Add virtual layout test for GPU accelerated decoding.

Add gpu accelerated config changes to SmokeTests.

SmokeTests are the subset of layout tests run on the android bots. This change
adds coverage to several historically problematic config-change scenarios.

These tests use a virtual suite that sets the flag "--use-gpu-in-tests". This
flag overrides the default test behavior of using MesaGL, which is not capable
of creating surfaces to use with gpu accelerated decoding (at least on Android).

This change also fixes some test/expectations rot.

BUG=555703
TESTS=Config change LayoutTests.

Committed: https://crrev.com/b8bf5eba6b6cd3623f16404c746251bf662d141d
Committed: https://crrev.com/b0a4caeabd005d3e5d3be751d52e1133758bed28
Cr-Original-Commit-Position: refs/heads/master@{#407938}
Cr-Commit-Position: refs/heads/master@{#408266}

[chcunningham, 2016-01-27 22:13:12]

chcunningham@chromium.org changed reviewers:
+ crouleau@chromium.org, wolenetz@chromium.org

[chcunningham, 2016-01-27 22:13:12]

These tests will probably still fail on the bots because I think ffmpeg branding
needs to be set to Chrome (not Chromium) in order for us to have the mp4
demuxers. 

This is is an issue to clarify with legal :)

[wolenetz, 2016-01-27 22:24:39]

I'm unfamiliar with the mapping of the virtual test suites:
1) For non-Android platforms, will the two additional flags always be set for
any test under .../http/tests/media/media-source?
2) For [Android] MSE bugs listed in the LayoutTests/TestExpectations file, why
don't they all now have the virtual test suite prefix?
3) How are virtual test suite tests enabled? I was expecting at least some
change to SmokeTests.

[wolenetz, 2016-02-24 22:20:22]

Answered a couple inline to myself :).
Also a couple new questions:

On 2016/01/27 22:24:39, wolenetz wrote:
> I'm unfamiliar with the mapping of the virtual test suites:
> 1) For non-Android platforms, will the two additional flags always be set for
> any test under .../http/tests/media/media-source?

Only for when the virtual/mse_unified_media_pipeline/... ones. And yes, for all
platforms within that virtual suite.

> 2) For [Android] MSE bugs listed in the LayoutTests/TestExpectations file, why
> don't they all now have the virtual test suite prefix?

Simple. Because Spitzer improves their behavior vs pre-Spitzer.

What's the long-term plan for the virtual test suite: is it expected to remain
while we have Android behavior conditioned by --enable-unified-media-pipeline?


> 3) How are virtual test suite tests enabled? I was expecting at least some
> change to SmokeTests.

Will such a SmokeTests change be in a subsequent CL?

[wolenetz, 2016-02-24 23:20:04]

From chat with chcunningham@

LGTM under the following assumptions:

Addition of the spitzer virtual suite test(s) to SmokeTests (and appropriate
Android platform-specific expectations) will come in a later CL.

Note that this CL adds H/W accel (gl) Spitzer (unified-media-pipeline)
testability, but only for manual runs currently, and without any expectations
update (which might be needed).

[wolenetz, 2016-02-24 23:22:55]

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

[commit-bot: I haz the power, 2016-02-24 23:25:35]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1644763002/1
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1644763002/1

[commit-bot: I haz the power, 2016-02-25 02:38:56]

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

[commit-bot: I haz the power, 2016-02-25 02:38:57]

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

[chcunningham, 2016-03-02 20:41:32]

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

[commit-bot: I haz the power, 2016-03-02 20:42:23]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1644763002/1
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1644763002/1

[commit-bot: I haz the power, 2016-03-02 21:42:54]

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

[commit-bot: I haz the power, 2016-03-02 21:42:55]

Dry run: This issue passed the CQ dry run.

[chcunningham, 2016-03-29 00:01:50]

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

[commit-bot: I haz the power, 2016-03-29 00:02:13]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1644763002/20001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1644763002/20001

[commit-bot: I haz the power, 2016-03-29 00:45:12]

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

[commit-bot: I haz the power, 2016-03-29 00:45:13]

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

[chcunningham, 2016-07-16 00:57:51]

https://codereview.chromium.org/1644763002/diff/40001/third_party/WebKit/Layo...
File
third_party/WebKit/LayoutTests/virtual/media-gpu-accelerated/http/tests/media/media-source/README.txt
(right):

https://codereview.chromium.org/1644763002/diff/40001/third_party/WebKit/Layo...
third_party/WebKit/LayoutTests/virtual/media-gpu-accelerated/http/tests/media/media-source/README.txt:3:
The media-gpu-accelerated prefix is common to a handful of virtual tests that
Note to self - fix this comment if we stick with one test (likely)

https://codereview.chromium.org/1644763002/diff/40001/third_party/WebKit/Layo...
third_party/WebKit/LayoutTests/virtual/media-gpu-accelerated/http/tests/media/media-source/README.txt:24:
WHY ADD THESE TESTS TO SMOKETESTS
Note to self - CL description is probably a better place for this.

