1401922 - [Bug 1400537] Crash in mozilla::MediaFormatReader::DecoderFactory::Wrapper::Init. Part2

Status: RESOLVED FIXED

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

Description

[Jean-Yves Avenard [:jya]]

Bug 1400537 didn't fully fixed the regressions introduced in bug 1400537.

This bug will have another go at it

Comment 1

[Jean-Yves Avenard [:jya]]

Attachment: Bug 1401922 - P1. Ensure an error is returned when no decoder could be created.

PDMFactory::CreateDecoder may not always modify CreateDecoderParams::mError as not all PDM handle this optional return value.

Review commit: https://reviewboard.mozilla.org/r/182162/diff/#index_header
See other reviews: https://reviewboard.mozilla.org/r/182162/

Comment 2

[Jean-Yves Avenard [:jya]]

Attachment: Bug 1401922 - P2. Only set error when non handled.

Review commit: https://reviewboard.mozilla.org/r/182164/diff/#index_header
See other reviews: https://reviewboard.mozilla.org/r/182164/

Comment 3

[Alastor Wu [:alwu]]

Comment on attachment 8910703 [details]
Bug 1401922 - P1. Ensure an error is returned when no decoder could be created.

https://reviewboard.mozilla.org/r/182162/#review187544

Thanks!

::: dom/media/MediaFormatReader.cpp:778
(Diff revision 1)
>    }
>  
>    if (aData.mDecoder) {
>      return NS_OK;
>    }
>  

Maybe we could add an assertion here to make sure the result won't be unexpectedly changed to NS_OK when creating fail?

Comment 4

[Alastor Wu [:alwu]]

Comment on attachment 8910704 [details]
Bug 1401922 - P2. Only set error when non handled.

https://reviewboard.mozilla.org/r/182164/#review187546

Comment 5

[Jean-Yves Avenard [:jya]]

Comment on attachment 8910703 [details]
Bug 1401922 - P1. Ensure an error is returned when no decoder could be created.

Review request updated; see interdiff: https://reviewboard.mozilla.org/r/182162/diff/1-2/

Comment 6

[Jean-Yves Avenard [:jya]]

Comment on attachment 8910704 [details]
Bug 1401922 - P2. Only set error when non handled.

Review request updated; see interdiff: https://reviewboard.mozilla.org/r/182164/diff/1-2/

Comment 7

[Jean-Yves Avenard [:jya]]

Comment on attachment 8910703 [details]
Bug 1401922 - P1. Ensure an error is returned when no decoder could be created.

https://reviewboard.mozilla.org/r/182162/#review187544

> Maybe we could add an assertion here to make sure the result won't be unexpectedly changed to NS_OK when creating fail?

good point. making it even a release_assert

Comment 8

[Jean-Yves Avenard [:jya]]

[Tracking Requested - why for this release]: recent regression in 57. Fairly certain these changes now fix it for good

Comment 9

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/907df99df1bb
P1. Ensure an error is returned when no decoder could be created. r=alwu
https://hg.mozilla.org/integration/autoland/rev/6d14932fb843
P2. Only set error when non handled. r=alwu

Comment 10

[Phil Ringnalda (:philor)]

https://hg.mozilla.org/mozilla-central/rev/907df99df1bb
https://hg.mozilla.org/mozilla-central/rev/6d14932fb843

Comment 11

[Julien Cristau [:jcristau] (away, back August 16)]

Please request uplift to beta when you get a chance.

Comment 12

[Calixte Denizet (:calixte)]

For information, there are 112 crashes for 17 installations in 57.0b0 (dev edition) for signature 'mozilla::EMEDecryptor::Init'.

Comment 13

[Julien Cristau [:jcristau] (away, back August 16)]

One crash on nightly with the 20170921220243 build, none since, as far as I can tell.

Comment 14

[Jean-Yves Avenard [:jya]]

Comment on attachment 8910703 [details]
Bug 1401922 - P1. Ensure an error is returned when no decoder could be created.

Approval Request Comment
[Feature/Bug causing the regression]: 1397141
[User impact if declined]: Crash
[Is this code covered by automated tests?]: no
[Has the fix been verified in Nightly?]: No crashes since fix went in in 58
[Needs manual test from QE? If yes, steps to reproduce]: I assume that if you attempt to play a DRM encumbered video when DRM is disabled you could reproduce it.
[List of other uplifts needed for the feature/fix]: both changes to be uplifted
[Is the change risky?]: no
[Why is the change risky/not risky?]: reversing to previous code logic.
[String changes made/needed]: none

Comment 15

[Sylvestre Ledru [:Sylvestre]]

Comment on attachment 8910703 [details]
Bug 1401922 - P1. Ensure an error is returned when no decoder could be created.

Fix a crash, taking it.
Should be in 57b4 (gtb tomorrow Thursday)

Comment 16

[[Back August 16] Ryan VanderMeulen]

https://hg.mozilla.org/releases/mozilla-beta/rev/7cd3c4b643b0
https://hg.mozilla.org/releases/mozilla-beta/rev/d2a6c60348a8