clang/test/OpenMP/amdgcn_target_device_vla.cpp

   1 // 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 _
   2 // REQUIRES: amdgpu-registered-target
   3
   4 // RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple x86_64-unknown-unknown -fopenmp-targets=amdgcn-amd-amdhsa -emit-llvm-bc %s -o %t-ppc-host.bc
   5 // RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple amdgcn-amd-amdhsa -fopenmp-targets=amdgcn-amd-amdhsa -emit-llvm %s -fopenmp-is-target-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o - | FileCheck %s
   6 // expected-no-diagnostics
   7 #ifndef HEADER
   8 #define HEADER
   9
  10 int foo1() {
  11   int sum = 0.0;
  12   #pragma omp target map(tofrom: sum)
  13   {
  14     int N = 10;
  15     int A[N];
  16
  17     for (int i = 0; i < N; i++)
  18       A[i] = i;
  19
  20     for (int i = 0; i < N; i++)
  21       sum += A[i];
  22   }
  23   return sum;
  24 }
  25
  26 int foo2() {
  27   int sum = 0.0;
  28   int M = 12;
  29   int result[M];
  30   #pragma omp target teams distribute parallel for map(from: result[:M])
  31   for (int i = 0; i < M; i++) {
  32     int N = 10;
  33     int A[N];
  34     result[i] = i;
  35
  36     for (int j = 0; j < N; j++)
  37       A[j] = j;
  38
  39     for (int j = 0; j < N; j++)
  40       result[i] += A[j];
  41   }
  42
  43   for (int i = 0; i < M; i++)
  44     sum += result[i];
  45   return sum;
  46 }
  47
  48 int foo3() {
  49   int sum = 0.0;
  50   int M = 12;
  51   int result[M];
  52   #pragma omp target teams distribute map(from: result[:M])
  53   for (int i = 0; i < M; i++) {
  54     int N = 10;
  55     int A[N];
  56     result[i] = i;
  57
  58     #pragma omp parallel for
  59     for (int j = 0; j < N; j++)
  60       A[j] = j;
  61
  62     for (int j = 0; j < N; j++)
  63       result[i] += A[j];
  64   }
  65
  66   for (int i = 0; i < M; i++)
  67     sum += result[i];
  68   return sum;
  69 }
  70
  71 int foo4() {
  72   int sum = 0.0;
  73   int M = 12;
  74   int result[M];
  75   int N = 10;
  76   #pragma omp target teams distribute map(from: result[:M])
  77   for (int i = 0; i < M; i++) {
  78     int A[N];
  79     result[i] = i;
  80
  81     #pragma omp parallel for
  82     for (int j = 0; j < N; j++)
  83       A[j] = j;
  84
  85     for (int j = 0; j < N; j++)
  86       result[i] += A[j];
  87   }
  88
  89   for (int i = 0; i < M; i++)
  90     sum += result[i];
  91   return sum;
  92 }
  93
  94 int main() {
  95   return foo1() + foo2() + foo3() + foo4();
  96 }
  97
  98 #endif
  99 // CHECK-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z4foo1v_l12
 100 // CHECK-SAME: (ptr noalias noundef [[DYN_PTR:%.*]], ptr noundef nonnull align 4 dereferenceable(4) [[SUM:%.*]]) #[[ATTR0:[0-9]+]] {
 101 // CHECK-NEXT:  entry:
 102 // CHECK-NEXT:    [[DYN_PTR_ADDR:%.*]] = alloca ptr, align 8, addrspace(5)
 103 // CHECK-NEXT:    [[SUM_ADDR:%.*]] = alloca ptr, align 8, addrspace(5)
 104 // CHECK-NEXT:    [[N:%.*]] = alloca i32, align 4, addrspace(5)
 105 // CHECK-NEXT:    [[__VLA_EXPR0:%.*]] = alloca i64, align 8, addrspace(5)
 106 // CHECK-NEXT:    [[I:%.*]] = alloca i32, align 4, addrspace(5)
 107 // CHECK-NEXT:    [[I1:%.*]] = alloca i32, align 4, addrspace(5)
 108 // CHECK-NEXT:    [[DYN_PTR_ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DYN_PTR_ADDR]] to ptr
 109 // CHECK-NEXT:    [[SUM_ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[SUM_ADDR]] to ptr
 110 // CHECK-NEXT:    [[N_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[N]] to ptr
 111 // CHECK-NEXT:    [[__VLA_EXPR0_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[__VLA_EXPR0]] to ptr
 112 // CHECK-NEXT:    [[I_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[I]] to ptr
 113 // CHECK-NEXT:    [[I1_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[I1]] to ptr
 114 // CHECK-NEXT:    store ptr [[DYN_PTR]], ptr [[DYN_PTR_ADDR_ASCAST]], align 8
 115 // CHECK-NEXT:    store ptr [[SUM]], ptr [[SUM_ADDR_ASCAST]], align 8
 116 // CHECK-NEXT:    [[TMP0:%.*]] = load ptr, ptr [[SUM_ADDR_ASCAST]], align 8
 117 // CHECK-NEXT:    [[TMP1:%.*]] = call i32 @__kmpc_target_init(ptr addrspacecast (ptr addrspace(1) @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z4foo1v_l12_kernel_environment to ptr), ptr [[DYN_PTR]])
 118 // CHECK-NEXT:    [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP1]], -1
 119 // CHECK-NEXT:    br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]]
 120 // CHECK:       user_code.entry:
 121 // CHECK-NEXT:    store i32 10, ptr [[N_ASCAST]], align 4
 122 // CHECK-NEXT:    [[TMP2:%.*]] = load i32, ptr [[N_ASCAST]], align 4
 123 // CHECK-NEXT:    [[TMP3:%.*]] = zext i32 [[TMP2]] to i64
 124 // CHECK-NEXT:    [[TMP4:%.*]] = mul nuw i64 [[TMP3]], 4
 125 // CHECK-NEXT:    [[TMP5:%.*]] = add nuw i64 [[TMP4]], 3
 126 // CHECK-NEXT:    [[TMP6:%.*]] = udiv i64 [[TMP5]], 4
 127 // CHECK-NEXT:    [[TMP7:%.*]] = mul nuw i64 [[TMP6]], 4
 128 // CHECK-NEXT:    [[A:%.*]] = call align 4 ptr @__kmpc_alloc_shared(i64 [[TMP7]])
 129 // CHECK-NEXT:    store i64 [[TMP3]], ptr [[__VLA_EXPR0_ASCAST]], align 8
 130 // CHECK-NEXT:    store i32 0, ptr [[I_ASCAST]], align 4
 131 // CHECK-NEXT:    br label [[FOR_COND:%.*]]
 132 // CHECK:       for.cond:
 133 // CHECK-NEXT:    [[TMP8:%.*]] = load i32, ptr [[I_ASCAST]], align 4
 134 // CHECK-NEXT:    [[TMP9:%.*]] = load i32, ptr [[N_ASCAST]], align 4
 135 // CHECK-NEXT:    [[CMP:%.*]] = icmp slt i32 [[TMP8]], [[TMP9]]
 136 // CHECK-NEXT:    br i1 [[CMP]], label [[FOR_BODY:%.*]], label [[FOR_END:%.*]]
 137 // CHECK:       for.body:
 138 // CHECK-NEXT:    [[TMP10:%.*]] = load i32, ptr [[I_ASCAST]], align 4
 139 // CHECK-NEXT:    [[TMP11:%.*]] = load i32, ptr [[I_ASCAST]], align 4
 140 // CHECK-NEXT:    [[IDXPROM:%.*]] = sext i32 [[TMP11]] to i64
 141 // CHECK-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds i32, ptr [[A]], i64 [[IDXPROM]]
 142 // CHECK-NEXT:    store i32 [[TMP10]], ptr [[ARRAYIDX]], align 4
 143 // CHECK-NEXT:    br label [[FOR_INC:%.*]]
 144 // CHECK:       for.inc:
 145 // CHECK-NEXT:    [[TMP12:%.*]] = load i32, ptr [[I_ASCAST]], align 4
 146 // CHECK-NEXT:    [[INC:%.*]] = add nsw i32 [[TMP12]], 1
 147 // CHECK-NEXT:    store i32 [[INC]], ptr [[I_ASCAST]], align 4
 148 // CHECK-NEXT:    br label [[FOR_COND]], !llvm.loop [[LOOP13:![0-9]+]]
 149 // CHECK:       worker.exit:
 150 // CHECK-NEXT:    ret void
 151 // CHECK:       for.end:
 152 // CHECK-NEXT:    store i32 0, ptr [[I1_ASCAST]], align 4
 153 // CHECK-NEXT:    br label [[FOR_COND2:%.*]]
 154 // CHECK:       for.cond2:
 155 // CHECK-NEXT:    [[TMP13:%.*]] = load i32, ptr [[I1_ASCAST]], align 4
 156 // CHECK-NEXT:    [[TMP14:%.*]] = load i32, ptr [[N_ASCAST]], align 4
 157 // CHECK-NEXT:    [[CMP3:%.*]] = icmp slt i32 [[TMP13]], [[TMP14]]
 158 // CHECK-NEXT:    br i1 [[CMP3]], label [[FOR_BODY4:%.*]], label [[FOR_END9:%.*]]
 159 // CHECK:       for.body4:
 160 // CHECK-NEXT:    [[TMP15:%.*]] = load i32, ptr [[I1_ASCAST]], align 4
 161 // CHECK-NEXT:    [[IDXPROM5:%.*]] = sext i32 [[TMP15]] to i64
 162 // CHECK-NEXT:    [[ARRAYIDX6:%.*]] = getelementptr inbounds i32, ptr [[A]], i64 [[IDXPROM5]]
 163 // CHECK-NEXT:    [[TMP16:%.*]] = load i32, ptr [[ARRAYIDX6]], align 4
 164 // CHECK-NEXT:    [[TMP17:%.*]] = load i32, ptr [[TMP0]], align 4
 165 // CHECK-NEXT:    [[ADD:%.*]] = add nsw i32 [[TMP17]], [[TMP16]]
 166 // CHECK-NEXT:    store i32 [[ADD]], ptr [[TMP0]], align 4
 167 // CHECK-NEXT:    br label [[FOR_INC7:%.*]]
 168 // CHECK:       for.inc7:
 169 // CHECK-NEXT:    [[TMP18:%.*]] = load i32, ptr [[I1_ASCAST]], align 4
 170 // CHECK-NEXT:    [[INC8:%.*]] = add nsw i32 [[TMP18]], 1
 171 // CHECK-NEXT:    store i32 [[INC8]], ptr [[I1_ASCAST]], align 4
 172 // CHECK-NEXT:    br label [[FOR_COND2]], !llvm.loop [[LOOP15:![0-9]+]]
 173 // CHECK:       for.end9:
 174 // CHECK-NEXT:    call void @__kmpc_free_shared(ptr [[A]], i64 [[TMP7]])
 175 // CHECK-NEXT:    call void @__kmpc_target_deinit()
 176 // CHECK-NEXT:    ret void
 177 //
 178 //
 179 // CHECK-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z4foo2v_l30
 180 // CHECK-SAME: (ptr noalias noundef [[DYN_PTR:%.*]], i64 noundef [[M:%.*]], i64 noundef [[VLA:%.*]], ptr noundef nonnull align 4 dereferenceable(4) [[RESULT:%.*]]) #[[ATTR0]] {
 181 // CHECK-NEXT:  entry:
 182 // CHECK-NEXT:    [[DYN_PTR_ADDR:%.*]] = alloca ptr, align 8, addrspace(5)
 183 // CHECK-NEXT:    [[M_ADDR:%.*]] = alloca i64, align 8, addrspace(5)
 184 // CHECK-NEXT:    [[VLA_ADDR:%.*]] = alloca i64, align 8, addrspace(5)
 185 // CHECK-NEXT:    [[RESULT_ADDR:%.*]] = alloca ptr, align 8, addrspace(5)
 186 // CHECK-NEXT:    [[M_CASTED:%.*]] = alloca i64, align 8, addrspace(5)
 187 // CHECK-NEXT:    [[DOTZERO_ADDR:%.*]] = alloca i32, align 4, addrspace(5)
 188 // CHECK-NEXT:    [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4, addrspace(5)
 189 // CHECK-NEXT:    [[DYN_PTR_ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DYN_PTR_ADDR]] to ptr
 190 // CHECK-NEXT:    [[M_ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[M_ADDR]] to ptr
 191 // CHECK-NEXT:    [[VLA_ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[VLA_ADDR]] to ptr
 192 // CHECK-NEXT:    [[RESULT_ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[RESULT_ADDR]] to ptr
 193 // CHECK-NEXT:    [[M_CASTED_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[M_CASTED]] to ptr
 194 // CHECK-NEXT:    [[DOTZERO_ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTZERO_ADDR]] to ptr
 195 // CHECK-NEXT:    [[DOTTHREADID_TEMP__ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTTHREADID_TEMP_]] to ptr
 196 // CHECK-NEXT:    store ptr [[DYN_PTR]], ptr [[DYN_PTR_ADDR_ASCAST]], align 8
 197 // CHECK-NEXT:    store i64 [[M]], ptr [[M_ADDR_ASCAST]], align 8
 198 // CHECK-NEXT:    store i64 [[VLA]], ptr [[VLA_ADDR_ASCAST]], align 8
 199 // CHECK-NEXT:    store ptr [[RESULT]], ptr [[RESULT_ADDR_ASCAST]], align 8
 200 // CHECK-NEXT:    [[TMP0:%.*]] = load i64, ptr [[VLA_ADDR_ASCAST]], align 8
 201 // CHECK-NEXT:    [[TMP1:%.*]] = load ptr, ptr [[RESULT_ADDR_ASCAST]], align 8
 202 // CHECK-NEXT:    [[TMP2:%.*]] = call i32 @__kmpc_target_init(ptr addrspacecast (ptr addrspace(1) @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z4foo2v_l30_kernel_environment to ptr), ptr [[DYN_PTR]])
 203 // CHECK-NEXT:    [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP2]], -1
 204 // CHECK-NEXT:    br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]]
 205 // CHECK:       user_code.entry:
 206 // CHECK-NEXT:    [[TMP3:%.*]] = call i32 @__kmpc_global_thread_num(ptr addrspacecast (ptr addrspace(1) @[[GLOB1:[0-9]+]] to ptr))
 207 // CHECK-NEXT:    [[TMP4:%.*]] = load i32, ptr [[M_ADDR_ASCAST]], align 4
 208 // CHECK-NEXT:    store i32 [[TMP4]], ptr [[M_CASTED_ASCAST]], align 4
 209 // CHECK-NEXT:    [[TMP5:%.*]] = load i64, ptr [[M_CASTED_ASCAST]], align 8
 210 // CHECK-NEXT:    store i32 0, ptr [[DOTZERO_ADDR_ASCAST]], align 4
 211 // CHECK-NEXT:    store i32 [[TMP3]], ptr [[DOTTHREADID_TEMP__ASCAST]], align 4
 212 // CHECK-NEXT:    call void @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z4foo2v_l30_omp_outlined(ptr [[DOTTHREADID_TEMP__ASCAST]], ptr [[DOTZERO_ADDR_ASCAST]], i64 [[TMP5]], i64 [[TMP0]], ptr [[TMP1]]) #[[ATTR4:[0-9]+]]
 213 // CHECK-NEXT:    call void @__kmpc_target_deinit()
 214 // CHECK-NEXT:    ret void
 215 // CHECK:       worker.exit:
 216 // CHECK-NEXT:    ret void
 217 //
 218 //
 219 // CHECK-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z4foo2v_l30_omp_outlined
 220 // CHECK-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]], i64 noundef [[M:%.*]], i64 noundef [[VLA:%.*]], ptr noundef nonnull align 4 dereferenceable(4) [[RESULT:%.*]]) #[[ATTR3:[0-9]+]] {
 221 // CHECK-NEXT:  entry:
 222 // CHECK-NEXT:    [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 8, addrspace(5)
 223 // CHECK-NEXT:    [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 8, addrspace(5)
 224 // CHECK-NEXT:    [[M_ADDR:%.*]] = alloca i64, align 8, addrspace(5)
