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 // expected-no-diagnostics
3 #ifndef HEADER
4 #define HEADER
5 
6 ///==========================================================================///
7 // RUN: %clang_cc1 -no-opaque-pointers -DCK4 -verify -fopenmp -fopenmp-targets=powerpc64le-ibm-linux-gnu -x c++ -triple powerpc64le-unknown-unknown -emit-llvm %s -o - | FileCheck -allow-deprecated-dag-overlap  %s --check-prefix=CHECK1
8 // RUN: %clang_cc1 -no-opaque-pointers -DCK4 -fopenmp -fopenmp-targets=powerpc64le-ibm-linux-gnu -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -emit-pch -o %t %s
9 // RUN: %clang_cc1 -no-opaque-pointers -fopenmp -fopenmp-targets=powerpc64le-ibm-linux-gnu -x c++ -triple powerpc64le-unknown-unknown -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck -allow-deprecated-dag-overlap  %s  --check-prefix=CHECK1
10 // RUN: %clang_cc1 -no-opaque-pointers -DCK4 -verify -fopenmp -fopenmp-targets=i386-pc-linux-gnu -x c++ -triple i386-unknown-unknown -emit-llvm %s -o - | FileCheck -allow-deprecated-dag-overlap  %s  --check-prefix=CHECK3
11 // RUN: %clang_cc1 -no-opaque-pointers -DCK4 -fopenmp -fopenmp-targets=i386-pc-linux-gnu -x c++ -std=c++11 -triple i386-unknown-unknown -emit-pch -o %t %s
12 // RUN: %clang_cc1 -no-opaque-pointers -fopenmp -fopenmp-targets=i386-pc-linux-gnu -x c++ -triple i386-unknown-unknown -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck -allow-deprecated-dag-overlap  %s  --check-prefix=CHECK3
13 
14 // RUN: %clang_cc1 -no-opaque-pointers -DCK4 -verify -fopenmp -fopenmp-version=45 -fopenmp-targets=powerpc64le-ibm-linux-gnu -x c++ -triple powerpc64le-unknown-unknown -emit-llvm %s -o - | FileCheck -allow-deprecated-dag-overlap  %s --check-prefix=CHECK1
15 // RUN: %clang_cc1 -no-opaque-pointers -DCK4 -fopenmp -fopenmp-version=45 -fopenmp-targets=powerpc64le-ibm-linux-gnu -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -emit-pch -o %t %s
16 // RUN: %clang_cc1 -no-opaque-pointers -fopenmp -fopenmp-version=45 -fopenmp-targets=powerpc64le-ibm-linux-gnu -x c++ -triple powerpc64le-unknown-unknown -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck -allow-deprecated-dag-overlap  %s  --check-prefix=CHECK1
17 // RUN: %clang_cc1 -no-opaque-pointers -DCK4 -verify -fopenmp -fopenmp-version=45 -fopenmp-targets=i386-pc-linux-gnu -x c++ -triple i386-unknown-unknown -emit-llvm %s -o - | FileCheck -allow-deprecated-dag-overlap  %s  --check-prefix=CHECK3
18 // RUN: %clang_cc1 -no-opaque-pointers -DCK4 -fopenmp -fopenmp-version=45 -fopenmp-targets=i386-pc-linux-gnu -x c++ -std=c++11 -triple i386-unknown-unknown -emit-pch -o %t %s
19 // RUN: %clang_cc1 -no-opaque-pointers -fopenmp -fopenmp-version=45 -fopenmp-targets=i386-pc-linux-gnu -x c++ -triple i386-unknown-unknown -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck -allow-deprecated-dag-overlap  %s  --check-prefix=CHECK3
20 
21 // RUN: %clang_cc1 -no-opaque-pointers -DCK4 -verify -fopenmp -fopenmp-version=50 -fopenmp-targets=powerpc64le-ibm-linux-gnu -x c++ -triple powerpc64le-unknown-unknown -emit-llvm %s -o - | FileCheck -allow-deprecated-dag-overlap  %s --check-prefix=CHECK1
22 // RUN: %clang_cc1 -no-opaque-pointers -DCK4 -fopenmp -fopenmp-version=50 -fopenmp-targets=powerpc64le-ibm-linux-gnu -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -emit-pch -o %t %s
