make events threadsafe #3042
Merged
make events threadsafe #3042
Conversation
This file contains 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
|
thanks for the fix. Do you have a minimal example that reproduces this segfault? |
|
PR looks great to me. Do we need to include the mutex lock in the destructor too? I've done it in 0096765. I've setup a branch to include changes in destructor, mutable for mutex and an experimental test that plays with 3 threads trying to force a possible race condition: a6e8999 These changes are in |
j-rivero
reviewed
Jul 29, 2021
|
@j-rivero thanks & merged in PR |
- initial connection (insert) and cleanup (remove) locked with mutex - signal (access) checks for object not null before attempting to access (not locking because would need recursive mutex)
Signed-off-by: Jose Luis Rivero <jrivero@osrfoundation.org>
Signed-off-by: Jose Luis Rivero <jrivero@osrfoundation.org>
Signed-off-by: Jose Luis Rivero <jrivero@osrfoundation.org>
|
Umm we have a problem on Mac: Need to look into it. |
I don't want to block this PR since it improves the current state of things. I've ticketed the issue of the test problem on Mac #3063. |
j-rivero
pushed a commit
that referenced
this pull request
Aug 13, 2021
- Initial connection (insert) and cleanup (remove) locked with mutex - Signal (access) checks for object not null before attempting to access (not locking because would need recursive mutex) - Declare mutex as mutable so EventCount const method can block it - Include race condition test to play with connections inside an event (contribution by Sonia Jin from AWS, back port of #3042) Co-authored-by: Jose Luis Rivero <jrivero@osrfoundation.org>
j-rivero
added a commit
that referenced
this pull request
Aug 16, 2021
- Initial connection (insert) and cleanup (remove) locked with mutex - Signal (access) checks for object not null before attempting to access (not locking because would need recursive mutex) - Declare mutex as mutable so EventCount const method can block it - Include race condition test to play with connections inside an event (contribution by Sonia Jin from AWS, back port of #3042) Co-authored-by: Jose Luis Rivero <jrivero@osrfoundation.org> Co-authored-by: sonia <lihui815@users.noreply.github.com>
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.
Fixes Per issue #2874, race conditions can occur between
EventT::ConnectandEventT::Signal. In order to address this, add lock onConnect.However, this only keeps thread-safety between
Connect(insert) and theCleanup(delete) referenced fromSignal. There is still potential race condition between the rest ofSignal(access) andConnect(insert). Adding a mutex lock guard here however will lock up the program as there appears to be aConnectcall from a callback, and using recursive mutexes will impact performance to some extent.Instead, since the conflict occurs between connection insertion and access, add a null check before accessing to each
Signaldefinition.