Bug 221457 – Enable audio capture for speech recognition in GPUProcess

• Status: RESOLVED
• Resolution: FIXED
• Component: WebRTC

[youenn fablet]

Enable audio capture for speech recognition in GPUProcess

[youenn fablet]

Created attachment 419381 [details]
Patch

[Radar WebKit Bug Importer]

<rdar://problem/74021657>

[Sihui Liu]

*** Bug 221452 has been marked as a duplicate of this bug. ***

[EWS]

commit-queue failed to commit attachment 419381 [details] to WebKit repository. To retry, please set cq+ flag again.

[Sihui Liu]

Comment on attachment 419381 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=419381&action=review

Nice! Was trying to add a SpeechRecognitionRealtimeMediaSourceManager to GPU process and messaging between GPU process and UI process; that works but requires more changes.

> Source/WebKit/WebProcess/Speech/SpeechRecognitionRealtimeMediaSourceManager.cpp:191
>      m_sandboxExtensionForTCCD = SandboxExtension::create(WTFMove(sandboxHandleForTCCD));
>      if (!m_sandboxExtensionForTCCD)
> -        LOG_ERROR("Failed to create sandbox extension for tccd");
> +        RELEASE_LOG_ERROR(Media, "Failed to create sandbox extension for tccd");
>      else
>          m_sandboxExtensionForTCCD->consume();
>  
>      m_sandboxExtensionForMicrophone = SandboxExtension::create(WTFMove(sandboxHandleForMicrophone));
>      if (!m_sandboxExtensionForMicrophone)
> -        LOG_ERROR("Failed to create sandbox extension for microphone");
> +        RELEASE_LOG_ERROR(Media, "Failed to create sandbox extension for microphone");

Should we also avoid granting sandbox extension to web process if capture is about to happen in GPU process?

[youenn fablet]

(In reply to Sihui Liu from comment #5)
> Comment on attachment 419381 [details]
> Patch
> 
> View in context:
> https://bugs.webkit.org/attachment.cgi?id=419381&action=review
> 
> Nice! Was trying to add a SpeechRecognitionRealtimeMediaSourceManager to GPU
> process and messaging between GPU process and UI process; that works but
> requires more changes.
> 
> > Source/WebKit/WebProcess/Speech/SpeechRecognitionRealtimeMediaSourceManager.cpp:191
> >      m_sandboxExtensionForTCCD = SandboxExtension::create(WTFMove(sandboxHandleForTCCD));
> >      if (!m_sandboxExtensionForTCCD)
> > -        LOG_ERROR("Failed to create sandbox extension for tccd");
> > +        RELEASE_LOG_ERROR(Media, "Failed to create sandbox extension for tccd");
> >      else
> >          m_sandboxExtensionForTCCD->consume();
> >  
> >      m_sandboxExtensionForMicrophone = SandboxExtension::create(WTFMove(sandboxHandleForMicrophone));
> >      if (!m_sandboxExtensionForMicrophone)
> > -        LOG_ERROR("Failed to create sandbox extension for microphone");
> > +        RELEASE_LOG_ERROR(Media, "Failed to create sandbox extension for microphone");
> 
> Should we also avoid granting sandbox extension to web process if capture is
> about to happen in GPU process?

Sounds good, will do that as a follow-up if you don't beat me to it.
Will land this change so that I can try landing enabling audio capture in GPU Process.

[EWS]

Committed r272434: <https://trac.webkit.org/changeset/272434>

All reviewed patches have been landed. Closing bug and clearing flags on attachment 419381 [details].