1353459 - Crash in java.lang.NullPointerException: at org.mozilla.gecko.media.CodecProxy.input(CodecProxy.java)

Status: RESOLVED FIXED

(Firefox for Android Graveyard :: Audio/Video, defect, P1)

Description

[James Willcox (:snorp) ([e-mail]) (he/him)]

I get this crash when trying to run Ski Safari in Fennec. The following log message appears many times as the game starts:

04-04 13:14:49.972 30178 30753 E GeckoRemoteCodec: Callbacks is dead
04-04 13:14:49.974 30178 30753 W System.err: android.media.MediaCodec$CodecException: Error 0xfffffff3
04-04 13:14:49.975 30178 30753 W System.err: 	at android.media.MediaCodec.releaseOutputBuffer(Native Method)
04-04 13:14:49.975 30178 30753 W System.err: 	at android.media.MediaCodec.releaseOutputBuffer(MediaCodec.java:2615)
04-04 13:14:49.975 30178 30753 W System.err: 	at org.mozilla.gecko.media.JellyBeanAsyncCodec.releaseOutputBuffer(JellyBeanAsyncCodec.java:395)
04-04 13:14:49.975 30178 30753 W System.err: 	at org.mozilla.gecko.media.Codec$OutputProcessor.reset(Codec.java:274)
04-04 13:14:49.975 30178 30753 W System.err: 	at org.mozilla.gecko.media.Codec$OutputProcessor.stop(Codec.java:295)
04-04 13:14:49.975 30178 30753 W System.err: 	at org.mozilla.gecko.media.Codec$OutputProcessor.access$1300(Codec.java:191)
04-04 13:14:49.975 30178 30753 W System.err: 	at org.mozilla.gecko.media.Codec.releaseCodec(Codec.java:395)
04-04 13:14:49.975 30178 30753 W System.err: 	at org.mozilla.gecko.media.Codec.release(Codec.java:517)
04-04 13:14:49.975 30178 30753 W System.err: 	at org.mozilla.gecko.media.Codec.binderDied(Codec.java:318)
04-04 13:14:49.975 30178 30753 W System.err: 	at android.os.BinderProxy.sendDeathNotice(Binder.java:688)
04-04 13:14:49.975 30178 30753 W System.err: android.os.DeadObjectException
04-04 13:14:49.975 30178 30753 W System.err: 	at android.os.BinderProxy.transactNative(Native Method)
04-04 13:14:49.975 30178 30753 W System.err: 	at android.os.BinderProxy.transact(Binder.java:615)
04-04 13:14:49.975 30178 30753 W System.err: 	at org.mozilla.gecko.media.ICodecCallbacks$Stub$Proxy.onError(ICodecCallbacks.java:156)
04-04 13:14:49.975 30178 30753 W System.err: 	at org.mozilla.gecko.media.Codec.reportError(Codec.java:445)
04-04 13:14:49.975 30178 30753 W System.err: 	at org.mozilla.gecko.media.Codec.releaseCodec(Codec.java:399)
04-04 13:14:49.975 30178 30753 W System.err: 	at org.mozilla.gecko.media.Codec.release(Codec.java:517)
04-04 13:14:49.975 30178 30753 W System.err: 	at org.mozilla.gecko.media.Codec.binderDied(Codec.java:318)
04-04 13:14:49.975 30178 30753 W System.err: 	at android.os.BinderProxy.sendDeathNotice(Binder.java:688)

My gut feeling is that it is quickly creating and destroying decoders (via <audio>?) and there is some lifecycle issue with the OOP stuff.

This bug is marked internal only because I'm not sure if we're allowed to share the game outside of Mozilla.

This bug was filed from the Socorro interface and is 
report bp-1ee36374-45d2-4fff-b8f4-5490f2170404.
=============================================================

Comment 1

[James Willcox (:snorp) ([e-mail]) (he/him)]

Andre has verified this URL is public.

Comment 2

[James Willcox (:snorp) ([e-mail]) (he/him)]

Blake, John, can you take a look at this?

Comment 3

[Blake Wu [:bwu][:blakewu]]

I can repro this bug on my Pixel easily by visiting the URL.

Comment 4

[John Lin [:jhlin][:jolin]]

There seems 2 issues:
1. Exceptions raised in Codec.queueInput() are not handled.
2. RemoteDataDecoder::Error() has no effect when there is no pending promise.

Comment 5

[John Lin [:jhlin][:jolin]]

Attachment: Bug 1353459 - handle queueInputBuffer exceptions.

Review commit: https://reviewboard.mozilla.org/r/129388/diff/#index_header
See other reviews: https://reviewboard.mozilla.org/r/129388/

Comment 6

[Eugen Sawin [:esawin]]

Comment on attachment 8857392 [details]
Bug 1353459 - handle queueInputBuffer exceptions.

https://reviewboard.mozilla.org/r/129388/#review132698

Comment 7

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/202eda8e82ce
handle queueInputBuffer exceptions. r=esawin

Comment 8

[Sebastian Hengst [:aryx] (needinfo me if it's about an intermittent or backout)]

https://hg.mozilla.org/mozilla-central/rev/202eda8e82ce

Comment 9

[Ryan VanderMeulen [:RyanVM]]

AFAICT, this goes back to 53 or so. Looks like we could use an Aurora approval here for 54?

Comment 10

[John Lin [:jhlin][:jolin]]

Comment on attachment 8857392 [details]
Bug 1353459 - handle queueInputBuffer exceptions.

Approval Request Comment
[Feature/Bug causing the regression]:bug 1257777
[User impact if declined]: browser crash when system out of shared memory
[Is this code covered by automated tests?]: no
[Has the fix been verified in Nightly?]: yes
[Needs manual test from QE? If yes, steps to reproduce]: yes. Link to the URL in the bug and see if it crash the browser
[List of other uplifts needed for the feature/fix]: none
[Is the change risky?]: no
[Why is the change risky/not risky?]: it catches more exceptions
[String changes made/needed]: none

Comment 11

[Gerry Chang [:gchang]]

Comment on attachment 8857392 [details]
Bug 1353459 - handle queueInputBuffer exceptions.

Fix a crash. Aurora54+.

Comment 12

[Iris Hsiao [:ihsiao]]

https://hg.mozilla.org/releases/mozilla-aurora/rev/b7cdc8cfc61f

Comment 13

[Carsten Book [:Tomcat]]

Hi, backed out of aurora because of failurs like https://treeherder.mozilla.org/logviewer.html#?job_id=92262250&repo=mozilla-aurora

Comment 14

[John Lin [:jhlin][:jolin]]

Attachment: Patch for 54

Sorry for the inconvenience. Please uplift this patch instead. Thanks a lot.

Comment 15

[John Lin [:jhlin][:jolin]]

Attachment: Patch for 54

The correct patch for 54.

Comment 16

[Iris Hsiao [:ihsiao]]

https://hg.mozilla.org/releases/mozilla-aurora/rev/7595a01c76d1

Comment 17

[Blake Wu [:bwu][:blakewu]]

Iris, 
It is great to have your help here!