[wip] Kill GPU if the memory buffer cannot be used as advertised.

[wip] Kill GPU if the memory buffer cannot be used as advertised.

...

BUG=
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel

[sadrul, 2016-09-02 00:59:19]

Description was changed from

==========
[wip] Kill GPU if the memory buffer cannot be used as advertised.

...

BUG=
==========

to

==========
[wip] Kill GPU if the memory buffer cannot be used as advertised.

...

BUG=
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel
==========

[sadrul, 2016-09-02 00:59:27]

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

[commit-bot: I haz the power, 2016-09-02 00:59:44]

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

[sadrul, 2016-09-02 01:11:22]

sadrul@chromium.org changed reviewers:
+ rjkroege@chromium.org

[sadrul, 2016-09-02 01:11:22]

This is what I was thinking the next step could be after
https://codereview.chromium.org/2302053003/. But I don't know if it's OK to
assume that gpu needs to be killed if GpuMemoryBuffer allocation fails.

[rjkroege, 2016-09-02 01:14:29]

but.. I think it would be useful to know why we're doing the force shutdown.
i.e.: the handle that came from the gpu process was invalid.

https://codereview.chromium.org/2302733003/diff/1/gpu/ipc/client/gpu_memory_b...
File gpu/ipc/client/gpu_memory_buffer_impl_ozone_native_pixmap.cc (right):

https://codereview.chromium.org/2302733003/diff/1/gpu/ipc/client/gpu_memory_b...
gpu/ipc/client/gpu_memory_buffer_impl_ozone_native_pixmap.cc:71: if
(!native_pixmap)
yeah. I think so.

[commit-bot: I haz the power, 2016-09-02 01:50:13]

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

[commit-bot: I haz the power, 2016-09-02 01:50:14]

Dry run: This issue passed the CQ dry run.

[sadrul, 2016-09-02 03:59:53]

On 2016/09/02 01:14:29, rjkroege wrote:
> but.. I think it would be useful to know why we're doing the force shutdown.
> i.e.: the handle that came from the gpu process was invalid.

Do you mean plumb-back the reason for the failure from here back into content?

https://codereview.chromium.org/2302733003/diff/1/gpu/ipc/client/gpu_memory_b...
File gpu/ipc/client/gpu_memory_buffer_impl_ozone_native_pixmap.cc (right):

https://codereview.chromium.org/2302733003/diff/1/gpu/ipc/client/gpu_memory_b...
gpu/ipc/client/gpu_memory_buffer_impl_ozone_native_pixmap.cc:71: if
(!native_pixmap)
On 2016/09/02 01:14:29, rjkroege wrote:
> yeah. I think so.

Not sure what you are agreeing with here :)

[rjkroege, 2016-09-02 19:38:49]

On 2016/09/02 03:59:53, sadrul wrote:
> On 2016/09/02 01:14:29, rjkroege wrote:
> > but.. I think it would be useful to know why we're doing the force shutdown.
> > i.e.: the handle that came from the gpu process was invalid.
> 
> Do you mean plumb-back the reason for the failure from here back into content?

Yes. We have an infrastructure in place to record the site responsible for a gpu
process crash. We should track which sites are open such that the GPU process
has sent us an invalid memory buffer handle.

> 
>
https://codereview.chromium.org/2302733003/diff/1/gpu/ipc/client/gpu_memory_b...
> File gpu/ipc/client/gpu_memory_buffer_impl_ozone_native_pixmap.cc (right):
> 
>
https://codereview.chromium.org/2302733003/diff/1/gpu/ipc/client/gpu_memory_b...
> gpu/ipc/client/gpu_memory_buffer_impl_ozone_native_pixmap.cc:71: if
> (!native_pixmap)
> On 2016/09/02 01:14:29, rjkroege wrote:
> > yeah. I think so.
> 
> Not sure what you are agreeing with here :)

The code. Maybe.