GPU context creation code duplication cleanup

GPU context creation code duplication cleanup

Use ContextCreationAttribHelper through the context creation in both
in-process and out-of-process cases, instead of duplicate structures such
as GLInProcessContextAttribs. Enums used for serializing the attributes
are now defined in only one place, instead of three.

Remove small pieces of dead code related to share groups.
share_resources_ attrib is removed since it is not in use anywhere.

gpu_unittests are changed to better model how bind_generates_resource
gets set, and new checks are added to make sure it is set consistently
everywhere.

TEST=WebGL conformance tests (with and without --in-process-gpu),
cc_unittests, gpu_unittests

BUG=

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=289269

[oetuaho-nv, 2014-08-08 12:59:26]

I started running into this duplicated code when trying to add an npot
validation setting to the command buffer. It needs a bit of cleanup to make it
more easily understandable, here's something to get started with that.

This change is still a bit WIP and there's some parts I'm not quite sure about
yet, particularly how bind_generates_resource gets set. Before there were
instances where it was set to true in the server by default but false on the
client - I changed these cases so that it's set to false on both. Does this need
to be the other way around in some cases? I'm particularly interested in getting
feedback on how this might affect Android Webview and Pepper. Are there good
ways to test these?

[piman, 2014-08-08 20:56:34]

This is a good change, LGTM.

Eventually I think we should stop passing a vector of ints via IPC, and just
pass something like ContextCreationAttribHelper directly - to avoiding
serialization/deserialization, and potential associated security bugs (we've had
some in the past). This gets us on the right path.

https://codereview.chromium.org/455783002/diff/1/android_webview/browser/hard...
File android_webview/browser/hardware_renderer.cc (right):

https://codereview.chromium.org/455783002/diff/1/android_webview/browser/hard...
android_webview/browser/hardware_renderer.cc:54:
attribs_for_gles2.bind_generates_resource_ = false;
nit: move this in WebGraphicsContext3DImpl::ConvertAttributes. In practice, any
user of WebGraphicsContext3D does bind_generates_resource = false, because the
only thing that uses it as true is Pepper3D (and possibly tests), that doesn't
use WebGraphicsContext3D.

[Zhenyao Mo, 2014-08-09 00:42:25]

Nice cleanup

LGTM with a few nits

https://codereview.chromium.org/455783002/diff/1/gpu/command_buffer/common/gl...
File gpu/command_buffer/common/gles2_cmd_utils.cc (left):

https://codereview.chromium.org/455783002/diff/1/gpu/command_buffer/common/gl...
gpu/command_buffer/common/gles2_cmd_utils.cc:802: const int32 kShareResources   
    = 0x10000;
nit: move the below three values one up.  It's weird to skip one value.

https://codereview.chromium.org/455783002/diff/1/gpu/command_buffer/common/gl...
File gpu/command_buffer/common/gles2_cmd_utils.h (right):

https://codereview.chromium.org/455783002/diff/1/gpu/command_buffer/common/gl...
gpu/command_buffer/common/gles2_cmd_utils.h:199: class GLES2_UTILS_EXPORT
ContextCreationAttribHelper {
Since you are cleaning things up in this: this should be a struct rather than
class, and all members should be named alpha_size instead of alpha_size_, etc.

[oetuaho-nv, 2014-08-12 07:19:27]

The CQ bit was checked by oetuaho@nvidia.com

[commit-bot: I haz the power, 2014-08-12 07:20:45]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/oetuaho@nvidia.com/455783002/40001

[oetuaho-nv, 2014-08-12 10:28:15]

The CQ bit was unchecked by oetuaho@nvidia.com

[oetuaho-nv, 2014-08-12 13:49:15]

boliu, benm: Can you approve the android_webview changes?

Also still working through missing deps for the component build, hopefully this
patch set will do the trick.

[boliu, 2014-08-12 15:18:14]

nice, android_webview lgtm

[oetuaho-nv, 2014-08-13 09:37:19]

The CQ bit was checked by oetuaho@nvidia.com

[commit-bot: I haz the power, 2014-08-13 09:38:53]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/oetuaho@nvidia.com/455783002/80001

[commit-bot: I haz the power, 2014-08-13 11:50:44]

Change committed as 289269