semaphore.h: handle spurious wakeups in TimedWait() on Linux - #1021
Merged
dconeybe merged 4 commits intoJul 11, 2022
Merged
Conversation
Copy link
Copy Markdown
Contributor
Sorry, something went wrong.
Copy link
Copy Markdown
❌ Integration test FAILEDRequested by @dconeybe on commit 26e918b
Add flaky tests to go/fpl-cpp-flake-tracker |
Sorry, something went wrong.
DellaBitta
reviewed
Jul 9, 2022
Comment thread
app/src/semaphore.h
DellaBitta
reviewed
Jul 9, 2022
Comment thread
app/src/semaphore.h
DellaBitta
requested changes
Jul 11, 2022
Comment thread
app/src/semaphore.h
DellaBitta
approved these changes
Jul 11, 2022
Comment thread
app/src/semaphore.h
dconeybe
marked this pull request as ready for review
July 11, 2022 14:40
jonsimantov
reviewed
Jul 11, 2022
Comment thread
app/src/semaphore.h
| // Return failure, since the timeout expired. | ||
| return false; | ||
| case EINVAL: | ||
| assert("sem_timedwait() failed with EINVAL" == 0); |
Copy link
Copy Markdown
Contributor
There was a problem hiding this comment.
Just making sure, you want this to NOT actually assert in release builds, yes? (the default assert behavior?)
Sorry, something went wrong.
Comment thread
app/src/semaphore.h
| #else // not windows and not mac - should be Linux. | ||
| timespec t = internal::MsToAbsoluteTimespec(milliseconds); | ||
| return sem_timedwait(semaphore_, &t) == 0; | ||
| while (true) { |
Copy link
Copy Markdown
Contributor
There was a problem hiding this comment.
Would it be possible to add a test exercising this failure/fix to semaphore_test.cc? (Or, even better, does re-enabling the disabled MultithreadedStressTest in that file now work?)
Sorry, something went wrong.
Copy link
Copy Markdown
Contributor
Author
There was a problem hiding this comment.
I'll take a look.
Sorry, something went wrong.
Copy link
Copy Markdown
Contributor
Author
Sorry, something went wrong.
Copy link
Copy Markdown
Contributor
Author
|
Vindication! One of the nightly test runs failed with this assertion failure: semaphore.h:189: bool firebase::Semaphore::TimedWait(int): Assertion `"sem_timedwait() failed with EINVAL" == 0' failed. https://github.com/firebase/firebase-cpp-sdk/runs/7338303645 According to https://linux.die.net/man/3/sem_timedwait, EINVAL occurs in one of two cases:
|
Sorry, something went wrong.
dconeybe
added a commit
that referenced
this pull request
Jul 20, 2022
…s with the TimedWait() fix in #1021
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 subscribe to this conversation on GitHub.
Already have an account?
Sign in.
Reviewers
Assignees
3 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.
This fixes a latent bug where Future::Wait(int timeout_milliseconds) would occasionally return prematurely, when neither the timeout had expired nor the Future been completed. This was due to the implementation of Semaphore::TimedWait(int milliseconds) which calls sem_timedwait() on Linux and Android and neglected to check if the errno was EINTR, in which case the wait should be restarted.
This bug surfaced as the integration tests for Firestore's TransactionTest.TestMaxAttempts flakily failing due to a call to Future.Await(int timeout_milliseconds) returning as if it had timed out when, in fact, no timeout had occurred.
Note that this fix only affects Linux and Android (which runs Linux under the hood).