Implement fallback from WMPI to WMPA.

Implement fallback from WMPI to WMPA.

With this change, RenderFrameImpl::createMediaPlayer() will select between WMPI and WMPA on Android. Specifically, WMPI will be used unless:
  - --disable-accelerated-video-decode is set.
  - MediaCodec is not supported (because the API level is too low or the device is blacklisted).
  - The URL indicates that the video is HLS.

Additionally, --disable-accelerated-video-decode is now automatically set whenever any of these flags are set:
  --disable-gpu
  --single-process
  --in-process-gpu

Since in these modes VDAs do not currently work (and these flags are not passed to the renderer).

BUG=516765
Committed: https://crrev.com/bcbd85c1c6067e05da5245b7fdabb084c48d5c32
Cr-Commit-Position: refs/heads/master@{#368475}

[sandersd (OOO until July 31), 2016-01-06 00:56:19]

sandersd@chromium.org changed reviewers:
+ watk@chromium.org

[sandersd (OOO until July 31), 2016-01-06 00:56:20]

https://codereview.chromium.org/1560953002/diff/1/content/browser/gpu/gpu_dat...
File content/browser/gpu/gpu_data_manager_impl_private.cc (right):

https://codereview.chromium.org/1560953002/diff/1/content/browser/gpu/gpu_dat...
content/browser/gpu/gpu_data_manager_impl_private.cc:920: bool
GpuDataManagerImplPrivate::ShouldDisableAcceleratedVideoDecode(
I'm not certain that this is the best place.

https://codereview.chromium.org/1560953002/diff/1/content/browser/gpu/gpu_dat...
content/browser/gpu/gpu_data_manager_impl_private.cc:937: if
(command_line->HasSwitch(switches::kDisableGpu) ||
This one should probably apply to all platforms?

[watk, 2016-01-06 01:17:21]

https://codereview.chromium.org/1560953002/diff/20001/content/renderer/render...
File content/renderer/render_frame_impl.cc (right):

https://codereview.chromium.org/1560953002/diff/20001/content/renderer/render...
content/renderer/render_frame_impl.cc:2326: static bool IsHLSURL(GURL& url) {
The style guide says refs must be const
(https://google.github.io/styleguide/cppguide.html#Reference_Arguments)

Did you consider making the isHLSURL in webmediaplayerandroid.cc visible and
calling that instead? Or otherwise sharing the implementation?

https://codereview.chromium.org/1560953002/diff/20001/content/renderer/render...
content/renderer/render_frame_impl.cc:2378: GURL gurl = url;
I'm guessing you can delete this and pass url directly if you make isHLSURL take
a const ref.

[sandersd (OOO until July 31), 2016-01-06 01:27:01]

https://codereview.chromium.org/1560953002/diff/20001/content/renderer/render...
File content/renderer/render_frame_impl.cc (right):

https://codereview.chromium.org/1560953002/diff/20001/content/renderer/render...
content/renderer/render_frame_impl.cc:2326: static bool IsHLSURL(GURL& url) {
On 2016/01/06 01:17:21, watk wrote:
> The style guide says refs must be const
> (https://google.github.io/styleguide/cppguide.html#Reference_Arguments)
> 
> Did you consider making the isHLSURL in webmediaplayerandroid.cc visible and
> calling that instead? Or otherwise sharing the implementation?

Done.

https://codereview.chromium.org/1560953002/diff/20001/content/renderer/render...
content/renderer/render_frame_impl.cc:2378: GURL gurl = url;
On 2016/01/06 01:17:20, watk wrote:
> I'm guessing you can delete this and pass url directly if you make isHLSURL
take
> a const ref. 

And the world makes sense once again. (Done)

[sandersd (OOO until July 31), 2016-01-06 20:05:19]

sandersd@chromium.org changed reviewers:
+ hubbe@chromium.org

[watk, 2016-01-06 20:06:48]

lgtm

https://codereview.chromium.org/1560953002/diff/40001/content/renderer/media/...
File content/renderer/media/android/webmediaplayer_android.cc (right):

https://codereview.chromium.org/1560953002/diff/40001/content/renderer/media/...
content/renderer/media/android/webmediaplayer_android.cc:189: 
extra newline or subtle style rule?

[hubbe, 2016-01-06 20:11:24]

lgtm

[sandersd (OOO until July 31), 2016-01-06 20:12:13]

sandersd@chromium.org changed reviewers:
+ dalecurtis@chromium.org

[sandersd (OOO until July 31), 2016-01-06 20:12:13]

+dalecurtis for good measure

[DaleCurtis, 2016-01-06 21:03:22]

Needs a meaningful description about what this is doing and why. Are there any
places you can add some tests for this?

https://codereview.chromium.org/1560953002/diff/60001/content/browser/gpu/gpu...
File content/browser/gpu/gpu_data_manager_impl_private.cc (right):

https://codereview.chromium.org/1560953002/diff/60001/content/browser/gpu/gpu...
content/browser/gpu/gpu_data_manager_impl_private.cc:943: if
(command_line->HasSwitch(switches::kDisableGpu) ||
This is going to mean no webview support since they use --single-process AFAIK.
--in-process-gpu is used by Android One variants I think. Others? I don't think
these are reasonable restrictions, so if you want to land this now these need to
be marked with a TODO() fix and have a bug attached.

