media: Add External Clear Key content browser test on Android

media: Add External Clear Key content browser test on Android

This CL adds External Clear Key support on Android for testing.
This is implemented by running AesDecryptor in the MojoCdmService,
which runs in the MojoMediaApplication in the GPU process. This is
only enabled when kExternalClearKeyForTesting feature is enabled.

A new content browser test is added to use External Clear Key on
Android. Since AesDecryptor doesn't support decoding, the media
pipeline is configured to do decrypt-only using the mojo CDM
(MojoDecryptor on MojoCdm), and then use the normal Android
pipeline (AVDA/MojoAudioDecoder) to decode encrypted audio/video.

Note that this is different from the default mode how Android
media pipeline works for encrypted content (decryption/decoding
both happens in the GPU process). When browser_tests are enabled
on Android, we should be able to have test coverage on that.

Here's what the new test covers:
- MojoCdm / MojoCdmService
- MojoDecryptor / MojoDecryptorService
- MojoMediaApplication
- Connection to services hosted in MojoMediaApplication through
  MojoShellContext.

This CL also fixes a bug in MojoCdm where the decryptor_ptr is bound to
one thread but is used on another thread.

BUG=581893
TEST=This CL adds a new test. No other functionality change.

Committed: https://crrev.com/bcd6858191102c4d4c949a617560231908480235
Cr-Commit-Position: refs/heads/master@{#420500}

[xhwang, 2016-08-24 00:44:43]

xhwang@chromium.org changed reviewers:
+ ddorwin@chromium.org, jrummell@chromium.org

[xhwang, 2016-08-24 00:44:43]

jrummell: Please review everything
ddorwin: Please do a high level review

[xhwang, 2016-08-24 00:45:57]

This CL is based on https://chromiumcodereview.appspot.com/2267283002/ where
renewal is moved to a separate test. On Android, we won't have renewal since
it's not supported in AesDecryptor.

[jrummell, 2016-08-24 20:36:53]

LG

https://codereview.chromium.org/2268283003/diff/1/components/cdm/renderer/ext...
File components/cdm/renderer/external_clear_key_key_system_properties.h (right):

https://codereview.chromium.org/2268283003/diff/1/components/cdm/renderer/ext...
components/cdm/renderer/external_clear_key_key_system_properties.h:39:
std::string GetPepperType() const override;
media::KeySystemProperties always specifies GetPepperType(). Shouldn't this be
defined, even if !ENABLE_PEPPER_CDMS? It looks like Widevine does this as well,
but all the calls aren't conditionally compiled in (although I guess the
assumption is that the properties won't be included if !ENABLE_PEPPER_CDMS).

https://codereview.chromium.org/2268283003/diff/1/content/browser/media/encry...
File content/browser/media/encrypted_media_browsertest.cc (right):

https://codereview.chromium.org/2268283003/diff/1/content/browser/media/encry...
content/browser/media/encrypted_media_browsertest.cc:21: const char
kExternalClearKeyKeySystem[] = "org.chromium.externalclearkey";
No complaints about unused value if !OS_ANDROID?

https://codereview.chromium.org/2268283003/diff/1/content/browser/media/encry...
content/browser/media/encrypted_media_browsertest.cc:213: // tested in
browser_tests.
If this is already supported, why not run it? The other tests are also mostly in
browser_tests, so why not simply run this additional test here?

https://codereview.chromium.org/2268283003/diff/1/content/browser/media/encry...
content/browser/media/encrypted_media_browsertest.cc:215:
RunSimpleEncryptedMediaTest("bear-320x240-av_enc-av.webm", kWebMAudioVideo,
The similar ClearKey test checks media::PlatformHasOpusSupport(). Should this
test do that as well?

[xhwang, 2016-08-24 23:29:51]

comments addressed

[xhwang, 2016-08-24 23:31:03]

Thanks! PTAL again.

https://codereview.chromium.org/2268283003/diff/1/components/cdm/renderer/ext...
File components/cdm/renderer/external_clear_key_key_system_properties.h (right):

https://codereview.chromium.org/2268283003/diff/1/components/cdm/renderer/ext...
components/cdm/renderer/external_clear_key_key_system_properties.h:39:
std::string GetPepperType() const override;
On 2016/08/24 20:36:52, jrummell wrote:
> media::KeySystemProperties always specifies GetPepperType(). Shouldn't this be
> defined, even if !ENABLE_PEPPER_CDMS? It looks like Widevine does this as
well,
> but all the calls aren't conditionally compiled in (although I guess the
> assumption is that the properties won't be included if !ENABLE_PEPPER_CDMS).

