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 volatile int &g1 = g; 23 float f; 24 char cnt; 25 26 // CHECK: [[S_FLOAT_TY:%.+]] = type { float } 27 // CHECK: [[S_INT_TY:%.+]] = type { i32 } 28 // CHECK-DAG: [[IMPLICIT_BARRIER_LOC:@.+]] = private unnamed_addr constant %{{.+}} { i32 0, i32 66, i32 0, i32 0, i8* 29 // CHECK-DAG: [[F:@.+]] = global float 0.0 30 // CHECK-DAG: [[CNT:@.+]] = global i8 0 31 template <typename T> 32 T tmain() { 33 S<T> test; 34 T *pvar = &test.f; 35 T &lvar = test.f; 36 #pragma omp parallel 37 #pragma omp for linear(pvar, lvar) 38 for (int i = 0; i < 2; ++i) { 39 ++pvar, ++lvar; 40 } 41 return T(); 42 } 43 44 int main() { 45 #ifdef LAMBDA 46 // LAMBDA: [[G:@.+]] = global i{{[0-9]+}} 1212, 47 // LAMBDA-LABEL: @main 48 // LAMBDA: call void [[OUTER_LAMBDA:@.+]]( 49 [&]() { 50 // LAMBDA: define{{.*}} internal{{.*}} void [[OUTER_LAMBDA]]( 51 // LAMBDA: call void {{.+}} @__kmpc_fork_call({{.+}}, i32 0, {{.+}}* [[OMP_REGION:@.+]] to {{.+}}) 52 #pragma omp parallel 53 #pragma omp for linear(g, g1:5) 54 for (int i = 0; i < 2; ++i) { 55 // LAMBDA: define{{.*}} internal{{.*}} void [[OMP_REGION]](i32* noalias %{{.+}}, i32* noalias %{{.+}}) 56 // LAMBDA: alloca i{{[0-9]+}}, 57 // LAMBDA: [[G_START_ADDR:%.+]] = alloca i{{[0-9]+}}, 58 // LAMBDA: 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: [[GTID_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** %{{.+}} 66 // LAMBDA: [[GTID:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[GTID_REF]] 67 // LAMBDA: call {{.+}} @__kmpc_for_static_init_4(%{{.+}}* @{{.+}}, i32 [[GTID]], i32 34, i32* [[IS_LAST_ADDR:%.+]], i32* %{{.+}}, i32* %{{.+}}, i32* %{{.+}}, i32 1, i32 1) 68 // LAMBDA: [[VAL:%.+]] = load i32, i32* [[G_START_ADDR]] 69 // LAMBDA: [[CNT:%.+]] = load i32, i32* 70 // LAMBDA: [[MUL:%.+]] = mul nsw i32 [[CNT]], 5 71 // LAMBDA: [[ADD:%.+]] = add nsw i32 [[VAL]], [[MUL]] 72 // LAMBDA: store i32 [[ADD]], i32* [[G_PRIVATE_ADDR]], 73 // LAMBDA: [[VAL:%.+]] = load i32, i32* [[G_PRIVATE_ADDR]], 74 // LAMBDA: [[ADD:%.+]] = add nsw i32 [[VAL]], 5 75 // LAMBDA: store i32 [[ADD]], i32* [[G_PRIVATE_ADDR]], 76 // LAMBDA: [[G_PRIVATE_ADDR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG:%.+]], i{{[0-9]+}} 0, i{{[0-9]+}} 0 77 // LAMBDA: store i{{[0-9]+}}* [[G_PRIVATE_ADDR]], i{{[0-9]+}}** [[G_PRIVATE_ADDR_REF]] 78 // LAMBDA: call void [[INNER_LAMBDA:@.+]](%{{.+}}* [[ARG]]) 79 // LAMBDA: call void @__kmpc_for_static_fini(%{{.+}}* @{{.+}}, i32 [[GTID]]) 80 g += 5; 81 g1 += 5; 82 // LAMBDA: call void @__kmpc_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 g1 = 2; 88 // LAMBDA: [[ARG_PTR:%.+]] = load %{{.+}}*, %{{.+}}** [[ARG_PTR_REF]] 89 // LAMBDA: [[G_PTR_REF:%.+]] = getelementptr inbounds %{{.+}}, %{{.+}}* [[ARG_PTR]], i{{[0-9]+}} 0, i{{[0-9]+}} 0 90 // LAMBDA: [[G_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[G_PTR_REF]] 91 // LAMBDA: store i{{[0-9]+}} 2, i{{[0-9]+}}* [[G_REF]] 92 }(); 93 } 94 }(); 95 return 0; 96 #elif defined(BLOCKS) 97 // BLOCKS: [[G:@.