Don't hold the channel lock across c-ares calls - #325
Open
dimbleby wants to merge 1 commit into
Open
Conversation
c-ares runs callbacks with its own channel lock held, and a callback may call back into the Channel from another thread. So holding the Python lock across a c-ares call can deadlock: a thread in wait() held the lock until the query queue drained, but draining it needed a callback which wanted the same lock. Hold the lock only while reading self._channel, and count the calls in flight instead, so that the shutdown thread can wait for the channel to fall idle before destroying it. Calls and waits are counted separately because they have to be drained either side of the cancel: a call may still submit a query, whereas a wait only returns once the cancel has happened. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
I do not love how complicated it is getting to avoid use-after-free and deadlock at shutdown etc; but I see no obvious simplification.
c-ares runs callbacks with its own channel lock held, and a callback may call back into the Channel from another thread. So holding the Python lock across a c-ares call can deadlock: a thread in wait() held the lock until the query queue drained, but draining it needed a callback which wanted the same lock.
Hold the lock only while reading self._channel, and count the calls in flight instead, so that the shutdown thread can wait for the channel to fall idle before destroying it. Calls and waits are counted separately because they have to be drained either side of the cancel: a call may still submit a query, whereas a wait only returns once the cancel has happened.