[chcunningham, 2016-07-16 19:01:34]

Description was changed from

==========
Add VirtualTestSuite for MSE with unified media pipeline.

Suite adds 2 command line flags:
--enable-unified-media-pipeline - make android use desktop media path
--use-gl=egl - override test behavior to use egl instead of mesagl.

Mesa is not compatible with android's SurfaceTexture, using in HW
accelerated video decoding.

With these flags, all tests are expected to pass excpet for
[Android] mediasource-config-change-mp4-v-framerate.html

A fix for this test should come soon.

BUG=555703
==========

to

==========
Create media-gpu-accelerated virtual test suite. Also update the
SmokeTests to use the new virtual tests, fix test rot, and update
test expectations.

This suite sets the flag "--use-gpu-in-tests" which is made to 
override the default test behavior of using MesaGL. MesaGL is not 
compatible Android's SurfaceTexture. 


Rationale for SmokeTests changes:

mediasource-config-change-mp4-av-video-framesize.html:
Tests mid-stream config changes. This is core to adaptive playback
and historically problematic.

mediasource-config-change-mp4-av-audio-bitrate.html:
Android's audio driver is distinct from desktop platforms so this test was
chosen to give coverage to that path. 

mediasource-config-change-mp4-v-framerate.html:
Test adaptive playback use case, varying the framerate of the video 
midstream. Its also nice coverage of video-without-audio playback.

BUG=555703
==========

[chcunningham, 2016-07-16 19:03:00]

Bots should all be happy now. I changed the mediasourc-play test to come from
the virtual test suite since this apparently uses gpu acceleration on the
android_blink_rel bot (while it uses libvpx on the nexus4 bot).

https://codereview.chromium.org/1644763002/diff/40001/third_party/WebKit/Layo...
File
third_party/WebKit/LayoutTests/virtual/media-gpu-accelerated/http/tests/media/media-source/README.txt
(right):

https://codereview.chromium.org/1644763002/diff/40001/third_party/WebKit/Layo...
third_party/WebKit/LayoutTests/virtual/media-gpu-accelerated/http/tests/media/media-source/README.txt:3:
The media-gpu-accelerated prefix is common to a handful of virtual tests that
On 2016/07/16 00:57:51, chcunningham wrote:
> Note to self - fix this comment if we stick with one test (likely)

Done.

https://codereview.chromium.org/1644763002/diff/40001/third_party/WebKit/Layo...
third_party/WebKit/LayoutTests/virtual/media-gpu-accelerated/http/tests/media/media-source/README.txt:24:
WHY ADD THESE TESTS TO SMOKETESTS
On 2016/07/16 00:57:51, chcunningham wrote:
> Note to self - CL description is probably a better place for this.

Done.

[chcunningham, 2016-07-16 19:03:26]

Description was changed from

==========
Create media-gpu-accelerated virtual test suite. Also update the
SmokeTests to use the new virtual tests, fix test rot, and update
test expectations.

This suite sets the flag "--use-gpu-in-tests" which is made to 
override the default test behavior of using MesaGL. MesaGL is not 
compatible Android's SurfaceTexture. 


Rationale for SmokeTests changes:

mediasource-config-change-mp4-av-video-framesize.html:
Tests mid-stream config changes. This is core to adaptive playback
and historically problematic.

mediasource-config-change-mp4-av-audio-bitrate.html:
Android's audio driver is distinct from desktop platforms so this test was
chosen to give coverage to that path. 

mediasource-config-change-mp4-v-framerate.html:
Test adaptive playback use case, varying the framerate of the video 
midstream. Its also nice coverage of video-without-audio playback.

BUG=555703
==========

to

==========
Create media-gpu-accelerated virtual test suite. Also update the
SmokeTests to use the new virtual tests, fix test rot, and update
test expectations.

This suite sets the flag "--use-gpu-in-tests" which is made to 
override the default test behavior of using MesaGL. MesaGL is not 
compatible Android's SurfaceTexture. 


Rationale for new SmokeTests:

mediasource-config-change-mp4-av-video-framesize.html:
Tests mid-stream config changes. This is core to adaptive playback
and historically problematic.

mediasource-config-change-mp4-av-audio-bitrate.html:
Android's audio driver is distinct from desktop platforms so this test was
chosen to give coverage to that path. 

mediasource-config-change-mp4-v-framerate.html:
Test adaptive playback use case, varying the framerate of the video 
midstream. Its also nice coverage of video-without-audio playback.

BUG=555703
==========

[chcunningham, 2016-07-16 19:03:43]

Description was changed from

==========
Create media-gpu-accelerated virtual test suite. Also update the
SmokeTests to use the new virtual tests, fix test rot, and update
test expectations.