There's a default implementation for this where it'll return an empty string:

https://cs.chromium.org/chromium/src/media/base/key_system_properties.cc?rcl=...

The default implementation also has a nice DCHECK that this shouldn't be called
if ENABLE_PEPPER_CDMS is not defined.

https://codereview.chromium.org/2268283003/diff/1/content/browser/media/encry...
File content/browser/media/encrypted_media_browsertest.cc (right):

https://codereview.chromium.org/2268283003/diff/1/content/browser/media/encry...
content/browser/media/encrypted_media_browsertest.cc:21: const char
kExternalClearKeyKeySystem[] = "org.chromium.externalclearkey";
On 2016/08/24 20:36:52, jrummell wrote:
> No complaints about unused value if !OS_ANDROID?

Done.

https://codereview.chromium.org/2268283003/diff/1/content/browser/media/encry...
content/browser/media/encrypted_media_browsertest.cc:213: // tested in
browser_tests.
On 2016/08/24 20:36:53, jrummell wrote:
> If this is already supported, why not run it? The other tests are also mostly
in
> browser_tests, so why not simply run this additional test here?

As discussed offline, ECK on desktop is not supported in content/ so we can't
run it here.

https://codereview.chromium.org/2268283003/diff/1/content/browser/media/encry...
content/browser/media/encrypted_media_browsertest.cc:215:
RunSimpleEncryptedMediaTest("bear-320x240-av_enc-av.webm", kWebMAudioVideo,
On 2016/08/24 20:36:52, jrummell wrote:
> The similar ClearKey test checks media::PlatformHasOpusSupport(). Should this
> test do that as well?

That check is for opus. I specifically chose a non-opus test content :)

[jrummell, 2016-08-25 00:14:13]

lgtm

https://codereview.chromium.org/2268283003/diff/1/content/browser/media/encry...
File content/browser/media/encrypted_media_browsertest.cc (right):

https://codereview.chromium.org/2268283003/diff/1/content/browser/media/encry...
content/browser/media/encrypted_media_browsertest.cc:215:
RunSimpleEncryptedMediaTest("bear-320x240-av_enc-av.webm", kWebMAudioVideo,
On 2016/08/24 23:31:03, xhwang (slow) wrote:
> On 2016/08/24 20:36:52, jrummell wrote:
> > The similar ClearKey test checks media::PlatformHasOpusSupport(). Should
this
> > test do that as well?
> 
> That check is for opus. I specifically chose a non-opus test content :)

I missed that. I just looked for the previous test that used "av_enc-av.webm",
and didn't see "opus" in the name.

[xhwang, 2016-08-25 04:04:05]

ddorwin: let me know if you also want to take a look :)

[xhwang, 2016-08-25 04:04:12]

The CQ bit was checked by xhwang@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-08-25 04:04:30]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-08-25 04:22:41]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-08-25 04:22:42]

Dry run: Try jobs failed on following builders:
  chromeos_x86-generic_chromium_compile_only_ng on
master.tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/chromeos_x86-ge...)

[xhwang, 2016-08-25 16:31:27]

The CQ bit was checked by xhwang@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-08-25 16:32:06]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-08-25 21:01:18]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-08-25 21:01:21]

Dry run: This issue passed the CQ dry run.

[ddorwin, 2016-09-14 01:06:26]

Why does ECK only do decryption and use the normal path for decoding. At least
for video, this is inconsistent with the MediaDrm flow. If the goal of this CL
is to exercise code, shouldn't it exercise those paths too? Maybe we can't
because there is no actual MediaDrm support and we'd have to spin up decoders on
the server side? Maybe note that somewhere in the description and/or code.

https://codereview.chromium.org/2268283003/diff/40001/content/browser/media/e...
File content/browser/media/encrypted_media_browsertest.cc (right):

https://codereview.chromium.org/2268283003/diff/40001/content/browser/media/e...
content/browser/media/encrypted_media_browsertest.cc:215: // On other platforms,
External Clear Key is supported in chrome/, so it is
This difference and all the ifdefs are unfortunate.

https://codereview.chromium.org/2268283003/diff/40001/media/base/android/andr...
File media/base/android/android_cdm_factory.cc (right):

https://codereview.chromium.org/2268283003/diff/40001/media/base/android/andr...
media/base/android/android_cdm_factory.cc:44: if
(IsExternalClearKey(key_system)) {
Let's discuss.

https://codereview.chromium.org/2268283003/diff/40001/media/base/key_systems.cc
File media/base/key_systems.cc (right):

https://codereview.chromium.org/2268283003/diff/40001/media/base/key_systems....
media/base/key_systems.cc:423: can_block = true;
Please add a comment. Maybe above with the general description.
We may also want to note that this is only a policy check and not a security
check since AesDecryptor runs in a different process, albeit the trusted browser
process.

