Compute GpuDriverBugWorkarounds only one time in the GPU process

Compute GpuDriverBugWorkarounds only one time in the GPU process

Previously the extraction of "--gpu-driver-bug-workarounds" from the
command-line was done in FeatureInfo's constructor. Since a new FeatureInfo
is created for each ContextGroup this extraction was done again whereas
the result is constant.

This CL moved owner ship of Workarounds from FeatureInfo to
InProcessCommandBuffer when running in gpu process mode. But moved
to GpuChannelManager otherwise.

class GpuDriverBugWorkarounds has been added within new files
gpu_driver_bug_workarounds.h/.cc.

BUG=599964
R=kbr@chromium.org, piman@chromium.org, rjkroege@chromium.org, zmo@chromium.org
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_optional_gpu_tests_rel;tryserver.chromium.mac:mac_optional_gpu_tests_rel;tryserver.chromium.win:win_optional_gpu_tests_rel

Committed: https://crrev.com/dab49af5278797db0a8374f2a7f06c6c33b8158b
Cr-Commit-Position: refs/heads/master@{#387569}

Committed: https://crrev.com/57b96c478dda56160c512fd85ee4e0f7437b524d
Cr-Commit-Position: refs/heads/master@{#387899}

Committed: https://crrev.com/265b2e56348b55fa622275e67a9863ad588f5183
Cr-Commit-Position: refs/heads/master@{#388157}

[Julien Isorce Samsung, 2016-04-07 17:29:35]

Description was changed from

==========
Compute gpu driver bug workarounds only one time in the GPU process

Previously the extraction of the gpu driver bug workarounds from the
command-line was done in FeatureInfo's constructor. Since a new FeatureInfo
is created for each ContextGroup this extraction was done again whereas
the result is constant.

This CL moved owner ship of Workarounds from FeatureInfo to
InProcessCommandBuffer when running in gpu process mode. But moved
to GpuChildThread otherwise.

struct GpuDriverBugWorkarounds has been added within new files
gpu_driver_bug_workarounds.h/.cc.

The refactoring showed that it has now same live cycle as gpu::GpuPreferences.
The difference is that gpu::GpuDriverBugWorkarounds is initialized in
the GPU process whereas gpu::GpuPreferences is sent from Browser process
to GPU process through GpuMsg_Initialize.

BUG=599964

R=kbr@chromium.org, zmo@chromium.org, piman@chromium.org
==========

to

==========
Compute gpu driver bug workarounds only one time in the GPU process

Previously the extraction of the gpu driver bug workarounds from the
command-line was done in FeatureInfo's constructor. Since a new FeatureInfo
is created for each ContextGroup this extraction was done again whereas
the result is constant.

This CL moved owner ship of Workarounds from FeatureInfo to
InProcessCommandBuffer when running in gpu process mode. But moved
to GpuChildThread otherwise.

struct GpuDriverBugWorkarounds has been added within new files
gpu_driver_bug_workarounds.h/.cc.

The refactoring showed that it has now same live cycle as gpu::GpuPreferences.
The difference is that gpu::GpuDriverBugWorkarounds is initialized in
the GPU process whereas gpu::GpuPreferences is sent from Browser process
to GPU process through GpuMsg_Initialize.

BUG=599964

R=kbr@chromium.org, zmo@chromium.org, piman@chromium.org
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_optional_gpu_tests_rel;tryserver.chromium.mac:mac_optional_gpu_tests_rel;tryserver.chromium.win:win_optional_gpu_tests_rel
==========

[piman, 2016-04-12 04:38:44]

I wouldn't go out of my way to use the GpuDriverBugWorkarounds by reference,
rather than by copy. The copy is cheap (just an inline memcpy), whereas
accessing the field has an extra cost (extra indirection). But most importantly,
in a lot of tests, you would get away with just a default-initialized
GpuDriverBugWorkarounds, since those won't have a command-line that sets
workarounds anyway. So no need to add an extra field (to hold the
GpuDriverBugWorkarounds), you could just pass a default one to (e.g.)
FeatureInfo, if it holds a concrete field rather than a reference.

https://codereview.chromium.org/1871613002/diff/1/gpu/command_buffer/service/...
File gpu/command_buffer/service/feature_info.h (right):

https://codereview.chromium.org/1871613002/diff/1/gpu/command_buffer/service/...
gpu/command_buffer/service/feature_info.h:94: FeatureInfo(const
GpuDriverBugWorkarounds& gpu_driver_bug_workarounds);
nit: explicit

https://codereview.chromium.org/1871613002/diff/1/gpu/config/gpu_driver_bug_w...
File gpu/config/gpu_driver_bug_workarounds.cc (right):

https://codereview.chromium.org/1871613002/diff/1/gpu/config/gpu_driver_bug_w...
gpu/config/gpu_driver_bug_workarounds.cc:25: : nullptr;
If we get nullptr, we'll crash below, won't we?

https://codereview.chromium.org/1871613002/diff/1/gpu/config/gpu_driver_bug_w...
File gpu/config/gpu_driver_bug_workarounds.h (right):

https://codereview.chromium.org/1871613002/diff/1/gpu/config/gpu_driver_bug_w...
gpu/config/gpu_driver_bug_workarounds.h:22: static void
Initialize(GpuDriverBugWorkarounds& workarounds);
If you make this (and below) like:
static GpuDriverBugWorkarounds Create();
Then a lot of code you added becomes simpler (e.g. you can initialize fields
with initializers).

[Julien Isorce Samsung, 2016-04-12 16:11:33]

https://codereview.chromium.org/1871613002/diff/1/gpu/command_buffer/service/...
File gpu/command_buffer/service/feature_info.h (right):

https://codereview.chromium.org/1871613002/diff/1/gpu/command_buffer/service/...
gpu/command_buffer/service/feature_info.h:94: FeatureInfo(const
GpuDriverBugWorkarounds& gpu_driver_bug_workarounds);
On 2016/04/12 04:38:44, piman wrote:
> nit: explicit

Done.

https://codereview.chromium.org/1871613002/diff/1/gpu/config/gpu_driver_bug_w...
File gpu/config/gpu_driver_bug_workarounds.cc (right):

https://codereview.chromium.org/1871613002/diff/1/gpu/config/gpu_driver_bug_w...
gpu/config/gpu_driver_bug_workarounds.cc:25: : nullptr;
On 2016/04/12 04:38:44, piman wrote:
> If we get nullptr, we'll crash below, won't we?
yes it will crash. I spotted that as well but I left it to keep same behavior
since it was present in FeatureInfo::FeatureInfo(). Not sure why
base::CommandLine::InitializedForCurrentProcess() is called in the first place.
For unit tests ?

