int bar(int n){
// 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]+]], i1 true, i1 false, i1 true)
+// CHECK1-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 true)
// 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:
@@ -63,7 +63,7 @@ int bar(int n){
// CHECK1-NEXT: store i8* [[TMP4]], i8** [[TMP3]], align 8
// CHECK1-NEXT: [[TMP5:%.*]] = bitcast [1 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**
// CHECK1-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB2]], i32 [[TMP2]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*, i16*)* @__omp_outlined__ to i8*), i8* null, i8** [[TMP5]], i64 1)
-// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 true)
+// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 true)
// CHECK1-NEXT: ret void
// CHECK1: worker.exit:
// CHECK1-NEXT: ret void
@@ -103,7 +103,7 @@ int bar(int n){
// 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]], i1 true, i1 false, i1 true)
+// CHECK1-NEXT: [[TMP3:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true)
// 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:
@@ -120,7 +120,7 @@ int bar(int n){
// CHECK1-NEXT: store i8* [[TMP11]], i8** [[TMP10]], align 8
// CHECK1-NEXT: [[TMP12:%.*]] = bitcast [3 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**
// CHECK1-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB2]], i32 [[TMP4]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*, i32*, i16*, [10 x i32]*)* @__omp_outlined__1 to i8*), i8* null, i8** [[TMP12]], i64 3)
-// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 true)
+// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 true)
// CHECK1-NEXT: ret void
// CHECK1: worker.exit:
// CHECK1-NEXT: ret void
@@ -164,7 +164,7 @@ int bar(int n){
// 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]+]], i1 true, i1 false, i1 true)
+// CHECK2-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 true)
// 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:
@@ -174,7 +174,7 @@ int bar(int n){
// CHECK2-NEXT: store i8* [[TMP4]], i8** [[TMP3]], align 4
// CHECK2-NEXT: [[TMP5:%.*]] = bitcast [1 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**
// CHECK2-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB2]], i32 [[TMP2]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*, i16*)* @__omp_outlined__ to i8*), i8* null, i8** [[TMP5]], i32 1)
-// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 true)
+// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 true)
// CHECK2-NEXT: ret void
// CHECK2: worker.exit:
// CHECK2-NEXT: ret void
@@ -213,7 +213,7 @@ int bar(int n){
// 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]], i1 true, i1 false, i1 true)
+// CHECK2-NEXT: [[TMP3:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true)
// 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:
@@ -230,7 +230,7 @@ int bar(int n){
// CHECK2-NEXT: store i8* [[TMP11]], i8** [[TMP10]], align 4
// CHECK2-NEXT: [[TMP12:%.*]] = bitcast [3 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**
// CHECK2-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB2]], i32 [[TMP4]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*, i32*, i16*, [10 x i32]*)* @__omp_outlined__1 to i8*), i8* null, i8** [[TMP12]], i32 3)
-// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 true)
+// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 true)
// CHECK2-NEXT: ret void
// CHECK2: worker.exit:
// CHECK2-NEXT: ret void
@@ -274,7 +274,7 @@ int bar(int n){
// 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]+]], i1 true, i1 false, i1 true)
+// CHECK3-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 true)
// 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:
@@ -284,7 +284,7 @@ int bar(int n){
// CHECK3-NEXT: store i8* [[TMP4]], i8** [[TMP3]], align 4
// CHECK3-NEXT: [[TMP5:%.*]] = bitcast [1 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**
// CHECK3-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB2]], i32 [[TMP2]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*, i16*)* @__omp_outlined__ to i8*), i8* null, i8** [[TMP5]], i32 1)
-// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 true)
+// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 true)
// CHECK3-NEXT: ret void
// CHECK3: worker.exit:
// CHECK3-NEXT: ret void
@@ -323,7 +323,7 @@ int bar(int n){
// 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]], i1 true, i1 false, i1 true)
+// CHECK3-NEXT: [[TMP3:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true)
// 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:
@@ -340,7 +340,7 @@ int bar(int n){
// CHECK3-NEXT: store i8* [[TMP11]], i8** [[TMP10]], align 4
// CHECK3-NEXT: [[TMP12:%.*]] = bitcast [3 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**
// CHECK3-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB2]], i32 [[TMP4]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*, i32*, i16*, [10 x i32]*)* @__omp_outlined__1 to i8*), i8* null, i8** [[TMP12]], i32 3)
-// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 true)
+// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 true)
// CHECK3-NEXT: ret void
// CHECK3: worker.exit:
// CHECK3-NEXT: ret void
@@ -384,7 +384,7 @@ int bar(int n){
// 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]+]], i1 true, i1 false, i1 true)
+// CHECK4-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 true)
// 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:
@@ -394,7 +394,7 @@ int bar(int n){
// CHECK4-NEXT: store i8* [[TMP4]], i8** [[TMP3]], align 8
// CHECK4-NEXT: [[TMP5:%.*]] = bitcast [1 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**
// CHECK4-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB2]], i32 [[TMP2]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*, i16*)* @__omp_outlined__ to i8*), i8* null, i8** [[TMP5]], i64 1)
-// CHECK4-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 true)
+// CHECK4-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 true)
// CHECK4-NEXT: ret void
// CHECK4: worker.exit:
// CHECK4-NEXT: ret void
@@ -434,7 +434,7 @@ int bar(int n){
// 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]], i1 true, i1 false, i1 true)
+// CHECK4-NEXT: [[TMP3:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true)
// 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:
@@ -451,7 +451,7 @@ int bar(int n){
// CHECK4-NEXT: store i8* [[TMP11]], i8** [[TMP10]], align 8
// CHECK4-NEXT: [[TMP12:%.*]] = bitcast [3 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**
// CHECK4-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB2]], i32 [[TMP4]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*, i32*, i16*, [10 x i32]*)* @__omp_outlined__1 to i8*), i8* null, i8** [[TMP12]], i64 3)
-// CHECK4-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 true)
+// CHECK4-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 true)
// CHECK4-NEXT: ret void
// CHECK4: worker.exit:
// CHECK4-NEXT: ret void
@@ -495,7 +495,7 @@ int bar(int n){
// 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]+]], i1 true, i1 false, i1 true)
+// CHECK5-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 true)
// 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:
@@ -505,7 +505,7 @@ int bar(int n){
// CHECK5-NEXT: store i8* [[TMP4]], i8** [[TMP3]], align 4
// CHECK5-NEXT: [[TMP5:%.*]] = bitcast [1 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**
// CHECK5-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB2]], i32 [[TMP2]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*, i16*)* @__omp_outlined__ to i8*), i8* null, i8** [[TMP5]], i32 1)
-// CHECK5-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 true)
+// CHECK5-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 true)
// CHECK5-NEXT: ret void
// CHECK5: worker.exit:
// CHECK5-NEXT: ret void
@@ -544,7 +544,7 @@ int bar(int n){
// 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]], i1 true, i1 false, i1 true)
+// CHECK5-NEXT: [[TMP3:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true)
// 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:
@@ -561,7 +561,7 @@ int bar(int n){
// CHECK5-NEXT: store i8* [[TMP11]], i8** [[TMP10]], align 4
// CHECK5-NEXT: [[TMP12:%.*]] = bitcast [3 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**
// CHECK5-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB2]], i32 [[TMP4]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*, i32*, i16*, [10 x i32]*)* @__omp_outlined__1 to i8*), i8* null, i8** [[TMP12]], i32 3)
-// CHECK5-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 true)
+// CHECK5-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 true)
// CHECK5-NEXT: ret void
// CHECK5: worker.exit:
// CHECK5-NEXT: ret void
@@ -605,7 +605,7 @@ int bar(int n){
// 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]+]], i1 true, i1 false, i1 true)
+// CHECK6-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 true)
// 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:
@@ -615,7 +615,7 @@ int bar(int n){
// CHECK6-NEXT: store i8* [[TMP4]], i8** [[TMP3]], align 4
// CHECK6-NEXT: [[TMP5:%.*]] = bitcast [1 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**
// CHECK6-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB2]], i32 [[TMP2]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*, i16*)* @__omp_outlined__ to i8*), i8* null, i8** [[TMP5]], i32 1)
-// CHECK6-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 true)
+// CHECK6-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 true)
// CHECK6-NEXT: ret void
// CHECK6: worker.exit:
// CHECK6-NEXT: ret void
@@ -654,7 +654,7 @@ int bar(int n){
// 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]], i1 true, i1 false, i1 true)
+// CHECK6-NEXT: [[TMP3:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true)
// 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:
@@ -671,7 +671,7 @@ int bar(int n){
// CHECK6-NEXT: store i8* [[TMP11]], i8** [[TMP10]], align 4
// CHECK6-NEXT: [[TMP12:%.*]] = bitcast [3 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**
// CHECK6-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB2]], i32 [[TMP4]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*, i32*, i16*, [10 x i32]*)* @__omp_outlined__1 to i8*), i8* null, i8** [[TMP12]], i32 3)
-// CHECK6-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 true)
+// CHECK6-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 true)
// CHECK6-NEXT: ret void
// CHECK6: worker.exit:
// CHECK6-NEXT: ret void
diff --git a/clang/test/OpenMP/nvptx_target_parallel_reduction_codegen.cpp b/clang/test/OpenMP/nvptx_target_parallel_reduction_codegen.cpp
index fdba4e95b329b..9bb6b15536d7e 100644
--- a/clang/test/OpenMP/nvptx_target_parallel_reduction_codegen.cpp
+++ b/clang/test/OpenMP/nvptx_target_parallel_reduction_codegen.cpp
@@ -54,8 +54,8 @@ int bar(int n){
// CHECK: define {{.*}}void {{@__omp_offloading_.+template.+l27}}(
//
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
-// CHECK: call void @__kmpc_target_deinit({{.*}}, i1 true, i1 true)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true)
+// CHECK: call void @__kmpc_target_deinit({{.*}}, i8 2, i1 true)
//
//
// define internal void [[PFN]](
@@ -233,8 +233,8 @@ int bar(int n){
// CHECK: define {{.*}}void {{@__omp_offloading_.+template.+l32}}(
//
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
-// CHECK: call void @__kmpc_target_deinit({{.*}}, i1 true, i1 true)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true)
+// CHECK: call void @__kmpc_target_deinit({{.*}}, i8 2, i1 true)
//
//
// define internal void [[PFN1]](
@@ -490,8 +490,8 @@ int bar(int n){
// CHECK: define {{.*}}void {{@__omp_offloading_.+template.+l38}}(
//
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 true)
-// CHECK: call void @__kmpc_target_deinit({{.*}}, i1 true, i1 true)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 true)
+// CHECK: call void @__kmpc_target_deinit({{.*}}, i8 2, i1 true)
//
//
// define internal void [[PFN2]](
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
index 1aeef86af7516..fcb3445d7e962 100644
--- a/clang/test/OpenMP/nvptx_target_parallel_reduction_codegen_tbaa_PR46146.cpp
+++ b/clang/test/OpenMP/nvptx_target_parallel_reduction_codegen_tbaa_PR46146.cpp
@@ -36,14 +36,14 @@ void test() {
// 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]+]], i1 false, i1 true, i1 true)
+// CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true)
// 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: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]])
// CHECK1-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4, !tbaa [[TBAA8:![0-9]+]]
// CHECK1-NEXT: call void @__omp_outlined__(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]]) #[[ATTR1:[0-9]+]]
-// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK1-NEXT: ret void
// CHECK1: worker.exit:
// CHECK1-NEXT: ret void
@@ -603,14 +603,14 @@ void test() {
// 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]], i1 false, i1 true, i1 true)
+// CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
// 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: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]])
// CHECK1-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4, !tbaa [[TBAA8]]
// CHECK1-NEXT: call void @__omp_outlined__2(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]]) #[[ATTR1]]
-// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK1-NEXT: ret void
// CHECK1: worker.exit:
// CHECK1-NEXT: ret void
@@ -1270,14 +1270,14 @@ void test() {
// 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]+]], i1 false, i1 true, i1 true)
+// CHECK2-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true)
// 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: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]])
// CHECK2-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4, !tbaa [[TBAA8:![0-9]+]]
// CHECK2-NEXT: call void @__omp_outlined__(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]]) #[[ATTR1:[0-9]+]]
-// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK2-NEXT: ret void
// CHECK2: worker.exit:
// CHECK2-NEXT: ret void
@@ -1837,14 +1837,14 @@ void test() {
// 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]], i1 false, i1 true, i1 true)
+// CHECK2-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
// 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: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]])
// CHECK2-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4, !tbaa [[TBAA8]]
// CHECK2-NEXT: call void @__omp_outlined__2(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]]) #[[ATTR1]]
-// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK2-NEXT: ret void
// CHECK2: worker.exit:
// CHECK2-NEXT: ret void
@@ -2504,14 +2504,14 @@ void test() {
// 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]+]], i1 false, i1 true, i1 true)
+// CHECK3-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true)
// 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: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]])
// CHECK3-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4, !tbaa [[TBAA8:![0-9]+]]
// CHECK3-NEXT: call void @__omp_outlined__(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]]) #[[ATTR1:[0-9]+]]
-// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK3-NEXT: ret void
// CHECK3: worker.exit:
// CHECK3-NEXT: ret void
@@ -3071,14 +3071,14 @@ void test() {
// 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]], i1 false, i1 true, i1 true)
+// CHECK3-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
// 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: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]])
// CHECK3-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4, !tbaa [[TBAA8]]
// CHECK3-NEXT: call void @__omp_outlined__2(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]]) #[[ATTR1]]
-// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK3-NEXT: ret void
// CHECK3: worker.exit:
// CHECK3-NEXT: ret void
diff --git a/clang/test/OpenMP/nvptx_target_printf_codegen.c b/clang/test/OpenMP/nvptx_target_printf_codegen.c
index 9e461944f8bc6..361943138c32c 100644
--- a/clang/test/OpenMP/nvptx_target_printf_codegen.c
+++ b/clang/test/OpenMP/nvptx_target_printf_codegen.c
@@ -48,7 +48,7 @@ void CheckAllocaIsInEntryBlock() {
// 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]+]], i1 false, 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)
// 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:
@@ -62,7 +62,7 @@ void CheckAllocaIsInEntryBlock() {
// CHECK-64-NEXT: store double 3.000000e+00, double* [[TMP4]], align 8
// CHECK-64-NEXT: [[TMP5:%.*]] = bitcast %printf_args* [[TMP]] to i8*
// CHECK-64-NEXT: [[TMP6:%.*]] = call i32 @vprintf(i8* [[TMP1]], i8* [[TMP5]])
-// CHECK-64-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK-64-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK-64-NEXT: ret void
// CHECK-64: worker.exit:
// CHECK-64-NEXT: ret void
@@ -71,12 +71,12 @@ void CheckAllocaIsInEntryBlock() {
// 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]], i1 false, i1 true, i1 true)
+// CHECK-64-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
// 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:
// CHECK-64-NEXT: [[TMP1:%.*]] = call i32 @vprintf(i8* getelementptr inbounds ([14 x i8], [14 x i8]* @.str1, i64 0, i64 0), i8* null)
-// CHECK-64-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK-64-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK-64-NEXT: ret void
// CHECK-64: worker.exit:
// CHECK-64-NEXT: ret void
@@ -89,7 +89,7 @@ void CheckAllocaIsInEntryBlock() {
// 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]], i1 false, i1 true, i1 true)
+// CHECK-64-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
// 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:
@@ -105,7 +105,7 @@ void CheckAllocaIsInEntryBlock() {
// CHECK-64: worker.exit:
// CHECK-64-NEXT: ret void
// CHECK-64: if.end:
-// CHECK-64-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK-64-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK-64-NEXT: ret void
//
//
@@ -117,7 +117,7 @@ void CheckAllocaIsInEntryBlock() {
// 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]+]], i1 false, 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)
// 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:
@@ -131,7 +131,7 @@ void CheckAllocaIsInEntryBlock() {
// CHECK-32-NEXT: store double 3.000000e+00, double* [[TMP4]], align 8
// CHECK-32-NEXT: [[TMP5:%.*]] = bitcast %printf_args* [[TMP]] to i8*
// CHECK-32-NEXT: [[TMP6:%.*]] = call i32 @vprintf(i8* [[TMP1]], i8* [[TMP5]])
-// CHECK-32-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK-32-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK-32-NEXT: ret void
// CHECK-32: worker.exit:
// CHECK-32-NEXT: ret void
@@ -140,12 +140,12 @@ void CheckAllocaIsInEntryBlock() {
// 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]], i1 false, i1 true, i1 true)
+// CHECK-32-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
// 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:
// CHECK-32-NEXT: [[TMP1:%.*]] = call i32 @vprintf(i8* getelementptr inbounds ([14 x i8], [14 x i8]* @.str1, i32 0, i32 0), i8* null)
-// CHECK-32-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK-32-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK-32-NEXT: ret void
// CHECK-32: worker.exit:
// CHECK-32-NEXT: ret void
@@ -157,7 +157,7 @@ void CheckAllocaIsInEntryBlock() {
// 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]], i1 false, i1 true, i1 true)
+// CHECK-32-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
// 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:
@@ -173,6 +173,6 @@ void CheckAllocaIsInEntryBlock() {
// CHECK-32: worker.exit:
// CHECK-32-NEXT: ret void
// CHECK-32: if.end:
-// CHECK-32-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK-32-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK-32-NEXT: ret void
//
diff --git a/clang/test/OpenMP/nvptx_target_simd_codegen.cpp b/clang/test/OpenMP/nvptx_target_simd_codegen.cpp
index 659f1e680b55f..6e4341f8b5606 100644
--- a/clang/test/OpenMP/nvptx_target_simd_codegen.cpp
+++ b/clang/test/OpenMP/nvptx_target_simd_codegen.cpp
@@ -61,32 +61,32 @@ int bar(int n){
}
// CHECK-LABEL: define {{.*}}void {{@__omp_offloading_.+l32}}(
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 false)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false)
// CHECK-NOT: call void @__kmpc_for_static_init
// CHECK-NOT: call void @__kmpc_for_static_fini
-// CHECK: call void @__kmpc_target_deinit({{.*}}, i1 true, i1 false)
+// 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({{.*}}, i1 true, i1 false, i1 false)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false)
// CHECK-NOT: call void @__kmpc_for_static_init
// CHECK-NOT: call void @__kmpc_for_static_fini
-// CHECK: call void @__kmpc_target_deinit({{.*}}, i1 true, i1 false)
+// 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({{.*}}, i1 true, i1 false, i1 false)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false)
// CHECK-NOT: call void @__kmpc_for_static_init
// CHECK-NOT: call void @__kmpc_for_static_fini
-// CHECK: call void @__kmpc_target_deinit({{.*}}, i1 true, i1 false)
+// 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({{.*}}, i1 true, i1 false, i1 false)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false)
// 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(
-// CHECK: call void @__kmpc_target_deinit({{.*}}, i1 true, i1 false)
+// CHECK: call void @__kmpc_target_deinit({{.*}}, i8 2, i1 false)
// CHECK: ret void
#endif
diff --git a/clang/test/OpenMP/nvptx_target_teams_codegen.cpp b/clang/test/OpenMP/nvptx_target_teams_codegen.cpp
index 6098f722e59da..2072961deb5f3 100644
--- a/clang/test/OpenMP/nvptx_target_teams_codegen.cpp
+++ b/clang/test/OpenMP/nvptx_target_teams_codegen.cpp
@@ -59,7 +59,7 @@ int bar(int n){
// 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]+]], i1 false, i1 true, i1 true)
+// CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true)
// 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:
@@ -70,7 +70,7 @@ int bar(int n){
// CHECK1-NEXT: [[TMP3:%.*]] = load i64, i64* [[A_CASTED]], align 8
// CHECK1-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
// CHECK1-NEXT: call void @__omp_outlined__(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i64 [[TMP3]]) #[[ATTR1:[0-9]+]]
-// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK1-NEXT: ret void
// CHECK1: worker.exit:
// CHECK1-NEXT: ret void
@@ -100,7 +100,7 @@ int bar(int n){
// 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]], i1 false, i1 true, i1 true)
+// CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
// 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:
@@ -111,7 +111,7 @@ int bar(int n){
// CHECK1-NEXT: [[TMP3:%.*]] = load i64, i64* [[AA_CASTED]], align 8
// CHECK1-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
// CHECK1-NEXT: call void @__omp_outlined__1(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i64 [[TMP3]]) #[[ATTR1]]
-// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK1-NEXT: ret void
// CHECK1: worker.exit:
// CHECK1-NEXT: ret void
@@ -141,7 +141,7 @@ int bar(int n){
// 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]], i1 true, i1 false, i1 true)
+// CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true)
// 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:
@@ -152,7 +152,7 @@ int bar(int n){
// CHECK1-NEXT: [[TMP3:%.*]] = load i64, i64* [[AA_CASTED]], align 8
// CHECK1-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
// CHECK1-NEXT: call void @__omp_outlined__2(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i64 [[TMP3]]) #[[ATTR1]]
-// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 true)
+// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 true)
// CHECK1-NEXT: ret void
// CHECK1: worker.exit:
// CHECK1-NEXT: ret void
@@ -224,7 +224,7 @@ int bar(int n){
// 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]+]], i1 false, i1 true, i1 true)
+// CHECK2-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true)
// 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:
@@ -235,7 +235,7 @@ int bar(int n){
// CHECK2-NEXT: [[TMP3:%.*]] = load i32, i32* [[A_CASTED]], align 4
// CHECK2-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
// CHECK2-NEXT: call void @__omp_outlined__(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i32 [[TMP3]]) #[[ATTR1:[0-9]+]]
-// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK2-NEXT: ret void
// CHECK2: worker.exit:
// CHECK2-NEXT: ret void
@@ -265,7 +265,7 @@ int bar(int n){
// 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]], i1 false, i1 true, i1 true)
+// CHECK2-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
// 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:
@@ -276,7 +276,7 @@ int bar(int n){
// CHECK2-NEXT: [[TMP3:%.*]] = load i32, i32* [[AA_CASTED]], align 4
// CHECK2-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
// CHECK2-NEXT: call void @__omp_outlined__1(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i32 [[TMP3]]) #[[ATTR1]]
-// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK2-NEXT: ret void
// CHECK2: worker.exit:
// CHECK2-NEXT: ret void
@@ -306,7 +306,7 @@ int bar(int n){
// 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]], i1 true, i1 false, i1 true)
+// CHECK2-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true)
// 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:
@@ -317,7 +317,7 @@ int bar(int n){
// CHECK2-NEXT: [[TMP3:%.*]] = load i32, i32* [[AA_CASTED]], align 4
// CHECK2-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
// CHECK2-NEXT: call void @__omp_outlined__2(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i32 [[TMP3]]) #[[ATTR1]]
-// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 true)
+// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 true)
// CHECK2-NEXT: ret void
// CHECK2: worker.exit:
// CHECK2-NEXT: ret void
@@ -389,7 +389,7 @@ int bar(int n){
// 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]+]], i1 false, i1 true, i1 true)
+// CHECK3-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true)
// 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:
@@ -400,7 +400,7 @@ int bar(int n){
// CHECK3-NEXT: [[TMP3:%.*]] = load i32, i32* [[A_CASTED]], align 4
// CHECK3-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
// CHECK3-NEXT: call void @__omp_outlined__(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i32 [[TMP3]]) #[[ATTR1:[0-9]+]]
-// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK3-NEXT: ret void
// CHECK3: worker.exit:
// CHECK3-NEXT: ret void
@@ -430,7 +430,7 @@ int bar(int n){
// 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]], i1 false, i1 true, i1 true)
+// CHECK3-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
// 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:
@@ -441,7 +441,7 @@ int bar(int n){
// CHECK3-NEXT: [[TMP3:%.*]] = load i32, i32* [[AA_CASTED]], align 4
// CHECK3-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
// CHECK3-NEXT: call void @__omp_outlined__1(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i32 [[TMP3]]) #[[ATTR1]]
-// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK3-NEXT: ret void
// CHECK3: worker.exit:
// CHECK3-NEXT: ret void
@@ -471,7 +471,7 @@ int bar(int n){
// 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]], i1 true, i1 false, i1 true)
+// CHECK3-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true)
// 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:
@@ -482,7 +482,7 @@ int bar(int n){
// CHECK3-NEXT: [[TMP3:%.*]] = load i32, i32* [[AA_CASTED]], align 4
// CHECK3-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
// CHECK3-NEXT: call void @__omp_outlined__2(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i32 [[TMP3]]) #[[ATTR1]]
-// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 true)
+// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 true)
// CHECK3-NEXT: ret void
// CHECK3: worker.exit:
// CHECK3-NEXT: ret void
diff --git a/clang/test/OpenMP/nvptx_target_teams_distribute_codegen.cpp b/clang/test/OpenMP/nvptx_target_teams_distribute_codegen.cpp
index e2af88f71eaac..0327ea356e59c 100644
--- a/clang/test/OpenMP/nvptx_target_teams_distribute_codegen.cpp
+++ b/clang/test/OpenMP/nvptx_target_teams_distribute_codegen.cpp
@@ -580,14 +580,14 @@ int bar(int n){
// 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]+]], i1 false, i1 true, i1 true)
+// CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true)
// 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: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]])
// CHECK1-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
-// CHECK1-NEXT: call void @__omp_outlined__(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]]) #[[ATTR1:[0-9]+]]
-// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK1-NEXT: call void @__omp_outlined__(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]]) #[[ATTR4:[0-9]+]]
+// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK1-NEXT: ret void
// CHECK1: worker.exit:
// CHECK1-NEXT: ret void
@@ -692,7 +692,7 @@ int bar(int n){
// CHECK1-NEXT: [[TMP3:%.*]] = getelementptr inbounds i8*, i8** [[TMP2]], i64 0
// CHECK1-NEXT: [[TMP4:%.*]] = bitcast i8** [[TMP3]] to i32**
// CHECK1-NEXT: [[TMP5:%.*]] = load i32*, i32** [[TMP4]], align 8
-// CHECK1-NEXT: call void @__omp_outlined__1(i32* [[DOTADDR1]], i32* [[DOTZERO_ADDR]], i32* [[TMP5]]) #[[ATTR1]]
+// CHECK1-NEXT: call void @__omp_outlined__1(i32* [[DOTADDR1]], i32* [[DOTZERO_ADDR]], i32* [[TMP5]]) #[[ATTR4]]
// CHECK1-NEXT: ret void
//
//
@@ -702,14 +702,14 @@ int bar(int n){
// 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]+]], i1 false, i1 true, i1 true)
+// CHECK2-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true)
// 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: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]])
// CHECK2-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
-// CHECK2-NEXT: call void @__omp_outlined__(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]]) #[[ATTR1:[0-9]+]]
-// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK2-NEXT: call void @__omp_outlined__(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]]) #[[ATTR4:[0-9]+]]
+// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK2-NEXT: ret void
// CHECK2: worker.exit:
// CHECK2-NEXT: ret void
@@ -814,7 +814,7 @@ int bar(int n){
// CHECK2-NEXT: [[TMP3:%.*]] = getelementptr inbounds i8*, i8** [[TMP2]], i32 0
// CHECK2-NEXT: [[TMP4:%.*]] = bitcast i8** [[TMP3]] to i32**
// CHECK2-NEXT: [[TMP5:%.*]] = load i32*, i32** [[TMP4]], align 4
-// CHECK2-NEXT: call void @__omp_outlined__1(i32* [[DOTADDR1]], i32* [[DOTZERO_ADDR]], i32* [[TMP5]]) #[[ATTR1]]
+// CHECK2-NEXT: call void @__omp_outlined__1(i32* [[DOTADDR1]], i32* [[DOTZERO_ADDR]], i32* [[TMP5]]) #[[ATTR4]]
// CHECK2-NEXT: ret void
//
//
@@ -824,14 +824,14 @@ int bar(int n){
// 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]+]], i1 false, i1 true, i1 true)
+// CHECK3-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true)
// 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: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]])
// CHECK3-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
-// CHECK3-NEXT: call void @__omp_outlined__(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]]) #[[ATTR1:[0-9]+]]
-// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK3-NEXT: call void @__omp_outlined__(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]]) #[[ATTR4:[0-9]+]]
+// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK3-NEXT: ret void
// CHECK3: worker.exit:
// CHECK3-NEXT: ret void
@@ -936,6 +936,6 @@ int bar(int n){
// CHECK3-NEXT: [[TMP3:%.*]] = getelementptr inbounds i8*, i8** [[TMP2]], i32 0
// CHECK3-NEXT: [[TMP4:%.*]] = bitcast i8** [[TMP3]] to i32**
// CHECK3-NEXT: [[TMP5:%.*]] = load i32*, i32** [[TMP4]], align 4
-// CHECK3-NEXT: call void @__omp_outlined__1(i32* [[DOTADDR1]], i32* [[DOTZERO_ADDR]], i32* [[TMP5]]) #[[ATTR1]]
+// CHECK3-NEXT: call void @__omp_outlined__1(i32* [[DOTADDR1]], i32* [[DOTZERO_ADDR]], i32* [[TMP5]]) #[[ATTR4]]
// CHECK3-NEXT: ret void
//
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
index 9818b6a2e617f..5d9f5108b7503 100644
--- 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 @@ int bar(int n){
// 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]+]], i1 true, i1 false, i1 false)
+// CHECK1-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 false)
// 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:
@@ -18498,8 +18498,8 @@ int bar(int n){
