Bug 219622 – [Media in GPU Process][MSE] The GPU Process may crash in the destructor of MediaPlayerPrivateMediaSourceAVFObjC

• Component: Media
• Status: RESOLVED FIXED

[Peng Liu]

[Media in GPU Process][MSE] The GPU Process may crash in the destructor of MediaPlayerPrivateMediaSourceAVFObjC

[Peng Liu]

Created attachment 415604 [details]
Patch

[youenn fablet]

Comment on attachment 415604 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=415604&action=review

> Source/WebKit/GPUProcess/media/cocoa/RemoteMediaPlayerProxyCocoa.mm:-60
> -    m_inlineLayerHostingContext->setRootLayer(m_player->platformLayer());

It is not clear why removing this line is fine.
Is it redundant? If so, can you add this info in change log?

[Peng Liu]

Comment on attachment 415604 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=415604&action=review

>> Source/WebKit/GPUProcess/media/cocoa/RemoteMediaPlayerProxyCocoa.mm:-60
>> -    m_inlineLayerHostingContext->setRootLayer(m_player->platformLayer());
> 
> It is not clear why removing this line is fine.
> Is it redundant? If so, can you add this info in change log?

The purpose of this callback is to notify the DOM that the player's rendering mode is changed, including the case that the layer is destroyed. So I think it was a bug to operate on the layer.
I just realized that we need to forward the callback to the Web process (HTMLMediaElement). I will upload a new patch.

[Peng Liu]

Created attachment 415651 [details]
Patch

[Peng Liu]

Created attachment 415653 [details]
Fix GTK/WPE build failures

[Peng Liu]

Created attachment 415657 [details]
Fix a build failure on the wincairo port

[EWS]

Committed r270563: <https://trac.webkit.org/changeset/270563>

All reviewed patches have been landed. Closing bug and clearing flags on attachment 415657 [details].

[Radar WebKit Bug Importer]

<rdar://problem/72115808>