1 // RUN: %clang_cc1 -verify -fopenmp -x c++ -triple x86_64-apple-darwin10 -emit-llvm %s -o - | FileCheck %s 2 // RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple x86_64-apple-darwin10 -emit-pch -o %t %s 3 // RUN: %clang_cc1 -fopenmp -x c++ -triple x86_64-apple-darwin10 -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s 4 // RUN: %clang_cc1 -verify -fopenmp -x c++ -std=c++11 -DLAMBDA -triple x86_64-apple-darwin10 -emit-llvm %s -o - | FileCheck -check-prefix=LAMBDA %s 5 // RUN: %clang_cc1 -verify -fopenmp -x c++ -fblocks -DBLOCKS -triple x86_64-apple-darwin10 -emit-llvm %s -o - | FileCheck -check-prefix=BLOCKS %s 6 // expected-no-diagnostics 7 // REQUIRES: x86-registered-target 8 #ifndef HEADER 9 #define HEADER 10 11 template <class T> 12 struct S { 13 T f; 14 S(T a) : f(a) {} 15 S() : f() {} 16 S<T> &operator=(const S<T> &); 17 operator T() { return T(); } 18 ~S() {} 19 }; 20 21 volatile int g = 1212; 22 float f; 23 char cnt; 24 25 // CHECK: [[S_FLOAT_TY:%.+]] = type { float } 26 // CHECK: [[CAP_MAIN_TY:%.+]] = type { float**, i64* } 27 // CHECK: [[S_INT_TY:%.+]] = type { i32 } 28 // CHECK: [[CAP_TMAIN_TY:%.+]] = type { i32**, i32* } 29 // CHECK-DAG: [[IMPLICIT_BARRIER_LOC:@.+]] = private unnamed_addr constant %{{.+}} { i32 0, i32 66, i32 0, i32 0, i8* 30 // CHECK-DAG: [[F:@.+]] = global float 0.0 31 // CHECK-DAG: [[CNT:@.+]] = global i8 0 32 template <typename T> 33 T tmain() { 34 S<T> test; 35 T *pvar = &test.f; 36 T lvar = T(); 37 #pragma omp parallel 38 #pragma omp for linear(pvar, lvar) 39 for (int i = 0; i < 2; ++i) { 40 ++pvar, ++lvar; 41 } 42 return T(); 43 } 44 45 int main() { 46 #ifdef LAMBDA 47 // LAMBDA: [[G:@.+]] = global i{{[0-9]+}} 1212, 48 // LAMBDA-LABEL: @main 49 // LAMBDA: call void [[OUTER_LAMBDA:@.+]]( 50 [&]() { 51 // LAMBDA: define{{.*}} internal{{.*}} void [[OUTER_LAMBDA]]( 52 // LAMBDA: call void {{.+}} @__kmpc_fork_call({{.+}}, i32 1, {{.+}}* [[OMP_REGION:@.+]] to {{.+}}, i8* %{{.+}}) 53 #pragma omp parallel 54 #pragma omp for linear(g:5) 55 for (int i = 0; i < 2; ++i) { 56 // LAMBDA: define{{.*}} internal{{.*}} void [[OMP_REGION]](i32* %{{.+}}, i32* %{{.+}}, %{{.+}}* [[ARG:%.+]]) 57 // LAMBDA: alloca i{{[0-9]+}}, 58 // LAMBDA: [[G_START_ADDR:%.+]] = alloca i{{[0-9]+}}, 59 // LAMBDA: alloca i{{[0-9]+}}, 60 // LAMBDA: alloca i{{[0-9]+}}, 61 // LAMBDA: alloca i{{[0-9]+}}, 62 // LAMBDA: alloca i{{[0-9]+}}, 63 // LAMBDA: alloca i{{[0-9]+}}, 64 // LAMBDA: [[G_PRIVATE_ADDR:%.+]] = alloca i{{[0-9]+}}, 65 // LAMBDA: store %{{.+}}* [[ARG]], %{{.+}}** [[ARG_REF:%.+]], 66 // LAMBDA: [[GTID_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** %{{.