[xhwang, 2016-09-16 18:27:08]

Description was changed from

==========
media: Add External Clear Key content browser test on Android

This CL adds External Clear Key support on Android for testing.
This is implemented by running AesDecryptor in the MojoCdmService,
which runs in the MojoMediaApplication in the GPU process.

A new content browser test is added to use External Clear Key on
Android. The media pipeline is configured to do decrypt-only using the
mojo CDM (MojoDecryptor on MojoCdm), and then use the normal Android
pipeline to decode decrypted audio/video.

This test adds test coverage to the following code paths:
- MojoCdm / MojoCdmService
- MojoDecryptor / MojoDecryptorService
- MojoMediaApplication
- Connection to services hosted in MojoMediaApplication through
  MojoShellContext.

This CL also fixes a bug in MojoCdm where the decryptor_ptr is bound to
one thread but is used on another thread.

BUG=581893
TEST=This CL adds a new test. No other functionality change.
==========

to

==========
media: Add External Clear Key content browser test on Android

This CL adds External Clear Key support on Android for testing.
This is implemented by running AesDecryptor in the MojoCdmService,
which runs in the MojoMediaApplication in the GPU process. This is
only enabled when kExternalClearKeyForTesting feature is enabled.

A new content browser test is added to use External Clear Key on
Android. Since AesDecryptor doesn't support decoding, the media
pipeline is configured to do decrypt-only using the mojo CDM
(MojoDecryptor on MojoCdm), and then use the normal Android
pipeline (AVDA/MojoAudioDecoder) to decode encrypted audio/video.

Note that this is different from the default mode how Android
media pipeline works for encrypted content (decryption/decoding
both happens in the GPU process). When browser_tests are enabled
on Android, we should be able to have test coverage on that.

Here's what the new test covers:
- MojoCdm / MojoCdmService
- MojoDecryptor / MojoDecryptorService
- MojoMediaApplication
- Connection to services hosted in MojoMediaApplication through
  MojoShellContext.

This CL also fixes a bug in MojoCdm where the decryptor_ptr is bound to
one thread but is used on another thread.

BUG=581893
TEST=This CL adds a new test. No other functionality change.
==========

[xhwang, 2016-09-16 18:49:58]

ddorwin: I updated the CL description based on the comments. Also I added the
feature check and fixed some comments in the code. PTAL again.

https://codereview.chromium.org/2268283003/diff/40001/content/browser/media/e...
File content/browser/media/encrypted_media_browsertest.cc (right):

https://codereview.chromium.org/2268283003/diff/40001/content/browser/media/e...
content/browser/media/encrypted_media_browsertest.cc:215: // On other platforms,
External Clear Key is supported in chrome/, so it is
On 2016/09/14 01:06:27, ddorwin wrote:
> This difference and all the ifdefs are unfortunate.

Acknowledged.

https://codereview.chromium.org/2268283003/diff/40001/media/base/android/andr...
File media/base/android/android_cdm_factory.cc (right):

https://codereview.chromium.org/2268283003/diff/40001/media/base/android/andr...
media/base/android/android_cdm_factory.cc:44: if
(IsExternalClearKey(key_system)) {
On 2016/09/14 01:06:27, ddorwin wrote:
> Let's discuss.

Done.

[xhwang, 2016-09-16 18:50:17]

https://codereview.chromium.org/2268283003/diff/40001/media/base/key_systems.cc
File media/base/key_systems.cc (right):

https://codereview.chromium.org/2268283003/diff/40001/media/base/key_systems....
media/base/key_systems.cc:423: can_block = true;
On 2016/09/14 01:06:27, ddorwin wrote:
> Please add a comment. Maybe above with the general description.
> We may also want to note that this is only a policy check and not a security
> check since AesDecryptor runs in a different process, albeit the trusted
browser
> process.