23 // RUN: %clang_cc1 -no-opaque-pointers -fopenmp -fopenmp-version=50 -fopenmp-targets=powerpc64le-ibm-linux-gnu -x c++ -triple powerpc64le-unknown-unknown -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck -allow-deprecated-dag-overlap  %s  --check-prefix=CHECK1
24 // RUN: %clang_cc1 -no-opaque-pointers -DCK4 -verify -fopenmp -fopenmp-version=50 -fopenmp-targets=i386-pc-linux-gnu -x c++ -triple i386-unknown-unknown -emit-llvm %s -o - | FileCheck -allow-deprecated-dag-overlap  %s  --check-prefix=CHECK3
25 // RUN: %clang_cc1 -no-opaque-pointers -DCK4 -fopenmp -fopenmp-version=50 -fopenmp-targets=i386-pc-linux-gnu -x c++ -std=c++11 -triple i386-unknown-unknown -emit-pch -o %t %s
26 // RUN: %clang_cc1 -no-opaque-pointers -fopenmp -fopenmp-version=50 -fopenmp-targets=i386-pc-linux-gnu -x c++ -triple i386-unknown-unknown -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck -allow-deprecated-dag-overlap  %s  --check-prefix=CHECK3
27 
28 // RUN: %clang_cc1 -no-opaque-pointers -DCK4 -verify -fopenmp-simd -fopenmp-targets=powerpc64le-ibm-linux-gnu -x c++ -triple powerpc64le-unknown-unknown -emit-llvm %s -o - | FileCheck %s --implicit-check-not="{{__kmpc|__tgt}}"
29 // RUN: %clang_cc1 -no-opaque-pointers -DCK4 -fopenmp-simd -fopenmp-targets=powerpc64le-ibm-linux-gnu -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -emit-pch -o %t %s
30 // RUN: %clang_cc1 -no-opaque-pointers -fopenmp-simd -fopenmp-targets=powerpc64le-ibm-linux-gnu -x c++ -triple powerpc64le-unknown-unknown -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --implicit-check-not="{{__kmpc|__tgt}}"
31 // RUN: %clang_cc1 -no-opaque-pointers -DCK4 -verify -fopenmp-simd -fopenmp-targets=i386-pc-linux-gnu -x c++ -triple i386-unknown-unknown -emit-llvm %s -o - | FileCheck %s --implicit-check-not="{{__kmpc|__tgt}}"
32 // RUN: %clang_cc1 -no-opaque-pointers -DCK4 -fopenmp-simd -fopenmp-targets=i386-pc-linux-gnu -x c++ -std=c++11 -triple i386-unknown-unknown -emit-pch -o %t %s
33 // RUN: %clang_cc1 -no-opaque-pointers -fopenmp-simd -fopenmp-targets=i386-pc-linux-gnu -x c++ -triple i386-unknown-unknown -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --implicit-check-not="{{__kmpc|__tgt}}"
34 #ifdef CK4
35 
36 
37 // Map types: OMP_MAP_PRIVATE_VAL | OMP_MAP_TARGET_PARAM | OMP_MAP_IMPLICIT = 800
38 
39 void implicit_maps_nested_integer (int a){
40   int i = a;
41 
42   // The captures in parallel are by reference. Only the capture in target is by
43   // copy.
44 
45   #pragma omp parallel
46   {
47 
48     #pragma omp target
49     {
50       #pragma omp parallel
51       {
52         ++i;
53       }
54     }
55   }
56 }
57 
58 #endif // CK4
59 #endif
60 // CHECK1-LABEL: define {{[^@]+}}@_Z28implicit_maps_nested_integeri
61 // CHECK1-SAME: (i32 noundef signext [[A:%.*]]) #[[ATTR0:[0-9]+]] {
62 // CHECK1-NEXT:  entry:
63 // CHECK1-NEXT:    [[A_ADDR:%.*]] = alloca i32, align 4
64 // CHECK1-NEXT:    [[I:%.*]] = alloca i32, align 4
65 // CHECK1-NEXT:    store i32 [[A]], i32* [[A_ADDR]], align 4
66 // CHECK1-NEXT:    [[TMP0:%.*]] = load i32, i32* [[A_ADDR]], align 4
67 // CHECK1-NEXT:    store i32 [[TMP0]], i32* [[I]], align 4
68 // CHECK1-NEXT:    call void (%struct.ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call(%struct.ident_t* @[[GLOB1:[0-9]+]], i32 1, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i32*)* @.omp_outlined. to void (i32*, i32*, ...)*), i32* [[I]])
69 // CHECK1-NEXT:    ret void
70 //
71 //
72 // CHECK1-LABEL: define {{[^@]+}}@.omp_outlined.