+}} 67 // LAMBDA: [[GTID:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[GTID_REF]] 68 // LAMBDA: call {{.+}} @__kmpc_for_static_init_4(%{{.+}}* @{{.+}}, i32 [[GTID]], i32 34, i32* [[IS_LAST_ADDR:%.+]], i32* %{{.+}}, i32* %{{.+}}, i32* %{{.+}}, i32 1, i32 1) 69 // LAMBDA: [[VAL:%.+]] = load i32, i32* [[G_START_ADDR]] 70 // LAMBDA: [[CNT:%.+]] = load i32, i32* 71 // LAMBDA: [[MUL:%.+]] = mul nsw i32 [[CNT]], 5 72 // LAMBDA: [[ADD:%.+]] = add nsw i32 [[VAL]], [[MUL]] 73 // LAMBDA: store i32 [[ADD]], i32* [[G_PRIVATE_ADDR]], 74 // LAMBDA: [[VAL:%.+]] = load i32, i32* [[G_PRIVATE_ADDR]], 75 // LAMBDA: [[ADD:%.+]] = add nsw i32 [[VAL]], 5 76 // LAMBDA: store i32 [[ADD]], i32* [[G_PRIVATE_ADDR]], 77 // LAMBDA: [[G_PRIVATE_ADDR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 0 78 // LAMBDA: store i{{[0-9]+}}* [[G_PRIVATE_ADDR]], i{{[0-9]+}}** [[G_PRIVATE_ADDR_REF]] 79 // LAMBDA: call void [[INNER_LAMBDA:@.+]](%{{.+}}* [[ARG]]) 80 // LAMBDA: call void @__kmpc_for_static_fini(%{{.+}}* @{{.+}}, i32 [[GTID]]) 81 g += 5; 82 // LAMBDA: call i32 @__kmpc_cancel_barrier(%{{.+}}* @{{.+}}, i{{[0-9]+}} [[GTID]]) 83 [&]() { 84 // LAMBDA: define {{.+}} void [[INNER_LAMBDA]](%{{.+}}* [[ARG_PTR:%.+]]) 85 // LAMBDA: store %{{.+}}* [[ARG_PTR]], %{{.+}}** [[ARG_PTR_REF:%.+]], 86 g = 2; 87 // LAMBDA: [[ARG_PTR:%.+]] = load %{{.+}}*, %{{.+}}** [[ARG_PTR_REF]] 88 // LAMBDA: [[G_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 0 89 // LAMBDA: [[G_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[G_PTR_REF]] 90 // LAMBDA: store i{{[0-9]+}} 2, i{{[0-9]+}}* [[G_REF]] 91 }(); 92 } 93 }(); 94 return 0; 95 #elif defined(BLOCKS) 96 // BLOCKS: [[G:@.+]] = global i{{[0-9]+}} 1212, 97 // BLOCKS-LABEL: @main 98 // BLOCKS: call void {{%.+}}(i8 99 ^{ 100 // BLOCKS: define{{.*}} internal{{.*}} void {{.+}}(i8* 101 // BLOCKS: call void {{.+}} @__kmpc_fork_call({{.+}}, i32 1, {{.+}}* [[OMP_REGION:@.+]] to {{.+}}, i8* %{{.+}}) 102 #pragma omp parallel 103 #pragma omp for linear(g:5) 104 for (int i = 0; i < 2; ++i) { 105 // BLOCKS: define{{.*}} internal{{.*}} void [[OMP_REGION]](i32* %{{.+}}, i32* %{{.+}}, %{{.+}}* [[ARG:%.+]]) 106 // BLOCKS: alloca i{{[0-9]+}}, 107 // BLOCKS: [[G_START_ADDR:%.+]] = alloca i{{[0-9]+}}, 108 // BLOCKS: alloca i{{[0-9]+}}, 109 // BLOCKS: alloca i{{[0-9]+}}, 110 // BLOCKS: alloca i{{[0-9]+}}, 111 // BLOCKS: alloca i{{[0-9]+}}, 112 // BLOCKS: alloca i{{[0-9]+}}, 113 // BLOCKS: [[G_PRIVATE_ADDR:%.+]] = alloca i{{[0-9]+}}, 114 // BLOCKS: store %{{.+}}* [[ARG]], %{{.+}}** [[ARG_REF:%.+]], 115 // BLOCKS: [[GTID_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** %{{.+}} 116 // BLOCKS: [[GTID:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[GTID_REF]] 117 // BLOCKS: call {{.