This suite sets the flag "--use-gpu-in-tests" which is made to 
override the default test behavior of using MesaGL. MesaGL is not 
compatible Android's SurfaceTexture. 


Rationale for new SmokeTests:

mediasource-config-change-mp4-av-video-framesize.html:
Tests mid-stream config changes. This is core to adaptive playback
and historically problematic.

mediasource-config-change-mp4-av-audio-bitrate.html:
Android's audio driver is distinct from desktop platforms so this test was
chosen to give coverage to that path. 

mediasource-config-change-mp4-v-framerate.html:
Test adaptive playback use case, varying the framerate of the video 
midstream. Its also nice coverage of video-without-audio playback.

BUG=555703
==========

to

==========
Create media-gpu-accelerated virtual test suite. Also update the
SmokeTests to use the new virtual tests, fix test rot, and update
test expectations.

This suite sets the flag "--use-gpu-in-tests" which is made to 
override the default test behavior of using MesaGL. MesaGL is not 
compatible Android's SurfaceTexture. 


Rationale for new SmokeTests:

mediasource-config-change-mp4-av-video-framesize.html:
Tests mid-stream config changes. This is core to adaptive playback
and historically problematic.

mediasource-config-change-mp4-av-audio-bitrate.html:
Android's audio driver is distinct from desktop platforms so this 
test was chosen to give coverage to that path. 

mediasource-config-change-mp4-v-framerate.html:
Test adaptive playback use case, varying the framerate of the video 
midstream. Its also nice coverage of video-without-audio playback.

BUG=555703
==========

[chcunningham, 2016-07-16 19:04:09]

Description was changed from

==========
Create media-gpu-accelerated virtual test suite. Also update the
SmokeTests to use the new virtual tests, fix test rot, and update
test expectations.

This suite sets the flag "--use-gpu-in-tests" which is made to 
override the default test behavior of using MesaGL. MesaGL is not 
compatible Android's SurfaceTexture. 


Rationale for new SmokeTests:

mediasource-config-change-mp4-av-video-framesize.html:
Tests mid-stream config changes. This is core to adaptive playback
and historically problematic.

mediasource-config-change-mp4-av-audio-bitrate.html:
Android's audio driver is distinct from desktop platforms so this 
test was chosen to give coverage to that path. 

mediasource-config-change-mp4-v-framerate.html:
Test adaptive playback use case, varying the framerate of the video 
midstream. Its also nice coverage of video-without-audio playback.

BUG=555703
==========

to

==========
Create media-gpu-accelerated virtual test suite. Also update the
SmokeTests to use the new virtual tests, fix test rot, and update
test expectations.

This suite sets the flag "--use-gpu-in-tests" which is made to 
override the default test behavior of using MesaGL. MesaGL is not 
compatible Android's SurfaceTexture. 


Rationale for new SmokeTests:

mediasource-config-change-mp4-av-video-framesize.html:
 Tests mid-stream config changes. This is core to adaptive playback
 and historically problematic.

mediasource-config-change-mp4-av-audio-bitrate.html:
 Android's audio driver is distinct from desktop platforms so this 
 test was chosen to give coverage to that path. 

mediasource-config-change-mp4-v-framerate.html:
 Test adaptive playback use case, varying the framerate of the video 
 midstream. Its also nice coverage of video-without-audio playback.

BUG=555703
==========

[chcunningham, 2016-07-16 20:19:39]

On 2016/02/24 23:20:04, wolenetz_OoO_July_15PM wrote:
> From chat with chcunningham@
> 
> LGTM under the following assumptions:
> 
> Addition of the spitzer virtual suite test(s) to SmokeTests (and appropriate
> Android platform-specific expectations) will come in a later CL.

Went ahead and did in this CL - the rot stops now :) 

> Note that this CL adds H/W accel (gl) Spitzer (unified-media-pipeline)
> testability, but only for manual runs currently, and without any expectations
> update (which might be needed).

I've confirmed that the expectations (and their platform specific versions) are
inherited by the virtual suite.

[chcunningham, 2016-07-17 18:38:43]

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

[commit-bot: I haz the power, 2016-07-17 18:38:50]

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, 2016-07-17 20:54:50]

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

[commit-bot: I haz the power, 2016-07-17 20:54:51]

Dry run: This issue passed the CQ dry run.

[chcunningham, 2016-07-19 20:03:51]

Matt pls take a look. Would love to land ASAP before it rots further.

[wolenetz, 2016-07-19 20:47:10]

looking pretty good; one question:

https://codereview.chromium.org/1644763002/diff/60001/third_party/WebKit/Layo...
File third_party/WebKit/LayoutTests/TestExpectations (right):

https://codereview.chromium.org/1644763002/diff/60001/third_party/WebKit/Layo...
third_party/WebKit/LayoutTests/TestExpectations:738: # These tests use gpu
accelerated decoding and need special flags to disable Mesa GL. See
media-gpu-accelerated VirtualTestSuites.
Why crash? Can we make these tests have platform-specific and flag-specific
failure expectations without crashing?
desktop: should pass
android with flags in virtual suite: should pass
android without flags (no virtual suite): should fail (but not crash)

Is this possible?