Done.

[ddorwin, 2016-09-16 19:40:07]

LGTM % comments

https://codereview.chromium.org/2268283003/diff/100001/content/shell/renderer...
File content/shell/renderer/shell_content_renderer_client.cc (right):

https://codereview.chromium.org/2268283003/diff/100001/content/shell/renderer...
content/shell/renderer/shell_content_renderer_client.cc:141: if
(!base::FeatureList::IsEnabled(media::kExternalClearKeyForTesting))
Is this really necessary?

https://codereview.chromium.org/2268283003/diff/100001/media/base/android/and...
File media/base/android/android_cdm_factory.cc (right):

https://codereview.chromium.org/2268283003/diff/100001/media/base/android/and...
media/base/android/android_cdm_factory.cc:45: if
(base::FeatureList::IsEnabled(media::kExternalClearKeyForTesting) &&
nit: This order seems strange.

https://codereview.chromium.org/2268283003/diff/100001/media/base/key_systems.cc
File media/base/key_systems.cc (right):

https://codereview.chromium.org/2268283003/diff/100001/media/base/key_systems...
media/base/key_systems.cc:414: // key system on Android, only enabled for
testing) assume that the CDM can
nit: It's not clear exactly what is only enabled for testing. AES decryptor is
not just for testing.

[xhwang, 2016-09-16 20:01:43]

comments addressed

[xhwang, 2016-09-16 20:02:16]

https://codereview.chromium.org/2268283003/diff/100001/content/shell/renderer...
File content/shell/renderer/shell_content_renderer_client.cc (right):

https://codereview.chromium.org/2268283003/diff/100001/content/shell/renderer...
content/shell/renderer/shell_content_renderer_client.cc:141: if
(!base::FeatureList::IsEnabled(media::kExternalClearKeyForTesting))
On 2016/09/16 19:40:07, ddorwin wrote:
> Is this really necessary?

Not really. But if ECK is not enabled, we don't register/enable it anywhere,
even in a test shell.

https://codereview.chromium.org/2268283003/diff/100001/media/base/android/and...
File media/base/android/android_cdm_factory.cc (right):

https://codereview.chromium.org/2268283003/diff/100001/media/base/android/and...
media/base/android/android_cdm_factory.cc:45: if
(base::FeatureList::IsEnabled(media::kExternalClearKeyForTesting) &&
On 2016/09/16 19:40:07, ddorwin wrote:
> nit: This order seems strange.

I see your point. What I am thinking is "if ECK is not supported, don't bother
to check whether the key system is ECK"...

https://codereview.chromium.org/2268283003/diff/100001/media/base/key_systems.cc
File media/base/key_systems.cc (right):

https://codereview.chromium.org/2268283003/diff/100001/media/base/key_systems...
media/base/key_systems.cc:414: // key system on Android, only enabled for
testing) assume that the CDM can
On 2016/09/16 19:40:07, ddorwin wrote:
> nit: It's not clear exactly what is only enabled for testing. AES decryptor is
> not just for testing.

Done.

[xhwang, 2016-09-16 20:02:21]

The CQ bit was checked by xhwang@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-09-16 20:02:55]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-09-16 21:20:00]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-09-16 21:20:02]

Dry run: This issue passed the CQ dry run.

[xhwang, 2016-09-16 21:39:41]

The CQ bit was checked by xhwang@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-09-16 21:40:16]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[xhwang, 2016-09-16 21:40:47]

xhwang@chromium.org changed reviewers:
+ mkwst@chromium.org

[xhwang, 2016-09-16 21:40:48]

mkwst@chromium.org: Please OWNERS review changes in content/shell/*

[commit-bot: I haz the power, 2016-09-16 21:47:13]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-09-16 21:47:18]

Dry run: This issue passed the CQ dry run.