73 // CHECK1-SAME: (i32* noalias noundef [[DOTGLOBAL_TID_:%.*]], i32* noalias noundef [[DOTBOUND_TID_:%.*]], i32* noundef nonnull align 4 dereferenceable(4) [[I:%.*]]) #[[ATTR1:[0-9]+]] {
74 // CHECK1-NEXT:  entry:
75 // CHECK1-NEXT:    [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8
76 // CHECK1-NEXT:    [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8
77 // CHECK1-NEXT:    [[I_ADDR:%.*]] = alloca i32*, align 8
78 // CHECK1-NEXT:    [[I_CASTED:%.*]] = alloca i64, align 8
79 // CHECK1-NEXT:    [[DOTOFFLOAD_BASEPTRS:%.*]] = alloca [1 x i8*], align 8
80 // CHECK1-NEXT:    [[DOTOFFLOAD_PTRS:%.*]] = alloca [1 x i8*], align 8
81 // CHECK1-NEXT:    [[DOTOFFLOAD_MAPPERS:%.*]] = alloca [1 x i8*], align 8
82 // CHECK1-NEXT:    store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 8
83 // CHECK1-NEXT:    store i32* [[DOTBOUND_TID_]], i32** [[DOTBOUND_TID__ADDR]], align 8
84 // CHECK1-NEXT:    store i32* [[I]], i32** [[I_ADDR]], align 8
85 // CHECK1-NEXT:    [[TMP0:%.*]] = load i32*, i32** [[I_ADDR]], align 8
86 // CHECK1-NEXT:    [[TMP1:%.*]] = load i32, i32* [[TMP0]], align 4
87 // CHECK1-NEXT:    [[CONV:%.*]] = bitcast i64* [[I_CASTED]] to i32*
88 // CHECK1-NEXT:    store i32 [[TMP1]], i32* [[CONV]], align 4
89 // CHECK1-NEXT:    [[TMP2:%.*]] = load i64, i64* [[I_CASTED]], align 8
90 // CHECK1-NEXT:    [[TMP3:%.*]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[DOTOFFLOAD_BASEPTRS]], i32 0, i32 0
91 // CHECK1-NEXT:    [[TMP4:%.*]] = bitcast i8** [[TMP3]] to i64*
92 // CHECK1-NEXT:    store i64 [[TMP2]], i64* [[TMP4]], align 8
93 // CHECK1-NEXT:    [[TMP5:%.*]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[DOTOFFLOAD_PTRS]], i32 0, i32 0
94 // CHECK1-NEXT:    [[TMP6:%.*]] = bitcast i8** [[TMP5]] to i64*
95 // CHECK1-NEXT:    store i64 [[TMP2]], i64* [[TMP6]], align 8
96 // CHECK1-NEXT:    [[TMP7:%.*]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[DOTOFFLOAD_MAPPERS]], i64 0, i64 0
97 // CHECK1-NEXT:    store i8* null, i8** [[TMP7]], align 8
98 // CHECK1-NEXT:    [[TMP8:%.*]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[DOTOFFLOAD_BASEPTRS]], i32 0, i32 0
99 // CHECK1-NEXT:    [[TMP9:%.*]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[DOTOFFLOAD_PTRS]], i32 0, i32 0
100 // CHECK1-NEXT:    [[KERNEL_ARGS:%.*]] = alloca [[STRUCT___TGT_KERNEL_ARGUMENTS:%.*]], align 8
101 // CHECK1-NEXT:    [[TMP10:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], %struct.__tgt_kernel_arguments* [[KERNEL_ARGS]], i32 0, i32 0
102 // CHECK1-NEXT:    store i32 1, i32* [[TMP10]], align 4
103 // CHECK1-NEXT:    [[TMP11:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], %struct.__tgt_kernel_arguments* [[KERNEL_ARGS]], i32 0, i32 1
104 // CHECK1-NEXT:    store i32 1, i32* [[TMP11]], align 4
105 // CHECK1-NEXT:    [[TMP12:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], %struct.__tgt_kernel_arguments* [[KERNEL_ARGS]], i32 0, i32 2
106 // CHECK1-NEXT:    store i8** [[TMP8]], i8*** [[TMP12]], align 8
