// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --function-signature --include-generated-funcs --replace-value-regex "__omp_offloading_[0-9a-z]+_[0-9a-z]+" "reduction_size[.].+[.]" "pl_cond[.].+[.|,]" --prefix-filecheck-ir-name _
// Test target codegen - host bc file has to be created first.
// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=nvptx64-nvidia-cuda -emit-llvm-bc %s -o %t-ppc-host.bc
// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple nvptx64-unknown-unknown -fopenmp-targets=nvptx64-nvidia-cuda -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o - | FileCheck %s --check-prefix=CHECK1
// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple i386-unknown-unknown -fopenmp-targets=nvptx-nvidia-cuda -emit-llvm-bc %s -o %t-x86-host.bc
// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -x c++ -triple nvptx-unknown-unknown -fopenmp-targets=nvptx-nvidia-cuda -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -o - | FileCheck %s --check-prefix=CHECK2
// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=45 -fexceptions -fcxx-exceptions -x c++ -triple nvptx-unknown-unknown -fopenmp-targets=nvptx-nvidia-cuda -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -o - | FileCheck %s --check-prefix=CHECK2

// RUN: %clang_cc1 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=nvptx64-nvidia-cuda -emit-llvm-bc %s -o %t-ppc-host.bc
// RUN: %clang_cc1 -verify -fopenmp -x c++ -triple nvptx64-unknown-unknown -fopenmp-targets=nvptx64-nvidia-cuda -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o - | FileCheck %s --check-prefix=CHECK1
// RUN: %clang_cc1 -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=nvptx-nvidia-cuda -emit-llvm-bc %s -o %t-x86-host.bc
// RUN: %clang_cc1 -verify -fopenmp -x c++ -triple nvptx-unknown-unknown -fopenmp-targets=nvptx-nvidia-cuda -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -o - | FileCheck %s --check-prefix=CHECK2
// RUN: %clang_cc1 -verify -fopenmp -fexceptions -fcxx-exceptions -x c++ -triple nvptx-unknown-unknown -fopenmp-targets=nvptx-nvidia-cuda -emit-llvm %s -fopenmp-is-device -fopenmp-host-ir-file-path %t-x86-host.bc -o - | FileCheck %s --check-prefix=CHECK2

// expected-no-diagnostics
#ifndef HEADER
#define HEADER

int a;

int foo(int *a);

int main(int argc, char **argv) {
#pragma omp target teams distribute parallel for map(tofrom:a) if(target:argc) schedule(static, a)
  for (int i= 0; i < argc; ++i)
    a = foo(&i) + foo(&a) + foo(&argc);
  return 0;
}


#endif
// CHECK1-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}_main_l24
// CHECK1-SAME: (i64 noundef [[ARGC:%.*]], ptr noundef nonnull align 4 dereferenceable(4) [[A:%.*]], i64 noundef [[DOTCAPTURE_EXPR_:%.*]]) #[[ATTR0:[0-9]+]] {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[ARGC_ADDR:%.*]] = alloca i64, align 8
// CHECK1-NEXT:    [[A_ADDR:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    [[DOTCAPTURE_EXPR__ADDR:%.*]] = alloca i64, align 8
// CHECK1-NEXT:    [[ARGC_CASTED:%.*]] = alloca i64, align 8
// CHECK1-NEXT:    [[DOTCAPTURE_EXPR__CASTED:%.*]] = alloca i64, align 8
// CHECK1-NEXT:    [[DOTZERO_ADDR:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    store i64 [[ARGC]], ptr [[ARGC_ADDR]], align 8
// CHECK1-NEXT:    store ptr [[A]], ptr [[A_ADDR]], align 8
// CHECK1-NEXT:    store i64 [[DOTCAPTURE_EXPR_]], ptr [[DOTCAPTURE_EXPR__ADDR]], align 8
// CHECK1-NEXT:    [[TMP0:%.*]] = load ptr, ptr [[A_ADDR]], align 8
// CHECK1-NEXT:    [[TMP1:%.*]] = call i32 @__kmpc_target_init(ptr @[[GLOB1:[0-9]+]], i8 2, 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(ptr @[[GLOB1]])
// CHECK1-NEXT:    [[TMP3:%.*]] = load i32, ptr [[ARGC_ADDR]], align 4
// CHECK1-NEXT:    store i32 [[TMP3]], ptr [[ARGC_CASTED]], align 4
// CHECK1-NEXT:    [[TMP4:%.*]] = load i64, ptr [[ARGC_CASTED]], align 8
// CHECK1-NEXT:    [[TMP5:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR__ADDR]], align 4
// CHECK1-NEXT:    store i32 [[TMP5]], ptr [[DOTCAPTURE_EXPR__CASTED]], align 4
// CHECK1-NEXT:    [[TMP6:%.*]] = load i64, ptr [[DOTCAPTURE_EXPR__CASTED]], align 8
// CHECK1-NEXT:    store i32 0, ptr [[DOTZERO_ADDR]], align 4
// CHECK1-NEXT:    store i32 [[TMP2]], ptr [[DOTTHREADID_TEMP_]], align 4
// CHECK1-NEXT:    call void @__omp_outlined__(ptr [[DOTTHREADID_TEMP_]], ptr [[DOTZERO_ADDR]], i64 [[TMP4]], ptr [[TMP0]], i64 [[TMP6]]) #[[ATTR3:[0-9]+]]
// CHECK1-NEXT:    call void @__kmpc_target_deinit(ptr @[[GLOB1]], i8 2)
// CHECK1-NEXT:    ret void
// CHECK1:       worker.exit:
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@__omp_outlined__
// CHECK1-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]], i64 noundef [[ARGC:%.*]], ptr noundef nonnull align 4 dereferenceable(4) [[A:%.*]], i64 noundef [[DOTCAPTURE_EXPR_:%.*]]) #[[ATTR1:[0-9]+]] {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    [[ARGC_ADDR:%.*]] = alloca i64, align 8
// CHECK1-NEXT:    [[A_ADDR:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    [[DOTCAPTURE_EXPR__ADDR:%.*]] = alloca i64, align 8
