Skip to content

[flang][cuda] Fix type mismatch in atomiccas #128548

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 1 commit into from
Feb 24, 2025
Merged

[flang][cuda] Fix type mismatch in atomiccas #128548

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
11 changes: 9 additions & 2 deletions flang/lib/Optimizer/Builder/IntrinsicCall.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2728,15 +2728,22 @@ mlir::Value IntrinsicLibrary::genAtomicOr(mlir::Type resultType,
mlir::Value IntrinsicLibrary::genAtomicCas(mlir::Type resultType,
llvm::ArrayRef<mlir::Value> args) {
assert(args.size() == 3);
assert(args[1].getType() == args[2].getType());
auto successOrdering = mlir::LLVM::AtomicOrdering::acq_rel;
auto failureOrdering = mlir::LLVM::AtomicOrdering::monotonic;
auto llvmPtrTy = mlir::LLVM::LLVMPointerType::get(resultType.getContext());

mlir::Value arg1 = args[1];
mlir::Value arg2 = args[2];
if (arg1.getType() != arg2.getType()) {
// arg1 and arg2 need to have the same type in AtomicCmpXchgOp.
arg2 = builder.createConvert(loc, arg1.getType(), arg2);
}

auto address =
builder.create<mlir::UnrealizedConversionCastOp>(loc, llvmPtrTy, args[0])
.getResult(0);
auto cmpxchg = builder.create<mlir::LLVM::AtomicCmpXchgOp>(
loc, address, args[1], args[2], successOrdering, failureOrdering);
loc, address, arg1, arg2, successOrdering, failureOrdering);
return builder.create<mlir::LLVM::ExtractValueOp>(loc, cmpxchg, 1);
}

Expand Down
10 changes: 10 additions & 0 deletions flang/test/Lower/CUDA/cuda-device-proc.cuf
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -162,3 +162,13 @@ end subroutine
! CHECK: llvm.atomicrmw _xor %{{.*}}, %{{.*}} seq_cst : !llvm.ptr, i32
! CHECK: %[[ADDR:.*]] = builtin.unrealized_conversion_cast %{{.*}}#1 : !fir.ref<i32> to !llvm.ptr
! CHECK: llvm.cmpxchg %[[ADDR]], %{{.*}}, %c14{{.*}} acq_rel monotonic : !llvm.ptr, i32

attributes(device) subroutine testAtomic2()
integer(8) :: a, i, istat
istat = atomiccas(a, i, 14)
end subroutine

! CHECK-LABEL: func.func @_QPtestatomic2()
! CHECK: %[[VAL:.*]] = fir.convert %c14{{.*}} : (i32) -> i64
! CHECK: %[[ADDR:.*]] = builtin.unrealized_conversion_cast %{{.*}}#1 : !fir.ref<i64> to !llvm.ptr
! CHECK: llvm.cmpxchg %{{.*}}, %{{.*}}, %[[VAL]] acq_rel monotonic : !llvm.ptr, i64
Loading