[chcunningham, 2016-07-19 23:21:38]

We chatted. For posterity:

> android without flags (no virtual suite): should fail (but not crash)

This is not possible ATM. They've added a check somewhere in AVDA that asserts
that the surface has certain needed qualities (I forget where exactly) and mesa
textures fail the check. Still, this crash encountered only if you manually run
these tests on android without the virtual suite flags - no bots will be doing
this.

[chcunningham, 2016-07-19 23:21:44]

The CQ bit was checked by chcunningham@chromium.org

[chcunningham, 2016-07-19 23:21:44]

The patchset sent to the CQ was uploaded after l-g-t-m from
wolenetz@chromium.org
Link to the patchset: https://codereview.chromium.org/1644763002/#ps60001
(title: "Use virtual mediasource-play in SmokeTests.")

[commit-bot: I haz the power, 2016-07-19 23:22:30]

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

[wolenetz, 2016-07-19 23:23:07]

On 2016/07/19 23:21:38, chcunningham wrote:
> We chatted. For posterity:
> 
> > android without flags (no virtual suite): should fail (but not crash)
> 
> This is not possible ATM. They've added a check somewhere in AVDA that asserts
> that the surface has certain needed qualities (I forget where exactly) and
mesa
> textures fail the check. Still, this crash encountered only if you manually
run
> these tests on android without the virtual suite flags - no bots will be doing
> this.

Ack LGTM

[commit-bot: I haz the power, 2016-07-19 23:33:22]

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

[commit-bot: I haz the power, 2016-07-19 23:33:23]

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...)

[chcunningham, 2016-07-19 23:40:07]

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

[chcunningham, 2016-07-19 23:40:08]

+jochen for content/

[chcunningham, 2016-07-19 23:43:00]

chcunningham@chromium.org changed reviewers:
+ dpranke@chromium.org

[chcunningham, 2016-07-19 23:43:01]

+dpranke for SmokeTests

[jochen (gone - plz use gerrit), 2016-07-20 11:05:56]

please format the CL description to match
https://sites.google.com/a/chromium.org/dev/developers/contributing-code?pli=...

[jochen (gone - plz use gerrit), 2016-07-20 11:06:16]

with that addressed, content/ lgtm

[chcunningham, 2016-07-20 19:09:17]

Description was changed from

==========
Create media-gpu-accelerated virtual test suite. Also update the
SmokeTests to use the new virtual tests, fix test rot, and update
test expectations.

This suite sets the flag "--use-gpu-in-tests" which is made to 
override the default test behavior of using MesaGL. MesaGL is not 
compatible Android's SurfaceTexture. 


Rationale for new SmokeTests:

mediasource-config-change-mp4-av-video-framesize.html:
 Tests mid-stream config changes. This is core to adaptive playback
 and historically problematic.

mediasource-config-change-mp4-av-audio-bitrate.html:
 Android's audio driver is distinct from desktop platforms so this 
 test was chosen to give coverage to that path. 

mediasource-config-change-mp4-v-framerate.html:
 Test adaptive playback use case, varying the framerate of the video 
 midstream. Its also nice coverage of video-without-audio playback.

BUG=555703
==========

to

==========
Add gpu accelerated config changes to SmokeTests.

SmokeTests are the subset of layout tests run on the android bots. This change
adds coverage to several historically problematic config-change scenarios.

These tests use a virtual suite that sets the flag "--use-gpu-in-tests". This 
flag overrides the default test behavior of using MesaGL, which is not capable
of creating surfaces to use with gpu accelerated decoding (at least on Android).

This change also fixes some test/expectations rot.

BUG=555703
TESTS=Config change LayoutTests.
==========

[chcunningham, 2016-07-20 19:09:38]

On 2016/07/20 11:06:16, jochen wrote:
> with that addressed, content/ lgtm

Done, thanks.

[chcunningham, 2016-07-21 17:32:59]

Dirk, can you take a look?

