[libcu++] Fix default make_shared_resource construction#9044
Merged
Conversation
Contributor
|
No actionable comments were generated in the recent review. 🎉 ℹ️ Recent review info⚙️ Run configurationConfiguration used: Path: .coderabbit.yaml Review profile: CHILL Plan: Enterprise Run ID: 📒 Files selected for processing (1)
🚧 Files skipped from review as they are similar to previous changes (1)
📝 WalkthroughSummary by CodeRabbit
suggestion: WalkthroughAdds an in-place constructor to __shared_block_ptr and updates shared_resource argument forwarding so make_shared_resource() with no arguments default-constructs T. Adds a test verifying a single construction and non-null allocate_sync. ChangesIn-place construction support for shared_resource
Assessment against linked issues
Possibly related PRs
Suggested reviewers
Comment |
This comment has been minimized.
This comment has been minimized.
Jacobfaib
approved these changes
May 16, 2026
This comment has been minimized.
This comment has been minimized.
Contributor
🥳 CI Workflow Results🟩 Finished in 2h 16m: Pass: 100%/116 | Total: 2d 13h | Max: 2h 14m | Hits: 81%/419684See results here. |
Contributor
|
Successfully created backport PR for |
miscco
reviewed
May 18, 2026
Comment on lines
+71
to
84
| //! @brief Constructs a new control block, forwarding arguments to the payload. | ||
| template <class... _Args> | ||
| _CCCL_HOST_API explicit __shared_block_ptr(::cuda::std::in_place_type_t<_Payload>, _Args&&... __args) | ||
| : __block_(new __block_t(::cuda::std::forward<_Args>(__args)...)) | ||
| {} | ||
|
|
||
| //! @brief Constructs a new control block, forwarding arguments to the payload. | ||
| _CCCL_TEMPLATE(class _Arg, class... _Rest) | ||
| _CCCL_REQUIRES((!::cuda::std::is_same_v<::cuda::std::remove_cvref_t<_Arg>, __shared_block_ptr>) ) | ||
| _CCCL_REQUIRES( | ||
| (!::cuda::std::is_same_v<::cuda::std::remove_cvref_t<_Arg>, __shared_block_ptr>) | ||
| && (!::cuda::std::is_same_v<::cuda::std::remove_cvref_t<_Arg>, ::cuda::std::in_place_type_t<_Payload>>) ) | ||
| _CCCL_HOST_API explicit __shared_block_ptr(_Arg&& __arg, _Rest&&... __rest) | ||
| : __block_(new __block_t(::cuda::std::forward<_Arg>(__arg), ::cuda::std::forward<_Rest>(__rest)...)) | ||
| {} |
Contributor
There was a problem hiding this comment.
I believe the additional constraint should not be necessary. We should probably just require everybody to use in_place
Contributor
Author
There was a problem hiding this comment.
Happy to do this in a follow-up but we need to land the backport first to get CI unblocked for RAPIDS.
GCC 7 jobs failed without the additional constraint (ambiguous with two forwarding constructors).
3 tasks
rapids-bot Bot
pushed a commit
to rapidsai/rmm
that referenced
this pull request
May 18, 2026
There is a bug in CCCL's `make_shared_resource<T>()` that doesn't work with types that have no constructor arguments. NVIDIA/cccl#9043 NVIDIA/cccl#9044 fixes it, but we need a workaround to unblock CI until that is backported (NVIDIA/cccl#9047) and rapids-cmake is updated. Authors: - Bradley Dice (https://github.com/bdice) Approvers: - David Wendt (https://github.com/davidwendt) URL: #2406
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Description
closes #9043
Adds coverage for zero-argument
cuda::mr::make_shared_resource<T>()with a default-constructible resource, and fixesshared_resource<T>(in_place_type<T>)so it constructs aTeven when no constructor arguments are provided.The fix preserves null default construction for
__shared_block_ptr<T>by adding an explicit in-place payload construction path.Tested with:
CCCL_BUILD_INFIX=opencode cmake --preset libcudacxx-cpp17 -DCMAKE_CUDA_ARCHITECTURES=native cmake --build /home/coder/cccl/build/opencode/libcudacxx-cpp17 --target libcudacxx.test.cuda.memory_resource.shared_resource -j ctest --test-dir /home/coder/cccl/build/opencode/libcudacxx-cpp17 -R '^libcudacxx\.test\.cuda\.memory_resource\.shared_resource$' --output-on-failureChecklist