// CHECK1-NEXT:    [[DOTOMP_IV:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[TMP:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[DOTCAPTURE_EXPR_1:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[DOTCAPTURE_EXPR_2:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[I:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[DOTOMP_COMB_LB:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[DOTOMP_COMB_UB:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[DOTOMP_STRIDE:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[DOTOMP_IS_LAST:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[I4:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[ARGC_CASTED:%.*]] = alloca i64, align 8
// CHECK1-NEXT:    [[DOTCAPTURE_EXPR__CASTED:%.*]] = alloca i64, align 8
// CHECK1-NEXT:    [[CAPTURED_VARS_ADDRS:%.*]] = alloca [5 x ptr], align 8
// CHECK1-NEXT:    store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK1-NEXT:    store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 8
// CHECK1-NEXT:    store i64 [[ARGC]], ptr [[ARGC_ADDR]], align 8
// CHECK1-NEXT:    store ptr [[A]], ptr [[A_ADDR]], align 8
// CHECK1-NEXT:    store i64 [[DOTCAPTURE_EXPR_]], ptr [[DOTCAPTURE_EXPR__ADDR]], align 8
// CHECK1-NEXT:    [[TMP0:%.*]] = load ptr, ptr [[A_ADDR]], align 8
// CHECK1-NEXT:    [[TMP1:%.*]] = load i32, ptr [[ARGC_ADDR]], align 4
// CHECK1-NEXT:    store i32 [[TMP1]], ptr [[DOTCAPTURE_EXPR_1]], align 4
// CHECK1-NEXT:    [[TMP2:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_1]], align 4
// CHECK1-NEXT:    [[SUB:%.*]] = sub nsw i32 [[TMP2]], 0
// CHECK1-NEXT:    [[DIV:%.*]] = sdiv i32 [[SUB]], 1
// CHECK1-NEXT:    [[SUB3:%.*]] = sub nsw i32 [[DIV]], 1
// CHECK1-NEXT:    store i32 [[SUB3]], ptr [[DOTCAPTURE_EXPR_2]], align 4
// CHECK1-NEXT:    store i32 0, ptr [[I]], align 4
// CHECK1-NEXT:    [[TMP3:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_1]], align 4
// CHECK1-NEXT:    [[CMP:%.*]] = icmp slt i32 0, [[TMP3]]
// CHECK1-NEXT:    br i1 [[CMP]], label [[OMP_PRECOND_THEN:%.*]], label [[OMP_PRECOND_END:%.*]]
// CHECK1:       omp.precond.then:
// CHECK1-NEXT:    store i32 0, ptr [[DOTOMP_COMB_LB]], align 4
// CHECK1-NEXT:    [[TMP4:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_2]], align 4
// CHECK1-NEXT:    store i32 [[TMP4]], ptr [[DOTOMP_COMB_UB]], align 4
// CHECK1-NEXT:    store i32 1, ptr [[DOTOMP_STRIDE]], align 4
// CHECK1-NEXT:    store i32 0, ptr [[DOTOMP_IS_LAST]], align 4
// CHECK1-NEXT:    [[NVPTX_NUM_THREADS:%.*]] = call i32 @__kmpc_get_hardware_num_threads_in_block()
// CHECK1-NEXT:    [[TMP5:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK1-NEXT:    [[TMP6:%.*]] = load i32, ptr [[TMP5]], align 4
// CHECK1-NEXT:    call void @__kmpc_distribute_static_init_4(ptr @[[GLOB2:[0-9]+]], i32 [[TMP6]], i32 91, ptr [[DOTOMP_IS_LAST]], ptr [[DOTOMP_COMB_LB]], ptr [[DOTOMP_COMB_UB]], ptr [[DOTOMP_STRIDE]], i32 1, i32 [[NVPTX_NUM_THREADS]])
// CHECK1-NEXT:    [[TMP7:%.*]] = load i32, ptr [[DOTOMP_COMB_UB]], align 4
// CHECK1-NEXT:    [[TMP8:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_2]], align 4
// CHECK1-NEXT:    [[CMP5:%.*]] = icmp sgt i32 [[TMP7]], [[TMP8]]
// CHECK1-NEXT:    br i1 [[CMP5]], label [[COND_TRUE:%.*]], label [[COND_FALSE:%.*]]
// CHECK1:       cond.true:
// CHECK1-NEXT:    [[TMP9:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_2]], align 4
// CHECK1-NEXT:    br label [[COND_END:%.*]]
// CHECK1:       cond.false:
// CHECK1-NEXT:    [[TMP10:%.*]] = load i32, ptr [[DOTOMP_COMB_UB]], align 4
// CHECK1-NEXT:    br label [[COND_END]]
// CHECK1:       cond.end:
// CHECK1-NEXT:    [[COND:%.*]] = phi i32 [ [[TMP9]], [[COND_TRUE]] ], [ [[TMP10]], [[COND_FALSE]] ]
// CHECK1-NEXT:    store i32 [[COND]], ptr [[DOTOMP_COMB_UB]], align 4
// CHECK1-NEXT:    [[TMP11:%.*]] = load i32, ptr [[DOTOMP_COMB_LB]], align 4
// CHECK1-NEXT:    store i32 [[TMP11]], ptr [[DOTOMP_IV]], align 4
// CHECK1-NEXT:    br label [[OMP_INNER_FOR_COND:%.*]]
// CHECK1:       omp.inner.for.cond:
// CHECK1-NEXT:    [[TMP12:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4
// CHECK1-NEXT:    [[TMP13:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_2]], align 4
// CHECK1-NEXT:    [[ADD:%.*]] = add nsw i32 [[TMP13]], 1
// CHECK1-NEXT:    [[CMP6:%.*]] = icmp slt i32 [[TMP12]], [[ADD]]
// CHECK1-NEXT:    br i1 [[CMP6]], label [[OMP_INNER_FOR_BODY:%.*]], label [[OMP_INNER_FOR_END:%.*]]
// CHECK1:       omp.inner.for.body:
// CHECK1-NEXT:    [[TMP14:%.*]] = load i32, ptr [[DOTOMP_COMB_LB]], align 4
// CHECK1-NEXT:    [[TMP15:%.*]] = zext i32 [[TMP14]] to i64
// CHECK1-NEXT:    [[TMP16:%.*]] = load i32, ptr [[DOTOMP_COMB_UB]], align 4
// CHECK1-NEXT:    [[TMP17:%.*]] = zext i32 [[TMP16]] to i64
// CHECK1-NEXT:    [[TMP18:%.*]] = load i32, ptr [[ARGC_ADDR]], align 4
// CHECK1-NEXT:    store i32 [[TMP18]], ptr [[ARGC_CASTED]], align 4
// CHECK1-NEXT:    [[TMP19:%.*]] = load i64, ptr [[ARGC_CASTED]], align 8
// CHECK1-NEXT:    [[TMP20:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR__ADDR]], align 4
// CHECK1-NEXT:    store i32 [[TMP20]], ptr [[DOTCAPTURE_EXPR__CASTED]], align 4
// CHECK1-NEXT:    [[TMP21:%.*]] = load i64, ptr [[DOTCAPTURE_EXPR__CASTED]], align 8
// CHECK1-NEXT:    [[TMP22:%.*]] = getelementptr inbounds [5 x ptr], ptr [[CAPTURED_VARS_ADDRS]], i64 0, i64 0
// CHECK1-NEXT:    [[TMP23:%.*]] = inttoptr i64 [[TMP15]] to ptr
// CHECK1-NEXT:    store ptr [[TMP23]], ptr [[TMP22]], align 8
// CHECK1-NEXT:    [[TMP24:%.*]] = getelementptr inbounds [5 x ptr], ptr [[CAPTURED_VARS_ADDRS]], i64 0, i64 1
