BUG: Prevent deadlock due to downstream importing NumPy in dlopen (#31303)#31335
Merged
Conversation
…mpy#31303) Downstream should probably never import numpy (i.e. activate the C-API) during dlopen (rather than the Python module init function). But if it does so, this check introduced a deadlock regression because dlopen has a recursive lock, but this breaks if the matmul is called and causes threads to spawn that also need to dlopen. So, as a hot-fix, just only run the check on MacOS, since we are not currently aware of reports on other system (possible or not), although this did happen with certain OpenBLAS versions (and not just accelerate). (I cannot fully rule out that there won't be regressions on some linux setups, but they were not yet reported. -- The "limit to 1 thread" solution is an alternative as well.)
Member
Author
|
Test failures are unrelated |
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
2 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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Backport of #31303.
Downstream should probably never import numpy (i.e. activate the C-API) during dlopen (rather than the Python module init function).
But if it does so, this check introduced a deadlock regression because dlopen has a recursive lock, but this breaks if the matmul is called and causes threads to spawn that also need to dlopen.
So, as a hot-fix, just only run the check on MacOS, since we are not currently aware of reports on other system (possible or not), although this did happen with certain OpenBLAS versions (and not just accelerate).
(I cannot fully rule out that there won't be regressions on some linux setups, but they were not yet reported. -- The "limit to 1 thread" solution is an alternative as well.)
EDIT: addresses gh-31284. Should be backported for 2.4.5