moonchen commented on PR #13362: URL: https://github.com/apache/trafficserver/pull/13362#issuecomment-5880401835
@saimayithri I discussed this with community members. We agreed that work protected by a NetHandler mutex should run on that NetHandler's owning network thread. The fix direction for #13743 and #13744 is: - If reenable is called from another thread, schedule the work on the connection/session's **specific owner** before acquiring its NetHandler mutex. - If already on the owner, keep synchronous execution with the required locks. This matters for TLS verification results. - Task-thread callers, including certifier, remain supported. ATS should handle the return to the owner. I updated both issues with this guidance. Other paths, including VIO reenable, also need checking for foreign-thread NetHandler mutex acquisition. For this PR, let's find and fix the first wrong-thread dispatch or foreign mutex acquisition. Moving only teardown can leave earlier work on the wrong thread. We still need to confirm whether either reported API issue caused these crashes. -- 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]