107 // CHECK1-NEXT:    [[TMP13:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], %struct.__tgt_kernel_arguments* [[KERNEL_ARGS]], i32 0, i32 3
108 // CHECK1-NEXT:    store i8** [[TMP9]], i8*** [[TMP13]], align 8
109 // CHECK1-NEXT:    [[TMP14:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], %struct.__tgt_kernel_arguments* [[KERNEL_ARGS]], i32 0, i32 4
110 // CHECK1-NEXT:    store i64* getelementptr inbounds ([1 x i64], [1 x i64]* @.offload_sizes, i32 0, i32 0), i64** [[TMP14]], align 8
111 // CHECK1-NEXT:    [[TMP15:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], %struct.__tgt_kernel_arguments* [[KERNEL_ARGS]], i32 0, i32 5
112 // CHECK1-NEXT:    store i64* getelementptr inbounds ([1 x i64], [1 x i64]* @.offload_maptypes, i32 0, i32 0), i64** [[TMP15]], align 8
113 // CHECK1-NEXT:    [[TMP16:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], %struct.__tgt_kernel_arguments* [[KERNEL_ARGS]], i32 0, i32 6
114 // CHECK1-NEXT:    store i8** null, i8*** [[TMP16]], align 8
115 // CHECK1-NEXT:    [[TMP17:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], %struct.__tgt_kernel_arguments* [[KERNEL_ARGS]], i32 0, i32 7
116 // CHECK1-NEXT:    store i8** null, i8*** [[TMP17]], align 8
117 // CHECK1-NEXT:    [[TMP18:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], %struct.__tgt_kernel_arguments* [[KERNEL_ARGS]], i32 0, i32 8
118 // CHECK1-NEXT:    store i64 0, i64* [[TMP18]], align 8
119 // CHECK1-NEXT:    [[TMP19:%.*]] = call i32 @__tgt_target_kernel(%struct.ident_t* @[[GLOB1]], i64 -1, i32 1, i32 0, i8* @.{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z28implicit_maps_nested_integeri_l48.region_id, %struct.__tgt_kernel_arguments* [[KERNEL_ARGS]])
120 // CHECK1-NEXT:    [[TMP20:%.*]] = icmp ne i32 [[TMP19]], 0
121 // CHECK1-NEXT:    br i1 [[TMP20]], label [[OMP_OFFLOAD_FAILED:%.*]], label [[OMP_OFFLOAD_CONT:%.*]]
122 // CHECK1:       omp_offload.failed:
123 // CHECK1-NEXT:    call void @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z28implicit_maps_nested_integeri_l48(i64 [[TMP2]]) #[[ATTR3:[0-9]+]]
124 // CHECK1-NEXT:    br label [[OMP_OFFLOAD_CONT]]
125 // CHECK1:       omp_offload.cont:
126 // CHECK1-NEXT:    ret void
127 //
128 //
129 // CHECK1-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z28implicit_maps_nested_integeri_l48
130 // CHECK1-SAME: (i64 noundef [[I:%.*]]) #[[ATTR2:[0-9]+]] {
131 // CHECK1-NEXT:  entry:
132 // CHECK1-NEXT:    [[I_ADDR:%.*]] = alloca i64, align 8
133 // CHECK1-NEXT:    store i64 [[I]], i64* [[I_ADDR]], align 8
134 // CHECK1-NEXT:    [[CONV:%.*]] = bitcast i64* [[I_ADDR]] to i32*
135 // CHECK1-NEXT:    call void (%struct.ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call(%struct.ident_t* @[[GLOB1]], i32 1, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i32*)* @.omp_outlined..1 to void (i32*, i32*, ...)*), i32* [[CONV]])
136 // CHECK1-NEXT:    ret void
137 //
138 //
139 // CHECK1-LABEL: define {{[^@]+}}@.omp_outlined..1
140 // CHECK1-SAME: (i32* noalias noundef [[DOTGLOBAL_TID_:%.*]], i32* noalias noundef [[DOTBOUND_TID_:%.*]], i32* noundef nonnull align 4 dereferenceable(4) [[I:%.*]]) #[[ATTR1]] {