// CHECK1-NEXT:    [[TMP25:%.*]] = inttoptr i64 [[TMP17]] to ptr
// CHECK1-NEXT:    store ptr [[TMP25]], ptr [[TMP24]], align 8
// CHECK1-NEXT:    [[TMP26:%.*]] = getelementptr inbounds [5 x ptr], ptr [[CAPTURED_VARS_ADDRS]], i64 0, i64 2
// CHECK1-NEXT:    [[TMP27:%.*]] = inttoptr i64 [[TMP19]] to ptr
// CHECK1-NEXT:    store ptr [[TMP27]], ptr [[TMP26]], align 8
// CHECK1-NEXT:    [[TMP28:%.*]] = getelementptr inbounds [5 x ptr], ptr [[CAPTURED_VARS_ADDRS]], i64 0, i64 3
// CHECK1-NEXT:    store ptr [[TMP0]], ptr [[TMP28]], align 8
// CHECK1-NEXT:    [[TMP29:%.*]] = getelementptr inbounds [5 x ptr], ptr [[CAPTURED_VARS_ADDRS]], i64 0, i64 4
// CHECK1-NEXT:    [[TMP30:%.*]] = inttoptr i64 [[TMP21]] to ptr
// CHECK1-NEXT:    store ptr [[TMP30]], ptr [[TMP29]], align 8
// CHECK1-NEXT:    [[TMP31:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK1-NEXT:    [[TMP32:%.*]] = load i32, ptr [[TMP31]], align 4
// CHECK1-NEXT:    call void @__kmpc_parallel_51(ptr @[[GLOB1]], i32 [[TMP32]], i32 1, i32 -1, i32 -1, ptr @__omp_outlined__1, ptr null, ptr [[CAPTURED_VARS_ADDRS]], i64 5)
// CHECK1-NEXT:    br label [[OMP_INNER_FOR_INC:%.*]]
// CHECK1:       omp.inner.for.inc:
// CHECK1-NEXT:    [[TMP33:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4
// CHECK1-NEXT:    [[TMP34:%.*]] = load i32, ptr [[DOTOMP_STRIDE]], align 4
// CHECK1-NEXT:    [[ADD7:%.*]] = add nsw i32 [[TMP33]], [[TMP34]]
// CHECK1-NEXT:    store i32 [[ADD7]], ptr [[DOTOMP_IV]], align 4
// CHECK1-NEXT:    [[TMP35:%.*]] = load i32, ptr [[DOTOMP_COMB_LB]], align 4
// CHECK1-NEXT:    [[TMP36:%.*]] = load i32, ptr [[DOTOMP_STRIDE]], align 4
// CHECK1-NEXT:    [[ADD8:%.*]] = add nsw i32 [[TMP35]], [[TMP36]]
// CHECK1-NEXT:    store i32 [[ADD8]], ptr [[DOTOMP_COMB_LB]], align 4
// CHECK1-NEXT:    [[TMP37:%.*]] = load i32, ptr [[DOTOMP_COMB_UB]], align 4
// CHECK1-NEXT:    [[TMP38:%.*]] = load i32, ptr [[DOTOMP_STRIDE]], align 4
// CHECK1-NEXT:    [[ADD9:%.*]] = add nsw i32 [[TMP37]], [[TMP38]]
// CHECK1-NEXT:    store i32 [[ADD9]], ptr [[DOTOMP_COMB_UB]], align 4
// CHECK1-NEXT:    [[TMP39:%.*]] = load i32, ptr [[DOTOMP_COMB_UB]], align 4
// CHECK1-NEXT:    [[TMP40:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_2]], align 4
// CHECK1-NEXT:    [[CMP10:%.*]] = icmp sgt i32 [[TMP39]], [[TMP40]]
// CHECK1-NEXT:    br i1 [[CMP10]], label [[COND_TRUE11:%.*]], label [[COND_FALSE12:%.*]]
// CHECK1:       cond.true11:
// CHECK1-NEXT:    [[TMP41:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_2]], align 4
// CHECK1-NEXT:    br label [[COND_END13:%.*]]
// CHECK1:       cond.false12:
// CHECK1-NEXT:    [[TMP42:%.*]] = load i32, ptr [[DOTOMP_COMB_UB]], align 4
// CHECK1-NEXT:    br label [[COND_END13]]
// CHECK1:       cond.end13:
// CHECK1-NEXT:    [[COND14:%.*]] = phi i32 [ [[TMP41]], [[COND_TRUE11]] ], [ [[TMP42]], [[COND_FALSE12]] ]
// CHECK1-NEXT:    store i32 [[COND14]], ptr [[DOTOMP_COMB_UB]], align 4
// CHECK1-NEXT:    [[TMP43:%.*]] = load i32, ptr [[DOTOMP_COMB_LB]], align 4
// CHECK1-NEXT:    store i32 [[TMP43]], ptr [[DOTOMP_IV]], align 4
// CHECK1-NEXT:    br label [[OMP_INNER_FOR_COND]]
// CHECK1:       omp.inner.for.end:
// CHECK1-NEXT:    br label [[OMP_LOOP_EXIT:%.*]]
// CHECK1:       omp.loop.exit:
// CHECK1-NEXT:    [[TMP44:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK1-NEXT:    [[TMP45:%.*]] = load i32, ptr [[TMP44]], align 4
// CHECK1-NEXT:    call void @__kmpc_distribute_static_fini(ptr @[[GLOB2]], i32 [[TMP45]])
// CHECK1-NEXT:    br label [[OMP_PRECOND_END]]
// CHECK1:       omp.precond.end:
// CHECK1-NEXT:    ret void
//
//
// CHECK1-LABEL: define {{[^@]+}}@__omp_outlined__1
// CHECK1-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]], i64 noundef [[DOTPREVIOUS_LB_:%.*]], i64 noundef [[DOTPREVIOUS_UB_:%.*]], i64 noundef [[ARGC:%.*]], ptr noundef nonnull align 4 dereferenceable(4) [[A:%.*]], i64 noundef [[DOTCAPTURE_EXPR_:%.*]]) #[[ATTR1]] {
// CHECK1-NEXT:  entry:
// CHECK1-NEXT:    [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    [[DOTPREVIOUS_LB__ADDR:%.*]] = alloca i64, align 8
// CHECK1-NEXT:    [[DOTPREVIOUS_UB__ADDR:%.*]] = alloca i64, align 8
// CHECK1-NEXT:    [[ARGC_ADDR:%.*]] = alloca i64, align 8
// CHECK1-NEXT:    [[A_ADDR:%.*]] = alloca ptr, align 8
// CHECK1-NEXT:    [[DOTCAPTURE_EXPR__ADDR:%.*]] = alloca i64, align 8
// CHECK1-NEXT:    [[DOTOMP_IV:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[TMP:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[DOTCAPTURE_EXPR_1:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[DOTCAPTURE_EXPR_2:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[I:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[DOTOMP_LB:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[DOTOMP_UB:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[DOTOMP_STRIDE:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[DOTOMP_IS_LAST:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    [[I5:%.*]] = alloca i32, align 4
// CHECK1-NEXT:    store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK1-NEXT:    store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 8
// CHECK1-NEXT:    store i64 [[DOTPREVIOUS_LB_]], ptr [[DOTPREVIOUS_LB__ADDR]], align 8
// CHECK1-NEXT:    store i64 [[DOTPREVIOUS_UB_]], ptr [[DOTPREVIOUS_UB__ADDR]], align 8
// CHECK1-NEXT:    store i64 [[ARGC]], ptr [[ARGC_ADDR]], align 8