[Julien Isorce Samsung, 2016-04-12 16:13:29]

On 2016/04/12 04:38:44, piman wrote:
> I wouldn't go out of my way to use the GpuDriverBugWorkarounds by reference,
> rather than by copy. The copy is cheap (just an inline memcpy), whereas
> accessing the field has an extra cost (extra indirection). But most
importantly,
> in a lot of tests, you would get away with just a default-initialized
> GpuDriverBugWorkarounds, since those won't have a command-line that sets
> workarounds anyway. So no need to add an extra field (to hold the
> GpuDriverBugWorkarounds), you could just pass a default one to (e.g.)
> FeatureInfo, if it holds a concrete field rather than a reference.


>
https://codereview.chromium.org/1871613002/diff/1/gpu/config/gpu_driver_bug_w...
> gpu/config/gpu_driver_bug_workarounds.h:22: static void
> Initialize(GpuDriverBugWorkarounds& workarounds);
> If you make this (and below) like:
> static GpuDriverBugWorkarounds Create();
> Then a lot of code you added becomes simpler (e.g. you can initialize fields
> with initializers).

Thx for the review. Good idea for these 2 points. I'll address these remark in
next Patch Set.

[Julien Isorce Samsung, 2016-04-13 00:03:17]

On 2016/04/12 16:13:29, j.isorce wrote:
> On 2016/04/12 04:38:44, piman wrote:
> > I wouldn't go out of my way to use the GpuDriverBugWorkarounds by reference,
> > rather than by copy. The copy is cheap (just an inline memcpy), whereas
> > accessing the field has an extra cost (extra indirection). But most
> importantly,
> > in a lot of tests, you would get away with just a default-initialized
> > GpuDriverBugWorkarounds, since those won't have a command-line that sets
> > workarounds anyway. So no need to add an extra field (to hold the
> > GpuDriverBugWorkarounds), you could just pass a default one to (e.g.)
> > FeatureInfo, if it holds a concrete field rather than a reference.
> 
> 
> >
>
https://codereview.chromium.org/1871613002/diff/1/gpu/config/gpu_driver_bug_w...
> > gpu/config/gpu_driver_bug_workarounds.h:22: static void
> > Initialize(GpuDriverBugWorkarounds& workarounds);
> > If you make this (and below) like:
> > static GpuDriverBugWorkarounds Create();
> > Then a lot of code you added becomes simpler (e.g. you can initialize fields
> > with initializers).
> 
> Thx for the review. Good idea for these 2 points. I'll address these remark in
> next Patch Set.

Done. Indeed it has simplified the CL :) Though I do not like the fact that
sometimes I have to pass GpuDriverBugWorkarounds() to ContextGroup constructor.
Maybe I should pass GpuChannelManager* if this is possible.

[Julien Isorce Samsung, 2016-04-13 12:49:09]

On 2016/04/13 00:03:17, j.isorce wrote:
> On 2016/04/12 16:13:29, j.isorce wrote:
> > On 2016/04/12 04:38:44, piman wrote:
> > > I wouldn't go out of my way to use the GpuDriverBugWorkarounds by
reference,
> > > rather than by copy. The copy is cheap (just an inline memcpy), whereas
> > > accessing the field has an extra cost (extra indirection). But most
> > importantly,
> > > in a lot of tests, you would get away with just a default-initialized
> > > GpuDriverBugWorkarounds, since those won't have a command-line that sets
> > > workarounds anyway. So no need to add an extra field (to hold the
> > > GpuDriverBugWorkarounds), you could just pass a default one to (e.g.)
> > > FeatureInfo, if it holds a concrete field rather than a reference.
> > 
> > 
> > >
> >
>
https://codereview.chromium.org/1871613002/diff/1/gpu/config/gpu_driver_bug_w...
> > > gpu/config/gpu_driver_bug_workarounds.h:22: static void
> > > Initialize(GpuDriverBugWorkarounds& workarounds);
> > > If you make this (and below) like:
> > > static GpuDriverBugWorkarounds Create();
> > > Then a lot of code you added becomes simpler (e.g. you can initialize
fields
> > > with initializers).
> > 
> > Thx for the review. Good idea for these 2 points. I'll address these remark
in
> > next Patch Set.
> 
> Done. Indeed it has simplified the CL :) Though I do not like the fact that
> sometimes I have to pass GpuDriverBugWorkarounds() to ContextGroup
constructor.
> Maybe I should pass GpuChannelManager* if this is possible.

So in Patch Set 5, I removed GpuDriverBugWorkarounds param from ContextGroup
constructor. This is something I added in earlier Patch Set anyway. But now the
param FeatureInfo of ContextGroup constructor is required to be not null. The
caller has to initialize it and provide gpu driver workarounds at this time.

[piman, 2016-04-14 02:42:09]

LGTM. I think this is nice, and we may be able to eventually initialize the
workarounds on the browser side rather than passing command-line flags.

[piman, 2016-04-14 02:42:13]

The CQ bit was checked by piman@chromium.org

[commit-bot: I haz the power, 2016-04-14 02:42:33]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1871613002/80001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1871613002/80001

[commit-bot: I haz the power, 2016-04-14 02:55:25]

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

[commit-bot: I haz the power, 2016-04-14 02:55:26]

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

[Julien Isorce Samsung, 2016-04-14 09:51:11]

Description was changed from

==========
Compute gpu driver bug workarounds only one time in the GPU process

Previously the extraction of the gpu driver bug workarounds from the
command-line was done in FeatureInfo's constructor. Since a new FeatureInfo
is created for each ContextGroup this extraction was done again whereas
the result is constant.

This CL moved owner ship of Workarounds from FeatureInfo to
InProcessCommandBuffer when running in gpu process mode. But moved
to GpuChildThread otherwise.

struct GpuDriverBugWorkarounds has been added within new files
gpu_driver_bug_workarounds.h/.cc.

The refactoring showed that it has now same live cycle as gpu::GpuPreferences.
The difference is that gpu::GpuDriverBugWorkarounds is initialized in
the GPU process whereas gpu::GpuPreferences is sent from Browser process
to GPU process through GpuMsg_Initialize.

BUG=599964

R=kbr@chromium.org, zmo@chromium.org, piman@chromium.org
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_optional_gpu_tests_rel;tryserver.chromium.mac:mac_optional_gpu_tests_rel;tryserver.chromium.win:win_optional_gpu_tests_rel
==========

to

==========
Compute GpuDriverBugWorkarounds only one time in the GPU process

Previously the extraction of "--gpu-driver-bug-workarounds" from the
command-line was done in FeatureInfo's constructor. Since a new FeatureInfo
is created for each ContextGroup this extraction was done again whereas
the result is constant.