[xhwang, 2016-09-19 19:18:24]

xhwang@chromium.org changed reviewers:
+ peter@chromium.org
- mkwst@chromium.org

[xhwang, 2016-09-19 19:18:24]

s/mkwst/peter since mkwst is OOO.

peter@chromium.org: Please OWNERS review changes in content/shell/*

[xhwang, 2016-09-20 16:28:07]

On 2016/09/19 19:18:24, xhwang (slow) wrote:
> s/mkwst/peter since mkwst is OOO.
> 
> mailto:peter@chromium.org: Please OWNERS review changes in content/shell/*

peter@: kindly ping :)

[Peter Beverloo, 2016-09-22 11:55:23]

//content/shell lgtm

https://codereview.chromium.org/2268283003/diff/120001/content/browser/media/...
File content/browser/media/encrypted_media_browsertest.cc (right):

https://codereview.chromium.org/2268283003/diff/120001/content/browser/media/...
content/browser/media/encrypted_media_browsertest.cc:231: // On Android,
External Clear Key is supported in shell/ by using mojo CDM
nit: consider using //content/shell/ instead of just shell/, since we have
various kinds of shells (e.g. Chrome Sync Shell, Extension Shell, and so on)

https://codereview.chromium.org/2268283003/diff/120001/content/shell/renderer...
File content/shell/renderer/shell_content_renderer_client.cc (right):

https://codereview.chromium.org/2268283003/diff/120001/content/shell/renderer...
content/shell/renderer/shell_content_renderer_client.cc:141: if
(!base::FeatureList::IsEnabled(media::kExternalClearKeyForTesting))
Out of interest, why do you use a base::Feature for this flag since it's only
meant for testing? You really only need the command line aspect of it?

[xhwang, 2016-09-22 16:54:53]

The CQ bit was checked by xhwang@chromium.org to run a CQ dry run

[xhwang, 2016-09-22 16:54:58]

https://codereview.chromium.org/2268283003/diff/120001/content/browser/media/...
File content/browser/media/encrypted_media_browsertest.cc (right):

https://codereview.chromium.org/2268283003/diff/120001/content/browser/media/...
content/browser/media/encrypted_media_browsertest.cc:231: // On Android,
External Clear Key is supported in shell/ by using mojo CDM
On 2016/09/22 11:55:23, Peter Beverloo wrote:
> nit: consider using //content/shell/ instead of just shell/, since we have
> various kinds of shells (e.g. Chrome Sync Shell, Extension Shell, and so on)

Good to know. Done.

https://codereview.chromium.org/2268283003/diff/120001/content/shell/renderer...
File content/shell/renderer/shell_content_renderer_client.cc (right):

https://codereview.chromium.org/2268283003/diff/120001/content/shell/renderer...
content/shell/renderer/shell_content_renderer_client.cc:141: if
(!base::FeatureList::IsEnabled(media::kExternalClearKeyForTesting))
On 2016/09/22 11:55:23, Peter Beverloo wrote:
> Out of interest, why do you use a base::Feature for this flag since it's only
> meant for testing? You really only need the command line aspect of it?

Yeah, I wanted to add a command line flag but was told that all
--enable/--disable flags should be a base::Feature now.

[commit-bot: I haz the power, 2016-09-22 16:55:09]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-09-22 18:00:44]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-09-22 18:00:45]

Dry run: Try jobs failed on following builders:
  linux_chromium_rel_ng on master.tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/linux_chromium_...)

[xhwang, 2016-09-22 21:51:23]

The CQ bit was checked by xhwang@chromium.org

[xhwang, 2016-09-22 21:51:24]

The patchset sent to the CQ was uploaded after l-g-t-m from
jrummell@chromium.org, ddorwin@chromium.org, peter@chromium.org
Link to the patchset:
https://chromiumcodereview.appspot.com/2268283003/#ps160001 (title: "nits")

[commit-bot: I haz the power, 2016-09-22 21:51:49]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-09-22 23:36:18]

Description was changed from

==========
media: Add External Clear Key content browser test on Android

This CL adds External Clear Key support on Android for testing.
This is implemented by running AesDecryptor in the MojoCdmService,
which runs in the MojoMediaApplication in the GPU process. This is
only enabled when kExternalClearKeyForTesting feature is enabled.

A new content browser test is added to use External Clear Key on
Android. Since AesDecryptor doesn't support decoding, the media
pipeline is configured to do decrypt-only using the mojo CDM
(MojoDecryptor on MojoCdm), and then use the normal Android
pipeline (AVDA/MojoAudioDecoder) to decode encrypted audio/video.

Note that this is different from the default mode how Android
media pipeline works for encrypted content (decryption/decoding
both happens in the GPU process). When browser_tests are enabled
on Android, we should be able to have test coverage on that.

Here's what the new test covers:
- MojoCdm / MojoCdmService
- MojoDecryptor / MojoDecryptorService
- MojoMediaApplication
- Connection to services hosted in MojoMediaApplication through
  MojoShellContext.