 225 // CHECK-NEXT:    [[VLA_ADDR:%.*]] = alloca i64, align 8, addrspace(5)
 226 // CHECK-NEXT:    [[RESULT_ADDR:%.*]] = alloca ptr, align 8, addrspace(5)
 227 // CHECK-NEXT:    [[DOTOMP_IV:%.*]] = alloca i32, align 4, addrspace(5)
 228 // CHECK-NEXT:    [[TMP:%.*]] = alloca i32, align 4, addrspace(5)
 229 // CHECK-NEXT:    [[DOTCAPTURE_EXPR_:%.*]] = alloca i32, align 4, addrspace(5)
 230 // CHECK-NEXT:    [[DOTCAPTURE_EXPR_1:%.*]] = alloca i32, align 4, addrspace(5)
 231 // CHECK-NEXT:    [[I:%.*]] = alloca i32, align 4, addrspace(5)
 232 // CHECK-NEXT:    [[DOTOMP_COMB_LB:%.*]] = alloca i32, align 4, addrspace(5)
 233 // CHECK-NEXT:    [[DOTOMP_COMB_UB:%.*]] = alloca i32, align 4, addrspace(5)
 234 // CHECK-NEXT:    [[DOTOMP_STRIDE:%.*]] = alloca i32, align 4, addrspace(5)
 235 // CHECK-NEXT:    [[DOTOMP_IS_LAST:%.*]] = alloca i32, align 4, addrspace(5)
 236 // CHECK-NEXT:    [[I3:%.*]] = alloca i32, align 4, addrspace(5)
 237 // CHECK-NEXT:    [[M_CASTED:%.*]] = alloca i64, align 8, addrspace(5)
 238 // CHECK-NEXT:    [[CAPTURED_VARS_ADDRS:%.*]] = alloca [5 x ptr], align 8, addrspace(5)
 239 // CHECK-NEXT:    [[DOTGLOBAL_TID__ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTGLOBAL_TID__ADDR]] to ptr
 240 // CHECK-NEXT:    [[DOTBOUND_TID__ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTBOUND_TID__ADDR]] to ptr
 241 // CHECK-NEXT:    [[M_ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[M_ADDR]] to ptr
 242 // CHECK-NEXT:    [[VLA_ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[VLA_ADDR]] to ptr
 243 // CHECK-NEXT:    [[RESULT_ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[RESULT_ADDR]] to ptr
 244 // CHECK-NEXT:    [[DOTOMP_IV_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTOMP_IV]] to ptr
 245 // CHECK-NEXT:    [[TMP_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[TMP]] to ptr
 246 // CHECK-NEXT:    [[DOTCAPTURE_EXPR__ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTCAPTURE_EXPR_]] to ptr
 247 // CHECK-NEXT:    [[DOTCAPTURE_EXPR_1_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTCAPTURE_EXPR_1]] to ptr
 248 // CHECK-NEXT:    [[I_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[I]] to ptr
 249 // CHECK-NEXT:    [[DOTOMP_COMB_LB_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTOMP_COMB_LB]] to ptr
 250 // CHECK-NEXT:    [[DOTOMP_COMB_UB_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTOMP_COMB_UB]] to ptr
 251 // CHECK-NEXT:    [[DOTOMP_STRIDE_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTOMP_STRIDE]] to ptr
 252 // CHECK-NEXT:    [[DOTOMP_IS_LAST_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTOMP_IS_LAST]] to ptr
 253 // CHECK-NEXT:    [[I3_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[I3]] to ptr
 254 // CHECK-NEXT:    [[M_CASTED_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[M_CASTED]] to ptr
 255 // CHECK-NEXT:    [[CAPTURED_VARS_ADDRS_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[CAPTURED_VARS_ADDRS]] to ptr
 256 // CHECK-NEXT:    store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR_ASCAST]], align 8
 257 // CHECK-NEXT:    store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR_ASCAST]], align 8
 258 // CHECK-NEXT:    store i64 [[M]], ptr [[M_ADDR_ASCAST]], align 8
 259 // CHECK-NEXT:    store i64 [[VLA]], ptr [[VLA_ADDR_ASCAST]], align 8
 260 // CHECK-NEXT:    store ptr [[RESULT]], ptr [[RESULT_ADDR_ASCAST]], align 8
 261 // CHECK-NEXT:    [[TMP0:%.*]] = load i64, ptr [[VLA_ADDR_ASCAST]], align 8
 262 // CHECK-NEXT:    [[TMP1:%.*]] = load ptr, ptr [[RESULT_ADDR_ASCAST]], align 8
 263 // CHECK-NEXT:    [[TMP2:%.*]] = load i32, ptr [[M_ADDR_ASCAST]], align 4
 264 // CHECK-NEXT:    store i32 [[TMP2]], ptr [[DOTCAPTURE_EXPR__ASCAST]], align 4
 265 // CHECK-NEXT:    [[TMP3:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR__ASCAST]], align 4
 266 // CHECK-NEXT:    [[SUB:%.*]] = sub nsw i32 [[TMP3]], 0
 267 // CHECK-NEXT:    [[DIV:%.*]] = sdiv i32 [[SUB]], 1
 268 // CHECK-NEXT:    [[SUB2:%.*]] = sub nsw i32 [[DIV]], 1
 269 // CHECK-NEXT:    store i32 [[SUB2]], ptr [[DOTCAPTURE_EXPR_1_ASCAST]], align 4
 270 // CHECK-NEXT:    store i32 0, ptr [[I_ASCAST]], align 4
 271 // CHECK-NEXT:    [[TMP4:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR__ASCAST]], align 4
 272 // CHECK-NEXT:    [[CMP:%.*]] = icmp slt i32 0, [[TMP4]]
 273 // CHECK-NEXT:    br i1 [[CMP]], label [[OMP_PRECOND_THEN:%.*]], label [[OMP_PRECOND_END:%.*]]
 274 // CHECK:       omp.precond.then:
 275 // CHECK-NEXT:    store i32 0, ptr [[DOTOMP_COMB_LB_ASCAST]], align 4
 276 // CHECK-NEXT:    [[TMP5:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_1_ASCAST]], align 4
 277 // CHECK-NEXT:    store i32 [[TMP5]], ptr [[DOTOMP_COMB_UB_ASCAST]], align 4
 278 // CHECK-NEXT:    store i32 1, ptr [[DOTOMP_STRIDE_ASCAST]], align 4
 279 // CHECK-NEXT:    store i32 0, ptr [[DOTOMP_IS_LAST_ASCAST]], align 4
 280 // CHECK-NEXT:    [[NVPTX_NUM_THREADS:%.*]] = call i32 @__kmpc_get_hardware_num_threads_in_block()
 281 // CHECK-NEXT:    [[TMP6:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR_ASCAST]], align 8
 282 // CHECK-NEXT:    [[TMP7:%.*]] = load i32, ptr [[TMP6]], align 4
 283 // CHECK-NEXT:    call void @__kmpc_distribute_static_init_4(ptr addrspacecast (ptr addrspace(1) @[[GLOB2:[0-9]+]] to ptr), i32 [[TMP7]], i32 91, ptr [[DOTOMP_IS_LAST_ASCAST]], ptr [[DOTOMP_COMB_LB_ASCAST]], ptr [[DOTOMP_COMB_UB_ASCAST]], ptr [[DOTOMP_STRIDE_ASCAST]], i32 1, i32 [[NVPTX_NUM_THREADS]])
 284 // CHECK-NEXT:    [[TMP8:%.*]] = load i32, ptr [[DOTOMP_COMB_UB_ASCAST]], align 4
 285 // CHECK-NEXT:    [[TMP9:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_1_ASCAST]], align 4
 286 // CHECK-NEXT:    [[CMP4:%.*]] = icmp sgt i32 [[TMP8]], [[TMP9]]
 287 // CHECK-NEXT:    br i1 [[CMP4]], label [[COND_TRUE:%.*]], label [[COND_FALSE:%.*]]
 288 // CHECK:       cond.true:
 289 // CHECK-NEXT:    [[TMP10:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_1_ASCAST]], align 4
 290 // CHECK-NEXT:    br label [[COND_END:%.*]]
 291 // CHECK:       cond.false:
 292 // CHECK-NEXT:    [[TMP11:%.*]] = load i32, ptr [[DOTOMP_COMB_UB_ASCAST]], align 4
 293 // CHECK-NEXT:    br label [[COND_END]]
 294 // CHECK:       cond.end:
 295 // CHECK-NEXT:    [[COND:%.*]] = phi i32 [ [[TMP10]], [[COND_TRUE]] ], [ [[TMP11]], [[COND_FALSE]] ]
 296 // CHECK-NEXT:    store i32 [[COND]], ptr [[DOTOMP_COMB_UB_ASCAST]], align 4
 297 // CHECK-NEXT:    [[TMP12:%.*]] = load i32, ptr [[DOTOMP_COMB_LB_ASCAST]], align 4
 298 // CHECK-NEXT:    store i32 [[TMP12]], ptr [[DOTOMP_IV_ASCAST]], align 4
 299 // CHECK-NEXT:    br label [[OMP_INNER_FOR_COND:%.*]]
 300 // CHECK:       omp.inner.for.cond:
 301 // CHECK-NEXT:    [[TMP13:%.*]] = load i32, ptr [[DOTOMP_IV_ASCAST]], align 4
 302 // CHECK-NEXT:    [[TMP14:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_1_ASCAST]], align 4
 303 // CHECK-NEXT:    [[ADD:%.*]] = add nsw i32 [[TMP14]], 1
 304 // CHECK-NEXT:    [[CMP5:%.*]] = icmp slt i32 [[TMP13]], [[ADD]]
 305 // CHECK-NEXT:    br i1 [[CMP5]], label [[OMP_INNER_FOR_BODY:%.*]], label [[OMP_INNER_FOR_END:%.*]]
 306 // CHECK:       omp.inner.for.body:
 307 // CHECK-NEXT:    [[TMP15:%.*]] = load i32, ptr [[DOTOMP_COMB_LB_ASCAST]], align 4
 308 // CHECK-NEXT:    [[TMP16:%.*]] = zext i32 [[TMP15]] to i64
 309 // CHECK-NEXT:    [[TMP17:%.*]] = load i32, ptr [[DOTOMP_COMB_UB_ASCAST]], align 4
 310 // CHECK-NEXT:    [[TMP18:%.*]] = zext i32 [[TMP17]] to i64
 311 // CHECK-NEXT:    [[TMP19:%.*]] = load i32, ptr [[M_ADDR_ASCAST]], align 4
 312 // CHECK-NEXT:    store i32 [[TMP19]], ptr [[M_CASTED_ASCAST]], align 4
 313 // CHECK-NEXT:    [[TMP20:%.*]] = load i64, ptr [[M_CASTED_ASCAST]], align 8
 314 // CHECK-NEXT:    [[TMP21:%.*]] = getelementptr inbounds [5 x ptr], ptr [[CAPTURED_VARS_ADDRS_ASCAST]], i64 0, i64 0
 315 // CHECK-NEXT:    [[TMP22:%.*]] = inttoptr i64 [[TMP16]] to ptr
 316 // CHECK-NEXT:    store ptr [[TMP22]], ptr [[TMP21]], align 8
 317 // CHECK-NEXT:    [[TMP23:%.*]] = getelementptr inbounds [5 x ptr], ptr [[CAPTURED_VARS_ADDRS_ASCAST]], i64 0, i64 1
 318 // CHECK-NEXT:    [[TMP24:%.*]] = inttoptr i64 [[TMP18]] to ptr
 319 // CHECK-NEXT:    store ptr [[TMP24]], ptr [[TMP23]], align 8
 320 // CHECK-NEXT:    [[TMP25:%.*]] = getelementptr inbounds [5 x ptr], ptr [[CAPTURED_VARS_ADDRS_ASCAST]], i64 0, i64 2
 321 // CHECK-NEXT:    [[TMP26:%.*]] = inttoptr i64 [[TMP20]] to ptr
 322 // CHECK-NEXT:    store ptr [[TMP26]], ptr [[TMP25]], align 8
 323 // CHECK-NEXT:    [[TMP27:%.*]] = getelementptr inbounds [5 x ptr], ptr [[CAPTURED_VARS_ADDRS_ASCAST]], i64 0, i64 3
 324 // CHECK-NEXT:    [[TMP28:%.*]] = inttoptr i64 [[TMP0]] to ptr
 325 // CHECK-NEXT:    store ptr [[TMP28]], ptr [[TMP27]], align 8
 326 // CHECK-NEXT:    [[TMP29:%.*]] = getelementptr inbounds [5 x ptr], ptr [[CAPTURED_VARS_ADDRS_ASCAST]], i64 0, i64 4
 327 // CHECK-NEXT:    store ptr [[TMP1]], ptr [[TMP29]], align 8
 328 // CHECK-NEXT:    [[TMP30:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR_ASCAST]], align 8
 329 // CHECK-NEXT:    [[TMP31:%.*]] = load i32, ptr [[TMP30]], align 4
 330 // CHECK-NEXT:    call void @__kmpc_parallel_51(ptr addrspacecast (ptr addrspace(1) @[[GLOB1]] to ptr), i32 [[TMP31]], i32 1, i32 -1, i32 -1, ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z4foo2v_l30_omp_outlined_omp_outlined, ptr null, ptr [[CAPTURED_VARS_ADDRS_ASCAST]], i64 5)
 331 // CHECK-NEXT:    br label [[OMP_INNER_FOR_INC:%.*]]
 332 // CHECK:       omp.inner.for.inc:
 333 // CHECK-NEXT:    [[TMP32:%.*]] = load i32, ptr [[DOTOMP_IV_ASCAST]], align 4
 334 // CHECK-NEXT:    [[TMP33:%.*]] = load i32, ptr [[DOTOMP_STRIDE_ASCAST]], align 4
 335 // CHECK-NEXT:    [[ADD6:%.*]] = add nsw i32 [[TMP32]], [[TMP33]]
 336 // CHECK-NEXT:    store i32 [[ADD6]], ptr [[DOTOMP_IV_ASCAST]], align 4
 337 // CHECK-NEXT:    [[TMP34:%.*]] = load i32, ptr [[DOTOMP_COMB_LB_ASCAST]], align 4
 338 // CHECK-NEXT:    [[TMP35:%.*]] = load i32, ptr [[DOTOMP_STRIDE_ASCAST]], align 4
 339 // CHECK-NEXT:    [[ADD7:%.*]] = add nsw i32 [[TMP34]], [[TMP35]]
 340 // CHECK-NEXT:    store i32 [[ADD7]], ptr [[DOTOMP_COMB_LB_ASCAST]], align 4
 341 // CHECK-NEXT:    [[TMP36:%.*]] = load i32, ptr [[DOTOMP_COMB_UB_ASCAST]], align 4
 342 // CHECK-NEXT:    [[TMP37:%.*]] = load i32, ptr [[DOTOMP_STRIDE_ASCAST]], align 4
 343 // CHECK-NEXT:    [[ADD8:%.*]] = add nsw i32 [[TMP36]], [[TMP37]]
 344 // CHECK-NEXT:    store i32 [[ADD8]], ptr [[DOTOMP_COMB_UB_ASCAST]], align 4
 345 // CHECK-NEXT:    [[TMP38:%.*]] = load i32, ptr [[DOTOMP_COMB_UB_ASCAST]], align 4
 346 // CHECK-NEXT:    [[TMP39:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_1_ASCAST]], align 4
 347 // CHECK-NEXT:    [[CMP9:%.*]] = icmp sgt i32 [[TMP38]], [[TMP39]]
 348 // CHECK-NEXT:    br i1 [[CMP9]], label [[COND_TRUE10:%.*]], label [[COND_FALSE11:%.*]]
 349 // CHECK:       cond.true10:
 350 // CHECK-NEXT:    [[TMP40:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_1_ASCAST]], align 4
 351 // CHECK-NEXT:    br label [[COND_END12:%.*]]
 352 // CHECK:       cond.false11:
 353 // CHECK-NEXT:    [[TMP41:%.*]] = load i32, ptr [[DOTOMP_COMB_UB_ASCAST]], align 4
 354 // CHECK-NEXT:    br label [[COND_END12]]
 355 // CHECK:       cond.end12:
 356 // CHECK-NEXT:    [[COND13:%.*]] = phi i32 [ [[TMP40]], [[COND_TRUE10]] ], [ [[TMP41]], [[COND_FALSE11]] ]
 357 // CHECK-NEXT:    store i32 [[COND13]], ptr [[DOTOMP_COMB_UB_ASCAST]], align 4
 358 // CHECK-NEXT:    [[TMP42:%.*]] = load i32, ptr [[DOTOMP_COMB_LB_ASCAST]], align 4
 359 // CHECK-NEXT:    store i32 [[TMP42]], ptr [[DOTOMP_IV_ASCAST]], align 4
 360 // CHECK-NEXT:    br label [[OMP_INNER_FOR_COND]]
 361 // CHECK:       omp.inner.for.end:
 362 // CHECK-NEXT:    br label [[OMP_LOOP_EXIT:%.*]]
 363 // CHECK:       omp.loop.exit:
 364 // CHECK-NEXT:    [[TMP43:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR_ASCAST]], align 8
 365 // CHECK-NEXT:    [[TMP44:%.*]] = load i32, ptr [[TMP43]], align 4
 366 // CHECK-NEXT:    call void @__kmpc_distribute_static_fini(ptr addrspacecast (ptr addrspace(1) @[[GLOB2]] to ptr), i32 [[TMP44]])