This CL moved owner ship of Workarounds from FeatureInfo to
InProcessCommandBuffer when running in gpu process mode. But moved
to GpuChannelManager otherwise.

class GpuDriverBugWorkarounds has been added within new files
gpu_driver_bug_workarounds.h/.cc.

BUG=599964

R=kbr@chromium.org, zmo@chromium.org, piman@chromium.org
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_optional_gpu_tests_rel;tryserver.chromium.mac:mac_optional_gpu_tests_rel;tryserver.chromium.win:win_optional_gpu_tests_rel
==========

[Julien Isorce Samsung, 2016-04-14 09:55:22]

I updated description to better match with "Patch Set 5". Also "Patch Set 6" I
just removed obsolete change in gpu/ipc/service/gpu_channel_test_common.h.

[Julien Isorce Samsung, 2016-04-14 09:55:37]

The CQ bit was checked by j.isorce@samsung.com to run a CQ dry run

[commit-bot: I haz the power, 2016-04-14 09:55:52]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1871613002/100001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1871613002/100001

[commit-bot: I haz the power, 2016-04-14 10:11:12]

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

[commit-bot: I haz the power, 2016-04-14 10:11:13]

Dry run: Try jobs failed on following builders:
  android_compile_dbg on tryserver.chromium.android (JOB_FAILED,
https://build.chromium.org/p/tryserver.chromium.android/builders/android_comp...)
  cast_shell_android on tryserver.chromium.android (JOB_FAILED,
https://build.chromium.org/p/tryserver.chromium.android/builders/cast_shell_a...)

[Julien Isorce Samsung, 2016-04-14 12:06:50]

The CQ bit was checked by j.isorce@samsung.com to run a CQ dry run

[commit-bot: I haz the power, 2016-04-14 12:07:02]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1871613002/120001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1871613002/120001

[commit-bot: I haz the power, 2016-04-14 12:11:07]

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

[commit-bot: I haz the power, 2016-04-14 12:11:08]

Dry run: Try jobs failed on following builders:
  android_chromium_gn_compile_dbg on tryserver.chromium.android (JOB_FAILED,
https://build.chromium.org/p/tryserver.chromium.android/builders/android_chro...)
  android_chromium_gn_compile_rel on tryserver.chromium.android (JOB_FAILED,
https://build.chromium.org/p/tryserver.chromium.android/builders/android_chro...)
  cast_shell_android on tryserver.chromium.android (JOB_FAILED,
https://build.chromium.org/p/tryserver.chromium.android/builders/cast_shell_a...)
  linux_chromium_clobber_rel_ng on tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/linux_chromium_...)
  linux_chromium_compile_dbg_ng on tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/linux_chromium_...)
  linux_chromium_gn_chromeos_rel on tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/linux_chromium_...)

[Julien Isorce Samsung, 2016-04-14 12:42:51]

The CQ bit was checked by j.isorce@samsung.com to run a CQ dry run

[commit-bot: I haz the power, 2016-04-14 12:43:01]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1871613002/140001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1871613002/140001

[commit-bot: I haz the power, 2016-04-14 12:47:00]

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

[commit-bot: I haz the power, 2016-04-14 12:47:01]

Dry run: Try jobs failed on following builders:
  android_chromium_gn_compile_dbg on tryserver.chromium.android (JOB_FAILED,
https://build.chromium.org/p/tryserver.chromium.android/builders/android_chro...)
  android_chromium_gn_compile_rel on tryserver.chromium.android (JOB_FAILED,
https://build.chromium.org/p/tryserver.chromium.android/builders/android_chro...)
  cast_shell_android on tryserver.chromium.android (JOB_FAILED,
https://build.chromium.org/p/tryserver.chromium.android/builders/cast_shell_a...)
  cast_shell_linux on tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/cast_shell_linu...)
  linux_chromium_clobber_rel_ng on tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/linux_chromium_...)
  linux_chromium_compile_dbg_ng on tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/linux_chromium_...)
  linux_chromium_gn_chromeos_rel on tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/linux_chromium_...)
  linux_chromium_rel_ng on tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/linux_chromium_...)

[Julien Isorce Samsung, 2016-04-14 13:07:10]

The CQ bit was checked by j.isorce@samsung.com to run a CQ dry run

[commit-bot: I haz the power, 2016-04-14 13:07:35]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1871613002/160001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1871613002/160001

[commit-bot: I haz the power, 2016-04-14 13:21:53]

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

[commit-bot: I haz the power, 2016-04-14 13:21:55]

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

[Julien Isorce Samsung, 2016-04-14 13:43:02]

The CQ bit was checked by j.isorce@samsung.com to run a CQ dry run

[commit-bot: I haz the power, 2016-04-14 13:43:25]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1871613002/180001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1871613002/180001

[commit-bot: I haz the power, 2016-04-14 14:44:13]

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

[commit-bot: I haz the power, 2016-04-14 14:44:15]

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

[Julien Isorce Samsung, 2016-04-14 14:54:10]

The CQ bit was checked by j.isorce@samsung.com to run a CQ dry run

[commit-bot: I haz the power, 2016-04-14 14:54:25]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1871613002/200001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1871613002/200001

[Julien Isorce Samsung, 2016-04-14 14:55:24]

Description was changed from

==========
Compute GpuDriverBugWorkarounds only one time in the GPU process

Previously the extraction of "--gpu-driver-bug-workarounds" from the
command-line was done in FeatureInfo's constructor. Since a new FeatureInfo
is created for each ContextGroup this extraction was done again whereas
the result is constant.

This CL moved owner ship of Workarounds from FeatureInfo to
InProcessCommandBuffer when running in gpu process mode. But moved
to GpuChannelManager otherwise.

class GpuDriverBugWorkarounds has been added within new files
gpu_driver_bug_workarounds.h/.cc.

BUG=599964

R=kbr@chromium.org, zmo@chromium.org, piman@chromium.org
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_optional_gpu_tests_rel;tryserver.chromium.mac:mac_optional_gpu_tests_rel;tryserver.chromium.win:win_optional_gpu_tests_rel
==========

to

==========
Compute GpuDriverBugWorkarounds only one time in the GPU process

Previously the extraction of "--gpu-driver-bug-workarounds" from the
command-line was done in FeatureInfo's constructor. Since a new FeatureInfo
is created for each ContextGroup this extraction was done again whereas
the result is constant.

This CL moved owner ship of Workarounds from FeatureInfo to
InProcessCommandBuffer when running in gpu process mode. But moved
to GpuChannelManager otherwise.

class GpuDriverBugWorkarounds has been added within new files
gpu_driver_bug_workarounds.h/.cc.