// CHECK1-NEXT: store i32 [[TMP5]], i32* [[CONV3]], align 4
// CHECK1-NEXT: [[TMP6:%.*]] = load i64, i64* [[L_CASTED]], align 8
// CHECK1-NEXT: store i32 [[TMP2]], i32* [[DOTTHREADID_TEMP_]], align 4
-// CHECK1-NEXT: call void @__omp_outlined__(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i64 [[TMP4]], [1000 x i32]* [[TMP0]], i64 [[TMP6]]) #[[ATTR1:[0-9]+]]
-// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 false)
+// CHECK1-NEXT: call void @__omp_outlined__(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i64 [[TMP4]], [1000 x i32]* [[TMP0]], i64 [[TMP6]]) #[[ATTR2:[0-9]+]]
+// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 false)
// CHECK1-NEXT: ret void
// CHECK1: worker.exit:
// CHECK1-NEXT: ret void
@@ -18795,7 +18795,7 @@ int bar(int n){
//
//
// CHECK1-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l34
-// CHECK1-SAME: (i64 [[N:%.*]], [1000 x i16]* nonnull align 2 dereferenceable(2000) [[AA:%.*]]) #[[ATTR2:[0-9]+]] {
+// CHECK1-SAME: (i64 [[N:%.*]], [1000 x i16]* nonnull align 2 dereferenceable(2000) [[AA:%.*]]) #[[ATTR4:[0-9]+]] {
// CHECK1-NEXT: entry:
// CHECK1-NEXT: [[N_ADDR:%.*]] = alloca i64, align 8
// CHECK1-NEXT: [[AA_ADDR:%.*]] = alloca [1000 x i16]*, align 8
@@ -18807,7 +18807,7 @@ int bar(int n){
// 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]], i1 true, i1 false, i1 false)
+// CHECK1-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false)
// 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:
@@ -18817,8 +18817,8 @@ int bar(int n){
// CHECK1-NEXT: store i32 [[TMP3]], i32* [[CONV1]], align 4
// CHECK1-NEXT: [[TMP4:%.*]] = load i64, i64* [[N_CASTED]], align 8
// CHECK1-NEXT: store i32 [[TMP2]], i32* [[DOTTHREADID_TEMP_]], align 4
-// CHECK1-NEXT: call void @__omp_outlined__2(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i64 [[TMP4]], [1000 x i16]* [[TMP0]]) #[[ATTR1]]
-// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 false)
+// CHECK1-NEXT: call void @__omp_outlined__2(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i64 [[TMP4]], [1000 x i16]* [[TMP0]]) #[[ATTR2]]
+// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 false)
// CHECK1-NEXT: ret void
// CHECK1: worker.exit:
// CHECK1-NEXT: ret void
@@ -19062,14 +19062,14 @@ int bar(int n){
// 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]], i1 true, i1 false, i1 false)
+// CHECK1-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false)
// 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* @[[GLOB4]])
// CHECK1-NEXT: store i32 [[TMP2]], i32* [[DOTTHREADID_TEMP_]], align 4
-// CHECK1-NEXT: call void @__omp_outlined__4(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], [10 x i32]* [[TMP0]]) #[[ATTR1]]
-// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 false)
+// CHECK1-NEXT: call void @__omp_outlined__4(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], [10 x i32]* [[TMP0]]) #[[ATTR2]]
+// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 false)
// CHECK1-NEXT: ret void
// CHECK1: worker.exit:
// CHECK1-NEXT: ret void
@@ -19253,7 +19253,7 @@ int bar(int n){
// 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]], i1 true, i1 false, i1 false)
+// CHECK1-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false)
// 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:
@@ -19263,8 +19263,8 @@ int bar(int n){
// CHECK1-NEXT: store i32 [[TMP3]], i32* [[CONV1]], align 4
// CHECK1-NEXT: [[TMP4:%.*]] = load i64, i64* [[F_CASTED]], align 8
// CHECK1-NEXT: store i32 [[TMP2]], i32* [[DOTTHREADID_TEMP_]], align 4
-// CHECK1-NEXT: call void @__omp_outlined__6(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], [10 x [10 x i32]]* [[TMP0]], i64 [[TMP4]]) #[[ATTR1]]
-// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 false)
+// CHECK1-NEXT: call void @__omp_outlined__6(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], [10 x [10 x i32]]* [[TMP0]], i64 [[TMP4]]) #[[ATTR2]]
+// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 false)
// CHECK1-NEXT: ret void
// CHECK1: worker.exit:
// CHECK1-NEXT: ret void
@@ -19486,7 +19486,7 @@ int bar(int n){
// 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]], i1 true, i1 false, i1 false)
+// CHECK1-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false)
// 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:
@@ -19496,8 +19496,8 @@ int bar(int n){
// CHECK1-NEXT: store i32 [[TMP3]], i32* [[CONV1]], align 4
// CHECK1-NEXT: [[TMP4:%.*]] = load i64, i64* [[N_CASTED]], align 8
// CHECK1-NEXT: store i32 [[TMP2]], i32* [[DOTTHREADID_TEMP_]], align 4
-// CHECK1-NEXT: call void @__omp_outlined__8(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i64 [[TMP4]], [10 x [10 x i32]]* [[TMP0]]) #[[ATTR1]]
-// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 false)
+// CHECK1-NEXT: call void @__omp_outlined__8(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i64 [[TMP4]], [10 x [10 x i32]]* [[TMP0]]) #[[ATTR2]]
+// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 false)
// CHECK1-NEXT: ret void
// CHECK1: worker.exit:
// CHECK1-NEXT: ret void
@@ -19805,7 +19805,7 @@ int bar(int n){
// 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]], i1 true, i1 false, i1 false)
+// CHECK1-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false)
// 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:
@@ -19816,8 +19816,8 @@ int bar(int n){
// CHECK1-NEXT: [[TMP4:%.*]] = load i64, i64* [[N_CASTED]], align 8
// CHECK1-NEXT: [[TMP5:%.*]] = load i32*, i32** [[V_ADDR]], align 8
// CHECK1-NEXT: store i32 [[TMP2]], i32* [[DOTTHREADID_TEMP_]], align 4
-// CHECK1-NEXT: call void @__omp_outlined__10(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i64 [[TMP4]], [1000 x i32]* [[TMP0]], i32* [[TMP5]]) #[[ATTR1]]
-// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 false)
+// CHECK1-NEXT: call void @__omp_outlined__10(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i64 [[TMP4]], [1000 x i32]* [[TMP0]], i32* [[TMP5]]) #[[ATTR2]]
+// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 false)
// CHECK1-NEXT: ret void
// CHECK1: worker.exit:
// CHECK1-NEXT: ret void
@@ -20078,7 +20078,7 @@ int bar(int n){
// 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]+]], i1 true, i1 false, i1 false)
+// CHECK2-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 false)
// 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:
@@ -20092,8 +20092,8 @@ int bar(int n){
// CHECK2-NEXT: store i32 [[TMP5]], i32* [[CONV3]], align 4
// CHECK2-NEXT: [[TMP6:%.*]] = load i64, i64* [[L_CASTED]], align 8
// CHECK2-NEXT: store i32 [[TMP2]], i32* [[DOTTHREADID_TEMP_]], align 4
-// CHECK2-NEXT: call void @__omp_outlined__(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i64 [[TMP4]], [1000 x i32]* [[TMP0]], i64 [[TMP6]]) #[[ATTR1:[0-9]+]]
-// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 false)
+// CHECK2-NEXT: call void @__omp_outlined__(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i64 [[TMP4]], [1000 x i32]* [[TMP0]], i64 [[TMP6]]) #[[ATTR2:[0-9]+]]
+// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 false)
// CHECK2-NEXT: ret void
// CHECK2: worker.exit:
// CHECK2-NEXT: ret void
@@ -20389,7 +20389,7 @@ int bar(int n){
//
//
// CHECK2-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l34
-// CHECK2-SAME: (i64 [[N:%.*]], [1000 x i16]* nonnull align 2 dereferenceable(2000) [[AA:%.*]]) #[[ATTR2:[0-9]+]] {
+// CHECK2-SAME: (i64 [[N:%.*]], [1000 x i16]* nonnull align 2 dereferenceable(2000) [[AA:%.*]]) #[[ATTR4:[0-9]+]] {
// CHECK2-NEXT: entry:
// CHECK2-NEXT: [[N_ADDR:%.*]] = alloca i64, align 8
// CHECK2-NEXT: [[AA_ADDR:%.*]] = alloca [1000 x i16]*, align 8
@@ -20401,7 +20401,7 @@ int bar(int n){
// 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]], i1 true, i1 false, i1 false)
+// CHECK2-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false)
// 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:
@@ -20411,8 +20411,8 @@ int bar(int n){
// CHECK2-NEXT: store i32 [[TMP3]], i32* [[CONV1]], align 4
// CHECK2-NEXT: [[TMP4:%.*]] = load i64, i64* [[N_CASTED]], align 8
// CHECK2-NEXT: store i32 [[TMP2]], i32* [[DOTTHREADID_TEMP_]], align 4
-// CHECK2-NEXT: call void @__omp_outlined__2(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i64 [[TMP4]], [1000 x i16]* [[TMP0]]) #[[ATTR1]]
-// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 false)
+// CHECK2-NEXT: call void @__omp_outlined__2(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i64 [[TMP4]], [1000 x i16]* [[TMP0]]) #[[ATTR2]]
+// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 false)
// CHECK2-NEXT: ret void
// CHECK2: worker.exit:
// CHECK2-NEXT: ret void
@@ -20656,14 +20656,14 @@ int bar(int n){
// 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]], i1 true, i1 false, i1 false)
+// CHECK2-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false)
// 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* @[[GLOB4]])
// CHECK2-NEXT: store i32 [[TMP2]], i32* [[DOTTHREADID_TEMP_]], align 4
-// CHECK2-NEXT: call void @__omp_outlined__4(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], [10 x i32]* [[TMP0]]) #[[ATTR1]]
-// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 false)
+// CHECK2-NEXT: call void @__omp_outlined__4(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], [10 x i32]* [[TMP0]]) #[[ATTR2]]
+// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 false)
// CHECK2-NEXT: ret void
// CHECK2: worker.exit:
// CHECK2-NEXT: ret void
@@ -20847,7 +20847,7 @@ int bar(int n){
// 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]], i1 true, i1 false, i1 false)
+// CHECK2-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false)
// 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:
@@ -20857,8 +20857,8 @@ int bar(int n){
// CHECK2-NEXT: store i32 [[TMP3]], i32* [[CONV1]], align 4
// CHECK2-NEXT: [[TMP4:%.*]] = load i64, i64* [[F_CASTED]], align 8
// CHECK2-NEXT: store i32 [[TMP2]], i32* [[DOTTHREADID_TEMP_]], align 4
-// CHECK2-NEXT: call void @__omp_outlined__6(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], [10 x [10 x i32]]* [[TMP0]], i64 [[TMP4]]) #[[ATTR1]]
-// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 false)
+// CHECK2-NEXT: call void @__omp_outlined__6(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], [10 x [10 x i32]]* [[TMP0]], i64 [[TMP4]]) #[[ATTR2]]
+// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 false)
// CHECK2-NEXT: ret void
// CHECK2: worker.exit:
// CHECK2-NEXT: ret void
@@ -21080,7 +21080,7 @@ int bar(int n){
// 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]], i1 true, i1 false, i1 false)
+// CHECK2-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false)
// 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:
@@ -21090,8 +21090,8 @@ int bar(int n){
// CHECK2-NEXT: store i32 [[TMP3]], i32* [[CONV1]], align 4
// CHECK2-NEXT: [[TMP4:%.*]] = load i64, i64* [[N_CASTED]], align 8
// CHECK2-NEXT: store i32 [[TMP2]], i32* [[DOTTHREADID_TEMP_]], align 4
-// CHECK2-NEXT: call void @__omp_outlined__8(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i64 [[TMP4]], [10 x [10 x i32]]* [[TMP0]]) #[[ATTR1]]
-// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 false)
+// CHECK2-NEXT: call void @__omp_outlined__8(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i64 [[TMP4]], [10 x [10 x i32]]* [[TMP0]]) #[[ATTR2]]
+// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 false)
// CHECK2-NEXT: ret void
// CHECK2: worker.exit:
// CHECK2-NEXT: ret void
@@ -21394,7 +21394,7 @@ int bar(int n){
// 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]], i1 true, i1 false, i1 false)
+// CHECK2-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false)
// 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:
@@ -21405,8 +21405,8 @@ int bar(int n){
// CHECK2-NEXT: [[TMP4:%.*]] = load i64, i64* [[N_CASTED]], align 8
// CHECK2-NEXT: [[TMP5:%.*]] = load i32*, i32** [[V_ADDR]], align 8
// CHECK2-NEXT: store i32 [[TMP2]], i32* [[DOTTHREADID_TEMP_]], align 4
-// CHECK2-NEXT: call void @__omp_outlined__10(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i64 [[TMP4]], [1000 x i32]* [[TMP0]], i32* [[TMP5]]) #[[ATTR1]]
-// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 false)
+// CHECK2-NEXT: call void @__omp_outlined__10(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i64 [[TMP4]], [1000 x i32]* [[TMP0]], i32* [[TMP5]]) #[[ATTR2]]
+// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 false)
// CHECK2-NEXT: ret void
// CHECK2: worker.exit:
// CHECK2-NEXT: ret void
@@ -21665,7 +21665,7 @@ int bar(int n){
// 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]+]], i1 true, i1 false, i1 false)
+// CHECK3-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 false)
// 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:
@@ -21677,8 +21677,8 @@ int bar(int n){
// CHECK3-NEXT: store i32 [[TMP5]], i32* [[L_CASTED]], align 4
// CHECK3-NEXT: [[TMP6:%.*]] = load i32, i32* [[L_CASTED]], align 4
// CHECK3-NEXT: store i32 [[TMP2]], i32* [[DOTTHREADID_TEMP_]], align 4
-// CHECK3-NEXT: call void @__omp_outlined__(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i32 [[TMP4]], [1000 x i32]* [[TMP0]], i32 [[TMP6]]) #[[ATTR1:[0-9]+]]
-// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 false)
+// CHECK3-NEXT: call void @__omp_outlined__(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i32 [[TMP4]], [1000 x i32]* [[TMP0]], i32 [[TMP6]]) #[[ATTR2:[0-9]+]]
+// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 false)
// CHECK3-NEXT: ret void
// CHECK3: worker.exit:
// CHECK3-NEXT: ret void
@@ -21961,7 +21961,7 @@ int bar(int n){
//
//
// CHECK3-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l34
-// CHECK3-SAME: (i32 [[N:%.*]], [1000 x i16]* nonnull align 2 dereferenceable(2000) [[AA:%.*]]) #[[ATTR2:[0-9]+]] {
+// CHECK3-SAME: (i32 [[N:%.*]], [1000 x i16]* nonnull align 2 dereferenceable(2000) [[AA:%.*]]) #[[ATTR4:[0-9]+]] {
// CHECK3-NEXT: entry:
// CHECK3-NEXT: [[N_ADDR:%.*]] = alloca i32, align 4
// CHECK3-NEXT: [[AA_ADDR:%.*]] = alloca [1000 x i16]*, align 4
@@ -21972,7 +21972,7 @@ int bar(int n){
// 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]], i1 true, i1 false, i1 false)
+// CHECK3-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false)
// 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:
@@ -21981,8 +21981,8 @@ int bar(int n){
// CHECK3-NEXT: store i32 [[TMP3]], i32* [[N_CASTED]], align 4
// CHECK3-NEXT: [[TMP4:%.*]] = load i32, i32* [[N_CASTED]], align 4
// CHECK3-NEXT: store i32 [[TMP2]], i32* [[DOTTHREADID_TEMP_]], align 4
-// CHECK3-NEXT: call void @__omp_outlined__2(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i32 [[TMP4]], [1000 x i16]* [[TMP0]]) #[[ATTR1]]
-// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 false)
+// CHECK3-NEXT: call void @__omp_outlined__2(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i32 [[TMP4]], [1000 x i16]* [[TMP0]]) #[[ATTR2]]
+// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 false)
// CHECK3-NEXT: ret void
// CHECK3: worker.exit:
// CHECK3-NEXT: ret void
@@ -22217,14 +22217,14 @@ int bar(int n){
// 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]], i1 true, i1 false, i1 false)
+// CHECK3-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false)
// 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* @[[GLOB4]])
// CHECK3-NEXT: store i32 [[TMP2]], i32* [[DOTTHREADID_TEMP_]], align 4
-// CHECK3-NEXT: call void @__omp_outlined__4(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], [10 x i32]* [[TMP0]]) #[[ATTR1]]
-// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 false)
+// CHECK3-NEXT: call void @__omp_outlined__4(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], [10 x i32]* [[TMP0]]) #[[ATTR2]]
+// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 false)
// CHECK3-NEXT: ret void
// CHECK3: worker.exit:
// CHECK3-NEXT: ret void
@@ -22401,7 +22401,7 @@ int bar(int n){
// 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]], i1 true, i1 false, i1 false)
+// CHECK3-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false)
// 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:
@@ -22410,8 +22410,8 @@ int bar(int n){
// CHECK3-NEXT: store i32 [[TMP3]], i32* [[F_CASTED]], align 4
// CHECK3-NEXT: [[TMP4:%.*]] = load i32, i32* [[F_CASTED]], align 4
// CHECK3-NEXT: store i32 [[TMP2]], i32* [[DOTTHREADID_TEMP_]], align 4
-// CHECK3-NEXT: call void @__omp_outlined__6(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], [10 x [10 x i32]]* [[TMP0]], i32 [[TMP4]]) #[[ATTR1]]
-// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 false)
+// CHECK3-NEXT: call void @__omp_outlined__6(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], [10 x [10 x i32]]* [[TMP0]], i32 [[TMP4]]) #[[ATTR2]]
+// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 false)
// CHECK3-NEXT: ret void
// CHECK3: worker.exit:
// CHECK3-NEXT: ret void
@@ -22622,7 +22622,7 @@ int bar(int n){
// 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]], i1 true, i1 false, i1 false)
+// CHECK3-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false)
// 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:
@@ -22631,8 +22631,8 @@ int bar(int n){
// CHECK3-NEXT: store i32 [[TMP3]], i32* [[N_CASTED]], align 4
// CHECK3-NEXT: [[TMP4:%.*]] = load i32, i32* [[N_CASTED]], align 4
// CHECK3-NEXT: store i32 [[TMP2]], i32* [[DOTTHREADID_TEMP_]], align 4
-// CHECK3-NEXT: call void @__omp_outlined__8(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i32 [[TMP4]], [10 x [10 x i32]]* [[TMP0]]) #[[ATTR1]]
-// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 false)
+// CHECK3-NEXT: call void @__omp_outlined__8(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i32 [[TMP4]], [10 x [10 x i32]]* [[TMP0]]) #[[ATTR2]]
+// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 false)
// CHECK3-NEXT: ret void
// CHECK3: worker.exit:
// CHECK3-NEXT: ret void
@@ -22939,7 +22939,7 @@ int bar(int n){
// 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]], i1 true, i1 false, i1 false)
+// CHECK3-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false)
// 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:
@@ -22949,8 +22949,8 @@ int bar(int n){
// CHECK3-NEXT: [[TMP4:%.*]] = load i32, i32* [[N_CASTED]], align 4
// CHECK3-NEXT: [[TMP5:%.*]] = load i32*, i32** [[V_ADDR]], align 4
// CHECK3-NEXT: store i32 [[TMP2]], i32* [[DOTTHREADID_TEMP_]], align 4
-// CHECK3-NEXT: call void @__omp_outlined__10(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i32 [[TMP4]], [1000 x i32]* [[TMP0]], i32* [[TMP5]]) #[[ATTR1]]
-// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 false)
+// CHECK3-NEXT: call void @__omp_outlined__10(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i32 [[TMP4]], [1000 x i32]* [[TMP0]], i32* [[TMP5]]) #[[ATTR2]]
+// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 false)
// CHECK3-NEXT: ret void
// CHECK3: worker.exit:
// CHECK3-NEXT: ret void
@@ -23199,7 +23199,7 @@ int bar(int n){
// 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]+]], i1 true, i1 false, i1 false)
+// CHECK4-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 false)
// 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:
@@ -23211,8 +23211,8 @@ int bar(int n){
// CHECK4-NEXT: store i32 [[TMP5]], i32* [[L_CASTED]], align 4
// CHECK4-NEXT: [[TMP6:%.*]] = load i32, i32* [[L_CASTED]], align 4
// CHECK4-NEXT: store i32 [[TMP2]], i32* [[DOTTHREADID_TEMP_]], align 4
-// CHECK4-NEXT: call void @__omp_outlined__(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i32 [[TMP4]], [1000 x i32]* [[TMP0]], i32 [[TMP6]]) #[[ATTR1:[0-9]+]]
-// CHECK4-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 false)
+// CHECK4-NEXT: call void @__omp_outlined__(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i32 [[TMP4]], [1000 x i32]* [[TMP0]], i32 [[TMP6]]) #[[ATTR2:[0-9]+]]
+// CHECK4-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 false)
// CHECK4-NEXT: ret void
// CHECK4: worker.exit:
// CHECK4-NEXT: ret void
@@ -23495,7 +23495,7 @@ int bar(int n){
//
//
// CHECK4-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l34
-// CHECK4-SAME: (i32 [[N:%.*]], [1000 x i16]* nonnull align 2 dereferenceable(2000) [[AA:%.*]]) #[[ATTR2:[0-9]+]] {
+// CHECK4-SAME: (i32 [[N:%.*]], [1000 x i16]* nonnull align 2 dereferenceable(2000) [[AA:%.*]]) #[[ATTR4:[0-9]+]] {
// CHECK4-NEXT: entry:
// CHECK4-NEXT: [[N_ADDR:%.*]] = alloca i32, align 4
// CHECK4-NEXT: [[AA_ADDR:%.*]] = alloca [1000 x i16]*, align 4
@@ -23506,7 +23506,7 @@ int bar(int n){
// 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]], i1 true, i1 false, i1 false)
+// CHECK4-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false)
// 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:
@@ -23515,8 +23515,8 @@ int bar(int n){
// CHECK4-NEXT: store i32 [[TMP3]], i32* [[N_CASTED]], align 4
// CHECK4-NEXT: [[TMP4:%.*]] = load i32, i32* [[N_CASTED]], align 4
// CHECK4-NEXT: store i32 [[TMP2]], i32* [[DOTTHREADID_TEMP_]], align 4
-// CHECK4-NEXT: call void @__omp_outlined__2(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i32 [[TMP4]], [1000 x i16]* [[TMP0]]) #[[ATTR1]]
-// CHECK4-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 false)
+// CHECK4-NEXT: call void @__omp_outlined__2(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i32 [[TMP4]], [1000 x i16]* [[TMP0]]) #[[ATTR2]]
+// CHECK4-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 false)
// CHECK4-NEXT: ret void
// CHECK4: worker.exit:
// CHECK4-NEXT: ret void
@@ -23751,14 +23751,14 @@ int bar(int n){
// 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]], i1 true, i1 false, i1 false)
+// CHECK4-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false)
// 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:
// CHECK4-NEXT: [[TMP2:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB4]])
// CHECK4-NEXT: store i32 [[TMP2]], i32* [[DOTTHREADID_TEMP_]], align 4
-// CHECK4-NEXT: call void @__omp_outlined__4(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], [10 x i32]* [[TMP0]]) #[[ATTR1]]
-// CHECK4-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 false)
+// CHECK4-NEXT: call void @__omp_outlined__4(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], [10 x i32]* [[TMP0]]) #[[ATTR2]]
+// CHECK4-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 false)
// CHECK4-NEXT: ret void
// CHECK4: worker.exit:
// CHECK4-NEXT: ret void
@@ -23935,7 +23935,7 @@ int bar(int n){
// 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]], i1 true, i1 false, i1 false)
+// CHECK4-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false)
// 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:
@@ -23944,8 +23944,8 @@ int bar(int n){
// CHECK4-NEXT: store i32 [[TMP3]], i32* [[F_CASTED]], align 4
// CHECK4-NEXT: [[TMP4:%.*]] = load i32, i32* [[F_CASTED]], align 4
// CHECK4-NEXT: store i32 [[TMP2]], i32* [[DOTTHREADID_TEMP_]], align 4
-// CHECK4-NEXT: call void @__omp_outlined__6(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], [10 x [10 x i32]]* [[TMP0]], i32 [[TMP4]]) #[[ATTR1]]
-// CHECK4-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 false)
+// CHECK4-NEXT: call void @__omp_outlined__6(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], [10 x [10 x i32]]* [[TMP0]], i32 [[TMP4]]) #[[ATTR2]]
+// CHECK4-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 false)
// CHECK4-NEXT: ret void
// CHECK4: worker.exit:
// CHECK4-NEXT: ret void
@@ -24156,7 +24156,7 @@ int bar(int n){
// 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]], i1 true, i1 false, i1 false)
+// CHECK4-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false)
// 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:
@@ -24165,8 +24165,8 @@ int bar(int n){
// CHECK4-NEXT: store i32 [[TMP3]], i32* [[N_CASTED]], align 4
// CHECK4-NEXT: [[TMP4:%.*]] = load i32, i32* [[N_CASTED]], align 4
// CHECK4-NEXT: store i32 [[TMP2]], i32* [[DOTTHREADID_TEMP_]], align 4
-// CHECK4-NEXT: call void @__omp_outlined__8(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i32 [[TMP4]], [10 x [10 x i32]]* [[TMP0]]) #[[ATTR1]]
-// CHECK4-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 false)
+// CHECK4-NEXT: call void @__omp_outlined__8(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i32 [[TMP4]], [10 x [10 x i32]]* [[TMP0]]) #[[ATTR2]]
+// CHECK4-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 false)
// CHECK4-NEXT: ret void
// CHECK4: worker.exit:
// CHECK4-NEXT: ret void
@@ -24473,7 +24473,7 @@ int bar(int n){
// 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]], i1 true, i1 false, i1 false)
+// CHECK4-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false)
// 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:
@@ -24483,8 +24483,8 @@ int bar(int n){
// CHECK4-NEXT: [[TMP4:%.*]] = load i32, i32* [[N_CASTED]], align 4
// CHECK4-NEXT: [[TMP5:%.*]] = load i32*, i32** [[V_ADDR]], align 4
// CHECK4-NEXT: store i32 [[TMP2]], i32* [[DOTTHREADID_TEMP_]], align 4
-// CHECK4-NEXT: call void @__omp_outlined__10(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i32 [[TMP4]], [1000 x i32]* [[TMP0]], i32* [[TMP5]]) #[[ATTR1]]
-// CHECK4-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 false)
+// CHECK4-NEXT: call void @__omp_outlined__10(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i32 [[TMP4]], [1000 x i32]* [[TMP0]], i32* [[TMP5]]) #[[ATTR2]]
+// CHECK4-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 false)
// CHECK4-NEXT: ret void
// CHECK4: worker.exit:
// CHECK4-NEXT: ret void
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
index 1a699d1cccb3f..bfd8b94b9e1d6 100644
--- 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 @@ int main(int argc, char **argv) {
// 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]+]], i1 true, i1 false, i1 false)
+// CHECK1-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 false)
// 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:
@@ -61,7 +61,7 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: [[TMP6:%.*]] = load i64, i64* [[DOTCAPTURE_EXPR__CASTED]], align 8
// CHECK1-NEXT: store i32 [[TMP2]], i32* [[DOTTHREADID_TEMP_]], align 4
// CHECK1-NEXT: call void @__omp_outlined__(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i64 [[TMP4]], i32* [[TMP0]], i64 [[TMP6]]) #[[ATTR3:[0-9]+]]
-// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 false)
+// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 false)
// CHECK1-NEXT: ret void
// CHECK1: worker.exit:
// CHECK1-NEXT: ret void
@@ -304,10 +304,10 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: [[MUL:%.*]] = mul nsw i32 [[TMP19]], 1
// CHECK1-NEXT: [[ADD:%.*]] = add nsw i32 0, [[MUL]]
// CHECK1-NEXT: store i32 [[ADD]], i32* [[I7]], align 4
-// CHECK1-NEXT: [[CALL:%.*]] = call i32 @_Z3fooPi(i32* [[I7]]) #[[ATTR4:[0-9]+]]
-// CHECK1-NEXT: [[CALL13:%.*]] = call i32 @_Z3fooPi(i32* [[TMP0]]) #[[ATTR4]]
+// CHECK1-NEXT: [[CALL:%.*]] = call i32 @_Z3fooPi(i32* [[I7]]) #[[ATTR5:[0-9]+]]
+// CHECK1-NEXT: [[CALL13:%.*]] = call i32 @_Z3fooPi(i32* [[TMP0]]) #[[ATTR5]]
// CHECK1-NEXT: [[ADD14:%.*]] = add nsw i32 [[CALL]], [[CALL13]]
-// CHECK1-NEXT: [[CALL15:%.*]] = call i32 @_Z3fooPi(i32* [[CONV]]) #[[ATTR4]]
+// CHECK1-NEXT: [[CALL15:%.*]] = call i32 @_Z3fooPi(i32* [[CONV]]) #[[ATTR5]]
// CHECK1-NEXT: [[ADD16:%.*]] = add nsw i32 [[ADD14]], [[CALL15]]
// CHECK1-NEXT: store i32 [[ADD16]], i32* [[TMP0]], align 4
// CHECK1-NEXT: br label [[OMP_BODY_CONTINUE:%.*]]
@@ -354,7 +354,7 @@ int main(int argc, char **argv) {
// 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]+]], i1 true, i1 false, i1 false)
+// CHECK2-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 false)
// 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:
@@ -367,7 +367,7 @@ int main(int argc, char **argv) {
// CHECK2-NEXT: [[TMP6:%.*]] = load i32, i32* [[DOTCAPTURE_EXPR__CASTED]], align 4
// CHECK2-NEXT: store i32 [[TMP2]], i32* [[DOTTHREADID_TEMP_]], align 4
// CHECK2-NEXT: call void @__omp_outlined__(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i32 [[TMP4]], i32* [[TMP0]], i32 [[TMP6]]) #[[ATTR3:[0-9]+]]
-// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 false)
+// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 false)
// CHECK2-NEXT: ret void
// CHECK2: worker.exit:
// CHECK2-NEXT: ret void
@@ -598,10 +598,10 @@ int main(int argc, char **argv) {
// CHECK2-NEXT: [[MUL:%.*]] = mul nsw i32 [[TMP19]], 1
// CHECK2-NEXT: [[ADD:%.*]] = add nsw i32 0, [[MUL]]
// CHECK2-NEXT: store i32 [[ADD]], i32* [[I4]], align 4
-// CHECK2-NEXT: [[CALL:%.*]] = call i32 @_Z3fooPi(i32* [[I4]]) #[[ATTR4:[0-9]+]]
-// CHECK2-NEXT: [[CALL8:%.*]] = call i32 @_Z3fooPi(i32* [[TMP0]]) #[[ATTR4]]
+// CHECK2-NEXT: [[CALL:%.*]] = call i32 @_Z3fooPi(i32* [[I4]]) #[[ATTR5:[0-9]+]]
+// CHECK2-NEXT: [[CALL8:%.*]] = call i32 @_Z3fooPi(i32* [[TMP0]]) #[[ATTR5]]
// CHECK2-NEXT: [[ADD9:%.*]] = add nsw i32 [[CALL]], [[CALL8]]
-// CHECK2-NEXT: [[CALL10:%.*]] = call i32 @_Z3fooPi(i32* [[ARGC_ADDR]]) #[[ATTR4]]
+// CHECK2-NEXT: [[CALL10:%.*]] = call i32 @_Z3fooPi(i32* [[ARGC_ADDR]]) #[[ATTR5]]
// CHECK2-NEXT: [[ADD11:%.*]] = add nsw i32 [[ADD9]], [[CALL10]]
// CHECK2-NEXT: store i32 [[ADD11]], i32* [[TMP0]], align 4
// CHECK2-NEXT: br label [[OMP_BODY_CONTINUE:%.*]]
@@ -648,7 +648,7 @@ int main(int argc, char **argv) {
// 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]+]], i1 true, i1 false, i1 false)
+// CHECK3-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 false)
// 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:
@@ -661,7 +661,7 @@ int main(int argc, char **argv) {
// CHECK3-NEXT: [[TMP6:%.*]] = load i32, i32* [[DOTCAPTURE_EXPR__CASTED]], align 4
// CHECK3-NEXT: store i32 [[TMP2]], i32* [[DOTTHREADID_TEMP_]], align 4
// CHECK3-NEXT: call void @__omp_outlined__(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i32 [[TMP4]], i32* [[TMP0]], i32 [[TMP6]]) #[[ATTR3:[0-9]+]]
-// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 false)
+// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 false)
// CHECK3-NEXT: ret void
// CHECK3: worker.exit:
// CHECK3-NEXT: ret void
@@ -892,10 +892,10 @@ int main(int argc, char **argv) {
// CHECK3-NEXT: [[MUL:%.*]] = mul nsw i32 [[TMP19]], 1
// CHECK3-NEXT: [[ADD:%.*]] = add nsw i32 0, [[MUL]]
// CHECK3-NEXT: store i32 [[ADD]], i32* [[I4]], align 4
-// CHECK3-NEXT: [[CALL:%.*]] = call i32 @_Z3fooPi(i32* [[I4]]) #[[ATTR4:[0-9]+]]
-// CHECK3-NEXT: [[CALL8:%.*]] = call i32 @_Z3fooPi(i32* [[TMP0]]) #[[ATTR4]]
+// CHECK3-NEXT: [[CALL:%.*]] = call i32 @_Z3fooPi(i32* [[I4]]) #[[ATTR5:[0-9]+]]
+// CHECK3-NEXT: [[CALL8:%.*]] = call i32 @_Z3fooPi(i32* [[TMP0]]) #[[ATTR5]]
// CHECK3-NEXT: [[ADD9:%.*]] = add nsw i32 [[CALL]], [[CALL8]]
-// CHECK3-NEXT: [[CALL10:%.*]] = call i32 @_Z3fooPi(i32* [[ARGC_ADDR]]) #[[ATTR4]]
+// CHECK3-NEXT: [[CALL10:%.*]] = call i32 @_Z3fooPi(i32* [[ARGC_ADDR]]) #[[ATTR5]]
// CHECK3-NEXT: [[ADD11:%.*]] = add nsw i32 [[ADD9]], [[CALL10]]
// CHECK3-NEXT: store i32 [[ADD11]], i32* [[TMP0]], align 4
// CHECK3-NEXT: br label [[OMP_BODY_CONTINUE:%.*]]
@@ -944,7 +944,7 @@ int main(int argc, char **argv) {
// 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]+]], i1 true, i1 false, i1 false)
+// CHECK4-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 false)
// 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:
@@ -959,7 +959,7 @@ int main(int argc, char **argv) {
// CHECK4-NEXT: [[TMP6:%.*]] = load i64, i64* [[DOTCAPTURE_EXPR__CASTED]], align 8
// CHECK4-NEXT: store i32 [[TMP2]], i32* [[DOTTHREADID_TEMP_]], align 4
// CHECK4-NEXT: call void @__omp_outlined__(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i64 [[TMP4]], i32* [[TMP0]], i64 [[TMP6]]) #[[ATTR3:[0-9]+]]
-// CHECK4-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 false)
+// CHECK4-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 false)
// CHECK4-NEXT: ret void
// CHECK4: worker.exit:
// CHECK4-NEXT: ret void
@@ -1202,10 +1202,10 @@ int main(int argc, char **argv) {
// CHECK4-NEXT: [[MUL:%.*]] = mul nsw i32 [[TMP19]], 1
// CHECK4-NEXT: [[ADD:%.*]] = add nsw i32 0, [[MUL]]
// CHECK4-NEXT: store i32 [[ADD]], i32* [[I7]], align 4
-// CHECK4-NEXT: [[CALL:%.*]] = call i32 @_Z3fooPi(i32* [[I7]]) #[[ATTR4:[0-9]+]]
-// CHECK4-NEXT: [[CALL13:%.*]] = call i32 @_Z3fooPi(i32* [[TMP0]]) #[[ATTR4]]
+// CHECK4-NEXT: [[CALL:%.*]] = call i32 @_Z3fooPi(i32* [[I7]]) #[[ATTR5:[0-9]+]]
+// CHECK4-NEXT: [[CALL13:%.*]] = call i32 @_Z3fooPi(i32* [[TMP0]]) #[[ATTR5]]
// CHECK4-NEXT: [[ADD14:%.*]] = add nsw i32 [[CALL]], [[CALL13]]
-// CHECK4-NEXT: [[CALL15:%.*]] = call i32 @_Z3fooPi(i32* [[CONV]]) #[[ATTR4]]
+// CHECK4-NEXT: [[CALL15:%.*]] = call i32 @_Z3fooPi(i32* [[CONV]]) #[[ATTR5]]
// CHECK4-NEXT: [[ADD16:%.*]] = add nsw i32 [[ADD14]], [[CALL15]]
// CHECK4-NEXT: store i32 [[ADD16]], i32* [[TMP0]], align 4
// CHECK4-NEXT: br label [[OMP_BODY_CONTINUE:%.*]]
