1248909 - crash in mozilla::TrackBuffersManager::CompleteResetParserState

Status: RESOLVED FIXED

(Core :: Audio/Video: Playback, defect)

Description

[[:philipp]]

This bug was filed from the Socorro interface and is 
report bp-6e8cd5f7-de5b-46a9-bf0b-7c0692160217.
=============================================================
Crashing Thread (78)
Frame 	Module 	Signature 	Source
0 	xul.dll 	mozilla::TrackBuffersManager::CompleteResetParserState() 	dom/media/mediasource/TrackBuffersManager.cpp
1 	xul.dll 	nsRunnableMethodImpl<void ( mozilla::WMFMediaDataDecoder::*)(void), 1>::Run() 	xpcom/glue/nsThreadUtils.h
2 	xul.dll 	mozilla::TaskQueue::Runner::Run() 	xpcom/threads/TaskQueue.cpp
3 	xul.dll 	nsThreadPool::Run() 	xpcom/threads/nsThreadPool.cpp
4 	xul.dll 	nsThread::ProcessNextEvent(bool, bool*) 	xpcom/threads/nsThread.cpp

this is a new signature since bug 1245463 landed.

Comment 1

[Jean-Yves Avenard [:jya]]

I don't understand how mozilla::WMFMediaDataDecoder::* could be in the crash signature :(

Comment 2

[Jean-Yves Avenard [:jya]]

Attachment: MozReview Request: Bug 1248909: [MSE] P2. Simplify diagnostic. r?gerald

Review commit: https://reviewboard.mozilla.org/r/35363/diff/#index_header
See other reviews: https://reviewboard.mozilla.org/r/35363/

Comment 3

[Jean-Yves Avenard [:jya]]

Going to change the MOZ_RELEASE_ASSERT into MOZ_DIAGNOSTIC_ASSERT for now until we can figure out what's happening.

At least this tells us an supposedly impossible solution is happening

Comment 4

[Gerald Squelart (he/him) (not at Mozilla since 2022-09-15)]

Comment on attachment 8720563 [details]
MozReview Request: Bug 1248909: [MSE] P2. Simplify diagnostic. r?gerald

https://reviewboard.mozilla.org/r/35363/#review32005

Comment 5

[Pulsebot]

https://hg.mozilla.org/integration/mozilla-inbound/rev/3a68483ad70c

Comment 6

[Jean-Yves Avenard [:jya]]

Comment on attachment 8720563 [details]
MozReview Request: Bug 1248909: [MSE] P2. Simplify diagnostic. r?gerald

Approval Request Comment
[Feature/regressing bug #]: 1248909
[User impact if declined]: Crashes
[Describe test coverage new/current, TreeHerder]: change of assertion level only
[Risks and why]: None, it's disabling a diagnostic assertion. We wanted to check if the problem was occurring, now we know
[String/UUID change made/needed]: none

Comment 7

[Jean-Yves Avenard [:jya]]

Comment on attachment 8720563 [details]
MozReview Request: Bug 1248909: [MSE] P2. Simplify diagnostic. r?gerald

Review request updated; see interdiff: https://reviewboard.mozilla.org/r/35363/diff/1-2/

Comment 8

[Gerald Squelart (he/him) (not at Mozilla since 2022-09-15)]

https://reviewboard.mozilla.org/r/35363/#review32049

r+ (again) for the code. Please review your check-in comment (confusing grammar).

Comment 9

[Jean-Yves Avenard [:jya]]

Attachment: [MSE] Use MOZ_DIAGNOSTIC_ASSERT. r=gerald

MozReview-Commit-ID: 5HPdLKopGGa

Comment 10

[Pulsebot]

https://hg.mozilla.org/integration/mozilla-inbound/rev/15485839d79f

Comment 11

[Jean-Yves Avenard [:jya]]

(In reply to Gerald Squelart [:gerald] from comment #8)
> https://reviewboard.mozilla.org/r/35363/#review32049
> 
> r+ (again) for the code. Please review your check-in comment (confusing
> grammar).

mozreview screwed up and removed the first patch and merged the 2nd together...

that or I just don't get how mozreview is supposed to work

Comment 12

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/mozilla-central/rev/3a68483ad70c

Comment 13

[Robert Kaiser]

(In reply to Jean-Yves Avenard [:jya] from comment #11)
> mozreview screwed up and removed the first patch and merged the 2nd
> together...

Could you please flag that to the MozReview devs (mcote/gps/...) to find out if there's a bug there that they can fix? Thanks!

Comment 14

[Sylvestre Ledru [:Sylvestre]]

Comment on attachment 8720563 [details]
MozReview Request: Bug 1248909: [MSE] P2. Simplify diagnostic. r?gerald

Taking it to improve the diagnostic.

Comment 15

[Wes Kocher (:KWierso) (Not reading bugmail; email directly if needed)]

I had to back this out in https://hg.mozilla.org/mozilla-central/rev/cdc8bb2a48ae because the patch caused a big spike in crashes like https://treeherder.mozilla.org/logviewer.html#?job_id=21908412&repo=mozilla-inbound

Comment 16

[Pulsebot]

https://hg.mozilla.org/mozilla-central/rev/d5daf10d3b74

Comment 17

[Wes Kocher (:KWierso) (Not reading bugmail; email directly if needed)]

Ugh, misread the commit history. The patch I backed out was fine, it was the patch that's still stuck on inbound that I needed to back out.

Relanding the original patch:
https://hg.mozilla.org/mozilla-central/rev/d5daf10d3b74

Backing out part 2: 
https://hg.mozilla.org/integration/mozilla-inbound/rev/f385fb62ec91

Comment 18

[Jean-Yves Avenard [:jya]]

(In reply to Wes Kocher (:KWierso) from comment #17)
> Ugh, misread the commit history. The patch I backed out was fine, it was the
> patch that's still stuck on inbound that I needed to back out.

you got me worried for a second, couldn't see how changing MOZ_RELEASE_ASSERT into MOZ_DIAGNOSTIC_ASSERT could lead to crashes !

Comment 19

[Pulsebot]

https://hg.mozilla.org/integration/mozilla-inbound/rev/3af8522fa44a

Comment 20

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/mozilla-central/rev/3af8522fa44a

Comment 21

[Carsten Book [:Tomcat]]

jean-yves so should uplift the commit from commit#17 or commit#20 to aurora and beta ?

Comment 22

[Jean-Yves Avenard [:jya]]

both.

The first prevent the assert for ever existing in beta.
And the 2nd completely remove the ability for this crash signature to exist.

Comment 23

[Gerald Squelart (he/him) (not at Mozilla since 2022-09-15)]

NI:myself to remember to monitor crash reports over the next few days.

Comment 24

[Carsten Book [:Tomcat]]

remote:   https://hg.mozilla.org/releases/mozilla-aurora/rev/916cf30525e8
remote:   https://hg.mozilla.org/releases/mozilla-aurora/rev/23819af3e092


remote:   https://hg.mozilla.org/releases/mozilla-beta/rev/26f6a1a3106b
remote:   https://hg.mozilla.org/releases/mozilla-beta/rev/7da232006437

Comment 25

[Jean-Yves Avenard [:jya]]

there's been no crash with this signature in the past week with updated browser. Every single crashes in the past 7 days were with build 20160215141016 which didn't include the fix