+}} @__kmpc_for_static_init_4(%{{.+}}* @{{.+}}, i32 [[GTID]], i32 34, i32* [[IS_LAST_ADDR:%.+]], i32* %{{.+}}, i32* %{{.+}}, i32* %{{.+}}, i32 1, i32 1) 118 // BLOCKS: [[VAL:%.+]] = load i32, i32* [[G_START_ADDR]] 119 // BLOCKS: [[CNT:%.+]] = load i32, i32* 120 // BLOCKS: [[MUL:%.+]] = mul nsw i32 [[CNT]], 5 121 // BLOCKS: [[ADD:%.+]] = add nsw i32 [[VAL]], [[MUL]] 122 // BLOCKS: store i32 [[ADD]], i32* [[G_PRIVATE_ADDR]], 123 // BLOCKS: [[VAL:%.+]] = load i32, i32* [[G_PRIVATE_ADDR]], 124 // BLOCKS: [[ADD:%.+]] = add nsw i32 [[VAL]], 5 125 // BLOCKS: store i32 [[ADD]], i32* [[G_PRIVATE_ADDR]], 126 // BLOCKS-NOT: [[G]]{{[[^:word:]]}} 127 // BLOCKS: i{{[0-9]+}}* [[G_PRIVATE_ADDR]] 128 // BLOCKS-NOT: [[G]]{{[[^:word:]]}} 129 // BLOCKS: call void {{%.+}}(i8 130 // BLOCKS: call void @__kmpc_for_static_fini(%{{.+}}* @{{.+}}, i32 [[GTID]]) 131 g += 5; 132 // BLOCKS: call i32 @__kmpc_cancel_barrier(%{{.+}}* @{{.+}}, i{{[0-9]+}} [[GTID]]) 133 g = 1; 134 ^{ 135 // BLOCKS: define {{.+}} void {{@.+}}(i8* 136 g = 2; 137 // BLOCKS-NOT: [[G]]{{[[^:word:]]}} 138 // BLOCKS: store i{{[0-9]+}} 2, i{{[0-9]+}}* 139 // BLOCKS-NOT: [[G]]{{[[^:word:]]}} 140 // BLOCKS: ret 141 }(); 142 } 143 }(); 144 return 0; 145 #else 146 S<float> test; 147 float *pvar = &test.f; 148 long long lvar = 0; 149 #pragma omp parallel 150 #pragma omp for linear(pvar, lvar : 3) 151 for (int i = 0; i < 2; ++i) { 152 pvar += 3, lvar += 3; 153 } 154 return tmain<int>(); 155 #endif 156 } 157 158 // CHECK: define i{{[0-9]+}} @main() 159 // CHECK: [[TEST:%.+]] = alloca [[S_FLOAT_TY]], 160 // CHECK: call {{.*}} [[S_FLOAT_TY_DEF_CONSTR:@.+]]([[S_FLOAT_TY]]* [[TEST]]) 161 // CHECK: %{{.+}} = bitcast [[CAP_MAIN_TY]]* 162 // CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{[0-9]+}} 1, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*, [[CAP_MAIN_TY]]*)* [[MAIN_MICROTASK:@.+]] to void 163 // CHECK: = call {{.+}} [[TMAIN_INT:@.+]]() 164 // CHECK: call void [[S_FLOAT_TY_DESTR:@.+]]([[S_FLOAT_TY]]* 165 // CHECK: ret 166 167 // CHECK: define internal void [[MAIN_MICROTASK]](i{{[0-9]+}}* [[GTID_ADDR:%.+]], i{{[0-9]+}}* %{{.+}}, [[CAP_MAIN_TY]]* %{{.+}}) 168 // CHECK: alloca i{{[0-9]+}}, 169 // CHECK: [[PVAR_START:%.+]] = alloca float*, 170 // CHECK: [[LVAR_START:%.+]] = alloca i64, 171 // CHECK: alloca i{{[0-9]+}}, 172 // CHECK: alloca i{{[0-9]+}}, 173 // CHECK: alloca i{{[0-9]+}}, 174 // CHECK: alloca i{{[0-9]+}}, 175 // CHECK: alloca i{{[0-9]+}}, 176 // CHECK: [[PVAR_PRIV:%.+]] = alloca float*, 177 // CHECK: [[LVAR_PRIV:%.+]] = alloca i64, 178 // CHECK: store i{{[0-9]+}}* [[GTID_ADDR]], i{{[0-9]+}}** [[GTID_ADDR_REF:%.