BUG=599964

R=kbr@chromium.org, piman@chromium.org, rjkroege@chromium.org, zmo@chromium.org
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_optional_gpu_tests_rel;tryserver.chromium.mac:mac_optional_gpu_tests_rel;tryserver.chromium.win:win_optional_gpu_tests_rel
==========

[Julien Isorce Samsung, 2016-04-14 14:55:24]

j.isorce@samsung.com changed reviewers:
+ rjkroege@chromium.org

[commit-bot: I haz the power, 2016-04-14 16:56:21]

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

[commit-bot: I haz the power, 2016-04-14 16:56:22]

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

[Julien Isorce Samsung, 2016-04-14 17:43:05]

On 2016/04/14 16:56:22, commit-bot: I haz the power wrote:
> Dry run: Try jobs failed on following builders:
>   linux_android_rel_ng on tryserver.chromium.android (JOB_FAILED,
>
https://build.chromium.org/p/tryserver.chromium.android/builders/linux_androi...)

So remaining error is a unit test failure in gl_tests on Android:

[==========] 183 tests ran.
[  PASSED  ] 182 tests.
[  FAILED  ] 1 test, listed below:
[  FAILED  ] GLChromiumFramebufferMultisampleTest.DrawAndResolve

This tests succeed on other try bots and on my local Linux machine. Also I
cannot spot anything obvious except this in gl_manager.cc:

void GLManager::SetupBaseContext() {
  if (use_count_) {
    #if defined(OS_ANDROID)
      base_share_group_ = new scoped_refptr<gfx::GLShareGroup>(
          new gfx::GLShareGroup);
      gfx::Size size(4, 4);
      base_surface_ = new scoped_refptr<gfx::GLSurface>(
          gfx::GLSurface::CreateOffscreenGLSurface(size));
      gfx::GpuPreference gpu_preference(gfx::PreferDiscreteGpu);
      base_context_ = new scoped_refptr<gfx::GLContext>(
          gfx::GLContext::CreateGLContext(base_share_group_->get(),
                                          base_surface_->get(),
                                          gpu_preference));
    #endif
  }
  ++use_count_;
}

It is done only on Android so maybe my patch does something wrong in that case.

Just in case, I submitted "Patch Set 12" which is just a rebase and I'll
re-trigger a dry run with that.

[Julien Isorce Samsung, 2016-04-14 17:43:29]

The CQ bit was checked by j.isorce@samsung.com to run a CQ dry run

[commit-bot: I haz the power, 2016-04-14 17:44:02]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1871613002/220001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1871613002/220001

[commit-bot: I haz the power, 2016-04-14 19:47:07]

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

[commit-bot: I haz the power, 2016-04-14 19:47:09]

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

[Julien Isorce Samsung, 2016-04-14 20:47:23]

The CQ bit was checked by j.isorce@samsung.com to run a CQ dry run

[commit-bot: I haz the power, 2016-04-14 20:47:51]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1871613002/260001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1871613002/260001

[Julien Isorce Samsung, 2016-04-14 22:14:56]

j.isorce@samsung.com changed reviewers:
+ ben@chromium.org, cpu@chromium.org

[Julien Isorce Samsung, 2016-04-14 22:14:56]

Hi Antoine, please have a look to the changes I made since your last review. So
from "Patch Set 6" (which should have been titled "Remove obsolete change in
gpu/ipc/service/gpu_channel_test_common.h")

Also I need a owner review for changes in components/mus/gles2.

Thx everyone.

[piman, 2016-04-14 22:25:22]

lgtm

[rjkroege, 2016-04-14 22:33:43]

mus: lgtm

[commit-bot: I haz the power, 2016-04-14 23:42:27]

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

[commit-bot: I haz the power, 2016-04-14 23:42:29]

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_...)

[Julien Isorce Samsung, 2016-04-15 08:57:17]

On 2016/04/14 23:42:29, commit-bot: I haz the power wrote:
> 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_...)

"Patch Set 16" should have been titled "Bring back setup of gpu driver bug
workarounds for GLTest.FeatureFlagsMatchCapabilities".

[Julien Isorce Samsung, 2016-04-15 08:57:34]

The CQ bit was checked by j.isorce@samsung.com to run a CQ dry run

[commit-bot: I haz the power, 2016-04-15 08:57:45]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1871613002/300001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1871613002/300001

[commit-bot: I haz the power, 2016-04-15 10:08:31]

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

[commit-bot: I haz the power, 2016-04-15 10:08:32]

Dry run: This issue passed the CQ dry run.

[Julien Isorce Samsung, 2016-04-15 10:20:58]

On 2016/04/15 10:08:32, commit-bot: I haz the power wrote:
> Dry run: This issue passed the CQ dry run.

Since the change from last review is minor:
https://codereview.chromium.org/1871613002/diff2/280001:300001/gpu/command_bu...
, I am going ahead to submit that CL to the CQ.

[Julien Isorce Samsung, 2016-04-15 10:21:27]

The CQ bit was checked by j.isorce@samsung.com

[Julien Isorce Samsung, 2016-04-15 10:21:28]

The patchset sent to the CQ was uploaded after l-g-t-m from
rjkroege@chromium.org, piman@chromium.org
Link to the patchset: https://codereview.chromium.org/1871613002/#ps300001
(title: " ")

[commit-bot: I haz the power, 2016-04-15 10:21:39]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1871613002/300001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1871613002/300001

[commit-bot: I haz the power, 2016-04-15 10:26:15]

Description was changed from

==========
Compute GpuDriverBugWorkarounds only one time in the GPU process

Previously the extraction of "--gpu-driver-bug-workarounds" from the
command-line was done in FeatureInfo's constructor. Since a new FeatureInfo
is created for each ContextGroup this extraction was done again whereas
the result is constant.

This CL moved owner ship of Workarounds from FeatureInfo to
InProcessCommandBuffer when running in gpu process mode. But moved
to GpuChannelManager otherwise.

class GpuDriverBugWorkarounds has been added within new files
gpu_driver_bug_workarounds.h/.cc.

BUG=599964

R=kbr@chromium.org, piman@chromium.org, rjkroege@chromium.org, zmo@chromium.org
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_optional_gpu_tests_rel;tryserver.chromium.mac:mac_optional_gpu_tests_rel;tryserver.chromium.win:win_optional_gpu_tests_rel
==========

to

==========
Compute GpuDriverBugWorkarounds only one time in the GPU process

Previously the extraction of "--gpu-driver-bug-workarounds" from the
command-line was done in FeatureInfo's constructor. Since a new FeatureInfo
is created for each ContextGroup this extraction was done again whereas
the result is constant.