This CL also fixes a bug in MojoCdm where the decryptor_ptr is bound to
one thread but is used on another thread.

BUG=581893
TEST=This CL adds a new test. No other functionality change.
==========

to

==========
media: Add External Clear Key content browser test on Android

This CL adds External Clear Key support on Android for testing.
This is implemented by running AesDecryptor in the MojoCdmService,
which runs in the MojoMediaApplication in the GPU process. This is
only enabled when kExternalClearKeyForTesting feature is enabled.

A new content browser test is added to use External Clear Key on
Android. Since AesDecryptor doesn't support decoding, the media
pipeline is configured to do decrypt-only using the mojo CDM
(MojoDecryptor on MojoCdm), and then use the normal Android
pipeline (AVDA/MojoAudioDecoder) to decode encrypted audio/video.

Note that this is different from the default mode how Android
media pipeline works for encrypted content (decryption/decoding
both happens in the GPU process). When browser_tests are enabled
on Android, we should be able to have test coverage on that.

Here's what the new test covers:
- MojoCdm / MojoCdmService
- MojoDecryptor / MojoDecryptorService
- MojoMediaApplication
- Connection to services hosted in MojoMediaApplication through
  MojoShellContext.

This CL also fixes a bug in MojoCdm where the decryptor_ptr is bound to
one thread but is used on another thread.

BUG=581893
TEST=This CL adds a new test. No other functionality change.
==========

[commit-bot: I haz the power, 2016-09-22 23:36:19]

Committed patchset #9 (id:160001)

[commit-bot: I haz the power, 2016-09-22 23:38:05]

Description was changed from

==========
media: Add External Clear Key content browser test on Android

This CL adds External Clear Key support on Android for testing.
This is implemented by running AesDecryptor in the MojoCdmService,
which runs in the MojoMediaApplication in the GPU process. This is
only enabled when kExternalClearKeyForTesting feature is enabled.

A new content browser test is added to use External Clear Key on
Android. Since AesDecryptor doesn't support decoding, the media
pipeline is configured to do decrypt-only using the mojo CDM
(MojoDecryptor on MojoCdm), and then use the normal Android
pipeline (AVDA/MojoAudioDecoder) to decode encrypted audio/video.

Note that this is different from the default mode how Android
media pipeline works for encrypted content (decryption/decoding
both happens in the GPU process). When browser_tests are enabled
on Android, we should be able to have test coverage on that.

Here's what the new test covers:
- MojoCdm / MojoCdmService
- MojoDecryptor / MojoDecryptorService
- MojoMediaApplication
- Connection to services hosted in MojoMediaApplication through
  MojoShellContext.

This CL also fixes a bug in MojoCdm where the decryptor_ptr is bound to
one thread but is used on another thread.

BUG=581893
TEST=This CL adds a new test. No other functionality change.
==========

to

==========
media: Add External Clear Key content browser test on Android

This CL adds External Clear Key support on Android for testing.
This is implemented by running AesDecryptor in the MojoCdmService,
which runs in the MojoMediaApplication in the GPU process. This is
only enabled when kExternalClearKeyForTesting feature is enabled.

A new content browser test is added to use External Clear Key on
Android. Since AesDecryptor doesn't support decoding, the media
pipeline is configured to do decrypt-only using the mojo CDM
(MojoDecryptor on MojoCdm), and then use the normal Android
pipeline (AVDA/MojoAudioDecoder) to decode encrypted audio/video.

Note that this is different from the default mode how Android
media pipeline works for encrypted content (decryption/decoding
both happens in the GPU process). When browser_tests are enabled
on Android, we should be able to have test coverage on that.

Here's what the new test covers:
- MojoCdm / MojoCdmService
- MojoDecryptor / MojoDecryptorService
- MojoMediaApplication
- Connection to services hosted in MojoMediaApplication through
  MojoShellContext.

This CL also fixes a bug in MojoCdm where the decryptor_ptr is bound to
one thread but is used on another thread.

BUG=581893
TEST=This CL adds a new test. No other functionality change.

Committed: https://crrev.com/bcd6858191102c4d4c949a617560231908480235
Cr-Commit-Position: refs/heads/master@{#420500}
==========

[commit-bot: I haz the power, 2016-09-22 23:38:07]

Patchset 9 (id:??) landed as
https://crrev.com/bcd6858191102c4d4c949a617560231908480235
Cr-Commit-Position: refs/heads/master@{#420500}