842332 - WebRTC Assertion failure: PR_FALSE == lock->locked, at nsprpub/pr/src/pthreads/ptsynch.c:158

Status: RESOLVED FIXED

(Core :: WebRTC, defect, P1)

Description

[Christoph Diehl [:posidron]]

Attachment: testcase

Load the testcase in the Browser.

Tested with m-i changeset: 122171:5e137a87e84f

Comment 1

[Christoph Diehl [:posidron]]

Attachment: testcase

Whoops. That was the wrong testcase. This one is correct.

Comment 2

[Ethan Hugg [:ehugg]]

Attachment: Stacktrace from Linux

bt from gdb on Linux64 with M-I rev listed above.  Latest M-C didn't do this.

Comment 3

[Ethan Hugg [:ehugg]]

Attachment: Fix 842332 Fix missing unlock in PeerConnectionMedia

Comment 4

[Ethan Hugg [:ehugg]]

Comment on attachment 715251 [details] [diff] [review]
Fix  842332 Fix missing unlock in PeerConnectionMedia


This appears to fix it on my build, ran the testcase for a couple times for several minutes each.

Comment 5

[Randell Jesup [:jesup] (needinfo me)]

Comment on attachment 715251 [details] [diff] [review]
Fix  842332 Fix missing unlock in PeerConnectionMedia

Review of attachment 715251 [details] [diff] [review]:
-----------------------------------------------------------------

r+

Why are we using PR_Lock directly and not Mutexes and MutexAutoLocks?  They make it much harder to make this mistake.

Comment 6

[Ethan Hugg [:ehugg]]

Comment on attachment 715251 [details] [diff] [review]
Fix  842332 Fix missing unlock in PeerConnectionMedia


http://hg.mozilla.org/integration/mozilla-inbound/rev/ea99e1b1d574

Comment 7

[Ethan Hugg [:ehugg]]

>Why are we using PR_Lock directly and not Mutexes and MutexAutoLocks?  They make it >much harder to make this mistake.

This code probably predates the availability of those.  I did a grep of PR_Lock and it only appears here and in Wrapper.h.  Should I file a new bug to change these to MutexAutoLock?

Comment 8

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/mozilla-central/rev/ea99e1b1d574