// CHECK1-NEXT:    store ptr [[A]], ptr [[A_ADDR]], align 8
// CHECK1-NEXT:    store i64 [[DOTCAPTURE_EXPR_]], ptr [[DOTCAPTURE_EXPR__ADDR]], align 8
// CHECK1-NEXT:    [[TMP0:%.*]] = load ptr, ptr [[A_ADDR]], align 8
// CHECK1-NEXT:    [[TMP1:%.*]] = load i32, ptr [[ARGC_ADDR]], align 4
// CHECK1-NEXT:    store i32 [[TMP1]], ptr [[DOTCAPTURE_EXPR_1]], align 4
// CHECK1-NEXT:    [[TMP2:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_1]], align 4
// CHECK1-NEXT:    [[SUB:%.*]] = sub nsw i32 [[TMP2]], 0
// CHECK1-NEXT:    [[DIV:%.*]] = sdiv i32 [[SUB]], 1
// CHECK1-NEXT:    [[SUB3:%.*]] = sub nsw i32 [[DIV]], 1
// CHECK1-NEXT:    store i32 [[SUB3]], ptr [[DOTCAPTURE_EXPR_2]], align 4
// CHECK1-NEXT:    store i32 0, ptr [[I]], align 4
// CHECK1-NEXT:    [[TMP3:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_1]], align 4
// CHECK1-NEXT:    [[CMP:%.*]] = icmp slt i32 0, [[TMP3]]
// CHECK1-NEXT:    br i1 [[CMP]], label [[OMP_PRECOND_THEN:%.*]], label [[OMP_PRECOND_END:%.*]]
// CHECK1:       omp.precond.then:
// CHECK1-NEXT:    store i32 0, ptr [[DOTOMP_LB]], align 4
// CHECK1-NEXT:    [[TMP4:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_2]], align 4
// CHECK1-NEXT:    store i32 [[TMP4]], ptr [[DOTOMP_UB]], align 4
// CHECK1-NEXT:    [[TMP5:%.*]] = load i64, ptr [[DOTPREVIOUS_LB__ADDR]], align 8
// CHECK1-NEXT:    [[CONV:%.*]] = trunc i64 [[TMP5]] to i32
// CHECK1-NEXT:    [[TMP6:%.*]] = load i64, ptr [[DOTPREVIOUS_UB__ADDR]], align 8
// CHECK1-NEXT:    [[CONV4:%.*]] = trunc i64 [[TMP6]] to i32
// CHECK1-NEXT:    store i32 [[CONV]], ptr [[DOTOMP_LB]], align 4
// CHECK1-NEXT:    store i32 [[CONV4]], ptr [[DOTOMP_UB]], align 4
// CHECK1-NEXT:    store i32 1, ptr [[DOTOMP_STRIDE]], align 4
// CHECK1-NEXT:    store i32 0, ptr [[DOTOMP_IS_LAST]], align 4
// CHECK1-NEXT:    [[TMP7:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR__ADDR]], align 4
// CHECK1-NEXT:    [[TMP8:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK1-NEXT:    [[TMP9:%.*]] = load i32, ptr [[TMP8]], align 4
// CHECK1-NEXT:    call void @__kmpc_for_static_init_4(ptr @[[GLOB3:[0-9]+]], i32 [[TMP9]], i32 33, ptr [[DOTOMP_IS_LAST]], ptr [[DOTOMP_LB]], ptr [[DOTOMP_UB]], ptr [[DOTOMP_STRIDE]], i32 1, i32 [[TMP7]])
// CHECK1-NEXT:    br label [[OMP_DISPATCH_COND:%.*]]
// CHECK1:       omp.dispatch.cond:
// CHECK1-NEXT:    [[TMP10:%.*]] = load i32, ptr [[DOTOMP_UB]], align 4
// CHECK1-NEXT:    [[TMP11:%.*]] = load i64, ptr [[DOTPREVIOUS_UB__ADDR]], align 8
// CHECK1-NEXT:    [[CONV6:%.*]] = trunc i64 [[TMP11]] to i32
// CHECK1-NEXT:    [[CMP7:%.*]] = icmp sgt i32 [[TMP10]], [[CONV6]]
// CHECK1-NEXT:    br i1 [[CMP7]], label [[COND_TRUE:%.*]], label [[COND_FALSE:%.*]]
// CHECK1:       cond.true:
// CHECK1-NEXT:    [[TMP12:%.*]] = load i64, ptr [[DOTPREVIOUS_UB__ADDR]], align 8
// CHECK1-NEXT:    [[CONV8:%.*]] = trunc i64 [[TMP12]] to i32
// CHECK1-NEXT:    br label [[COND_END:%.*]]
// CHECK1:       cond.false:
// CHECK1-NEXT:    [[TMP13:%.*]] = load i32, ptr [[DOTOMP_UB]], align 4
// CHECK1-NEXT:    br label [[COND_END]]
// CHECK1:       cond.end:
// CHECK1-NEXT:    [[COND:%.*]] = phi i32 [ [[CONV8]], [[COND_TRUE]] ], [ [[TMP13]], [[COND_FALSE]] ]
// CHECK1-NEXT:    store i32 [[COND]], ptr [[DOTOMP_UB]], align 4
// CHECK1-NEXT:    [[TMP14:%.*]] = load i32, ptr [[DOTOMP_LB]], align 4
// CHECK1-NEXT:    store i32 [[TMP14]], ptr [[DOTOMP_IV]], align 4
// CHECK1-NEXT:    [[TMP15:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4
// CHECK1-NEXT:    [[TMP16:%.*]] = load i32, ptr [[DOTOMP_UB]], align 4
// CHECK1-NEXT:    [[CMP9:%.*]] = icmp sle i32 [[TMP15]], [[TMP16]]
// CHECK1-NEXT:    br i1 [[CMP9]], label [[OMP_DISPATCH_BODY:%.*]], label [[OMP_DISPATCH_END:%.*]]
// CHECK1:       omp.dispatch.body:
// CHECK1-NEXT:    br label [[OMP_INNER_FOR_COND:%.*]]
// CHECK1:       omp.inner.for.cond:
// CHECK1-NEXT:    [[TMP17:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4
// CHECK1-NEXT:    [[TMP18:%.*]] = load i32, ptr [[DOTOMP_UB]], align 4
// CHECK1-NEXT:    [[CMP10:%.*]] = icmp sle i32 [[TMP17]], [[TMP18]]
// CHECK1-NEXT:    br i1 [[CMP10]], label [[OMP_INNER_FOR_BODY:%.*]], label [[OMP_INNER_FOR_END:%.*]]
// CHECK1:       omp.inner.for.body:
// CHECK1-NEXT:    [[TMP19:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4
// CHECK1-NEXT:    [[MUL:%.*]] = mul nsw i32 [[TMP19]], 1
// CHECK1-NEXT:    [[ADD:%.*]] = add nsw i32 0, [[MUL]]
// CHECK1-NEXT:    store i32 [[ADD]], ptr [[I5]], align 4
// CHECK1-NEXT:    [[CALL:%.*]] = call noundef i32 @_Z3fooPi(ptr noundef [[I5]]) #[[ATTR5:[0-9]+]]
// CHECK1-NEXT:    [[CALL11:%.*]] = call noundef i32 @_Z3fooPi(ptr noundef [[TMP0]]) #[[ATTR5]]
// CHECK1-NEXT:    [[ADD12:%.*]] = add nsw i32 [[CALL]], [[CALL11]]
// CHECK1-NEXT:    [[CALL13:%.*]] = call noundef i32 @_Z3fooPi(ptr noundef [[ARGC_ADDR]]) #[[ATTR5]]
// CHECK1-NEXT:    [[ADD14:%.*]] = add nsw i32 [[ADD12]], [[CALL13]]
// CHECK1-NEXT:    store i32 [[ADD14]], ptr [[TMP0]], align 4
// CHECK1-NEXT:    br label [[OMP_BODY_CONTINUE:%.*]]
// CHECK1:       omp.body.continue:
// CHECK1-NEXT:    br label [[OMP_INNER_FOR_INC:%.*]]
// CHECK1:       omp.inner.for.inc:
// CHECK1-NEXT:    [[TMP20:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4
// CHECK1-NEXT:    [[ADD15:%.*]] = add nsw i32 [[TMP20]], 1