This CL moved owner ship of Workarounds from FeatureInfo to
InProcessCommandBuffer when running in gpu process mode. But moved
to GpuChannelManager otherwise.

class GpuDriverBugWorkarounds has been added within new files
gpu_driver_bug_workarounds.h/.cc.

BUG=599964

R=kbr@chromium.org, piman@chromium.org, rjkroege@chromium.org, zmo@chromium.org
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_optional_gpu_tests_rel;tryserver.chromium.mac:mac_optional_gpu_tests_rel;tryserver.chromium.win:win_optional_gpu_tests_rel
==========

[commit-bot: I haz the power, 2016-04-15 10:26:17]

Committed patchset #16 (id:300001)

[commit-bot: I haz the power, 2016-04-15 10:27:56]

Description was changed from

==========
Compute GpuDriverBugWorkarounds only one time in the GPU process

Previously the extraction of "--gpu-driver-bug-workarounds" from the
command-line was done in FeatureInfo's constructor. Since a new FeatureInfo
is created for each ContextGroup this extraction was done again whereas
the result is constant.

This CL moved owner ship of Workarounds from FeatureInfo to
InProcessCommandBuffer when running in gpu process mode. But moved
to GpuChannelManager otherwise.

class GpuDriverBugWorkarounds has been added within new files
gpu_driver_bug_workarounds.h/.cc.

BUG=599964

R=kbr@chromium.org, piman@chromium.org, rjkroege@chromium.org, zmo@chromium.org
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_optional_gpu_tests_rel;tryserver.chromium.mac:mac_optional_gpu_tests_rel;tryserver.chromium.win:win_optional_gpu_tests_rel
==========

to

==========
Compute GpuDriverBugWorkarounds only one time in the GPU process

Previously the extraction of "--gpu-driver-bug-workarounds" from the
command-line was done in FeatureInfo's constructor. Since a new FeatureInfo
is created for each ContextGroup this extraction was done again whereas
the result is constant.

This CL moved owner ship of Workarounds from FeatureInfo to
InProcessCommandBuffer when running in gpu process mode. But moved
to GpuChannelManager otherwise.

class GpuDriverBugWorkarounds has been added within new files
gpu_driver_bug_workarounds.h/.cc.

BUG=599964

R=kbr@chromium.org, piman@chromium.org, rjkroege@chromium.org, zmo@chromium.org
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_optional_gpu_tests_rel;tryserver.chromium.mac:mac_optional_gpu_tests_rel;tryserver.chromium.win:win_optional_gpu_tests_rel

Committed: https://crrev.com/dab49af5278797db0a8374f2a7f06c6c33b8158b
Cr-Commit-Position: refs/heads/master@{#387569}
==========

[commit-bot: I haz the power, 2016-04-15 10:27:58]

Patchset 16 (id:??) landed as
https://crrev.com/dab49af5278797db0a8374f2a7f06c6c33b8158b
Cr-Commit-Position: refs/heads/master@{#387569}

[Corentin Wallez, 2016-04-15 13:28:38]

A revert of this CL (patchset #16 id:300001) has been created in
https://codereview.chromium.org/1892483004/ by cwallez@chromium.org.

The reason for reverting is: This commit breaks the gles2_conform tests on all
the GPU FYI Linux and Mac bots, see
https://build.chromium.org/p/chromium.gpu.fyi/builders/Linux%20Release%20%28N....

[Julien Isorce Samsung, 2016-04-15 13:53:09]

Description was changed from

==========
Compute GpuDriverBugWorkarounds only one time in the GPU process

Previously the extraction of "--gpu-driver-bug-workarounds" from the
command-line was done in FeatureInfo's constructor. Since a new FeatureInfo
is created for each ContextGroup this extraction was done again whereas
the result is constant.

This CL moved owner ship of Workarounds from FeatureInfo to
InProcessCommandBuffer when running in gpu process mode. But moved
to GpuChannelManager otherwise.

class GpuDriverBugWorkarounds has been added within new files
gpu_driver_bug_workarounds.h/.cc.

BUG=599964

R=kbr@chromium.org, piman@chromium.org, rjkroege@chromium.org, zmo@chromium.org
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_optional_gpu_tests_rel;tryserver.chromium.mac:mac_optional_gpu_tests_rel;tryserver.chromium.win:win_optional_gpu_tests_rel

Committed: https://crrev.com/dab49af5278797db0a8374f2a7f06c6c33b8158b
Cr-Commit-Position: refs/heads/master@{#387569}
==========

to

==========
Compute GpuDriverBugWorkarounds only one time in the GPU process

Previously the extraction of "--gpu-driver-bug-workarounds" from the
command-line was done in FeatureInfo's constructor. Since a new FeatureInfo
is created for each ContextGroup this extraction was done again whereas
the result is constant.

This CL moved owner ship of Workarounds from FeatureInfo to
InProcessCommandBuffer when running in gpu process mode. But moved
to GpuChannelManager otherwise.

class GpuDriverBugWorkarounds has been added within new files
gpu_driver_bug_workarounds.h/.cc.

BUG=599964

R=kbr@chromium.org, piman@chromium.org, rjkroege@chromium.org, zmo@chromium.org
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_optional_gpu_tests_rel;tryserver.chromium.mac:mac_optional_gpu_tests_rel;tryserver.chromium.win:win_optional_gpu_tests_rel

Committed: https://crrev.com/dab49af5278797db0a8374f2a7f06c6c33b8158b
Cr-Commit-Position: refs/heads/master@{#387569}
==========

[Julien Isorce Samsung, 2016-04-15 13:53:28]

On 2016/04/15 13:28:38, Corentin Wallez wrote:
> A revert of this CL (patchset #16 id:300001) has been created in
> https://codereview.chromium.org/1892483004/ by mailto:cwallez@chromium.org.
> 
> The reason for reverting is: This commit breaks the gles2_conform tests on all
> the GPU FYI Linux and Mac bots, see
>
https://build.chromium.org/p/chromium.gpu.fyi/builders/Linux%20Release%20%28N....

Sorry about that, how can I run gles2_conform_test ? Because on my machine,
"./out/Release/gles2_conform_test --use-gpu-in-tests
--gtest_also_run_disabled_tests" just outputs "0 tests run".

[Julien Isorce Samsung, 2016-04-15 14:19:21]

The CQ bit was checked by j.isorce@samsung.com to run a CQ dry run

[commit-bot: I haz the power, 2016-04-15 14:19:35]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1871613002/310001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1871613002/310001

[commit-bot: I haz the power, 2016-04-15 15:17:42]

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

[commit-bot: I haz the power, 2016-04-15 15:17:44]

Dry run: This issue passed the CQ dry run.

