Add nullptr check for GpuProcessHost when allocating GpuMemoryBuffer

Add nullptr check for GpuProcessHost when allocating GpuMemoryBuffer

This is appearing as a minor crasher.

BUG=723698

[ccameron, 2017-06-26 21:01:02]

Description was changed from

==========
Add nullptr check for GpuProcessHost when allocating GpuMemoryBuffer

This is appearing as a minor crasher.

BUG=723698
==========

to

==========
Add nullptr check for GpuProcessHost when allocating GpuMemoryBuffer

This is appearing as a minor crasher.

BUG=723698
==========

[ccameron, 2017-06-26 21:01:02]

ccameron@chromium.org changed reviewers:
+ zmo@chromium.org

[ccameron, 2017-06-26 21:01:19]

ptal -- small fix.

[Zhenyao Mo, 2017-06-26 21:04:20]

lgtm

[ccameron, 2017-06-26 21:05:03]

The CQ bit was checked by ccameron@chromium.org

[commit-bot: I haz the power, 2017-06-26 21:05:25]

CQ is trying da patch.

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

[reveman, 2017-06-26 22:12:18]

reveman@chromium.org changed reviewers:
+ reveman@chromium.org

[reveman, 2017-06-26 22:12:19]

Can you add a test for this? Or at least explain in the description of the CL
when this can happen?

[reveman, 2017-06-26 22:12:51]

The CQ bit was unchecked by reveman@chromium.org

[ccameron, 2017-06-26 22:42:06]

On 2017/06/26 22:12:19, reveman wrote:
> Can you add a test for this? Or at least explain in the description of the CL
> when this can happen?

Every single other instance of GpuProcessHost::Get() has a null check, so this
is bringing this in line with all of the other instances, which presumably had a
motivation for their existence.

I took on this bug because it was mis-attributed to a change I made, and I don't
plan to spend any more time on it. If you'd like more research into this, I can
assign the bug to you.

[reveman, 2017-06-26 22:59:57]

On 2017/06/26 at 22:42:06, ccameron wrote:
> On 2017/06/26 22:12:19, reveman wrote:
> > Can you add a test for this? Or at least explain in the description of the
CL
> > when this can happen?
> 
> Every single other instance of GpuProcessHost::Get() has a null check, so this
is bringing this in line with all of the other instances, which presumably had a
motivation for their existence.

Ok, if you mention that in the description and point to another place where we
have a null check then this lgtm.

> 
> I took on this bug because it was mis-attributed to a change I made, and I
don't plan to spend any more time on it. If you'd like more research into this,
I can assign the bug to you.