Added mapping for pointers captured in Lambdas in OpenMP target regions,

by David Truby.

Summary:
This adds a zero length array section mapping for each pointer captured by a lambda that is used in a target region, as per section 2.19.7.1 of the OpenMP 5 specification.

Reviewers: ABataev

Reviewed By: ABataev

Subscribers: guansong, jdoerfert, cfe-commits

Tags: #clang

Differential Revision: https://reviews.llvm.org/D64558

llvm-svn: 365777
This commit is contained in:
Alexey Bataev 2019-07-11 13:54:26 +00:00
parent 605e772abc
commit af6725baa4
2 changed files with 144 additions and 9 deletions

View File

@ -8093,20 +8093,30 @@ public:
OMP_MAP_MEMBER_OF | OMP_MAP_IMPLICIT);
}
for (const LambdaCapture &LC : RD->captures()) {
if (LC.getCaptureKind() != LCK_ByRef)
if (!LC.capturesVariable())
continue;
const VarDecl *VD = LC.getCapturedVar();
if (LC.getCaptureKind() != LCK_ByRef && !VD->getType()->isPointerType())
continue;
auto It = Captures.find(VD);
assert(It != Captures.end() && "Found lambda capture without field.");
LValue VarLVal = CGF.EmitLValueForFieldInitialization(VDLVal, It->second);
LValue VarLValVal = CGF.EmitLValueForField(VDLVal, It->second);
LambdaPointers.try_emplace(VarLVal.getPointer(), VDLVal.getPointer());
BasePointers.push_back(VarLVal.getPointer());
Pointers.push_back(VarLValVal.getPointer());
Sizes.push_back(CGF.Builder.CreateIntCast(
CGF.getTypeSize(
VD->getType().getCanonicalType().getNonReferenceType()),
CGF.Int64Ty, /*isSigned=*/true));
if (LC.getCaptureKind() == LCK_ByRef) {
LValue VarLValVal = CGF.EmitLValueForField(VDLVal, It->second);
LambdaPointers.try_emplace(VarLVal.getPointer(), VDLVal.getPointer());
BasePointers.push_back(VarLVal.getPointer());
Pointers.push_back(VarLValVal.getPointer());
Sizes.push_back(CGF.Builder.CreateIntCast(
CGF.getTypeSize(
VD->getType().getCanonicalType().getNonReferenceType()),
CGF.Int64Ty, /*isSigned=*/true));
} else {
RValue VarRVal = CGF.EmitLoadOfLValue(VarLVal, RD->getLocation());
LambdaPointers.try_emplace(VarLVal.getPointer(), VDLVal.getPointer());
BasePointers.push_back(VarLVal.getPointer());
Pointers.push_back(VarRVal.getScalarVal());
Sizes.push_back(llvm::ConstantInt::get(CGF.Int64Ty, 0));
}
Types.push_back(OMP_MAP_PTR_AND_OBJ | OMP_MAP_LITERAL |
OMP_MAP_MEMBER_OF | OMP_MAP_IMPLICIT);
}