[Julien Isorce Samsung, 2016-04-15 16:09:28]

The CQ bit was checked by j.isorce@samsung.com to run a CQ dry run

[commit-bot: I haz the power, 2016-04-15 16:09:39]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1871613002/330001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1871613002/330001

[commit-bot: I haz the power, 2016-04-15 18:48:35]

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

[commit-bot: I haz the power, 2016-04-15 18:48:37]

Dry run: This issue passed the CQ dry run.

[Julien Isorce Samsung, 2016-04-15 20:38:41]

On 2016/04/15 18:48:37, commit-bot: I haz the power wrote:
> Dry run: This issue passed the CQ dry run.

So locally it fails to generate the gles2 conform tests because I cannot fetch
third_party/gles2_conform. I guess this is an internal repo (not visible here
https://chromium.googlesource.com/chromium/src.git/+/master/third_party/). So
even if I add it in my .gclient I just cannot fetch it. And ading
internal_gles2_conform_tests=1 to GYP then fails because of missing
../../third_party/gles2_conform/GTF_ES/glsl/GTF.
 If it si a bug it has been reported here
https://bugs.chromium.org/p/chromium/issues/detail?id=394423 .

Other solution for me would have been to select "chromium.gpu.fyi/builders/Linux
Release (New Intel)" in the try bot list but it is not present.

In any case the "Patch Set 17" should fix the issue about failure on
gles2_conform tests. So I am tempted to submit to the CQ, what do you think ?

[Ken Russell (switch to Gerrit), 2016-04-16 00:10:55]

On 2016/04/15 20:38:41, j.isorce wrote:
> On 2016/04/15 18:48:37, commit-bot: I haz the power wrote:
> > Dry run: This issue passed the CQ dry run.
> 
> So locally it fails to generate the gles2 conform tests because I cannot fetch
> third_party/gles2_conform. I guess this is an internal repo (not visible here
> https://chromium.googlesource.com/chromium/src.git/+/master/third_party/). So
> even if I add it in my .gclient I just cannot fetch it. And ading
> internal_gles2_conform_tests=1 to GYP then fails because of missing
> ../../third_party/gles2_conform/GTF_ES/glsl/GTF.
>  If it si a bug it has been reported here
> https://bugs.chromium.org/p/chromium/issues/detail?id=394423 .
> 
> Other solution for me would have been to select
"chromium.gpu.fyi/builders/Linux
> Release (New Intel)" in the try bot list but it is not present.
> 
> In any case the "Patch Set 17" should fix the issue about failure on
> gles2_conform tests. So I am tempted to submit to the CQ, what do you think ?

How about doing that on Monday in case it passes the CQ, breaks the FYI bots and
they're red all weekend?

[Julien Isorce Samsung, 2016-04-18 08:53:21]

The CQ bit was checked by j.isorce@samsung.com

[Julien Isorce Samsung, 2016-04-18 08:53:21]

The patchset sent to the CQ was uploaded after l-g-t-m from
rjkroege@chromium.org, piman@chromium.org
Link to the patchset: https://codereview.chromium.org/1871613002/#ps350001
(title: "Rebase")

[commit-bot: I haz the power, 2016-04-18 08:53:31]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1871613002/350001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1871613002/350001

[commit-bot: I haz the power, 2016-04-18 08:55:21]

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

[commit-bot: I haz the power, 2016-04-18 08:55:23]

Try jobs failed on following builders:
  mac_chromium_gn_rel on tryserver.chromium.mac (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.mac/builders/mac_chromium_gn_r...)
  mac_chromium_rel_ng on tryserver.chromium.mac (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.mac/builders/mac_chromium_rel_...)
  mac_optional_gpu_tests_rel on tryserver.chromium.mac (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.mac/builders/mac_optional_gpu_...)

[Julien Isorce Samsung, 2016-04-18 11:07:20]

The CQ bit was checked by j.isorce@samsung.com to run a CQ dry run

[commit-bot: I haz the power, 2016-04-18 11:07:40]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1871613002/350001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1871613002/350001

[commit-bot: I haz the power, 2016-04-18 11:09:49]

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

[commit-bot: I haz the power, 2016-04-18 11:09:51]

Dry run: Try jobs failed on following builders:
  ios_dbg_simulator_gn on tryserver.chromium.mac (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.mac/builders/ios_dbg_simulator...)

[Julien Isorce Samsung, 2016-04-18 12:34:09]

The CQ bit was checked by j.isorce@samsung.com to run a CQ dry run

[commit-bot: I haz the power, 2016-04-18 12:34:23]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1871613002/370001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1871613002/370001

[commit-bot: I haz the power, 2016-04-18 13:50:20]

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

[commit-bot: I haz the power, 2016-04-18 13:50:22]

Dry run: This issue passed the CQ dry run.

[Julien Isorce Samsung, 2016-04-18 13:58:49]

The CQ bit was checked by j.isorce@samsung.com

[Julien Isorce Samsung, 2016-04-18 13:58:49]

The patchset sent to the CQ was uploaded after l-g-t-m from
rjkroege@chromium.org, piman@chromium.org
Link to the patchset: https://codereview.chromium.org/1871613002/#ps370001
(title: "Rebase")

[commit-bot: I haz the power, 2016-04-18 13:59:04]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1871613002/370001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1871613002/370001

[commit-bot: I haz the power, 2016-04-18 14:11:42]

Description was changed from

==========
Compute GpuDriverBugWorkarounds only one time in the GPU process

Previously the extraction of "--gpu-driver-bug-workarounds" from the
command-line was done in FeatureInfo's constructor. Since a new FeatureInfo
is created for each ContextGroup this extraction was done again whereas
the result is constant.

This CL moved owner ship of Workarounds from FeatureInfo to
InProcessCommandBuffer when running in gpu process mode. But moved
to GpuChannelManager otherwise.

class GpuDriverBugWorkarounds has been added within new files
gpu_driver_bug_workarounds.h/.cc.

BUG=599964

R=kbr@chromium.org, piman@chromium.org, rjkroege@chromium.org, zmo@chromium.org
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_optional_gpu_tests_rel;tryserver.chromium.mac:mac_optional_gpu_tests_rel;tryserver.chromium.win:win_optional_gpu_tests_rel

Committed: https://crrev.com/dab49af5278797db0a8374f2a7f06c6c33b8158b
Cr-Commit-Position: refs/heads/master@{#387569}
==========

to

==========
Compute GpuDriverBugWorkarounds only one time in the GPU process

Previously the extraction of "--gpu-driver-bug-workarounds" from the
command-line was done in FeatureInfo's constructor. Since a new FeatureInfo
is created for each ContextGroup this extraction was done again whereas
the result is constant.

