fix member access to packed CUDA struct #26691
Merged
+2
−1
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Outdated suggestions cannot be applied.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.
The mosaic code uses the CUDA
CUpti_ActivityKernel9
struct, which (at least as of CUDA 12.6) has been declared with__attribute__((packed))
. This makes it UB to access a reference to a potentially unaligned field, causing GCC 13 to raise an error:Work around this by separating out the member access, which avoids the packed struct reference issue because then it's just a copy of the pointer, not a reference. We needed this to unblock the distribution of jax in conda-forge (where we're currently on GCC 13.3 & CUDA 12.6): conda-forge/jaxlib-feedstock#295