View File

@ -0,0 +1,125 @@
// Test host codegen only.
// RUN: %clang_cc1 -verify -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -emit-pch -o %t %s
// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple powerpc64le-unknown-unknown -fopenmp-targets=powerpc64le-ibm-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-64
// RUN: %clang_cc1 -verify -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-llvm %s -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -emit-pch -o %t %s
// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple i386-unknown-unknown -fopenmp-targets=i386-pc-linux-gnu -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s --check-prefix CHECK --check-prefix CHECK-32
// expected-no-diagnostics
#ifndef HEADER
#define HEADER
// CHECK: [[ANON_T:%.+]] = type { i32*, i32* }
// CHECK-DAG: [[SIZES_TEMPLATE:@.+]] = private {{.+}} constant [5 x i[[PTRSZ:32|64]]] [i{{32|64}} 4, i{{32|64}} 4, i{{32|64}} {{8|16}}, i{{32|64}} 0, i{{32|64}} 0]
// CHECK-DAG: [[TYPES_TEMPLATE:@.+]] = private {{.+}} constant [5 x i64] [i64 800, i64 800, i64 673, i64 844424930132752, i64 844424930132752]
// CHECK-DAG: [[SIZES:@.+]] = private {{.+}} constant [3 x i[[PTRSZ:32|64]]] [i{{32|64}} {{8|16}}, i{{32|64}} 0, i{{32|64}} 0]
// CHECK-DAG: [[TYPES:@.+]] = private {{.+}} constant [3 x i64] [i64 673, i64 281474976711440, i64 281474976711440]
template <typename F>
void omp_loop(int start, int end, F body) {
#pragma omp target teams distribute parallel for
for (int i = start; i < end; ++i) {
body(i);
}
}
// CHECK: define {{.*}}[[MAIN:@.+]](
int main()
{
int* p = new int[100];
int* q = new int[100];
auto body = [=](int i){
p[i] = q[i];
};
#pragma omp target teams distribute parallel for
for (int i = 0; i < 100; ++i) {
body(i);
}
// CHECK: [[BASE_PTRS:%.+]] = alloca [3 x i8*]{{.+}}
// CHECK: [[PTRS:%.+]] = alloca [3 x i8*]{{.+}}
// First gep of pointers inside lambdas to store the values across function call need to be ignored
// CHECK: {{%.+}} = getelementptr inbounds [[ANON_T]], [[ANON_T]]* %{{.+}}, i{{.+}} 0, i{{.+}} 0
// CHECK: {{%.+}} = getelementptr inbounds [[ANON_T]], [[ANON_T]]* %{{.+}}, i{{.+}} 0, i{{.+}} 1
// access of pointers inside lambdas
// CHECK: [[BASE_PTR1:%.+]] = getelementptr inbounds [[ANON_T]], [[ANON_T]]* %{{.+}}, i{{.+}} 0, i{{.+}} 0
// CHECK: [[PTR1:%.+]] = load i32*, i32** [[BASE_PTR1]]
// CHECK: [[BASE_PTR2:%.+]] = getelementptr inbounds [[ANON_T]], [[ANON_T]]* %{{.+}}, i{{.+}} 0, i{{.+}} 1
// CHECK: [[PTR2:%.+]] = load i32*, i32** [[BASE_PTR2]]
// storage of pointers in baseptrs and ptrs arrays
// CHECK: [[LOC_LAMBDA:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[BASE_PTRS]], i{{.+}} 0, i{{.+}} 0
// CHECK: [[CAST_LAMBDA:%.+]] = bitcast i8** [[LOC_LAMBDA]] to [[ANON_T]]**
// CHECK: store [[ANON_T]]* %{{.+}}, [[ANON_T]]** [[CAST_LAMBDA]]{{.+}}
// CHECK: [[LOC_LAMBDA:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[PTRS]], i{{.+}} 0, i{{.+}} 0
// CHECK: [[CAST_LAMBDA:%.+]] = bitcast i8** [[LOC_LAMBDA]] to [[ANON_T]]**
// CHECK: store [[ANON_T]]* %{{.+}}, [[ANON_T]]** [[CAST_LAMBDA]]{{.+}}
// CHECK: [[LOC_PTR1:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[BASE_PTRS]], i{{.+}} 0, i{{.+}} 1
// CHECK: [[CAST_PTR1:%.+]] = bitcast i8** [[LOC_PTR1]] to i32***
// CHECK: store i32** [[BASE_PTR1]], i32*** [[CAST_PTR1]]{{.+}}
// CHECK: [[LOC_PTR1:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[PTRS]], i{{.+}} 0, i{{.+}} 1
// CHECK: [[CAST_PTR1:%.+]] = bitcast i8** [[LOC_PTR1]] to i32**
// CHECK: store i32* [[PTR1]], i32** [[CAST_PTR1]]{{.+}}