 367 // CHECK-NEXT:    br label [[OMP_PRECOND_END]]
 368 // CHECK:       omp.precond.end:
 369 // CHECK-NEXT:    ret void
 370 //
 371 //
 372 // CHECK-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z4foo2v_l30_omp_outlined_omp_outlined
 373 // CHECK-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]], i64 noundef [[DOTPREVIOUS_LB_:%.*]], i64 noundef [[DOTPREVIOUS_UB_:%.*]], i64 noundef [[M:%.*]], i64 noundef [[VLA:%.*]], ptr noundef nonnull align 4 dereferenceable(4) [[RESULT:%.*]]) #[[ATTR3]] {
 374 // CHECK-NEXT:  entry:
 375 // CHECK-NEXT:    [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 8, addrspace(5)
 376 // CHECK-NEXT:    [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 8, addrspace(5)
 377 // CHECK-NEXT:    [[DOTPREVIOUS_LB__ADDR:%.*]] = alloca i64, align 8, addrspace(5)
 378 // CHECK-NEXT:    [[DOTPREVIOUS_UB__ADDR:%.*]] = alloca i64, align 8, addrspace(5)
 379 // CHECK-NEXT:    [[M_ADDR:%.*]] = alloca i64, align 8, addrspace(5)
 380 // CHECK-NEXT:    [[VLA_ADDR:%.*]] = alloca i64, align 8, addrspace(5)
 381 // CHECK-NEXT:    [[RESULT_ADDR:%.*]] = alloca ptr, align 8, addrspace(5)
 382 // CHECK-NEXT:    [[DOTOMP_IV:%.*]] = alloca i32, align 4, addrspace(5)
 383 // CHECK-NEXT:    [[TMP:%.*]] = alloca i32, align 4, addrspace(5)
 384 // CHECK-NEXT:    [[DOTCAPTURE_EXPR_:%.*]] = alloca i32, align 4, addrspace(5)
 385 // CHECK-NEXT:    [[DOTCAPTURE_EXPR_1:%.*]] = alloca i32, align 4, addrspace(5)
 386 // CHECK-NEXT:    [[I:%.*]] = alloca i32, align 4, addrspace(5)
 387 // CHECK-NEXT:    [[DOTOMP_LB:%.*]] = alloca i32, align 4, addrspace(5)
 388 // CHECK-NEXT:    [[DOTOMP_UB:%.*]] = alloca i32, align 4, addrspace(5)
 389 // CHECK-NEXT:    [[DOTOMP_STRIDE:%.*]] = alloca i32, align 4, addrspace(5)
 390 // CHECK-NEXT:    [[DOTOMP_IS_LAST:%.*]] = alloca i32, align 4, addrspace(5)
 391 // CHECK-NEXT:    [[I4:%.*]] = alloca i32, align 4, addrspace(5)
 392 // CHECK-NEXT:    [[N:%.*]] = alloca i32, align 4, addrspace(5)
 393 // CHECK-NEXT:    [[SAVED_STACK:%.*]] = alloca ptr addrspace(5), align 4, addrspace(5)
 394 // CHECK-NEXT:    [[__VLA_EXPR0:%.*]] = alloca i64, align 8, addrspace(5)
 395 // CHECK-NEXT:    [[J:%.*]] = alloca i32, align 4, addrspace(5)
 396 // CHECK-NEXT:    [[J11:%.*]] = alloca i32, align 4, addrspace(5)
 397 // CHECK-NEXT:    [[DOTGLOBAL_TID__ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTGLOBAL_TID__ADDR]] to ptr
 398 // CHECK-NEXT:    [[DOTBOUND_TID__ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTBOUND_TID__ADDR]] to ptr
 399 // CHECK-NEXT:    [[DOTPREVIOUS_LB__ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTPREVIOUS_LB__ADDR]] to ptr
 400 // CHECK-NEXT:    [[DOTPREVIOUS_UB__ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTPREVIOUS_UB__ADDR]] to ptr
 401 // CHECK-NEXT:    [[M_ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[M_ADDR]] to ptr
 402 // CHECK-NEXT:    [[VLA_ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[VLA_ADDR]] to ptr
 403 // CHECK-NEXT:    [[RESULT_ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[RESULT_ADDR]] to ptr
 404 // CHECK-NEXT:    [[DOTOMP_IV_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTOMP_IV]] to ptr
 405 // CHECK-NEXT:    [[TMP_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[TMP]] to ptr
 406 // CHECK-NEXT:    [[DOTCAPTURE_EXPR__ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTCAPTURE_EXPR_]] to ptr
 407 // CHECK-NEXT:    [[DOTCAPTURE_EXPR_1_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTCAPTURE_EXPR_1]] to ptr
 408 // CHECK-NEXT:    [[I_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[I]] to ptr
 409 // CHECK-NEXT:    [[DOTOMP_LB_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTOMP_LB]] to ptr
 410 // CHECK-NEXT:    [[DOTOMP_UB_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTOMP_UB]] to ptr
 411 // CHECK-NEXT:    [[DOTOMP_STRIDE_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTOMP_STRIDE]] to ptr
 412 // CHECK-NEXT:    [[DOTOMP_IS_LAST_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTOMP_IS_LAST]] to ptr
 413 // CHECK-NEXT:    [[I4_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[I4]] to ptr
 414 // CHECK-NEXT:    [[N_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[N]] to ptr
 415 // CHECK-NEXT:    [[SAVED_STACK_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[SAVED_STACK]] to ptr
 416 // CHECK-NEXT:    [[__VLA_EXPR0_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[__VLA_EXPR0]] to ptr
 417 // CHECK-NEXT:    [[J_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[J]] to ptr
 418 // CHECK-NEXT:    [[J11_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[J11]] to ptr
 419 // CHECK-NEXT:    store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR_ASCAST]], align 8
 420 // CHECK-NEXT:    store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR_ASCAST]], align 8
 421 // CHECK-NEXT:    store i64 [[DOTPREVIOUS_LB_]], ptr [[DOTPREVIOUS_LB__ADDR_ASCAST]], align 8
 422 // CHECK-NEXT:    store i64 [[DOTPREVIOUS_UB_]], ptr [[DOTPREVIOUS_UB__ADDR_ASCAST]], align 8
 423 // CHECK-NEXT:    store i64 [[M]], ptr [[M_ADDR_ASCAST]], align 8
 424 // CHECK-NEXT:    store i64 [[VLA]], ptr [[VLA_ADDR_ASCAST]], align 8
 425 // CHECK-NEXT:    store ptr [[RESULT]], ptr [[RESULT_ADDR_ASCAST]], align 8
 426 // CHECK-NEXT:    [[TMP0:%.*]] = load i64, ptr [[VLA_ADDR_ASCAST]], align 8
 427 // CHECK-NEXT:    [[TMP1:%.*]] = load ptr, ptr [[RESULT_ADDR_ASCAST]], align 8
 428 // CHECK-NEXT:    [[TMP2:%.*]] = load i32, ptr [[M_ADDR_ASCAST]], align 4
 429 // CHECK-NEXT:    store i32 [[TMP2]], ptr [[DOTCAPTURE_EXPR__ASCAST]], align 4
 430 // CHECK-NEXT:    [[TMP3:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR__ASCAST]], align 4
 431 // CHECK-NEXT:    [[SUB:%.*]] = sub nsw i32 [[TMP3]], 0
 432 // CHECK-NEXT:    [[DIV:%.*]] = sdiv i32 [[SUB]], 1
 433 // CHECK-NEXT:    [[SUB2:%.*]] = sub nsw i32 [[DIV]], 1
 434 // CHECK-NEXT:    store i32 [[SUB2]], ptr [[DOTCAPTURE_EXPR_1_ASCAST]], align 4
 435 // CHECK-NEXT:    store i32 0, ptr [[I_ASCAST]], align 4
 436 // CHECK-NEXT:    [[TMP4:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR__ASCAST]], align 4
 437 // CHECK-NEXT:    [[CMP:%.*]] = icmp slt i32 0, [[TMP4]]
 438 // CHECK-NEXT:    br i1 [[CMP]], label [[OMP_PRECOND_THEN:%.*]], label [[OMP_PRECOND_END:%.*]]
 439 // CHECK:       omp.precond.then:
 440 // CHECK-NEXT:    store i32 0, ptr [[DOTOMP_LB_ASCAST]], align 4
 441 // CHECK-NEXT:    [[TMP5:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_1_ASCAST]], align 4
 442 // CHECK-NEXT:    store i32 [[TMP5]], ptr [[DOTOMP_UB_ASCAST]], align 4
 443 // CHECK-NEXT:    [[TMP6:%.*]] = load i64, ptr [[DOTPREVIOUS_LB__ADDR_ASCAST]], align 8
 444 // CHECK-NEXT:    [[CONV:%.*]] = trunc i64 [[TMP6]] to i32
 445 // CHECK-NEXT:    [[TMP7:%.*]] = load i64, ptr [[DOTPREVIOUS_UB__ADDR_ASCAST]], align 8
 446 // CHECK-NEXT:    [[CONV3:%.*]] = trunc i64 [[TMP7]] to i32
 447 // CHECK-NEXT:    store i32 [[CONV]], ptr [[DOTOMP_LB_ASCAST]], align 4
 448 // CHECK-NEXT:    store i32 [[CONV3]], ptr [[DOTOMP_UB_ASCAST]], align 4
 449 // CHECK-NEXT:    store i32 1, ptr [[DOTOMP_STRIDE_ASCAST]], align 4
 450 // CHECK-NEXT:    store i32 0, ptr [[DOTOMP_IS_LAST_ASCAST]], align 4
 451 // CHECK-NEXT:    [[TMP8:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR_ASCAST]], align 8
 452 // CHECK-NEXT:    [[TMP9:%.*]] = load i32, ptr [[TMP8]], align 4
 453 // CHECK-NEXT:    call void @__kmpc_for_static_init_4(ptr addrspacecast (ptr addrspace(1) @[[GLOB3:[0-9]+]] to ptr), i32 [[TMP9]], i32 33, ptr [[DOTOMP_IS_LAST_ASCAST]], ptr [[DOTOMP_LB_ASCAST]], ptr [[DOTOMP_UB_ASCAST]], ptr [[DOTOMP_STRIDE_ASCAST]], i32 1, i32 1)
 454 // CHECK-NEXT:    [[TMP10:%.*]] = load i32, ptr [[DOTOMP_LB_ASCAST]], align 4
 455 // CHECK-NEXT:    store i32 [[TMP10]], ptr [[DOTOMP_IV_ASCAST]], align 4
 456 // CHECK-NEXT:    br label [[OMP_INNER_FOR_COND:%.*]]
 457 // CHECK:       omp.inner.for.cond:
 458 // CHECK-NEXT:    [[TMP11:%.*]] = load i32, ptr [[DOTOMP_IV_ASCAST]], align 4
 459 // CHECK-NEXT:    [[CONV5:%.*]] = sext i32 [[TMP11]] to i64
 460 // CHECK-NEXT:    [[TMP12:%.*]] = load i64, ptr [[DOTPREVIOUS_UB__ADDR_ASCAST]], align 8
 461 // CHECK-NEXT:    [[CMP6:%.*]] = icmp ule i64 [[CONV5]], [[TMP12]]
 462 // CHECK-NEXT:    br i1 [[CMP6]], label [[OMP_INNER_FOR_BODY:%.*]], label [[OMP_INNER_FOR_END:%.*]]
 463 // CHECK:       omp.inner.for.body:
 464 // CHECK-NEXT:    [[TMP13:%.*]] = load i32, ptr [[DOTOMP_IV_ASCAST]], align 4
 465 // CHECK-NEXT:    [[MUL:%.*]] = mul nsw i32 [[TMP13]], 1
 466 // CHECK-NEXT:    [[ADD:%.*]] = add nsw i32 0, [[MUL]]
 467 // CHECK-NEXT:    store i32 [[ADD]], ptr [[I4_ASCAST]], align 4
 468 // CHECK-NEXT:    store i32 10, ptr [[N_ASCAST]], align 4
 469 // CHECK-NEXT:    [[TMP14:%.*]] = load i32, ptr [[N_ASCAST]], align 4
 470 // CHECK-NEXT:    [[TMP15:%.*]] = zext i32 [[TMP14]] to i64
 471 // CHECK-NEXT:    [[TMP16:%.*]] = call ptr addrspace(5) @llvm.stacksave.p5()
 472 // CHECK-NEXT:    store ptr addrspace(5) [[TMP16]], ptr [[SAVED_STACK_ASCAST]], align 4
 473 // CHECK-NEXT:    [[VLA7:%.*]] = alloca i32, i64 [[TMP15]], align 4, addrspace(5)
 474 // CHECK-NEXT:    [[VLA7_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[VLA7]] to ptr
 475 // CHECK-NEXT:    store i64 [[TMP15]], ptr [[__VLA_EXPR0_ASCAST]], align 8
 476 // CHECK-NEXT:    [[TMP17:%.*]] = load i32, ptr [[I4_ASCAST]], align 4
 477 // CHECK-NEXT:    [[TMP18:%.*]] = load i32, ptr [[I4_ASCAST]], align 4
 478 // CHECK-NEXT:    [[IDXPROM:%.*]] = sext i32 [[TMP18]] to i64
 479 // CHECK-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds i32, ptr [[TMP1]], i64 [[IDXPROM]]
 480 // CHECK-NEXT:    store i32 [[TMP17]], ptr [[ARRAYIDX]], align 4
 481 // CHECK-NEXT:    store i32 0, ptr [[J_ASCAST]], align 4
 482 // CHECK-NEXT:    br label [[FOR_COND:%.*]]
 483 // CHECK:       for.cond:
 484 // CHECK-NEXT:    [[TMP19:%.*]] = load i32, ptr [[J_ASCAST]], align 4
 485 // CHECK-NEXT:    [[TMP20:%.*]] = load i32, ptr [[N_ASCAST]], align 4
 486 // CHECK-NEXT:    [[CMP8:%.*]] = icmp slt i32 [[TMP19]], [[TMP20]]
 487 // CHECK-NEXT:    br i1 [[CMP8]], label [[FOR_BODY:%.*]], label [[FOR_END:%.*]]
 488 // CHECK:       for.body:
 489 // CHECK-NEXT:    [[TMP21:%.*]] = load i32, ptr [[J_ASCAST]], align 4
 490 // CHECK-NEXT:    [[TMP22:%.*]] = load i32, ptr [[J_ASCAST]], align 4
 491 // CHECK-NEXT:    [[IDXPROM9:%.*]] = sext i32 [[TMP22]] to i64
 492 // CHECK-NEXT:    [[ARRAYIDX10:%.*]] = getelementptr inbounds i32, ptr [[VLA7_ASCAST]], i64 [[IDXPROM9]]
 493 // CHECK-NEXT:    store i32 [[TMP21]], ptr [[ARRAYIDX10]], align 4
 494 // CHECK-NEXT:    br label [[FOR_INC:%.*]]
 495 // CHECK:       for.inc:
 496 // CHECK-NEXT:    [[TMP23:%.*]] = load i32, ptr [[J_ASCAST]], align 4
 497 // CHECK-NEXT:    [[INC:%.*]] = add nsw i32 [[TMP23]], 1
 498 // CHECK-NEXT:    store i32 [[INC]], ptr [[J_ASCAST]], align 4
 499 // CHECK-NEXT:    br label [[FOR_COND]], !llvm.loop [[LOOP16:![0-9]+]]
 500 // CHECK:       for.end:
 501 // CHECK-NEXT:    store i32 0, ptr [[J11_ASCAST]], align 4
 502 // CHECK-NEXT:    br label [[FOR_COND12:%.*]]
 503 // CHECK:       for.cond12:
 504 // CHECK-NEXT:    [[TMP24:%.*]] = load i32, ptr [[J11_ASCAST]], align 4
 505 // CHECK-NEXT:    [[TMP25:%.*]] = load i32, ptr [[N_ASCAST]], align 4
 506 // CHECK-NEXT:    [[CMP13:%.*]] = icmp slt i32 [[TMP24]], [[TMP25]]
 507 // CHECK-NEXT:    br i1 [[CMP13]], label [[FOR_BODY14:%.*]], label [[FOR_END22:%.*]]
 508 // CHECK:       for.body14:
 509 // CHECK-NEXT:    [[TMP26:%.*]] = load i32, ptr [[J11_ASCAST]], align 4
 510 // CHECK-NEXT:    [[IDXPROM15:%.*]] = sext i32 [[TMP26]] to i64
 511 // CHECK-NEXT:    [[ARRAYIDX16:%.*]] = getelementptr inbounds i32, ptr [[VLA7_ASCAST]], i64 [[IDXPROM15]]
 512 // CHECK-NEXT:    [[TMP27:%.*]] = load i32, ptr [[ARRAYIDX16]], align 4
 513 // CHECK-NEXT:    [[TMP28:%.*]] = load i32, ptr [[I4_ASCAST]], align 4
 514 // CHECK-NEXT:    [[IDXPROM17:%.*]] = sext i32 [[TMP28]] to i64
 515 // CHECK-NEXT:    [[ARRAYIDX18:%.*]] = getelementptr inbounds i32, ptr [[TMP1]], i64 [[IDXPROM17]]
