fix(s2n-quic-dc): synchronize dispatch queue closure #2553
Merged
Conversation
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
Mark-Simulacrum
approved these changes
Mar 17, 2025
| drop(inner); | ||
|
|
||
| // drop wakers after the lock to avoid potential mutex poisoning | ||
| wakers.dont_wake(); |
There was a problem hiding this comment.
Are we waking somewhere else? Otherwise it seems like this will hang a potential read(...) when the stream closes? Or maybe that doesn't actually get hit in practice?
(I think this is not changed in this PR, so not super relevant, but might be another deadlock risk)
There was a problem hiding this comment.
This function is only called from the receivers themselves so in almost all cases they shouldn't be set. This is mostly for if one of the receivers drops prematurely before getting notified of a new packet.
Mark-Simulacrum
pushed a commit
to Mark-Simulacrum/s2n-quic
that referenced
this pull request
Mar 25, 2025
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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 of changes:
After doing some testing, the current dispatch queue closing logic has a potential race condition where the queue can end up being freed twice. This change fixes that by locking both receivers in the same order (to avoid a deadlock) and observing the other side's status before deciding to free the queue or not.
Call-outs:
I did a little more clean up around tracing events by creating actual probes instead. I've also tried to avoid calling any of the probes while holding a mutex.
Testing:
I added a stress test that was able to reproduce the original issue and show that the changes fixed.
By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license.