[CalebRouleau, 2016-07-21 18:29:41]

crouleau@chromium.org changed reviewers:
- crouleau@chromium.org

[Dirk Pranke, 2016-07-21 20:45:05]

https://codereview.chromium.org/1644763002/diff/60001/content/browser/gpu/gpu...
File content/browser/gpu/gpu_process_host.cc (right):

https://codereview.chromium.org/1644763002/diff/60001/content/browser/gpu/gpu...
content/browser/gpu/gpu_process_host.cc:111:
switches::kDisableAcceleratedVideoDecode,
is this indentation change correct?

https://codereview.chromium.org/1644763002/diff/60001/third_party/WebKit/Layo...
File third_party/WebKit/LayoutTests/TestExpectations (right):

https://codereview.chromium.org/1644763002/diff/60001/third_party/WebKit/Layo...
third_party/WebKit/LayoutTests/TestExpectations:738: # These tests use gpu
accelerated decoding and need special flags to disable Mesa GL. See
media-gpu-accelerated VirtualTestSuites.
On 2016/07/19 20:47:10, wolenetz wrote:
> Why crash? Can we make these tests have platform-specific and flag-specific
> failure expectations without crashing?
> desktop: should pass
> android with flags in virtual suite: should pass
> android without flags (no virtual suite): should fail (but not crash)
> 
> Is this possible?

If they're always going to fail, why run them at all?

[chcunningham, 2016-07-21 21:04:36]

https://codereview.chromium.org/1644763002/diff/60001/content/browser/gpu/gpu...
File content/browser/gpu/gpu_process_host.cc (right):

https://codereview.chromium.org/1644763002/diff/60001/content/browser/gpu/gpu...
content/browser/gpu/gpu_process_host.cc:111:
switches::kDisableAcceleratedVideoDecode,
On 2016/07/21 20:45:05, Dirk Pranke wrote:
> is this indentation change correct?

This was what clang-format generated. I vote we go with it.

https://codereview.chromium.org/1644763002/diff/60001/third_party/WebKit/Layo...
File third_party/WebKit/LayoutTests/TestExpectations (right):

https://codereview.chromium.org/1644763002/diff/60001/third_party/WebKit/Layo...
third_party/WebKit/LayoutTests/TestExpectations:738: # These tests use gpu
accelerated decoding and need special flags to disable Mesa GL. See
media-gpu-accelerated VirtualTestSuites.
On 2016/07/21 20:45:05, Dirk Pranke wrote:
> On 2016/07/19 20:47:10, wolenetz wrote:
> > Why crash? Can we make these tests have platform-specific and flag-specific
> > failure expectations without crashing?
> > desktop: should pass
> > android with flags in virtual suite: should pass
> > android without flags (no virtual suite): should fail (but not crash)
> > 
> > Is this possible?
> 
> If they're always going to fail, why run them at all?

They will always fail, but they aren't actually being run anywhere - this serves
to document what is expected if someone attempts to manually run. The android
bots will instead run a subset of these via the virtual test suite - and those
are all expected to pass.

[Dirk Pranke, 2016-07-21 21:07:31]

https://codereview.chromium.org/1644763002/diff/60001/third_party/WebKit/Layo...
File third_party/WebKit/LayoutTests/TestExpectations (right):

https://codereview.chromium.org/1644763002/diff/60001/third_party/WebKit/Layo...
third_party/WebKit/LayoutTests/TestExpectations:738: # These tests use gpu
accelerated decoding and need special flags to disable Mesa GL. See
media-gpu-accelerated VirtualTestSuites.
On 2016/07/21 21:04:36, chcunningham wrote:
> On 2016/07/21 20:45:05, Dirk Pranke wrote:
> > On 2016/07/19 20:47:10, wolenetz wrote:
> > > Why crash? Can we make these tests have platform-specific and
flag-specific
> > > failure expectations without crashing?
> > > desktop: should pass
> > > android with flags in virtual suite: should pass
> > > android without flags (no virtual suite): should fail (but not crash)
> > > 
> > > Is this possible?
> > 
> > If they're always going to fail, why run them at all?
> 
> They will always fail, but they aren't actually being run anywhere - this
serves
> to document what is expected if someone attempts to manually run. The android
> bots will instead run a subset of these via the virtual test suite - and those
> are all expected to pass.

Ah, then move these entries to NeverFixTests.

[chcunningham, 2016-07-22 22:12:59]

On 2016/07/21 21:07:31, Dirk Pranke wrote:
>
https://codereview.chromium.org/1644763002/diff/60001/third_party/WebKit/Layo...
> File third_party/WebKit/LayoutTests/TestExpectations (right):
> 
>
https://codereview.chromium.org/1644763002/diff/60001/third_party/WebKit/Layo...
> third_party/WebKit/LayoutTests/TestExpectations:738: # These tests use gpu
> accelerated decoding and need special flags to disable Mesa GL. See
> media-gpu-accelerated VirtualTestSuites.
> On 2016/07/21 21:04:36, chcunningham wrote:
> > On 2016/07/21 20:45:05, Dirk Pranke wrote:
> > > On 2016/07/19 20:47:10, wolenetz wrote:
> > > > Why crash? Can we make these tests have platform-specific and
> flag-specific
> > > > failure expectations without crashing?
> > > > desktop: should pass
> > > > android with flags in virtual suite: should pass
> > > > android without flags (no virtual suite): should fail (but not crash)
> > > > 
> > > > Is this possible?
> > > 
> > > If they're always going to fail, why run them at all?
> > 
> > They will always fail, but they aren't actually being run anywhere - this
> serves
> > to document what is expected if someone attempts to manually run. The
android
> > bots will instead run a subset of these via the virtual test suite - and
those
> > are all expected to pass.
> 
> Ah, then move these entries to NeverFixTests.