141 // CHECK1-NEXT:  entry:
142 // CHECK1-NEXT:    [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8
143 // CHECK1-NEXT:    [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8
144 // CHECK1-NEXT:    [[I_ADDR:%.*]] = alloca i32*, align 8
145 // CHECK1-NEXT:    store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 8
146 // CHECK1-NEXT:    store i32* [[DOTBOUND_TID_]], i32** [[DOTBOUND_TID__ADDR]], align 8
147 // CHECK1-NEXT:    store i32* [[I]], i32** [[I_ADDR]], align 8
148 // CHECK1-NEXT:    [[TMP0:%.*]] = load i32*, i32** [[I_ADDR]], align 8
149 // CHECK1-NEXT:    [[TMP1:%.*]] = load i32, i32* [[TMP0]], align 4
150 // CHECK1-NEXT:    [[INC:%.*]] = add nsw i32 [[TMP1]], 1
151 // CHECK1-NEXT:    store i32 [[INC]], i32* [[TMP0]], align 4
152 // CHECK1-NEXT:    ret void
153 //
154 //
155 // CHECK1-LABEL: define {{[^@]+}}@.omp_offloading.requires_reg
156 // CHECK1-SAME: () #[[ATTR4:[0-9]+]] {
157 // CHECK1-NEXT:  entry:
158 // CHECK1-NEXT:    call void @__tgt_register_requires(i64 1)
159 // CHECK1-NEXT:    ret void
160 //
161 //
162 // CHECK3-LABEL: define {{[^@]+}}@_Z28implicit_maps_nested_integeri
163 // CHECK3-SAME: (i32 noundef [[A:%.*]]) #[[ATTR0:[0-9]+]] {
164 // CHECK3-NEXT:  entry:
165 // CHECK3-NEXT:    [[A_ADDR:%.*]] = alloca i32, align 4
166 // CHECK3-NEXT:    [[I:%.*]] = alloca i32, align 4
167 // CHECK3-NEXT:    store i32 [[A]], i32* [[A_ADDR]], align 4
168 // CHECK3-NEXT:    [[TMP0:%.*]] = load i32, i32* [[A_ADDR]], align 4
169 // CHECK3-NEXT:    store i32 [[TMP0]], i32* [[I]], align 4
170 // CHECK3-NEXT:    call void (%struct.ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call(%struct.ident_t* @[[GLOB1:[0-9]+]], i32 1, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i32*)* @.omp_outlined. to void (i32*, i32*, ...)*), i32* [[I]])
171 // CHECK3-NEXT:    ret void
172 //
173 //
174 // CHECK3-LABEL: define {{[^@]+}}@.omp_outlined.
175 // CHECK3-SAME: (i32* noalias noundef [[DOTGLOBAL_TID_:%.*]], i32* noalias noundef [[DOTBOUND_TID_:%.*]], i32* noundef nonnull align 4 dereferenceable(4) [[I:%.*]]) #[[ATTR1:[0-9]+]] {
176 // CHECK3-NEXT:  entry:
177 // CHECK3-NEXT:    [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 4
178 // CHECK3-NEXT:    [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 4
179 // CHECK3-NEXT:    [[I_ADDR:%.*]] = alloca i32*, align 4
180 // CHECK3-NEXT:    [[I_CASTED:%.*]] = alloca i32, align 4
181 // CHECK3-NEXT:    [[DOTOFFLOAD_BASEPTRS:%.*]] = alloca [1 x i8*], align 4
182 // CHECK3-NEXT:    [[DOTOFFLOAD_PTRS:%.*]] = alloca [1 x i8*], align 4
183 // CHECK3-NEXT:    [[DOTOFFLOAD_MAPPERS:%.*]] = alloca [1 x i8*], align 4
184 // CHECK3-NEXT:    store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 4
185 // CHECK3-NEXT:    store i32* [[DOTBOUND_TID_]], i32** [[DOTBOUND_TID__ADDR]], align 4
186 // CHECK3-NEXT:    store i32* [[I]], i32** [[I_ADDR]], align 4
187 // CHECK3-NEXT:    [[TMP0:%.*]] = load i32*, i32** [[I_ADDR]], align 4
