1193669 - Make MediaDataDecoder::Shutdown() return a promise

Status: RESOLVED WONTFIX

(Core :: Audio/Video: Playback, defect, P5)

Description

[Jean-Yves Avenard [:jya]]

This would greatly simply shutting down decoders ; and we remove the need for the use of wait().

This depends on the removal of the SharedDecoderManager

Comment 1

[JW Wang [:jwwang] [:jw_wang]]

(In reply to Jean-Yves Avenard [:jya] from comment #0)
> This would greatly simply shutting down decoders ; and we remove the need
> for the use of wait().
Where is the use of wait()?

Comment 2

[Jean-Yves Avenard [:jya]]

(In reply to JW Wang [:jwwang] from comment #1)
> (In reply to Jean-Yves Avenard [:jya] from comment #0)
> > This would greatly simply shutting down decoders ; and we remove the need
> > for the use of wait().
> Where is the use of wait()?

This bug was created at the time we used the MP4Reader which had a SharedDecoderManager ; when shutting down the SharedDecoderManager ; it would go through the list of decoder and call shutdown while holding a lock / wait.

We had plenty of deadlocks then.

Comment 3

[JW Wang [:jwwang] [:jw_wang]]

Now we have no SharedDecoderManager and no wait() code. Is this bug still valid?

Comment 4

[Jean-Yves Avenard [:jya]]

(In reply to JW Wang [:jwwang] from comment #3)
> Now we have no SharedDecoderManager and no wait() code. Is this bug still
> valid?

Well, I guess it would help ensuring there's no pending task going ; Shutdown could simply wait for all tasks to complete , things like that.

but if you think it's unnecessary, feel free to close it.

Comment 5

[JW Wang [:jwwang] [:jw_wang]]

MFR assumes MediaDataDecoder::Shutdown() is sync and continues its call flow after Calling Shutdown(). We will need to also change MFR's assumption if MediaDataDecoder::Shutdown() returns a promise.

When we allow a MediaDataDecoder to create its own task queue, things will be much easier since 2 MediaDataDecoders do not share the same task queue and they will not interfere with each other.