gh-132641: Fix race in lru_cache due to inconsistent critical section… #133787
+5
−4
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Outdated suggestions cannot be applied.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.
… use.
The bounded_lru_cache code was using a critical section on the lru cache object to protect dictionary accesses in some code paths, but using the critical section on the dictionary itself to protect accesses in other code paths. This led to races since not all threads agreed on which lock they needed to be holding.
Instead, always use a critical section on the underlying dictionary, rather than the lru cache object itself.
Fixes #132641
I do not have a small Python reproducer for this problem, but a test in the JAX test suite fails under 3.14 with a TSAN warning 27 out of 50 runs without this fix and 0 out of 50 runs with it.