+]] 179 180 // Check for default initialization. 181 // CHECK: [[PVAR_PTR_REF:%.+]] = getelementptr inbounds [[CAP_MAIN_TY]], [[CAP_MAIN_TY]]* %{{.+}}, i{{[0-9]+}} 0, i{{[0-9]+}} 0 182 // CHECK: [[PVAR_REF:%.+]] = load float**, float*** [[PVAR_PTR_REF]], 183 // CHECK: [[PVAR_VAL:%.+]] = load float*, float** [[PVAR_REF]], 184 // CHECK: store float* [[PVAR_VAL]], float** [[PVAR_START]], 185 // CHECK: [[LVAR_PTR_REF:%.+]] = getelementptr inbounds [[CAP_MAIN_TY]], [[CAP_MAIN_TY]]* %{{.+}}, i{{[0-9]+}} 0, i{{[0-9]+}} 1 186 // CHECK: [[LVAR_REF:%.+]] = load i64*, i64** [[LVAR_PTR_REF]], 187 // CHECK: [[LVAR_VAL:%.+]] = load i64, i64* [[LVAR_REF]], 188 // CHECK: store i64 [[LVAR_VAL]], i64* [[LVAR_START]], 189 // CHECK: call {{.+}} @__kmpc_for_static_init_4(%{{.+}}* @{{.+}}, i32 [[GTID:%.+]], i32 34, i32* [[IS_LAST_ADDR:%.+]], i32* %{{.+}}, i32* %{{.+}}, i32* %{{.+}}, i32 1, i32 1) 190 // CHECK: [[PVAR_VAL:%.+]] = load float*, float** [[PVAR_START]], 191 // CHECK: [[CNT:%.+]] = load i32, i32* 192 // CHECK: [[MUL:%.+]] = mul nsw i32 [[CNT]], 3 193 // CHECK: [[IDX:%.+]] = sext i32 [[MUL]] to i64 194 // CHECK: [[PTR:%.+]] = getelementptr inbounds float, float* [[PVAR_VAL]], i64 [[IDX]] 195 // CHECK: store float* [[PTR]], float** [[PVAR_PRIV]], 196 // CHECK: [[LVAR_VAL:%.+]] = load i64, i64* [[LVAR_START]], 197 // CHECK: [[CNT:%.+]] = load i32, i32* 198 // CHECK: [[MUL:%.+]] = mul nsw i32 [[CNT]], 3 199 // CHECK: [[CONV:%.+]] = sext i32 [[MUL]] to i64 200 // CHECK: [[VAL:%.+]] = add nsw i64 [[LVAR_VAL]], [[CONV]] 201 // CHECK: store i64 [[VAL]], i64* [[LVAR_PRIV]], 202 // CHECK: [[PVAR_VAL:%.+]] = load float*, float** [[PVAR_PRIV]] 203 // CHECK: [[PTR:%.+]] = getelementptr inbounds float, float* [[PVAR_VAL]], i64 3 204 // CHECK: store float* [[PTR]], float** [[PVAR_PRIV]], 205 // CHECK: [[LVAR_VAL:%.+]] = load i64, i64* [[LVAR_PRIV]], 206 // CHECK: [[ADD:%.+]] = add nsw i64 [[LVAR_VAL]], 3 207 // CHECK: store i64 [[ADD]], i64* [[LVAR_PRIV]], 208 // CHECK: call void @__kmpc_for_static_fini(%{{.+}}* @{{.+}}, i32 %{{.+}}) 209 // CHECK: call i32 @__kmpc_cancel_barrier(%{{.+}}* [[IMPLICIT_BARRIER_LOC]], i{{[0-9]+}} [[GTID]]) 210 // CHECK: ret void 211 212 // CHECK: define {{.*}} i{{[0-9]+}} [[TMAIN_INT]]() 213 // CHECK: [[TEST:%.+]] = alloca [[S_INT_TY]], 214 // CHECK: call {{.*}} [[S_INT_TY_DEF_CONSTR:@.+]]([[S_INT_TY]]* [[TEST]]) 215 // CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{[0-9]+}} 1, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*, [[CAP_TMAIN_TY]]*)* [[TMAIN_MICROTASK:@.+]] to void 216 // CHECK: call void [[S_INT_TY_DESTR:@.+]]([[S_INT_TY]]* 217 // CHECK: ret 218 // 219 // CHECK: define internal void [[TMAIN_MICROTASK]](i{{[0-9]+}}* [[GTID_ADDR:%.