+]] = global i{{[0-9]+}} 1212, 98 // BLOCKS-LABEL: @main 99 // BLOCKS: call void {{%.+}}(i8 100 ^{ 101 // BLOCKS: define{{.*}} internal{{.*}} void {{.+}}(i8* 102 // BLOCKS: call void {{.+}} @__kmpc_fork_call({{.+}}, i32 0, {{.+}}* [[OMP_REGION:@.+]] to {{.+}}) 103 #pragma omp parallel 104 #pragma omp for linear(g, g1:5) 105 for (int i = 0; i < 2; ++i) { 106 // BLOCKS: define{{.*}} internal{{.*}} void [[OMP_REGION]](i32* noalias %{{.+}}, i32* noalias %{{.+}}) 107 // BLOCKS: alloca i{{[0-9]+}}, 108 // BLOCKS: [[G_START_ADDR:%.+]] = 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: alloca i{{[0-9]+}}, 114 // BLOCKS: alloca i{{[0-9]+}}, 115 // BLOCKS: [[G_PRIVATE_ADDR:%.+]] = alloca i{{[0-9]+}}, 116 // BLOCKS: [[GTID_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** %{{.+}} 117 // BLOCKS: [[GTID:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[GTID_REF]] 118 // BLOCKS: call {{.+}} @__kmpc_for_static_init_4(%{{.+}}* @{{.+}}, i32 [[GTID]], i32 34, i32* [[IS_LAST_ADDR:%.+]], i32* %{{.+}}, i32* %{{.+}}, i32* %{{.+}}, i32 1, i32 1) 119 // BLOCKS: [[VAL:%.+]] = load i32, i32* [[G_START_ADDR]] 120 // BLOCKS: [[CNT:%.+]] = load i32, i32* 121 // BLOCKS: [[MUL:%.+]] = mul nsw i32 [[CNT]], 5 122 // BLOCKS: [[ADD:%.+]] = add nsw i32 [[VAL]], [[MUL]] 123 // BLOCKS: store i32 [[ADD]], i32* [[G_PRIVATE_ADDR]], 124 // BLOCKS: [[VAL:%.+]] = load i32, i32* [[G_PRIVATE_ADDR]], 125 // BLOCKS: [[ADD:%.+]] = add nsw i32 [[VAL]], 5 126 // BLOCKS: store i32 [[ADD]], i32* [[G_PRIVATE_ADDR]], 127 // BLOCKS-NOT: [[G]]{{[[^:word:]]}} 128 // BLOCKS: i{{[0-9]+}}* [[G_PRIVATE_ADDR]] 129 // BLOCKS-NOT: [[G]]{{[[^:word:]]}} 130 // BLOCKS: call void {{%.+}}(i8 131 // BLOCKS: call void @__kmpc_for_static_fini(%{{.+}}* @{{.+}}, i32 [[GTID]]) 132 g += 5; 133 g1 += 5; 134 // BLOCKS: call void @__kmpc_barrier(%{{.+}}* @{{.+}}, i{{[0-9]+}} [[GTID]]) 135 g = 1; 136 g1 = 5; 137 ^{ 138 // BLOCKS: define {{.+}} void {{@.+}}(i8* 139 g = 2; 140 g1 = 2; 141 // BLOCKS-NOT: [[G]]{{[[^:word:]]}} 142 // BLOCKS: store i{{[0-9]+}} 2, i{{[0-9]+}}* 143 // BLOCKS-NOT: [[G]]{{[[^:word:]]}} 144 // BLOCKS: ret 145 }(); 146 } 147 }(); 148 return 0; 149 #else 150 S<float> test; 151 float *pvar = &test.f; 152 long long lvar = 0; 153 #pragma omp parallel 154 #pragma omp for linear(pvar, lvar : 3) 155 for (int i = 0; i < 2; ++i) { 156 pvar += 3, lvar += 3; 157 } 158 return tmain<int>(); 159 #endif 160 } 161 162 // CHECK: define i{{[0-9]+}} @main() 163 // CHECK: [[TEST:%.+]] = alloca [[S_FLOAT_TY]], 164 // CHECK: call {{.*}} [[S_FLOAT_TY_DEF_CONSTR:@.+]]([[S_FLOAT_TY]]* [[TEST]]) 165 // CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{[0-9]+}} 2, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*, float**, i64*)* [[MAIN_MICROTASK:@.+]] to void 166 // CHECK: = call {{.