Add SurfaceTexture::release and call it from AVDA.

Add SurfaceTexture::release and call it from AVDA.

SurfaceTexture::release() discards unused buffers and tears down the
SurfaceTexture, leaving only the front buffer if it's bound to a GL
texture.  This CL plumbs that down to native.

This CL also uses it for AVDA, so that back buffers don't take up
memory when the deferred strategy is shut down.

As part of that, it splits strategy cleanup into two parts.  The
first is called before flush, while the latter is called after.
This prevents transitioning the SurfaceTexture into the released
state while the flush is still trying to talk to MediaCodec.  It's
unclear that would be safe.

BUG=619658
Committed: https://crrev.com/2c238bbd17a40008c3d0a2b437f36fd8b0455f0e
Cr-Commit-Position: refs/heads/master@{#402187}

[liberato (no reviews please), 2016-06-24 17:40:11]

Description was changed from

==========
Add SurfaceTexture::release and call it from AVDA.

SurfaceTexture::release() discards unused buffers and tears down the
SurfaceTexture, leaving only the front buffer if it's bound to a GL
texture.  This CL plumbs that down to native.

This CL also uses it for AVDA, so that back buffers don't take up
memory when the deferred strategy is shut down.

As part of that, it splits strategy cleanup into two parts.  The
first is called before flush, while the latter is called after.
This prevents transitioning the SurfaceTexture into the released
state while the flush is still trying to talk to MediaCodec.  It's
unclear that would be safe.

BUG=619658
==========

to

==========
Add SurfaceTexture::release and call it from AVDA.

SurfaceTexture::release() discards unused buffers and tears down the
SurfaceTexture, leaving only the front buffer if it's bound to a GL
texture.  This CL plumbs that down to native.

This CL also uses it for AVDA, so that back buffers don't take up
memory when the deferred strategy is shut down.

As part of that, it splits strategy cleanup into two parts.  The
first is called before flush, while the latter is called after.
This prevents transitioning the SurfaceTexture into the released
state while the flush is still trying to talk to MediaCodec.  It's
unclear that would be safe.

BUG=619658
==========

[liberato (no reviews please), 2016-06-24 17:40:12]

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

[liberato (no reviews please), 2016-06-24 17:40:20]

https://codereview.chromium.org/2095873005/diff/1/media/gpu/android_deferred_...
File media/gpu/android_deferred_rendering_backing_strategy.cc (right):

https://codereview.chromium.org/2095873005/diff/1/media/gpu/android_deferred_...
media/gpu/android_deferred_rendering_backing_strategy.cc:83:
surface_texture_->ReleaseSurfaceTexture();
i was on the fence about whether we could just do this in Cleanup(), before the
flush completes.  probably it's safe, since we won't be releasing with
rendering.  i was worried that MediaCodec might be doing something weird with
SurfaceTexture back buffers that would quit working.

it definitely seems safe to do after.

[watk, 2016-06-24 22:04:01]

Nice! I assume you tested this manually by backgrounding chrome, etc? lgtm

[liberato (no reviews please), 2016-06-24 23:09:44]

The CQ bit was checked by liberato@chromium.org

[commit-bot: I haz the power, 2016-06-24 23:10:18]

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, 2016-06-24 23:18:46]

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

[commit-bot: I haz the power, 2016-06-24 23:18:47]

Try jobs failed on following builders:
  chromium_presubmit on master.tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/chromium_presub...)