Done

[chcunningham, 2016-07-22 22:13:09]

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

[commit-bot: I haz the power, 2016-07-22 22:14:03]

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, 2016-07-22 23:56:01]

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

[commit-bot: I haz the power, 2016-07-22 23:56:02]

Dry run: 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_...)

[chcunningham, 2016-07-25 18:49:30]

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

[chcunningham, 2016-07-25 18:49:47]

Dirk, lgty?

[commit-bot: I haz the power, 2016-07-25 18:50:03]

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, 2016-07-25 19:11:15]

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

[commit-bot: I haz the power, 2016-07-25 19:11:16]

Dry run: Try jobs failed on following builders:
  android_arm64_dbg_recipe on master.tryserver.chromium.android (JOB_FAILED,
https://build.chromium.org/p/tryserver.chromium.android/builders/android_arm6...)

[Dirk Pranke, 2016-07-25 21:03:29]

lgtm, though I still think the indentation change in the gpu file is weird :).

[chcunningham, 2016-07-25 23:15:47]

The CQ bit was checked by chcunningham@chromium.org

[chcunningham, 2016-07-25 23:15:47]

The patchset sent to the CQ was uploaded after l-g-t-m from jochen@chromium.org,
wolenetz@chromium.org, dpranke@chromium.org
Link to the patchset: https://codereview.chromium.org/1644763002/#ps120001
(title: "Fix formatting for new addition to switches")

[commit-bot: I haz the power, 2016-07-25 23:16:31]

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

[chcunningham, 2016-07-25 23:16:50]

Thanks everyone :)

On 2016/07/25 21:03:29, Dirk Pranke wrote:
> lgtm, though I still think the indentation change in the gpu file is weird :).

Just checked it again (new switch was added) -- confirmed clang-format prefers
it this way.

[chcunningham, 2016-07-26 00:03:46]

The CQ bit was unchecked by chcunningham@chromium.org

[chcunningham, 2016-07-26 00:09:34]

The CQ bit was checked by chcunningham@chromium.org

[chcunningham, 2016-07-26 00:09:35]

The patchset sent to the CQ was uploaded after l-g-t-m from
dpranke@chromium.org, jochen@chromium.org, wolenetz@chromium.org
Link to the patchset: https://codereview.chromium.org/1644763002/#ps140001
(title: "Removing enabled-unified-media-pipeline flag - no longer
needed/meaningful")

[commit-bot: I haz the power, 2016-07-26 00:10:04]

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, 2016-07-26 02:13:07]

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

[commit-bot: I haz the power, 2016-07-26 02:13:09]

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_...)

[chcunningham, 2016-07-26 18:46:38]

The CQ bit was checked by chcunningham@chromium.org

[commit-bot: I haz the power, 2016-07-26 18:47:14]

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, 2016-07-26 22:30:04]

Description was changed from

==========
Add gpu accelerated config changes to SmokeTests.

SmokeTests are the subset of layout tests run on the android bots. This change
adds coverage to several historically problematic config-change scenarios.

These tests use a virtual suite that sets the flag "--use-gpu-in-tests". This 
flag overrides the default test behavior of using MesaGL, which is not capable
of creating surfaces to use with gpu accelerated decoding (at least on Android).

This change also fixes some test/expectations rot.

BUG=555703
TESTS=Config change LayoutTests.
==========

to

==========
Add gpu accelerated config changes to SmokeTests.

SmokeTests are the subset of layout tests run on the android bots. This change
adds coverage to several historically problematic config-change scenarios.

These tests use a virtual suite that sets the flag "--use-gpu-in-tests". This 
flag overrides the default test behavior of using MesaGL, which is not capable
of creating surfaces to use with gpu accelerated decoding (at least on Android).

This change also fixes some test/expectations rot.

BUG=555703
TESTS=Config change LayoutTests.
==========

[commit-bot: I haz the power, 2016-07-26 22:30:05]

Committed patchset #8 (id:140001)

[commit-bot: I haz the power, 2016-07-26 22:32:57]

Description was changed from

==========
Add gpu accelerated config changes to SmokeTests.

SmokeTests are the subset of layout tests run on the android bots. This change
adds coverage to several historically problematic config-change scenarios.

These tests use a virtual suite that sets the flag "--use-gpu-in-tests". This 
flag overrides the default test behavior of using MesaGL, which is not capable
of creating surfaces to use with gpu accelerated decoding (at least on Android).

This change also fixes some test/expectations rot.