Why is this necessary at all? Won't we just fallback if possible during
initialize since the factories would be null in the gpu decoding case.

https://codereview.chromium.org/1560953002/diff/60001/content/renderer/render...
File content/renderer/render_frame_impl.cc (right):

https://codereview.chromium.org/1560953002/diff/60001/content/renderer/render...
content/renderer/render_frame_impl.cc:2370:
WebMediaPlayerAndroid::IsHLSPath(url)) {
Should this function be in MediaCodecUtil instead?

[sandersd (OOO until July 31), 2016-01-06 22:20:38]

https://codereview.chromium.org/1560953002/diff/60001/content/browser/gpu/gpu...
File content/browser/gpu/gpu_data_manager_impl_private.cc (right):

https://codereview.chromium.org/1560953002/diff/60001/content/browser/gpu/gpu...
content/browser/gpu/gpu_data_manager_impl_private.cc:943: if
(command_line->HasSwitch(switches::kDisableGpu) ||
On 2016/01/06 21:03:22, DaleCurtis wrote:
> This is going to mean no webview support since they use --single-process
AFAIK.
> --in-process-gpu is used by Android One variants I think. Others? I don't
think
> these are reasonable restrictions, so if you want to land this now these need
to
> be marked with a TODO() fix and have a bug attached.

I agree, but I don't think this is going to get fixed soon. TODO added.

> Why is this necessary at all? Won't we just fallback if possible during
> initialize since the factories would be null in the gpu decoding case.

No, by that time we already picked WMPI and it's too late.

https://codereview.chromium.org/1560953002/diff/60001/content/renderer/render...
File content/renderer/render_frame_impl.cc (right):

https://codereview.chromium.org/1560953002/diff/60001/content/renderer/render...
content/renderer/render_frame_impl.cc:2370:
WebMediaPlayerAndroid::IsHLSPath(url)) {
On 2016/01/06 21:03:22, DaleCurtis wrote:
> Should this function be in MediaCodecUtil instead?

Done.

[DaleCurtis, 2016-01-06 22:54:40]

lgtm % comment q.

https://codereview.chromium.org/1560953002/diff/80001/content/browser/gpu/gpu...
File content/browser/gpu/gpu_data_manager_impl_private.cc (right):

https://codereview.chromium.org/1560953002/diff/80001/content/browser/gpu/gpu...
content/browser/gpu/gpu_data_manager_impl_private.cc:944: // --in-process-gpu,
at least on Android (necessary to support WebView).
Does WebView only use --in-process-gpu and not --single-process?

[DaleCurtis, 2016-01-06 22:54:53]

(Still needs a beefier description though)

[sandersd (OOO until July 31), 2016-01-07 00:08:38]

https://codereview.chromium.org/1560953002/diff/80001/content/browser/gpu/gpu...
File content/browser/gpu/gpu_data_manager_impl_private.cc (right):

https://codereview.chromium.org/1560953002/diff/80001/content/browser/gpu/gpu...
content/browser/gpu/gpu_data_manager_impl_private.cc:944: // --in-process-gpu,
at least on Android (necessary to support WebView).
On 2016/01/06 22:54:40, DaleCurtis wrote:
> Does WebView only use --in-process-gpu and not --single-process?

WebView may use either, depending on other settings and the device. They are
slightly different though, in --single-process mode compositing is synchronous
(which I understand to be an exposed API feature of WebView).

[sandersd (OOO until July 31), 2016-01-07 00:16:11]

Description was changed from

==========
Implement MediaPlayer fallback.

BUG=516765
==========

to

==========
Implement fallback from WMPI to WMPA.

With this change, RenderFrameImpl::createMediaPlayer() will select between WMPI
and WMPA on Android. Specifically, WMPI will be used unless:
  - --disable-accelerated-video-decode is set.
  - MediaCodec is not supported (because the API level is too low or the device
is blacklisted).
  - The URL indicates that the video is HLS.

Additionally, --disable-accelerated-video-decode is now automatically set
whenever any of these flags are set:
  --disable-gpu
  --single-process
  --in-process-gpu

Since in these modes VDAs do not currently work (and these flags are not passed
to the renderer).