@@ -1252,7 +1252,7 @@ int main(int argc, char **argv) {
// 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]+]], i1 true, i1 false, i1 false)
+// CHECK5-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 false)
// 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:
@@ -1265,7 +1265,7 @@ int main(int argc, char **argv) {
// CHECK5-NEXT: [[TMP6:%.*]] = load i32, i32* [[DOTCAPTURE_EXPR__CASTED]], align 4
// CHECK5-NEXT: store i32 [[TMP2]], i32* [[DOTTHREADID_TEMP_]], align 4
// CHECK5-NEXT: call void @__omp_outlined__(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i32 [[TMP4]], i32* [[TMP0]], i32 [[TMP6]]) #[[ATTR3:[0-9]+]]
-// CHECK5-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 false)
+// CHECK5-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 false)
// CHECK5-NEXT: ret void
// CHECK5: worker.exit:
// CHECK5-NEXT: ret void
@@ -1496,10 +1496,10 @@ int main(int argc, char **argv) {
// CHECK5-NEXT: [[MUL:%.*]] = mul nsw i32 [[TMP19]], 1
// CHECK5-NEXT: [[ADD:%.*]] = add nsw i32 0, [[MUL]]
// CHECK5-NEXT: store i32 [[ADD]], i32* [[I4]], align 4
-// CHECK5-NEXT: [[CALL:%.*]] = call i32 @_Z3fooPi(i32* [[I4]]) #[[ATTR4:[0-9]+]]
-// CHECK5-NEXT: [[CALL8:%.*]] = call i32 @_Z3fooPi(i32* [[TMP0]]) #[[ATTR4]]
+// CHECK5-NEXT: [[CALL:%.*]] = call i32 @_Z3fooPi(i32* [[I4]]) #[[ATTR5:[0-9]+]]
+// CHECK5-NEXT: [[CALL8:%.*]] = call i32 @_Z3fooPi(i32* [[TMP0]]) #[[ATTR5]]
// CHECK5-NEXT: [[ADD9:%.*]] = add nsw i32 [[CALL]], [[CALL8]]
-// CHECK5-NEXT: [[CALL10:%.*]] = call i32 @_Z3fooPi(i32* [[ARGC_ADDR]]) #[[ATTR4]]
+// CHECK5-NEXT: [[CALL10:%.*]] = call i32 @_Z3fooPi(i32* [[ARGC_ADDR]]) #[[ATTR5]]
// CHECK5-NEXT: [[ADD11:%.*]] = add nsw i32 [[ADD9]], [[CALL10]]
// CHECK5-NEXT: store i32 [[ADD11]], i32* [[TMP0]], align 4
// CHECK5-NEXT: br label [[OMP_BODY_CONTINUE:%.*]]
@@ -1546,7 +1546,7 @@ int main(int argc, char **argv) {
// 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]+]], i1 true, i1 false, i1 false)
+// CHECK6-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 false)
// 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:
@@ -1559,7 +1559,7 @@ int main(int argc, char **argv) {
// CHECK6-NEXT: [[TMP6:%.*]] = load i32, i32* [[DOTCAPTURE_EXPR__CASTED]], align 4
// CHECK6-NEXT: store i32 [[TMP2]], i32* [[DOTTHREADID_TEMP_]], align 4
// CHECK6-NEXT: call void @__omp_outlined__(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i32 [[TMP4]], i32* [[TMP0]], i32 [[TMP6]]) #[[ATTR3:[0-9]+]]
-// CHECK6-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 false)
+// CHECK6-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 false)
// CHECK6-NEXT: ret void
// CHECK6: worker.exit:
// CHECK6-NEXT: ret void
@@ -1790,10 +1790,10 @@ int main(int argc, char **argv) {
// CHECK6-NEXT: [[MUL:%.*]] = mul nsw i32 [[TMP19]], 1
// CHECK6-NEXT: [[ADD:%.*]] = add nsw i32 0, [[MUL]]
// CHECK6-NEXT: store i32 [[ADD]], i32* [[I4]], align 4
-// CHECK6-NEXT: [[CALL:%.*]] = call i32 @_Z3fooPi(i32* [[I4]]) #[[ATTR4:[0-9]+]]
-// CHECK6-NEXT: [[CALL8:%.*]] = call i32 @_Z3fooPi(i32* [[TMP0]]) #[[ATTR4]]
+// CHECK6-NEXT: [[CALL:%.*]] = call i32 @_Z3fooPi(i32* [[I4]]) #[[ATTR5:[0-9]+]]
+// CHECK6-NEXT: [[CALL8:%.*]] = call i32 @_Z3fooPi(i32* [[TMP0]]) #[[ATTR5]]
// CHECK6-NEXT: [[ADD9:%.*]] = add nsw i32 [[CALL]], [[CALL8]]
-// CHECK6-NEXT: [[CALL10:%.*]] = call i32 @_Z3fooPi(i32* [[ARGC_ADDR]]) #[[ATTR4]]
+// CHECK6-NEXT: [[CALL10:%.*]] = call i32 @_Z3fooPi(i32* [[ARGC_ADDR]]) #[[ATTR5]]
// CHECK6-NEXT: [[ADD11:%.*]] = add nsw i32 [[ADD9]], [[CALL10]]
// CHECK6-NEXT: store i32 [[ADD11]], i32* [[TMP0]], align 4
// CHECK6-NEXT: br label [[OMP_BODY_CONTINUE:%.*]]
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
index 4ecb670f46fe5..1a7369eea94cd 100644
--- 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 @@ int bar(int n){
// 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]+]], i1 true, i1 false, i1 false)
+// CHECK1-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 false)
// 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:
@@ -9411,8 +9411,8 @@ int bar(int n){
// CHECK1-NEXT: store i32 [[TMP5]], i32* [[CONV3]], align 4
// CHECK1-NEXT: [[TMP6:%.*]] = load i64, i64* [[L_CASTED]], align 8
// CHECK1-NEXT: store i32 [[TMP2]], i32* [[DOTTHREADID_TEMP_]], align 4
-// CHECK1-NEXT: call void @__omp_outlined__(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i64 [[TMP4]], [1000 x i32]* [[TMP0]], i64 [[TMP6]]) #[[ATTR1:[0-9]+]]
-// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 false)
+// CHECK1-NEXT: call void @__omp_outlined__(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i64 [[TMP4]], [1000 x i32]* [[TMP0]], i64 [[TMP6]]) #[[ATTR2:[0-9]+]]
+// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 false)
// CHECK1-NEXT: ret void
// CHECK1: worker.exit:
// CHECK1-NEXT: ret void
@@ -9732,7 +9732,7 @@ int bar(int n){
//
//
// CHECK1-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l32
-// CHECK1-SAME: (i64 [[N:%.*]], [1000 x i16]* nonnull align 2 dereferenceable(2000) [[AA:%.*]]) #[[ATTR2:[0-9]+]] {
+// CHECK1-SAME: (i64 [[N:%.*]], [1000 x i16]* nonnull align 2 dereferenceable(2000) [[AA:%.*]]) #[[ATTR4:[0-9]+]] {
// CHECK1-NEXT: entry:
// CHECK1-NEXT: [[N_ADDR:%.*]] = alloca i64, align 8
// CHECK1-NEXT: [[AA_ADDR:%.*]] = alloca [1000 x i16]*, align 8
@@ -9744,7 +9744,7 @@ int bar(int n){
// 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]], i1 true, i1 false, i1 false)
+// CHECK1-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false)
// 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:
@@ -9754,8 +9754,8 @@ int bar(int n){
// CHECK1-NEXT: store i32 [[TMP3]], i32* [[CONV1]], align 4
// CHECK1-NEXT: [[TMP4:%.*]] = load i64, i64* [[N_CASTED]], align 8
// CHECK1-NEXT: store i32 [[TMP2]], i32* [[DOTTHREADID_TEMP_]], align 4
-// CHECK1-NEXT: call void @__omp_outlined__2(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i64 [[TMP4]], [1000 x i16]* [[TMP0]]) #[[ATTR1]]
-// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 false)
+// CHECK1-NEXT: call void @__omp_outlined__2(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i64 [[TMP4]], [1000 x i16]* [[TMP0]]) #[[ATTR2]]
+// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 false)
// CHECK1-NEXT: ret void
// CHECK1: worker.exit:
// CHECK1-NEXT: ret void
@@ -10023,14 +10023,14 @@ int bar(int n){
// 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]], i1 true, i1 false, i1 false)
+// CHECK1-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false)
// 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* @[[GLOB4]])
// CHECK1-NEXT: store i32 [[TMP2]], i32* [[DOTTHREADID_TEMP_]], align 4
-// CHECK1-NEXT: call void @__omp_outlined__4(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], [10 x i32]* [[TMP0]]) #[[ATTR1]]
-// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 false)
+// CHECK1-NEXT: call void @__omp_outlined__4(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], [10 x i32]* [[TMP0]]) #[[ATTR2]]
+// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 false)
// CHECK1-NEXT: ret void
// CHECK1: worker.exit:
// CHECK1-NEXT: ret void
@@ -10228,7 +10228,7 @@ int bar(int n){
// 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]], i1 true, i1 false, i1 false)
+// CHECK1-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false)
// 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:
@@ -10238,8 +10238,8 @@ int bar(int n){
// CHECK1-NEXT: store i32 [[TMP3]], i32* [[CONV1]], align 4
// CHECK1-NEXT: [[TMP4:%.*]] = load i64, i64* [[F_CASTED]], align 8
// CHECK1-NEXT: store i32 [[TMP2]], i32* [[DOTTHREADID_TEMP_]], align 4
-// CHECK1-NEXT: call void @__omp_outlined__6(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], [10 x [10 x i32]]* [[TMP0]], i64 [[TMP4]]) #[[ATTR1]]
-// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 false)
+// CHECK1-NEXT: call void @__omp_outlined__6(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], [10 x [10 x i32]]* [[TMP0]], i64 [[TMP4]]) #[[ATTR2]]
+// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 false)
// CHECK1-NEXT: ret void
// CHECK1: worker.exit:
// CHECK1-NEXT: ret void
@@ -10479,7 +10479,7 @@ int bar(int n){
// 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]+]], i1 true, i1 false, i1 false)
+// CHECK2-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 false)
// 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:
@@ -10491,8 +10491,8 @@ int bar(int n){
// CHECK2-NEXT: store i32 [[TMP5]], i32* [[L_CASTED]], align 4
// CHECK2-NEXT: [[TMP6:%.*]] = load i32, i32* [[L_CASTED]], align 4
// CHECK2-NEXT: store i32 [[TMP2]], i32* [[DOTTHREADID_TEMP_]], align 4
-// CHECK2-NEXT: call void @__omp_outlined__(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i32 [[TMP4]], [1000 x i32]* [[TMP0]], i32 [[TMP6]]) #[[ATTR1:[0-9]+]]
-// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 false)
+// CHECK2-NEXT: call void @__omp_outlined__(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i32 [[TMP4]], [1000 x i32]* [[TMP0]], i32 [[TMP6]]) #[[ATTR2:[0-9]+]]
+// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 false)
// CHECK2-NEXT: ret void
// CHECK2: worker.exit:
// CHECK2-NEXT: ret void
@@ -10799,7 +10799,7 @@ int bar(int n){
//
//
// CHECK2-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l32
-// CHECK2-SAME: (i32 [[N:%.*]], [1000 x i16]* nonnull align 2 dereferenceable(2000) [[AA:%.*]]) #[[ATTR2:[0-9]+]] {
+// CHECK2-SAME: (i32 [[N:%.*]], [1000 x i16]* nonnull align 2 dereferenceable(2000) [[AA:%.*]]) #[[ATTR4:[0-9]+]] {
// CHECK2-NEXT: entry:
// CHECK2-NEXT: [[N_ADDR:%.*]] = alloca i32, align 4
// CHECK2-NEXT: [[AA_ADDR:%.*]] = alloca [1000 x i16]*, align 4
@@ -10810,7 +10810,7 @@ int bar(int n){
// 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]], i1 true, i1 false, i1 false)
+// CHECK2-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false)
// 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:
@@ -10819,8 +10819,8 @@ int bar(int n){
// CHECK2-NEXT: store i32 [[TMP3]], i32* [[N_CASTED]], align 4
// CHECK2-NEXT: [[TMP4:%.*]] = load i32, i32* [[N_CASTED]], align 4
// CHECK2-NEXT: store i32 [[TMP2]], i32* [[DOTTHREADID_TEMP_]], align 4
-// CHECK2-NEXT: call void @__omp_outlined__2(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i32 [[TMP4]], [1000 x i16]* [[TMP0]]) #[[ATTR1]]
-// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 false)
+// CHECK2-NEXT: call void @__omp_outlined__2(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i32 [[TMP4]], [1000 x i16]* [[TMP0]]) #[[ATTR2]]
+// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 false)
// CHECK2-NEXT: ret void
// CHECK2: worker.exit:
// CHECK2-NEXT: ret void
@@ -11079,14 +11079,14 @@ int bar(int n){
// 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]], i1 true, i1 false, i1 false)
+// CHECK2-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false)
// 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* @[[GLOB4]])
// CHECK2-NEXT: store i32 [[TMP2]], i32* [[DOTTHREADID_TEMP_]], align 4
-// CHECK2-NEXT: call void @__omp_outlined__4(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], [10 x i32]* [[TMP0]]) #[[ATTR1]]
-// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 false)
+// CHECK2-NEXT: call void @__omp_outlined__4(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], [10 x i32]* [[TMP0]]) #[[ATTR2]]
+// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 false)
// CHECK2-NEXT: ret void
// CHECK2: worker.exit:
// CHECK2-NEXT: ret void
@@ -11277,7 +11277,7 @@ int bar(int n){
// 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]], i1 true, i1 false, i1 false)
+// CHECK2-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false)
// 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:
@@ -11286,8 +11286,8 @@ int bar(int n){
// CHECK2-NEXT: store i32 [[TMP3]], i32* [[F_CASTED]], align 4
// CHECK2-NEXT: [[TMP4:%.*]] = load i32, i32* [[F_CASTED]], align 4
// CHECK2-NEXT: store i32 [[TMP2]], i32* [[DOTTHREADID_TEMP_]], align 4
-// CHECK2-NEXT: call void @__omp_outlined__6(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], [10 x [10 x i32]]* [[TMP0]], i32 [[TMP4]]) #[[ATTR1]]
-// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 false)
+// CHECK2-NEXT: call void @__omp_outlined__6(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], [10 x [10 x i32]]* [[TMP0]], i32 [[TMP4]]) #[[ATTR2]]
+// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 false)
// CHECK2-NEXT: ret void
// CHECK2: worker.exit:
// CHECK2-NEXT: ret void
@@ -11517,7 +11517,7 @@ int bar(int n){
// 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]+]], i1 true, i1 false, i1 false)
+// CHECK3-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 false)
// 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:
@@ -11529,8 +11529,8 @@ int bar(int n){
// CHECK3-NEXT: store i32 [[TMP5]], i32* [[L_CASTED]], align 4
// CHECK3-NEXT: [[TMP6:%.*]] = load i32, i32* [[L_CASTED]], align 4
// CHECK3-NEXT: store i32 [[TMP2]], i32* [[DOTTHREADID_TEMP_]], align 4
-// CHECK3-NEXT: call void @__omp_outlined__(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i32 [[TMP4]], [1000 x i32]* [[TMP0]], i32 [[TMP6]]) #[[ATTR1:[0-9]+]]
-// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 false)
+// CHECK3-NEXT: call void @__omp_outlined__(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i32 [[TMP4]], [1000 x i32]* [[TMP0]], i32 [[TMP6]]) #[[ATTR2:[0-9]+]]
+// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 false)
// CHECK3-NEXT: ret void
// CHECK3: worker.exit:
// CHECK3-NEXT: ret void
@@ -11837,7 +11837,7 @@ int bar(int n){
//
//
// CHECK3-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z9ftemplateIiET_i_l32
-// CHECK3-SAME: (i32 [[N:%.*]], [1000 x i16]* nonnull align 2 dereferenceable(2000) [[AA:%.*]]) #[[ATTR2:[0-9]+]] {
+// CHECK3-SAME: (i32 [[N:%.*]], [1000 x i16]* nonnull align 2 dereferenceable(2000) [[AA:%.*]]) #[[ATTR4:[0-9]+]] {
// CHECK3-NEXT: entry:
// CHECK3-NEXT: [[N_ADDR:%.*]] = alloca i32, align 4
// CHECK3-NEXT: [[AA_ADDR:%.*]] = alloca [1000 x i16]*, align 4
@@ -11848,7 +11848,7 @@ int bar(int n){
// 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]], i1 true, i1 false, i1 false)
+// CHECK3-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false)
// 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:
@@ -11857,8 +11857,8 @@ int bar(int n){
// CHECK3-NEXT: store i32 [[TMP3]], i32* [[N_CASTED]], align 4
// CHECK3-NEXT: [[TMP4:%.*]] = load i32, i32* [[N_CASTED]], align 4
// CHECK3-NEXT: store i32 [[TMP2]], i32* [[DOTTHREADID_TEMP_]], align 4
-// CHECK3-NEXT: call void @__omp_outlined__2(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i32 [[TMP4]], [1000 x i16]* [[TMP0]]) #[[ATTR1]]
-// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 false)
+// CHECK3-NEXT: call void @__omp_outlined__2(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i32 [[TMP4]], [1000 x i16]* [[TMP0]]) #[[ATTR2]]
+// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 false)
// CHECK3-NEXT: ret void
// CHECK3: worker.exit:
// CHECK3-NEXT: ret void
@@ -12117,14 +12117,14 @@ int bar(int n){
// 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]], i1 true, i1 false, i1 false)
+// CHECK3-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false)
// 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* @[[GLOB4]])
// CHECK3-NEXT: store i32 [[TMP2]], i32* [[DOTTHREADID_TEMP_]], align 4
-// CHECK3-NEXT: call void @__omp_outlined__4(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], [10 x i32]* [[TMP0]]) #[[ATTR1]]
-// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 false)
+// CHECK3-NEXT: call void @__omp_outlined__4(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], [10 x i32]* [[TMP0]]) #[[ATTR2]]
+// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 false)
// CHECK3-NEXT: ret void
// CHECK3: worker.exit:
// CHECK3-NEXT: ret void
@@ -12315,7 +12315,7 @@ int bar(int n){
// 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]], i1 true, i1 false, i1 false)
+// CHECK3-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false)
// 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:
@@ -12324,8 +12324,8 @@ int bar(int n){
// CHECK3-NEXT: store i32 [[TMP3]], i32* [[F_CASTED]], align 4
// CHECK3-NEXT: [[TMP4:%.*]] = load i32, i32* [[F_CASTED]], align 4
// CHECK3-NEXT: store i32 [[TMP2]], i32* [[DOTTHREADID_TEMP_]], align 4
-// CHECK3-NEXT: call void @__omp_outlined__6(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], [10 x [10 x i32]]* [[TMP0]], i32 [[TMP4]]) #[[ATTR1]]
-// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 false)
+// CHECK3-NEXT: call void @__omp_outlined__6(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], [10 x [10 x i32]]* [[TMP0]], i32 [[TMP4]]) #[[ATTR2]]
+// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 false)
// CHECK3-NEXT: ret void
// CHECK3: worker.exit:
// CHECK3-NEXT: ret void
diff --git a/clang/test/OpenMP/nvptx_target_teams_distribute_simd_codegen.cpp b/clang/test/OpenMP/nvptx_target_teams_distribute_simd_codegen.cpp
index f8cabbeb11eb3..1d3433e61316d 100644
--- a/clang/test/OpenMP/nvptx_target_teams_distribute_simd_codegen.cpp
+++ b/clang/test/OpenMP/nvptx_target_teams_distribute_simd_codegen.cpp
@@ -70,24 +70,24 @@ int bar(int n){
}
// CHECK-LABEL: define {{.*}}void {{@__omp_offloading_.+}}_l37(
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 false)
-// CHECK: call void @__kmpc_target_deinit({{.*}}, i1 true, i1 false)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false)
+// CHECK: call void @__kmpc_target_deinit({{.*}}, i8 2, i1 false)
// CHECK: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, {{.+}} 91,
// CHECK: call void @__kmpc_for_static_fini(
// CHECK: ret void
// CHECK-LABEL: define {{.*}}void {{@__omp_offloading_.+}}_l43(
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 false)
-// CHECK: call void @__kmpc_target_deinit({{.*}}, i1 true, i1 false)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false)
+// CHECK: call void @__kmpc_target_deinit({{.*}}, i8 2, i1 false)
// CHECK: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, {{.+}} 91,
// CHECK: call void @__kmpc_for_static_fini(
// CHECK: ret void
// CHECK-LABEL: define {{.*}}void {{@__omp_offloading_.+}}_l48(
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 false)
-// CHECK: call void @__kmpc_target_deinit({{.*}}, i1 true, i1 false)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false)
+// CHECK: call void @__kmpc_target_deinit({{.*}}, i8 2, i1 false)
// CHECK: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, {{.+}} 91,
// CHECK: call void @__kmpc_for_static_fini(
@@ -95,8 +95,8 @@ int bar(int n){
// CHECK: define {{.*}}void {{@__omp_offloading_.+}}_l53({{.+}}, i{{32|64}} [[F_IN:%.+]])
// CHECK: store {{.+}} [[F_IN]], {{.+}}* {{.+}},
-// CHECK: call i32 @__kmpc_target_init({{.*}}, i1 true, i1 false, i1 false)
-// CHECK: call void @__kmpc_target_deinit({{.*}}, i1 true, i1 false)
+// CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false)
+// CHECK: call void @__kmpc_target_deinit({{.*}}, i8 2, i1 false)
// CHECK: store {{.+}} 99, {{.+}}* [[COMB_UB:%.+]], align
// CHECK: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, {{.+}} 91, {{.+}}, {{.+}}, {{.+}}* [[COMB_UB]],
diff --git a/clang/test/OpenMP/nvptx_teams_codegen.cpp b/clang/test/OpenMP/nvptx_teams_codegen.cpp
index 8006fecb42bbe..1e87175d719c7 100644
--- a/clang/test/OpenMP/nvptx_teams_codegen.cpp
+++ b/clang/test/OpenMP/nvptx_teams_codegen.cpp
@@ -899,7 +899,7 @@ int main (int argc, char **argv) {
// 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]+]], i1 false, i1 true, i1 true)
+// CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true)
// 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:
@@ -909,9 +909,9 @@ int main (int argc, char **argv) {
// CHECK1-NEXT: store i32 [[TMP1]], i32* [[ARGC_ON_STACK]], align 4
// CHECK1-NEXT: [[TMP2:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]])
// CHECK1-NEXT: store i32 [[TMP2]], i32* [[DOTTHREADID_TEMP_]], align 4
-// CHECK1-NEXT: call void @__omp_outlined__(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i32* [[ARGC_ON_STACK]]) #[[ATTR1:[0-9]+]]
+// CHECK1-NEXT: call void @__omp_outlined__(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i32* [[ARGC_ON_STACK]]) #[[ATTR2:[0-9]+]]
// CHECK1-NEXT: call void @__kmpc_free_shared(i8* [[ARGC1]], i64 4)
-// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK1-NEXT: ret void
// CHECK1: worker.exit:
// CHECK1-NEXT: ret void
@@ -939,7 +939,7 @@ int main (int argc, char **argv) {
// 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]], i1 false, i1 true, i1 true)
+// CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
// 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:
@@ -949,9 +949,9 @@ int main (int argc, char **argv) {
// CHECK1-NEXT: store i8** [[TMP1]], i8*** [[ARGC_ON_STACK]], align 8
// CHECK1-NEXT: [[TMP2:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]])
// CHECK1-NEXT: store i32 [[TMP2]], i32* [[DOTTHREADID_TEMP_]], align 4
-// CHECK1-NEXT: call void @__omp_outlined__1(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i8*** [[ARGC_ON_STACK]]) #[[ATTR1]]
+// CHECK1-NEXT: call void @__omp_outlined__1(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i8*** [[ARGC_ON_STACK]]) #[[ATTR2]]
// CHECK1-NEXT: call void @__kmpc_free_shared(i8* [[ARGC1]], i64 8)
-// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK1-NEXT: ret void
// CHECK1: worker.exit:
// CHECK1-NEXT: ret void
@@ -979,7 +979,7 @@ int main (int argc, char **argv) {
// 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]+]], i1 false, i1 true, i1 true)
+// CHECK2-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true)
// 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:
@@ -989,9 +989,9 @@ int main (int argc, char **argv) {
// CHECK2-NEXT: store i32 [[TMP1]], i32* [[ARGC_ON_STACK]], align 4
// CHECK2-NEXT: [[TMP2:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]])
// CHECK2-NEXT: store i32 [[TMP2]], i32* [[DOTTHREADID_TEMP_]], align 4
-// CHECK2-NEXT: call void @__omp_outlined__(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i32* [[ARGC_ON_STACK]]) #[[ATTR1:[0-9]+]]
+// CHECK2-NEXT: call void @__omp_outlined__(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i32* [[ARGC_ON_STACK]]) #[[ATTR2:[0-9]+]]
// CHECK2-NEXT: call void @__kmpc_free_shared(i8* [[ARGC1]], i32 4)
-// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK2-NEXT: ret void
// CHECK2: worker.exit:
// CHECK2-NEXT: ret void
@@ -1019,7 +1019,7 @@ int main (int argc, char **argv) {
// 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]], i1 false, i1 true, i1 true)
+// CHECK2-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
// 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:
@@ -1029,9 +1029,9 @@ int main (int argc, char **argv) {
// CHECK2-NEXT: store i8** [[TMP1]], i8*** [[ARGC_ON_STACK]], align 4
// CHECK2-NEXT: [[TMP2:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]])
// CHECK2-NEXT: store i32 [[TMP2]], i32* [[DOTTHREADID_TEMP_]], align 4
-// CHECK2-NEXT: call void @__omp_outlined__1(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i8*** [[ARGC_ON_STACK]]) #[[ATTR1]]
+// CHECK2-NEXT: call void @__omp_outlined__1(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i8*** [[ARGC_ON_STACK]]) #[[ATTR2]]
// CHECK2-NEXT: call void @__kmpc_free_shared(i8* [[ARGC1]], i32 4)
-// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK2-NEXT: ret void
// CHECK2: worker.exit:
// CHECK2-NEXT: ret void
@@ -1066,7 +1066,7 @@ int main (int argc, char **argv) {
// 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]+]], i1 false, i1 true, i1 true)
+// CHECK3-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true)
// 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:
@@ -1076,9 +1076,9 @@ int main (int argc, char **argv) {
// CHECK3-NEXT: store i32 [[TMP1]], i32* [[ARGC_ON_STACK]], align 4
// CHECK3-NEXT: [[TMP2:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]])
// CHECK3-NEXT: store i32 [[TMP2]], i32* [[DOTTHREADID_TEMP_]], align 4
-// CHECK3-NEXT: call void @__omp_outlined__(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i32* [[ARGC_ON_STACK]]) #[[ATTR1:[0-9]+]]
+// CHECK3-NEXT: call void @__omp_outlined__(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i32* [[ARGC_ON_STACK]]) #[[ATTR2:[0-9]+]]
// CHECK3-NEXT: call void @__kmpc_free_shared(i8* [[ARGC3]], i64 4)
-// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK3-NEXT: ret void
// CHECK3: worker.exit:
// CHECK3-NEXT: ret void
@@ -1112,7 +1112,7 @@ int main (int argc, char **argv) {
// 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]], i1 false, i1 true, i1 true)
+// CHECK3-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
// 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:
@@ -1122,9 +1122,9 @@ int main (int argc, char **argv) {
// CHECK3-NEXT: store i8** [[TMP1]], i8*** [[ARGC_ON_STACK]], align 8
// CHECK3-NEXT: [[TMP2:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]])
// CHECK3-NEXT: store i32 [[TMP2]], i32* [[DOTTHREADID_TEMP_]], align 4
-// CHECK3-NEXT: call void @__omp_outlined__1(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i8*** [[ARGC_ON_STACK]]) #[[ATTR1]]
+// CHECK3-NEXT: call void @__omp_outlined__1(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i8*** [[ARGC_ON_STACK]]) #[[ATTR2]]
// CHECK3-NEXT: call void @__kmpc_free_shared(i8* [[ARGC2]], i64 8)
-// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK3-NEXT: ret void
// CHECK3: worker.exit:
// CHECK3-NEXT: ret void
@@ -1156,7 +1156,7 @@ int main (int argc, char **argv) {
// 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]+]], i1 false, i1 true, i1 true)
+// CHECK4-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true)
// 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:
@@ -1166,9 +1166,9 @@ int main (int argc, char **argv) {
// CHECK4-NEXT: store i32 [[TMP1]], i32* [[ARGC_ON_STACK]], align 4
// CHECK4-NEXT: [[TMP2:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]])
// CHECK4-NEXT: store i32 [[TMP2]], i32* [[DOTTHREADID_TEMP_]], align 4
-// CHECK4-NEXT: call void @__omp_outlined__(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i32* [[ARGC_ON_STACK]]) #[[ATTR1:[0-9]+]]
+// CHECK4-NEXT: call void @__omp_outlined__(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i32* [[ARGC_ON_STACK]]) #[[ATTR2:[0-9]+]]
// CHECK4-NEXT: call void @__kmpc_free_shared(i8* [[ARGC1]], i32 4)
-// CHECK4-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK4-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK4-NEXT: ret void
// CHECK4: worker.exit:
// CHECK4-NEXT: ret void
@@ -1200,7 +1200,7 @@ int main (int argc, char **argv) {
// 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]], i1 false, i1 true, i1 true)
+// CHECK4-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
// 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:
@@ -1210,9 +1210,9 @@ int main (int argc, char **argv) {
// CHECK4-NEXT: store i8** [[TMP1]], i8*** [[ARGC_ON_STACK]], align 4
// CHECK4-NEXT: [[TMP2:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]])
// CHECK4-NEXT: store i32 [[TMP2]], i32* [[DOTTHREADID_TEMP_]], align 4
-// CHECK4-NEXT: call void @__omp_outlined__1(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i8*** [[ARGC_ON_STACK]]) #[[ATTR1]]
+// CHECK4-NEXT: call void @__omp_outlined__1(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i8*** [[ARGC_ON_STACK]]) #[[ATTR2]]
// CHECK4-NEXT: call void @__kmpc_free_shared(i8* [[ARGC1]], i32 4)
-// CHECK4-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK4-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK4-NEXT: ret void
// CHECK4: worker.exit:
// CHECK4-NEXT: ret void
diff --git a/clang/test/OpenMP/nvptx_teams_reduction_codegen.cpp b/clang/test/OpenMP/nvptx_teams_reduction_codegen.cpp
index f7a9c0b3081e0..973c5b3293eb0 100644
--- a/clang/test/OpenMP/nvptx_teams_reduction_codegen.cpp
+++ b/clang/test/OpenMP/nvptx_teams_reduction_codegen.cpp
@@ -4215,7 +4215,7 @@ int bar(int n){
// 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]+]], i1 false, i1 true, i1 true)
+// CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true)
// 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:
@@ -4225,9 +4225,9 @@ int bar(int n){
// CHECK1-NEXT: store double [[TMP1]], double* [[E_ON_STACK]], align 8
// CHECK1-NEXT: [[TMP2:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]])
// CHECK1-NEXT: store i32 [[TMP2]], i32* [[DOTTHREADID_TEMP_]], align 4
-// CHECK1-NEXT: call void @__omp_outlined__(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], double* [[E_ON_STACK]]) #[[ATTR1:[0-9]+]]
+// CHECK1-NEXT: call void @__omp_outlined__(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], double* [[E_ON_STACK]]) #[[ATTR4:[0-9]+]]
// CHECK1-NEXT: call void @__kmpc_free_shared(i8* [[E1]], i64 8)
-// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK1-NEXT: ret void
// CHECK1: worker.exit:
// CHECK1-NEXT: ret void
@@ -4323,7 +4323,7 @@ int bar(int n){
// CHECK1: then:
// CHECK1-NEXT: [[TMP35:%.*]] = bitcast [1 x i8*]* [[TMP5]] to i8*
// CHECK1-NEXT: [[TMP36:%.*]] = bitcast [1 x i8*]* [[DOTOMP_REDUCTION_REMOTE_REDUCE_LIST]] to i8*
-// CHECK1-NEXT: call void @"_omp$reduction$reduction_func"(i8* [[TMP35]], i8* [[TMP36]]) #[[ATTR1]]
+// CHECK1-NEXT: call void @"_omp$reduction$reduction_func"(i8* [[TMP35]], i8* [[TMP36]]) #[[ATTR4]]
// CHECK1-NEXT: br label [[IFCONT:%.*]]
// CHECK1: else:
// CHECK1-NEXT: br label [[IFCONT]]
@@ -4453,7 +4453,7 @@ int bar(int n){
// CHECK1-NEXT: store i8* [[TMP8]], i8** [[TMP6]], align 8
// CHECK1-NEXT: [[TMP9:%.*]] = bitcast [1 x i8*]* [[DOTOMP_REDUCTION_RED_LIST]] to i8*
// CHECK1-NEXT: [[TMP10:%.*]] = load i8*, i8** [[DOTADDR2]], align 8
-// CHECK1-NEXT: call void @"_omp$reduction$reduction_func"(i8* [[TMP9]], i8* [[TMP10]]) #[[ATTR1]]
+// CHECK1-NEXT: call void @"_omp$reduction$reduction_func"(i8* [[TMP9]], i8* [[TMP10]]) #[[ATTR4]]
// CHECK1-NEXT: ret void
//
//
@@ -4501,7 +4501,7 @@ int bar(int n){
// CHECK1-NEXT: store i8* [[TMP8]], i8** [[TMP6]], align 8
// CHECK1-NEXT: [[TMP9:%.*]] = bitcast [1 x i8*]* [[DOTOMP_REDUCTION_RED_LIST]] to i8*
// CHECK1-NEXT: [[TMP10:%.*]] = load i8*, i8** [[DOTADDR2]], align 8
-// CHECK1-NEXT: call void @"_omp$reduction$reduction_func"(i8* [[TMP10]], i8* [[TMP9]]) #[[ATTR1]]
+// CHECK1-NEXT: call void @"_omp$reduction$reduction_func"(i8* [[TMP10]], i8* [[TMP9]]) #[[ATTR4]]
// CHECK1-NEXT: ret void
//
//
@@ -4517,7 +4517,7 @@ int bar(int n){
// 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]], i1 false, i1 true, i1 true)
+// CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
// 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:
@@ -4530,10 +4530,10 @@ int bar(int n){
// CHECK1-NEXT: store float [[TMP2]], float* [[D_ON_STACK]], align 4
// CHECK1-NEXT: [[TMP3:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]])
// CHECK1-NEXT: store i32 [[TMP3]], i32* [[DOTTHREADID_TEMP_]], align 4
-// CHECK1-NEXT: call void @__omp_outlined__1(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i8* [[C2]], float* [[D_ON_STACK]]) #[[ATTR1]]
+// CHECK1-NEXT: call void @__omp_outlined__1(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i8* [[C2]], float* [[D_ON_STACK]]) #[[ATTR4]]
// CHECK1-NEXT: call void @__kmpc_free_shared(i8* [[D3]], i64 4)
// CHECK1-NEXT: call void @__kmpc_free_shared(i8* [[C2]], i64 1)
-// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK1-NEXT: ret void
// CHECK1: worker.exit:
// CHECK1-NEXT: ret void
@@ -4664,7 +4664,7 @@ int bar(int n){
// CHECK1: then:
// CHECK1-NEXT: [[TMP46:%.*]] = bitcast [2 x i8*]* [[TMP5]] to i8*
// CHECK1-NEXT: [[TMP47:%.*]] = bitcast [2 x i8*]* [[DOTOMP_REDUCTION_REMOTE_REDUCE_LIST]] to i8*
-// CHECK1-NEXT: call void @"_omp$reduction$reduction_func2"(i8* [[TMP46]], i8* [[TMP47]]) #[[ATTR1]]
+// CHECK1-NEXT: call void @"_omp$reduction$reduction_func2"(i8* [[TMP46]], i8* [[TMP47]]) #[[ATTR4]]
// CHECK1-NEXT: br label [[IFCONT:%.*]]
// CHECK1: else:
// CHECK1-NEXT: br label [[IFCONT]]
@@ -4825,7 +4825,7 @@ int bar(int n){
// CHECK1-NEXT: store i8* [[TMP10]], i8** [[TMP8]], align 8
// CHECK1-NEXT: [[TMP11:%.*]] = bitcast [2 x i8*]* [[DOTOMP_REDUCTION_RED_LIST]] to i8*
// CHECK1-NEXT: [[TMP12:%.*]] = load i8*, i8** [[DOTADDR2]], align 8
-// CHECK1-NEXT: call void @"_omp$reduction$reduction_func2"(i8* [[TMP11]], i8* [[TMP12]]) #[[ATTR1]]
+// CHECK1-NEXT: call void @"_omp$reduction$reduction_func2"(i8* [[TMP11]], i8* [[TMP12]]) #[[ATTR4]]
// CHECK1-NEXT: ret void
//
//
@@ -4883,7 +4883,7 @@ int bar(int n){
// CHECK1-NEXT: store i8* [[TMP10]], i8** [[TMP8]], align 8
// CHECK1-NEXT: [[TMP11:%.*]] = bitcast [2 x i8*]* [[DOTOMP_REDUCTION_RED_LIST]] to i8*
// CHECK1-NEXT: [[TMP12:%.*]] = load i8*, i8** [[DOTADDR2]], align 8