+]], i{{[0-9]+}}* %{{.+}}, [[CAP_TMAIN_TY]]* %{{.+}}) 220 // CHECK: alloca i{{[0-9]+}}, 221 // CHECK: [[PVAR_START:%.+]] = alloca i32*, 222 // CHECK: [[LVAR_START:%.+]] = alloca i32, 223 // CHECK: alloca i{{[0-9]+}}, 224 // CHECK: alloca i{{[0-9]+}}, 225 // CHECK: alloca i{{[0-9]+}}, 226 // CHECK: alloca i{{[0-9]+}}, 227 // CHECK: alloca i{{[0-9]+}}, 228 // CHECK: [[PVAR_PRIV:%.+]] = alloca i32*, 229 // CHECK: [[LVAR_PRIV:%.+]] = alloca i32, 230 // CHECK: store i{{[0-9]+}}* [[GTID_ADDR]], i{{[0-9]+}}** [[GTID_ADDR_REF:%.+]] 231 232 // Check for default initialization. 233 // CHECK: [[PVAR_PTR_REF:%.+]] = getelementptr inbounds [[CAP_TMAIN_TY]], [[CAP_TMAIN_TY]]* %{{.+}}, i{{[0-9]+}} 0, i{{[0-9]+}} 0 234 // CHECK: [[PVAR_REF:%.+]] = load i32**, i32*** [[PVAR_PTR_REF]], 235 // CHECK: [[PVAR_VAL:%.+]] = load i32*, i32** [[PVAR_REF]], 236 // CHECK: store i32* [[PVAR_VAL]], i32** [[PVAR_START]], 237 // CHECK: [[LVAR_PTR_REF:%.+]] = getelementptr inbounds [[CAP_TMAIN_TY]], [[CAP_TMAIN_TY]]* %{{.+}}, i{{[0-9]+}} 0, i{{[0-9]+}} 1 238 // CHECK: [[LVAR_REF:%.+]] = load i32*, i32** [[LVAR_PTR_REF]], 239 // CHECK: [[LVAR_VAL:%.+]] = load i32, i32* [[LVAR_REF]], 240 // CHECK: store i32 [[LVAR_VAL]], i32* [[LVAR_START]], 241 // CHECK: call {{.+}} @__kmpc_for_static_init_4(%{{.+}}* @{{.+}}, i32 [[GTID:%.+]], i32 34, i32* [[IS_LAST_ADDR:%.+]], i32* %{{.+}}, i32* %{{.+}}, i32* %{{.+}}, i32 1, i32 1) 242 // CHECK: [[PVAR_VAL:%.+]] = load i32*, i32** [[PVAR_START]], 243 // CHECK: [[CNT:%.+]] = load i32, i32* 244 // CHECK: [[MUL:%.+]] = mul nsw i32 [[CNT]], 1 245 // CHECK: [[IDX:%.+]] = sext i32 [[MUL]] to i64 246 // CHECK: [[PTR:%.+]] = getelementptr inbounds i32, i32* [[PVAR_VAL]], i64 [[IDX]] 247 // CHECK: store i32* [[PTR]], i32** [[PVAR_PRIV]], 248 // CHECK: [[LVAR_VAL:%.+]] = load i32, i32* [[LVAR_START]], 249 // CHECK: [[CNT:%.+]] = load i32, i32* 250 // CHECK: [[MUL:%.+]] = mul nsw i32 [[CNT]], 1 251 // CHECK: [[VAL:%.+]] = add nsw i32 [[LVAR_VAL]], [[MUL]] 252 // CHECK: store i32 [[VAL]], i32* [[LVAR_PRIV]], 253 // CHECK: [[PVAR_VAL:%.+]] = load i32*, i32** [[PVAR_PRIV]] 254 // CHECK: [[PTR:%.+]] = getelementptr inbounds i32, i32* [[PVAR_VAL]], i32 1 255 // CHECK: store i32* [[PTR]], i32** [[PVAR_PRIV]], 256 // CHECK: [[LVAR_VAL:%.+]] = load i32, i32* [[LVAR_PRIV]], 257 // CHECK: [[ADD:%.+]] = add nsw i32 [[LVAR_VAL]], 1 258 // CHECK: store i32 [[ADD]], i32* [[LVAR_PRIV]], 259 // CHECK: call void @__kmpc_for_static_fini(%{{.+}}* @{{.+}}, i32 %{{.+}}) 260 // CHECK: call i32 @__kmpc_cancel_barrier(%{{.+}}* [[IMPLICIT_BARRIER_LOC]], i{{[0-9]+}} [[GTID]]) 261 // CHECK: ret void 262 #endif 263 264