BUG=555703
TESTS=Config change LayoutTests.
==========

to

==========
Add gpu accelerated config changes to SmokeTests.

SmokeTests are the subset of layout tests run on the android bots. This change
adds coverage to several historically problematic config-change scenarios.

These tests use a virtual suite that sets the flag "--use-gpu-in-tests". This
flag overrides the default test behavior of using MesaGL, which is not capable
of creating surfaces to use with gpu accelerated decoding (at least on Android).

This change also fixes some test/expectations rot.

BUG=555703
TESTS=Config change LayoutTests.

Committed: https://crrev.com/b8bf5eba6b6cd3623f16404c746251bf662d141d
Cr-Commit-Position: refs/heads/master@{#407938}
==========

[commit-bot: I haz the power, 2016-07-26 22:32:59]

Patchset 8 (id:??) landed as
https://crrev.com/b8bf5eba6b6cd3623f16404c746251bf662d141d
Cr-Commit-Position: refs/heads/master@{#407938}

[robliao, 2016-07-27 01:29:06]

A revert of this CL (patchset #8 id:140001) has been created in
https://codereview.chromium.org/2184003002/ by robliao@chromium.org.

The reason for reverting is: Suspected in breaking webkit tests:
https://build.chromium.org/p/chromium.webkit/builders/WebKit%20Linux%20MSAN/b...


unexpected_failures:
virtual/media-gpu-accelerated/http/tests/media/media-source/mediasource-sourcebufferlist.html
virtual/media-gpu-accelerated/http/tests/media/media-source/mediasource-appendstream-quota-exceeded.html
virtual/media-gpu-accelerated/http/tests/media/media-source/mediasource-getvideoplaybackquality.html
and more in mediasource-*.html.

[chcunningham, 2016-07-27 19:54:59]

Description was changed from

==========
Add gpu accelerated config changes to SmokeTests.

SmokeTests are the subset of layout tests run on the android bots. This change
adds coverage to several historically problematic config-change scenarios.

These tests use a virtual suite that sets the flag "--use-gpu-in-tests". This
flag overrides the default test behavior of using MesaGL, which is not capable
of creating surfaces to use with gpu accelerated decoding (at least on Android).

This change also fixes some test/expectations rot.

BUG=555703
TESTS=Config change LayoutTests.

Committed: https://crrev.com/b8bf5eba6b6cd3623f16404c746251bf662d141d
Cr-Commit-Position: refs/heads/master@{#407938}
==========

to

==========
Add gpu accelerated config changes to SmokeTests.

SmokeTests are the subset of layout tests run on the android bots. This change
adds coverage to several historically problematic config-change scenarios.

These tests use a virtual suite that sets the flag "--use-gpu-in-tests". This
flag overrides the default test behavior of using MesaGL, which is not capable
of creating surfaces to use with gpu accelerated decoding (at least on Android).

This change also fixes some test/expectations rot.

BUG=555703
TESTS=Config change LayoutTests.

Committed: https://crrev.com/b8bf5eba6b6cd3623f16404c746251bf662d141d
Cr-Commit-Position: refs/heads/master@{#407938}
==========

[chcunningham, 2016-07-27 20:05:04]

Now skipping the tests in msan. Turns out to be a known limitation for msan -
cannot easily instrument the gl library.

Also, this made me aware of "skip" option, which I've applied for the desktop
operating systems for these new virtual tests. Unlike android, desktop does not
currently have proprietary gpu accelerated codecs turned on, so there's no
benefit to running the virtual tests on those platforms.

[chcunningham, 2016-07-27 20:06:03]

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

[commit-bot: I haz the power, 2016-07-27 20:06:20]

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

[chcunningham, 2016-07-27 20:06:37]

Dirk, can you take a look at these latest PS. I'll assume the other reviewers
LGTM still stands.

[commit-bot: I haz the power, 2016-07-27 21:16:23]

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

[commit-bot: I haz the power, 2016-07-27 21:16:24]

Dry run: This issue passed the CQ dry run.

[Dirk Pranke, 2016-07-27 22:55:34]

lgtm

[wolenetz, 2016-07-27 22:59:38]

lgtm

[chcunningham1, 2016-07-27 23:06:24]

The CQ bit was checked by chcunningham@google.com

[chcunningham1, 2016-07-27 23:06:24]

The patchset sent to the CQ was uploaded after l-g-t-m from jochen@chromium.org
Link to the patchset: https://codereview.chromium.org/1644763002/#ps160001
(title: "Skip desktop virtual tests")

[commit-bot: I haz the power, 2016-07-27 23:07:12]

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, 2016-07-27 23:12:06]

Description was changed from

==========
Add gpu accelerated config changes to SmokeTests.

SmokeTests are the subset of layout tests run on the android bots. This change
adds coverage to several historically problematic config-change scenarios.