188 // CHECK3-NEXT:    [[TMP1:%.*]] = load i32, i32* [[TMP0]], align 4
189 // CHECK3-NEXT:    store i32 [[TMP1]], i32* [[I_CASTED]], align 4
190 // CHECK3-NEXT:    [[TMP2:%.*]] = load i32, i32* [[I_CASTED]], align 4
191 // CHECK3-NEXT:    [[TMP3:%.*]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[DOTOFFLOAD_BASEPTRS]], i32 0, i32 0
192 // CHECK3-NEXT:    [[TMP4:%.*]] = bitcast i8** [[TMP3]] to i32*
193 // CHECK3-NEXT:    store i32 [[TMP2]], i32* [[TMP4]], align 4
194 // CHECK3-NEXT:    [[TMP5:%.*]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[DOTOFFLOAD_PTRS]], i32 0, i32 0
195 // CHECK3-NEXT:    [[TMP6:%.*]] = bitcast i8** [[TMP5]] to i32*
196 // CHECK3-NEXT:    store i32 [[TMP2]], i32* [[TMP6]], align 4
197 // CHECK3-NEXT:    [[TMP7:%.*]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[DOTOFFLOAD_MAPPERS]], i32 0, i32 0
198 // CHECK3-NEXT:    store i8* null, i8** [[TMP7]], align 4
199 // CHECK3-NEXT:    [[TMP8:%.*]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[DOTOFFLOAD_BASEPTRS]], i32 0, i32 0
200 // CHECK3-NEXT:    [[TMP9:%.*]] = getelementptr inbounds [1 x i8*], [1 x i8*]* [[DOTOFFLOAD_PTRS]], i32 0, i32 0
201 // CHECK3-NEXT:    [[KERNEL_ARGS:%.*]] = alloca [[STRUCT___TGT_KERNEL_ARGUMENTS:%.*]], align 8
202 // CHECK3-NEXT:    [[TMP10:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], %struct.__tgt_kernel_arguments* [[KERNEL_ARGS]], i32 0, i32 0
203 // CHECK3-NEXT:    store i32 1, i32* [[TMP10]], align 4
204 // CHECK3-NEXT:    [[TMP11:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], %struct.__tgt_kernel_arguments* [[KERNEL_ARGS]], i32 0, i32 1
205 // CHECK3-NEXT:    store i32 1, i32* [[TMP11]], align 4
206 // CHECK3-NEXT:    [[TMP12:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], %struct.__tgt_kernel_arguments* [[KERNEL_ARGS]], i32 0, i32 2
207 // CHECK3-NEXT:    store i8** [[TMP8]], i8*** [[TMP12]], align 4
208 // CHECK3-NEXT:    [[TMP13:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], %struct.__tgt_kernel_arguments* [[KERNEL_ARGS]], i32 0, i32 3
209 // CHECK3-NEXT:    store i8** [[TMP9]], i8*** [[TMP13]], align 4
210 // CHECK3-NEXT:    [[TMP14:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], %struct.__tgt_kernel_arguments* [[KERNEL_ARGS]], i32 0, i32 4
211 // CHECK3-NEXT:    store i64* getelementptr inbounds ([1 x i64], [1 x i64]* @.offload_sizes, i32 0, i32 0), i64** [[TMP14]], align 4
212 // CHECK3-NEXT:    [[TMP15:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], %struct.__tgt_kernel_arguments* [[KERNEL_ARGS]], i32 0, i32 5
213 // CHECK3-NEXT:    store i64* getelementptr inbounds ([1 x i64], [1 x i64]* @.offload_maptypes, i32 0, i32 0), i64** [[TMP15]], align 4
214 // CHECK3-NEXT:    [[TMP16:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], %struct.__tgt_kernel_arguments* [[KERNEL_ARGS]], i32 0, i32 6
215 // CHECK3-NEXT:    store i8** null, i8*** [[TMP16]], align 4
216 // CHECK3-NEXT:    [[TMP17:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], %struct.__tgt_kernel_arguments* [[KERNEL_ARGS]], i32 0, i32 7