This CL moved owner ship of Workarounds from FeatureInfo to
InProcessCommandBuffer when running in gpu process mode. But moved
to GpuChannelManager otherwise.

class GpuDriverBugWorkarounds has been added within new files
gpu_driver_bug_workarounds.h/.cc.

BUG=599964

R=kbr@chromium.org, piman@chromium.org, rjkroege@chromium.org, zmo@chromium.org
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_optional_gpu_tests_rel;tryserver.chromium.mac:mac_optional_gpu_tests_rel;tryserver.chromium.win:win_optional_gpu_tests_rel

Committed: https://crrev.com/dab49af5278797db0a8374f2a7f06c6c33b8158b
Cr-Commit-Position: refs/heads/master@{#387569}
==========

[commit-bot: I haz the power, 2016-04-18 14:11:44]

Committed patchset #20 (id:370001)

[commit-bot: I haz the power, 2016-04-18 14:13:14]

Description was changed from

==========
Compute GpuDriverBugWorkarounds only one time in the GPU process

Previously the extraction of "--gpu-driver-bug-workarounds" from the
command-line was done in FeatureInfo's constructor. Since a new FeatureInfo
is created for each ContextGroup this extraction was done again whereas
the result is constant.

This CL moved owner ship of Workarounds from FeatureInfo to
InProcessCommandBuffer when running in gpu process mode. But moved
to GpuChannelManager otherwise.

class GpuDriverBugWorkarounds has been added within new files
gpu_driver_bug_workarounds.h/.cc.

BUG=599964

R=kbr@chromium.org, piman@chromium.org, rjkroege@chromium.org, zmo@chromium.org
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_optional_gpu_tests_rel;tryserver.chromium.mac:mac_optional_gpu_tests_rel;tryserver.chromium.win:win_optional_gpu_tests_rel

Committed: https://crrev.com/dab49af5278797db0a8374f2a7f06c6c33b8158b
Cr-Commit-Position: refs/heads/master@{#387569}
==========

to

==========
Compute GpuDriverBugWorkarounds only one time in the GPU process

Previously the extraction of "--gpu-driver-bug-workarounds" from the
command-line was done in FeatureInfo's constructor. Since a new FeatureInfo
is created for each ContextGroup this extraction was done again whereas
the result is constant.

This CL moved owner ship of Workarounds from FeatureInfo to
InProcessCommandBuffer when running in gpu process mode. But moved
to GpuChannelManager otherwise.

class GpuDriverBugWorkarounds has been added within new files
gpu_driver_bug_workarounds.h/.cc.

BUG=599964

R=kbr@chromium.org, piman@chromium.org, rjkroege@chromium.org, zmo@chromium.org
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_optional_gpu_tests_rel;tryserver.chromium.mac:mac_optional_gpu_tests_rel;tryserver.chromium.win:win_optional_gpu_tests_rel

Committed: https://crrev.com/dab49af5278797db0a8374f2a7f06c6c33b8158b
Cr-Commit-Position: refs/heads/master@{#387569}

Committed: https://crrev.com/57b96c478dda56160c512fd85ee4e0f7437b524d
Cr-Commit-Position: refs/heads/master@{#387899}
==========

[commit-bot: I haz the power, 2016-04-18 14:13:16]

Patchset 20 (id:??) landed as
https://crrev.com/57b96c478dda56160c512fd85ee4e0f7437b524d
Cr-Commit-Position: refs/heads/master@{#387899}

[Zhenyao Mo, 2016-04-18 17:27:25]

A revert of this CL (patchset #20 id:370001) has been created in
https://codereview.chromium.org/1902453002/ by zmo@chromium.org.

The reason for reverting is: Sorry, but I have to revert this one before I can
revert https://codereview.chromium.org/1714883002/ due to display.h and
display.cc conflicts.

Feel free to reland after rebasing and testing it out carefully..

[Julien Isorce Samsung, 2016-04-18 20:07:56]

Description was changed from

==========
Compute GpuDriverBugWorkarounds only one time in the GPU process

Previously the extraction of "--gpu-driver-bug-workarounds" from the
command-line was done in FeatureInfo's constructor. Since a new FeatureInfo
is created for each ContextGroup this extraction was done again whereas
the result is constant.

This CL moved owner ship of Workarounds from FeatureInfo to
InProcessCommandBuffer when running in gpu process mode. But moved
to GpuChannelManager otherwise.

class GpuDriverBugWorkarounds has been added within new files
gpu_driver_bug_workarounds.h/.cc.

BUG=599964

R=kbr@chromium.org, piman@chromium.org, rjkroege@chromium.org, zmo@chromium.org
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_optional_gpu_tests_rel;tryserver.chromium.mac:mac_optional_gpu_tests_rel;tryserver.chromium.win:win_optional_gpu_tests_rel

Committed: https://crrev.com/dab49af5278797db0a8374f2a7f06c6c33b8158b
Cr-Commit-Position: refs/heads/master@{#387569}

Committed: https://crrev.com/57b96c478dda56160c512fd85ee4e0f7437b524d
Cr-Commit-Position: refs/heads/master@{#387899}
==========

to

==========
Compute GpuDriverBugWorkarounds only one time in the GPU process

Previously the extraction of "--gpu-driver-bug-workarounds" from the
command-line was done in FeatureInfo's constructor. Since a new FeatureInfo
is created for each ContextGroup this extraction was done again whereas
the result is constant.

This CL moved owner ship of Workarounds from FeatureInfo to
InProcessCommandBuffer when running in gpu process mode. But moved
to GpuChannelManager otherwise.

class GpuDriverBugWorkarounds has been added within new files
gpu_driver_bug_workarounds.h/.cc.

BUG=599964

R=kbr@chromium.org, piman@chromium.org, rjkroege@chromium.org, zmo@chromium.org
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_optional_gpu_tests_rel;tryserver.chromium.mac:mac_optional_gpu_tests_rel;tryserver.chromium.win:win_optional_gpu_tests_rel

Committed: https://crrev.com/dab49af5278797db0a8374f2a7f06c6c33b8158b
Cr-Commit-Position: refs/heads/master@{#387569}

Committed: https://crrev.com/57b96c478dda56160c512fd85ee4e0f7437b524d
Cr-Commit-Position: refs/heads/master@{#387899}
==========

[Julien Isorce Samsung, 2016-04-18 20:21:13]

The CQ bit was checked by j.isorce@samsung.com to run a CQ dry run

[commit-bot: I haz the power, 2016-04-18 20:21:37]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1871613002/390001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1871613002/390001

[commit-bot: I haz the power, 2016-04-18 20:25:22]

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

[commit-bot: I haz the power, 2016-04-18 20:25:24]

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

[Julien Isorce Samsung, 2016-04-18 21:36:57]