 516 // CHECK-NEXT:    [[TMP29:%.*]] = load i32, ptr [[ARRAYIDX18]], align 4
 517 // CHECK-NEXT:    [[ADD19:%.*]] = add nsw i32 [[TMP29]], [[TMP27]]
 518 // CHECK-NEXT:    store i32 [[ADD19]], ptr [[ARRAYIDX18]], align 4
 519 // CHECK-NEXT:    br label [[FOR_INC20:%.*]]
 520 // CHECK:       for.inc20:
 521 // CHECK-NEXT:    [[TMP30:%.*]] = load i32, ptr [[J11_ASCAST]], align 4
 522 // CHECK-NEXT:    [[INC21:%.*]] = add nsw i32 [[TMP30]], 1
 523 // CHECK-NEXT:    store i32 [[INC21]], ptr [[J11_ASCAST]], align 4
 524 // CHECK-NEXT:    br label [[FOR_COND12]], !llvm.loop [[LOOP17:![0-9]+]]
 525 // CHECK:       for.end22:
 526 // CHECK-NEXT:    [[TMP31:%.*]] = load ptr addrspace(5), ptr [[SAVED_STACK_ASCAST]], align 4
 527 // CHECK-NEXT:    call void @llvm.stackrestore.p5(ptr addrspace(5) [[TMP31]])
 528 // CHECK-NEXT:    br label [[OMP_BODY_CONTINUE:%.*]]
 529 // CHECK:       omp.body.continue:
 530 // CHECK-NEXT:    br label [[OMP_INNER_FOR_INC:%.*]]
 531 // CHECK:       omp.inner.for.inc:
 532 // CHECK-NEXT:    [[TMP32:%.*]] = load i32, ptr [[DOTOMP_IV_ASCAST]], align 4
 533 // CHECK-NEXT:    [[TMP33:%.*]] = load i32, ptr [[DOTOMP_STRIDE_ASCAST]], align 4
 534 // CHECK-NEXT:    [[ADD23:%.*]] = add nsw i32 [[TMP32]], [[TMP33]]
 535 // CHECK-NEXT:    store i32 [[ADD23]], ptr [[DOTOMP_IV_ASCAST]], align 4
 536 // CHECK-NEXT:    br label [[OMP_INNER_FOR_COND]]
 537 // CHECK:       omp.inner.for.end:
 538 // CHECK-NEXT:    br label [[OMP_LOOP_EXIT:%.*]]
 539 // CHECK:       omp.loop.exit:
 540 // CHECK-NEXT:    [[TMP34:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR_ASCAST]], align 8
 541 // CHECK-NEXT:    [[TMP35:%.*]] = load i32, ptr [[TMP34]], align 4
 542 // CHECK-NEXT:    call void @__kmpc_for_static_fini(ptr addrspacecast (ptr addrspace(1) @[[GLOB3]] to ptr), i32 [[TMP35]])
 543 // CHECK-NEXT:    br label [[OMP_PRECOND_END]]
 544 // CHECK:       omp.precond.end:
 545 // CHECK-NEXT:    ret void
 546 //
 547 //
 548 // CHECK-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z4foo3v_l52
 549 // CHECK-SAME: (ptr noalias noundef [[DYN_PTR:%.*]], i64 noundef [[M:%.*]], i64 noundef [[VLA:%.*]], ptr noundef nonnull align 4 dereferenceable(4) [[RESULT:%.*]]) #[[ATTR0]] {
 550 // CHECK-NEXT:  entry:
 551 // CHECK-NEXT:    [[DYN_PTR_ADDR:%.*]] = alloca ptr, align 8, addrspace(5)
 552 // CHECK-NEXT:    [[M_ADDR:%.*]] = alloca i64, align 8, addrspace(5)
 553 // CHECK-NEXT:    [[VLA_ADDR:%.*]] = alloca i64, align 8, addrspace(5)
 554 // CHECK-NEXT:    [[RESULT_ADDR:%.*]] = alloca ptr, align 8, addrspace(5)
 555 // CHECK-NEXT:    [[M_CASTED:%.*]] = alloca i64, align 8, addrspace(5)
 556 // CHECK-NEXT:    [[DOTZERO_ADDR:%.*]] = alloca i32, align 4, addrspace(5)
 557 // CHECK-NEXT:    [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4, addrspace(5)
 558 // CHECK-NEXT:    [[DYN_PTR_ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DYN_PTR_ADDR]] to ptr
 559 // CHECK-NEXT:    [[M_ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[M_ADDR]] to ptr
 560 // CHECK-NEXT:    [[VLA_ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[VLA_ADDR]] to ptr
 561 // CHECK-NEXT:    [[RESULT_ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[RESULT_ADDR]] to ptr
 562 // CHECK-NEXT:    [[M_CASTED_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[M_CASTED]] to ptr
 563 // CHECK-NEXT:    [[DOTZERO_ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTZERO_ADDR]] to ptr
 564 // CHECK-NEXT:    [[DOTTHREADID_TEMP__ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTTHREADID_TEMP_]] to ptr
 565 // CHECK-NEXT:    store ptr [[DYN_PTR]], ptr [[DYN_PTR_ADDR_ASCAST]], align 8
 566 // CHECK-NEXT:    store i64 [[M]], ptr [[M_ADDR_ASCAST]], align 8
 567 // CHECK-NEXT:    store i64 [[VLA]], ptr [[VLA_ADDR_ASCAST]], align 8
 568 // CHECK-NEXT:    store ptr [[RESULT]], ptr [[RESULT_ADDR_ASCAST]], align 8
 569 // CHECK-NEXT:    [[TMP0:%.*]] = load i64, ptr [[VLA_ADDR_ASCAST]], align 8
 570 // CHECK-NEXT:    [[TMP1:%.*]] = load ptr, ptr [[RESULT_ADDR_ASCAST]], align 8
 571 // CHECK-NEXT:    [[TMP2:%.*]] = call i32 @__kmpc_target_init(ptr addrspacecast (ptr addrspace(1) @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z4foo3v_l52_kernel_environment to ptr), ptr [[DYN_PTR]])
 572 // CHECK-NEXT:    [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP2]], -1
 573 // CHECK-NEXT:    br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]]
 574 // CHECK:       user_code.entry:
 575 // CHECK-NEXT:    [[TMP3:%.*]] = call i32 @__kmpc_global_thread_num(ptr addrspacecast (ptr addrspace(1) @[[GLOB1]] to ptr))
 576 // CHECK-NEXT:    [[TMP4:%.*]] = load i32, ptr [[M_ADDR_ASCAST]], align 4
 577 // CHECK-NEXT:    store i32 [[TMP4]], ptr [[M_CASTED_ASCAST]], align 4
 578 // CHECK-NEXT:    [[TMP5:%.*]] = load i64, ptr [[M_CASTED_ASCAST]], align 8
 579 // CHECK-NEXT:    store i32 0, ptr [[DOTZERO_ADDR_ASCAST]], align 4
 580 // CHECK-NEXT:    store i32 [[TMP3]], ptr [[DOTTHREADID_TEMP__ASCAST]], align 4
 581 // CHECK-NEXT:    call void @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z4foo3v_l52_omp_outlined(ptr [[DOTTHREADID_TEMP__ASCAST]], ptr [[DOTZERO_ADDR_ASCAST]], i64 [[TMP5]], i64 [[TMP0]], ptr [[TMP1]]) #[[ATTR4]]
 582 // CHECK-NEXT:    call void @__kmpc_target_deinit()
 583 // CHECK-NEXT:    ret void
 584 // CHECK:       worker.exit:
 585 // CHECK-NEXT:    ret void
 586 //
 587 //
 588 // CHECK-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z4foo3v_l52_omp_outlined
 589 // CHECK-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]], i64 noundef [[M:%.*]], i64 noundef [[VLA:%.*]], ptr noundef nonnull align 4 dereferenceable(4) [[RESULT:%.*]]) #[[ATTR3]] {
 590 // CHECK-NEXT:  entry:
 591 // CHECK-NEXT:    [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 8, addrspace(5)
 592 // CHECK-NEXT:    [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 8, addrspace(5)
 593 // CHECK-NEXT:    [[M_ADDR:%.*]] = alloca i64, align 8, addrspace(5)
 594 // CHECK-NEXT:    [[VLA_ADDR:%.*]] = alloca i64, align 8, addrspace(5)
 595 // CHECK-NEXT:    [[RESULT_ADDR:%.*]] = alloca ptr, align 8, addrspace(5)
 596 // CHECK-NEXT:    [[DOTOMP_IV:%.*]] = alloca i32, align 4, addrspace(5)
 597 // CHECK-NEXT:    [[TMP:%.*]] = alloca i32, align 4, addrspace(5)
 598 // CHECK-NEXT:    [[DOTCAPTURE_EXPR_:%.*]] = alloca i32, align 4, addrspace(5)
 599 // CHECK-NEXT:    [[DOTCAPTURE_EXPR_1:%.*]] = alloca i32, align 4, addrspace(5)
 600 // CHECK-NEXT:    [[I:%.*]] = alloca i32, align 4, addrspace(5)
 601 // CHECK-NEXT:    [[DOTOMP_LB:%.*]] = alloca i32, align 4, addrspace(5)
 602 // CHECK-NEXT:    [[DOTOMP_UB:%.*]] = alloca i32, align 4, addrspace(5)
 603 // CHECK-NEXT:    [[DOTOMP_STRIDE:%.*]] = alloca i32, align 4, addrspace(5)
 604 // CHECK-NEXT:    [[DOTOMP_IS_LAST:%.*]] = alloca i32, align 4, addrspace(5)
 605 // CHECK-NEXT:    [[I3:%.*]] = alloca i32, align 4, addrspace(5)
 606 // CHECK-NEXT:    [[__VLA_EXPR0:%.*]] = alloca i64, align 8, addrspace(5)
 607 // CHECK-NEXT:    [[CAPTURED_VARS_ADDRS:%.*]] = alloca [3 x ptr], align 8, addrspace(5)
 608 // CHECK-NEXT:    [[J:%.*]] = alloca i32, align 4, addrspace(5)
 609 // CHECK-NEXT:    [[DOTGLOBAL_TID__ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTGLOBAL_TID__ADDR]] to ptr
 610 // CHECK-NEXT:    [[DOTBOUND_TID__ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTBOUND_TID__ADDR]] to ptr
 611 // CHECK-NEXT:    [[M_ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[M_ADDR]] to ptr
 612 // CHECK-NEXT:    [[VLA_ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[VLA_ADDR]] to ptr
 613 // CHECK-NEXT:    [[RESULT_ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[RESULT_ADDR]] to ptr
 614 // CHECK-NEXT:    [[DOTOMP_IV_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTOMP_IV]] to ptr
 615 // CHECK-NEXT:    [[TMP_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[TMP]] to ptr
 616 // CHECK-NEXT:    [[DOTCAPTURE_EXPR__ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTCAPTURE_EXPR_]] to ptr
 617 // CHECK-NEXT:    [[DOTCAPTURE_EXPR_1_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTCAPTURE_EXPR_1]] to ptr
 618 // CHECK-NEXT:    [[I_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[I]] to ptr
 619 // CHECK-NEXT:    [[DOTOMP_LB_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTOMP_LB]] to ptr
 620 // CHECK-NEXT:    [[DOTOMP_UB_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTOMP_UB]] to ptr
 621 // CHECK-NEXT:    [[DOTOMP_STRIDE_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTOMP_STRIDE]] to ptr
 622 // CHECK-NEXT:    [[DOTOMP_IS_LAST_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTOMP_IS_LAST]] to ptr
 623 // CHECK-NEXT:    [[I3_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[I3]] to ptr
 624 // CHECK-NEXT:    [[__VLA_EXPR0_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[__VLA_EXPR0]] to ptr
 625 // CHECK-NEXT:    [[CAPTURED_VARS_ADDRS_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[CAPTURED_VARS_ADDRS]] to ptr
 626 // CHECK-NEXT:    [[J_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[J]] to ptr
 627 // CHECK-NEXT:    store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR_ASCAST]], align 8
 628 // CHECK-NEXT:    store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR_ASCAST]], align 8
 629 // CHECK-NEXT:    store i64 [[M]], ptr [[M_ADDR_ASCAST]], align 8
 630 // CHECK-NEXT:    store i64 [[VLA]], ptr [[VLA_ADDR_ASCAST]], align 8
 631 // CHECK-NEXT:    store ptr [[RESULT]], ptr [[RESULT_ADDR_ASCAST]], align 8
 632 // CHECK-NEXT:    [[TMP0:%.*]] = load i64, ptr [[VLA_ADDR_ASCAST]], align 8
 633 // CHECK-NEXT:    [[TMP1:%.*]] = load ptr, ptr [[RESULT_ADDR_ASCAST]], align 8
 634 // CHECK-NEXT:    [[N:%.*]] = call align 8 ptr @__kmpc_alloc_shared(i64 4)
 635 // CHECK-NEXT:    [[TMP2:%.*]] = load i32, ptr [[M_ADDR_ASCAST]], align 4
 636 // CHECK-NEXT:    store i32 [[TMP2]], ptr [[DOTCAPTURE_EXPR__ASCAST]], align 4
 637 // CHECK-NEXT:    [[TMP3:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR__ASCAST]], align 4
 638 // CHECK-NEXT:    [[SUB:%.*]] = sub nsw i32 [[TMP3]], 0
 639 // CHECK-NEXT:    [[DIV:%.*]] = sdiv i32 [[SUB]], 1
 640 // CHECK-NEXT:    [[SUB2:%.*]] = sub nsw i32 [[DIV]], 1
 641 // CHECK-NEXT:    store i32 [[SUB2]], ptr [[DOTCAPTURE_EXPR_1_ASCAST]], align 4
 642 // CHECK-NEXT:    store i32 0, ptr [[I_ASCAST]], align 4
 643 // CHECK-NEXT:    [[TMP4:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR__ASCAST]], align 4
 644 // CHECK-NEXT:    [[CMP:%.*]] = icmp slt i32 0, [[TMP4]]
 645 // CHECK-NEXT:    br i1 [[CMP]], label [[OMP_PRECOND_THEN:%.*]], label [[OMP_PRECOND_END:%.*]]
 646 // CHECK:       omp.precond.then:
 647 // CHECK-NEXT:    store i32 0, ptr [[DOTOMP_LB_ASCAST]], align 4
 648 // CHECK-NEXT:    [[TMP5:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_1_ASCAST]], align 4
 649 // CHECK-NEXT:    store i32 [[TMP5]], ptr [[DOTOMP_UB_ASCAST]], align 4
 650 // CHECK-NEXT:    store i32 1, ptr [[DOTOMP_STRIDE_ASCAST]], align 4
 651 // CHECK-NEXT:    store i32 0, ptr [[DOTOMP_IS_LAST_ASCAST]], align 4
 652 // CHECK-NEXT:    [[TMP6:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR_ASCAST]], align 8
 653 // CHECK-NEXT:    [[TMP7:%.*]] = load i32, ptr [[TMP6]], align 4
 654 // CHECK-NEXT:    call void @__kmpc_distribute_static_init_4(ptr addrspacecast (ptr addrspace(1) @[[GLOB2]] to ptr), i32 [[TMP7]], i32 92, ptr [[DOTOMP_IS_LAST_ASCAST]], ptr [[DOTOMP_LB_ASCAST]], ptr [[DOTOMP_UB_ASCAST]], ptr [[DOTOMP_STRIDE_ASCAST]], i32 1, i32 1)
 655 // CHECK-NEXT:    [[TMP8:%.*]] = load i32, ptr [[DOTOMP_UB_ASCAST]], align 4
 656 // CHECK-NEXT:    [[TMP9:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_1_ASCAST]], align 4
 657 // CHECK-NEXT:    [[CMP4:%.*]] = icmp sgt i32 [[TMP8]], [[TMP9]]
 658 // CHECK-NEXT:    br i1 [[CMP4]], label [[COND_TRUE:%.*]], label [[COND_FALSE:%.*]]
 659 // CHECK:       cond.true:
 660 // CHECK-NEXT:    [[TMP10:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_1_ASCAST]], align 4