-// CHECK1-NEXT: call void @"_omp$reduction$reduction_func2"(i8* [[TMP12]], i8* [[TMP11]]) #[[ATTR1]]
+// CHECK1-NEXT: call void @"_omp$reduction$reduction_func2"(i8* [[TMP12]], i8* [[TMP11]]) #[[ATTR4]]
// CHECK1-NEXT: ret void
//
//
@@ -4899,14 +4899,14 @@ int bar(int n){
// 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]], i1 true, i1 false, i1 true)
+// CHECK1-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true)
// 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: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB3:[0-9]+]])
// CHECK1-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
-// CHECK1-NEXT: call void @__omp_outlined__9(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i32* [[CONV]], i16* [[CONV1]]) #[[ATTR1]]
-// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 true)
+// CHECK1-NEXT: call void @__omp_outlined__9(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i32* [[CONV]], i16* [[CONV1]]) #[[ATTR4]]
+// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 true)
// CHECK1-NEXT: ret void
// CHECK1: worker.exit:
// CHECK1-NEXT: ret void
@@ -5118,7 +5118,7 @@ int bar(int n){
// CHECK1: then:
// CHECK1-NEXT: [[TMP47:%.*]] = bitcast [2 x i8*]* [[TMP5]] to i8*
// CHECK1-NEXT: [[TMP48:%.*]] = bitcast [2 x i8*]* [[DOTOMP_REDUCTION_REMOTE_REDUCE_LIST]] to i8*
-// CHECK1-NEXT: call void @"_omp$reduction$reduction_func11"(i8* [[TMP47]], i8* [[TMP48]]) #[[ATTR1]]
+// CHECK1-NEXT: call void @"_omp$reduction$reduction_func11"(i8* [[TMP47]], i8* [[TMP48]]) #[[ATTR4]]
// CHECK1-NEXT: br label [[IFCONT:%.*]]
// CHECK1: else:
// CHECK1-NEXT: br label [[IFCONT]]
@@ -5296,7 +5296,7 @@ int bar(int n){
// CHECK1: then:
// CHECK1-NEXT: [[TMP47:%.*]] = bitcast [2 x i8*]* [[TMP5]] to i8*
// CHECK1-NEXT: [[TMP48:%.*]] = bitcast [2 x i8*]* [[DOTOMP_REDUCTION_REMOTE_REDUCE_LIST]] to i8*
-// CHECK1-NEXT: call void @"_omp$reduction$reduction_func14"(i8* [[TMP47]], i8* [[TMP48]]) #[[ATTR1]]
+// CHECK1-NEXT: call void @"_omp$reduction$reduction_func14"(i8* [[TMP47]], i8* [[TMP48]]) #[[ATTR4]]
// CHECK1-NEXT: br label [[IFCONT:%.*]]
// CHECK1: else:
// CHECK1-NEXT: br label [[IFCONT]]
@@ -5463,7 +5463,7 @@ int bar(int n){
// CHECK1-NEXT: store i8* [[TMP11]], i8** [[TMP9]], align 8
// CHECK1-NEXT: [[TMP12:%.*]] = bitcast [2 x i8*]* [[DOTOMP_REDUCTION_RED_LIST]] to i8*
// CHECK1-NEXT: [[TMP13:%.*]] = load i8*, i8** [[DOTADDR2]], align 8
-// CHECK1-NEXT: call void @"_omp$reduction$reduction_func14"(i8* [[TMP12]], i8* [[TMP13]]) #[[ATTR1]]
+// CHECK1-NEXT: call void @"_omp$reduction$reduction_func14"(i8* [[TMP12]], i8* [[TMP13]]) #[[ATTR4]]
// CHECK1-NEXT: ret void
//
//
@@ -5523,7 +5523,7 @@ int bar(int n){
// CHECK1-NEXT: store i8* [[TMP11]], i8** [[TMP9]], align 8
// CHECK1-NEXT: [[TMP12:%.*]] = bitcast [2 x i8*]* [[DOTOMP_REDUCTION_RED_LIST]] to i8*
// CHECK1-NEXT: [[TMP13:%.*]] = load i8*, i8** [[DOTADDR2]], align 8
-// CHECK1-NEXT: call void @"_omp$reduction$reduction_func14"(i8* [[TMP13]], i8* [[TMP12]]) #[[ATTR1]]
+// CHECK1-NEXT: call void @"_omp$reduction$reduction_func14"(i8* [[TMP13]], i8* [[TMP12]]) #[[ATTR4]]
// CHECK1-NEXT: ret void
//
//
@@ -5537,7 +5537,7 @@ int bar(int n){
// 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]+]], i1 false, i1 true, i1 true)
+// CHECK2-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true)
// 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:
@@ -5545,8 +5545,8 @@ int bar(int n){
// CHECK2-NEXT: [[TMP3:%.*]] = load double, double* [[TMP0]], align 8
// CHECK2-NEXT: store double [[TMP3]], double* [[E1]], align 8
// CHECK2-NEXT: store i32 [[TMP2]], i32* [[DOTTHREADID_TEMP_]], align 4
-// CHECK2-NEXT: call void @__omp_outlined__(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], double* [[E1]]) #[[ATTR1:[0-9]+]]
-// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK2-NEXT: call void @__omp_outlined__(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], double* [[E1]]) #[[ATTR4:[0-9]+]]
+// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK2-NEXT: ret void
// CHECK2: worker.exit:
// CHECK2-NEXT: ret void
@@ -5642,7 +5642,7 @@ int bar(int n){
// CHECK2: then:
// CHECK2-NEXT: [[TMP35:%.*]] = bitcast [1 x i8*]* [[TMP5]] to i8*
// CHECK2-NEXT: [[TMP36:%.*]] = bitcast [1 x i8*]* [[DOTOMP_REDUCTION_REMOTE_REDUCE_LIST]] to i8*
-// CHECK2-NEXT: call void @"_omp$reduction$reduction_func"(i8* [[TMP35]], i8* [[TMP36]]) #[[ATTR1]]
+// CHECK2-NEXT: call void @"_omp$reduction$reduction_func"(i8* [[TMP35]], i8* [[TMP36]]) #[[ATTR4]]
// CHECK2-NEXT: br label [[IFCONT:%.*]]
// CHECK2: else:
// CHECK2-NEXT: br label [[IFCONT]]
@@ -5772,7 +5772,7 @@ int bar(int n){
// CHECK2-NEXT: store i8* [[TMP8]], i8** [[TMP6]], align 4
// CHECK2-NEXT: [[TMP9:%.*]] = bitcast [1 x i8*]* [[DOTOMP_REDUCTION_RED_LIST]] to i8*
// CHECK2-NEXT: [[TMP10:%.*]] = load i8*, i8** [[DOTADDR2]], align 4
-// CHECK2-NEXT: call void @"_omp$reduction$reduction_func"(i8* [[TMP9]], i8* [[TMP10]]) #[[ATTR1]]
+// CHECK2-NEXT: call void @"_omp$reduction$reduction_func"(i8* [[TMP9]], i8* [[TMP10]]) #[[ATTR4]]
// CHECK2-NEXT: ret void
//
//
@@ -5820,7 +5820,7 @@ int bar(int n){
// CHECK2-NEXT: store i8* [[TMP8]], i8** [[TMP6]], align 4
// CHECK2-NEXT: [[TMP9:%.*]] = bitcast [1 x i8*]* [[DOTOMP_REDUCTION_RED_LIST]] to i8*
// CHECK2-NEXT: [[TMP10:%.*]] = load i8*, i8** [[DOTADDR2]], align 4
-// CHECK2-NEXT: call void @"_omp$reduction$reduction_func"(i8* [[TMP10]], i8* [[TMP9]]) #[[ATTR1]]
+// CHECK2-NEXT: call void @"_omp$reduction$reduction_func"(i8* [[TMP10]], i8* [[TMP9]]) #[[ATTR4]]
// CHECK2-NEXT: ret void
//
//
@@ -5836,7 +5836,7 @@ int bar(int n){
// 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]], i1 false, i1 true, i1 true)
+// CHECK2-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
// 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:
@@ -5849,10 +5849,10 @@ int bar(int n){
// CHECK2-NEXT: store float [[TMP2]], float* [[D_ON_STACK]], align 4
// CHECK2-NEXT: [[TMP3:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]])
// CHECK2-NEXT: store i32 [[TMP3]], i32* [[DOTTHREADID_TEMP_]], align 4
-// CHECK2-NEXT: call void @__omp_outlined__1(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i8* [[C2]], float* [[D_ON_STACK]]) #[[ATTR1]]
+// CHECK2-NEXT: call void @__omp_outlined__1(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i8* [[C2]], float* [[D_ON_STACK]]) #[[ATTR4]]
// CHECK2-NEXT: call void @__kmpc_free_shared(i8* [[D3]], i32 4)
// CHECK2-NEXT: call void @__kmpc_free_shared(i8* [[C2]], i32 1)
-// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK2-NEXT: ret void
// CHECK2: worker.exit:
// CHECK2-NEXT: ret void
@@ -5983,7 +5983,7 @@ int bar(int n){
// CHECK2: then:
// CHECK2-NEXT: [[TMP46:%.*]] = bitcast [2 x i8*]* [[TMP5]] to i8*
// CHECK2-NEXT: [[TMP47:%.*]] = bitcast [2 x i8*]* [[DOTOMP_REDUCTION_REMOTE_REDUCE_LIST]] to i8*
-// CHECK2-NEXT: call void @"_omp$reduction$reduction_func2"(i8* [[TMP46]], i8* [[TMP47]]) #[[ATTR1]]
+// CHECK2-NEXT: call void @"_omp$reduction$reduction_func2"(i8* [[TMP46]], i8* [[TMP47]]) #[[ATTR4]]
// CHECK2-NEXT: br label [[IFCONT:%.*]]
// CHECK2: else:
// CHECK2-NEXT: br label [[IFCONT]]
@@ -6144,7 +6144,7 @@ int bar(int n){
// CHECK2-NEXT: store i8* [[TMP10]], i8** [[TMP8]], align 4
// CHECK2-NEXT: [[TMP11:%.*]] = bitcast [2 x i8*]* [[DOTOMP_REDUCTION_RED_LIST]] to i8*
// CHECK2-NEXT: [[TMP12:%.*]] = load i8*, i8** [[DOTADDR2]], align 4
-// CHECK2-NEXT: call void @"_omp$reduction$reduction_func2"(i8* [[TMP11]], i8* [[TMP12]]) #[[ATTR1]]
+// CHECK2-NEXT: call void @"_omp$reduction$reduction_func2"(i8* [[TMP11]], i8* [[TMP12]]) #[[ATTR4]]
// CHECK2-NEXT: ret void
//
//
@@ -6202,7 +6202,7 @@ int bar(int n){
// CHECK2-NEXT: store i8* [[TMP10]], i8** [[TMP8]], align 4
// CHECK2-NEXT: [[TMP11:%.*]] = bitcast [2 x i8*]* [[DOTOMP_REDUCTION_RED_LIST]] to i8*
// CHECK2-NEXT: [[TMP12:%.*]] = load i8*, i8** [[DOTADDR2]], align 4
-// CHECK2-NEXT: call void @"_omp$reduction$reduction_func2"(i8* [[TMP12]], i8* [[TMP11]]) #[[ATTR1]]
+// CHECK2-NEXT: call void @"_omp$reduction$reduction_func2"(i8* [[TMP12]], i8* [[TMP11]]) #[[ATTR4]]
// CHECK2-NEXT: ret void
//
//
@@ -6217,14 +6217,14 @@ int bar(int n){
// 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]], i1 true, i1 false, i1 true)
+// CHECK2-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true)
// 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: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB3:[0-9]+]])
// CHECK2-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
-// CHECK2-NEXT: call void @__omp_outlined__9(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i32* [[A_ADDR]], i16* [[CONV]]) #[[ATTR1]]
-// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 true)
+// CHECK2-NEXT: call void @__omp_outlined__9(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i32* [[A_ADDR]], i16* [[CONV]]) #[[ATTR4]]
+// CHECK2-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 true)
// CHECK2-NEXT: ret void
// CHECK2: worker.exit:
// CHECK2-NEXT: ret void
@@ -6436,7 +6436,7 @@ int bar(int n){
// CHECK2: then:
// CHECK2-NEXT: [[TMP47:%.*]] = bitcast [2 x i8*]* [[TMP5]] to i8*
// CHECK2-NEXT: [[TMP48:%.*]] = bitcast [2 x i8*]* [[DOTOMP_REDUCTION_REMOTE_REDUCE_LIST]] to i8*
-// CHECK2-NEXT: call void @"_omp$reduction$reduction_func11"(i8* [[TMP47]], i8* [[TMP48]]) #[[ATTR1]]
+// CHECK2-NEXT: call void @"_omp$reduction$reduction_func11"(i8* [[TMP47]], i8* [[TMP48]]) #[[ATTR4]]
// CHECK2-NEXT: br label [[IFCONT:%.*]]
// CHECK2: else:
// CHECK2-NEXT: br label [[IFCONT]]
@@ -6614,7 +6614,7 @@ int bar(int n){
// CHECK2: then:
// CHECK2-NEXT: [[TMP47:%.*]] = bitcast [2 x i8*]* [[TMP5]] to i8*
// CHECK2-NEXT: [[TMP48:%.*]] = bitcast [2 x i8*]* [[DOTOMP_REDUCTION_REMOTE_REDUCE_LIST]] to i8*
-// CHECK2-NEXT: call void @"_omp$reduction$reduction_func14"(i8* [[TMP47]], i8* [[TMP48]]) #[[ATTR1]]
+// CHECK2-NEXT: call void @"_omp$reduction$reduction_func14"(i8* [[TMP47]], i8* [[TMP48]]) #[[ATTR4]]
// CHECK2-NEXT: br label [[IFCONT:%.*]]
// CHECK2: else:
// CHECK2-NEXT: br label [[IFCONT]]
@@ -6781,7 +6781,7 @@ int bar(int n){
// CHECK2-NEXT: store i8* [[TMP11]], i8** [[TMP9]], align 4
// CHECK2-NEXT: [[TMP12:%.*]] = bitcast [2 x i8*]* [[DOTOMP_REDUCTION_RED_LIST]] to i8*
// CHECK2-NEXT: [[TMP13:%.*]] = load i8*, i8** [[DOTADDR2]], align 4
-// CHECK2-NEXT: call void @"_omp$reduction$reduction_func14"(i8* [[TMP12]], i8* [[TMP13]]) #[[ATTR1]]
+// CHECK2-NEXT: call void @"_omp$reduction$reduction_func14"(i8* [[TMP12]], i8* [[TMP13]]) #[[ATTR4]]
// CHECK2-NEXT: ret void
//
//
@@ -6841,7 +6841,7 @@ int bar(int n){
// CHECK2-NEXT: store i8* [[TMP11]], i8** [[TMP9]], align 4
// CHECK2-NEXT: [[TMP12:%.*]] = bitcast [2 x i8*]* [[DOTOMP_REDUCTION_RED_LIST]] to i8*
// CHECK2-NEXT: [[TMP13:%.*]] = load i8*, i8** [[DOTADDR2]], align 4
-// CHECK2-NEXT: call void @"_omp$reduction$reduction_func14"(i8* [[TMP13]], i8* [[TMP12]]) #[[ATTR1]]
+// CHECK2-NEXT: call void @"_omp$reduction$reduction_func14"(i8* [[TMP13]], i8* [[TMP12]]) #[[ATTR4]]
// CHECK2-NEXT: ret void
//
//
@@ -6855,7 +6855,7 @@ int bar(int n){
// 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]+]], i1 false, i1 true, i1 true)
+// CHECK3-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 true, i1 true)
// 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:
@@ -6863,8 +6863,8 @@ int bar(int n){
// CHECK3-NEXT: [[TMP3:%.*]] = load double, double* [[TMP0]], align 8
// CHECK3-NEXT: store double [[TMP3]], double* [[E1]], align 8
// CHECK3-NEXT: store i32 [[TMP2]], i32* [[DOTTHREADID_TEMP_]], align 4
-// CHECK3-NEXT: call void @__omp_outlined__(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], double* [[E1]]) #[[ATTR1:[0-9]+]]
-// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK3-NEXT: call void @__omp_outlined__(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], double* [[E1]]) #[[ATTR4:[0-9]+]]
+// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK3-NEXT: ret void
// CHECK3: worker.exit:
// CHECK3-NEXT: ret void
@@ -6960,7 +6960,7 @@ int bar(int n){
// CHECK3: then:
// CHECK3-NEXT: [[TMP35:%.*]] = bitcast [1 x i8*]* [[TMP5]] to i8*
// CHECK3-NEXT: [[TMP36:%.*]] = bitcast [1 x i8*]* [[DOTOMP_REDUCTION_REMOTE_REDUCE_LIST]] to i8*
-// CHECK3-NEXT: call void @"_omp$reduction$reduction_func"(i8* [[TMP35]], i8* [[TMP36]]) #[[ATTR1]]
+// CHECK3-NEXT: call void @"_omp$reduction$reduction_func"(i8* [[TMP35]], i8* [[TMP36]]) #[[ATTR4]]
// CHECK3-NEXT: br label [[IFCONT:%.*]]
// CHECK3: else:
// CHECK3-NEXT: br label [[IFCONT]]
@@ -7090,7 +7090,7 @@ int bar(int n){
// CHECK3-NEXT: store i8* [[TMP8]], i8** [[TMP6]], align 4
// CHECK3-NEXT: [[TMP9:%.*]] = bitcast [1 x i8*]* [[DOTOMP_REDUCTION_RED_LIST]] to i8*
// CHECK3-NEXT: [[TMP10:%.*]] = load i8*, i8** [[DOTADDR2]], align 4
-// CHECK3-NEXT: call void @"_omp$reduction$reduction_func"(i8* [[TMP9]], i8* [[TMP10]]) #[[ATTR1]]
+// CHECK3-NEXT: call void @"_omp$reduction$reduction_func"(i8* [[TMP9]], i8* [[TMP10]]) #[[ATTR4]]
// CHECK3-NEXT: ret void
//
//
@@ -7138,7 +7138,7 @@ int bar(int n){
// CHECK3-NEXT: store i8* [[TMP8]], i8** [[TMP6]], align 4
// CHECK3-NEXT: [[TMP9:%.*]] = bitcast [1 x i8*]* [[DOTOMP_REDUCTION_RED_LIST]] to i8*
// CHECK3-NEXT: [[TMP10:%.*]] = load i8*, i8** [[DOTADDR2]], align 4
-// CHECK3-NEXT: call void @"_omp$reduction$reduction_func"(i8* [[TMP10]], i8* [[TMP9]]) #[[ATTR1]]
+// CHECK3-NEXT: call void @"_omp$reduction$reduction_func"(i8* [[TMP10]], i8* [[TMP9]]) #[[ATTR4]]
// CHECK3-NEXT: ret void
//
//
@@ -7154,7 +7154,7 @@ int bar(int n){
// 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]], i1 false, i1 true, i1 true)
+// CHECK3-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
// 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:
@@ -7167,10 +7167,10 @@ int bar(int n){
// CHECK3-NEXT: store float [[TMP2]], float* [[D_ON_STACK]], align 4
// CHECK3-NEXT: [[TMP3:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]])
// CHECK3-NEXT: store i32 [[TMP3]], i32* [[DOTTHREADID_TEMP_]], align 4
-// CHECK3-NEXT: call void @__omp_outlined__1(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i8* [[C2]], float* [[D_ON_STACK]]) #[[ATTR1]]
+// CHECK3-NEXT: call void @__omp_outlined__1(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i8* [[C2]], float* [[D_ON_STACK]]) #[[ATTR4]]
// CHECK3-NEXT: call void @__kmpc_free_shared(i8* [[D3]], i32 4)
// CHECK3-NEXT: call void @__kmpc_free_shared(i8* [[C2]], i32 1)
-// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
// CHECK3-NEXT: ret void
// CHECK3: worker.exit:
// CHECK3-NEXT: ret void
@@ -7301,7 +7301,7 @@ int bar(int n){
// CHECK3: then:
// CHECK3-NEXT: [[TMP46:%.*]] = bitcast [2 x i8*]* [[TMP5]] to i8*
// CHECK3-NEXT: [[TMP47:%.*]] = bitcast [2 x i8*]* [[DOTOMP_REDUCTION_REMOTE_REDUCE_LIST]] to i8*
-// CHECK3-NEXT: call void @"_omp$reduction$reduction_func2"(i8* [[TMP46]], i8* [[TMP47]]) #[[ATTR1]]
+// CHECK3-NEXT: call void @"_omp$reduction$reduction_func2"(i8* [[TMP46]], i8* [[TMP47]]) #[[ATTR4]]
// CHECK3-NEXT: br label [[IFCONT:%.*]]
// CHECK3: else:
// CHECK3-NEXT: br label [[IFCONT]]
@@ -7462,7 +7462,7 @@ int bar(int n){
// CHECK3-NEXT: store i8* [[TMP10]], i8** [[TMP8]], align 4
// CHECK3-NEXT: [[TMP11:%.*]] = bitcast [2 x i8*]* [[DOTOMP_REDUCTION_RED_LIST]] to i8*
// CHECK3-NEXT: [[TMP12:%.*]] = load i8*, i8** [[DOTADDR2]], align 4
-// CHECK3-NEXT: call void @"_omp$reduction$reduction_func2"(i8* [[TMP11]], i8* [[TMP12]]) #[[ATTR1]]
+// CHECK3-NEXT: call void @"_omp$reduction$reduction_func2"(i8* [[TMP11]], i8* [[TMP12]]) #[[ATTR4]]
// CHECK3-NEXT: ret void
//
//
@@ -7520,7 +7520,7 @@ int bar(int n){
// CHECK3-NEXT: store i8* [[TMP10]], i8** [[TMP8]], align 4
// CHECK3-NEXT: [[TMP11:%.*]] = bitcast [2 x i8*]* [[DOTOMP_REDUCTION_RED_LIST]] to i8*
// CHECK3-NEXT: [[TMP12:%.*]] = load i8*, i8** [[DOTADDR2]], align 4
-// CHECK3-NEXT: call void @"_omp$reduction$reduction_func2"(i8* [[TMP12]], i8* [[TMP11]]) #[[ATTR1]]
+// CHECK3-NEXT: call void @"_omp$reduction$reduction_func2"(i8* [[TMP12]], i8* [[TMP11]]) #[[ATTR4]]
// CHECK3-NEXT: ret void
//
//
@@ -7535,14 +7535,14 @@ int bar(int n){
// 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]], i1 true, i1 false, i1 true)
+// CHECK3-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 true)
// 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: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB3:[0-9]+]])
// CHECK3-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
-// CHECK3-NEXT: call void @__omp_outlined__9(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i32* [[A_ADDR]], i16* [[CONV]]) #[[ATTR1]]
-// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 true)
+// CHECK3-NEXT: call void @__omp_outlined__9(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTZERO_ADDR]], i32* [[A_ADDR]], i16* [[CONV]]) #[[ATTR4]]
+// CHECK3-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 true)
// CHECK3-NEXT: ret void
// CHECK3: worker.exit:
// CHECK3-NEXT: ret void
@@ -7754,7 +7754,7 @@ int bar(int n){
// CHECK3: then:
// CHECK3-NEXT: [[TMP47:%.*]] = bitcast [2 x i8*]* [[TMP5]] to i8*
// CHECK3-NEXT: [[TMP48:%.*]] = bitcast [2 x i8*]* [[DOTOMP_REDUCTION_REMOTE_REDUCE_LIST]] to i8*
-// CHECK3-NEXT: call void @"_omp$reduction$reduction_func11"(i8* [[TMP47]], i8* [[TMP48]]) #[[ATTR1]]
+// CHECK3-NEXT: call void @"_omp$reduction$reduction_func11"(i8* [[TMP47]], i8* [[TMP48]]) #[[ATTR4]]
// CHECK3-NEXT: br label [[IFCONT:%.*]]
// CHECK3: else:
// CHECK3-NEXT: br label [[IFCONT]]
@@ -7932,7 +7932,7 @@ int bar(int n){
// CHECK3: then:
// CHECK3-NEXT: [[TMP47:%.*]] = bitcast [2 x i8*]* [[TMP5]] to i8*
// CHECK3-NEXT: [[TMP48:%.*]] = bitcast [2 x i8*]* [[DOTOMP_REDUCTION_REMOTE_REDUCE_LIST]] to i8*
-// CHECK3-NEXT: call void @"_omp$reduction$reduction_func14"(i8* [[TMP47]], i8* [[TMP48]]) #[[ATTR1]]
+// CHECK3-NEXT: call void @"_omp$reduction$reduction_func14"(i8* [[TMP47]], i8* [[TMP48]]) #[[ATTR4]]
// CHECK3-NEXT: br label [[IFCONT:%.*]]
// CHECK3: else:
// CHECK3-NEXT: br label [[IFCONT]]
@@ -8099,7 +8099,7 @@ int bar(int n){
// CHECK3-NEXT: store i8* [[TMP11]], i8** [[TMP9]], align 4
// CHECK3-NEXT: [[TMP12:%.*]] = bitcast [2 x i8*]* [[DOTOMP_REDUCTION_RED_LIST]] to i8*
// CHECK3-NEXT: [[TMP13:%.*]] = load i8*, i8** [[DOTADDR2]], align 4
-// CHECK3-NEXT: call void @"_omp$reduction$reduction_func14"(i8* [[TMP12]], i8* [[TMP13]]) #[[ATTR1]]
+// CHECK3-NEXT: call void @"_omp$reduction$reduction_func14"(i8* [[TMP12]], i8* [[TMP13]]) #[[ATTR4]]
// CHECK3-NEXT: ret void
//
//
@@ -8159,6 +8159,6 @@ int bar(int n){
// CHECK3-NEXT: store i8* [[TMP11]], i8** [[TMP9]], align 4
// CHECK3-NEXT: [[TMP12:%.*]] = bitcast [2 x i8*]* [[DOTOMP_REDUCTION_RED_LIST]] to i8*
// CHECK3-NEXT: [[TMP13:%.*]] = load i8*, i8** [[DOTADDR2]], align 4
-// CHECK3-NEXT: call void @"_omp$reduction$reduction_func14"(i8* [[TMP13]], i8* [[TMP12]]) #[[ATTR1]]
+// CHECK3-NEXT: call void @"_omp$reduction$reduction_func14"(i8* [[TMP13]], i8* [[TMP12]]) #[[ATTR4]]
// CHECK3-NEXT: ret void
//
diff --git a/clang/test/OpenMP/target_parallel_debug_codegen.cpp b/clang/test/OpenMP/target_parallel_debug_codegen.cpp
index a457ed701d0e7..81d1b14750c86 100644
--- a/clang/test/OpenMP/target_parallel_debug_codegen.cpp
+++ b/clang/test/OpenMP/target_parallel_debug_codegen.cpp
@@ -95,7 +95,7 @@ int main() {
// 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]+]], i1 true, 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), !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:
@@ -117,7 +117,7 @@ int main() {
// CHECK1-NEXT: store i8* [[TMP7]], i8** [[TMP18]], align 8, !dbg [[DBG48]]
// CHECK1-NEXT: [[TMP19:%.*]] = bitcast [4 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**, !dbg [[DBG48]]
// CHECK1-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB3]], i32 [[TMP9]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*, [10 x [10 x [10 x i32]]]*, i64, [10 x [10 x i32]]*, i8*)* @__omp_outlined__ to i8*), i8* null, i8** [[TMP19]], i64 4), !dbg [[DBG48]]
-// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB5:[0-9]+]], i1 true, i1 true), !dbg [[DBG49:![0-9]+]]
+// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB5:[0-9]+]], i8 2, i1 true), !dbg [[DBG49:![0-9]+]]
// CHECK1-NEXT: ret void, !dbg [[DBG51:![0-9]+]]
// CHECK1: worker.exit:
// CHECK1-NEXT: ret void, !dbg [[DBG47]]
@@ -316,7 +316,7 @@ int main() {
// 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]+]], i1 true, 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), !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:
@@ -338,7 +338,7 @@ int main() {
// CHECK1-NEXT: store i8* [[TMP8]], i8** [[TMP19]], align 8, !dbg [[DBG138]]
// CHECK1-NEXT: [[TMP20:%.*]] = bitcast [4 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**, !dbg [[DBG138]]
// CHECK1-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB9]], i32 [[TMP10]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*, [10 x [10 x [10 x i32]]]*, i64, [10 x [10 x i32]]*, i8*)* @__omp_outlined__2 to i8*), i8* null, i8** [[TMP20]], i64 4), !dbg [[DBG138]]
-// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB11:[0-9]+]], i1 true, i1 true), !dbg [[DBG139:![0-9]+]]
+// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB11:[0-9]+]], i8 2, i1 true), !dbg [[DBG139:![0-9]+]]
// CHECK1-NEXT: ret void, !dbg [[DBG141:![0-9]+]]
// CHECK1: worker.exit:
// CHECK1-NEXT: ret void, !dbg [[DBG137]]
@@ -531,7 +531,7 @@ int main() {
// 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]+]], i1 true, 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), !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:
@@ -549,7 +549,7 @@ int main() {
// CHECK1-NEXT: store i8* [[TMP11]], i8** [[TMP20]], align 8, !dbg [[DBG213]]
// CHECK1-NEXT: [[TMP21:%.*]] = bitcast [4 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**, !dbg [[DBG213]]
// CHECK1-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB15]], i32 [[TMP13]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*, [10 x [10 x [10 x i32]]]*, i32*, [10 x [10 x i32]]*, i8*)* @__omp_outlined__4 to i8*), i8* null, i8** [[TMP21]], i64 4), !dbg [[DBG213]]
-// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB17:[0-9]+]], i1 true, i1 true), !dbg [[DBG214:![0-9]+]]
+// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB17:[0-9]+]], i8 2, i1 true), !dbg [[DBG214:![0-9]+]]
// CHECK1-NEXT: ret void, !dbg [[DBG216:![0-9]+]]
// CHECK1: worker.exit:
// CHECK1-NEXT: ret void, !dbg [[DBG212]]
diff --git a/clang/test/OpenMP/target_parallel_for_debug_codegen.cpp b/clang/test/OpenMP/target_parallel_for_debug_codegen.cpp
index 651b646427b8f..87d1dfdd170f7 100644
--- a/clang/test/OpenMP/target_parallel_for_debug_codegen.cpp
+++ b/clang/test/OpenMP/target_parallel_for_debug_codegen.cpp
@@ -89,7 +89,7 @@ int main() {
// 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]+]], i1 true, 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), !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:
@@ -114,7 +114,7 @@ int main() {
// CHECK1-NEXT: [[TMP20:%.*]] = zext i1 [[TOBOOL]] to i32, !dbg [[DBG42]]
// CHECK1-NEXT: [[TMP21:%.*]] = bitcast [4 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**, !dbg [[DBG42]]
// CHECK1-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB6]], i32 [[TMP9]], i32 [[TMP20]], i32 -1, i32 -1, i8* bitcast (void (i32*, i32*, [10 x [10 x [10 x i32]]]*, i64, [10 x [10 x i32]]*, i8*)* @__omp_outlined__ to i8*), i8* null, i8** [[TMP21]], i64 4), !dbg [[DBG42]]
-// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB8:[0-9]+]], i1 true, i1 false), !dbg [[DBG45:![0-9]+]]
+// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB8:[0-9]+]], i8 2, i1 false), !dbg [[DBG45:![0-9]+]]
// CHECK1-NEXT: ret void, !dbg [[DBG46:![0-9]+]]
// CHECK1: worker.exit:
// CHECK1-NEXT: ret void, !dbg [[DBG41]]
@@ -392,7 +392,7 @@ int main() {
// 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]+]], i1 true, 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), !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:
@@ -414,7 +414,7 @@ int main() {
// CHECK1-NEXT: store i8* [[TMP8]], i8** [[TMP19]], align 8, !dbg [[DBG147]]
// CHECK1-NEXT: [[TMP20:%.*]] = bitcast [4 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**, !dbg [[DBG147]]
// CHECK1-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB15]], i32 [[TMP10]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*, [10 x [10 x [10 x i32]]]*, i64, [10 x [10 x i32]]*, i8*)* @__omp_outlined__2 to i8*), i8* null, i8** [[TMP20]], i64 4), !dbg [[DBG147]]
-// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB17:[0-9]+]], i1 true, i1 false), !dbg [[DBG148:![0-9]+]]
+// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB17:[0-9]+]], i8 2, i1 false), !dbg [[DBG148:![0-9]+]]
// CHECK1-NEXT: ret void, !dbg [[DBG150:![0-9]+]]
// CHECK1: worker.exit:
// CHECK1-NEXT: ret void, !dbg [[DBG146]]
@@ -680,7 +680,7 @@ int main() {
// 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]+]], i1 true, 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), !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:
@@ -698,7 +698,7 @@ int main() {
// CHECK1-NEXT: store i8* [[TMP11]], i8** [[TMP20]], align 8, !dbg [[DBG237]]
// CHECK1-NEXT: [[TMP21:%.*]] = bitcast [4 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**, !dbg [[DBG237]]
// CHECK1-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* @[[GLOB24]], i32 [[TMP13]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*, [10 x [10 x [10 x i32]]]*, i32*, [10 x [10 x i32]]*, i8*)* @__omp_outlined__4 to i8*), i8* null, i8** [[TMP21]], i64 4), !dbg [[DBG237]]
-// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB26:[0-9]+]], i1 true, i1 false), !dbg [[DBG238:![0-9]+]]
+// CHECK1-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB26:[0-9]+]], i8 2, i1 false), !dbg [[DBG238:![0-9]+]]
// CHECK1-NEXT: ret void, !dbg [[DBG240:![0-9]+]]
// CHECK1: worker.exit:
// CHECK1-NEXT: ret void, !dbg [[DBG236]]
diff --git a/llvm/include/llvm/Frontend/OpenMP/OMPKinds.def b/llvm/include/llvm/Frontend/OpenMP/OMPKinds.def
index 098892a9b59c8..7cb28232b788c 100644
--- a/llvm/include/llvm/Frontend/OpenMP/OMPKinds.def
+++ b/llvm/include/llvm/Frontend/OpenMP/OMPKinds.def
@@ -414,8 +414,8 @@ __OMP_RTL(__kmpc_task_allow_completion_event, false, VoidPtr, IdentPtr,
/* Int */ Int32, /* kmp_task_t */ VoidPtr)
/// OpenMP Device runtime functions
-__OMP_RTL(__kmpc_target_init, false, Int32, IdentPtr, Int1, Int1, Int1)
-__OMP_RTL(__kmpc_target_deinit, false, Void, IdentPtr, Int1, Int1)
+__OMP_RTL(__kmpc_target_init, false, Int32, IdentPtr, Int8, Int1, Int1)
+__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,
VoidPtr, VoidPtr, VoidPtrPtr, SizeTy)
diff --git a/llvm/lib/Frontend/OpenMP/OMPIRBuilder.cpp b/llvm/lib/Frontend/OpenMP/OMPIRBuilder.cpp
index d812f22ff1c65..94217c632e436 100644
--- a/llvm/lib/Frontend/OpenMP/OMPIRBuilder.cpp
+++ b/llvm/lib/Frontend/OpenMP/OMPIRBuilder.cpp
@@ -2761,7 +2761,9 @@ OpenMPIRBuilder::createTargetInit(const LocationDescription &Loc, bool IsSPMD,
Constant *SrcLocStr = getOrCreateSrcLocStr(Loc);
Value *Ident = getOrCreateIdent(SrcLocStr);
- ConstantInt *IsSPMDVal = ConstantInt::getBool(Int32->getContext(), IsSPMD);
+ ConstantInt *IsSPMDVal = ConstantInt::getSigned(
+ IntegerType::getInt8Ty(Int8->getContext()),
+ IsSPMD ? OMP_TGT_EXEC_MODE_SPMD : OMP_TGT_EXEC_MODE_GENERIC);
ConstantInt *UseGenericStateMachine =
ConstantInt::getBool(Int32->getContext(), !IsSPMD);
ConstantInt *RequiresFullRuntimeVal =
@@ -2812,7 +2814,9 @@ void OpenMPIRBuilder::createTargetDeinit(const LocationDescription &Loc,
Constant *SrcLocStr = getOrCreateSrcLocStr(Loc);
Value *Ident = getOrCreateIdent(SrcLocStr);
- ConstantInt *IsSPMDVal = ConstantInt::getBool(Int32->getContext(), IsSPMD);
+ ConstantInt *IsSPMDVal = ConstantInt::getSigned(
+ IntegerType::getInt8Ty(Int8->getContext()),
+ IsSPMD ? OMP_TGT_EXEC_MODE_SPMD : OMP_TGT_EXEC_MODE_GENERIC);
ConstantInt *RequiresFullRuntimeVal =
ConstantInt::getBool(Int32->getContext(), RequiresFullRuntime);
diff --git a/llvm/lib/Transforms/IPO/OpenMPOpt.cpp b/llvm/lib/Transforms/IPO/OpenMPOpt.cpp
index 23d0d8aae09f2..11db2ad0727f8 100644
--- a/llvm/lib/Transforms/IPO/OpenMPOpt.cpp
+++ b/llvm/lib/Transforms/IPO/OpenMPOpt.cpp
@@ -2580,10 +2580,9 @@ ChangeStatus AAExecutionDomainFunction::updateImpl(Attributor &A) {
CB = CB ? OpenMPOpt::getCallIfRegularCall(*CB, &RFI) : nullptr;
if (!CB)
return false;
- const int InitIsSPMDArgNo = 1;
- auto *IsSPMDModeCI =
- dyn_cast<ConstantInt>(CB->getOperand(InitIsSPMDArgNo));
- return IsSPMDModeCI && IsSPMDModeCI->isZero();
+ const int InitModeArgNo = 1;
+ auto *ModeCI = dyn_cast<ConstantInt>(CB->getOperand(InitModeArgNo));
+ return ModeCI && (ModeCI->getSExtValue() & OMP_TGT_EXEC_MODE_GENERIC);
}
if (C->isZero()) {
@@ -2941,7 +2940,7 @@ struct AAKernelInfoFunction : AAKernelInfo {
return FalseVal;
};
- Attributor::SimplifictionCallbackTy IsSPMDModeSimplifyCB =
+ Attributor::SimplifictionCallbackTy ModeSimplifyCB =
[&](const IRPosition &IRP, const AbstractAttribute *AA,
bool &UsedAssumedInformation) -> Optional<Value *> {
// IRP represents the "SPMDCompatibilityTracker" argument of an
@@ -2957,8 +2956,10 @@ struct AAKernelInfoFunction : AAKernelInfo {
} else {
UsedAssumedInformation = false;
}
- auto *Val = ConstantInt::getBool(IRP.getAnchorValue().getContext(),
- SPMDCompatibilityTracker.isAssumed());
+ auto *Val = ConstantInt::getSigned(
+ IntegerType::getInt8Ty(IRP.getAnchorValue().getContext()),
+ SPMDCompatibilityTracker.isAssumed() ? OMP_TGT_EXEC_MODE_SPMD
+ : OMP_TGT_EXEC_MODE_GENERIC);
return Val;
};
@@ -2983,8 +2984,8 @@ struct AAKernelInfoFunction : AAKernelInfo {
return Val;
};
- constexpr const int InitIsSPMDArgNo = 1;
- constexpr const int DeinitIsSPMDArgNo = 1;
+ constexpr const int InitModeArgNo = 1;
+ constexpr const int DeinitModeArgNo = 1;
constexpr const int InitUseStateMachineArgNo = 2;
constexpr const int InitRequiresFullRuntimeArgNo = 3;
constexpr const int DeinitRequiresFullRuntimeArgNo = 2;
@@ -2992,11 +2993,11 @@ struct AAKernelInfoFunction : AAKernelInfo {
IRPosition::callsite_argument(*KernelInitCB, InitUseStateMachineArgNo),
StateMachineSimplifyCB);
A.registerSimplificationCallback(
- IRPosition::callsite_argument(*KernelInitCB, InitIsSPMDArgNo),
- IsSPMDModeSimplifyCB);
+ IRPosition::callsite_argument(*KernelInitCB, InitModeArgNo),
+ ModeSimplifyCB);
A.registerSimplificationCallback(
- IRPosition::callsite_argument(*KernelDeinitCB, DeinitIsSPMDArgNo),
- IsSPMDModeSimplifyCB);
+ IRPosition::callsite_argument(*KernelDeinitCB, DeinitModeArgNo),
+ ModeSimplifyCB);
A.registerSimplificationCallback(
IRPosition::callsite_argument(*KernelInitCB,
InitRequiresFullRuntimeArgNo),
@@ -3007,9 +3008,9 @@ struct AAKernelInfoFunction : AAKernelInfo {
IsGenericModeSimplifyCB);
// Check if we know we are in SPMD-mode already.
- ConstantInt *IsSPMDArg =
- dyn_cast<ConstantInt>(KernelInitCB->getArgOperand(InitIsSPMDArgNo));
- if (IsSPMDArg && !IsSPMDArg->isZero())
+ ConstantInt *ModeArg =
+ dyn_cast<ConstantInt>(KernelInitCB->getArgOperand(InitModeArgNo));
+ if (ModeArg && (ModeArg->getSExtValue() & OMP_TGT_EXEC_MODE_SPMD))
SPMDCompatibilityTracker.indicateOptimisticFixpoint();
// This is a generic region but SPMDization is disabled so stop tracking.
else if (DisableOpenMPOptSPMDization)
@@ -3298,21 +3299,24 @@ struct AAKernelInfoFunction : AAKernelInfo {
ExecModeVal | OMP_TGT_EXEC_MODE_GENERIC_SPMD));
// Next rewrite the init and deinit calls to indicate we use SPMD-mode now.
- const int InitIsSPMDArgNo = 1;
- const int DeinitIsSPMDArgNo = 1;
+ const int InitModeArgNo = 1;
+ const int DeinitModeArgNo = 1;
const int InitUseStateMachineArgNo = 2;
const int InitRequiresFullRuntimeArgNo = 3;
const int DeinitRequiresFullRuntimeArgNo = 2;
auto &Ctx = getAnchorValue().getContext();
- A.changeUseAfterManifest(KernelInitCB->getArgOperandUse(InitIsSPMDArgNo),
- *ConstantInt::getBool(Ctx, 1));
+ A.changeUseAfterManifest(
+ KernelInitCB->getArgOperandUse(InitModeArgNo),
+ *ConstantInt::getSigned(IntegerType::getInt8Ty(Ctx),
+ OMP_TGT_EXEC_MODE_SPMD));
A.changeUseAfterManifest(
KernelInitCB->getArgOperandUse(InitUseStateMachineArgNo),
*ConstantInt::getBool(Ctx, 0));
A.changeUseAfterManifest(
- KernelDeinitCB->getArgOperandUse(DeinitIsSPMDArgNo),
- *ConstantInt::getBool(Ctx, 1));
+ KernelDeinitCB->getArgOperandUse(DeinitModeArgNo),
+ *ConstantInt::getSigned(IntegerType::getInt8Ty(Ctx),
+ OMP_TGT_EXEC_MODE_SPMD));
A.changeUseAfterManifest(
KernelInitCB->getArgOperandUse(InitRequiresFullRuntimeArgNo),
*ConstantInt::getBool(Ctx, 0));
@@ -3337,7 +3341,7 @@ struct AAKernelInfoFunction : AAKernelInfo {
assert(ReachedKnownParallelRegions.isValidState() &&
"Custom state machine with invalid parallel region states?");
- const int InitIsSPMDArgNo = 1;
+ const int InitModeArgNo = 1;
const int InitUseStateMachineArgNo = 2;
// Check if the current configuration is non-SPMD and generic state machine.
