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