 661 // CHECK-NEXT:    br label [[COND_END:%.*]]
 662 // CHECK:       cond.false:
 663 // CHECK-NEXT:    [[TMP11:%.*]] = load i32, ptr [[DOTOMP_UB_ASCAST]], align 4
 664 // CHECK-NEXT:    br label [[COND_END]]
 665 // CHECK:       cond.end:
 666 // CHECK-NEXT:    [[COND:%.*]] = phi i32 [ [[TMP10]], [[COND_TRUE]] ], [ [[TMP11]], [[COND_FALSE]] ]
 667 // CHECK-NEXT:    store i32 [[COND]], ptr [[DOTOMP_UB_ASCAST]], align 4
 668 // CHECK-NEXT:    [[TMP12:%.*]] = load i32, ptr [[DOTOMP_LB_ASCAST]], align 4
 669 // CHECK-NEXT:    store i32 [[TMP12]], ptr [[DOTOMP_IV_ASCAST]], align 4
 670 // CHECK-NEXT:    br label [[OMP_INNER_FOR_COND:%.*]]
 671 // CHECK:       omp.inner.for.cond:
 672 // CHECK-NEXT:    [[TMP13:%.*]] = load i32, ptr [[DOTOMP_IV_ASCAST]], align 4
 673 // CHECK-NEXT:    [[TMP14:%.*]] = load i32, ptr [[DOTOMP_UB_ASCAST]], align 4
 674 // CHECK-NEXT:    [[CMP5:%.*]] = icmp sle i32 [[TMP13]], [[TMP14]]
 675 // CHECK-NEXT:    br i1 [[CMP5]], label [[OMP_INNER_FOR_BODY:%.*]], label [[OMP_INNER_FOR_END:%.*]]
 676 // CHECK:       omp.inner.for.body:
 677 // CHECK-NEXT:    [[TMP15:%.*]] = load i32, ptr [[DOTOMP_IV_ASCAST]], align 4
 678 // CHECK-NEXT:    [[MUL:%.*]] = mul nsw i32 [[TMP15]], 1
 679 // CHECK-NEXT:    [[ADD:%.*]] = add nsw i32 0, [[MUL]]
 680 // CHECK-NEXT:    store i32 [[ADD]], ptr [[I3_ASCAST]], align 4
 681 // CHECK-NEXT:    store i32 10, ptr [[N]], align 4
 682 // CHECK-NEXT:    [[TMP16:%.*]] = load i32, ptr [[N]], align 4
 683 // CHECK-NEXT:    [[TMP17:%.*]] = zext i32 [[TMP16]] to i64
 684 // CHECK-NEXT:    [[TMP18:%.*]] = mul nuw i64 [[TMP17]], 4
 685 // CHECK-NEXT:    [[TMP19:%.*]] = add nuw i64 [[TMP18]], 3
 686 // CHECK-NEXT:    [[TMP20:%.*]] = udiv i64 [[TMP19]], 4
 687 // CHECK-NEXT:    [[TMP21:%.*]] = mul nuw i64 [[TMP20]], 4
 688 // CHECK-NEXT:    [[A:%.*]] = call align 4 ptr @__kmpc_alloc_shared(i64 [[TMP21]])
 689 // CHECK-NEXT:    store i64 [[TMP17]], ptr [[__VLA_EXPR0_ASCAST]], align 8
 690 // CHECK-NEXT:    [[TMP22:%.*]] = load i32, ptr [[I3_ASCAST]], align 4
 691 // CHECK-NEXT:    [[TMP23:%.*]] = load i32, ptr [[I3_ASCAST]], align 4
 692 // CHECK-NEXT:    [[IDXPROM:%.*]] = sext i32 [[TMP23]] to i64
 693 // CHECK-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds i32, ptr [[TMP1]], i64 [[IDXPROM]]
 694 // CHECK-NEXT:    store i32 [[TMP22]], ptr [[ARRAYIDX]], align 4
 695 // CHECK-NEXT:    [[TMP24:%.*]] = getelementptr inbounds [3 x ptr], ptr [[CAPTURED_VARS_ADDRS_ASCAST]], i64 0, i64 0
 696 // CHECK-NEXT:    store ptr [[N]], ptr [[TMP24]], align 8
 697 // CHECK-NEXT:    [[TMP25:%.*]] = getelementptr inbounds [3 x ptr], ptr [[CAPTURED_VARS_ADDRS_ASCAST]], i64 0, i64 1
 698 // CHECK-NEXT:    [[TMP26:%.*]] = inttoptr i64 [[TMP17]] to ptr
 699 // CHECK-NEXT:    store ptr [[TMP26]], ptr [[TMP25]], align 8
 700 // CHECK-NEXT:    [[TMP27:%.*]] = getelementptr inbounds [3 x ptr], ptr [[CAPTURED_VARS_ADDRS_ASCAST]], i64 0, i64 2
 701 // CHECK-NEXT:    store ptr [[A]], ptr [[TMP27]], align 8
 702 // CHECK-NEXT:    [[TMP28:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR_ASCAST]], align 8
 703 // CHECK-NEXT:    [[TMP29:%.*]] = load i32, ptr [[TMP28]], align 4
 704 // CHECK-NEXT:    call void @__kmpc_parallel_51(ptr addrspacecast (ptr addrspace(1) @[[GLOB1]] to ptr), i32 [[TMP29]], i32 1, i32 -1, i32 -1, ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z4foo3v_l52_omp_outlined_omp_outlined, ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z4foo3v_l52_omp_outlined_omp_outlined_wrapper, ptr [[CAPTURED_VARS_ADDRS_ASCAST]], i64 3)
 705 // CHECK-NEXT:    store i32 0, ptr [[J_ASCAST]], align 4
 706 // CHECK-NEXT:    br label [[FOR_COND:%.*]]
 707 // CHECK:       for.cond:
 708 // CHECK-NEXT:    [[TMP30:%.*]] = load i32, ptr [[J_ASCAST]], align 4
 709 // CHECK-NEXT:    [[TMP31:%.*]] = load i32, ptr [[N]], align 4
 710 // CHECK-NEXT:    [[CMP6:%.*]] = icmp slt i32 [[TMP30]], [[TMP31]]
 711 // CHECK-NEXT:    br i1 [[CMP6]], label [[FOR_BODY:%.*]], label [[FOR_END:%.*]]
 712 // CHECK:       for.body:
 713 // CHECK-NEXT:    [[TMP32:%.*]] = load i32, ptr [[J_ASCAST]], align 4
 714 // CHECK-NEXT:    [[IDXPROM7:%.*]] = sext i32 [[TMP32]] to i64
 715 // CHECK-NEXT:    [[ARRAYIDX8:%.*]] = getelementptr inbounds i32, ptr [[A]], i64 [[IDXPROM7]]
 716 // CHECK-NEXT:    [[TMP33:%.*]] = load i32, ptr [[ARRAYIDX8]], align 4
 717 // CHECK-NEXT:    [[TMP34:%.*]] = load i32, ptr [[I3_ASCAST]], align 4
 718 // CHECK-NEXT:    [[IDXPROM9:%.*]] = sext i32 [[TMP34]] to i64
 719 // CHECK-NEXT:    [[ARRAYIDX10:%.*]] = getelementptr inbounds i32, ptr [[TMP1]], i64 [[IDXPROM9]]
 720 // CHECK-NEXT:    [[TMP35:%.*]] = load i32, ptr [[ARRAYIDX10]], align 4
 721 // CHECK-NEXT:    [[ADD11:%.*]] = add nsw i32 [[TMP35]], [[TMP33]]
 722 // CHECK-NEXT:    store i32 [[ADD11]], ptr [[ARRAYIDX10]], align 4
 723 // CHECK-NEXT:    br label [[FOR_INC:%.*]]
 724 // CHECK:       for.inc:
 725 // CHECK-NEXT:    [[TMP36:%.*]] = load i32, ptr [[J_ASCAST]], align 4
 726 // CHECK-NEXT:    [[INC:%.*]] = add nsw i32 [[TMP36]], 1
 727 // CHECK-NEXT:    store i32 [[INC]], ptr [[J_ASCAST]], align 4
 728 // CHECK-NEXT:    br label [[FOR_COND]], !llvm.loop [[LOOP18:![0-9]+]]
 729 // CHECK:       for.end:
 730 // CHECK-NEXT:    call void @__kmpc_free_shared(ptr [[A]], i64 [[TMP21]])
 731 // CHECK-NEXT:    br label [[OMP_BODY_CONTINUE:%.*]]
 732 // CHECK:       omp.body.continue:
 733 // CHECK-NEXT:    br label [[OMP_INNER_FOR_INC:%.*]]
 734 // CHECK:       omp.inner.for.inc:
 735 // CHECK-NEXT:    [[TMP37:%.*]] = load i32, ptr [[DOTOMP_IV_ASCAST]], align 4
 736 // CHECK-NEXT:    [[ADD12:%.*]] = add nsw i32 [[TMP37]], 1
 737 // CHECK-NEXT:    store i32 [[ADD12]], ptr [[DOTOMP_IV_ASCAST]], align 4
 738 // CHECK-NEXT:    br label [[OMP_INNER_FOR_COND]]
 739 // CHECK:       omp.inner.for.end:
 740 // CHECK-NEXT:    br label [[OMP_LOOP_EXIT:%.*]]
 741 // CHECK:       omp.loop.exit:
 742 // CHECK-NEXT:    [[TMP38:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR_ASCAST]], align 8
 743 // CHECK-NEXT:    [[TMP39:%.*]] = load i32, ptr [[TMP38]], align 4
 744 // CHECK-NEXT:    call void @__kmpc_distribute_static_fini(ptr addrspacecast (ptr addrspace(1) @[[GLOB2]] to ptr), i32 [[TMP39]])
 745 // CHECK-NEXT:    br label [[OMP_PRECOND_END]]
 746 // CHECK:       omp.precond.end:
 747 // CHECK-NEXT:    call void @__kmpc_free_shared(ptr [[N]], i64 4)
 748 // CHECK-NEXT:    ret void
 749 //
 750 //
 751 // CHECK-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z4foo3v_l52_omp_outlined_omp_outlined
 752 // CHECK-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]], ptr noundef nonnull align 4 dereferenceable(4) [[N:%.*]], i64 noundef [[VLA:%.*]], ptr noundef nonnull align 4 dereferenceable(4) [[A:%.*]]) #[[ATTR3]] {
 753 // CHECK-NEXT:  entry:
 754 // CHECK-NEXT:    [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 8, addrspace(5)
 755 // CHECK-NEXT:    [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 8, addrspace(5)
 756 // CHECK-NEXT:    [[N_ADDR:%.*]] = alloca ptr, align 8, addrspace(5)
 757 // CHECK-NEXT:    [[VLA_ADDR:%.*]] = alloca i64, align 8, addrspace(5)
 758 // CHECK-NEXT:    [[A_ADDR:%.*]] = alloca ptr, align 8, addrspace(5)
 759 // CHECK-NEXT:    [[DOTOMP_IV:%.*]] = alloca i32, align 4, addrspace(5)
 760 // CHECK-NEXT:    [[TMP:%.*]] = alloca i32, align 4, addrspace(5)
 761 // CHECK-NEXT:    [[DOTCAPTURE_EXPR_:%.*]] = alloca i32, align 4, addrspace(5)
 762 // CHECK-NEXT:    [[DOTCAPTURE_EXPR_1:%.*]] = alloca i32, align 4, addrspace(5)
 763 // CHECK-NEXT:    [[J:%.*]] = alloca i32, align 4, addrspace(5)
 764 // CHECK-NEXT:    [[DOTOMP_LB:%.*]] = alloca i32, align 4, addrspace(5)
 765 // CHECK-NEXT:    [[DOTOMP_UB:%.*]] = alloca i32, align 4, addrspace(5)
 766 // CHECK-NEXT:    [[DOTOMP_STRIDE:%.*]] = alloca i32, align 4, addrspace(5)
 767 // CHECK-NEXT:    [[DOTOMP_IS_LAST:%.*]] = alloca i32, align 4, addrspace(5)
 768 // CHECK-NEXT:    [[J3:%.*]] = alloca i32, align 4, addrspace(5)
 769 // CHECK-NEXT:    [[DOTGLOBAL_TID__ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTGLOBAL_TID__ADDR]] to ptr
 770 // CHECK-NEXT:    [[DOTBOUND_TID__ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTBOUND_TID__ADDR]] to ptr
 771 // CHECK-NEXT:    [[N_ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[N_ADDR]] to ptr
 772 // CHECK-NEXT:    [[VLA_ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[VLA_ADDR]] to ptr
 773 // CHECK-NEXT:    [[A_ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[A_ADDR]] to ptr
 774 // CHECK-NEXT:    [[DOTOMP_IV_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTOMP_IV]] to ptr
 775 // CHECK-NEXT:    [[TMP_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[TMP]] to ptr
 776 // CHECK-NEXT:    [[DOTCAPTURE_EXPR__ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTCAPTURE_EXPR_]] to ptr
 777 // CHECK-NEXT:    [[DOTCAPTURE_EXPR_1_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTCAPTURE_EXPR_1]] to ptr
 778 // CHECK-NEXT:    [[J_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[J]] to ptr
 779 // CHECK-NEXT:    [[DOTOMP_LB_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTOMP_LB]] to ptr
 780 // CHECK-NEXT:    [[DOTOMP_UB_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTOMP_UB]] to ptr
 781 // CHECK-NEXT:    [[DOTOMP_STRIDE_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTOMP_STRIDE]] to ptr
 782 // CHECK-NEXT:    [[DOTOMP_IS_LAST_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTOMP_IS_LAST]] to ptr
 783 // CHECK-NEXT:    [[J3_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[J3]] to ptr
 784 // CHECK-NEXT:    store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR_ASCAST]], align 8
 785 // CHECK-NEXT:    store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR_ASCAST]], align 8
 786 // CHECK-NEXT:    store ptr [[N]], ptr [[N_ADDR_ASCAST]], align 8
 787 // CHECK-NEXT:    store i64 [[VLA]], ptr [[VLA_ADDR_ASCAST]], align 8
 788 // CHECK-NEXT:    store ptr [[A]], ptr [[A_ADDR_ASCAST]], align 8
 789 // CHECK-NEXT:    [[TMP0:%.*]] = load ptr, ptr [[N_ADDR_ASCAST]], align 8
 790 // CHECK-NEXT:    [[TMP1:%.*]] = load i64, ptr [[VLA_ADDR_ASCAST]], align 8
 791 // CHECK-NEXT:    [[TMP2:%.*]] = load ptr, ptr [[A_ADDR_ASCAST]], align 8
 792 // CHECK-NEXT:    [[TMP3:%.*]] = load i32, ptr [[TMP0]], align 4
 793 // CHECK-NEXT:    store i32 [[TMP3]], ptr [[DOTCAPTURE_EXPR__ASCAST]], align 4
 794 // CHECK-NEXT:    [[TMP4:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR__ASCAST]], align 4
 795 // CHECK-NEXT:    [[SUB:%.*]] = sub nsw i32 [[TMP4]], 0
 796 // CHECK-NEXT:    [[DIV:%.*]] = sdiv i32 [[SUB]], 1
 797 // CHECK-NEXT:    [[SUB2:%.*]] = sub nsw i32 [[DIV]], 1
 798 // CHECK-NEXT:    store i32 [[SUB2]], ptr [[DOTCAPTURE_EXPR_1_ASCAST]], align 4
 799 // CHECK-NEXT:    store i32 0, ptr [[J_ASCAST]], align 4
 800 // CHECK-NEXT:    [[TMP5:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR__ASCAST]], align 4
 801 // CHECK-NEXT:    [[CMP:%.*]] = icmp slt i32 0, [[TMP5]]
 802 // CHECK-NEXT:    br i1 [[CMP]], label [[OMP_PRECOND_THEN:%.*]], label [[OMP_PRECOND_END:%.*]]
 803 // CHECK:       omp.precond.then:
 804 // CHECK-NEXT:    store i32 0, ptr [[DOTOMP_LB_ASCAST]], align 4
 805 // CHECK-NEXT:    [[TMP6:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_1_ASCAST]], align 4
 806 // CHECK-NEXT:    store i32 [[TMP6]], ptr [[DOTOMP_UB_ASCAST]], align 4
 807 // CHECK-NEXT:    store i32 1, ptr [[DOTOMP_STRIDE_ASCAST]], align 4
 808 // CHECK-NEXT:    store i32 0, ptr [[DOTOMP_IS_LAST_ASCAST]], align 4
 809 // CHECK-NEXT:    [[TMP7:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR_ASCAST]], align 8
 810 // CHECK-NEXT:    [[TMP8:%.*]] = load i32, ptr [[TMP7]], align 4
 811 // CHECK-NEXT:    call void @__kmpc_for_static_init_4(ptr addrspacecast (ptr addrspace(1) @[[GLOB3]] to ptr), i32 [[TMP8]], i32 33, ptr [[DOTOMP_IS_LAST_ASCAST]], ptr [[DOTOMP_LB_ASCAST]], ptr [[DOTOMP_UB_ASCAST]], ptr [[DOTOMP_STRIDE_ASCAST]], i32 1, i32 1)
 812 // CHECK-NEXT:    br label [[OMP_DISPATCH_COND:%.*]]
 813 // CHECK:       omp.dispatch.cond:
 814 // CHECK-NEXT:    [[TMP9:%.*]] = load i32, ptr [[DOTOMP_UB_ASCAST]], align 4
 815 // CHECK-NEXT:    [[TMP10:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_1_ASCAST]], align 4
 816 // CHECK-NEXT:    [[CMP4:%.*]] = icmp sgt i32 [[TMP9]], [[TMP10]]
 817 // CHECK-NEXT:    br i1 [[CMP4]], label [[COND_TRUE:%.*]], label [[COND_FALSE:%.*]]
 818 // CHECK:       cond.true:
 819 // CHECK-NEXT:    [[TMP11:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_1_ASCAST]], align 4
 820 // CHECK-NEXT:    br label [[COND_END:%.*]]
 821 // CHECK:       cond.false:
 822 // CHECK-NEXT:    [[TMP12:%.*]] = load i32, ptr [[DOTOMP_UB_ASCAST]], align 4
 823 // CHECK-NEXT:    br label [[COND_END]]
 824 // CHECK:       cond.end:
 825 // CHECK-NEXT:    [[COND:%.*]] = phi i32 [ [[TMP11]], [[COND_TRUE]] ], [ [[TMP12]], [[COND_FALSE]] ]
 826 // CHECK-NEXT:    store i32 [[COND]], ptr [[DOTOMP_UB_ASCAST]], align 4
 827 // CHECK-NEXT:    [[TMP13:%.*]] = load i32, ptr [[DOTOMP_LB_ASCAST]], align 4
 828 // CHECK-NEXT:    store i32 [[TMP13]], ptr [[DOTOMP_IV_ASCAST]], align 4
 829 // CHECK-NEXT:    [[TMP14:%.*]] = load i32, ptr [[DOTOMP_IV_ASCAST]], align 4
 830 // CHECK-NEXT:    [[TMP15:%.*]] = load i32, ptr [[DOTOMP_UB_ASCAST]], align 4
 831 // CHECK-NEXT:    [[CMP5:%.*]] = icmp sle i32 [[TMP14]], [[TMP15]]
 832 // CHECK-NEXT:    br i1 [[CMP5]], label [[OMP_DISPATCH_BODY:%.*]], label [[OMP_DISPATCH_END:%.*]]
 833 // CHECK:       omp.dispatch.body:
 834 // CHECK-NEXT:    br label [[OMP_INNER_FOR_COND:%.*]]
 835 // CHECK:       omp.inner.for.cond:
 836 // CHECK-NEXT:    [[TMP16:%.*]] = load i32, ptr [[DOTOMP_IV_ASCAST]], align 4
 837 // CHECK-NEXT:    [[TMP17:%.*]] = load i32, ptr [[DOTOMP_UB_ASCAST]], align 4
 838 // CHECK-NEXT:    [[CMP6:%.*]] = icmp sle i32 [[TMP16]], [[TMP17]]
 839 // CHECK-NEXT:    br i1 [[CMP6]], label [[OMP_INNER_FOR_BODY:%.*]], label [[OMP_INNER_FOR_END:%.*]]
 840 // CHECK:       omp.inner.for.body:
 841 // CHECK-NEXT:    [[TMP18:%.*]] = load i32, ptr [[DOTOMP_IV_ASCAST]], align 4
 842 // CHECK-NEXT:    [[MUL:%.*]] = mul nsw i32 [[TMP18]], 1
 843 // CHECK-NEXT:    [[ADD:%.*]] = add nsw i32 0, [[MUL]]
 844 // CHECK-NEXT:    store i32 [[ADD]], ptr [[J3_ASCAST]], align 4
 845 // CHECK-NEXT:    [[TMP19:%.*]] = load i32, ptr [[J3_ASCAST]], align 4
 846 // CHECK-NEXT:    [[TMP20:%.*]] = load i32, ptr [[J3_ASCAST]], align 4
 847 // CHECK-NEXT:    [[IDXPROM:%.*]] = sext i32 [[TMP20]] to i64
 848 // CHECK-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds i32, ptr [[TMP2]], i64 [[IDXPROM]]
 849 // CHECK-NEXT:    store i32 [[TMP19]], ptr [[ARRAYIDX]], align 4
 850 // CHECK-NEXT:    br label [[OMP_BODY_CONTINUE:%.*]]
 851 // CHECK:       omp.body.continue:
 852 // CHECK-NEXT:    br label [[OMP_INNER_FOR_INC:%.*]]
 853 // CHECK:       omp.inner.for.inc:
 854 // CHECK-NEXT:    [[TMP21:%.*]] = load i32, ptr [[DOTOMP_IV_ASCAST]], align 4
 855 // CHECK-NEXT:    [[ADD7:%.*]] = add nsw i32 [[TMP21]], 1
 856 // CHECK-NEXT:    store i32 [[ADD7]], ptr [[DOTOMP_IV_ASCAST]], align 4
 857 // CHECK-NEXT:    br label [[OMP_INNER_FOR_COND]]
 858 // CHECK:       omp.inner.for.end:
 859 // CHECK-NEXT:    br label [[OMP_DISPATCH_INC:%.*]]
 860 // CHECK:       omp.dispatch.inc:
 861 // CHECK-NEXT:    [[TMP22:%.*]] = load i32, ptr [[DOTOMP_LB_ASCAST]], align 4
 862 // CHECK-NEXT:    [[TMP23:%.*]] = load i32, ptr [[DOTOMP_STRIDE_ASCAST]], align 4
 863 // CHECK-NEXT:    [[ADD8:%.*]] = add nsw i32 [[TMP22]], [[TMP23]]
 864 // CHECK-NEXT:    store i32 [[ADD8]], ptr [[DOTOMP_LB_ASCAST]], align 4
 865 // CHECK-NEXT:    [[TMP24:%.*]] = load i32, ptr [[DOTOMP_UB_ASCAST]], align 4
 866 // CHECK-NEXT:    [[TMP25:%.*]] = load i32, ptr [[DOTOMP_STRIDE_ASCAST]], align 4
 867 // CHECK-NEXT:    [[ADD9:%.*]] = add nsw i32 [[TMP24]], [[TMP25]]
 868 // CHECK-NEXT:    store i32 [[ADD9]], ptr [[DOTOMP_UB_ASCAST]], align 4
 869 // CHECK-NEXT:    br label [[OMP_DISPATCH_COND]]
 870 // CHECK:       omp.dispatch.end:
 871 // CHECK-NEXT:    [[TMP26:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR_ASCAST]], align 8
 872 // CHECK-NEXT:    [[TMP27:%.*]] = load i32, ptr [[TMP26]], align 4
 873 // CHECK-NEXT:    call void @__kmpc_for_static_fini(ptr addrspacecast (ptr addrspace(1) @[[GLOB3]] to ptr), i32 [[TMP27]])
 874 // CHECK-NEXT:    br label [[OMP_PRECOND_END]]
 875 // CHECK:       omp.precond.end:
 876 // CHECK-NEXT:    ret void
 877 //
 878 //
 879 // CHECK-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z4foo3v_l52_omp_outlined_omp_outlined_wrapper
 880 // CHECK-SAME: (i16 noundef zeroext [[TMP0:%.*]], i32 noundef [[TMP1:%.*]]) #[[ATTR7:[0-9]+]] {
 881 // CHECK-NEXT:  entry:
 882 // CHECK-NEXT:    [[DOTADDR:%.*]] = alloca i16, align 2, addrspace(5)
 883 // CHECK-NEXT:    [[DOTADDR1:%.*]] = alloca i32, align 4, addrspace(5)
 884 // CHECK-NEXT:    [[DOTZERO_ADDR:%.*]] = alloca i32, align 4, addrspace(5)
 885 // CHECK-NEXT:    [[GLOBAL_ARGS:%.*]] = alloca ptr, align 8, addrspace(5)
 886 // CHECK-NEXT:    [[DOTADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTADDR]] to ptr
 887 // CHECK-NEXT:    [[DOTADDR1_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTADDR1]] to ptr
 888 // CHECK-NEXT:    [[DOTZERO_ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTZERO_ADDR]] to ptr
 889 // CHECK-NEXT:    [[GLOBAL_ARGS_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[GLOBAL_ARGS]] to ptr
 890 // CHECK-NEXT:    store i16 [[TMP0]], ptr [[DOTADDR_ASCAST]], align 2
 891 // CHECK-NEXT:    store i32 [[TMP1]], ptr [[DOTADDR1_ASCAST]], align 4
 892 // CHECK-NEXT:    store i32 0, ptr [[DOTZERO_ADDR_ASCAST]], align 4
 893 // CHECK-NEXT:    call void @__kmpc_get_shared_variables(ptr [[GLOBAL_ARGS_ASCAST]])
 894 // CHECK-NEXT:    [[TMP2:%.*]] = load ptr, ptr [[GLOBAL_ARGS_ASCAST]], align 8
 895 // CHECK-NEXT:    [[TMP3:%.*]] = getelementptr inbounds ptr, ptr [[TMP2]], i64 0
 896 // CHECK-NEXT:    [[TMP4:%.*]] = load ptr, ptr [[TMP3]], align 8
 897 // CHECK-NEXT:    [[TMP5:%.*]] = getelementptr inbounds ptr, ptr [[TMP2]], i64 1
 898 // CHECK-NEXT:    [[TMP6:%.*]] = load i64, ptr [[TMP5]], align 8
 899 // CHECK-NEXT:    [[TMP7:%.*]] = getelementptr inbounds ptr, ptr [[TMP2]], i64 2
 900 // CHECK-NEXT:    [[TMP8:%.*]] = load ptr, ptr [[TMP7]], align 8
 901 // CHECK-NEXT:    call void @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z4foo3v_l52_omp_outlined_omp_outlined(ptr [[DOTADDR1_ASCAST]], ptr [[DOTZERO_ADDR_ASCAST]], ptr [[TMP4]], i64 [[TMP6]], ptr [[TMP8]]) #[[ATTR4]]
 902 // CHECK-NEXT:    ret void
 903 //
 904 //
 905 // CHECK-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z4foo4v_l76
 906 // CHECK-SAME: (ptr noalias noundef [[DYN_PTR:%.*]], i64 noundef [[M:%.*]], i64 noundef [[N:%.*]], i64 noundef [[VLA:%.*]], ptr noundef nonnull align 4 dereferenceable(4) [[RESULT:%.*]]) #[[ATTR0]] {
 907 // CHECK-NEXT:  entry:
 908 // CHECK-NEXT:    [[DYN_PTR_ADDR:%.*]] = alloca ptr, align 8, addrspace(5)
 909 // CHECK-NEXT:    [[M_ADDR:%.*]] = alloca i64, align 8, addrspace(5)
 910 // CHECK-NEXT:    [[N_ADDR:%.*]] = alloca i64, align 8, addrspace(5)
 911 // CHECK-NEXT:    [[VLA_ADDR:%.*]] = alloca i64, align 8, addrspace(5)
 912 // CHECK-NEXT:    [[RESULT_ADDR:%.*]] = alloca ptr, align 8, addrspace(5)
 913 // CHECK-NEXT:    [[M_CASTED:%.*]] = alloca i64, align 8, addrspace(5)
 914 // CHECK-NEXT:    [[N_CASTED:%.*]] = alloca i64, align 8, addrspace(5)
 915 // CHECK-NEXT:    [[DOTZERO_ADDR:%.*]] = alloca i32, align 4, addrspace(5)
 916 // CHECK-NEXT:    [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4, addrspace(5)
 917 // CHECK-NEXT:    [[DYN_PTR_ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DYN_PTR_ADDR]] to ptr
 918 // CHECK-NEXT:    [[M_ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[M_ADDR]] to ptr
 919 // CHECK-NEXT:    [[N_ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[N_ADDR]] to ptr
 920 // CHECK-NEXT:    [[VLA_ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[VLA_ADDR]] to ptr
 921 // CHECK-NEXT:    [[RESULT_ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[RESULT_ADDR]] to ptr
 922 // CHECK-NEXT:    [[M_CASTED_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[M_CASTED]] to ptr
 923 // CHECK-NEXT:    [[N_CASTED_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[N_CASTED]] to ptr
 924 // CHECK-NEXT:    [[DOTZERO_ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTZERO_ADDR]] to ptr
 925 // CHECK-NEXT:    [[DOTTHREADID_TEMP__ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTTHREADID_TEMP_]] to ptr
 926 // CHECK-NEXT:    store ptr [[DYN_PTR]], ptr [[DYN_PTR_ADDR_ASCAST]], align 8
 927 // CHECK-NEXT:    store i64 [[M]], ptr [[M_ADDR_ASCAST]], align 8
 928 // CHECK-NEXT:    store i64 [[N]], ptr [[N_ADDR_ASCAST]], align 8
 929 // CHECK-NEXT:    store i64 [[VLA]], ptr [[VLA_ADDR_ASCAST]], align 8
 930 // CHECK-NEXT:    store ptr [[RESULT]], ptr [[RESULT_ADDR_ASCAST]], align 8
 931 // CHECK-NEXT:    [[TMP0:%.*]] = load i64, ptr [[VLA_ADDR_ASCAST]], align 8
 932 // CHECK-NEXT:    [[TMP1:%.*]] = load ptr, ptr [[RESULT_ADDR_ASCAST]], align 8
 933 // CHECK-NEXT:    [[TMP2:%.*]] = call i32 @__kmpc_target_init(ptr addrspacecast (ptr addrspace(1) @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z4foo4v_l76_kernel_environment to ptr), ptr [[DYN_PTR]])
 934 // CHECK-NEXT:    [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP2]], -1
 935 // CHECK-NEXT:    br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[WORKER_EXIT:%.*]]
 936 // CHECK:       user_code.entry:
 937 // CHECK-NEXT:    [[TMP3:%.*]] = call i32 @__kmpc_global_thread_num(ptr addrspacecast (ptr addrspace(1) @[[GLOB1]] to ptr))
 938 // CHECK-NEXT:    [[TMP4:%.*]] = load i32, ptr [[M_ADDR_ASCAST]], align 4
 939 // CHECK-NEXT:    store i32 [[TMP4]], ptr [[M_CASTED_ASCAST]], align 4
 940 // CHECK-NEXT:    [[TMP5:%.*]] = load i64, ptr [[M_CASTED_ASCAST]], align 8
 941 // CHECK-NEXT:    [[TMP6:%.*]] = load i32, ptr [[N_ADDR_ASCAST]], align 4
 942 // CHECK-NEXT:    store i32 [[TMP6]], ptr [[N_CASTED_ASCAST]], align 4
 943 // CHECK-NEXT:    [[TMP7:%.*]] = load i64, ptr [[N_CASTED_ASCAST]], align 8
 944 // CHECK-NEXT:    store i32 0, ptr [[DOTZERO_ADDR_ASCAST]], align 4
 945 // CHECK-NEXT:    store i32 [[TMP3]], ptr [[DOTTHREADID_TEMP__ASCAST]], align 4
 946 // CHECK-NEXT:    call void @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z4foo4v_l76_omp_outlined(ptr [[DOTTHREADID_TEMP__ASCAST]], ptr [[DOTZERO_ADDR_ASCAST]], i64 [[TMP5]], i64 [[TMP7]], i64 [[TMP0]], ptr [[TMP1]]) #[[ATTR4]]
 947 // CHECK-NEXT:    call void @__kmpc_target_deinit()
 948 // CHECK-NEXT:    ret void
 949 // CHECK:       worker.exit:
 950 // CHECK-NEXT:    ret void
 951 //
 952 //
 953 // CHECK-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z4foo4v_l76_omp_outlined
 954 // CHECK-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]], i64 noundef [[M:%.*]], i64 noundef [[N:%.*]], i64 noundef [[VLA:%.*]], ptr noundef nonnull align 4 dereferenceable(4) [[RESULT:%.*]]) #[[ATTR3]] {
 955 // CHECK-NEXT:  entry:
 956 // CHECK-NEXT:    [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 8, addrspace(5)
 957 // CHECK-NEXT:    [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 8, addrspace(5)
 958 // CHECK-NEXT:    [[M_ADDR:%.*]] = alloca i64, align 8, addrspace(5)
 959 // CHECK-NEXT:    [[N_ADDR:%.*]] = alloca i64, align 8, addrspace(5)
 960 // CHECK-NEXT:    [[VLA_ADDR:%.*]] = alloca i64, align 8, addrspace(5)
 961 // CHECK-NEXT:    [[RESULT_ADDR:%.*]] = alloca ptr, align 8, addrspace(5)
 962 // CHECK-NEXT:    [[DOTOMP_IV:%.*]] = alloca i32, align 4, addrspace(5)
 963 // CHECK-NEXT:    [[TMP:%.*]] = alloca i32, align 4, addrspace(5)
 964 // CHECK-NEXT:    [[DOTCAPTURE_EXPR_:%.*]] = alloca i32, align 4, addrspace(5)
 965 // CHECK-NEXT:    [[DOTCAPTURE_EXPR_2:%.*]] = alloca i32, align 4, addrspace(5)
 966 // CHECK-NEXT:    [[I:%.*]] = alloca i32, align 4, addrspace(5)
 967 // CHECK-NEXT:    [[DOTOMP_LB:%.*]] = alloca i32, align 4, addrspace(5)
 968 // CHECK-NEXT:    [[DOTOMP_UB:%.*]] = alloca i32, align 4, addrspace(5)
 969 // CHECK-NEXT:    [[DOTOMP_STRIDE:%.*]] = alloca i32, align 4, addrspace(5)