@@ -3346,14 +3350,14 @@ struct AAKernelInfoFunction : AAKernelInfo {
// we give up.
ConstantInt *UseStateMachine = dyn_cast<ConstantInt>(
KernelInitCB->getArgOperand(InitUseStateMachineArgNo));
- ConstantInt *IsSPMD =
- dyn_cast<ConstantInt>(KernelInitCB->getArgOperand(InitIsSPMDArgNo));
+ ConstantInt *Mode =
+ dyn_cast<ConstantInt>(KernelInitCB->getArgOperand(InitModeArgNo));
// If we are stuck with generic mode, try to create a custom device (=GPU)
// state machine which is specialized for the parallel regions that are
// reachable by the kernel.
- if (!UseStateMachine || UseStateMachine->isZero() || !IsSPMD ||
- !IsSPMD->isZero())
+ if (!UseStateMachine || UseStateMachine->isZero() || !Mode ||
+ (Mode->getSExtValue() & OMP_TGT_EXEC_MODE_SPMD))
return ChangeStatus::UNCHANGED;
// If not SPMD mode, indicate we use a custom state machine now.
diff --git a/llvm/test/Transforms/OpenMP/always_inline_device.ll b/llvm/test/Transforms/OpenMP/always_inline_device.ll
index 2aa797ee28886..7ff4541a740c3 100644
--- a/llvm/test/Transforms/OpenMP/always_inline_device.ll
+++ b/llvm/test/Transforms/OpenMP/always_inline_device.ll
@@ -12,32 +12,32 @@ define weak void @__omp_offloading_fd02_c0934fc2_foo_l4() #0 {
; 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]+]], i1 false, i1 false, i1 true)
+; CHECK-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 false, i1 true)
; 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:
-; CHECK-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+; CHECK-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
; CHECK-NEXT: ret void
; CHECK: worker.exit:
; CHECK-NEXT: ret void
;
entry:
- %0 = call i32 @__kmpc_target_init(%struct.ident_t* @1, i1 false, i1 true, i1 true)
+ %0 = call i32 @__kmpc_target_init(%struct.ident_t* @1, i8 1, i1 true, i1 true)
%exec_user_code = icmp eq i32 %0, -1
br i1 %exec_user_code, label %user_code.entry, label %worker.exit
user_code.entry: ; preds = %entry
call void @bar() #2
- call void @__kmpc_target_deinit(%struct.ident_t* @1, i1 false, i1 true)
+ call void @__kmpc_target_deinit(%struct.ident_t* @1, i8 1, i1 true)
ret void
worker.exit: ; preds = %entry
ret void
}
-declare i32 @__kmpc_target_init(%struct.ident_t*, i1, i1, i1)
+declare i32 @__kmpc_target_init(%struct.ident_t*, i8, i1, i1)
-declare void @__kmpc_target_deinit(%struct.ident_t*, i1, i1)
+declare void @__kmpc_target_deinit(%struct.ident_t*, i8, i1)
; Function Attrs: convergent nounwind
define hidden void @bar() #1 {
diff --git a/llvm/test/Transforms/OpenMP/custom_state_machines.ll b/llvm/test/Transforms/OpenMP/custom_state_machines.ll
index e95a5f662c65a..d2c2985902207 100644
--- a/llvm/test/Transforms/OpenMP/custom_state_machines.ll
+++ b/llvm/test/Transforms/OpenMP/custom_state_machines.ll
@@ -141,7 +141,7 @@ entry:
%.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, i1 false, i1 true, i1 true)
+ %0 = call i32 @__kmpc_target_init(%struct.ident_t* @1, i8 1, i1 true, i1 true)
%exec_user_code = icmp eq i32 %0, -1
br i1 %exec_user_code, label %user_code.entry, label %worker.exit
@@ -149,14 +149,14 @@ user_code.entry: ; preds = %entry
%1 = call i32 @__kmpc_global_thread_num(%struct.ident_t* @1)
store i32 %1, i32* %.threadid_temp., align 4
call void @__omp_outlined__(i32* %.threadid_temp., i32* %.zero.addr) #3
- call void @__kmpc_target_deinit(%struct.ident_t* @1, i1 false, i1 true)
+ call void @__kmpc_target_deinit(%struct.ident_t* @1, i8 1, i1 true)
ret void
worker.exit: ; preds = %entry
ret void
}
-declare i32 @__kmpc_target_init(%struct.ident_t*, i1, i1, i1)
+declare i32 @__kmpc_target_init(%struct.ident_t*, i8, i1, i1)
define internal void @__omp_outlined__(i32* noalias %.global_tid., i32* noalias %.bound_tid.) #0 {
entry:
@@ -190,14 +190,14 @@ declare void @unknown_no_openmp() #2
declare i32 @__kmpc_global_thread_num(%struct.ident_t*) #3
-declare void @__kmpc_target_deinit(%struct.ident_t*, i1, i1)
+declare void @__kmpc_target_deinit(%struct.ident_t*, i8, i1)
define weak void @__omp_offloading_14_a36502b_simple_state_machine_l22() #0 {
entry:
%.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, i1 false, i1 true, i1 true)
+ %0 = call i32 @__kmpc_target_init(%struct.ident_t* @1, i8 1, i1 true, i1 true)
%exec_user_code = icmp eq i32 %0, -1
br i1 %exec_user_code, label %user_code.entry, label %worker.exit
@@ -205,7 +205,7 @@ user_code.entry: ; preds = %entry
%1 = call i32 @__kmpc_global_thread_num(%struct.ident_t* @1)
store i32 %1, i32* %.threadid_temp., align 4
call void @__omp_outlined__1(i32* %.threadid_temp., i32* %.zero.addr) #3
- call void @__kmpc_target_deinit(%struct.ident_t* @1, i1 false, i1 true)
+ call void @__kmpc_target_deinit(%struct.ident_t* @1, i8 1, i1 true)
ret void
worker.exit: ; preds = %entry
@@ -292,7 +292,7 @@ entry:
%.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, i1 false, i1 true, i1 true)
+ %0 = call i32 @__kmpc_target_init(%struct.ident_t* @1, i8 1, i1 true, i1 true)
%exec_user_code = icmp eq i32 %0, -1
br i1 %exec_user_code, label %user_code.entry, label %worker.exit
@@ -300,7 +300,7 @@ user_code.entry: ; preds = %entry
%1 = call i32 @__kmpc_global_thread_num(%struct.ident_t* @1)
store i32 %1, i32* %.threadid_temp., align 4
call void @__omp_outlined__4(i32* %.threadid_temp., i32* %.zero.addr) #3
- call void @__kmpc_target_deinit(%struct.ident_t* @1, i1 false, i1 true)
+ call void @__kmpc_target_deinit(%struct.ident_t* @1, i8 1, i1 true)
ret void
worker.exit: ; preds = %entry
@@ -372,7 +372,7 @@ entry:
%.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, i1 false, i1 true, i1 true)
+ %0 = call i32 @__kmpc_target_init(%struct.ident_t* @1, i8 1, i1 true, i1 true)
%exec_user_code = icmp eq i32 %0, -1
br i1 %exec_user_code, label %user_code.entry, label %worker.exit
@@ -380,7 +380,7 @@ user_code.entry: ; preds = %entry
%1 = call i32 @__kmpc_global_thread_num(%struct.ident_t* @1)
store i32 %1, i32* %.threadid_temp., align 4
call void @__omp_outlined__6(i32* %.threadid_temp., i32* %.zero.addr) #3
- call void @__kmpc_target_deinit(%struct.ident_t* @1, i1 false, i1 true)
+ call void @__kmpc_target_deinit(%struct.ident_t* @1, i8 1, i1 true)
ret void
worker.exit: ; preds = %entry
@@ -460,7 +460,7 @@ entry:
%.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, i1 false, i1 true, i1 true)
+ %0 = call i32 @__kmpc_target_init(%struct.ident_t* @1, i8 1, i1 true, i1 true)
%exec_user_code = icmp eq i32 %0, -1
br i1 %exec_user_code, label %user_code.entry, label %worker.exit
@@ -468,7 +468,7 @@ user_code.entry: ; preds = %entry
%1 = call i32 @__kmpc_global_thread_num(%struct.ident_t* @1)
store i32 %1, i32* %.threadid_temp., align 4
call void @__omp_outlined__9(i32* %.threadid_temp., i32* %.zero.addr) #3
- call void @__kmpc_target_deinit(%struct.ident_t* @1, i1 false, i1 true)
+ call void @__kmpc_target_deinit(%struct.ident_t* @1, i8 1, i1 true)
ret void
worker.exit: ; preds = %entry
@@ -546,7 +546,7 @@ entry:
%.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, i1 false, i1 true, i1 true)
+ %0 = call i32 @__kmpc_target_init(%struct.ident_t* @1, i8 1, i1 true, i1 true)
%exec_user_code = icmp eq i32 %0, -1
br i1 %exec_user_code, label %user_code.entry, label %worker.exit
@@ -554,7 +554,7 @@ user_code.entry: ; preds = %entry
%1 = call i32 @__kmpc_global_thread_num(%struct.ident_t* @1)
store i32 %1, i32* %.threadid_temp., align 4
call void @__omp_outlined__12(i32* %.threadid_temp., i32* %.zero.addr) #3
- call void @__kmpc_target_deinit(%struct.ident_t* @1, i1 false, i1 true)
+ call void @__kmpc_target_deinit(%struct.ident_t* @1, i8 1, i1 true)
ret void
worker.exit: ; preds = %entry
@@ -635,7 +635,7 @@ entry:
%.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, i1 false, i1 true, i1 true)
+ %0 = call i32 @__kmpc_target_init(%struct.ident_t* @1, i8 1, i1 true, i1 true)
%exec_user_code = icmp eq i32 %0, -1
br i1 %exec_user_code, label %user_code.entry, label %worker.exit
@@ -643,7 +643,7 @@ user_code.entry: ; preds = %entry
%1 = call i32 @__kmpc_global_thread_num(%struct.ident_t* @1)
store i32 %1, i32* %.threadid_temp., align 4
call void @__omp_outlined__15(i32* %.threadid_temp., i32* %.zero.addr) #3
- call void @__kmpc_target_deinit(%struct.ident_t* @1, i1 false, i1 true)
+ call void @__kmpc_target_deinit(%struct.ident_t* @1, i8 1, i1 true)
ret void
worker.exit: ; preds = %entry
@@ -690,7 +690,7 @@ entry:
%.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, i1 false, i1 true, i1 true)
+ %0 = call i32 @__kmpc_target_init(%struct.ident_t* @1, i8 1, i1 true, i1 true)
%exec_user_code = icmp eq i32 %0, -1
br i1 %exec_user_code, label %user_code.entry, label %worker.exit
@@ -698,7 +698,7 @@ user_code.entry: ; preds = %entry
%1 = call i32 @__kmpc_global_thread_num(%struct.ident_t* @1)
store i32 %1, i32* %.threadid_temp., align 4
call void @__omp_outlined__16(i32* %.threadid_temp., i32* %.zero.addr) #3
- call void @__kmpc_target_deinit(%struct.ident_t* @1, i1 false, i1 true)
+ call void @__kmpc_target_deinit(%struct.ident_t* @1, i8 1, i1 true)
ret void
worker.exit: ; preds = %entry
@@ -847,13 +847,13 @@ attributes #9 = { convergent nounwind readonly willreturn }
; 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]+]], i1 false, i1 false, i1 true)
+; AMDGPU-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 false, i1 true)
; 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:
; AMDGPU-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR3:[0-9]+]]
; AMDGPU-NEXT: call void @__omp_outlined__(i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR3]]
-; AMDGPU-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+; AMDGPU-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
; AMDGPU-NEXT: ret void
; AMDGPU: worker.exit:
; AMDGPU-NEXT: ret void
@@ -911,7 +911,7 @@ attributes #9 = { convergent nounwind readonly willreturn }
; 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]], i1 false, i1 false, i1 true)
+; AMDGPU-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true)
; 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:
@@ -952,7 +952,7 @@ attributes #9 = { convergent nounwind readonly willreturn }
; AMDGPU-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR3]]
; AMDGPU-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
; AMDGPU-NEXT: call void @__omp_outlined__1(i32* noalias nocapture noundef nonnull readonly align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR3]]
-; AMDGPU-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+; AMDGPU-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
; AMDGPU-NEXT: ret void
; AMDGPU: worker.exit:
; AMDGPU-NEXT: ret void
@@ -1036,7 +1036,7 @@ attributes #9 = { convergent nounwind readonly willreturn }
; 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]], i1 false, i1 false, i1 true)
+; AMDGPU-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true)
; 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:
@@ -1083,7 +1083,7 @@ attributes #9 = { convergent nounwind readonly willreturn }
; AMDGPU-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR3]]
; AMDGPU-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
; AMDGPU-NEXT: call void @__omp_outlined__4(i32* noalias nocapture noundef nonnull readonly align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR3]]
-; AMDGPU-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+; AMDGPU-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
; AMDGPU-NEXT: ret void
; AMDGPU: worker.exit:
; AMDGPU-NEXT: ret void
@@ -1184,7 +1184,7 @@ attributes #9 = { convergent nounwind readonly willreturn }
; 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]], i1 false, i1 false, i1 true)
+; AMDGPU-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true)
; 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:
@@ -1227,7 +1227,7 @@ attributes #9 = { convergent nounwind readonly willreturn }
; AMDGPU-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR3]]
; AMDGPU-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
; AMDGPU-NEXT: call void @__omp_outlined__6(i32* noalias nocapture noundef nonnull readonly align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR3]]
-; AMDGPU-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+; AMDGPU-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
; AMDGPU-NEXT: ret void
; AMDGPU: worker.exit:
; AMDGPU-NEXT: ret void
@@ -1310,7 +1310,7 @@ attributes #9 = { convergent nounwind readonly willreturn }
; 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]], i1 false, i1 false, i1 true)
+; AMDGPU-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true)
; 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:
@@ -1351,7 +1351,7 @@ attributes #9 = { convergent nounwind readonly willreturn }
; AMDGPU-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR3]]
; AMDGPU-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
; AMDGPU-NEXT: call void @__omp_outlined__9(i32* noalias nocapture noundef nonnull readonly align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR3]]
-; AMDGPU-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+; AMDGPU-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
; AMDGPU-NEXT: ret void
; AMDGPU: worker.exit:
; AMDGPU-NEXT: ret void
@@ -1434,7 +1434,7 @@ attributes #9 = { convergent nounwind readonly willreturn }
; 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]], i1 false, i1 false, i1 true)
+; AMDGPU-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true)
; 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:
@@ -1475,7 +1475,7 @@ attributes #9 = { convergent nounwind readonly willreturn }
; AMDGPU-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR3]]
; AMDGPU-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
; AMDGPU-NEXT: call void @__omp_outlined__12(i32* noalias nocapture noundef nonnull readonly align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR3]]
-; AMDGPU-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+; AMDGPU-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
; AMDGPU-NEXT: ret void
; AMDGPU: worker.exit:
; AMDGPU-NEXT: ret void
@@ -1558,7 +1558,7 @@ attributes #9 = { convergent nounwind readonly willreturn }
; 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]], i1 false, i1 false, i1 true)
+; AMDGPU-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true)
; 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:
@@ -1594,7 +1594,7 @@ attributes #9 = { convergent nounwind readonly willreturn }
; AMDGPU: user_code.entry:
; AMDGPU-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR3]]
; AMDGPU-NEXT: call void @__omp_outlined__15(i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR3]]
-; AMDGPU-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+; AMDGPU-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
; AMDGPU-NEXT: ret void
; AMDGPU: worker.exit:
; AMDGPU-NEXT: ret void
@@ -1658,7 +1658,7 @@ attributes #9 = { convergent nounwind readonly willreturn }
; 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]], i1 false, i1 false, i1 true)
+; AMDGPU-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true)
; 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:
@@ -1688,7 +1688,7 @@ attributes #9 = { convergent nounwind readonly willreturn }
; AMDGPU: user_code.entry:
; AMDGPU-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR3]]
; AMDGPU-NEXT: call void @__omp_outlined__16(i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR3]]
-; AMDGPU-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+; AMDGPU-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
; AMDGPU-NEXT: ret void
; AMDGPU: worker.exit:
; AMDGPU-NEXT: ret void
@@ -1817,13 +1817,13 @@ attributes #9 = { convergent nounwind readonly willreturn }
; 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]+]], i1 false, i1 false, i1 true)
+; NVPTX-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 1, i1 false, i1 true)
; 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:
; NVPTX-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR3:[0-9]+]]
; NVPTX-NEXT: call void @__omp_outlined__(i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR3]]
-; NVPTX-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+; NVPTX-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
; NVPTX-NEXT: ret void
; NVPTX: worker.exit:
; NVPTX-NEXT: ret void
@@ -1881,7 +1881,7 @@ attributes #9 = { convergent nounwind readonly willreturn }
; 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]], i1 false, i1 false, i1 true)
+; NVPTX-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true)
; 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:
@@ -1921,7 +1921,7 @@ attributes #9 = { convergent nounwind readonly willreturn }
; NVPTX-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR3]]
; NVPTX-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
; NVPTX-NEXT: call void @__omp_outlined__1(i32* noalias nocapture noundef nonnull readonly align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR3]]
-; NVPTX-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+; NVPTX-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
; NVPTX-NEXT: ret void
; NVPTX: worker.exit:
; NVPTX-NEXT: ret void
@@ -2005,7 +2005,7 @@ attributes #9 = { convergent nounwind readonly willreturn }
; 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]], i1 false, i1 false, i1 true)
+; NVPTX-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true)
; 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:
@@ -2051,7 +2051,7 @@ attributes #9 = { convergent nounwind readonly willreturn }
; NVPTX-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR3]]
; NVPTX-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
; NVPTX-NEXT: call void @__omp_outlined__4(i32* noalias nocapture noundef nonnull readonly align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR3]]
-; NVPTX-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+; NVPTX-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
; NVPTX-NEXT: ret void
; NVPTX: worker.exit:
; NVPTX-NEXT: ret void
@@ -2152,7 +2152,7 @@ attributes #9 = { convergent nounwind readonly willreturn }
; 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]], i1 false, i1 false, i1 true)
+; NVPTX-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true)
; 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:
@@ -2194,7 +2194,7 @@ attributes #9 = { convergent nounwind readonly willreturn }
; NVPTX-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR3]]
; NVPTX-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
; NVPTX-NEXT: call void @__omp_outlined__6(i32* noalias nocapture noundef nonnull readonly align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR3]]
-; NVPTX-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+; NVPTX-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
; NVPTX-NEXT: ret void
; NVPTX: worker.exit:
; NVPTX-NEXT: ret void
@@ -2277,7 +2277,7 @@ attributes #9 = { convergent nounwind readonly willreturn }
; 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]], i1 false, i1 false, i1 true)
+; NVPTX-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true)
; 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:
@@ -2317,7 +2317,7 @@ attributes #9 = { convergent nounwind readonly willreturn }
; NVPTX-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR3]]
; NVPTX-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
; NVPTX-NEXT: call void @__omp_outlined__9(i32* noalias nocapture noundef nonnull readonly align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR3]]
-; NVPTX-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+; NVPTX-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
; NVPTX-NEXT: ret void
; NVPTX: worker.exit:
; NVPTX-NEXT: ret void
@@ -2400,7 +2400,7 @@ attributes #9 = { convergent nounwind readonly willreturn }
; 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]], i1 false, i1 false, i1 true)
+; NVPTX-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true)
; 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:
@@ -2440,7 +2440,7 @@ attributes #9 = { convergent nounwind readonly willreturn }
; NVPTX-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR3]]
; NVPTX-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
; NVPTX-NEXT: call void @__omp_outlined__12(i32* noalias nocapture noundef nonnull readonly align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR3]]
-; NVPTX-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+; NVPTX-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
; NVPTX-NEXT: ret void
; NVPTX: worker.exit:
; NVPTX-NEXT: ret void
@@ -2523,7 +2523,7 @@ attributes #9 = { convergent nounwind readonly willreturn }
; 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]], i1 false, i1 false, i1 true)
+; NVPTX-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true)
; 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:
@@ -2558,7 +2558,7 @@ attributes #9 = { convergent nounwind readonly willreturn }
; NVPTX: user_code.entry:
; NVPTX-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR3]]
; NVPTX-NEXT: call void @__omp_outlined__15(i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR3]]
-; NVPTX-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+; NVPTX-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
; NVPTX-NEXT: ret void
; NVPTX: worker.exit:
; NVPTX-NEXT: ret void
@@ -2622,7 +2622,7 @@ attributes #9 = { convergent nounwind readonly willreturn }
; 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]], i1 false, i1 false, i1 true)
+; NVPTX-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true)
; 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:
@@ -2651,7 +2651,7 @@ attributes #9 = { convergent nounwind readonly willreturn }
; NVPTX: user_code.entry:
; NVPTX-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR3]]
; NVPTX-NEXT: call void @__omp_outlined__16(i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR3]]
-; NVPTX-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+; NVPTX-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
; NVPTX-NEXT: ret void
; NVPTX: worker.exit:
; NVPTX-NEXT: ret void
@@ -2780,13 +2780,13 @@ attributes #9 = { convergent nounwind readonly willreturn }
; 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]+]], i1 false, 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)
; 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:
; AMDGPU-DISABLED-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR3:[0-9]+]]
; AMDGPU-DISABLED-NEXT: call void @__omp_outlined__(i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR3]]
-; AMDGPU-DISABLED-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+; AMDGPU-DISABLED-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
; AMDGPU-DISABLED-NEXT: ret void
; AMDGPU-DISABLED: worker.exit:
; AMDGPU-DISABLED-NEXT: ret void
@@ -2843,14 +2843,14 @@ attributes #9 = { convergent nounwind readonly willreturn }
; 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]], i1 false, i1 true, i1 true)
+; AMDGPU-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
; 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:
; AMDGPU-DISABLED-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR3]]
; AMDGPU-DISABLED-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
; AMDGPU-DISABLED-NEXT: call void @__omp_outlined__1(i32* noalias nocapture noundef nonnull readonly align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR3]]
-; AMDGPU-DISABLED-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+; AMDGPU-DISABLED-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
; AMDGPU-DISABLED-NEXT: ret void
; AMDGPU-DISABLED: worker.exit:
; AMDGPU-DISABLED-NEXT: ret void
@@ -2933,14 +2933,14 @@ attributes #9 = { convergent nounwind readonly willreturn }
; 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]], i1 false, i1 true, i1 true)
+; AMDGPU-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
; 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:
; AMDGPU-DISABLED-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR3]]
; AMDGPU-DISABLED-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
; AMDGPU-DISABLED-NEXT: call void @__omp_outlined__4(i32* noalias nocapture noundef nonnull readonly align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR3]]
-; AMDGPU-DISABLED-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+; AMDGPU-DISABLED-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
; AMDGPU-DISABLED-NEXT: ret void
; AMDGPU-DISABLED: worker.exit:
; AMDGPU-DISABLED-NEXT: ret void
@@ -3040,14 +3040,14 @@ attributes #9 = { convergent nounwind readonly willreturn }
; 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]], i1 false, i1 true, i1 true)
+; AMDGPU-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
; 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:
; AMDGPU-DISABLED-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR3]]
; AMDGPU-DISABLED-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
; AMDGPU-DISABLED-NEXT: call void @__omp_outlined__6(i32* noalias nocapture noundef nonnull readonly align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR3]]
-; AMDGPU-DISABLED-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+; AMDGPU-DISABLED-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
; AMDGPU-DISABLED-NEXT: ret void
; AMDGPU-DISABLED: worker.exit:
; AMDGPU-DISABLED-NEXT: ret void
@@ -3129,14 +3129,14 @@ attributes #9 = { convergent nounwind readonly willreturn }
; 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]], i1 false, i1 true, i1 true)
+; AMDGPU-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
; 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:
; AMDGPU-DISABLED-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR3]]
; AMDGPU-DISABLED-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
; AMDGPU-DISABLED-NEXT: call void @__omp_outlined__9(i32* noalias nocapture noundef nonnull readonly align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR3]]
-; AMDGPU-DISABLED-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+; AMDGPU-DISABLED-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
; AMDGPU-DISABLED-NEXT: ret void
; AMDGPU-DISABLED: worker.exit:
; AMDGPU-DISABLED-NEXT: ret void
@@ -3218,14 +3218,14 @@ attributes #9 = { convergent nounwind readonly willreturn }
; 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]], i1 false, i1 true, i1 true)
+; AMDGPU-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
; 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:
; AMDGPU-DISABLED-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR3]]
; AMDGPU-DISABLED-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
; AMDGPU-DISABLED-NEXT: call void @__omp_outlined__12(i32* noalias nocapture noundef nonnull readonly align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR3]]
-; AMDGPU-DISABLED-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+; AMDGPU-DISABLED-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
; AMDGPU-DISABLED-NEXT: ret void
; AMDGPU-DISABLED: worker.exit:
; AMDGPU-DISABLED-NEXT: ret void
@@ -3307,13 +3307,13 @@ attributes #9 = { convergent nounwind readonly willreturn }
; 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]], i1 false, i1 true, i1 true)
+; AMDGPU-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
; 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:
; AMDGPU-DISABLED-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR3]]
; AMDGPU-DISABLED-NEXT: call void @__omp_outlined__15(i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR3]]
-; AMDGPU-DISABLED-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+; AMDGPU-DISABLED-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
; AMDGPU-DISABLED-NEXT: ret void
; AMDGPU-DISABLED: worker.exit:
; AMDGPU-DISABLED-NEXT: ret void
@@ -3376,13 +3376,13 @@ attributes #9 = { convergent nounwind readonly willreturn }
; 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]], i1 false, i1 true, i1 true)
+; AMDGPU-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
; 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:
; AMDGPU-DISABLED-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR3]]
; AMDGPU-DISABLED-NEXT: call void @__omp_outlined__16(i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR3]]
-; AMDGPU-DISABLED-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+; AMDGPU-DISABLED-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
; AMDGPU-DISABLED-NEXT: ret void
; AMDGPU-DISABLED: worker.exit:
; AMDGPU-DISABLED-NEXT: ret void
@@ -3511,13 +3511,13 @@ attributes #9 = { convergent nounwind readonly willreturn }
; 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]+]], i1 false, 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)
; 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:
; NVPTX-DISABLED-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR3:[0-9]+]]
; NVPTX-DISABLED-NEXT: call void @__omp_outlined__(i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR3]]
-; NVPTX-DISABLED-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+; NVPTX-DISABLED-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
; NVPTX-DISABLED-NEXT: ret void
; NVPTX-DISABLED: worker.exit:
; NVPTX-DISABLED-NEXT: ret void
@@ -3574,14 +3574,14 @@ attributes #9 = { convergent nounwind readonly willreturn }
; 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]], i1 false, i1 true, i1 true)
+; NVPTX-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
; 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:
; NVPTX-DISABLED-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR3]]
; NVPTX-DISABLED-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
; NVPTX-DISABLED-NEXT: call void @__omp_outlined__1(i32* noalias nocapture noundef nonnull readonly align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR3]]
-; NVPTX-DISABLED-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+; NVPTX-DISABLED-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
; NVPTX-DISABLED-NEXT: ret void
; NVPTX-DISABLED: worker.exit:
; NVPTX-DISABLED-NEXT: ret void
@@ -3664,14 +3664,14 @@ attributes #9 = { convergent nounwind readonly willreturn }
; 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]], i1 false, i1 true, i1 true)
+; NVPTX-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
; 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:
; NVPTX-DISABLED-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR3]]
; NVPTX-DISABLED-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
; NVPTX-DISABLED-NEXT: call void @__omp_outlined__4(i32* noalias nocapture noundef nonnull readonly align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR3]]
-; NVPTX-DISABLED-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+; NVPTX-DISABLED-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
; NVPTX-DISABLED-NEXT: ret void
; NVPTX-DISABLED: worker.exit:
; NVPTX-DISABLED-NEXT: ret void
@@ -3771,14 +3771,14 @@ attributes #9 = { convergent nounwind readonly willreturn }
; 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]], i1 false, i1 true, i1 true)
+; NVPTX-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
; 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:
; NVPTX-DISABLED-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR3]]
; NVPTX-DISABLED-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
; NVPTX-DISABLED-NEXT: call void @__omp_outlined__6(i32* noalias nocapture noundef nonnull readonly align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR3]]
-; NVPTX-DISABLED-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+; NVPTX-DISABLED-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
; NVPTX-DISABLED-NEXT: ret void
; NVPTX-DISABLED: worker.exit:
; NVPTX-DISABLED-NEXT: ret void
@@ -3860,14 +3860,14 @@ attributes #9 = { convergent nounwind readonly willreturn }
; 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]], i1 false, i1 true, i1 true)
+; NVPTX-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
; 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:
; NVPTX-DISABLED-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR3]]
; NVPTX-DISABLED-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
; NVPTX-DISABLED-NEXT: call void @__omp_outlined__9(i32* noalias nocapture noundef nonnull readonly align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR3]]
-; NVPTX-DISABLED-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+; NVPTX-DISABLED-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
; NVPTX-DISABLED-NEXT: ret void
; NVPTX-DISABLED: worker.exit:
; NVPTX-DISABLED-NEXT: ret void
@@ -3949,14 +3949,14 @@ attributes #9 = { convergent nounwind readonly willreturn }
; 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]], i1 false, i1 true, i1 true)
+; NVPTX-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
; 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:
; NVPTX-DISABLED-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR3]]
; NVPTX-DISABLED-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
; NVPTX-DISABLED-NEXT: call void @__omp_outlined__12(i32* noalias nocapture noundef nonnull readonly align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR3]]
-; NVPTX-DISABLED-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+; NVPTX-DISABLED-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
; NVPTX-DISABLED-NEXT: ret void
; NVPTX-DISABLED: worker.exit:
; NVPTX-DISABLED-NEXT: ret void
@@ -4038,13 +4038,13 @@ attributes #9 = { convergent nounwind readonly willreturn }
; 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]], i1 false, i1 true, i1 true)
+; NVPTX-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
; 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:
; NVPTX-DISABLED-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR3]]
; NVPTX-DISABLED-NEXT: call void @__omp_outlined__15(i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR3]]
-; NVPTX-DISABLED-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+; NVPTX-DISABLED-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
; NVPTX-DISABLED-NEXT: ret void
; NVPTX-DISABLED: worker.exit:
; NVPTX-DISABLED-NEXT: ret void
@@ -4107,13 +4107,13 @@ attributes #9 = { convergent nounwind readonly willreturn }
; 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]], i1 false, i1 true, i1 true)
+; NVPTX-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
; 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:
; NVPTX-DISABLED-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR3]]
; NVPTX-DISABLED-NEXT: call void @__omp_outlined__16(i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR3]]
-; NVPTX-DISABLED-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+; NVPTX-DISABLED-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
; NVPTX-DISABLED-NEXT: ret void
; NVPTX-DISABLED: worker.exit:
; NVPTX-DISABLED-NEXT: ret void
diff --git a/llvm/test/Transforms/OpenMP/custom_state_machines_remarks.ll b/llvm/test/Transforms/OpenMP/custom_state_machines_remarks.ll
index a4af1be603e7a..ae269ca39626e 100644
--- a/llvm/test/Transforms/OpenMP/custom_state_machines_remarks.ll
+++ b/llvm/test/Transforms/OpenMP/custom_state_machines_remarks.ll
@@ -14,7 +14,7 @@ target triple = "nvptx64"
;; unknown();
;; }
;; }
-;;
+;;
;; void test_fallback(void) {
;; #pragma omp target teams
;; {
@@ -60,7 +60,7 @@ target triple = "nvptx64"
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, i1 false, i1 true, i1 true) #3, !dbg !18
+ %0 = call i32 @__kmpc_target_init(%struct.ident_t* nonnull @1, i8 1, i1 true, i1 true) #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
@@ -77,11 +77,11 @@ user_code.entry: ; preds = %entry
call void @__kmpc_parallel_51(%struct.ident_t* noundef nonnull @13, i32 %3, i32 noundef 1, i32 noundef -1, i32 noundef -1, i8* noundef bitcast (void (i32*, i32*)* @__omp_outlined__2 to i8*), i8* noundef bitcast (void (i16, i32)* @__omp_outlined__2_wrapper to i8*), i8** noundef nonnull %4, i64 noundef 0) #3, !dbg !23
call void @llvm.lifetime.end.p0i8(i64 0, i8* nonnull %2) #3, !dbg !26
call void @unknown() #6, !dbg !27
- call void @__kmpc_target_deinit(%struct.ident_t* nonnull @5, i1 false, i1 true) #3, !dbg !28
+ call void @__kmpc_target_deinit(%struct.ident_t* nonnull @5, i8 1, i1 true) #3, !dbg !28
br label %common.ret
}
-declare i32 @__kmpc_target_init(%struct.ident_t*, i1, i1, i1) local_unnamed_addr
+declare i32 @__kmpc_target_init(%struct.ident_t*, i8, i1, i1) local_unnamed_addr
; Function Attrs: convergent
declare void @unknown() local_unnamed_addr #1
@@ -99,13 +99,13 @@ entry:
; Function Attrs: nounwind
declare i32 @__kmpc_global_thread_num(%struct.ident_t*) local_unnamed_addr #3
-declare void @__kmpc_target_deinit(%struct.ident_t*, i1, i1) local_unnamed_addr
+declare void @__kmpc_target_deinit(%struct.ident_t*, i8, i1) local_unnamed_addr
; Function Attrs: norecurse nounwind
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, i1 false, i1 true, i1 true) #3, !dbg !33
+ %0 = call i32 @__kmpc_target_init(%struct.ident_t* nonnull @7, i8 1, i1 true, i1 true) #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
@@ -130,7 +130,7 @@ user_code.entry: ; preds = %entry
call void @llvm.lifetime.end.p0i8(i64 0, i8* nonnull %2) #3, !dbg !45
call void @no_openmp()
call void @no_parallelism()
- call void @__kmpc_target_deinit(%struct.ident_t* nonnull @11, i1 false, i1 true) #3, !dbg !46
+ call void @__kmpc_target_deinit(%struct.ident_t* nonnull @11, i8 1, i1 true) #3, !dbg !46
br label %common.ret
}
diff --git a/llvm/test/Transforms/OpenMP/deduplication_target.ll b/llvm/test/Transforms/OpenMP/deduplication_target.ll
index 4b9b86559ed69..64ec1a688bfff 100644
--- a/llvm/test/Transforms/OpenMP/deduplication_target.ll
+++ b/llvm/test/Transforms/OpenMP/deduplication_target.ll
@@ -20,37 +20,37 @@ define weak void @__omp_offloading_50_a3e09bf8_foo_l2() #0 {
; 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]+]], i1 true, i1 false, i1 true)
+; CHECK-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1:[0-9]+]], i8 2, i1 false, i1 true)
; 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:%.*]]
; CHECK: user_code.entry:
-; CHECK-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 true)
+; CHECK-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 true)
; CHECK-NEXT: ret void
; CHECK: worker.exit:
; CHECK-NEXT: ret void
;
entry:
%captured_vars_addrs = alloca [0 x i8*], align 8
- %0 = call i32 @__kmpc_target_init(%struct.ident_t* @1, i1 true, i1 false, i1 true)
+ %0 = call i32 @__kmpc_target_init(%struct.ident_t* @1, i8 2, i1 false, i1 true)
%exec_user_code = icmp eq i32 %0, -1
br i1 %exec_user_code, label %user_code.entry, label %worker.exit
user_code.entry: ; preds = %entry
%1 = call i32 @__kmpc_global_thread_num(%struct.ident_t* @2)
%2 = call i32 @__kmpc_global_thread_num(%struct.ident_t* @2)
- call void @__kmpc_target_deinit(%struct.ident_t* @1, i1 true, i1 true)
+ call void @__kmpc_target_deinit(%struct.ident_t* @1, i8 2, i1 true)
ret void
worker.exit: ; preds = %entry
ret void
}
-declare i32 @__kmpc_target_init(%struct.ident_t*, i1, i1, i1)
+declare i32 @__kmpc_target_init(%struct.ident_t*, i8, i1, i1)
declare i32 @__kmpc_global_thread_num(%struct.ident_t*) #1
-declare void @__kmpc_target_deinit(%struct.ident_t*, i1, i1)
+declare void @__kmpc_target_deinit(%struct.ident_t*, i8, i1)
attributes #0 = { convergent noinline norecurse nounwind "frame-pointer"="none" "min-legal-vector-width"="0" "no-trapping-math"="true" "stack-protector-buffer-size"="8" "target-features"="+ptx32,+sm_20" }
attributes #1 = { nounwind }
diff --git a/llvm/test/Transforms/OpenMP/fold_generic_main_thread.ll b/llvm/test/Transforms/OpenMP/fold_generic_main_thread.ll
index df1f49e53bbcc..ac644142adad0 100644
--- 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]+]], i1 false, 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)
; CHECK-NEXT: [[CMP:%.*]] = icmp eq i32 [[CALL]], -1
; CHECK-NEXT: br i1 [[CMP]], label [[IF_THEN:%.*]], label [[IF_ELSE:%.*]]
; CHECK: if.then:
@@ -20,11 +20,11 @@ define void @kernel() {
; CHECK-NEXT: call void @bar()
; CHECK-NEXT: br label [[IF_END]]
; CHECK: if.end:
-; CHECK-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* null, i1 false, i1 true)
+; CHECK-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* null, i8 1, i1 true)
; 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]+]], i1 false, 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)
; 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:
@@ -34,10 +34,10 @@ define void @kernel() {
; CHECK-DISABLED-NEXT: call void @bar()
; CHECK-DISABLED-NEXT: br label [[IF_END]]
; CHECK-DISABLED: if.end:
-; CHECK-DISABLED-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* null, i1 false, i1 true)
+; 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, i1 false, i1 false, i1 false)
+ %call = call i32 @__kmpc_target_init(%struct.ident_t* nonnull @1, i8 1, i1 false, i1 false)
%cmp = icmp eq i32 %call, -1
br i1 %cmp, label %if.then, label %if.else
if.then:
@@ -47,7 +47,7 @@ if.else:
call void @bar()
br label %if.end
if.end:
- call void @__kmpc_target_deinit(%struct.ident_t* null, i1 false, i1 true)
+ call void @__kmpc_target_deinit(%struct.ident_t* null, i8 1, i1 true)
ret void
}
@@ -135,9 +135,9 @@ declare i8 @__kmpc_is_generic_main_thread_id(i32)
declare i32 @__kmpc_get_hardware_thread_id()
-declare i32 @__kmpc_target_init(%struct.ident_t*, i1, i1, i1)
+declare i32 @__kmpc_target_init(%struct.ident_t*, i8, i1, i1)
-declare void @__kmpc_target_deinit(%struct.ident_t*, i1, i1)
+declare void @__kmpc_target_deinit(%struct.ident_t*, i8, i1)
!llvm.dbg.cu = !{!0}
!llvm.module.flags = !{!3, !4, !5, !6}
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
index d929d0bd6e63f..a817c384719a6 100644
--- 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
@@ -18,18 +18,18 @@ target triple = "nvptx64"
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, i1 true, i1 false, i1 false)
+; CHECK-NEXT: [[I:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* null, i8 2, i1 false, i1 false)
; 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, i1 true, i1 false)
+; 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, i1 true, i1 false, i1 false)
+ %i = call i32 @__kmpc_target_init(%struct.ident_t* null, i8 2, i1 false, i1 false)
call void @helper0()
call void @helper1()
call void @helper2()
- call void @__kmpc_target_deinit(%struct.ident_t* null, i1 true, i1 false)
+ call void @__kmpc_target_deinit(%struct.ident_t* null, i8 2, i1 false)
ret void
}
@@ -38,14 +38,14 @@ define weak void @kernel0() #0 {
define weak void @kernel1() #0 {
; CHECK-LABEL: define {{[^@]+}}@kernel1
; CHECK-SAME: () #[[ATTR0]] {
-; CHECK-NEXT: [[I:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* null, i1 true, i1 false, i1 false)
+; CHECK-NEXT: [[I:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* null, i8 2, i1 false, i1 false)
; CHECK-NEXT: call void @helper1() #[[ATTR1]]
-; CHECK-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* null, i1 false, i1 false)
+; 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, i1 true, i1 false, i1 false)
+ %i = call i32 @__kmpc_target_init(%struct.ident_t* null, i8 2, i1 false, i1 false)
call void @helper1()
- call void @__kmpc_target_deinit(%struct.ident_t* null, i1 false, i1 false)
+ call void @__kmpc_target_deinit(%struct.ident_t* null, i8 2, i1 false)
ret void
}
@@ -56,7 +56,7 @@ define weak void @kernel2() #0 {
; 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, i1 true, i1 false, i1 false)
+; CHECK-NEXT: [[I:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* null, i8 2, i1 false, i1 false)
; 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:
@@ -68,12 +68,12 @@ define weak void @kernel2() #0 {
; CHECK-NEXT: call void @helper1() #[[ATTR1]]
; CHECK-NEXT: call void @helper2() #[[ATTR1]]
; CHECK-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* null, i32 [[TMP0]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*)* @__omp_outlined__ to i8*), i8* bitcast (void (i16, i32)* @__omp_outlined___wrapper to i8*), i8** [[TMP1]], i64 0)
-; CHECK-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* null, i1 true, i1 false)
+; CHECK-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* null, i8 2, i1 false)
; CHECK-NEXT: ret void
;
entry:
%captured_vars_addrs = alloca [0 x i8*], align 8
- %i = call i32 @__kmpc_target_init(%struct.ident_t* null, i1 false, i1 true, i1 true)
+ %i = call i32 @__kmpc_target_init(%struct.ident_t* null, i8 1, i1 true, i1 true)
%exec_user_code = icmp eq i32 %i, -1
br i1 %exec_user_code, label %user_code.entry, label %common.ret
@@ -87,7 +87,7 @@ user_code.entry:
call void @helper1()
call void @helper2()
call void @__kmpc_parallel_51(%struct.ident_t* null, i32 %0, i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*)* @__omp_outlined__ to i8*), i8* bitcast (void (i16, i32)* @__omp_outlined___wrapper to i8*), i8** %1, i64 0)
- call void @__kmpc_target_deinit(%struct.ident_t* null, i1 false, i1 true)
+ call void @__kmpc_target_deinit(%struct.ident_t* null, i8 1, i1 true)
ret void
}
@@ -179,8 +179,8 @@ entry:
}
declare i32 @__kmpc_get_hardware_num_threads_in_block()
-declare i32 @__kmpc_target_init(%struct.ident_t*, i1 zeroext, i1 zeroext, i1 zeroext) #1
-declare void @__kmpc_target_deinit(%struct.ident_t* nocapture readnone, i1 zeroext, i1 zeroext) #1
+declare i32 @__kmpc_target_init(%struct.ident_t*, i8, i1 zeroext, i1 zeroext) #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
index 45ce9f011f3ff..dbc44c4d723ef 100644
--- 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, i1 true, i1 false, i1 false) #0
+ %0 = tail call i32 @__kmpc_target_init(%struct.ident_t* nonnull @1, i8 2, i1 false, i1 false) #0
%exec_user_code = icmp eq i32 %0, -1
br i1 %exec_user_code, label %user_code.entry, label %common.ret
@@ -29,13 +29,13 @@ region.guarded:
region.barrier:
tail call void @__kmpc_barrier_simple_spmd(%struct.ident_t* nonnull @1, i32 %2)
- tail call void @__kmpc_target_deinit(%struct.ident_t* nonnull @1, i1 true, i1 false) #0
+ tail call void @__kmpc_target_deinit(%struct.ident_t* nonnull @1, i8 2, i1 false) #0
br label %common.ret
}
-declare i32 @__kmpc_target_init(%struct.ident_t*, i1, i1, i1) local_unnamed_addr
+declare i32 @__kmpc_target_init(%struct.ident_t*, i8, i1, i1) local_unnamed_addr
-declare void @__kmpc_target_deinit(%struct.ident_t*, i1, i1) local_unnamed_addr
+declare void @__kmpc_target_deinit(%struct.ident_t*, i8, i1) local_unnamed_addr
define internal void @__omp_offloading__fd02_85283c04_Device_l6_ctor() {
entry:
@@ -78,7 +78,7 @@ attributes #1 = { convergent nounwind }
; 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]+]], i1 true, 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) #[[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:
@@ -93,6 +93,6 @@ attributes #1 = { convergent nounwind }
; CHECK-NEXT: br label [[REGION_BARRIER]]
; CHECK: region.barrier:
; CHECK-NEXT: tail call void @__kmpc_barrier_simple_spmd(%struct.ident_t* nonnull @[[GLOB1]], i32 [[TMP2]]) #[[ATTR1]]
-; CHECK-NEXT: tail call void @__kmpc_target_deinit(%struct.ident_t* nonnull @[[GLOB1]], i1 true, i1 false) #[[ATTR1]]
+; CHECK-NEXT: tail call void @__kmpc_target_deinit(%struct.ident_t* nonnull @[[GLOB1]], i8 2, i1 false) #[[ATTR1]]
; CHECK-NEXT: br label [[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
index 1c9e67cf39f62..f597e0557f949 100644
--- 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 @@ entry:
%.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, i1 false, i1 true, i1 true)
+ %0 = call i32 @__kmpc_target_init(%struct.ident_t* @1, i8 1, i1 true, i1 true)
%exec_user_code = icmp eq i32 %0, -1
br i1 %exec_user_code, label %user_code.entry, label %worker.exit
@@ -56,14 +56,14 @@ user_code.entry: ; preds = %entry
%1 = call i32 @__kmpc_global_thread_num(%struct.ident_t* @1)
store i32 %1, i32* %.threadid_temp., align 4
call void @__omp_outlined__(i32* %.threadid_temp., i32* %.zero.addr)
- call void @__kmpc_target_deinit(%struct.ident_t* @1, i1 false, i1 true)
+ call void @__kmpc_target_deinit(%struct.ident_t* @1, i8 1, i1 true)
ret void
worker.exit: ; preds = %entry
ret void
}
-declare i32 @__kmpc_target_init(%struct.ident_t*, i1, i1, i1)
+declare i32 @__kmpc_target_init(%struct.ident_t*, i8, i1, i1)
declare void @unknown()
define internal void @__omp_outlined__(i32* noalias %.global_tid., i32* noalias %.bound_tid.) {
@@ -146,7 +146,7 @@ entry:
declare i32 @__kmpc_global_thread_num(%struct.ident_t*)
-declare void @__kmpc_target_deinit(%struct.ident_t*, i1, i1)
+declare void @__kmpc_target_deinit(%struct.ident_t*, i8, i1)
define internal void @__omp_outlined__3(i32* noalias %.global_tid., i32* noalias %.bound_tid.) {
entry:
diff --git a/llvm/test/Transforms/OpenMP/is_spmd_exec_mode_fold.ll b/llvm/test/Transforms/OpenMP/is_spmd_exec_mode_fold.ll
index d4069d83a793f..383cba27d446a 100644
--- a/llvm/test/Transforms/OpenMP/is_spmd_exec_mode_fold.ll
+++ b/llvm/test/Transforms/OpenMP/is_spmd_exec_mode_fold.ll
@@ -21,18 +21,18 @@ target triple = "nvptx64"
;.