[liberato (no reviews please), 2016-06-24 23:20:19]

liberato@chromium.org changed reviewers:
+ sievers@chromium.org

[liberato (no reviews please), 2016-06-24 23:24:59]

sievers, PTAL at surface_texture.* and SurfaceTexturePlatformWrapper.java .  i
forgot that i modified non-media stuff.

watk: i let it suspend after timeout while paused.

thanks
-fl

[no sievers, 2016-06-25 00:03:15]

On 2016/06/24 23:24:59, liberato wrote:
> sievers, PTAL at surface_texture.* and SurfaceTexturePlatformWrapper.java .  i
> forgot that i modified non-media stuff.
> 
> watk: i let it suspend after timeout while paused.
> 
> thanks
> -fl

lgtm

[liberato (no reviews please), 2016-06-25 00:11:25]

The CQ bit was checked by liberato@chromium.org

[commit-bot: I haz the power, 2016-06-25 00:11:43]

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, 2016-06-25 00:41:13]

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

[commit-bot: I haz the power, 2016-06-25 00:41:14]

Try jobs failed on following builders:
  win_chromium_rel_ng on master.tryserver.chromium.win (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.win/builders/win_chromium_rel_...)

[liberato (no reviews please), 2016-06-27 14:19:49]

The CQ bit was checked by liberato@chromium.org

[commit-bot: I haz the power, 2016-06-27 14:20:03]

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, 2016-06-27 15:41:06]

Description was changed from

==========
Add SurfaceTexture::release and call it from AVDA.

SurfaceTexture::release() discards unused buffers and tears down the
SurfaceTexture, leaving only the front buffer if it's bound to a GL
texture.  This CL plumbs that down to native.

This CL also uses it for AVDA, so that back buffers don't take up
memory when the deferred strategy is shut down.

As part of that, it splits strategy cleanup into two parts.  The
first is called before flush, while the latter is called after.
This prevents transitioning the SurfaceTexture into the released
state while the flush is still trying to talk to MediaCodec.  It's
unclear that would be safe.

BUG=619658
==========

to

==========
Add SurfaceTexture::release and call it from AVDA.

SurfaceTexture::release() discards unused buffers and tears down the
SurfaceTexture, leaving only the front buffer if it's bound to a GL
texture.  This CL plumbs that down to native.

This CL also uses it for AVDA, so that back buffers don't take up
memory when the deferred strategy is shut down.

As part of that, it splits strategy cleanup into two parts.  The
first is called before flush, while the latter is called after.
This prevents transitioning the SurfaceTexture into the released
state while the flush is still trying to talk to MediaCodec.  It's
unclear that would be safe.

BUG=619658
==========

[commit-bot: I haz the power, 2016-06-27 15:41:07]

Committed patchset #1 (id:1)

[commit-bot: I haz the power, 2016-06-27 15:45:00]

Description was changed from

==========
Add SurfaceTexture::release and call it from AVDA.

SurfaceTexture::release() discards unused buffers and tears down the
SurfaceTexture, leaving only the front buffer if it's bound to a GL
texture.  This CL plumbs that down to native.

This CL also uses it for AVDA, so that back buffers don't take up
memory when the deferred strategy is shut down.

As part of that, it splits strategy cleanup into two parts.  The
first is called before flush, while the latter is called after.
This prevents transitioning the SurfaceTexture into the released
state while the flush is still trying to talk to MediaCodec.  It's
unclear that would be safe.

BUG=619658
==========

to

==========
Add SurfaceTexture::release and call it from AVDA.

SurfaceTexture::release() discards unused buffers and tears down the
SurfaceTexture, leaving only the front buffer if it's bound to a GL
texture.  This CL plumbs that down to native.

This CL also uses it for AVDA, so that back buffers don't take up
memory when the deferred strategy is shut down.

As part of that, it splits strategy cleanup into two parts.  The
first is called before flush, while the latter is called after.
This prevents transitioning the SurfaceTexture into the released
state while the flush is still trying to talk to MediaCodec.  It's
unclear that would be safe.

BUG=619658

Committed: https://crrev.com/2c238bbd17a40008c3d0a2b437f36fd8b0455f0e
Cr-Commit-Position: refs/heads/master@{#402187}
==========

[commit-bot: I haz the power, 2016-06-27 15:45:01]

Patchset 1 (id:??) landed as
https://crrev.com/2c238bbd17a40008c3d0a2b437f36fd8b0455f0e
Cr-Commit-Position: refs/heads/master@{#402187}