 970 // CHECK-NEXT:    [[DOTOMP_IS_LAST:%.*]] = alloca i32, align 4, addrspace(5)
 971 // CHECK-NEXT:    [[I4:%.*]] = alloca i32, align 4, addrspace(5)
 972 // CHECK-NEXT:    [[__VLA_EXPR0:%.*]] = alloca i64, align 8, addrspace(5)
 973 // CHECK-NEXT:    [[CAPTURED_VARS_ADDRS:%.*]] = alloca [3 x ptr], align 8, addrspace(5)
 974 // CHECK-NEXT:    [[J:%.*]] = alloca i32, align 4, addrspace(5)
 975 // CHECK-NEXT:    [[DOTGLOBAL_TID__ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTGLOBAL_TID__ADDR]] to ptr
 976 // CHECK-NEXT:    [[DOTBOUND_TID__ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTBOUND_TID__ADDR]] to ptr
 977 // CHECK-NEXT:    [[M_ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[M_ADDR]] to ptr
 978 // CHECK-NEXT:    [[N_ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[N_ADDR]] to ptr
 979 // CHECK-NEXT:    [[VLA_ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[VLA_ADDR]] to ptr
 980 // CHECK-NEXT:    [[RESULT_ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[RESULT_ADDR]] to ptr
 981 // CHECK-NEXT:    [[DOTOMP_IV_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTOMP_IV]] to ptr
 982 // CHECK-NEXT:    [[TMP_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[TMP]] to ptr
 983 // CHECK-NEXT:    [[DOTCAPTURE_EXPR__ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTCAPTURE_EXPR_]] to ptr
 984 // CHECK-NEXT:    [[DOTCAPTURE_EXPR_2_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTCAPTURE_EXPR_2]] to ptr
 985 // CHECK-NEXT:    [[I_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[I]] to ptr
 986 // CHECK-NEXT:    [[DOTOMP_LB_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTOMP_LB]] to ptr
 987 // CHECK-NEXT:    [[DOTOMP_UB_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTOMP_UB]] to ptr
 988 // CHECK-NEXT:    [[DOTOMP_STRIDE_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTOMP_STRIDE]] to ptr
 989 // CHECK-NEXT:    [[DOTOMP_IS_LAST_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTOMP_IS_LAST]] to ptr
 990 // CHECK-NEXT:    [[I4_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[I4]] to ptr
 991 // CHECK-NEXT:    [[__VLA_EXPR0_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[__VLA_EXPR0]] to ptr
 992 // CHECK-NEXT:    [[CAPTURED_VARS_ADDRS_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[CAPTURED_VARS_ADDRS]] to ptr
 993 // CHECK-NEXT:    [[J_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[J]] to ptr
 994 // CHECK-NEXT:    store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR_ASCAST]], align 8
 995 // CHECK-NEXT:    store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR_ASCAST]], align 8
 996 // CHECK-NEXT:    store i64 [[M]], ptr [[M_ADDR_ASCAST]], align 8
 997 // CHECK-NEXT:    store i64 [[N]], ptr [[N_ADDR_ASCAST]], align 8
 998 // CHECK-NEXT:    store i64 [[VLA]], ptr [[VLA_ADDR_ASCAST]], align 8
 999 // CHECK-NEXT:    store ptr [[RESULT]], ptr [[RESULT_ADDR_ASCAST]], align 8
1000 // CHECK-NEXT:    [[TMP0:%.*]] = load i64, ptr [[VLA_ADDR_ASCAST]], align 8
1001 // CHECK-NEXT:    [[TMP1:%.*]] = load ptr, ptr [[RESULT_ADDR_ASCAST]], align 8
1002 // CHECK-NEXT:    [[TMP2:%.*]] = load i32, ptr [[N_ADDR_ASCAST]], align 4
1003 // CHECK-NEXT:    [[N1:%.*]] = call align 8 ptr @__kmpc_alloc_shared(i64 4)
1004 // CHECK-NEXT:    store i32 [[TMP2]], ptr [[N1]], align 4
1005 // CHECK-NEXT:    [[TMP3:%.*]] = load i32, ptr [[M_ADDR_ASCAST]], align 4
1006 // CHECK-NEXT:    store i32 [[TMP3]], ptr [[DOTCAPTURE_EXPR__ASCAST]], align 4
1007 // CHECK-NEXT:    [[TMP4:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR__ASCAST]], align 4
1008 // CHECK-NEXT:    [[SUB:%.*]] = sub nsw i32 [[TMP4]], 0
1009 // CHECK-NEXT:    [[DIV:%.*]] = sdiv i32 [[SUB]], 1
1010 // CHECK-NEXT:    [[SUB3:%.*]] = sub nsw i32 [[DIV]], 1
1011 // CHECK-NEXT:    store i32 [[SUB3]], ptr [[DOTCAPTURE_EXPR_2_ASCAST]], align 4
1012 // CHECK-NEXT:    store i32 0, ptr [[I_ASCAST]], align 4
1013 // CHECK-NEXT:    [[TMP5:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR__ASCAST]], align 4
1014 // CHECK-NEXT:    [[CMP:%.*]] = icmp slt i32 0, [[TMP5]]
1015 // CHECK-NEXT:    br i1 [[CMP]], label [[OMP_PRECOND_THEN:%.*]], label [[OMP_PRECOND_END:%.*]]
1016 // CHECK:       omp.precond.then:
1017 // CHECK-NEXT:    store i32 0, ptr [[DOTOMP_LB_ASCAST]], align 4
1018 // CHECK-NEXT:    [[TMP6:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_2_ASCAST]], align 4
1019 // CHECK-NEXT:    store i32 [[TMP6]], ptr [[DOTOMP_UB_ASCAST]], align 4
1020 // CHECK-NEXT:    store i32 1, ptr [[DOTOMP_STRIDE_ASCAST]], align 4
1021 // CHECK-NEXT:    store i32 0, ptr [[DOTOMP_IS_LAST_ASCAST]], align 4
1022 // CHECK-NEXT:    [[TMP7:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR_ASCAST]], align 8
1023 // CHECK-NEXT:    [[TMP8:%.*]] = load i32, ptr [[TMP7]], align 4
1024 // CHECK-NEXT:    call void @__kmpc_distribute_static_init_4(ptr addrspacecast (ptr addrspace(1) @[[GLOB2]] to ptr), i32 [[TMP8]], i32 92, ptr [[DOTOMP_IS_LAST_ASCAST]], ptr [[DOTOMP_LB_ASCAST]], ptr [[DOTOMP_UB_ASCAST]], ptr [[DOTOMP_STRIDE_ASCAST]], i32 1, i32 1)
1025 // CHECK-NEXT:    [[TMP9:%.*]] = load i32, ptr [[DOTOMP_UB_ASCAST]], align 4
1026 // CHECK-NEXT:    [[TMP10:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_2_ASCAST]], align 4
1027 // CHECK-NEXT:    [[CMP5:%.*]] = icmp sgt i32 [[TMP9]], [[TMP10]]
1028 // CHECK-NEXT:    br i1 [[CMP5]], label [[COND_TRUE:%.*]], label [[COND_FALSE:%.*]]
1029 // CHECK:       cond.true:
1030 // CHECK-NEXT:    [[TMP11:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_2_ASCAST]], align 4
1031 // CHECK-NEXT:    br label [[COND_END:%.*]]
1032 // CHECK:       cond.false:
1033 // CHECK-NEXT:    [[TMP12:%.*]] = load i32, ptr [[DOTOMP_UB_ASCAST]], align 4
1034 // CHECK-NEXT:    br label [[COND_END]]
1035 // CHECK:       cond.end:
1036 // CHECK-NEXT:    [[COND:%.*]] = phi i32 [ [[TMP11]], [[COND_TRUE]] ], [ [[TMP12]], [[COND_FALSE]] ]
1037 // CHECK-NEXT:    store i32 [[COND]], ptr [[DOTOMP_UB_ASCAST]], align 4
1038 // CHECK-NEXT:    [[TMP13:%.*]] = load i32, ptr [[DOTOMP_LB_ASCAST]], align 4
1039 // CHECK-NEXT:    store i32 [[TMP13]], ptr [[DOTOMP_IV_ASCAST]], align 4
1040 // CHECK-NEXT:    br label [[OMP_INNER_FOR_COND:%.*]]
1041 // CHECK:       omp.inner.for.cond:
1042 // CHECK-NEXT:    [[TMP14:%.*]] = load i32, ptr [[DOTOMP_IV_ASCAST]], align 4
1043 // CHECK-NEXT:    [[TMP15:%.*]] = load i32, ptr [[DOTOMP_UB_ASCAST]], align 4
1044 // CHECK-NEXT:    [[CMP6:%.*]] = icmp sle i32 [[TMP14]], [[TMP15]]
1045 // CHECK-NEXT:    br i1 [[CMP6]], label [[OMP_INNER_FOR_BODY:%.*]], label [[OMP_INNER_FOR_END:%.*]]
1046 // CHECK:       omp.inner.for.body:
1047 // CHECK-NEXT:    [[TMP16:%.*]] = load i32, ptr [[DOTOMP_IV_ASCAST]], align 4
1048 // CHECK-NEXT:    [[MUL:%.*]] = mul nsw i32 [[TMP16]], 1
1049 // CHECK-NEXT:    [[ADD:%.*]] = add nsw i32 0, [[MUL]]
1050 // CHECK-NEXT:    store i32 [[ADD]], ptr [[I4_ASCAST]], align 4
1051 // CHECK-NEXT:    [[TMP17:%.*]] = load i32, ptr [[N1]], align 4
1052 // CHECK-NEXT:    [[TMP18:%.*]] = zext i32 [[TMP17]] to i64
1053 // CHECK-NEXT:    [[TMP19:%.*]] = mul nuw i64 [[TMP18]], 4
1054 // CHECK-NEXT:    [[TMP20:%.*]] = add nuw i64 [[TMP19]], 3
1055 // CHECK-NEXT:    [[TMP21:%.*]] = udiv i64 [[TMP20]], 4
1056 // CHECK-NEXT:    [[TMP22:%.*]] = mul nuw i64 [[TMP21]], 4
1057 // CHECK-NEXT:    [[A:%.*]] = call align 4 ptr @__kmpc_alloc_shared(i64 [[TMP22]])
1058 // CHECK-NEXT:    store i64 [[TMP18]], ptr [[__VLA_EXPR0_ASCAST]], align 8
1059 // CHECK-NEXT:    [[TMP23:%.*]] = load i32, ptr [[I4_ASCAST]], align 4
1060 // CHECK-NEXT:    [[TMP24:%.*]] = load i32, ptr [[I4_ASCAST]], align 4
1061 // CHECK-NEXT:    [[IDXPROM:%.*]] = sext i32 [[TMP24]] to i64
1062 // CHECK-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds i32, ptr [[TMP1]], i64 [[IDXPROM]]
1063 // CHECK-NEXT:    store i32 [[TMP23]], ptr [[ARRAYIDX]], align 4
1064 // CHECK-NEXT:    [[TMP25:%.*]] = getelementptr inbounds [3 x ptr], ptr [[CAPTURED_VARS_ADDRS_ASCAST]], i64 0, i64 0
1065 // CHECK-NEXT:    store ptr [[N1]], ptr [[TMP25]], align 8
1066 // CHECK-NEXT:    [[TMP26:%.*]] = getelementptr inbounds [3 x ptr], ptr [[CAPTURED_VARS_ADDRS_ASCAST]], i64 0, i64 1
1067 // CHECK-NEXT:    [[TMP27:%.*]] = inttoptr i64 [[TMP18]] to ptr
1068 // CHECK-NEXT:    store ptr [[TMP27]], ptr [[TMP26]], align 8
1069 // CHECK-NEXT:    [[TMP28:%.*]] = getelementptr inbounds [3 x ptr], ptr [[CAPTURED_VARS_ADDRS_ASCAST]], i64 0, i64 2
1070 // CHECK-NEXT:    store ptr [[A]], ptr [[TMP28]], align 8
1071 // CHECK-NEXT:    [[TMP29:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR_ASCAST]], align 8
1072 // CHECK-NEXT:    [[TMP30:%.*]] = load i32, ptr [[TMP29]], align 4
1073 // CHECK-NEXT:    call void @__kmpc_parallel_51(ptr addrspacecast (ptr addrspace(1) @[[GLOB1]] to ptr), i32 [[TMP30]], i32 1, i32 -1, i32 -1, ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z4foo4v_l76_omp_outlined_omp_outlined, ptr @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z4foo4v_l76_omp_outlined_omp_outlined_wrapper, ptr [[CAPTURED_VARS_ADDRS_ASCAST]], i64 3)
1074 // CHECK-NEXT:    store i32 0, ptr [[J_ASCAST]], align 4
1075 // CHECK-NEXT:    br label [[FOR_COND:%.*]]
1076 // CHECK:       for.cond:
1077 // CHECK-NEXT:    [[TMP31:%.*]] = load i32, ptr [[J_ASCAST]], align 4
1078 // CHECK-NEXT:    [[TMP32:%.*]] = load i32, ptr [[N1]], align 4
1079 // CHECK-NEXT:    [[CMP7:%.*]] = icmp slt i32 [[TMP31]], [[TMP32]]
1080 // CHECK-NEXT:    br i1 [[CMP7]], label [[FOR_BODY:%.*]], label [[FOR_END:%.*]]
1081 // CHECK:       for.body:
1082 // CHECK-NEXT:    [[TMP33:%.*]] = load i32, ptr [[J_ASCAST]], align 4
1083 // CHECK-NEXT:    [[IDXPROM8:%.*]] = sext i32 [[TMP33]] to i64
1084 // CHECK-NEXT:    [[ARRAYIDX9:%.*]] = getelementptr inbounds i32, ptr [[A]], i64 [[IDXPROM8]]
1085 // CHECK-NEXT:    [[TMP34:%.*]] = load i32, ptr [[ARRAYIDX9]], align 4
1086 // CHECK-NEXT:    [[TMP35:%.*]] = load i32, ptr [[I4_ASCAST]], align 4
1087 // CHECK-NEXT:    [[IDXPROM10:%.*]] = sext i32 [[TMP35]] to i64
1088 // CHECK-NEXT:    [[ARRAYIDX11:%.*]] = getelementptr inbounds i32, ptr [[TMP1]], i64 [[IDXPROM10]]
1089 // CHECK-NEXT:    [[TMP36:%.*]] = load i32, ptr [[ARRAYIDX11]], align 4
1090 // CHECK-NEXT:    [[ADD12:%.*]] = add nsw i32 [[TMP36]], [[TMP34]]
1091 // CHECK-NEXT:    store i32 [[ADD12]], ptr [[ARRAYIDX11]], align 4
1092 // CHECK-NEXT:    br label [[FOR_INC:%.*]]
1093 // CHECK:       for.inc:
1094 // CHECK-NEXT:    [[TMP37:%.*]] = load i32, ptr [[J_ASCAST]], align 4
1095 // CHECK-NEXT:    [[INC:%.*]] = add nsw i32 [[TMP37]], 1
1096 // CHECK-NEXT:    store i32 [[INC]], ptr [[J_ASCAST]], align 4
1097 // CHECK-NEXT:    br label [[FOR_COND]], !llvm.loop [[LOOP19:![0-9]+]]
1098 // CHECK:       for.end:
1099 // CHECK-NEXT:    call void @__kmpc_free_shared(ptr [[A]], i64 [[TMP22]])
1100 // CHECK-NEXT:    br label [[OMP_BODY_CONTINUE:%.*]]
1101 // CHECK:       omp.body.continue:
1102 // CHECK-NEXT:    br label [[OMP_INNER_FOR_INC:%.*]]
1103 // CHECK:       omp.inner.for.inc:
1104 // CHECK-NEXT:    [[TMP38:%.*]] = load i32, ptr [[DOTOMP_IV_ASCAST]], align 4
1105 // CHECK-NEXT:    [[ADD13:%.*]] = add nsw i32 [[TMP38]], 1
1106 // CHECK-NEXT:    store i32 [[ADD13]], ptr [[DOTOMP_IV_ASCAST]], align 4
1107 // CHECK-NEXT:    br label [[OMP_INNER_FOR_COND]]
1108 // CHECK:       omp.inner.for.end:
1109 // CHECK-NEXT:    br label [[OMP_LOOP_EXIT:%.*]]
1110 // CHECK:       omp.loop.exit:
1111 // CHECK-NEXT:    [[TMP39:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR_ASCAST]], align 8
1112 // CHECK-NEXT:    [[TMP40:%.*]] = load i32, ptr [[TMP39]], align 4
1113 // CHECK-NEXT:    call void @__kmpc_distribute_static_fini(ptr addrspacecast (ptr addrspace(1) @[[GLOB2]] to ptr), i32 [[TMP40]])
1114 // CHECK-NEXT:    br label [[OMP_PRECOND_END]]
1115 // CHECK:       omp.precond.end:
1116 // CHECK-NEXT:    call void @__kmpc_free_shared(ptr [[N1]], i64 4)
1117 // CHECK-NEXT:    ret void
1118 //
1119 //
1120 // CHECK-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z4foo4v_l76_omp_outlined_omp_outlined