// CHECK: [[LOC_PTR2:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[BASE_PTRS]], i{{.+}} 0, i{{.+}} 2
// CHECK: [[CAST_PTR2:%.+]] = bitcast i8** [[LOC_PTR2]] to i32***
// CHECK: store i32** [[BASE_PTR2]], i32*** [[CAST_PTR2]]{{.+}}
// CHECK: [[LOC_PTR2:%.+]] = getelementptr inbounds [3 x i8*], [3 x i8*]* [[PTRS]], i{{.+}} 0, i{{.+}} 2
// CHECK: [[CAST_PTR2:%.+]] = bitcast i8** [[LOC_PTR2]] to i32**
// CHECK: store i32* [[PTR2]], i32** [[CAST_PTR2]]{{.+}}
// actual target invocation
// CHECK: [[BASES_GEP:%.+]] = getelementptr {{.+}} [3 x {{.+}}*], [3 x {{.+}}*]* [[BASE_PTRS]], {{.+}} 0, {{.+}} 0
// CHECK: [[PTRS_GEP:%.+]] = getelementptr {{.+}} [3 x {{.+}}*], [3 x {{.+}}*]* [[PTRS]], {{.+}} 0, {{.+}} 0
// CHECK: {{%.+}} = call{{.+}} @__tgt_target_teams({{.+}}, {{.+}}, {{.+}}, i8** [[BASES_GEP]], i8** [[PTRS_GEP]], i[[PTRSZ]]* getelementptr inbounds ([3 x i{{.+}}], [3 x i{{.+}}]* [[SIZES]], i{{.+}} 0, i{{.+}} 0), i64* getelementptr inbounds ([3 x i64], [3 x i64]* [[TYPES]], i{{.+}} 0, i{{.+}} 0), {{.+}}, {{.+}})
omp_loop(0,100,body);
}
// CHECK: [[BASE_PTRS:%.+]] = alloca [5 x i8*]{{.+}}
// CHECK: [[PTRS:%.+]] = alloca [5 x i8*]{{.+}}
// access of pointers inside lambdas
// CHECK: [[BASE_PTR1:%.+]] = getelementptr inbounds [[ANON_T]], [[ANON_T]]* %{{.+}}, i{{.+}} 0, i{{.+}} 0
// CHECK: [[PTR1:%.+]] = load i32*, i32** [[BASE_PTR1]]
// CHECK: [[BASE_PTR2:%.+]] = getelementptr inbounds [[ANON_T]], [[ANON_T]]* %{{.+}}, i{{.+}} 0, i{{.+}} 1
// CHECK: [[PTR2:%.+]] = load i32*, i32** [[BASE_PTR2]]
// storage of pointers in baseptrs and ptrs arrays
// CHECK: [[LOC_LAMBDA:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[BASE_PTRS]], i{{.+}} 0, i{{.+}} 2
// CHECK: [[CAST_LAMBDA:%.+]] = bitcast i8** [[LOC_LAMBDA]] to [[ANON_T]]**
// CHECK: store [[ANON_T]]* %{{.+}}, [[ANON_T]]** [[CAST_LAMBDA]]{{.+}}
// CHECK: [[LOC_LAMBDA:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[PTRS]], i{{.+}} 0, i{{.+}} 2
// CHECK: [[CAST_LAMBDA:%.+]] = bitcast i8** [[LOC_LAMBDA]] to [[ANON_T]]**
// CHECK: store [[ANON_T]]* %{{.+}}, [[ANON_T]]** [[CAST_LAMBDA]]{{.+}}
// CHECK: [[LOC_PTR1:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[BASE_PTRS]], i{{.+}} 0, i{{.+}} 3
// CHECK: [[CAST_PTR1:%.+]] = bitcast i8** [[LOC_PTR1]] to i32***
// CHECK: store i32** [[BASE_PTR1]], i32*** [[CAST_PTR1]]{{.+}}
// CHECK: [[LOC_PTR1:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[PTRS]], i{{.+}} 0, i{{.+}} 3
// CHECK: [[CAST_PTR1:%.+]] = bitcast i8** [[LOC_PTR1]] to i32**
// CHECK: store i32* [[PTR1]], i32** [[CAST_PTR1]]{{.+}}
// CHECK: [[LOC_PTR2:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[BASE_PTRS]], i{{.+}} 0, i{{.+}} 4
// CHECK: [[CAST_PTR2:%.+]] = bitcast i8** [[LOC_PTR2]] to i32***
// CHECK: store i32** [[BASE_PTR2]], i32*** [[CAST_PTR2]]{{.+}}
// CHECK: [[LOC_PTR2:%.+]] = getelementptr inbounds [5 x i8*], [5 x i8*]* [[PTRS]], i{{.+}} 0, i{{.+}} 4
// CHECK: [[CAST_PTR2:%.+]] = bitcast i8** [[LOC_PTR2]] to i32**
// CHECK: store i32* [[PTR2]], i32** [[CAST_PTR2]]{{.+}}
// actual target invocation
// CHECK: [[BASES_GEP:%.+]] = getelementptr {{.+}} [5 x {{.+}}*], [5 x {{.+}}*]* [[BASE_PTRS]], {{.+}} 0, {{.+}} 0
// CHECK: [[PTRS_GEP:%.+]] = getelementptr {{.+}} [5 x {{.+}}*], [5 x {{.+}}*]* [[PTRS]], {{.+}} 0, {{.+}} 0
// CHECK: {{%.+}} = call{{.+}} @__tgt_target_teams({{.+}}, {{.+}}, {{.+}}, i8** [[BASES_GEP]], i8** [[PTRS_GEP]], i[[PTRSZ]]* getelementptr inbounds ([5 x i{{.+}}], [5 x i{{.+}}]* [[SIZES_TEMPLATE]], i{{.+}} 0, i{{.+}} 0), i64* getelementptr inbounds ([5 x i64], [5 x i64]* [[TYPES_TEMPLATE]], i{{.+}} 0, i{{.+}} 0), {{.+}}, {{.+}})
#endif