BUG=516765
==========

[sandersd (OOO until July 31), 2016-01-07 00:19:37]

sandersd@chromium.org changed reviewers:
+ jochen@chromium.org

[sandersd (OOO until July 31), 2016-01-07 00:19:37]

jochen@chromium.org: Please review changes in content/

[jochen (gone - plz use gerrit), 2016-01-08 14:19:57]

lgtm

[sandersd (OOO until July 31), 2016-01-08 21:12:50]

The CQ bit was checked by sandersd@chromium.org

[sandersd (OOO until July 31), 2016-01-08 21:12:50]

The patchset sent to the CQ was uploaded after l-g-t-m from watk@chromium.org,
hubbe@chromium.org
Link to the patchset: https://codereview.chromium.org/1560953002/#ps80001
(title: "Move IsHLS* to MediaCodecUtil")

[commit-bot: I haz the power, 2016-01-08 21:13:35]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1560953002/80001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1560953002/80001

[commit-bot: I haz the power, 2016-01-08 21:24:08]

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

[commit-bot: I haz the power, 2016-01-08 21:24:09]

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

[sandersd (OOO until July 31), 2016-01-08 21:52:00]

The CQ bit was checked by sandersd@chromium.org

[sandersd (OOO until July 31), 2016-01-08 21:52:00]

The patchset sent to the CQ was uploaded after l-g-t-m from watk@chromium.org,
jochen@chromium.org, hubbe@chromium.org, dalecurtis@chromium.org
Link to the patchset: https://codereview.chromium.org/1560953002/#ps100001
(title: "Only include MediaCodecUtil on Android.")

[commit-bot: I haz the power, 2016-01-08 21:52:44]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1560953002/100001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1560953002/100001

[sandersd (OOO until July 31), 2016-01-08 22:29:26]

The CQ bit was unchecked by sandersd@chromium.org

[sandersd (OOO until July 31), 2016-01-08 22:42:12]

The CQ bit was checked by sandersd@chromium.org

[commit-bot: I haz the power, 2016-01-08 22:44:19]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1560953002/100001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1560953002/100001

[commit-bot: I haz the power, 2016-01-09 00:39:25]

Description was changed from

==========
Implement fallback from WMPI to WMPA.

With this change, RenderFrameImpl::createMediaPlayer() will select between WMPI
and WMPA on Android. Specifically, WMPI will be used unless:
  - --disable-accelerated-video-decode is set.
  - MediaCodec is not supported (because the API level is too low or the device
is blacklisted).
  - The URL indicates that the video is HLS.

Additionally, --disable-accelerated-video-decode is now automatically set
whenever any of these flags are set:
  --disable-gpu
  --single-process
  --in-process-gpu

Since in these modes VDAs do not currently work (and these flags are not passed
to the renderer).

BUG=516765
==========

to

==========
Implement fallback from WMPI to WMPA.

With this change, RenderFrameImpl::createMediaPlayer() will select between WMPI
and WMPA on Android. Specifically, WMPI will be used unless:
  - --disable-accelerated-video-decode is set.
  - MediaCodec is not supported (because the API level is too low or the device
is blacklisted).
  - The URL indicates that the video is HLS.

Additionally, --disable-accelerated-video-decode is now automatically set
whenever any of these flags are set:
  --disable-gpu
  --single-process
  --in-process-gpu

Since in these modes VDAs do not currently work (and these flags are not passed
to the renderer).

BUG=516765
==========

[commit-bot: I haz the power, 2016-01-09 00:39:26]

Committed patchset #6 (id:100001)

[commit-bot: I haz the power, 2016-01-09 00:40:47]

Description was changed from

==========
Implement fallback from WMPI to WMPA.

With this change, RenderFrameImpl::createMediaPlayer() will select between WMPI
and WMPA on Android. Specifically, WMPI will be used unless:
  - --disable-accelerated-video-decode is set.
  - MediaCodec is not supported (because the API level is too low or the device
is blacklisted).
  - The URL indicates that the video is HLS.

Additionally, --disable-accelerated-video-decode is now automatically set
whenever any of these flags are set:
  --disable-gpu
  --single-process
  --in-process-gpu

Since in these modes VDAs do not currently work (and these flags are not passed
to the renderer).

BUG=516765
==========

to

==========
Implement fallback from WMPI to WMPA.

With this change, RenderFrameImpl::createMediaPlayer() will select between WMPI
and WMPA on Android. Specifically, WMPI will be used unless:
  - --disable-accelerated-video-decode is set.
  - MediaCodec is not supported (because the API level is too low or the device
is blacklisted).
  - The URL indicates that the video is HLS.

Additionally, --disable-accelerated-video-decode is now automatically set
whenever any of these flags are set:
  --disable-gpu
  --single-process
  --in-process-gpu

Since in these modes VDAs do not currently work (and these flags are not passed
to the renderer).

BUG=516765

Committed: https://crrev.com/bcbd85c1c6067e05da5245b7fdabb084c48d5c32
Cr-Commit-Position: refs/heads/master@{#368475}
==========

[commit-bot: I haz the power, 2016-01-09 00:40:48]

Patchset 6 (id:??) landed as
https://crrev.com/bcbd85c1c6067e05da5245b7fdabb084c48d5c32
Cr-Commit-Position: refs/heads/master@{#368475}