1044738 - [EME] Hook up CDMProxy to a PlatformDecoderModule/MP4Reader

Status: RESOLVED FIXED

(Core :: Audio/Video, defect)

Description

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

We need a PDM which uses the CDMProxy to decrypt EME content. We can wrap the platforms PDM and use that for decoding the samples decrypted by the CDMProxy.

In future we'll use the CDMProxy for decoding as well.

Based on patch from CDMProxy from bug 1043147, and on Bug 1042373 which landed on m-i today.

Comment 1

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

Attachment: Patch v1

* Create a PDM which decorates an existing PDM, and uses the CDMProxy to decrypt the input MP4Samples before sending them to the decorated PDM for decoding.
* Add the various machinery in the rest of the media stack to make it work.

Comment 2

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

Once we have this bug and bug 1043147 and Bug 1042373 landed, we can write a EME GMP for clearkey, and get mochitests going.

Comment 3

[Edwin Flores [inactive from 2016-12-01] [:eflores] [:edwin]]

Comment on attachment 8463215 [details] [diff] [review]
Patch v1

Review of attachment 8463215 [details] [diff] [review]:
-----------------------------------------------------------------

Annoyingly spotless. Just a few tiny things.

::: content/media/fmp4/MP4Decoder.cpp
@@ +22,5 @@
> +extern PRLogModuleInfo* GetDemuxerLog();
> +#define LOG(...) PR_LOG(GetDemuxerLog(), PR_LOG_DEBUG, (__VA_ARGS__))
> +#else
> +#define LOG(...)
> +#endif

This isn't used in this file.

@@ +37,5 @@
> +{
> +  nsresult rv = MediaDecoder::SetCDMProxy(aProxy);
> +  NS_ENSURE_SUCCESS(rv, rv);
> +  {
> +    // The MP4Reader can't decrypting EME content until it has a CDMProxy,

nit: "can't decrypting"

::: content/media/fmp4/MP4Reader.cpp
@@ +169,5 @@
>  
>    mLayersBackendType = layerManager->GetCompositorBackendType();
>  }
>  
> +static bool sSetupPrefCache = false;

Best have this next to its (only) use so it's clear it's only used in the one place.

::: content/media/fmp4/MP4Reader.h
@@ +150,3 @@
>    // decoder.
>    nsAutoPtr<mp4_demuxer::MP4Sample> mQueuedVideoSample;
> +  nsAutoPtr<mp4_demuxer::MP4Sample> mQueuedAudioSample;

mQueuedAudioSample is never assigned.

::: content/media/fmp4/PlatformDecoderModule.h
@@ +66,5 @@
>    // This is called on the decode thread.
>    static PlatformDecoderModule* Create();
>  
> +  // Creates a PlatformDecoderModule that uses a CDMProxy to decrypt or
> +  // decrypt-and-decode EME encrypted content. If the CDM only decryptr and

nit: "decryptr"

::: content/media/fmp4/eme/EMEDecoderModule.cpp
@@ +184,5 @@
> +}
> +
> +nsresult
> +EMEDecoderModule::Shutdown()
> +{

Assert that we're on main thread.

Comment 4

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

https://tbpl.mozilla.org/?tree=Try&rev=4620444071d6

Comment 5

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

Comment on attachment 8463215 [details] [diff] [review]
Patch v1

Review of attachment 8463215 [details] [diff] [review]:
-----------------------------------------------------------------

::: content/media/fmp4/eme/EMEDecoderModule.cpp
@@ +141,5 @@
> +    mTaskQueue->SyncDispatch(
> +      NS_NewRunnableMethod(
> +        mDecoder,
> +        &MediaDataDecoder::Drain));
> +    mCallback->DrainComplete();

We should not call DrainComplete() here, we should let the wrapped MediaDataDecoder do that. We can then make the dispatch to MediaDataDecoder::Drain async too then.

Comment 6

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

https://hg.mozilla.org/integration/mozilla-inbound/rev/82606a995121

Comment 7

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/mozilla-central/rev/82606a995121