define weak void @is_spmd() {
; CHECK-LABEL: define {{[^@]+}}@is_spmd() {
-; CHECK-NEXT: [[I:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* null, i1 true, i1 false, i1 false)
+; CHECK-NEXT: [[I:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* null, i8 2, i1 false, i1 false)
; 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, i1 true, i1 false)
+; 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, i1 true, i1 false, i1 false)
+ %i = call i32 @__kmpc_target_init(%struct.ident_t* null, i8 2, i1 false, i1 false)
call void @is_spmd_helper1()
call void @is_spmd_helper2()
call void @is_mixed_helper()
- call void @__kmpc_target_deinit(%struct.ident_t* null, i1 true, i1 false)
+ call void @__kmpc_target_deinit(%struct.ident_t* null, i8 2, i1 false)
ret void
}
@@ -40,7 +40,7 @@ define weak void @will_be_spmd() {
; 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, i1 true, i1 false, i1 false)
+; CHECK-NEXT: [[I:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* null, i8 2, i1 false, i1 false)
; 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:
@@ -50,12 +50,12 @@ define weak void @will_be_spmd() {
; CHECK-NEXT: [[TMP1:%.*]] = bitcast [0 x i8*]* [[CAPTURED_VARS_ADDRS]] to i8**
; CHECK-NEXT: call void @is_spmd_helper2()
; CHECK-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* null, i32 [[TMP0]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*)* @__omp_outlined__ to i8*), i8* bitcast (void (i16, i32)* @__omp_outlined___wrapper to i8*), i8** [[TMP1]], i64 0)
-; CHECK-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* null, i1 true, i1 false)
+; CHECK-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* null, i8 2, i1 false)
; CHECK-NEXT: ret void
;
entry:
%captured_vars_addrs = alloca [0 x i8*], align 8
- %i = call i32 @__kmpc_target_init(%struct.ident_t* null, i1 false, i1 true, i1 true)
+ %i = call i32 @__kmpc_target_init(%struct.ident_t* null, i8 1, i1 true, i1 true)
%exec_user_code = icmp eq i32 %i, -1
br i1 %exec_user_code, label %user_code.entry, label %common.ret
@@ -67,41 +67,41 @@ user_code.entry:
%1 = bitcast [0 x i8*]* %captured_vars_addrs to i8**
call void @is_spmd_helper2()
call void @__kmpc_parallel_51(%struct.ident_t* null, i32 %0, i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*)* @__omp_outlined__ to i8*), i8* bitcast (void (i16, i32)* @__omp_outlined___wrapper to i8*), i8** %1, i64 0)
- call void @__kmpc_target_deinit(%struct.ident_t* null, i1 false, i1 true)
+ call void @__kmpc_target_deinit(%struct.ident_t* null, i8 1, i1 true)
ret void
}
define weak void @non_spmd() {
; CHECK-LABEL: define {{[^@]+}}@non_spmd() {
-; CHECK-NEXT: [[I:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* null, i1 false, i1 false, i1 false)
+; CHECK-NEXT: [[I:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* null, i8 1, i1 false, i1 false)
; 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, i1 false, i1 false)
+; 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, i1 false, i1 false, i1 false)
+ %i = call i32 @__kmpc_target_init(%struct.ident_t* null, i8 1, i1 false, i1 false)
call void @is_generic_helper1()
call void @is_generic_helper2()
call void @is_mixed_helper()
- call void @__kmpc_target_deinit(%struct.ident_t* null, i1 false, i1 false)
+ call void @__kmpc_target_deinit(%struct.ident_t* null, i8 1, i1 false)
ret void
}
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, i1 false, i1 false, i1 false)
+; CHECK-NEXT: [[I:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* null, i8 1, i1 false, i1 false)
; 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, i1 false, i1 false)
+; 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, i1 false, i1 false, i1 false)
+ %i = call i32 @__kmpc_target_init(%struct.ident_t* null, i8 1, i1 false, i1 false)
call void @is_generic_helper1()
call void @is_generic_helper2()
call void @is_mixed_helper()
- call void @__kmpc_target_deinit(%struct.ident_t* null, i1 false, i1 false)
+ call void @__kmpc_target_deinit(%struct.ident_t* null, i8 1, i1 false)
ret void
}
@@ -197,8 +197,8 @@ entry:
declare void @spmd_compatible() "llvm.assume"="ompx_spmd_amenable"
declare i8 @__kmpc_is_spmd_exec_mode()
-declare i32 @__kmpc_target_init(%struct.ident_t*, i1 zeroext, i1 zeroext, i1 zeroext)
-declare void @__kmpc_target_deinit(%struct.ident_t* nocapture readnone, i1 zeroext, i1 zeroext)
+declare i32 @__kmpc_target_init(%struct.ident_t*, i8, i1 zeroext, i1 zeroext)
+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*)
declare void @foo()
diff --git a/llvm/test/Transforms/OpenMP/parallel_level_fold.ll b/llvm/test/Transforms/OpenMP/parallel_level_fold.ll
index f951b270aa55a..1b3dc9b754baa 100644
--- a/llvm/test/Transforms/OpenMP/parallel_level_fold.ll
+++ b/llvm/test/Transforms/OpenMP/parallel_level_fold.ll
@@ -19,46 +19,46 @@ target triple = "nvptx64"
;.
define weak void @none_spmd() {
; CHECK-LABEL: define {{[^@]+}}@none_spmd() {
-; CHECK-NEXT: [[I:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* null, i1 false, i1 false, i1 false)
+; CHECK-NEXT: [[I:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* null, i8 1, i1 false, i1 false)
; CHECK-NEXT: call void @none_spmd_helper()
; CHECK-NEXT: call void @mixed_helper()
-; CHECK-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* null, i1 false, i1 false)
+; 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, i1 false, i1 false, i1 false)
+ %i = call i32 @__kmpc_target_init(%struct.ident_t* null, i8 1, i1 false, i1 false)
call void @none_spmd_helper()
call void @mixed_helper()
- call void @__kmpc_target_deinit(%struct.ident_t* null, i1 false, i1 false)
+ call void @__kmpc_target_deinit(%struct.ident_t* null, i8 1, i1 false)
ret void
}
define weak void @spmd() {
; CHECK-LABEL: define {{[^@]+}}@spmd() {
-; CHECK-NEXT: [[I:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* null, i1 true, i1 false, i1 false)
+; CHECK-NEXT: [[I:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* null, i8 2, i1 false, i1 false)
; CHECK-NEXT: call void @spmd_helper()
; CHECK-NEXT: call void @mixed_helper()
-; CHECK-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* null, i1 true, i1 false)
+; 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, i1 true, i1 false, i1 false)
+ %i = call i32 @__kmpc_target_init(%struct.ident_t* null, i8 2, i1 false, i1 false)
call void @spmd_helper()
call void @mixed_helper()
- call void @__kmpc_target_deinit(%struct.ident_t* null, i1 true, i1 false)
+ call void @__kmpc_target_deinit(%struct.ident_t* null, i8 2, i1 false)
ret void
}
define weak void @parallel() {
; CHECK-LABEL: define {{[^@]+}}@parallel() {
-; CHECK-NEXT: [[I:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* align 1073741824 null, i1 true, i1 false, i1 false)
+; CHECK-NEXT: [[I:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* align 1073741824 null, i8 2, i1 false, i1 false)
; 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, i1 true, i1 false)
+; 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, i1 true, i1 false, i1 false)
+ %i = call i32 @__kmpc_target_init(%struct.ident_t* null, i8 2, i1 false, i1 false)
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, i1 true, i1 false)
+ call void @__kmpc_target_deinit(%struct.ident_t* null, i8 2, i1 false)
ret void
}
@@ -130,8 +130,8 @@ define internal void @parallel_helper() {
declare void @foo()
declare void @bar()
declare i8 @__kmpc_parallel_level()
-declare i32 @__kmpc_target_init(%struct.ident_t*, i1 zeroext, i1 zeroext, i1 zeroext) #1
-declare void @__kmpc_target_deinit(%struct.ident_t* nocapture readnone, i1 zeroext, i1 zeroext) #1
+declare i32 @__kmpc_target_init(%struct.ident_t*, i8 zeroext, i1 zeroext, i1 zeroext) #1
+declare void @__kmpc_target_deinit(%struct.ident_t* nocapture readnone, i8 zeroext, i1 zeroext) #1
!llvm.module.flags = !{!0, !1}
!nvvm.annotations = !{!2, !3, !4}
diff --git a/llvm/test/Transforms/OpenMP/remove_globalization.ll b/llvm/test/Transforms/OpenMP/remove_globalization.ll
index 56f7a0f71900c..e718db3751f48 100644
--- a/llvm/test/Transforms/OpenMP/remove_globalization.ll
+++ b/llvm/test/Transforms/OpenMP/remove_globalization.ll
@@ -14,8 +14,8 @@ target triple = "nvptx64"
%struct.ident_t = type { i32, i32, i32, i32, i8* }
-declare i32 @__kmpc_target_init(%struct.ident_t*, i1, i1, i1)
-declare void @__kmpc_target_deinit(%struct.ident_t*, i1, i1)
+declare i32 @__kmpc_target_init(%struct.ident_t*, i8, i1, i1)
+declare void @__kmpc_target_deinit(%struct.ident_t*, i8, i1)
;.
; CHECK: @[[S:[a-zA-Z0-9_$"\\.-]+]] = external local_unnamed_addr global i8*
@@ -25,28 +25,28 @@ declare void @__kmpc_target_deinit(%struct.ident_t*, i1, i1)
define void @kernel() {
; CHECK-LABEL: define {{[^@]+}}@kernel() {
; CHECK-NEXT: entry:
-; CHECK-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* nonnull null, i1 false, i1 false, i1 true)
+; CHECK-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* nonnull null, i8 1, i1 false, i1 true)
; CHECK-NEXT: call void @foo() #[[ATTR4:[0-9]+]]
; CHECK-NEXT: call void @bar() #[[ATTR4]]
; CHECK-NEXT: call void @unknown_no_openmp()
-; CHECK-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* nonnull null, i1 false, i1 true)
+; CHECK-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* nonnull null, i8 1, i1 true)
; CHECK-NEXT: ret void
;
; CHECK-DISABLED-LABEL: define {{[^@]+}}@kernel() {
; CHECK-DISABLED-NEXT: entry:
-; CHECK-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* nonnull null, i1 false, i1 false, i1 true)
+; CHECK-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* nonnull null, i8 1, i1 false, i1 true)
; CHECK-DISABLED-NEXT: call void @foo() #[[ATTR4:[0-9]+]]
; CHECK-DISABLED-NEXT: call void @bar() #[[ATTR4]]
; CHECK-DISABLED-NEXT: call void @unknown_no_openmp()
-; CHECK-DISABLED-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* nonnull null, i1 false, i1 true)
+; CHECK-DISABLED-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* nonnull null, i8 1, i1 true)
; CHECK-DISABLED-NEXT: ret void
;
entry:
- %0 = call i32 @__kmpc_target_init(%struct.ident_t* nonnull null, i1 false, i1 true, i1 true)
+ %0 = call i32 @__kmpc_target_init(%struct.ident_t* nonnull null, i8 1, i1 true, i1 true)
call void @foo()
call void @bar()
call void @unknown_no_openmp()
- call void @__kmpc_target_deinit(%struct.ident_t* nonnull null, i1 false, i1 true)
+ call void @__kmpc_target_deinit(%struct.ident_t* nonnull null, i8 1, i1 true)
ret void
}
diff --git a/llvm/test/Transforms/OpenMP/replace_globalization.ll b/llvm/test/Transforms/OpenMP/replace_globalization.ll
index 9e74787c3889c..032ea4eb8d1ca 100644
--- a/llvm/test/Transforms/OpenMP/replace_globalization.ll
+++ b/llvm/test/Transforms/OpenMP/replace_globalization.ll
@@ -17,30 +17,30 @@ target triple = "nvptx64"
define dso_local void @foo() {
entry:
- %c = call i32 @__kmpc_target_init(%struct.ident_t* @1, i1 false, i1 true, i1 true)
+ %c = call i32 @__kmpc_target_init(%struct.ident_t* @1, i8 1, i1 true, i1 true)
%x = call i8* @__kmpc_alloc_shared(i64 4)
call void @unknown_no_openmp()
%x_on_stack = bitcast i8* %x to i32*
%0 = bitcast i32* %x_on_stack to i8*
call void @use(i8* %0)
call void @__kmpc_free_shared(i8* %x, i64 4)
- call void @__kmpc_target_deinit(%struct.ident_t* @1, i1 false, i1 true)
+ call void @__kmpc_target_deinit(%struct.ident_t* @1, i8 1, i1 true)
ret void
}
define void @bar() {
- %c = call i32 @__kmpc_target_init(%struct.ident_t* @1, i1 false, i1 true, i1 true)
+ %c = call i32 @__kmpc_target_init(%struct.ident_t* @1, i8 1, i1 true, i1 true)
call void @unknown_no_openmp()
call void @baz()
call void @qux()
call void @negative_qux_spmd()
- call void @__kmpc_target_deinit(%struct.ident_t* @1, i1 false, i1 true)
+ call void @__kmpc_target_deinit(%struct.ident_t* @1, i8 1, i1 true)
ret void
}
define internal void @baz() {
entry:
- %call = call i32 @__kmpc_target_init(%struct.ident_t* nonnull @1, i1 false, i1 false, i1 true)
+ %call = call i32 @__kmpc_target_init(%struct.ident_t* nonnull @1, i8 1, i1 false, i1 true)
call void @unknown_no_openmp()
%cmp = icmp eq i32 %call, -1
br i1 %cmp, label %master, label %exit
@@ -57,7 +57,7 @@ exit:
define internal void @qux() {
entry:
- %call = call i32 @__kmpc_target_init(%struct.ident_t* nonnull @1, i1 false, i1 true, i1 true)
+ %call = call i32 @__kmpc_target_init(%struct.ident_t* nonnull @1, i8 1, i1 true, i1 true)
call void @unknown_no_openmp()
%0 = icmp eq i32 %call, -1
br i1 %0, label %master, label %exit
@@ -74,7 +74,7 @@ exit:
define internal void @negative_qux_spmd() {
entry:
- %call = call i32 @__kmpc_target_init(%struct.ident_t* nonnull @1, i1 true, i1 true, i1 true)
+ %call = call i32 @__kmpc_target_init(%struct.ident_t* nonnull @1, i8 2, i1 true, i1 true)
call void @unknown_no_openmp()
%0 = icmp eq i32 %call, -1
br i1 %0, label %master, label %exit
@@ -106,9 +106,9 @@ declare i32 @llvm.nvvm.read.ptx.sreg.ntid.x()
declare i32 @llvm.nvvm.read.ptx.sreg.warpsize()
-declare i32 @__kmpc_target_init(%struct.ident_t*, i1, i1, i1)
+declare i32 @__kmpc_target_init(%struct.ident_t*, i8, i1, i1)
-declare void @__kmpc_target_deinit(%struct.ident_t*, i1, i1)
+declare void @__kmpc_target_deinit(%struct.ident_t*, i8, i1)
declare void @unknown_no_openmp() "llvm.assume"="omp_no_openmp"
@@ -138,28 +138,28 @@ declare void @unknown_no_openmp() "llvm.assume"="omp_no_openmp"
;.
; CHECK-LABEL: define {{[^@]+}}@foo() {
; CHECK-NEXT: entry:
-; CHECK-NEXT: [[C:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i1 false, i1 false, i1 true)
+; CHECK-NEXT: [[C:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true)
; 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]+]]
; CHECK-NEXT: call void @__kmpc_free_shared(i8* [[X]], i64 4) #[[ATTR4]]
-; CHECK-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+; CHECK-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
; CHECK-NEXT: ret void
;
;
; CHECK-LABEL: define {{[^@]+}}@bar() {
-; CHECK-NEXT: [[C:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i1 false, i1 true, i1 true)
+; CHECK-NEXT: [[C:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 true, i1 true)
; CHECK-NEXT: call void @unknown_no_openmp()
; CHECK-NEXT: call void @baz()
; CHECK-NEXT: call void @qux()
; CHECK-NEXT: call void @negative_qux_spmd()
-; CHECK-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+; CHECK-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
; CHECK-NEXT: ret void
;
;
; CHECK-LABEL: define {{[^@]+}}@baz() {
; CHECK-NEXT: entry:
-; CHECK-NEXT: [[CALL:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* noundef nonnull @[[GLOB1]], i1 noundef false, i1 noundef false, i1 noundef true)
+; CHECK-NEXT: [[CALL:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* noundef nonnull @[[GLOB1]], i8 noundef 1, i1 noundef false, i1 noundef true)
; CHECK-NEXT: call void @unknown_no_openmp()
; CHECK-NEXT: [[CMP:%.*]] = icmp eq i32 [[CALL]], -1
; CHECK-NEXT: br i1 [[CMP]], label [[MASTER:%.*]], label [[EXIT:%.*]]
@@ -172,7 +172,7 @@ declare void @unknown_no_openmp() "llvm.assume"="omp_no_openmp"
;
; CHECK-LABEL: define {{[^@]+}}@qux() {
; CHECK-NEXT: entry:
-; CHECK-NEXT: [[CALL:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* noundef nonnull @[[GLOB1]], i1 noundef false, i1 noundef true, i1 noundef true)
+; CHECK-NEXT: [[CALL:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* noundef nonnull @[[GLOB1]], i8 noundef 1, i1 noundef true, i1 noundef true)
; CHECK-NEXT: call void @unknown_no_openmp()
; CHECK-NEXT: [[TMP0:%.*]] = icmp eq i32 [[CALL]], -1
; CHECK-NEXT: br i1 [[TMP0]], label [[MASTER:%.*]], label [[EXIT:%.*]]
@@ -185,7 +185,7 @@ declare void @unknown_no_openmp() "llvm.assume"="omp_no_openmp"
;
; CHECK-LABEL: define {{[^@]+}}@negative_qux_spmd() {
; CHECK-NEXT: entry:
-; CHECK-NEXT: [[CALL:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* noundef nonnull @[[GLOB1]], i1 noundef true, i1 noundef true, i1 noundef true)
+; CHECK-NEXT: [[CALL:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* noundef nonnull @[[GLOB1]], i8 noundef 2, i1 noundef true, i1 noundef true)
; CHECK-NEXT: call void @unknown_no_openmp()
; CHECK-NEXT: [[TMP0:%.*]] = icmp eq i32 [[CALL]], -1
; CHECK-NEXT: br i1 [[TMP0]], label [[MASTER:%.*]], label [[EXIT:%.*]]
diff --git a/llvm/test/Transforms/OpenMP/single_threaded_execution.ll b/llvm/test/Transforms/OpenMP/single_threaded_execution.ll
index e3b1ee297c9e7..7fb7d7292a7ec 100644
--- 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, i1 false, i1 false, i1 false)
+ %call = call i32 @__kmpc_target_init(%struct.ident_t* nonnull @1, i8 1, i1 false, i1 false)
%cmp = icmp eq i32 %call, -1
br i1 %cmp, label %if.then, label %if.else
if.then:
@@ -22,7 +22,7 @@ if.then:
if.else:
br label %if.end
if.end:
- call void @__kmpc_target_deinit(%struct.ident_t* null, i1 false, i1 true)
+ call void @__kmpc_target_deinit(%struct.ident_t* null, i8 1, i1 true)
ret void
}
@@ -105,9 +105,9 @@ declare i32 @llvm.amdgcn.workitem.id.x()
declare void @__kmpc_kernel_prepare_parallel(i8*)
-declare i32 @__kmpc_target_init(%struct.ident_t*, i1, i1, i1)
+declare i32 @__kmpc_target_init(%struct.ident_t*, i8, i1, i1)
-declare void @__kmpc_target_deinit(%struct.ident_t*, i1, i1)
+declare void @__kmpc_target_deinit(%struct.ident_t*, i8, i1)
attributes #0 = { cold noinline }
diff --git a/llvm/test/Transforms/OpenMP/spmdization.ll b/llvm/test/Transforms/OpenMP/spmdization.ll
index 3cf26fc88e746..baf61407e700c 100644
--- a/llvm/test/Transforms/OpenMP/spmdization.ll
+++ b/llvm/test/Transforms/OpenMP/spmdization.ll
@@ -147,14 +147,14 @@ define weak void @__omp_offloading_14_a34ca11_sequential_loop_l5() #0 {
; 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]], i1 true, i1 false, i1 false)
+; AMDGPU-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false)
; 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:
; AMDGPU-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR4:[0-9]+]]
; AMDGPU-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
; AMDGPU-NEXT: call void @__omp_outlined__(i32* noalias nocapture noundef nonnull readonly align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR4]]
-; AMDGPU-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 false)
+; AMDGPU-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 false)
; AMDGPU-NEXT: ret void
; AMDGPU: worker.exit:
; AMDGPU-NEXT: ret void
@@ -164,14 +164,14 @@ define weak void @__omp_offloading_14_a34ca11_sequential_loop_l5() #0 {
; 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]], i1 true, i1 false, i1 false)
+; NVPTX-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false)
; 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:
; NVPTX-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR4:[0-9]+]]
; NVPTX-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
; NVPTX-NEXT: call void @__omp_outlined__(i32* noalias nocapture noundef nonnull readonly align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR4]]
-; NVPTX-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 false)
+; NVPTX-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 false)
; NVPTX-NEXT: ret void
; NVPTX: worker.exit:
; NVPTX-NEXT: ret void
@@ -183,7 +183,7 @@ define weak void @__omp_offloading_14_a34ca11_sequential_loop_l5() #0 {
; 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]], i1 false, i1 false, i1 true)
+; AMDGPU-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true)
; 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:
@@ -220,7 +220,7 @@ define weak void @__omp_offloading_14_a34ca11_sequential_loop_l5() #0 {
; AMDGPU-DISABLED-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR4:[0-9]+]]
; AMDGPU-DISABLED-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
; AMDGPU-DISABLED-NEXT: call void @__omp_outlined__(i32* noalias nocapture noundef nonnull readonly align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR4]]
-; AMDGPU-DISABLED-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+; AMDGPU-DISABLED-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
; AMDGPU-DISABLED-NEXT: ret void
; AMDGPU-DISABLED: worker.exit:
; AMDGPU-DISABLED-NEXT: ret void
@@ -232,7 +232,7 @@ define weak void @__omp_offloading_14_a34ca11_sequential_loop_l5() #0 {
; 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]], i1 false, i1 false, i1 true)
+; NVPTX-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true)
; 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:
@@ -268,7 +268,7 @@ define weak void @__omp_offloading_14_a34ca11_sequential_loop_l5() #0 {
; NVPTX-DISABLED-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR4:[0-9]+]]
; NVPTX-DISABLED-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
; NVPTX-DISABLED-NEXT: call void @__omp_outlined__(i32* noalias nocapture noundef nonnull readonly align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR4]]
-; NVPTX-DISABLED-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+; NVPTX-DISABLED-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
; NVPTX-DISABLED-NEXT: ret void
; NVPTX-DISABLED: worker.exit:
; NVPTX-DISABLED-NEXT: ret void
@@ -277,7 +277,7 @@ entry:
%.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, i1 false, i1 true, i1 true)
+ %0 = call i32 @__kmpc_target_init(%struct.ident_t* @1, i8 1, i1 true, i1 true)
%exec_user_code = icmp eq i32 %0, -1
br i1 %exec_user_code, label %user_code.entry, label %worker.exit
@@ -285,14 +285,14 @@ user_code.entry: ; preds = %entry
%1 = call i32 @__kmpc_global_thread_num(%struct.ident_t* @1)
store i32 %1, i32* %.threadid_temp., align 4
call void @__omp_outlined__(i32* %.threadid_temp., i32* %.zero.addr) #3
- call void @__kmpc_target_deinit(%struct.ident_t* @1, i1 false, i1 true)
+ call void @__kmpc_target_deinit(%struct.ident_t* @1, i8 1, i1 true)
ret void
worker.exit: ; preds = %entry
ret void
}
-declare i32 @__kmpc_target_init(%struct.ident_t*, i1, i1, i1)
+declare i32 @__kmpc_target_init(%struct.ident_t*, i8, i1, i1)
define internal void @__omp_outlined__(i32* noalias %.global_tid., i32* noalias %.bound_tid.) #0 {
;
@@ -322,7 +322,7 @@ define internal void @__omp_outlined__(i32* noalias %.global_tid., i32* noalias
; AMDGPU-NEXT: store i32 [[INC]], i32* [[I]], align 4
; AMDGPU-NEXT: br label [[FOR_COND]], !llvm.loop [[LOOP13:![0-9]+]]
; AMDGPU: for.end:
-; AMDGPU-NEXT: call void @spmd_amenable() #[[ATTR6:[0-9]+]]
+; AMDGPU-NEXT: call void @spmd_amenable() #[[ATTR7:[0-9]+]]
; AMDGPU-NEXT: ret void
;
; NVPTX-LABEL: define {{[^@]+}}@__omp_outlined__
@@ -350,7 +350,7 @@ define internal void @__omp_outlined__(i32* noalias %.global_tid., i32* noalias
; NVPTX-NEXT: store i32 [[INC]], i32* [[I]], align 4
; NVPTX-NEXT: br label [[FOR_COND]], !llvm.loop [[LOOP13:![0-9]+]]
; NVPTX: for.end:
-; NVPTX-NEXT: call void @spmd_amenable() #[[ATTR6:[0-9]+]]
+; NVPTX-NEXT: call void @spmd_amenable() #[[ATTR7:[0-9]+]]
; NVPTX-NEXT: ret void
;
; AMDGPU-DISABLED-LABEL: define {{[^@]+}}@__omp_outlined__
@@ -378,7 +378,7 @@ define internal void @__omp_outlined__(i32* noalias %.global_tid., i32* noalias
; AMDGPU-DISABLED-NEXT: store i32 [[INC]], i32* [[I]], align 4
; AMDGPU-DISABLED-NEXT: br label [[FOR_COND]], !llvm.loop [[LOOP13:![0-9]+]]
; AMDGPU-DISABLED: for.end:
-; AMDGPU-DISABLED-NEXT: call void @spmd_amenable() #[[ATTR6:[0-9]+]]
+; AMDGPU-DISABLED-NEXT: call void @spmd_amenable() #[[ATTR7:[0-9]+]]
; AMDGPU-DISABLED-NEXT: ret void
;
; NVPTX-DISABLED-LABEL: define {{[^@]+}}@__omp_outlined__
@@ -406,7 +406,7 @@ define internal void @__omp_outlined__(i32* noalias %.global_tid., i32* noalias
; NVPTX-DISABLED-NEXT: store i32 [[INC]], i32* [[I]], align 4
; NVPTX-DISABLED-NEXT: br label [[FOR_COND]], !llvm.loop [[LOOP13:![0-9]+]]
; NVPTX-DISABLED: for.end:
-; NVPTX-DISABLED-NEXT: call void @spmd_amenable() #[[ATTR6:[0-9]+]]
+; NVPTX-DISABLED-NEXT: call void @spmd_amenable() #[[ATTR7:[0-9]+]]
; NVPTX-DISABLED-NEXT: ret void
;
entry:
@@ -450,7 +450,7 @@ define internal void @__omp_outlined__1(i32* noalias %.global_tid., i32* noalias
; AMDGPU-NEXT: entry:
; AMDGPU-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8
; AMDGPU-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8
-; AMDGPU-NEXT: call void @unknown() #[[ATTR7:[0-9]+]]
+; AMDGPU-NEXT: call void @unknown() #[[ATTR8:[0-9]+]]
; AMDGPU-NEXT: ret void
;
; NVPTX-LABEL: define {{[^@]+}}@__omp_outlined__1
@@ -458,7 +458,7 @@ define internal void @__omp_outlined__1(i32* noalias %.global_tid., i32* noalias
; NVPTX-NEXT: entry:
; NVPTX-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8
; NVPTX-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8
-; NVPTX-NEXT: call void @unknown() #[[ATTR7:[0-9]+]]
+; NVPTX-NEXT: call void @unknown() #[[ATTR8:[0-9]+]]
; NVPTX-NEXT: ret void
;
; AMDGPU-DISABLED-LABEL: define {{[^@]+}}@__omp_outlined__1
@@ -466,7 +466,7 @@ define internal void @__omp_outlined__1(i32* noalias %.global_tid., i32* noalias
; AMDGPU-DISABLED-NEXT: entry:
; AMDGPU-DISABLED-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8
; AMDGPU-DISABLED-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8
-; AMDGPU-DISABLED-NEXT: call void @unknown() #[[ATTR7:[0-9]+]]
+; AMDGPU-DISABLED-NEXT: call void @unknown() #[[ATTR8:[0-9]+]]
; AMDGPU-DISABLED-NEXT: ret void
;
; NVPTX-DISABLED-LABEL: define {{[^@]+}}@__omp_outlined__1
@@ -474,7 +474,7 @@ define internal void @__omp_outlined__1(i32* noalias %.global_tid., i32* noalias
; NVPTX-DISABLED-NEXT: entry:
; NVPTX-DISABLED-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8
; NVPTX-DISABLED-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8
-; NVPTX-DISABLED-NEXT: call void @unknown() #[[ATTR7:[0-9]+]]
+; NVPTX-DISABLED-NEXT: call void @unknown() #[[ATTR8:[0-9]+]]
; NVPTX-DISABLED-NEXT: ret void
;
entry:
@@ -568,7 +568,7 @@ declare void @spmd_amenable() #2
declare i32 @__kmpc_global_thread_num(%struct.ident_t*) #3
-declare void @__kmpc_target_deinit(%struct.ident_t*, i1, i1)
+declare void @__kmpc_target_deinit(%struct.ident_t*, i8, i1)
define weak void @__omp_offloading_14_a34ca11_sequential_loop_to_stack_var_l20() #0 {
;
@@ -578,14 +578,14 @@ define weak void @__omp_offloading_14_a34ca11_sequential_loop_to_stack_var_l20()
; 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]], i1 true, i1 false, i1 false)
+; AMDGPU-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false)
; 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:
; AMDGPU-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR4]]
; AMDGPU-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
; AMDGPU-NEXT: call void @__omp_outlined__2(i32* noalias nocapture noundef nonnull readonly align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR4]]
-; AMDGPU-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 false)
+; AMDGPU-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 false)
; AMDGPU-NEXT: ret void
; AMDGPU: worker.exit:
; AMDGPU-NEXT: ret void
@@ -595,14 +595,14 @@ define weak void @__omp_offloading_14_a34ca11_sequential_loop_to_stack_var_l20()
; 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]], i1 true, i1 false, i1 false)
+; NVPTX-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false)
; 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:
; NVPTX-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR4]]
; NVPTX-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
; NVPTX-NEXT: call void @__omp_outlined__2(i32* noalias nocapture noundef nonnull readonly align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR4]]
-; NVPTX-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 false)
+; NVPTX-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 false)
; NVPTX-NEXT: ret void
; NVPTX: worker.exit:
; NVPTX-NEXT: ret void
@@ -614,7 +614,7 @@ define weak void @__omp_offloading_14_a34ca11_sequential_loop_to_stack_var_l20()
; 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]], i1 false, i1 false, i1 true)
+; AMDGPU-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true)
; 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:
@@ -651,7 +651,7 @@ define weak void @__omp_offloading_14_a34ca11_sequential_loop_to_stack_var_l20()
; AMDGPU-DISABLED-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR4]]
; AMDGPU-DISABLED-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
; AMDGPU-DISABLED-NEXT: call void @__omp_outlined__2(i32* noalias nocapture noundef nonnull readonly align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR4]]
-; AMDGPU-DISABLED-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+; AMDGPU-DISABLED-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
; AMDGPU-DISABLED-NEXT: ret void
; AMDGPU-DISABLED: worker.exit:
; AMDGPU-DISABLED-NEXT: ret void
@@ -663,7 +663,7 @@ define weak void @__omp_offloading_14_a34ca11_sequential_loop_to_stack_var_l20()
; 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]], i1 false, i1 false, i1 true)
+; NVPTX-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true)
; 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:
@@ -699,7 +699,7 @@ define weak void @__omp_offloading_14_a34ca11_sequential_loop_to_stack_var_l20()
; NVPTX-DISABLED-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR4]]
; NVPTX-DISABLED-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
; NVPTX-DISABLED-NEXT: call void @__omp_outlined__2(i32* noalias nocapture noundef nonnull readonly align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR4]]
-; NVPTX-DISABLED-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+; NVPTX-DISABLED-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
; NVPTX-DISABLED-NEXT: ret void
; NVPTX-DISABLED: worker.exit:
; NVPTX-DISABLED-NEXT: ret void
@@ -708,7 +708,7 @@ entry:
%.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, i1 false, i1 true, i1 true)
+ %0 = call i32 @__kmpc_target_init(%struct.ident_t* @1, i8 1, i1 true, i1 true)
%exec_user_code = icmp eq i32 %0, -1
br i1 %exec_user_code, label %user_code.entry, label %worker.exit
@@ -716,7 +716,7 @@ user_code.entry: ; preds = %entry
%1 = call i32 @__kmpc_global_thread_num(%struct.ident_t* @1)
store i32 %1, i32* %.threadid_temp., align 4
call void @__omp_outlined__2(i32* %.threadid_temp., i32* %.zero.addr) #3
- call void @__kmpc_target_deinit(%struct.ident_t* @1, i1 false, i1 true)
+ call void @__kmpc_target_deinit(%struct.ident_t* @1, i8 1, i1 true)
ret void
worker.exit: ; preds = %entry
@@ -734,7 +734,7 @@ define internal void @__omp_outlined__2(i32* noalias %.global_tid., i32* noalias
; AMDGPU-NEXT: store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 8
; AMDGPU-NEXT: [[TMP0:%.*]] = alloca i8, i64 4, align 1
; AMDGPU-NEXT: [[X_ON_STACK:%.*]] = bitcast i8* [[TMP0]] to i32*
-; AMDGPU-NEXT: call void @use(i32* nocapture [[X_ON_STACK]]) #[[ATTR6]]
+; AMDGPU-NEXT: call void @use(i32* nocapture [[X_ON_STACK]]) #[[ATTR7]]
; AMDGPU-NEXT: store i32 0, i32* [[I]], align 4
; AMDGPU-NEXT: br label [[FOR_COND:%.*]]
; AMDGPU: for.cond:
@@ -752,7 +752,7 @@ define internal void @__omp_outlined__2(i32* noalias %.global_tid., i32* noalias
; AMDGPU-NEXT: store i32 [[INC]], i32* [[I]], align 4
; AMDGPU-NEXT: br label [[FOR_COND]], !llvm.loop [[LOOP15:![0-9]+]]
; AMDGPU: for.end:
-; AMDGPU-NEXT: call void @spmd_amenable() #[[ATTR6]]
+; AMDGPU-NEXT: call void @spmd_amenable() #[[ATTR7]]
; AMDGPU-NEXT: ret void
;
; NVPTX-LABEL: define {{[^@]+}}@__omp_outlined__2
@@ -765,7 +765,7 @@ define internal void @__omp_outlined__2(i32* noalias %.global_tid., i32* noalias
; NVPTX-NEXT: store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 8
; NVPTX-NEXT: [[TMP0:%.*]] = alloca i8, i64 4, align 1
; NVPTX-NEXT: [[X_ON_STACK:%.*]] = bitcast i8* [[TMP0]] to i32*
-; NVPTX-NEXT: call void @use(i32* nocapture [[X_ON_STACK]]) #[[ATTR6]]
+; NVPTX-NEXT: call void @use(i32* nocapture [[X_ON_STACK]]) #[[ATTR7]]
; NVPTX-NEXT: store i32 0, i32* [[I]], align 4
; NVPTX-NEXT: br label [[FOR_COND:%.*]]
; NVPTX: for.cond:
@@ -783,7 +783,7 @@ define internal void @__omp_outlined__2(i32* noalias %.global_tid., i32* noalias
; NVPTX-NEXT: store i32 [[INC]], i32* [[I]], align 4
; NVPTX-NEXT: br label [[FOR_COND]], !llvm.loop [[LOOP15:![0-9]+]]
; NVPTX: for.end:
-; NVPTX-NEXT: call void @spmd_amenable() #[[ATTR6]]
+; NVPTX-NEXT: call void @spmd_amenable() #[[ATTR7]]
; NVPTX-NEXT: ret void
;
; AMDGPU-DISABLED-LABEL: define {{[^@]+}}@__omp_outlined__2
@@ -796,7 +796,7 @@ define internal void @__omp_outlined__2(i32* noalias %.global_tid., i32* noalias
; AMDGPU-DISABLED-NEXT: store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 8
; AMDGPU-DISABLED-NEXT: [[TMP0:%.*]] = alloca i8, i64 4, align 1
; AMDGPU-DISABLED-NEXT: [[X_ON_STACK:%.*]] = bitcast i8* [[TMP0]] to i32*
-; AMDGPU-DISABLED-NEXT: call void @use(i32* nocapture [[X_ON_STACK]]) #[[ATTR6]]
+; AMDGPU-DISABLED-NEXT: call void @use(i32* nocapture [[X_ON_STACK]]) #[[ATTR7]]
; AMDGPU-DISABLED-NEXT: store i32 0, i32* [[I]], align 4
; AMDGPU-DISABLED-NEXT: br label [[FOR_COND:%.*]]
; AMDGPU-DISABLED: for.cond:
@@ -814,7 +814,7 @@ define internal void @__omp_outlined__2(i32* noalias %.global_tid., i32* noalias
; AMDGPU-DISABLED-NEXT: store i32 [[INC]], i32* [[I]], align 4
; AMDGPU-DISABLED-NEXT: br label [[FOR_COND]], !llvm.loop [[LOOP15:![0-9]+]]
; AMDGPU-DISABLED: for.end:
-; AMDGPU-DISABLED-NEXT: call void @spmd_amenable() #[[ATTR6]]
+; AMDGPU-DISABLED-NEXT: call void @spmd_amenable() #[[ATTR7]]
; AMDGPU-DISABLED-NEXT: ret void
;
; NVPTX-DISABLED-LABEL: define {{[^@]+}}@__omp_outlined__2
@@ -827,7 +827,7 @@ define internal void @__omp_outlined__2(i32* noalias %.global_tid., i32* noalias
; NVPTX-DISABLED-NEXT: store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 8
; NVPTX-DISABLED-NEXT: [[TMP0:%.*]] = alloca i8, i64 4, align 1
; NVPTX-DISABLED-NEXT: [[X_ON_STACK:%.*]] = bitcast i8* [[TMP0]] to i32*
-; NVPTX-DISABLED-NEXT: call void @use(i32* nocapture [[X_ON_STACK]]) #[[ATTR6]]
+; NVPTX-DISABLED-NEXT: call void @use(i32* nocapture [[X_ON_STACK]]) #[[ATTR7]]
; NVPTX-DISABLED-NEXT: store i32 0, i32* [[I]], align 4
; NVPTX-DISABLED-NEXT: br label [[FOR_COND:%.*]]
; NVPTX-DISABLED: for.cond:
@@ -845,7 +845,7 @@ define internal void @__omp_outlined__2(i32* noalias %.global_tid., i32* noalias
; NVPTX-DISABLED-NEXT: store i32 [[INC]], i32* [[I]], align 4
; NVPTX-DISABLED-NEXT: br label [[FOR_COND]], !llvm.loop [[LOOP15:![0-9]+]]
; NVPTX-DISABLED: for.end:
-; NVPTX-DISABLED-NEXT: call void @spmd_amenable() #[[ATTR6]]
+; NVPTX-DISABLED-NEXT: call void @spmd_amenable() #[[ATTR7]]
; NVPTX-DISABLED-NEXT: ret void
;
entry:
@@ -897,7 +897,7 @@ define internal void @__omp_outlined__3(i32* noalias %.global_tid., i32* noalias
; AMDGPU-NEXT: entry:
; AMDGPU-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8
; AMDGPU-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8
-; AMDGPU-NEXT: call void @unknown() #[[ATTR7]]
+; AMDGPU-NEXT: call void @unknown() #[[ATTR8]]
; AMDGPU-NEXT: ret void
;
; NVPTX-LABEL: define {{[^@]+}}@__omp_outlined__3
@@ -905,7 +905,7 @@ define internal void @__omp_outlined__3(i32* noalias %.global_tid., i32* noalias
; NVPTX-NEXT: entry:
; NVPTX-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8
; NVPTX-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8
-; NVPTX-NEXT: call void @unknown() #[[ATTR7]]
+; NVPTX-NEXT: call void @unknown() #[[ATTR8]]
; NVPTX-NEXT: ret void
;
; AMDGPU-DISABLED-LABEL: define {{[^@]+}}@__omp_outlined__3
@@ -913,7 +913,7 @@ define internal void @__omp_outlined__3(i32* noalias %.global_tid., i32* noalias
; AMDGPU-DISABLED-NEXT: entry:
; AMDGPU-DISABLED-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8
; AMDGPU-DISABLED-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8
-; AMDGPU-DISABLED-NEXT: call void @unknown() #[[ATTR7]]
+; AMDGPU-DISABLED-NEXT: call void @unknown() #[[ATTR8]]
; AMDGPU-DISABLED-NEXT: ret void
;
; NVPTX-DISABLED-LABEL: define {{[^@]+}}@__omp_outlined__3
@@ -921,7 +921,7 @@ define internal void @__omp_outlined__3(i32* noalias %.global_tid., i32* noalias
; NVPTX-DISABLED-NEXT: entry:
; NVPTX-DISABLED-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8
; NVPTX-DISABLED-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8
-; NVPTX-DISABLED-NEXT: call void @unknown() #[[ATTR7]]
+; NVPTX-DISABLED-NEXT: call void @unknown() #[[ATTR8]]
; NVPTX-DISABLED-NEXT: ret void
;
entry:
@@ -1015,14 +1015,14 @@ define weak void @__omp_offloading_14_a34ca11_sequential_loop_to_shared_var_l35(
; 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]], i1 true, i1 false, i1 false)
+; AMDGPU-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false)
; 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:
; AMDGPU-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR4]]
; AMDGPU-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
; AMDGPU-NEXT: call void @__omp_outlined__4(i32* noalias nocapture noundef nonnull readonly align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR4]]
-; AMDGPU-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 false)
+; AMDGPU-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 false)
; AMDGPU-NEXT: ret void
; AMDGPU: worker.exit:
; AMDGPU-NEXT: ret void
@@ -1032,14 +1032,14 @@ define weak void @__omp_offloading_14_a34ca11_sequential_loop_to_shared_var_l35(
; 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]], i1 true, i1 false, i1 false)
+; NVPTX-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false)
; 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:
; NVPTX-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR4]]
; NVPTX-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
; NVPTX-NEXT: call void @__omp_outlined__4(i32* noalias nocapture noundef nonnull readonly align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR4]]
-; NVPTX-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 false)
+; NVPTX-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 false)
; NVPTX-NEXT: ret void
; NVPTX: worker.exit:
; NVPTX-NEXT: ret void
@@ -1051,7 +1051,7 @@ define weak void @__omp_offloading_14_a34ca11_sequential_loop_to_shared_var_l35(
; 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]], i1 false, i1 false, i1 true)
+; AMDGPU-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true)
; 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:
@@ -1088,7 +1088,7 @@ define weak void @__omp_offloading_14_a34ca11_sequential_loop_to_shared_var_l35(
; AMDGPU-DISABLED-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR4]]
; AMDGPU-DISABLED-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
; AMDGPU-DISABLED-NEXT: call void @__omp_outlined__4(i32* noalias nocapture noundef nonnull readonly align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR4]]
-; AMDGPU-DISABLED-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+; AMDGPU-DISABLED-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
; AMDGPU-DISABLED-NEXT: ret void
; AMDGPU-DISABLED: worker.exit:
; AMDGPU-DISABLED-NEXT: ret void
@@ -1100,7 +1100,7 @@ define weak void @__omp_offloading_14_a34ca11_sequential_loop_to_shared_var_l35(
; 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]], i1 false, i1 false, i1 true)
+; NVPTX-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true)
; 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:
@@ -1136,7 +1136,7 @@ define weak void @__omp_offloading_14_a34ca11_sequential_loop_to_shared_var_l35(
; NVPTX-DISABLED-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR4]]
; NVPTX-DISABLED-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
; NVPTX-DISABLED-NEXT: call void @__omp_outlined__4(i32* noalias nocapture noundef nonnull readonly align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR4]]
-; NVPTX-DISABLED-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+; NVPTX-DISABLED-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
; NVPTX-DISABLED-NEXT: ret void
; NVPTX-DISABLED: worker.exit:
; NVPTX-DISABLED-NEXT: ret void
@@ -1145,7 +1145,7 @@ entry:
%.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, i1 false, i1 true, i1 true)
+ %0 = call i32 @__kmpc_target_init(%struct.ident_t* @1, i8 1, i1 true, i1 true)
%exec_user_code = icmp eq i32 %0, -1
br i1 %exec_user_code, label %user_code.entry, label %worker.exit
@@ -1153,7 +1153,7 @@ user_code.entry: ; preds = %entry
%1 = call i32 @__kmpc_global_thread_num(%struct.ident_t* @1)
store i32 %1, i32* %.threadid_temp., align 4
call void @__omp_outlined__4(i32* %.threadid_temp., i32* %.zero.addr) #3
- call void @__kmpc_target_deinit(%struct.ident_t* @1, i1 false, i1 true)
+ call void @__kmpc_target_deinit(%struct.ident_t* @1, i8 1, i1 true)
ret void
worker.exit: ; preds = %entry
@@ -1190,7 +1190,7 @@ define internal void @__omp_outlined__4(i32* noalias %.global_tid., i32* noalias
; AMDGPU-NEXT: store i32 [[INC]], i32* [[I]], align 4
; AMDGPU-NEXT: br label [[FOR_COND]], !llvm.loop [[LOOP16:![0-9]+]]
; AMDGPU: for.end:
-; AMDGPU-NEXT: call void @spmd_amenable() #[[ATTR6]]
+; AMDGPU-NEXT: call void @spmd_amenable() #[[ATTR7]]
; AMDGPU-NEXT: ret void
;
; NVPTX-LABEL: define {{[^@]+}}@__omp_outlined__4
@@ -1220,7 +1220,7 @@ define internal void @__omp_outlined__4(i32* noalias %.global_tid., i32* noalias
; NVPTX-NEXT: store i32 [[INC]], i32* [[I]], align 4
; NVPTX-NEXT: br label [[FOR_COND]], !llvm.loop [[LOOP16:![0-9]+]]
; NVPTX: for.end:
-; NVPTX-NEXT: call void @spmd_amenable() #[[ATTR6]]
+; NVPTX-NEXT: call void @spmd_amenable() #[[ATTR7]]
; NVPTX-NEXT: ret void
;
; AMDGPU-DISABLED-LABEL: define {{[^@]+}}@__omp_outlined__4
@@ -1250,7 +1250,7 @@ define internal void @__omp_outlined__4(i32* noalias %.global_tid., i32* noalias
; AMDGPU-DISABLED-NEXT: store i32 [[INC]], i32* [[I]], align 4
; AMDGPU-DISABLED-NEXT: br label [[FOR_COND]], !llvm.loop [[LOOP16:![0-9]+]]
; AMDGPU-DISABLED: for.end:
-; AMDGPU-DISABLED-NEXT: call void @spmd_amenable() #[[ATTR6]]
+; AMDGPU-DISABLED-NEXT: call void @spmd_amenable() #[[ATTR7]]
; AMDGPU-DISABLED-NEXT: ret void
;
; NVPTX-DISABLED-LABEL: define {{[^@]+}}@__omp_outlined__4
@@ -1280,7 +1280,7 @@ define internal void @__omp_outlined__4(i32* noalias %.global_tid., i32* noalias
; NVPTX-DISABLED-NEXT: store i32 [[INC]], i32* [[I]], align 4
; NVPTX-DISABLED-NEXT: br label [[FOR_COND]], !llvm.loop [[LOOP16:![0-9]+]]
; NVPTX-DISABLED: for.end:
-; NVPTX-DISABLED-NEXT: call void @spmd_amenable() #[[ATTR6]]
+; NVPTX-DISABLED-NEXT: call void @spmd_amenable() #[[ATTR7]]
; NVPTX-DISABLED-NEXT: ret void
;
entry:
@@ -1335,7 +1335,7 @@ define internal void @__omp_outlined__5(i32* noalias %.global_tid., i32* noalias
; AMDGPU-NEXT: [[TMP0:%.*]] = load i32, i32* [[X]], align 4
; AMDGPU-NEXT: [[INC:%.*]] = add nsw i32 [[TMP0]], 1
; AMDGPU-NEXT: store i32 [[INC]], i32* [[X]], align 4
-; AMDGPU-NEXT: call void @unknown() #[[ATTR7]]
+; AMDGPU-NEXT: call void @unknown() #[[ATTR8]]
; AMDGPU-NEXT: ret void
;
; NVPTX-LABEL: define {{[^@]+}}@__omp_outlined__5
@@ -1348,7 +1348,7 @@ define internal void @__omp_outlined__5(i32* noalias %.global_tid., i32* noalias
; NVPTX-NEXT: [[TMP0:%.*]] = load i32, i32* [[X]], align 4
; NVPTX-NEXT: [[INC:%.*]] = add nsw i32 [[TMP0]], 1
; NVPTX-NEXT: store i32 [[INC]], i32* [[X]], align 4
-; NVPTX-NEXT: call void @unknown() #[[ATTR7]]
+; NVPTX-NEXT: call void @unknown() #[[ATTR8]]
; NVPTX-NEXT: ret void
;
; AMDGPU-DISABLED-LABEL: define {{[^@]+}}@__omp_outlined__5
@@ -1361,7 +1361,7 @@ define internal void @__omp_outlined__5(i32* noalias %.global_tid., i32* noalias
; AMDGPU-DISABLED-NEXT: [[TMP0:%.*]] = load i32, i32* [[X]], align 4
; AMDGPU-DISABLED-NEXT: [[INC:%.*]] = add nsw i32 [[TMP0]], 1
; AMDGPU-DISABLED-NEXT: store i32 [[INC]], i32* [[X]], align 4
-; AMDGPU-DISABLED-NEXT: call void @unknown() #[[ATTR7]]
+; AMDGPU-DISABLED-NEXT: call void @unknown() #[[ATTR8]]
; AMDGPU-DISABLED-NEXT: ret void
;
; NVPTX-DISABLED-LABEL: define {{[^@]+}}@__omp_outlined__5
@@ -1374,7 +1374,7 @@ define internal void @__omp_outlined__5(i32* noalias %.global_tid., i32* noalias
; NVPTX-DISABLED-NEXT: [[TMP0:%.*]] = load i32, i32* [[X]], align 4
; NVPTX-DISABLED-NEXT: [[INC:%.*]] = add nsw i32 [[TMP0]], 1
; NVPTX-DISABLED-NEXT: store i32 [[INC]], i32* [[X]], align 4
-; NVPTX-DISABLED-NEXT: call void @unknown() #[[ATTR7]]
+; NVPTX-DISABLED-NEXT: call void @unknown() #[[ATTR8]]
; NVPTX-DISABLED-NEXT: ret void
;
entry:
@@ -1492,14 +1492,14 @@ define weak void @__omp_offloading_14_a34ca11_sequential_loop_to_shared_var_guar
; 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]], i1 true, i1 false, i1 false)
+; AMDGPU-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false)
; 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:
; AMDGPU-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR4]]
; AMDGPU-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
; AMDGPU-NEXT: call void @__omp_outlined__6(i32* noalias nocapture noundef nonnull readonly align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR4]]
-; AMDGPU-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 false)
+; AMDGPU-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 false)
; AMDGPU-NEXT: ret void
; AMDGPU: worker.exit:
; AMDGPU-NEXT: ret void
@@ -1509,14 +1509,14 @@ define weak void @__omp_offloading_14_a34ca11_sequential_loop_to_shared_var_guar
; 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]], i1 true, i1 false, i1 false)
+; NVPTX-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 2, i1 false, i1 false)
; 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:
; NVPTX-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR4]]
; NVPTX-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
; NVPTX-NEXT: call void @__omp_outlined__6(i32* noalias nocapture noundef nonnull readonly align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR4]]
-; NVPTX-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 true, i1 false)
+; NVPTX-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 2, i1 false)
; NVPTX-NEXT: ret void
; NVPTX: worker.exit:
; NVPTX-NEXT: ret void
@@ -1528,7 +1528,7 @@ define weak void @__omp_offloading_14_a34ca11_sequential_loop_to_shared_var_guar
; 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]], i1 false, i1 false, i1 true)
+; AMDGPU-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true)
; 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:
@@ -1565,7 +1565,7 @@ define weak void @__omp_offloading_14_a34ca11_sequential_loop_to_shared_var_guar
; AMDGPU-DISABLED-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR4]]
; AMDGPU-DISABLED-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
; AMDGPU-DISABLED-NEXT: call void @__omp_outlined__6(i32* noalias nocapture noundef nonnull readonly align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR4]]
-; AMDGPU-DISABLED-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+; AMDGPU-DISABLED-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
; AMDGPU-DISABLED-NEXT: ret void
; AMDGPU-DISABLED: worker.exit:
; AMDGPU-DISABLED-NEXT: ret void
@@ -1577,7 +1577,7 @@ define weak void @__omp_offloading_14_a34ca11_sequential_loop_to_shared_var_guar
; 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]], i1 false, i1 false, i1 true)
+; NVPTX-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true)
; 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:
@@ -1613,7 +1613,7 @@ define weak void @__omp_offloading_14_a34ca11_sequential_loop_to_shared_var_guar
; NVPTX-DISABLED-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR4]]
; NVPTX-DISABLED-NEXT: store i32 [[TMP1]], i32* [[DOTTHREADID_TEMP_]], align 4
; NVPTX-DISABLED-NEXT: call void @__omp_outlined__6(i32* noalias nocapture noundef nonnull readonly align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR4]]
-; NVPTX-DISABLED-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+; NVPTX-DISABLED-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
; NVPTX-DISABLED-NEXT: ret void
; NVPTX-DISABLED: worker.exit:
; NVPTX-DISABLED-NEXT: ret void
@@ -1622,7 +1622,7 @@ entry:
%.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, i1 false, i1 true, i1 true)
+ %0 = call i32 @__kmpc_target_init(%struct.ident_t* @1, i8 1, i1 true, i1 true)
%exec_user_code = icmp eq i32 %0, -1
br i1 %exec_user_code, label %user_code.entry, label %worker.exit
@@ -1630,7 +1630,7 @@ user_code.entry: ; preds = %entry
%1 = call i32 @__kmpc_global_thread_num(%struct.ident_t* @1)
store i32 %1, i32* %.threadid_temp., align 4
call void @__omp_outlined__6(i32* %.threadid_temp., i32* %.zero.addr) #3
- call void @__kmpc_target_deinit(%struct.ident_t* @1, i1 false, i1 true)
+ call void @__kmpc_target_deinit(%struct.ident_t* @1, i8 1, i1 true)
ret void
worker.exit: ; preds = %entry
@@ -1682,7 +1682,7 @@ define internal void @__omp_outlined__6(i32* noalias %.global_tid., i32* noalias
; AMDGPU-NEXT: store i32 [[INC]], i32* [[I]], align 4
; AMDGPU-NEXT: br label [[FOR_COND]], !llvm.loop [[LOOP17:![0-9]+]]
; AMDGPU: for.end:
-; AMDGPU-NEXT: call void @spmd_amenable() #[[ATTR6]]
+; AMDGPU-NEXT: call void @spmd_amenable() #[[ATTR7]]
; AMDGPU-NEXT: ret void
;
; NVPTX-LABEL: define {{[^@]+}}@__omp_outlined__6
@@ -1727,7 +1727,7 @@ define internal void @__omp_outlined__6(i32* noalias %.global_tid., i32* noalias
; NVPTX-NEXT: store i32 [[INC]], i32* [[I]], align 4
; NVPTX-NEXT: br label [[FOR_COND]], !llvm.loop [[LOOP17:![0-9]+]]
; NVPTX: for.end:
-; NVPTX-NEXT: call void @spmd_amenable() #[[ATTR6]]
+; NVPTX-NEXT: call void @spmd_amenable() #[[ATTR7]]
; NVPTX-NEXT: ret void
;
; AMDGPU-DISABLED-LABEL: define {{[^@]+}}@__omp_outlined__6
@@ -1759,7 +1759,7 @@ define internal void @__omp_outlined__6(i32* noalias %.global_tid., i32* noalias
; AMDGPU-DISABLED-NEXT: store i32 [[INC]], i32* [[I]], align 4
; AMDGPU-DISABLED-NEXT: br label [[FOR_COND]], !llvm.loop [[LOOP17:![0-9]+]]
; AMDGPU-DISABLED: for.end:
-; AMDGPU-DISABLED-NEXT: call void @spmd_amenable() #[[ATTR6]]
+; AMDGPU-DISABLED-NEXT: call void @spmd_amenable() #[[ATTR7]]
; AMDGPU-DISABLED-NEXT: ret void
;
; NVPTX-DISABLED-LABEL: define {{[^@]+}}@__omp_outlined__6
@@ -1791,7 +1791,7 @@ define internal void @__omp_outlined__6(i32* noalias %.global_tid., i32* noalias
; NVPTX-DISABLED-NEXT: store i32 [[INC]], i32* [[I]], align 4
; NVPTX-DISABLED-NEXT: br label [[FOR_COND]], !llvm.loop [[LOOP17:![0-9]+]]
; NVPTX-DISABLED: for.end:
-; NVPTX-DISABLED-NEXT: call void @spmd_amenable() #[[ATTR6]]
+; NVPTX-DISABLED-NEXT: call void @spmd_amenable() #[[ATTR7]]
; NVPTX-DISABLED-NEXT: ret void
;
entry:
@@ -1847,7 +1847,7 @@ define internal void @__omp_outlined__7(i32* noalias %.global_tid., i32* noalias
; AMDGPU-NEXT: [[TMP0:%.*]] = load i32, i32* [[X]], align 4
; AMDGPU-NEXT: [[INC:%.*]] = add nsw i32 [[TMP0]], 1
; AMDGPU-NEXT: store i32 [[INC]], i32* [[X]], align 4
-; AMDGPU-NEXT: call void @unknown() #[[ATTR7]]
+; AMDGPU-NEXT: call void @unknown() #[[ATTR8]]
; AMDGPU-NEXT: ret void
;
; NVPTX-LABEL: define {{[^@]+}}@__omp_outlined__7
@@ -1860,7 +1860,7 @@ define internal void @__omp_outlined__7(i32* noalias %.global_tid., i32* noalias
; NVPTX-NEXT: [[TMP0:%.*]] = load i32, i32* [[X]], align 4
; NVPTX-NEXT: [[INC:%.*]] = add nsw i32 [[TMP0]], 1
; NVPTX-NEXT: store i32 [[INC]], i32* [[X]], align 4
-; NVPTX-NEXT: call void @unknown() #[[ATTR7]]
+; NVPTX-NEXT: call void @unknown() #[[ATTR8]]
; NVPTX-NEXT: ret void
;
; AMDGPU-DISABLED-LABEL: define {{[^@]+}}@__omp_outlined__7
@@ -1873,7 +1873,7 @@ define internal void @__omp_outlined__7(i32* noalias %.global_tid., i32* noalias
; AMDGPU-DISABLED-NEXT: [[TMP0:%.*]] = load i32, i32* [[X]], align 4
; AMDGPU-DISABLED-NEXT: [[INC:%.*]] = add nsw i32 [[TMP0]], 1
; AMDGPU-DISABLED-NEXT: store i32 [[INC]], i32* [[X]], align 4
-; AMDGPU-DISABLED-NEXT: call void @unknown() #[[ATTR7]]
+; AMDGPU-DISABLED-NEXT: call void @unknown() #[[ATTR8]]
; AMDGPU-DISABLED-NEXT: ret void
;
; NVPTX-DISABLED-LABEL: define {{[^@]+}}@__omp_outlined__7
@@ -1886,7 +1886,7 @@ define internal void @__omp_outlined__7(i32* noalias %.global_tid., i32* noalias
; NVPTX-DISABLED-NEXT: [[TMP0:%.*]] = load i32, i32* [[X]], align 4
; NVPTX-DISABLED-NEXT: [[INC:%.*]] = add nsw i32 [[TMP0]], 1
; NVPTX-DISABLED-NEXT: store i32 [[INC]], i32* [[X]], align 4
-; NVPTX-DISABLED-NEXT: call void @unknown() #[[ATTR7]]
+; NVPTX-DISABLED-NEXT: call void @unknown() #[[ATTR8]]
; NVPTX-DISABLED-NEXT: ret void
;
entry:
@@ -2005,7 +2005,7 @@ define weak void @__omp_offloading_14_a34ca11_do_not_spmdize_target_l65() #0 {
; 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]], i1 false, i1 false, i1 true)
+; AMDGPU-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true)
; 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:
@@ -2035,7 +2035,7 @@ define weak void @__omp_offloading_14_a34ca11_do_not_spmdize_target_l65() #0 {
; AMDGPU: user_code.entry:
; AMDGPU-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR4]]
; AMDGPU-NEXT: call void @__omp_outlined__8(i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR4]]
-; AMDGPU-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+; AMDGPU-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
; AMDGPU-NEXT: ret void
; AMDGPU: worker.exit:
; AMDGPU-NEXT: ret void
@@ -2046,7 +2046,7 @@ define weak void @__omp_offloading_14_a34ca11_do_not_spmdize_target_l65() #0 {
; 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]], i1 false, i1 false, i1 true)
+; NVPTX-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true)
; 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:
@@ -2075,7 +2075,7 @@ define weak void @__omp_offloading_14_a34ca11_do_not_spmdize_target_l65() #0 {
; NVPTX: user_code.entry:
; NVPTX-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR4]]
; NVPTX-NEXT: call void @__omp_outlined__8(i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR4]]
-; NVPTX-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+; NVPTX-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
; NVPTX-NEXT: ret void
; NVPTX: worker.exit:
; NVPTX-NEXT: ret void
@@ -2087,7 +2087,7 @@ define weak void @__omp_offloading_14_a34ca11_do_not_spmdize_target_l65() #0 {
; 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]], i1 false, i1 false, i1 true)
+; AMDGPU-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true)
; 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:
@@ -2117,7 +2117,7 @@ define weak void @__omp_offloading_14_a34ca11_do_not_spmdize_target_l65() #0 {
; AMDGPU-DISABLED: user_code.entry:
; AMDGPU-DISABLED-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR4]]
; AMDGPU-DISABLED-NEXT: call void @__omp_outlined__8(i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR4]]
-; AMDGPU-DISABLED-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+; AMDGPU-DISABLED-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
; AMDGPU-DISABLED-NEXT: ret void
; AMDGPU-DISABLED: worker.exit:
; AMDGPU-DISABLED-NEXT: ret void
@@ -2129,7 +2129,7 @@ define weak void @__omp_offloading_14_a34ca11_do_not_spmdize_target_l65() #0 {
; 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]], i1 false, i1 false, i1 true)
+; NVPTX-DISABLED-NEXT: [[TMP0:%.*]] = call i32 @__kmpc_target_init(%struct.ident_t* @[[GLOB1]], i8 1, i1 false, i1 true)
; 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:
@@ -2158,7 +2158,7 @@ define weak void @__omp_offloading_14_a34ca11_do_not_spmdize_target_l65() #0 {
; NVPTX-DISABLED: user_code.entry:
; NVPTX-DISABLED-NEXT: [[TMP1:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1]]) #[[ATTR4]]
; NVPTX-DISABLED-NEXT: call void @__omp_outlined__8(i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTTHREADID_TEMP_]], i32* noalias nocapture noundef nonnull readnone align 4 dereferenceable(4) [[DOTZERO_ADDR]]) #[[ATTR4]]
-; NVPTX-DISABLED-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i1 false, i1 true)
+; NVPTX-DISABLED-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* @[[GLOB1]], i8 1, i1 true)
; NVPTX-DISABLED-NEXT: ret void
; NVPTX-DISABLED: worker.exit:
; NVPTX-DISABLED-NEXT: ret void
@@ -2167,7 +2167,7 @@ entry:
%.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, i1 false, i1 true, i1 true)
+ %0 = call i32 @__kmpc_target_init(%struct.ident_t* @1, i8 1, i1 true, i1 true)
%exec_user_code = icmp eq i32 %0, -1
br i1 %exec_user_code, label %user_code.entry, label %worker.exit
@@ -2175,7 +2175,7 @@ user_code.entry: ; preds = %entry
%1 = call i32 @__kmpc_global_thread_num(%struct.ident_t* @1)
store i32 %1, i32* %.threadid_temp., align 4
call void @__omp_outlined__8(i32* %.threadid_temp., i32* %.zero.addr) #3
- call void @__kmpc_target_deinit(%struct.ident_t* @1, i1 false, i1 true)
+ call void @__kmpc_target_deinit(%struct.ident_t* @1, i8 1, i1 true)
ret void
worker.exit: ; preds = %entry
@@ -2190,7 +2190,7 @@ define internal void @__omp_outlined__8(i32* noalias %.global_tid., i32* noalias
; AMDGPU-NEXT: entry:
; AMDGPU-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8
; AMDGPU-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8
-; AMDGPU-NEXT: call void @unknown() #[[ATTR7]]
+; AMDGPU-NEXT: call void @unknown() #[[ATTR8]]
; AMDGPU-NEXT: ret void
;
; NVPTX-LABEL: define {{[^@]+}}@__omp_outlined__8
@@ -2198,7 +2198,7 @@ define internal void @__omp_outlined__8(i32* noalias %.global_tid., i32* noalias
; NVPTX-NEXT: entry:
; NVPTX-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8
; NVPTX-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8
-; NVPTX-NEXT: call void @unknown() #[[ATTR7]]
+; NVPTX-NEXT: call void @unknown() #[[ATTR8]]
; NVPTX-NEXT: ret void
;
; AMDGPU-DISABLED-LABEL: define {{[^@]+}}@__omp_outlined__8
@@ -2206,7 +2206,7 @@ define internal void @__omp_outlined__8(i32* noalias %.global_tid., i32* noalias
; AMDGPU-DISABLED-NEXT: entry:
; AMDGPU-DISABLED-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8
; AMDGPU-DISABLED-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8
-; AMDGPU-DISABLED-NEXT: call void @unknown() #[[ATTR7]]
+; AMDGPU-DISABLED-NEXT: call void @unknown() #[[ATTR8]]
; AMDGPU-DISABLED-NEXT: ret void
;
; NVPTX-DISABLED-LABEL: define {{[^@]+}}@__omp_outlined__8
@@ -2214,7 +2214,7 @@ define internal void @__omp_outlined__8(i32* noalias %.global_tid., i32* noalias
; NVPTX-DISABLED-NEXT: entry:
; NVPTX-DISABLED-NEXT: [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8
; NVPTX-DISABLED-NEXT: [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8
-; NVPTX-DISABLED-NEXT: call void @unknown() #[[ATTR7]]
+; NVPTX-DISABLED-NEXT: call void @unknown() #[[ATTR8]]
; NVPTX-DISABLED-NEXT: ret void
;
entry:
@@ -2261,36 +2261,40 @@ attributes #5 = { convergent }
; AMDGPU: attributes #[[ATTR2:[0-9]+]] = { alwaysinline }
; AMDGPU: attributes #[[ATTR3:[0-9]+]] = { convergent "frame-pointer"="none" "llvm.assume"="ompx_spmd_amenable" "no-trapping-math"="true" "stack-protector-buffer-size"="8" "target-features"="+ptx32,+sm_20" }
; AMDGPU: attributes #[[ATTR4]] = { nounwind }
-; AMDGPU: attributes #[[ATTR5:[0-9]+]] = { convergent nounwind }
-; AMDGPU: attributes #[[ATTR6]] = { convergent "llvm.assume"="ompx_spmd_amenable" }
-; AMDGPU: attributes #[[ATTR7]] = { convergent }
+; AMDGPU: attributes #[[ATTR5:[0-9]+]] = { nosync nounwind }
+; AMDGPU: attributes #[[ATTR6:[0-9]+]] = { convergent nounwind }
+; AMDGPU: attributes #[[ATTR7]] = { convergent "llvm.assume"="ompx_spmd_amenable" }
+; AMDGPU: attributes #[[ATTR8]] = { convergent }
;.