These tests use a virtual suite that sets the flag "--use-gpu-in-tests". This
flag overrides the default test behavior of using MesaGL, which is not capable
of creating surfaces to use with gpu accelerated decoding (at least on Android).

This change also fixes some test/expectations rot.

BUG=555703
TESTS=Config change LayoutTests.

Committed: https://crrev.com/b8bf5eba6b6cd3623f16404c746251bf662d141d
Cr-Commit-Position: refs/heads/master@{#407938}
==========

to

==========
Add gpu accelerated config changes to SmokeTests.

SmokeTests are the subset of layout tests run on the android bots. This change
adds coverage to several historically problematic config-change scenarios.

These tests use a virtual suite that sets the flag "--use-gpu-in-tests". This
flag overrides the default test behavior of using MesaGL, which is not capable
of creating surfaces to use with gpu accelerated decoding (at least on Android).

This change also fixes some test/expectations rot.

BUG=555703
TESTS=Config change LayoutTests.

Committed: https://crrev.com/b8bf5eba6b6cd3623f16404c746251bf662d141d
Cr-Commit-Position: refs/heads/master@{#407938}
==========

[commit-bot: I haz the power, 2016-07-27 23:12:09]

Committed patchset #9 (id:160001)

[commit-bot: I haz the power, 2016-07-27 23:14:33]

Description was changed from

==========
Add gpu accelerated config changes to SmokeTests.

SmokeTests are the subset of layout tests run on the android bots. This change
adds coverage to several historically problematic config-change scenarios.

These tests use a virtual suite that sets the flag "--use-gpu-in-tests". This
flag overrides the default test behavior of using MesaGL, which is not capable
of creating surfaces to use with gpu accelerated decoding (at least on Android).

This change also fixes some test/expectations rot.

BUG=555703
TESTS=Config change LayoutTests.

Committed: https://crrev.com/b8bf5eba6b6cd3623f16404c746251bf662d141d
Cr-Commit-Position: refs/heads/master@{#407938}
==========

to

==========
Add gpu accelerated config changes to SmokeTests.

SmokeTests are the subset of layout tests run on the android bots. This change
adds coverage to several historically problematic config-change scenarios.

These tests use a virtual suite that sets the flag "--use-gpu-in-tests". This
flag overrides the default test behavior of using MesaGL, which is not capable
of creating surfaces to use with gpu accelerated decoding (at least on Android).

This change also fixes some test/expectations rot.

BUG=555703
TESTS=Config change LayoutTests.

Committed: https://crrev.com/b8bf5eba6b6cd3623f16404c746251bf662d141d
Committed: https://crrev.com/b0a4caeabd005d3e5d3be751d52e1133758bed28
Cr-Original-Commit-Position: refs/heads/master@{#407938}
Cr-Commit-Position: refs/heads/master@{#408266}
==========

[commit-bot: I haz the power, 2016-07-27 23:14:35]

Patchset 9 (id:??) landed as
https://crrev.com/b0a4caeabd005d3e5d3be751d52e1133758bed28
Cr-Commit-Position: refs/heads/master@{#408266}

[chcunningham, 2016-07-28 22:16:24]

Description was changed from

==========
Add gpu accelerated config changes to SmokeTests.

SmokeTests are the subset of layout tests run on the android bots. This change
adds coverage to several historically problematic config-change scenarios.

These tests use a virtual suite that sets the flag "--use-gpu-in-tests". This
flag overrides the default test behavior of using MesaGL, which is not capable
of creating surfaces to use with gpu accelerated decoding (at least on Android).

This change also fixes some test/expectations rot.

BUG=555703
TESTS=Config change LayoutTests.

Committed: https://crrev.com/b8bf5eba6b6cd3623f16404c746251bf662d141d
Committed: https://crrev.com/b0a4caeabd005d3e5d3be751d52e1133758bed28
Cr-Original-Commit-Position: refs/heads/master@{#407938}
Cr-Commit-Position: refs/heads/master@{#408266}
==========

to

==========
Add gpu accelerated config changes to SmokeTests.

SmokeTests are the subset of layout tests run on the android bots. This change
adds coverage to several historically problematic config-change scenarios.

These tests use a virtual suite that sets the flag "--use-gpu-in-tests". This
flag overrides the default test behavior of using MesaGL, which is not capable
of creating surfaces to use with gpu accelerated decoding (at least on Android).

This change also fixes some test/expectations rot.

BUG=555703
TESTS=Config change LayoutTests.

Committed: https://crrev.com/b8bf5eba6b6cd3623f16404c746251bf662d141d
Committed: https://crrev.com/b0a4caeabd005d3e5d3be751d52e1133758bed28
Cr-Original-Commit-Position: refs/heads/master@{#407938}
Cr-Commit-Position: refs/heads/master@{#408266}
==========

[chcunningham, 2016-07-28 22:16:24]

chcunningham@chromium.org changed reviewers:
- chcunningham@google.com