// CHECK1-NEXT:    store i32 [[ADD15]], ptr [[DOTOMP_IV]], align 4
// CHECK1-NEXT:    br label [[OMP_INNER_FOR_COND]]
// CHECK1:       omp.inner.for.end:
// CHECK1-NEXT:    br label [[OMP_DISPATCH_INC:%.*]]
// CHECK1:       omp.dispatch.inc:
// CHECK1-NEXT:    [[TMP21:%.*]] = load i32, ptr [[DOTOMP_LB]], align 4
// CHECK1-NEXT:    [[TMP22:%.*]] = load i32, ptr [[DOTOMP_STRIDE]], align 4
// CHECK1-NEXT:    [[ADD16:%.*]] = add nsw i32 [[TMP21]], [[TMP22]]
// CHECK1-NEXT:    store i32 [[ADD16]], ptr [[DOTOMP_LB]], align 4
// CHECK1-NEXT:    [[TMP23:%.*]] = load i32, ptr [[DOTOMP_UB]], align 4
// CHECK1-NEXT:    [[TMP24:%.*]] = load i32, ptr [[DOTOMP_STRIDE]], align 4
// CHECK1-NEXT:    [[ADD17:%.*]] = add nsw i32 [[TMP23]], [[TMP24]]
// CHECK1-NEXT:    store i32 [[ADD17]], ptr [[DOTOMP_UB]], align 4
// CHECK1-NEXT:    br label [[OMP_DISPATCH_COND]]
// CHECK1:       omp.dispatch.end:
// CHECK1-NEXT:    [[TMP25:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 8
// CHECK1-NEXT:    [[TMP26:%.*]] = load i32, ptr [[TMP25]], align 4
// CHECK1-NEXT:    call void @__kmpc_distribute_static_fini(ptr @[[GLOB2]], i32 [[TMP26]])
// CHECK1-NEXT:    br label [[OMP_PRECOND_END]]
// CHECK1:       omp.precond.end:
// CHECK1-NEXT:    ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}_main_l24
// CHECK2-SAME: (i32 noundef [[ARGC:%.*]], ptr noundef nonnull align 4 dereferenceable(4) [[A:%.*]], i32 noundef [[DOTCAPTURE_EXPR_:%.*]]) #[[ATTR0:[0-9]+]] {
// CHECK2-NEXT:  entry:
// CHECK2-NEXT:    [[ARGC_ADDR:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[A_ADDR:%.*]] = alloca ptr, align 4
// CHECK2-NEXT:    [[DOTCAPTURE_EXPR__ADDR:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[ARGC_CASTED:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[DOTCAPTURE_EXPR__CASTED:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[DOTZERO_ADDR:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    store i32 [[ARGC]], ptr [[ARGC_ADDR]], align 4
// CHECK2-NEXT:    store ptr [[A]], ptr [[A_ADDR]], align 4
// CHECK2-NEXT:    store i32 [[DOTCAPTURE_EXPR_]], ptr [[DOTCAPTURE_EXPR__ADDR]], align 4
// CHECK2-NEXT:    [[TMP0:%.*]] = load ptr, ptr [[A_ADDR]], align 4
// CHECK2-NEXT:    [[TMP1:%.*]] = call i32 @__kmpc_target_init(ptr @[[GLOB1:[0-9]+]], i8 2, 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(ptr @[[GLOB1]])
// CHECK2-NEXT:    [[TMP3:%.*]] = load i32, ptr [[ARGC_ADDR]], align 4
// CHECK2-NEXT:    store i32 [[TMP3]], ptr [[ARGC_CASTED]], align 4
// CHECK2-NEXT:    [[TMP4:%.*]] = load i32, ptr [[ARGC_CASTED]], align 4
// CHECK2-NEXT:    [[TMP5:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR__ADDR]], align 4
// CHECK2-NEXT:    store i32 [[TMP5]], ptr [[DOTCAPTURE_EXPR__CASTED]], align 4
// CHECK2-NEXT:    [[TMP6:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR__CASTED]], align 4
// CHECK2-NEXT:    store i32 0, ptr [[DOTZERO_ADDR]], align 4
// CHECK2-NEXT:    store i32 [[TMP2]], ptr [[DOTTHREADID_TEMP_]], align 4
// CHECK2-NEXT:    call void @__omp_outlined__(ptr [[DOTTHREADID_TEMP_]], ptr [[DOTZERO_ADDR]], i32 [[TMP4]], ptr [[TMP0]], i32 [[TMP6]]) #[[ATTR3:[0-9]+]]
// CHECK2-NEXT:    call void @__kmpc_target_deinit(ptr @[[GLOB1]], i8 2)
// CHECK2-NEXT:    ret void
// CHECK2:       worker.exit:
// CHECK2-NEXT:    ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@__omp_outlined__
// CHECK2-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]], i32 noundef [[ARGC:%.*]], ptr noundef nonnull align 4 dereferenceable(4) [[A:%.*]], i32 noundef [[DOTCAPTURE_EXPR_:%.*]]) #[[ATTR1:[0-9]+]] {
// CHECK2-NEXT:  entry:
// CHECK2-NEXT:    [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 4
// CHECK2-NEXT:    [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 4
// CHECK2-NEXT:    [[ARGC_ADDR:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[A_ADDR:%.*]] = alloca ptr, align 4
// CHECK2-NEXT:    [[DOTCAPTURE_EXPR__ADDR:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[DOTOMP_IV:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[TMP:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[DOTCAPTURE_EXPR_1:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[DOTCAPTURE_EXPR_2:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[I:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[DOTOMP_COMB_LB:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[DOTOMP_COMB_UB:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[DOTOMP_STRIDE:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[DOTOMP_IS_LAST:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[I4:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[ARGC_CASTED:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[DOTCAPTURE_EXPR__CASTED:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[CAPTURED_VARS_ADDRS:%.*]] = alloca [5 x ptr], align 4
// CHECK2-NEXT:    store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 4
// CHECK2-NEXT:    store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 4
// CHECK2-NEXT:    store i32 [[ARGC]], ptr [[ARGC_ADDR]], align 4
// CHECK2-NEXT:    store ptr [[A]], ptr [[A_ADDR]], align 4
// CHECK2-NEXT:    store i32 [[DOTCAPTURE_EXPR_]], ptr [[DOTCAPTURE_EXPR__ADDR]], align 4
// CHECK2-NEXT:    [[TMP0:%.*]] = load ptr, ptr [[A_ADDR]], align 4
// CHECK2-NEXT:    [[TMP1:%.*]] = load i32, ptr [[ARGC_ADDR]], align 4