; NVPTX: attributes #[[ATTR0]] = { convergent noinline norecurse nounwind "frame-pointer"="none" "min-legal-vector-width"="0" "no-trapping-math"="true" "stack-protector-buffer-size"="8" "target-features"="+ptx32,+sm_20" }
; NVPTX: attributes #[[ATTR1:[0-9]+]] = { convergent "frame-pointer"="none" "no-trapping-math"="true" "stack-protector-buffer-size"="8" "target-features"="+ptx32,+sm_20" }
; NVPTX: attributes #[[ATTR2:[0-9]+]] = { alwaysinline }
; NVPTX: attributes #[[ATTR3:[0-9]+]] = { convergent "frame-pointer"="none" "llvm.assume"="ompx_spmd_amenable" "no-trapping-math"="true" "stack-protector-buffer-size"="8" "target-features"="+ptx32,+sm_20" }
; NVPTX: attributes #[[ATTR4]] = { nounwind }
-; NVPTX: attributes #[[ATTR5:[0-9]+]] = { convergent nounwind }
-; NVPTX: attributes #[[ATTR6]] = { convergent "llvm.assume"="ompx_spmd_amenable" }
-; NVPTX: attributes #[[ATTR7]] = { convergent }
+; NVPTX: attributes #[[ATTR5:[0-9]+]] = { nosync nounwind }
+; NVPTX: attributes #[[ATTR6:[0-9]+]] = { convergent nounwind }
+; NVPTX: attributes #[[ATTR7]] = { convergent "llvm.assume"="ompx_spmd_amenable" }
+; NVPTX: attributes #[[ATTR8]] = { convergent }
;.
; AMDGPU-DISABLED: attributes #[[ATTR0]] = { convergent noinline norecurse nounwind "frame-pointer"="none" "min-legal-vector-width"="0" "no-trapping-math"="true" "stack-protector-buffer-size"="8" "target-features"="+ptx32,+sm_20" }
; AMDGPU-DISABLED: attributes #[[ATTR1:[0-9]+]] = { convergent "frame-pointer"="none" "no-trapping-math"="true" "stack-protector-buffer-size"="8" "target-features"="+ptx32,+sm_20" }
; AMDGPU-DISABLED: attributes #[[ATTR2:[0-9]+]] = { alwaysinline }
; AMDGPU-DISABLED: attributes #[[ATTR3:[0-9]+]] = { convergent "frame-pointer"="none" "llvm.assume"="ompx_spmd_amenable" "no-trapping-math"="true" "stack-protector-buffer-size"="8" "target-features"="+ptx32,+sm_20" }
; AMDGPU-DISABLED: attributes #[[ATTR4]] = { nounwind }
-; AMDGPU-DISABLED: attributes #[[ATTR5:[0-9]+]] = { convergent nounwind }
-; AMDGPU-DISABLED: attributes #[[ATTR6]] = { convergent "llvm.assume"="ompx_spmd_amenable" }
-; AMDGPU-DISABLED: attributes #[[ATTR7]] = { convergent }
+; AMDGPU-DISABLED: attributes #[[ATTR5:[0-9]+]] = { nosync nounwind }
+; AMDGPU-DISABLED: attributes #[[ATTR6:[0-9]+]] = { convergent nounwind }
+; AMDGPU-DISABLED: attributes #[[ATTR7]] = { convergent "llvm.assume"="ompx_spmd_amenable" }
+; AMDGPU-DISABLED: attributes #[[ATTR8]] = { convergent }
;.
; NVPTX-DISABLED: attributes #[[ATTR0]] = { convergent noinline norecurse nounwind "frame-pointer"="none" "min-legal-vector-width"="0" "no-trapping-math"="true" "stack-protector-buffer-size"="8" "target-features"="+ptx32,+sm_20" }
; NVPTX-DISABLED: attributes #[[ATTR1:[0-9]+]] = { convergent "frame-pointer"="none" "no-trapping-math"="true" "stack-protector-buffer-size"="8" "target-features"="+ptx32,+sm_20" }
; NVPTX-DISABLED: attributes #[[ATTR2:[0-9]+]] = { alwaysinline }
; NVPTX-DISABLED: attributes #[[ATTR3:[0-9]+]] = { convergent "frame-pointer"="none" "llvm.assume"="ompx_spmd_amenable" "no-trapping-math"="true" "stack-protector-buffer-size"="8" "target-features"="+ptx32,+sm_20" }
; NVPTX-DISABLED: attributes #[[ATTR4]] = { nounwind }
-; NVPTX-DISABLED: attributes #[[ATTR5:[0-9]+]] = { convergent nounwind }
-; NVPTX-DISABLED: attributes #[[ATTR6]] = { convergent "llvm.assume"="ompx_spmd_amenable" }
-; NVPTX-DISABLED: attributes #[[ATTR7]] = { convergent }
+; NVPTX-DISABLED: attributes #[[ATTR5:[0-9]+]] = { nosync nounwind }
+; NVPTX-DISABLED: attributes #[[ATTR6:[0-9]+]] = { convergent nounwind }
+; NVPTX-DISABLED: attributes #[[ATTR7]] = { convergent "llvm.assume"="ompx_spmd_amenable" }
+; NVPTX-DISABLED: attributes #[[ATTR8]] = { convergent }
;.
; AMDGPU: [[META0:![0-9]+]] = !{i32 0, i32 20, i32 171231761, !"sequential_loop_to_stack_var", i32 20, i32 1}
; AMDGPU: [[META1:![0-9]+]] = !{i32 0, i32 20, i32 171231761, !"sequential_loop", i32 5, i32 0}
diff --git a/llvm/test/Transforms/OpenMP/spmdization_assumes.ll b/llvm/test/Transforms/OpenMP/spmdization_assumes.ll
index 950e5fcf3f787..a0d43feecc3c0 100644
--- a/llvm/test/Transforms/OpenMP/spmdization_assumes.ll
+++ b/llvm/test/Transforms/OpenMP/spmdization_assumes.ll
@@ -31,7 +31,7 @@ define weak void @__omp_offloading_fd02_404433c2_main_l5(double* nonnull align 8
; 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]], i1 true, 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) #[[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:
@@ -55,12 +55,12 @@ define weak void @__omp_offloading_fd02_404433c2_main_l5(double* nonnull align 8
; CHECK: region.exit:
; CHECK-NEXT: [[TMP4:%.*]] = getelementptr inbounds [0 x i8*], [0 x i8*]* [[CAPTURED_VARS_ADDRS]], i64 0, i64 0
; CHECK-NEXT: call void @__kmpc_parallel_51(%struct.ident_t* nonnull @[[GLOB1]], i32 [[TMP1]], i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*)* @__omp_outlined__ to i8*), i8* bitcast (void (i16, i32)* @__omp_outlined___wrapper to i8*), i8** nonnull [[TMP4]], i64 0) #[[ATTR3]]
-; CHECK-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* nonnull @[[GLOB1]], i1 true, i1 false) #[[ATTR3]]
+; CHECK-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* nonnull @[[GLOB1]], i8 2, i1 false) #[[ATTR3]]
; CHECK-NEXT: br label [[COMMON_RET]]
;
entry:
%captured_vars_addrs = alloca [0 x i8*], align 8
- %0 = call i32 @__kmpc_target_init(%struct.ident_t* nonnull @1, i1 false, i1 true, i1 true) #3
+ %0 = call i32 @__kmpc_target_init(%struct.ident_t* nonnull @1, i8 1, i1 true, i1 true) #3
%exec_user_code = icmp eq i32 %0, -1
br i1 %exec_user_code, label %user_code.entry, label %common.ret
@@ -73,11 +73,11 @@ user_code.entry: ; preds = %entry
store double %call.i, double* %x, align 8, !tbaa !8
%2 = getelementptr inbounds [0 x i8*], [0 x i8*]* %captured_vars_addrs, i64 0, i64 0
call void @__kmpc_parallel_51(%struct.ident_t* nonnull @1, i32 %1, i32 1, i32 -1, i32 -1, i8* bitcast (void (i32*, i32*)* @__omp_outlined__ to i8*), i8* bitcast (void (i16, i32)* @__omp_outlined___wrapper to i8*), i8** nonnull %2, i64 0) #3
- call void @__kmpc_target_deinit(%struct.ident_t* nonnull @1, i1 false, i1 true) #3
+ call void @__kmpc_target_deinit(%struct.ident_t* nonnull @1, i8 1, i1 true) #3
br label %common.ret
}
-declare i32 @__kmpc_target_init(%struct.ident_t*, i1, i1, i1) local_unnamed_addr
+declare i32 @__kmpc_target_init(%struct.ident_t*, i8, i1, i1) 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 {
@@ -113,7 +113,7 @@ declare i32 @__kmpc_global_thread_num(%struct.ident_t*) local_unnamed_addr #3
; Function Attrs: alwaysinline
declare void @__kmpc_parallel_51(%struct.ident_t*, i32, i32, i32, i32, i8*, i8*, i8**, i64) local_unnamed_addr #4
-declare void @__kmpc_target_deinit(%struct.ident_t*, i1, i1) local_unnamed_addr
+declare void @__kmpc_target_deinit(%struct.ident_t*, i8, i1) local_unnamed_addr
; Function Attrs: convergent
declare double @__nv_sin(double) local_unnamed_addr #5
diff --git a/llvm/test/Transforms/OpenMP/spmdization_guarding.ll b/llvm/test/Transforms/OpenMP/spmdization_guarding.ll
index d77398de41408..a2a6d6de13729 100644
--- a/llvm/test/Transforms/OpenMP/spmdization_guarding.ll
+++ b/llvm/test/Transforms/OpenMP/spmdization_guarding.ll
@@ -59,7 +59,7 @@ define weak void @__omp_offloading_2a_fbfa7a_sequential_loop_l6(i32* %x, i64 %N)
; 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]], i1 true, 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) #[[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:
@@ -166,7 +166,7 @@ define weak void @__omp_offloading_2a_fbfa7a_sequential_loop_l6(i32* %x, i64 %N)
; CHECK-NEXT: [[CALL14_I:%.*]] = call i32 @no_openmp(i32* nonnull [[X]]) #[[ATTR7]], !noalias !8
; CHECK-NEXT: [[CALL15_I:%.*]] = call i32 @no_openmp(i32* nonnull [[X]]) #[[ATTR7]], !noalias !8
; CHECK-NEXT: [[CALL16_I:%.*]] = call i32 @no_openmp(i32* nonnull [[X]]) #[[ATTR7]], !noalias !8
-; CHECK-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* nonnull @[[GLOB1]], i1 true, i1 false) #[[ATTR4]]
+; CHECK-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* nonnull @[[GLOB1]], i8 2, i1 false) #[[ATTR4]]
; CHECK-NEXT: ret void
; CHECK: worker.exit:
; CHECK-NEXT: ret void
@@ -176,7 +176,7 @@ define weak void @__omp_offloading_2a_fbfa7a_sequential_loop_l6(i32* %x, i64 %N)
; 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]], i1 false, 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) #[[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:
@@ -245,14 +245,14 @@ define weak void @__omp_offloading_2a_fbfa7a_sequential_loop_l6(i32* %x, i64 %N)
; CHECK-DISABLED-NEXT: [[CALL14_I:%.*]] = call i32 @no_openmp(i32* nonnull [[X]]) #[[ATTR7]], !noalias !8
; CHECK-DISABLED-NEXT: [[CALL15_I:%.*]] = call i32 @no_openmp(i32* nonnull [[X]]) #[[ATTR7]], !noalias !8
; CHECK-DISABLED-NEXT: [[CALL16_I:%.*]] = call i32 @no_openmp(i32* nonnull [[X]]) #[[ATTR7]], !noalias !8
-; CHECK-DISABLED-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* nonnull @[[GLOB1]], i1 false, i1 true) #[[ATTR4]]
+; CHECK-DISABLED-NEXT: call void @__kmpc_target_deinit(%struct.ident_t* nonnull @[[GLOB1]], i8 1, i1 true) #[[ATTR4]]
; CHECK-DISABLED-NEXT: ret void
; CHECK-DISABLED: worker.exit:
; CHECK-DISABLED-NEXT: ret void
;
entry:
%N.addr.sroa.0.0.extract.trunc = trunc i64 %N to i32
- %0 = call i32 @__kmpc_target_init(%struct.ident_t* nonnull @1, i1 false, i1 true, i1 true) #3
+ %0 = call i32 @__kmpc_target_init(%struct.ident_t* nonnull @1, i8 1, i1 true, i1 true) #3
%exec_user_code = icmp eq i32 %0, -1
br i1 %exec_user_code, label %user_code.entry, label %worker.exit
@@ -298,7 +298,7 @@ __omp_outlined__.exit: ; preds = %for.cond.i
%call14.i = call i32 @no_openmp(i32* nonnull %x) #5, !noalias !8
%call15.i = call i32 @no_openmp(i32* nonnull %x) #5, !noalias !8
%call16.i = call i32 @no_openmp(i32* nonnull %x) #5, !noalias !8
- call void @__kmpc_target_deinit(%struct.ident_t* nonnull @1, i1 false, i1 true) #3
+ call void @__kmpc_target_deinit(%struct.ident_t* nonnull @1, i8 1, i1 true) #3
ret void
worker.exit: ; preds = %entry
@@ -330,7 +330,7 @@ define internal void @__omp_outlined__1_wrapper(i16 zeroext %0, i32 %1) {
declare void @__kmpc_parallel_51(%struct.ident_t*, i32, i32, i32, i32, i8*, i8*, i8**, i64)
-declare i32 @__kmpc_target_init(%struct.ident_t*, i1, i1, i1)
+declare i32 @__kmpc_target_init(%struct.ident_t*, i8, i1, i1)
; Function Attrs: convergent
declare i32 @no_openmp(i32*) #1
@@ -341,7 +341,7 @@ declare void @pure() #2
; Function Attrs: nounwind
declare i32 @__kmpc_global_thread_num(%struct.ident_t*) #3
-declare void @__kmpc_target_deinit(%struct.ident_t*, i1, i1)
+declare void @__kmpc_target_deinit(%struct.ident_t*, i8, i1)
; Function Attrs: inaccessiblememonly nofree nosync nounwind willreturn
declare void @llvm.experimental.noalias.scope.decl(metadata) #4
diff --git a/llvm/test/Transforms/OpenMP/spmdization_remarks.ll b/llvm/test/Transforms/OpenMP/spmdization_remarks.ll
index df60ddc2e1f3b..8f63e2c68f7a0 100644
--- a/llvm/test/Transforms/OpenMP/spmdization_remarks.ll
+++ b/llvm/test/Transforms/OpenMP/spmdization_remarks.ll
@@ -16,7 +16,7 @@ target triple = "nvptx64"
;; unknown();
;; }
;; }
-;;
+;;
;; void test_fallback(void) {
;; #pragma omp target teams
;; {
@@ -62,7 +62,7 @@ target triple = "nvptx64"
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, i1 false, i1 true, i1 true) #3, !dbg !18
+ %0 = call i32 @__kmpc_target_init(%struct.ident_t* nonnull @1, i8 1, i1 true, i1 true) #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
@@ -79,11 +79,11 @@ user_code.entry: ; preds = %entry
call void @__kmpc_parallel_51(%struct.ident_t* noundef nonnull @13, i32 %3, i32 noundef 1, i32 noundef -1, i32 noundef -1, i8* noundef bitcast (void (i32*, i32*)* @__omp_outlined__2 to i8*), i8* noundef bitcast (void (i16, i32)* @__omp_outlined__2_wrapper to i8*), i8** noundef nonnull %4, i64 noundef 0) #3, !dbg !23
call void @llvm.lifetime.end.p0i8(i64 0, i8* nonnull %2) #3, !dbg !26
call void @unknown() #6, !dbg !27
- call void @__kmpc_target_deinit(%struct.ident_t* nonnull @5, i1 false, i1 true) #3, !dbg !28
+ call void @__kmpc_target_deinit(%struct.ident_t* nonnull @5, i8 1, i1 true) #3, !dbg !28
br label %common.ret
}
-declare i32 @__kmpc_target_init(%struct.ident_t*, i1, i1, i1) local_unnamed_addr
+declare i32 @__kmpc_target_init(%struct.ident_t*, i8, i1, i1) local_unnamed_addr
; Function Attrs: convergent
declare void @unknown() local_unnamed_addr #1
@@ -101,13 +101,13 @@ entry:
; Function Attrs: nounwind
declare i32 @__kmpc_global_thread_num(%struct.ident_t*) local_unnamed_addr #3
-declare void @__kmpc_target_deinit(%struct.ident_t*, i1, i1) local_unnamed_addr
+declare void @__kmpc_target_deinit(%struct.ident_t*, i8, i1) local_unnamed_addr
; Function Attrs: norecurse nounwind
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, i1 false, i1 true, i1 true) #3, !dbg !33
+ %0 = call i32 @__kmpc_target_init(%struct.ident_t* nonnull @7, i8 1, i1 true, i1 true) #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
@@ -131,7 +131,7 @@ user_code.entry: ; preds = %entry
call void @__kmpc_parallel_51(%struct.ident_t* noundef nonnull @13, i32 %6, i32 noundef 1, i32 noundef -1, i32 noundef -1, i8* noundef bitcast (void (i32*, i32*)* @__omp_outlined__2 to i8*), i8* noundef bitcast (void (i16, i32)* @__omp_outlined__2_wrapper to i8*), i8** noundef nonnull %4, i64 noundef 0) #3, !dbg !43
call void @llvm.lifetime.end.p0i8(i64 0, i8* nonnull %2) #3, !dbg !45
call void @spmd_amenable()
- call void @__kmpc_target_deinit(%struct.ident_t* nonnull @11, i1 false, i1 true) #3, !dbg !46
+ call void @__kmpc_target_deinit(%struct.ident_t* nonnull @11, i8 1, i1 true) #3, !dbg !46
br label %common.ret
}
diff --git a/openmp/libomptarget/DeviceRTL/include/Interface.h b/openmp/libomptarget/DeviceRTL/include/Interface.h
index 027d2c9e13645..1d41acf5dad6b 100644
--- a/openmp/libomptarget/DeviceRTL/include/Interface.h
+++ b/openmp/libomptarget/DeviceRTL/include/Interface.h
@@ -206,10 +206,10 @@ uint32_t __kmpc_get_hardware_thread_id_in_block();
///{
int8_t __kmpc_is_spmd_exec_mode();
-int32_t __kmpc_target_init(IdentTy *Ident, bool IsSPMD,
+int32_t __kmpc_target_init(IdentTy *Ident, int8_t Mode,
bool UseGenericStateMachine, bool);
-void __kmpc_target_deinit(IdentTy *Ident, bool IsSPMD, bool);
+void __kmpc_target_deinit(IdentTy *Ident, int8_t Mode, bool);
///}
diff --git a/openmp/libomptarget/DeviceRTL/include/Types.h b/openmp/libomptarget/DeviceRTL/include/Types.h
index 5b15778d5a2eb..0ff0cee66b3f3 100644
--- a/openmp/libomptarget/DeviceRTL/include/Types.h
+++ b/openmp/libomptarget/DeviceRTL/include/Types.h
@@ -178,6 +178,11 @@ typedef enum omp_allocator_handle_t {
KMP_ALLOCATOR_MAX_HANDLE = ~(0U)
} omp_allocator_handle_t;
+enum OMPTgtExecModeFlags : int8_t {
+ OMP_TGT_EXEC_MODE_GENERIC = 1 << 0,
+ OMP_TGT_EXEC_MODE_SPMD = 1 << 1,
+};
+
#define __PRAGMA(STR) _Pragma(#STR)
#define OMP_PRAGMA(STR) __PRAGMA(omp STR)
diff --git a/openmp/libomptarget/DeviceRTL/src/Kernel.cpp b/openmp/libomptarget/DeviceRTL/src/Kernel.cpp
index 4959177d42931..ead6085ae373c 100644
--- a/openmp/libomptarget/DeviceRTL/src/Kernel.cpp
+++ b/openmp/libomptarget/DeviceRTL/src/Kernel.cpp
@@ -64,8 +64,9 @@ extern "C" {
///
/// \param Ident Source location identification, can be NULL.
///
-int32_t __kmpc_target_init(IdentTy *Ident, bool IsSPMD,
+int32_t __kmpc_target_init(IdentTy *Ident, int8_t Mode,
bool UseGenericStateMachine, bool) {
+ const bool IsSPMD = Mode & OMP_TGT_EXEC_MODE_SPMD;
if (IsSPMD) {
inititializeRuntime(/* IsSPMD */ true);
synchronize::threads();
@@ -96,7 +97,8 @@ int32_t __kmpc_target_init(IdentTy *Ident, bool IsSPMD,
///
/// \param Ident Source location identification, can be NULL.
///
-void __kmpc_target_deinit(IdentTy *Ident, bool IsSPMD, bool) {
+void __kmpc_target_deinit(IdentTy *Ident, int8_t Mode, bool) {
+ const bool IsSPMD = Mode & OMP_TGT_EXEC_MODE_SPMD;
state::assumeInitialState(IsSPMD);
if (IsSPMD)
return;
diff --git a/openmp/libomptarget/deviceRTLs/common/include/target.h b/openmp/libomptarget/deviceRTLs/common/include/target.h
index 997e93b924e24..fbb72f9900779 100644
--- a/openmp/libomptarget/deviceRTLs/common/include/target.h
+++ b/openmp/libomptarget/deviceRTLs/common/include/target.h
@@ -36,7 +36,7 @@ typedef struct ident ident_t;
///
/// \code
/// void kernel(...) {
-/// ThreadKind = __kmpc_target_init(Ident, /* IsSPMD */ false,
+/// ThreadKind = __kmpc_target_init(Ident, /* Mode */ 1,
/// /* UseGenericStateMachine */ true,
/// /* RequiresFullRuntime */ ... );
/// if (ThreadKind == -1) {
@@ -50,7 +50,7 @@ typedef struct ident ident_t;
///
/// \code
/// void kernel(...) {
-/// ThreadKind = __kmpc_target_init(Ident, /* IsSPMD */ false,
+/// ThreadKind = __kmpc_target_init(Ident, /* Mode */ 1,
/// /* UseGenericStateMachine */ false,
/// /* RequiresFullRuntime */ ... );
/// if (ThreadKind == -1) {
@@ -72,7 +72,7 @@ typedef struct ident ident_t;
///
/// \param Ident Source location identification, can be NULL.
///
-int32_t __kmpc_target_init(ident_t *Ident, bool IsSPMD,
+int32_t __kmpc_target_init(ident_t *Ident, int8_t Mode,
bool UseGenericStateMachine,
bool RequiresFullRuntime);
@@ -86,7 +86,7 @@ int32_t __kmpc_target_init(ident_t *Ident, bool IsSPMD,
///
/// \param Ident Source location identification, can be NULL.
///
-void __kmpc_target_deinit(ident_t *Ident, bool IsSPMD,
+void __kmpc_target_deinit(ident_t *Ident, int8_t Mode,
bool RequiresFullRuntime);
///}
diff --git a/openmp/libomptarget/deviceRTLs/common/omptarget.h b/openmp/libomptarget/deviceRTLs/common/omptarget.h
index 1464959fb6114..d4f8678ee1b35 100644
--- a/openmp/libomptarget/deviceRTLs/common/omptarget.h
+++ b/openmp/libomptarget/deviceRTLs/common/omptarget.h
@@ -256,7 +256,7 @@ extern uint16_t EXTERN_SHARED(nThreads);
extern omptarget_nvptx_ThreadPrivateContext *
EXTERN_SHARED(omptarget_nvptx_threadPrivateContext);
-extern uint32_t EXTERN_SHARED(execution_param);
+extern int8_t EXTERN_SHARED(execution_param);
extern void *EXTERN_SHARED(ReductionScratchpadPtr);
////////////////////////////////////////////////////////////////////////////////
diff --git a/openmp/libomptarget/deviceRTLs/common/src/omp_data.cu b/openmp/libomptarget/deviceRTLs/common/src/omp_data.cu
index e31217e1df80a..d420fad42b62d 100644
--- a/openmp/libomptarget/deviceRTLs/common/src/omp_data.cu
+++ b/openmp/libomptarget/deviceRTLs/common/src/omp_data.cu
@@ -56,7 +56,7 @@ omptarget_nvptx_WorkFn SHARED(omptarget_nvptx_workFn);
////////////////////////////////////////////////////////////////////////////////
// OpenMP kernel execution parameters
////////////////////////////////////////////////////////////////////////////////
-uint32_t SHARED(execution_param);
+int8_t SHARED(execution_param);
////////////////////////////////////////////////////////////////////////////////
// Scratchpad for teams reduction.
diff --git a/openmp/libomptarget/deviceRTLs/common/src/omptarget.cu b/openmp/libomptarget/deviceRTLs/common/src/omptarget.cu
index 7f7982cbc3b8f..78bd5d7cd62b1 100644
--- a/openmp/libomptarget/deviceRTLs/common/src/omptarget.cu
+++ b/openmp/libomptarget/deviceRTLs/common/src/omptarget.cu
@@ -38,7 +38,7 @@ static void __kmpc_generic_kernel_init() {
if (threadIdInBlock != GetMasterThreadID())
return;
- setExecutionParameters(Generic, RuntimeInitialized);
+ setExecutionParameters(OMP_TGT_EXEC_MODE_GENERIC, OMP_TGT_RUNTIME_INITIALIZED);
ASSERT0(LT_FUSSY, threadIdInBlock == GetMasterThreadID(),
"__kmpc_kernel_init() must be called by team master warp only!");
PRINT0(LD_IO, "call to __kmpc_kernel_init for master\n");
@@ -85,8 +85,9 @@ static void __kmpc_generic_kernel_deinit() {
static void __kmpc_spmd_kernel_init(bool RequiresFullRuntime) {
PRINT0(LD_IO, "call to __kmpc_spmd_kernel_init\n");
- setExecutionParameters(Spmd, RequiresFullRuntime ? RuntimeInitialized
- : RuntimeUninitialized);
+ setExecutionParameters(OMP_TGT_EXEC_MODE_SPMD,
+ RequiresFullRuntime ? OMP_TGT_RUNTIME_INITIALIZED
+ : OMP_TGT_RUNTIME_UNINITIALIZED);
int threadId = __kmpc_get_hardware_thread_id_in_block();
if (threadId == 0) {
usedSlotIdx = __kmpc_impl_smid() % MAX_SM;
@@ -160,7 +161,7 @@ static void __kmpc_spmd_kernel_deinit(bool RequiresFullRuntime) {
// Return true if the current target region is executed in SPMD mode.
EXTERN int8_t __kmpc_is_spmd_exec_mode() {
- return (execution_param & ModeMask) == Spmd;
+ return execution_param & OMP_TGT_EXEC_MODE_SPMD;
}
EXTERN int8_t __kmpc_is_generic_main_thread(kmp_int32 Tid) {
@@ -202,9 +203,10 @@ static void __kmpc_target_region_state_machine(ident_t *Ident) {
}
EXTERN
-int32_t __kmpc_target_init(ident_t *Ident, bool IsSPMD,
+int32_t __kmpc_target_init(ident_t *Ident, int8_t Mode,
bool UseGenericStateMachine,
bool RequiresFullRuntime) {
+ const bool IsSPMD = Mode & OMP_TGT_EXEC_MODE_SPMD;
int TId = __kmpc_get_hardware_thread_id_in_block();
if (IsSPMD)
__kmpc_spmd_kernel_init(RequiresFullRuntime);
@@ -226,13 +228,13 @@ int32_t __kmpc_target_init(ident_t *Ident, bool IsSPMD,
}
EXTERN
-void __kmpc_target_deinit(ident_t *Ident, bool IsSPMD,
- bool RequiresFullRuntime) {
+void __kmpc_target_deinit(ident_t *Ident, int8_t Mode,
+ bool RequiresFullRuntime) {
+ const bool IsSPMD = Mode & OMP_TGT_EXEC_MODE_SPMD;
if (IsSPMD)
__kmpc_spmd_kernel_deinit(RequiresFullRuntime);
else
__kmpc_generic_kernel_deinit();
}
-
#pragma omp end declare target
diff --git a/openmp/libomptarget/deviceRTLs/common/src/support.cu b/openmp/libomptarget/deviceRTLs/common/src/support.cu
index a0a8116bb825f..6dd591325c0bf 100644
--- a/openmp/libomptarget/deviceRTLs/common/src/support.cu
+++ b/openmp/libomptarget/deviceRTLs/common/src/support.cu
@@ -19,19 +19,18 @@
// Execution Parameters
////////////////////////////////////////////////////////////////////////////////
-void setExecutionParameters(ExecutionMode EMode, RuntimeMode RMode) {
+void setExecutionParameters(OMPTgtExecModeFlags EMode,
+ OMPTgtRuntimeModeFlags RMode) {
execution_param = EMode;
execution_param |= RMode;
}
-bool isGenericMode() { return (execution_param & ModeMask) == Generic; }
+bool isGenericMode() { return execution_param & OMP_TGT_EXEC_MODE_GENERIC; }
-bool isRuntimeUninitialized() {
- return (execution_param & RuntimeMask) == RuntimeUninitialized;
-}
+bool isRuntimeUninitialized() { return !isRuntimeInitialized(); }
bool isRuntimeInitialized() {
- return (execution_param & RuntimeMask) == RuntimeInitialized;
+ return execution_param & OMP_TGT_RUNTIME_INITIALIZED;
}
////////////////////////////////////////////////////////////////////////////////
diff --git a/openmp/libomptarget/deviceRTLs/common/support.h b/openmp/libomptarget/deviceRTLs/common/support.h
index 7a4e46feead24..dfaf08339e894 100644
--- a/openmp/libomptarget/deviceRTLs/common/support.h
+++ b/openmp/libomptarget/deviceRTLs/common/support.h
@@ -19,19 +19,18 @@
////////////////////////////////////////////////////////////////////////////////
// Execution Parameters
////////////////////////////////////////////////////////////////////////////////
-enum ExecutionMode {
- Spmd = 0x00u,
- Generic = 0x01u,
- ModeMask = 0x01u,
+enum OMPTgtExecModeFlags : int8_t {
+ OMP_TGT_EXEC_MODE_GENERIC = 1 << 0,
+ OMP_TGT_EXEC_MODE_SPMD = 1 << 1
};
-enum RuntimeMode {
- RuntimeInitialized = 0x00u,
- RuntimeUninitialized = 0x02u,
- RuntimeMask = 0x02u,
+enum OMPTgtRuntimeModeFlags : int8_t {
+ OMP_TGT_RUNTIME_UNINITIALIZED = 0,
+ OMP_TGT_RUNTIME_INITIALIZED = 1 << 2
};
-void setExecutionParameters(ExecutionMode EMode, RuntimeMode RMode);
+void setExecutionParameters(OMPTgtExecModeFlags EMode,
+ OMPTgtRuntimeModeFlags RMode);
bool isGenericMode();
bool isRuntimeUninitialized();
bool isRuntimeInitialized();
diff --git a/openmp/libomptarget/deviceRTLs/interface.h b/openmp/libomptarget/deviceRTLs/interface.h
index 1986902031e09..4bee5db03641a 100644
--- a/openmp/libomptarget/deviceRTLs/interface.h
+++ b/openmp/libomptarget/deviceRTLs/interface.h
@@ -416,10 +416,10 @@ EXTERN int32_t __kmpc_cancel(kmp_Ident *loc, int32_t global_tid,
int32_t cancelVal);
// non standard
-EXTERN int32_t __kmpc_target_init(ident_t *Ident, bool IsSPMD,
+EXTERN int32_t __kmpc_target_init(ident_t *Ident, int8_t Mode,
bool UseGenericStateMachine,
bool RequiresFullRuntime);
-EXTERN void __kmpc_target_deinit(ident_t *Ident, bool IsSPMD,
+EXTERN void __kmpc_target_deinit(ident_t *Ident, int8_t Mode,
bool RequiresFullRuntime);
EXTERN void __kmpc_kernel_prepare_parallel(void *WorkFn);
EXTERN bool __kmpc_kernel_parallel(void **WorkFn);
More information about the cfe-commits
mailing list