+}} [[TMAIN_INT:@.+]]() 167 // CHECK: call void [[S_FLOAT_TY_DESTR:@.+]]([[S_FLOAT_TY]]* 168 // CHECK: ret 169 170 // CHECK: define internal void [[MAIN_MICROTASK]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}}, float** dereferenceable(8) %{{.+}}, i64* dereferenceable(8) %{{.+}}) 171 // CHECK: alloca i{{[0-9]+}}, 172 // CHECK: [[PVAR_START:%.+]] = alloca float*, 173 // CHECK: [[LVAR_START:%.+]] = alloca i64, 174 // CHECK: alloca i{{[0-9]+}}, 175 // CHECK: alloca i{{[0-9]+}}, 176 // CHECK: alloca i{{[0-9]+}}, 177 // CHECK: alloca i{{[0-9]+}}, 178 // CHECK: alloca i{{[0-9]+}}, 179 // CHECK: [[PVAR_PRIV:%.+]] = alloca float*, 180 // CHECK: [[LVAR_PRIV:%.+]] = alloca i64, 181 // CHECK: store i{{[0-9]+}}* [[GTID_ADDR]], i{{[0-9]+}}** [[GTID_ADDR_REF:%.+]] 182 183 // Check for default initialization. 184 // CHECK: [[PVAR_REF:%.+]] = load float**, float*** % 185 // CHECK: [[LVAR_REF:%.+]] = load i64*, i64** % 186 // CHECK: [[PVAR_VAL:%.+]] = load float*, float** [[PVAR_REF]], 187 // CHECK: store float* [[PVAR_VAL]], float** [[PVAR_START]], 188 // CHECK: [[LVAR_VAL:%.+]] = load i64, i64* [[LVAR_REF]], 189 // CHECK: store i64 [[LVAR_VAL]], i64* [[LVAR_START]], 190 // CHECK: call {{.+}} @__kmpc_for_static_init_4(%{{.+}}* @{{.+}}, i32 [[GTID:%.+]], i32 34, i32* [[IS_LAST_ADDR:%.+]], i32* %{{.+}}, i32* %{{.+}}, i32* %{{.+}}, i32 1, i32 1) 191 // CHECK: [[PVAR_VAL:%.+]] = load float*, float** [[PVAR_START]], 192 // CHECK: [[CNT:%.+]] = load i32, i32* 193 // CHECK: [[MUL:%.+]] = mul nsw i32 [[CNT]], 3 194 // CHECK: [[IDX:%.+]] = sext i32 [[MUL]] to i64 195 // CHECK: [[PTR:%.+]] = getelementptr inbounds float, float* [[PVAR_VAL]], i64 [[IDX]] 196 // CHECK: store float* [[PTR]], float** [[PVAR_PRIV]], 197 // CHECK: [[LVAR_VAL:%.+]] = load i64, i64* [[LVAR_START]], 198 // CHECK: [[CNT:%.+]] = load i32, i32* 199 // CHECK: [[MUL:%.+]] = mul nsw i32 [[CNT]], 3 200 // CHECK: [[CONV:%.+]] = sext i32 [[MUL]] to i64 201 // CHECK: [[VAL:%.+]] = add nsw i64 [[LVAR_VAL]], [[CONV]] 202 // CHECK: store i64 [[VAL]], i64* [[LVAR_PRIV]], 203 // CHECK: [[PVAR_VAL:%.+]] = load float*, float** [[PVAR_PRIV]] 204 // CHECK: [[PTR:%.+]] = getelementptr inbounds float, float* [[PVAR_VAL]], i64 3 205 // CHECK: store float* [[PTR]], float** [[PVAR_PRIV]], 206 // CHECK: [[LVAR_VAL:%.+]] = load i64, i64* [[LVAR_PRIV]], 207 // CHECK: [[ADD:%.+]] = add nsw i64 [[LVAR_VAL]], 3 208 // CHECK: store i64 [[ADD]], i64* [[LVAR_PRIV]], 209 // CHECK: call void @__kmpc_for_static_fini(%{{.+}}* @{{.+}}, i32 %{{.+}}) 210 // CHECK: call void @__kmpc_barrier(%{{.+}}* [[IMPLICIT_BARRIER_LOC]], i{{[0-9]+}} [[GTID]]) 211 // CHECK: ret void 212 213 // CHECK: define {{.*}} i{{[0-9]+}} [[TMAIN_INT]]() 214 // CHECK: [[TEST:%.+]] = alloca [[S_INT_TY]], 215 // CHECK: call {{.*}} [[S_INT_TY_DEF_CONSTR:@.+]]([[S_INT_TY]]* [[TEST]]) 216 // CHECK: call void (%{{.