// CHECK2-NEXT:    store i32 [[TMP1]], ptr [[DOTCAPTURE_EXPR_1]], align 4
// CHECK2-NEXT:    [[TMP2:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_1]], align 4
// CHECK2-NEXT:    [[SUB:%.*]] = sub nsw i32 [[TMP2]], 0
// CHECK2-NEXT:    [[DIV:%.*]] = sdiv i32 [[SUB]], 1
// CHECK2-NEXT:    [[SUB3:%.*]] = sub nsw i32 [[DIV]], 1
// CHECK2-NEXT:    store i32 [[SUB3]], ptr [[DOTCAPTURE_EXPR_2]], align 4
// CHECK2-NEXT:    store i32 0, ptr [[I]], align 4
// CHECK2-NEXT:    [[TMP3:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_1]], align 4
// CHECK2-NEXT:    [[CMP:%.*]] = icmp slt i32 0, [[TMP3]]
// CHECK2-NEXT:    br i1 [[CMP]], label [[OMP_PRECOND_THEN:%.*]], label [[OMP_PRECOND_END:%.*]]
// CHECK2:       omp.precond.then:
// CHECK2-NEXT:    store i32 0, ptr [[DOTOMP_COMB_LB]], align 4
// CHECK2-NEXT:    [[TMP4:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_2]], align 4
// CHECK2-NEXT:    store i32 [[TMP4]], ptr [[DOTOMP_COMB_UB]], align 4
// CHECK2-NEXT:    store i32 1, ptr [[DOTOMP_STRIDE]], align 4
// CHECK2-NEXT:    store i32 0, ptr [[DOTOMP_IS_LAST]], align 4
// CHECK2-NEXT:    [[NVPTX_NUM_THREADS:%.*]] = call i32 @__kmpc_get_hardware_num_threads_in_block()
// CHECK2-NEXT:    [[TMP5:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 4
// CHECK2-NEXT:    [[TMP6:%.*]] = load i32, ptr [[TMP5]], align 4
// CHECK2-NEXT:    call void @__kmpc_distribute_static_init_4(ptr @[[GLOB2:[0-9]+]], i32 [[TMP6]], i32 91, ptr [[DOTOMP_IS_LAST]], ptr [[DOTOMP_COMB_LB]], ptr [[DOTOMP_COMB_UB]], ptr [[DOTOMP_STRIDE]], i32 1, i32 [[NVPTX_NUM_THREADS]])
// CHECK2-NEXT:    [[TMP7:%.*]] = load i32, ptr [[DOTOMP_COMB_UB]], align 4
// CHECK2-NEXT:    [[TMP8:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_2]], align 4
// CHECK2-NEXT:    [[CMP5:%.*]] = icmp sgt i32 [[TMP7]], [[TMP8]]
// CHECK2-NEXT:    br i1 [[CMP5]], label [[COND_TRUE:%.*]], label [[COND_FALSE:%.*]]
// CHECK2:       cond.true:
// CHECK2-NEXT:    [[TMP9:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_2]], align 4
// CHECK2-NEXT:    br label [[COND_END:%.*]]
// CHECK2:       cond.false:
// CHECK2-NEXT:    [[TMP10:%.*]] = load i32, ptr [[DOTOMP_COMB_UB]], align 4
// CHECK2-NEXT:    br label [[COND_END]]
// CHECK2:       cond.end:
// CHECK2-NEXT:    [[COND:%.*]] = phi i32 [ [[TMP9]], [[COND_TRUE]] ], [ [[TMP10]], [[COND_FALSE]] ]
// CHECK2-NEXT:    store i32 [[COND]], ptr [[DOTOMP_COMB_UB]], align 4
// CHECK2-NEXT:    [[TMP11:%.*]] = load i32, ptr [[DOTOMP_COMB_LB]], align 4
// CHECK2-NEXT:    store i32 [[TMP11]], ptr [[DOTOMP_IV]], align 4
// CHECK2-NEXT:    br label [[OMP_INNER_FOR_COND:%.*]]
// CHECK2:       omp.inner.for.cond:
// CHECK2-NEXT:    [[TMP12:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4
// CHECK2-NEXT:    [[TMP13:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_2]], align 4
// CHECK2-NEXT:    [[ADD:%.*]] = add nsw i32 [[TMP13]], 1
// CHECK2-NEXT:    [[CMP6:%.*]] = icmp slt i32 [[TMP12]], [[ADD]]
// CHECK2-NEXT:    br i1 [[CMP6]], label [[OMP_INNER_FOR_BODY:%.*]], label [[OMP_INNER_FOR_END:%.*]]
// CHECK2:       omp.inner.for.body:
// CHECK2-NEXT:    [[TMP14:%.*]] = load i32, ptr [[DOTOMP_COMB_LB]], align 4
// CHECK2-NEXT:    [[TMP15:%.*]] = load i32, ptr [[DOTOMP_COMB_UB]], align 4
// CHECK2-NEXT:    [[TMP16:%.*]] = load i32, ptr [[ARGC_ADDR]], align 4
// CHECK2-NEXT:    store i32 [[TMP16]], ptr [[ARGC_CASTED]], align 4
// CHECK2-NEXT:    [[TMP17:%.*]] = load i32, ptr [[ARGC_CASTED]], align 4
// CHECK2-NEXT:    [[TMP18:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR__ADDR]], align 4
// CHECK2-NEXT:    store i32 [[TMP18]], ptr [[DOTCAPTURE_EXPR__CASTED]], align 4
// CHECK2-NEXT:    [[TMP19:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR__CASTED]], align 4
// CHECK2-NEXT:    [[TMP20:%.*]] = getelementptr inbounds [5 x ptr], ptr [[CAPTURED_VARS_ADDRS]], i32 0, i32 0
// CHECK2-NEXT:    [[TMP21:%.*]] = inttoptr i32 [[TMP14]] to ptr
// CHECK2-NEXT:    store ptr [[TMP21]], ptr [[TMP20]], align 4
// CHECK2-NEXT:    [[TMP22:%.*]] = getelementptr inbounds [5 x ptr], ptr [[CAPTURED_VARS_ADDRS]], i32 0, i32 1
// CHECK2-NEXT:    [[TMP23:%.*]] = inttoptr i32 [[TMP15]] to ptr
// CHECK2-NEXT:    store ptr [[TMP23]], ptr [[TMP22]], align 4
// CHECK2-NEXT:    [[TMP24:%.*]] = getelementptr inbounds [5 x ptr], ptr [[CAPTURED_VARS_ADDRS]], i32 0, i32 2
// CHECK2-NEXT:    [[TMP25:%.*]] = inttoptr i32 [[TMP17]] to ptr
// CHECK2-NEXT:    store ptr [[TMP25]], ptr [[TMP24]], align 4
// CHECK2-NEXT:    [[TMP26:%.*]] = getelementptr inbounds [5 x ptr], ptr [[CAPTURED_VARS_ADDRS]], i32 0, i32 3
// CHECK2-NEXT:    store ptr [[TMP0]], ptr [[TMP26]], align 4
// CHECK2-NEXT:    [[TMP27:%.*]] = getelementptr inbounds [5 x ptr], ptr [[CAPTURED_VARS_ADDRS]], i32 0, i32 4
// CHECK2-NEXT:    [[TMP28:%.*]] = inttoptr i32 [[TMP19]] to ptr
// CHECK2-NEXT:    store ptr [[TMP28]], ptr [[TMP27]], align 4
// CHECK2-NEXT:    [[TMP29:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 4
// CHECK2-NEXT:    [[TMP30:%.*]] = load i32, ptr [[TMP29]], align 4
// CHECK2-NEXT:    call void @__kmpc_parallel_51(ptr @[[GLOB1]], i32 [[TMP30]], i32 1, i32 -1, i32 -1, ptr @__omp_outlined__1, ptr null, ptr [[CAPTURED_VARS_ADDRS]], i32 5)
// CHECK2-NEXT:    br label [[OMP_INNER_FOR_INC:%.*]]
// CHECK2:       omp.inner.for.inc:
// CHECK2-NEXT:    [[TMP31:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4
// CHECK2-NEXT:    [[TMP32:%.*]] = load i32, ptr [[DOTOMP_STRIDE]], align 4
// CHECK2-NEXT:    [[ADD7:%.*]] = add nsw i32 [[TMP31]], [[TMP32]]
// CHECK2-NEXT:    store i32 [[ADD7]], ptr [[DOTOMP_IV]], align 4
// CHECK2-NEXT:    [[TMP33:%.*]] = load i32, ptr [[DOTOMP_COMB_LB]], align 4
// CHECK2-NEXT:    [[TMP34:%.*]] = load i32, ptr [[DOTOMP_STRIDE]], align 4
// CHECK2-NEXT:    [[ADD8:%.*]] = add nsw i32 [[TMP33]], [[TMP34]]
// CHECK2-NEXT:    store i32 [[ADD8]], ptr [[DOTOMP_COMB_LB]], align 4
// CHECK2-NEXT:    [[TMP35:%.*]] = load i32, ptr [[DOTOMP_COMB_UB]], align 4
// CHECK2-NEXT:    [[TMP36:%.*]] = load i32, ptr [[DOTOMP_STRIDE]], align 4
// CHECK2-NEXT:    [[ADD9:%.*]] = add nsw i32 [[TMP35]], [[TMP36]]
// CHECK2-NEXT:    store i32 [[ADD9]], ptr [[DOTOMP_COMB_UB]], align 4
// CHECK2-NEXT:    [[TMP37:%.*]] = load i32, ptr [[DOTOMP_COMB_UB]], align 4
// CHECK2-NEXT:    [[TMP38:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_2]], align 4
// CHECK2-NEXT:    [[CMP10:%.*]] = icmp sgt i32 [[TMP37]], [[TMP38]]
// CHECK2-NEXT:    br i1 [[CMP10]], label [[COND_TRUE11:%.*]], label [[COND_FALSE12:%.*]]
// CHECK2:       cond.true11:
// CHECK2-NEXT:    [[TMP39:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_2]], align 4
// CHECK2-NEXT:    br label [[COND_END13:%.*]]
// CHECK2:       cond.false12:
// CHECK2-NEXT:    [[TMP40:%.*]] = load i32, ptr [[DOTOMP_COMB_UB]], align 4
// CHECK2-NEXT:    br label [[COND_END13]]
// CHECK2:       cond.end13:
// CHECK2-NEXT:    [[COND14:%.*]] = phi i32 [ [[TMP39]], [[COND_TRUE11]] ], [ [[TMP40]], [[COND_FALSE12]] ]
// CHECK2-NEXT:    store i32 [[COND14]], ptr [[DOTOMP_COMB_UB]], align 4
// CHECK2-NEXT:    [[TMP41:%.*]] = load i32, ptr [[DOTOMP_COMB_LB]], align 4
// CHECK2-NEXT:    store i32 [[TMP41]], ptr [[DOTOMP_IV]], align 4
// CHECK2-NEXT:    br label [[OMP_INNER_FOR_COND]]
// CHECK2:       omp.inner.for.end:
// CHECK2-NEXT:    br label [[OMP_LOOP_EXIT:%.*]]
// CHECK2:       omp.loop.exit:
// CHECK2-NEXT:    [[TMP42:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 4
// CHECK2-NEXT:    [[TMP43:%.*]] = load i32, ptr [[TMP42]], align 4
// CHECK2-NEXT:    call void @__kmpc_distribute_static_fini(ptr @[[GLOB2]], i32 [[TMP43]])
// CHECK2-NEXT:    br label [[OMP_PRECOND_END]]
// CHECK2:       omp.precond.end:
// CHECK2-NEXT:    ret void
//
//
// CHECK2-LABEL: define {{[^@]+}}@__omp_outlined__1
// CHECK2-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]], i32 noundef [[DOTPREVIOUS_LB_:%.*]], i32 noundef [[DOTPREVIOUS_UB_:%.*]], i32 noundef [[ARGC:%.*]], ptr noundef nonnull align 4 dereferenceable(4) [[A:%.*]], i32 noundef [[DOTCAPTURE_EXPR_:%.*]]) #[[ATTR1]] {
// CHECK2-NEXT:  entry:
// CHECK2-NEXT:    [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 4
// CHECK2-NEXT:    [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 4
// CHECK2-NEXT:    [[DOTPREVIOUS_LB__ADDR:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[DOTPREVIOUS_UB__ADDR:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[ARGC_ADDR:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[A_ADDR:%.*]] = alloca ptr, align 4
// CHECK2-NEXT:    [[DOTCAPTURE_EXPR__ADDR:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[DOTOMP_IV:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[TMP:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[DOTCAPTURE_EXPR_1:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[DOTCAPTURE_EXPR_2:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[I:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[DOTOMP_LB:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[DOTOMP_UB:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[DOTOMP_STRIDE:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[DOTOMP_IS_LAST:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    [[I4:%.*]] = alloca i32, align 4
// CHECK2-NEXT:    store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR]], align 4
// CHECK2-NEXT:    store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR]], align 4
// CHECK2-NEXT:    store i32 [[DOTPREVIOUS_LB_]], ptr [[DOTPREVIOUS_LB__ADDR]], align 4
// CHECK2-NEXT:    store i32 [[DOTPREVIOUS_UB_]], ptr [[DOTPREVIOUS_UB__ADDR]], align 4
// CHECK2-NEXT:    store i32 [[ARGC]], ptr [[ARGC_ADDR]], align 4
// CHECK2-NEXT:    store ptr [[A]], ptr [[A_ADDR]], align 4
// CHECK2-NEXT:    store i32 [[DOTCAPTURE_EXPR_]], ptr [[DOTCAPTURE_EXPR__ADDR]], align 4
// CHECK2-NEXT:    [[TMP0:%.*]] = load ptr, ptr [[A_ADDR]], align 4
// CHECK2-NEXT:    [[TMP1:%.*]] = load i32, ptr [[ARGC_ADDR]], align 4
// CHECK2-NEXT:    store i32 [[TMP1]], ptr [[DOTCAPTURE_EXPR_1]], align 4
// CHECK2-NEXT:    [[TMP2:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_1]], align 4
// CHECK2-NEXT:    [[SUB:%.*]] = sub nsw i32 [[TMP2]], 0
// CHECK2-NEXT:    [[DIV:%.*]] = sdiv i32 [[SUB]], 1
// CHECK2-NEXT:    [[SUB3:%.*]] = sub nsw i32 [[DIV]], 1
// CHECK2-NEXT:    store i32 [[SUB3]], ptr [[DOTCAPTURE_EXPR_2]], align 4
// CHECK2-NEXT:    store i32 0, ptr [[I]], align 4
// CHECK2-NEXT:    [[TMP3:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_1]], align 4
// CHECK2-NEXT:    [[CMP:%.*]] = icmp slt i32 0, [[TMP3]]
// CHECK2-NEXT:    br i1 [[CMP]], label [[OMP_PRECOND_THEN:%.*]], label [[OMP_PRECOND_END:%.*]]
