1366907 - Remove BufferDecoder's GMPCrashHelper

Status: RESOLVED FIXED

(Core :: Web Audio, enhancement, P3)

Description

[Chris Pearce [:cpearce (Not reading bugmail)]]

We no longer support use of GMPs for decoding of non-EME content. So we can remove the GMPCrashHelper from the WebAudio BufferDecoder class, since we can't be using GMPs here.

This was added in bug 1267918.

Comment 1

[Chris Pearce [:cpearce (Not reading bugmail)]]

Attachment: Bug 1366907 - Remove BufferDecoder's GMPCrashHelper.

We no longer support decoding non-EME content via GMPs, so it's not possible
for a GMP to be used for decoding content that passes through a WebAudio
BufferDecoder. So we don't need to keep the GMPCrashHelper around in the
WebAudio code. If EME content is being used, there will be a crash helper on
the MediaKeys object anyway.

Review commit: https://reviewboard.mozilla.org/r/141648/diff/#index_header
See other reviews: https://reviewboard.mozilla.org/r/141648/

Comment 2

[Jean-Yves Avenard [:jya]]

Why no support for GMP? Since when?

I need this for webrtc use of MediaDataDecoder with OpenH264

Comment 3

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

I think the relevant portion here is "from the WebAudio BufferDecoder class" - so this is probably fine.

Chris: can you clarify?  OpenH264 is non-EME, and we've talked about moving things like VC1 decoding (and encoding once we do that) into a GMP process for security reasons, especially for interactive use.

Comment 4

[Chris Pearce [:cpearce (Not reading bugmail)]]

(In reply to Jean-Yves Avenard [:jya] from comment #2)
> Why no support for GMP? Since when?

When we removed Adobe's GMP in January, we lost the only GMP that had any useful audio decoders, so I removed the GMP audio decoding path.

> I need this for webrtc use of MediaDataDecoder with OpenH264

The code I'm removing is only called by AudioContent.decodeAudioData(). WebRTC and OpenH264 are not involved here at all.

Comment 5

[Chris Pearce [:cpearce (Not reading bugmail)]]

(In reply to Randell Jesup [:jesup] from comment #3)
> I think the relevant portion here is "from the WebAudio BufferDecoder class"
> - so this is probably fine.

Exactly.

> Chris: can you clarify?  OpenH264 is non-EME, and we've talked about moving
> things like VC1 decoding (and encoding once we do that) into a GMP process
> for security reasons, especially for interactive use.

We can decode VC1?

The existing GMPVideoDecoder/MediaDataDecoder still has a GMPCrashHelper. Nothing I'm doing here has any effect on decoding video via GMP. WebAudio is for audio, right?

Comment 6

[Karl Tomlinson (:karlt)]

Comment on attachment 8870197 [details]
Bug 1366907 - Remove BufferDecoder's GMPCrashHelper.

https://reviewboard.mozilla.org/r/141648/#review145464

(In reply to Chris Pearce (:cpearce) from comment #5)
> WebAudio is for audio, right?

Correct (though I haven't actually checked what happens if there is video muxed with the audio).

::: commit-message-4c7ee:3
(Diff revision 1)
> +Bug 1366907 - Remove BufferDecoder's GMPCrashHelper. r?karlt
> +
> +We no longer support decoding non-EME content via GMPs, so it's not possible

Sounds like there may still be GMP decoding of some non-EME "content", but not audio.

If appropriate please be specific, with for example s/content/audio/.

Comment 7

[Chris Pearce [:cpearce (Not reading bugmail)]]

Comment on attachment 8870197 [details]
Bug 1366907 - Remove BufferDecoder's GMPCrashHelper.

Review request updated; see interdiff: https://reviewboard.mozilla.org/r/141648/diff/1-2/

Comment 8

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/53561331ff27
Remove BufferDecoder's GMPCrashHelper. r=karlt

Comment 9

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

https://hg.mozilla.org/mozilla-central/rev/53561331ff27