1434950 - Crash in _EH_prolog3

Status: RESOLVED WORKSFORME

(Core :: Graphics, defect, P2)

Description

[Calixte Denizet (:calixte)]

This bug was filed from the Socorro interface and is
report bp-ea25097f-92f9-408d-8bd9-c83850180201.
=============================================================

Top 10 frames of crashing thread:

0 mmdevapi.dll _EH_prolog3 
1 audioses.dll CCrossProcessBaseEndpoint::GetNextTimestampMessage 
2 mmdevapi.dll UnregisterMediaCallback 
3 audioses.dll ATL::CHandle::Close 
4 audioses.dll ATL::CComObject<CAudioSessionControl>::~CComObject<CAudioSessionControl> 
5 audioses.dll ATL::CComObject<CAudioSessionControl>::`scalar deleting destructor' 
6 audioses.dll ATL::CComObject<CAudioSessionControl>::Release 
7 xul.dll mozilla::widget::AudioSession::StopInternal widget/windows/AudioSession.cpp:304
8 xul.dll mozilla::widget::AudioSession::Stop widget/windows/AudioSession.cpp:321
9 xul.dll mozilla::dom::ContentChild::ShutdownInternal dom/ipc/ContentChild.cpp:2992

=============================================================

There is 1 crash (from 1 installation) in nightly 60 with buildid 20180131220303. In analyzing the backtrace, the regression may have been introduced by patch [1] to fix bug 1425559.

[1] https://hg.mozilla.org/mozilla-central/rev?node=939a770548c4

Comment 1

[Andrea Marchesini [:baku]]

bug 142559 is unrelated. It's landed 6 days ago and we have crash reports older than that. For instance:
https://crash-stats.mozilla.com/report/index/ea25097f-92f9-408d-8bd9-c83850180201

Comment 2

[Marion Daly [:mdaly]]

:baku, worth looking into either way I think

Comment 3

[Andrea Marchesini [:baku]]

It seems related to audio on windows. I'm not familiar with it.
padenot, do you know who can be help here?

Comment 4

[Paul Adenot (:padenot)]

bug 1358372 looks too close and too recent for it to be a coincidence.

David, Jim, would you mind having a look at this one?

Comment 5

[David Parks [:handyman]]

The fix for bug 1358372 went in on 2/7 and I don't see any of these crashes in builds after 2/1.

Only a few of the crashes under this signature seem related to AudioSession.  There isn't much info as to what was happening but if I were to guess I say it has something to do with a COM or a COM endpoint.  Honestly, since the stacks look like the AudioSession was valid, there isn't much else we could be responsible for in the crash.  IAudioSessionControl is quite opaque.  The volume slider bug you referenced is due to some bad message timing between Firefox and Windows' sndvol32.exe.  I believe it uses COM.  Thats a lot of speculation -- hope it helps.

Comment 6

[Marcia Knous [:marcia]]

Discussed during triage. No crashes on 60 that I can see since the original bug was filed. This feels like a won't fix for 60.

Comment 7

[Jens Stutte [:jstutte]]

From the recent stack trace I see no connection with DOM:Worker. Maybe a coincidence in the signature?

Comment 8

[BugBot [:suhaib / :marco/ :calixte]]

Since the crash volume is low (less than 5 per week), the severity is downgraded to S3. Feel free to change it back if you think the bug is still critical.

For more information, please visit auto_nag documentation.

Comment 9

[BugBot [:suhaib / :marco/ :calixte]]

Closing because no crashes reported for 12 weeks.