+}}*, i{{[0-9]+}}, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)*, ...) @__kmpc_fork_call(%{{.+}}* @{{.+}}, i{{[0-9]+}} 2, void (i{{[0-9]+}}*, i{{[0-9]+}}*, ...)* bitcast (void (i{{[0-9]+}}*, i{{[0-9]+}}*, i32**, i32*)* [[TMAIN_MICROTASK:@.+]] to void 217 // CHECK: call void [[S_INT_TY_DESTR:@.+]]([[S_INT_TY]]* 218 // CHECK: ret 219 // 220 // CHECK: define internal void [[TMAIN_MICROTASK]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}}, i32** dereferenceable(8) %{{.+}}, i32* dereferenceable(4) %{{.+}}) 221 // CHECK: alloca i{{[0-9]+}}, 222 // CHECK: [[PVAR_START:%.+]] = alloca i32*, 223 // CHECK: [[LVAR_START:%.+]] = alloca i32, 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: alloca i{{[0-9]+}}, 229 // CHECK: [[PVAR_PRIV:%.+]] = alloca i32*, 230 // CHECK: [[LVAR_PRIV:%.+]] = alloca i32, 231 // CHECK: [[LVAR_PRIV_REF:%.+]] = alloca i32*, 232 // CHECK: store i{{[0-9]+}}* [[GTID_ADDR]], i{{[0-9]+}}** [[GTID_ADDR_REF:%.+]] 233 234 // Check for default initialization. 235 // CHECK: [[PVAR_REF:%.+]] = load i32**, i32*** % 236 // CHECK: [[PVAR_VAL:%.+]] = load i32*, i32** [[PVAR_REF]], 237 // CHECK: store i32* [[PVAR_VAL]], i32** [[PVAR_START]], 238 // CHECK: [[LVAR_REF:%.+]] = load i32*, i32** % 239 // CHECK: [[LVAR_VAL:%.+]] = load i32, i32* [[LVAR_REF]], 240 // CHECK: store i32 [[LVAR_VAL]], i32* [[LVAR_START]], 241 // CHECK: store i32* [[LVAR_PRIV]], i32** [[LVAR_PRIV_REF]], 242 243 // CHECK: call {{.+}} @__kmpc_for_static_init_4(%{{.+}}* @{{.+}}, i32 [[GTID:%.+]], i32 34, i32* [[IS_LAST_ADDR:%.+]], i32* %{{.+}}, i32* %{{.+}}, i32* %{{.+}}, i32 1, i32 1) 244 // CHECK: [[PVAR_VAL:%.+]] = load i32*, i32** [[PVAR_START]], 245 // CHECK: [[CNT:%.+]] = load i32, i32* 246 // CHECK: [[MUL:%.+]] = mul nsw i32 [[CNT]], 1 247 // CHECK: [[IDX:%.+]] = sext i32 [[MUL]] to i64 248 // CHECK: [[PTR:%.+]] = getelementptr inbounds i32, i32* [[PVAR_VAL]], i64 [[IDX]] 249 // CHECK: store i32* [[PTR]], i32** [[PVAR_PRIV]], 250 // CHECK: [[LVAR_VAL:%.+]] = load i32, i32* [[LVAR_START]], 251 // CHECK: [[CNT:%.+]] = load i32, i32* 252 // CHECK: [[MUL:%.+]] = mul nsw i32 [[CNT]], 1 253 // CHECK: [[VAL:%.+]] = add nsw i32 [[LVAR_VAL]], [[MUL]] 254 // CHECK: store i32 [[VAL]], i32* [[LVAR_PRIV]], 255 // CHECK: [[PVAR_VAL:%.+]] = load i32*, i32** [[PVAR_PRIV]] 256 // CHECK: [[PTR:%.+]] = getelementptr inbounds i32, i32* [[PVAR_VAL]], i32 1 257 // CHECK: store i32* [[PTR]], i32** [[PVAR_PRIV]], 258 // CHECK: [[LVAR_PRIV:%.+]] = load i32*, i32** [[LVAR_PRIV_REF]], 259 // CHECK: [[LVAR_VAL:%.+]] = load i32, i32* [[LVAR_PRIV]], 260 // CHECK: [[ADD:%.+]] = add nsw i32 [[LVAR_VAL]], 1 261 // CHECK: store i32 [[ADD]], i32* [[LVAR_PRIV]], 262 // CHECK: call void @__kmpc_for_static_fini(%{{.+}}* @{{.+}}, i32 %{{.+}}) 263 // CHECK: call void @__kmpc_barrier(%{{.+}}* [[IMPLICIT_BARRIER_LOC]], i{{[0-9]+}} [[GTID]]) 264 // CHECK: ret void 265 #endif 266 267