217 // CHECK3-NEXT:    store i8** null, i8*** [[TMP17]], align 4
218 // CHECK3-NEXT:    [[TMP18:%.*]] = getelementptr inbounds [[STRUCT___TGT_KERNEL_ARGUMENTS]], %struct.__tgt_kernel_arguments* [[KERNEL_ARGS]], i32 0, i32 8
219 // CHECK3-NEXT:    store i64 0, i64* [[TMP18]], align 8
220 // CHECK3-NEXT:    [[TMP19:%.*]] = call i32 @__tgt_target_kernel(%struct.ident_t* @[[GLOB1]], i64 -1, i32 1, i32 0, i8* @.{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z28implicit_maps_nested_integeri_l48.region_id, %struct.__tgt_kernel_arguments* [[KERNEL_ARGS]])
221 // CHECK3-NEXT:    [[TMP20:%.*]] = icmp ne i32 [[TMP19]], 0
222 // CHECK3-NEXT:    br i1 [[TMP20]], label [[OMP_OFFLOAD_FAILED:%.*]], label [[OMP_OFFLOAD_CONT:%.*]]
223 // CHECK3:       omp_offload.failed:
224 // CHECK3-NEXT:    call void @{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z28implicit_maps_nested_integeri_l48(i32 [[TMP2]]) #[[ATTR3:[0-9]+]]
225 // CHECK3-NEXT:    br label [[OMP_OFFLOAD_CONT]]
226 // CHECK3:       omp_offload.cont:
227 // CHECK3-NEXT:    ret void
228 //
229 //
230 // CHECK3-LABEL: define {{[^@]+}}@{{__omp_offloading_[0-9a-z]+_[0-9a-z]+}}__Z28implicit_maps_nested_integeri_l48
231 // CHECK3-SAME: (i32 noundef [[I:%.*]]) #[[ATTR2:[0-9]+]] {
232 // CHECK3-NEXT:  entry:
233 // CHECK3-NEXT:    [[I_ADDR:%.*]] = alloca i32, align 4
234 // CHECK3-NEXT:    store i32 [[I]], i32* [[I_ADDR]], align 4
235 // CHECK3-NEXT:    call void (%struct.ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call(%struct.ident_t* @[[GLOB1]], i32 1, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i32*)* @.omp_outlined..1 to void (i32*, i32*, ...)*), i32* [[I_ADDR]])
236 // CHECK3-NEXT:    ret void
237 //
238 //
239 // CHECK3-LABEL: define {{[^@]+}}@.omp_outlined..1
240 // CHECK3-SAME: (i32* noalias noundef [[DOTGLOBAL_TID_:%.*]], i32* noalias noundef [[DOTBOUND_TID_:%.*]], i32* noundef nonnull align 4 dereferenceable(4) [[I:%.*]]) #[[ATTR1]] {
241 // CHECK3-NEXT:  entry:
242 // CHECK3-NEXT:    [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 4
243 // CHECK3-NEXT:    [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 4
244 // CHECK3-NEXT:    [[I_ADDR:%.*]] = alloca i32*, align 4
245 // CHECK3-NEXT:    store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 4
246 // CHECK3-NEXT:    store i32* [[DOTBOUND_TID_]], i32** [[DOTBOUND_TID__ADDR]], align 4
247 // CHECK3-NEXT:    store i32* [[I]], i32** [[I_ADDR]], align 4
248 // CHECK3-NEXT:    [[TMP0:%.*]] = load i32*, i32** [[I_ADDR]], align 4
249 // CHECK3-NEXT:    [[TMP1:%.*]] = load i32, i32* [[TMP0]], align 4
250 // CHECK3-NEXT:    [[INC:%.*]] = add nsw i32 [[TMP1]], 1
251 // CHECK3-NEXT:    store i32 [[INC]], i32* [[TMP0]], align 4
252 // CHECK3-NEXT:    ret void
253 //
254 //
255 // CHECK3-LABEL: define {{[^@]+}}@.omp_offloading.requires_reg
256 // CHECK3-SAME: () #[[ATTR4:[0-9]+]] {
257 // CHECK3-NEXT:  entry:
258 // CHECK3-NEXT:    call void @__tgt_register_requires(i64 1)
259 // CHECK3-NEXT:    ret void
260 //
261