// CHECK2:       omp.precond.then:
// CHECK2-NEXT:    store i32 0, ptr [[DOTOMP_LB]], align 4
// CHECK2-NEXT:    [[TMP4:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_2]], align 4
// CHECK2-NEXT:    store i32 [[TMP4]], ptr [[DOTOMP_UB]], align 4
// CHECK2-NEXT:    [[TMP5:%.*]] = load i32, ptr [[DOTPREVIOUS_LB__ADDR]], align 4
// CHECK2-NEXT:    [[TMP6:%.*]] = load i32, ptr [[DOTPREVIOUS_UB__ADDR]], align 4
// CHECK2-NEXT:    store i32 [[TMP5]], ptr [[DOTOMP_LB]], align 4
// CHECK2-NEXT:    store i32 [[TMP6]], ptr [[DOTOMP_UB]], align 4
// CHECK2-NEXT:    store i32 1, ptr [[DOTOMP_STRIDE]], align 4
// CHECK2-NEXT:    store i32 0, ptr [[DOTOMP_IS_LAST]], align 4
// CHECK2-NEXT:    [[TMP7:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR__ADDR]], align 4
// CHECK2-NEXT:    [[TMP8:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 4
// CHECK2-NEXT:    [[TMP9:%.*]] = load i32, ptr [[TMP8]], align 4
// CHECK2-NEXT:    call void @__kmpc_for_static_init_4(ptr @[[GLOB3:[0-9]+]], i32 [[TMP9]], i32 33, ptr [[DOTOMP_IS_LAST]], ptr [[DOTOMP_LB]], ptr [[DOTOMP_UB]], ptr [[DOTOMP_STRIDE]], i32 1, i32 [[TMP7]])
// CHECK2-NEXT:    br label [[OMP_DISPATCH_COND:%.*]]
// CHECK2:       omp.dispatch.cond:
// CHECK2-NEXT:    [[TMP10:%.*]] = load i32, ptr [[DOTOMP_UB]], align 4
// CHECK2-NEXT:    [[TMP11:%.*]] = load i32, ptr [[DOTPREVIOUS_UB__ADDR]], align 4
// CHECK2-NEXT:    [[CMP5:%.*]] = icmp sgt i32 [[TMP10]], [[TMP11]]
// CHECK2-NEXT:    br i1 [[CMP5]], label [[COND_TRUE:%.*]], label [[COND_FALSE:%.*]]
// CHECK2:       cond.true:
// CHECK2-NEXT:    [[TMP12:%.*]] = load i32, ptr [[DOTPREVIOUS_UB__ADDR]], align 4
// CHECK2-NEXT:    br label [[COND_END:%.*]]
// CHECK2:       cond.false:
// CHECK2-NEXT:    [[TMP13:%.*]] = load i32, ptr [[DOTOMP_UB]], align 4
// CHECK2-NEXT:    br label [[COND_END]]
// CHECK2:       cond.end:
// CHECK2-NEXT:    [[COND:%.*]] = phi i32 [ [[TMP12]], [[COND_TRUE]] ], [ [[TMP13]], [[COND_FALSE]] ]
// CHECK2-NEXT:    store i32 [[COND]], ptr [[DOTOMP_UB]], align 4
// CHECK2-NEXT:    [[TMP14:%.*]] = load i32, ptr [[DOTOMP_LB]], align 4
// CHECK2-NEXT:    store i32 [[TMP14]], ptr [[DOTOMP_IV]], align 4
// CHECK2-NEXT:    [[TMP15:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4
// CHECK2-NEXT:    [[TMP16:%.*]] = load i32, ptr [[DOTOMP_UB]], align 4
// CHECK2-NEXT:    [[CMP6:%.*]] = icmp sle i32 [[TMP15]], [[TMP16]]
// CHECK2-NEXT:    br i1 [[CMP6]], label [[OMP_DISPATCH_BODY:%.*]], label [[OMP_DISPATCH_END:%.*]]
// CHECK2:       omp.dispatch.body:
// CHECK2-NEXT:    br label [[OMP_INNER_FOR_COND:%.*]]
// CHECK2:       omp.inner.for.cond:
// CHECK2-NEXT:    [[TMP17:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4
// CHECK2-NEXT:    [[TMP18:%.*]] = load i32, ptr [[DOTOMP_UB]], align 4
// CHECK2-NEXT:    [[CMP7:%.*]] = icmp sle i32 [[TMP17]], [[TMP18]]
// CHECK2-NEXT:    br i1 [[CMP7]], label [[OMP_INNER_FOR_BODY:%.*]], label [[OMP_INNER_FOR_END:%.*]]
// CHECK2:       omp.inner.for.body:
// CHECK2-NEXT:    [[TMP19:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4
// CHECK2-NEXT:    [[MUL:%.*]] = mul nsw i32 [[TMP19]], 1
// CHECK2-NEXT:    [[ADD:%.*]] = add nsw i32 0, [[MUL]]
// CHECK2-NEXT:    store i32 [[ADD]], ptr [[I4]], align 4
// CHECK2-NEXT:    [[CALL:%.*]] = call noundef i32 @_Z3fooPi(ptr noundef [[I4]]) #[[ATTR5:[0-9]+]]
// CHECK2-NEXT:    [[CALL8:%.*]] = call noundef i32 @_Z3fooPi(ptr noundef [[TMP0]]) #[[ATTR5]]
// CHECK2-NEXT:    [[ADD9:%.*]] = add nsw i32 [[CALL]], [[CALL8]]
// CHECK2-NEXT:    [[CALL10:%.*]] = call noundef i32 @_Z3fooPi(ptr noundef [[ARGC_ADDR]]) #[[ATTR5]]
// CHECK2-NEXT:    [[ADD11:%.*]] = add nsw i32 [[ADD9]], [[CALL10]]
// CHECK2-NEXT:    store i32 [[ADD11]], ptr [[TMP0]], align 4
// CHECK2-NEXT:    br label [[OMP_BODY_CONTINUE:%.*]]
// CHECK2:       omp.body.continue:
// CHECK2-NEXT:    br label [[OMP_INNER_FOR_INC:%.*]]
// CHECK2:       omp.inner.for.inc:
// CHECK2-NEXT:    [[TMP20:%.*]] = load i32, ptr [[DOTOMP_IV]], align 4
// CHECK2-NEXT:    [[ADD12:%.*]] = add nsw i32 [[TMP20]], 1
// CHECK2-NEXT:    store i32 [[ADD12]], ptr [[DOTOMP_IV]], align 4
// CHECK2-NEXT:    br label [[OMP_INNER_FOR_COND]]
// CHECK2:       omp.inner.for.end:
// CHECK2-NEXT:    br label [[OMP_DISPATCH_INC:%.*]]
// CHECK2:       omp.dispatch.inc:
// CHECK2-NEXT:    [[TMP21:%.*]] = load i32, ptr [[DOTOMP_LB]], align 4
// CHECK2-NEXT:    [[TMP22:%.*]] = load i32, ptr [[DOTOMP_STRIDE]], align 4
// CHECK2-NEXT:    [[ADD13:%.*]] = add nsw i32 [[TMP21]], [[TMP22]]
// CHECK2-NEXT:    store i32 [[ADD13]], ptr [[DOTOMP_LB]], align 4
// CHECK2-NEXT:    [[TMP23:%.*]] = load i32, ptr [[DOTOMP_UB]], align 4
// CHECK2-NEXT:    [[TMP24:%.*]] = load i32, ptr [[DOTOMP_STRIDE]], align 4
// CHECK2-NEXT:    [[ADD14:%.*]] = add nsw i32 [[TMP23]], [[TMP24]]
// CHECK2-NEXT:    store i32 [[ADD14]], ptr [[DOTOMP_UB]], align 4
// CHECK2-NEXT:    br label [[OMP_DISPATCH_COND]]
// CHECK2:       omp.dispatch.end:
// CHECK2-NEXT:    [[TMP25:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR]], align 4
// CHECK2-NEXT:    [[TMP26:%.*]] = load i32, ptr [[TMP25]], align 4
// CHECK2-NEXT:    call void @__kmpc_distribute_static_fini(ptr @[[GLOB2]], i32 [[TMP26]])
// CHECK2-NEXT:    br label [[OMP_PRECOND_END]]
// CHECK2:       omp.precond.end:
// CHECK2-NEXT:    ret void
//