haiyangsun-db opened a new pull request, #58525:
URL: https://github.com/apache/spark/pull/58525

   ### What changes were proposed in this pull request?
   
   Capture and clear an interrupt that races with lifecycle notification in
   `DirectWorkerDispatcher.awaitInFlightOperations()`.
   
   The interrupt is checked after leaving the `lifecycleLock.synchronized` 
block.
   `Thread.interrupted()` is evaluated first so the interrupt flag is always
   cleared. `close()` restores the interrupt after cleanup completes.
   
   ### Why are the changes needed?
   
   `lifecycleLock.wait()` may return normally after notification while an
   overlapping interrupt leaves the thread's interrupt flag set. Subsequent
   cleanup may then fail with `InterruptedException` before it completes.
   
   ### Does this PR introduce _any_ user-facing change?
   
   No. The affected UDF worker functionality is unreleased and experimental.
   
   ### How was this patch tested?
   
   The existing regression test passed 40 consecutive runs:
   
   `DirectWorkerDispatcherSuite: interrupted close completes cleanup before 
restoring the interrupt`
   
   The full `DirectWorkerDispatcherSuite` also passed with all 34 tests 
successful.
   
   ### Was this patch authored or co-authored using generative AI tooling?
   
   Yes.
   


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]


---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to