1121 // CHECK-SAME: (ptr noalias noundef [[DOTGLOBAL_TID_:%.*]], ptr noalias noundef [[DOTBOUND_TID_:%.*]], ptr noundef nonnull align 4 dereferenceable(4) [[N:%.*]], i64 noundef [[VLA:%.*]], ptr noundef nonnull align 4 dereferenceable(4) [[A:%.*]]) #[[ATTR3]] {
1122 // CHECK-NEXT:  entry:
1123 // CHECK-NEXT:    [[DOTGLOBAL_TID__ADDR:%.*]] = alloca ptr, align 8, addrspace(5)
1124 // CHECK-NEXT:    [[DOTBOUND_TID__ADDR:%.*]] = alloca ptr, align 8, addrspace(5)
1125 // CHECK-NEXT:    [[N_ADDR:%.*]] = alloca ptr, align 8, addrspace(5)
1126 // CHECK-NEXT:    [[VLA_ADDR:%.*]] = alloca i64, align 8, addrspace(5)
1127 // CHECK-NEXT:    [[A_ADDR:%.*]] = alloca ptr, align 8, addrspace(5)
1128 // CHECK-NEXT:    [[DOTOMP_IV:%.*]] = alloca i32, align 4, addrspace(5)
1129 // CHECK-NEXT:    [[TMP:%.*]] = alloca i32, align 4, addrspace(5)
1130 // CHECK-NEXT:    [[DOTCAPTURE_EXPR_:%.*]] = alloca i32, align 4, addrspace(5)
1131 // CHECK-NEXT:    [[DOTCAPTURE_EXPR_1:%.*]] = alloca i32, align 4, addrspace(5)
1132 // CHECK-NEXT:    [[J:%.*]] = alloca i32, align 4, addrspace(5)
1133 // CHECK-NEXT:    [[DOTOMP_LB:%.*]] = alloca i32, align 4, addrspace(5)
1134 // CHECK-NEXT:    [[DOTOMP_UB:%.*]] = alloca i32, align 4, addrspace(5)
1135 // CHECK-NEXT:    [[DOTOMP_STRIDE:%.*]] = alloca i32, align 4, addrspace(5)
1136 // CHECK-NEXT:    [[DOTOMP_IS_LAST:%.*]] = alloca i32, align 4, addrspace(5)
1137 // CHECK-NEXT:    [[J3:%.*]] = alloca i32, align 4, addrspace(5)
1138 // CHECK-NEXT:    [[DOTGLOBAL_TID__ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTGLOBAL_TID__ADDR]] to ptr
1139 // CHECK-NEXT:    [[DOTBOUND_TID__ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTBOUND_TID__ADDR]] to ptr
1140 // CHECK-NEXT:    [[N_ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[N_ADDR]] to ptr
1141 // CHECK-NEXT:    [[VLA_ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[VLA_ADDR]] to ptr
1142 // CHECK-NEXT:    [[A_ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[A_ADDR]] to ptr
1143 // CHECK-NEXT:    [[DOTOMP_IV_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTOMP_IV]] to ptr
1144 // CHECK-NEXT:    [[TMP_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[TMP]] to ptr
1145 // CHECK-NEXT:    [[DOTCAPTURE_EXPR__ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTCAPTURE_EXPR_]] to ptr
1146 // CHECK-NEXT:    [[DOTCAPTURE_EXPR_1_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTCAPTURE_EXPR_1]] to ptr
1147 // CHECK-NEXT:    [[J_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[J]] to ptr
1148 // CHECK-NEXT:    [[DOTOMP_LB_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTOMP_LB]] to ptr
1149 // CHECK-NEXT:    [[DOTOMP_UB_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTOMP_UB]] to ptr
1150 // CHECK-NEXT:    [[DOTOMP_STRIDE_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTOMP_STRIDE]] to ptr
1151 // CHECK-NEXT:    [[DOTOMP_IS_LAST_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTOMP_IS_LAST]] to ptr
1152 // CHECK-NEXT:    [[J3_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[J3]] to ptr
1153 // CHECK-NEXT:    store ptr [[DOTGLOBAL_TID_]], ptr [[DOTGLOBAL_TID__ADDR_ASCAST]], align 8
1154 // CHECK-NEXT:    store ptr [[DOTBOUND_TID_]], ptr [[DOTBOUND_TID__ADDR_ASCAST]], align 8
1155 // CHECK-NEXT:    store ptr [[N]], ptr [[N_ADDR_ASCAST]], align 8
1156 // CHECK-NEXT:    store i64 [[VLA]], ptr [[VLA_ADDR_ASCAST]], align 8
1157 // CHECK-NEXT:    store ptr [[A]], ptr [[A_ADDR_ASCAST]], align 8
1158 // CHECK-NEXT:    [[TMP0:%.*]] = load ptr, ptr [[N_ADDR_ASCAST]], align 8
1159 // CHECK-NEXT:    [[TMP1:%.*]] = load i64, ptr [[VLA_ADDR_ASCAST]], align 8
1160 // CHECK-NEXT:    [[TMP2:%.*]] = load ptr, ptr [[A_ADDR_ASCAST]], align 8
1161 // CHECK-NEXT:    [[TMP3:%.*]] = load i32, ptr [[TMP0]], align 4
1162 // CHECK-NEXT:    store i32 [[TMP3]], ptr [[DOTCAPTURE_EXPR__ASCAST]], align 4
1163 // CHECK-NEXT:    [[TMP4:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR__ASCAST]], align 4
1164 // CHECK-NEXT:    [[SUB:%.*]] = sub nsw i32 [[TMP4]], 0
1165 // CHECK-NEXT:    [[DIV:%.*]] = sdiv i32 [[SUB]], 1
1166 // CHECK-NEXT:    [[SUB2:%.*]] = sub nsw i32 [[DIV]], 1
1167 // CHECK-NEXT:    store i32 [[SUB2]], ptr [[DOTCAPTURE_EXPR_1_ASCAST]], align 4
1168 // CHECK-NEXT:    store i32 0, ptr [[J_ASCAST]], align 4
1169 // CHECK-NEXT:    [[TMP5:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR__ASCAST]], align 4
1170 // CHECK-NEXT:    [[CMP:%.*]] = icmp slt i32 0, [[TMP5]]
1171 // CHECK-NEXT:    br i1 [[CMP]], label [[OMP_PRECOND_THEN:%.*]], label [[OMP_PRECOND_END:%.*]]
1172 // CHECK:       omp.precond.then:
1173 // CHECK-NEXT:    store i32 0, ptr [[DOTOMP_LB_ASCAST]], align 4
1174 // CHECK-NEXT:    [[TMP6:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_1_ASCAST]], align 4
1175 // CHECK-NEXT:    store i32 [[TMP6]], ptr [[DOTOMP_UB_ASCAST]], align 4
1176 // CHECK-NEXT:    store i32 1, ptr [[DOTOMP_STRIDE_ASCAST]], align 4
1177 // CHECK-NEXT:    store i32 0, ptr [[DOTOMP_IS_LAST_ASCAST]], align 4
1178 // CHECK-NEXT:    [[TMP7:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR_ASCAST]], align 8
1179 // CHECK-NEXT:    [[TMP8:%.*]] = load i32, ptr [[TMP7]], align 4
1180 // CHECK-NEXT:    call void @__kmpc_for_static_init_4(ptr addrspacecast (ptr addrspace(1) @[[GLOB3]] to ptr), i32 [[TMP8]], i32 33, ptr [[DOTOMP_IS_LAST_ASCAST]], ptr [[DOTOMP_LB_ASCAST]], ptr [[DOTOMP_UB_ASCAST]], ptr [[DOTOMP_STRIDE_ASCAST]], i32 1, i32 1)
1181 // CHECK-NEXT:    br label [[OMP_DISPATCH_COND:%.*]]
1182 // CHECK:       omp.dispatch.cond:
1183 // CHECK-NEXT:    [[TMP9:%.*]] = load i32, ptr [[DOTOMP_UB_ASCAST]], align 4
1184 // CHECK-NEXT:    [[TMP10:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_1_ASCAST]], align 4
1185 // CHECK-NEXT:    [[CMP4:%.*]] = icmp sgt i32 [[TMP9]], [[TMP10]]
1186 // CHECK-NEXT:    br i1 [[CMP4]], label [[COND_TRUE:%.*]], label [[COND_FALSE:%.*]]
1187 // CHECK:       cond.true:
1188 // CHECK-NEXT:    [[TMP11:%.*]] = load i32, ptr [[DOTCAPTURE_EXPR_1_ASCAST]], align 4
1189 // CHECK-NEXT:    br label [[COND_END:%.*]]
1190 // CHECK:       cond.false:
1191 // CHECK-NEXT:    [[TMP12:%.*]] = load i32, ptr [[DOTOMP_UB_ASCAST]], align 4
1192 // CHECK-NEXT:    br label [[COND_END]]
1193 // CHECK:       cond.end:
1194 // CHECK-NEXT:    [[COND:%.*]] = phi i32 [ [[TMP11]], [[COND_TRUE]] ], [ [[TMP12]], [[COND_FALSE]] ]
1195 // CHECK-NEXT:    store i32 [[COND]], ptr [[DOTOMP_UB_ASCAST]], align 4
1196 // CHECK-NEXT:    [[TMP13:%.*]] = load i32, ptr [[DOTOMP_LB_ASCAST]], align 4
1197 // CHECK-NEXT:    store i32 [[TMP13]], ptr [[DOTOMP_IV_ASCAST]], align 4
1198 // CHECK-NEXT:    [[TMP14:%.*]] = load i32, ptr [[DOTOMP_IV_ASCAST]], align 4
1199 // CHECK-NEXT:    [[TMP15:%.*]] = load i32, ptr [[DOTOMP_UB_ASCAST]], align 4
1200 // CHECK-NEXT:    [[CMP5:%.*]] = icmp sle i32 [[TMP14]], [[TMP15]]
1201 // CHECK-NEXT:    br i1 [[CMP5]], label [[OMP_DISPATCH_BODY:%.*]], label [[OMP_DISPATCH_END:%.*]]
1202 // CHECK:       omp.dispatch.body:
1203 // CHECK-NEXT:    br label [[OMP_INNER_FOR_COND:%.*]]
1204 // CHECK:       omp.inner.for.cond:
1205 // CHECK-NEXT:    [[TMP16:%.*]] = load i32, ptr [[DOTOMP_IV_ASCAST]], align 4
1206 // CHECK-NEXT:    [[TMP17:%.*]] = load i32, ptr [[DOTOMP_UB_ASCAST]], align 4
1207 // CHECK-NEXT:    [[CMP6:%.*]] = icmp sle i32 [[TMP16]], [[TMP17]]
1208 // CHECK-NEXT:    br i1 [[CMP6]], label [[OMP_INNER_FOR_BODY:%.*]], label [[OMP_INNER_FOR_END:%.*]]
1209 // CHECK:       omp.inner.for.body:
1210 // CHECK-NEXT:    [[TMP18:%.*]] = load i32, ptr [[DOTOMP_IV_ASCAST]], align 4
1211 // CHECK-NEXT:    [[MUL:%.*]] = mul nsw i32 [[TMP18]], 1
1212 // CHECK-NEXT:    [[ADD:%.*]] = add nsw i32 0, [[MUL]]
1213 // CHECK-NEXT:    store i32 [[ADD]], ptr [[J3_ASCAST]], align 4
1214 // CHECK-NEXT:    [[TMP19:%.*]] = load i32, ptr [[J3_ASCAST]], align 4
1215 // CHECK-NEXT:    [[TMP20:%.*]] = load i32, ptr [[J3_ASCAST]], align 4
1216 // CHECK-NEXT:    [[IDXPROM:%.*]] = sext i32 [[TMP20]] to i64
1217 // CHECK-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds i32, ptr [[TMP2]], i64 [[IDXPROM]]
1218 // CHECK-NEXT:    store i32 [[TMP19]], ptr [[ARRAYIDX]], align 4
1219 // CHECK-NEXT:    br label [[OMP_BODY_CONTINUE:%.*]]
1220 // CHECK:       omp.body.continue:
1221 // CHECK-NEXT:    br label [[OMP_INNER_FOR_INC:%.*]]
1222 // CHECK:       omp.inner.for.inc:
1223 // CHECK-NEXT:    [[TMP21:%.*]] = load i32, ptr [[DOTOMP_IV_ASCAST]], align 4
1224 // CHECK-NEXT:    [[ADD7:%.*]] = add nsw i32 [[TMP21]], 1
1225 // CHECK-NEXT:    store i32 [[ADD7]], ptr [[DOTOMP_IV_ASCAST]], align 4
1226 // CHECK-NEXT:    br label [[OMP_INNER_FOR_COND]]
1227 // CHECK:       omp.inner.for.end:
1228 // CHECK-NEXT:    br label [[OMP_DISPATCH_INC:%.*]]
1229 // CHECK:       omp.dispatch.inc:
1230 // CHECK-NEXT:    [[TMP22:%.*]] = load i32, ptr [[DOTOMP_LB_ASCAST]], align 4
1231 // CHECK-NEXT:    [[TMP23:%.*]] = load i32, ptr [[DOTOMP_STRIDE_ASCAST]], align 4
1232 // CHECK-NEXT:    [[ADD8:%.*]] = add nsw i32 [[TMP22]], [[TMP23]]
1233 // CHECK-NEXT:    store i32 [[ADD8]], ptr [[DOTOMP_LB_ASCAST]], align 4
1234 // CHECK-NEXT:    [[TMP24:%.*]] = load i32, ptr [[DOTOMP_UB_ASCAST]], align 4
1235 // CHECK-NEXT:    [[TMP25:%.*]] = load i32, ptr [[DOTOMP_STRIDE_ASCAST]], align 4
1236 // CHECK-NEXT:    [[ADD9:%.*]] = add nsw i32 [[TMP24]], [[TMP25]]
1237 // CHECK-NEXT:    store i32 [[ADD9]], ptr [[DOTOMP_UB_ASCAST]], align 4
1238 // CHECK-NEXT:    br label [[OMP_DISPATCH_COND]]
1239 // CHECK:       omp.dispatch.end:
1240 // CHECK-NEXT:    [[TMP26:%.*]] = load ptr, ptr [[DOTGLOBAL_TID__ADDR_ASCAST]], align 8
1241 // CHECK-NEXT:    [[TMP27:%.*]] = load i32, ptr [[TMP26]], align 4
1242 // CHECK-NEXT:    call void @__kmpc_for_static_fini(ptr addrspacecast (ptr addrspace(1) @[[GLOB3]] to ptr), i32 [[TMP27]])
1243 // CHECK-NEXT:    br label [[OMP_PRECOND_END]]
1244 // CHECK:       omp.precond.end:
1245 // CHECK-NEXT:    ret void
1246 //
1247 //
1248 // CHECK-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z4foo4v_l76_omp_outlined_omp_outlined_wrapper
1249 // CHECK-SAME: (i16 noundef zeroext [[TMP0:%.*]], i32 noundef [[TMP1:%.*]]) #[[ATTR7]] {
1250 // CHECK-NEXT:  entry:
1251 // CHECK-NEXT:    [[DOTADDR:%.*]] = alloca i16, align 2, addrspace(5)
1252 // CHECK-NEXT:    [[DOTADDR1:%.*]] = alloca i32, align 4, addrspace(5)
1253 // CHECK-NEXT:    [[DOTZERO_ADDR:%.*]] = alloca i32, align 4, addrspace(5)
1254 // CHECK-NEXT:    [[GLOBAL_ARGS:%.*]] = alloca ptr, align 8, addrspace(5)
1255 // CHECK-NEXT:    [[DOTADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTADDR]] to ptr
1256 // CHECK-NEXT:    [[DOTADDR1_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTADDR1]] to ptr
1257 // CHECK-NEXT:    [[DOTZERO_ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[DOTZERO_ADDR]] to ptr
1258 // CHECK-NEXT:    [[GLOBAL_ARGS_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[GLOBAL_ARGS]] to ptr
1259 // CHECK-NEXT:    store i16 [[TMP0]], ptr [[DOTADDR_ASCAST]], align 2
1260 // CHECK-NEXT:    store i32 [[TMP1]], ptr [[DOTADDR1_ASCAST]], align 4
1261 // CHECK-NEXT:    store i32 0, ptr [[DOTZERO_ADDR_ASCAST]], align 4
1262 // CHECK-NEXT:    call void @__kmpc_get_shared_variables(ptr [[GLOBAL_ARGS_ASCAST]])
1263 // CHECK-NEXT:    [[TMP2:%.*]] = load ptr, ptr [[GLOBAL_ARGS_ASCAST]], align 8
1264 // CHECK-NEXT:    [[TMP3:%.*]] = getelementptr inbounds ptr, ptr [[TMP2]], i64 0
1265 // CHECK-NEXT:    [[TMP4:%.*]] = load ptr, ptr [[TMP3]], align 8
1266 // CHECK-NEXT:    [[TMP5:%.*]] = getelementptr inbounds ptr, ptr [[TMP2]], i64 1
1267 // CHECK-NEXT:    [[TMP6:%.*]] = load i64, ptr [[TMP5]], align 8
1268 // CHECK-NEXT:    [[TMP7:%.*]] = getelementptr inbounds ptr, ptr [[TMP2]], i64 2
1269 // CHECK-NEXT:    [[TMP8:%.*]] = load ptr, ptr [[TMP7]], align 8
1270 // CHECK-NEXT:    call void @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z4foo4v_l76_omp_outlined_omp_outlined(ptr [[DOTADDR1_ASCAST]], ptr [[DOTZERO_ADDR_ASCAST]], ptr [[TMP4]], i64 [[TMP6]], ptr [[TMP8]]) #[[ATTR4]]
1271 // CHECK-NEXT:    ret void
1272 //