The CQ bit was checked by j.isorce@samsung.com to run a CQ dry run

[commit-bot: I haz the power, 2016-04-18 21:37:18]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1871613002/410001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1871613002/410001

[commit-bot: I haz the power, 2016-04-18 22:56:42]

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

[commit-bot: I haz the power, 2016-04-18 22:56:43]

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

[Julien Isorce Samsung, 2016-04-19 05:22:17]

The CQ bit was checked by j.isorce@samsung.com to run a CQ dry run

[commit-bot: I haz the power, 2016-04-19 05:22:29]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1871613002/430001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1871613002/430001

[commit-bot: I haz the power, 2016-04-19 05:24:03]

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

[commit-bot: I haz the power, 2016-04-19 05:24:05]

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

[Julien Isorce Samsung, 2016-04-19 06:46:09]

The CQ bit was checked by j.isorce@samsung.com

[Julien Isorce Samsung, 2016-04-19 06:46:09]

The patchset sent to the CQ was uploaded after l-g-t-m from
rjkroege@chromium.org, piman@chromium.org
Link to the patchset: https://codereview.chromium.org/1871613002/#ps430001
(title: "Rebase")

[commit-bot: I haz the power, 2016-04-19 06:46:22]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1871613002/430001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1871613002/430001

[commit-bot: I haz the power, 2016-04-19 07:40:11]

Description was changed from

==========
Compute GpuDriverBugWorkarounds only one time in the GPU process

Previously the extraction of "--gpu-driver-bug-workarounds" from the
command-line was done in FeatureInfo's constructor. Since a new FeatureInfo
is created for each ContextGroup this extraction was done again whereas
the result is constant.

This CL moved owner ship of Workarounds from FeatureInfo to
InProcessCommandBuffer when running in gpu process mode. But moved
to GpuChannelManager otherwise.

class GpuDriverBugWorkarounds has been added within new files
gpu_driver_bug_workarounds.h/.cc.

BUG=599964

R=kbr@chromium.org, piman@chromium.org, rjkroege@chromium.org, zmo@chromium.org
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_optional_gpu_tests_rel;tryserver.chromium.mac:mac_optional_gpu_tests_rel;tryserver.chromium.win:win_optional_gpu_tests_rel

Committed: https://crrev.com/dab49af5278797db0a8374f2a7f06c6c33b8158b
Cr-Commit-Position: refs/heads/master@{#387569}

Committed: https://crrev.com/57b96c478dda56160c512fd85ee4e0f7437b524d
Cr-Commit-Position: refs/heads/master@{#387899}
==========

to

==========
Compute GpuDriverBugWorkarounds only one time in the GPU process

Previously the extraction of "--gpu-driver-bug-workarounds" from the
command-line was done in FeatureInfo's constructor. Since a new FeatureInfo
is created for each ContextGroup this extraction was done again whereas
the result is constant.

This CL moved owner ship of Workarounds from FeatureInfo to
InProcessCommandBuffer when running in gpu process mode. But moved
to GpuChannelManager otherwise.

class GpuDriverBugWorkarounds has been added within new files
gpu_driver_bug_workarounds.h/.cc.

BUG=599964

R=kbr@chromium.org, piman@chromium.org, rjkroege@chromium.org, zmo@chromium.org
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_optional_gpu_tests_rel;tryserver.chromium.mac:mac_optional_gpu_tests_rel;tryserver.chromium.win:win_optional_gpu_tests_rel

Committed: https://crrev.com/dab49af5278797db0a8374f2a7f06c6c33b8158b
Cr-Commit-Position: refs/heads/master@{#387569}

Committed: https://crrev.com/57b96c478dda56160c512fd85ee4e0f7437b524d
Cr-Commit-Position: refs/heads/master@{#387899}
==========

[commit-bot: I haz the power, 2016-04-19 07:40:13]

Committed patchset #23 (id:430001)

[commit-bot: I haz the power, 2016-04-19 07:41:10]

Description was changed from

==========
Compute GpuDriverBugWorkarounds only one time in the GPU process

Previously the extraction of "--gpu-driver-bug-workarounds" from the
command-line was done in FeatureInfo's constructor. Since a new FeatureInfo
is created for each ContextGroup this extraction was done again whereas
the result is constant.

This CL moved owner ship of Workarounds from FeatureInfo to
InProcessCommandBuffer when running in gpu process mode. But moved
to GpuChannelManager otherwise.

class GpuDriverBugWorkarounds has been added within new files
gpu_driver_bug_workarounds.h/.cc.

BUG=599964

R=kbr@chromium.org, piman@chromium.org, rjkroege@chromium.org, zmo@chromium.org
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_optional_gpu_tests_rel;tryserver.chromium.mac:mac_optional_gpu_tests_rel;tryserver.chromium.win:win_optional_gpu_tests_rel

Committed: https://crrev.com/dab49af5278797db0a8374f2a7f06c6c33b8158b
Cr-Commit-Position: refs/heads/master@{#387569}

Committed: https://crrev.com/57b96c478dda56160c512fd85ee4e0f7437b524d
Cr-Commit-Position: refs/heads/master@{#387899}
==========

to

==========
Compute GpuDriverBugWorkarounds only one time in the GPU process

Previously the extraction of "--gpu-driver-bug-workarounds" from the
command-line was done in FeatureInfo's constructor. Since a new FeatureInfo
is created for each ContextGroup this extraction was done again whereas
the result is constant.

This CL moved owner ship of Workarounds from FeatureInfo to
InProcessCommandBuffer when running in gpu process mode. But moved
to GpuChannelManager otherwise.

class GpuDriverBugWorkarounds has been added within new files
gpu_driver_bug_workarounds.h/.cc.

BUG=599964

R=kbr@chromium.org, piman@chromium.org, rjkroege@chromium.org, zmo@chromium.org
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_optional_gpu_tests_rel;tryserver.chromium.mac:mac_optional_gpu_tests_rel;tryserver.chromium.win:win_optional_gpu_tests_rel

Committed: https://crrev.com/dab49af5278797db0a8374f2a7f06c6c33b8158b
Cr-Commit-Position: refs/heads/master@{#387569}

Committed: https://crrev.com/57b96c478dda56160c512fd85ee4e0f7437b524d
Cr-Commit-Position: refs/heads/master@{#387899}

Committed: https://crrev.com/265b2e56348b55fa622275e67a9863ad588f5183
Cr-Commit-Position: refs/heads/master@{#388157}
==========

[commit-bot: I haz the power, 2016-04-19 07:41:12]

Patchset 23 (id:??) landed as
https://crrev.com/265b2e56348b55fa622275e67a9863ad588f5183
Cr-Commit-Position: refs/heads/master@{#388157}