Add Sync related APIs to GPU command buffer: Part I

Add Sync related APIs to GPU command buffer.

glFenceSync
glIsFence
glDeleteFence

Most of the work is to deal with the fact that GLsync is not like any other ID types (GLuint).

BUG=429053
TEST=gpu_unittests

Committed: https://crrev.com/b5e2c68d2572cf5a38a0329e5daaee5fc27f32fe
Cr-Commit-Position: refs/heads/master@{#312544}

[Zhenyao Mo, 2015-01-21 01:25:13]

Patchset #3 (id:40001) has been deleted

[Zhenyao Mo, 2015-01-21 01:25:21]

Patchset #2 (id:20001) has been deleted

[Zhenyao Mo, 2015-01-21 01:25:28]

Patchset #1 (id:1) has been deleted

[Zhenyao Mo, 2015-01-21 01:45:25]

Patchset #2 (id:80001) has been deleted

[Zhenyao Mo, 2015-01-21 18:51:21]

Patchset #2 (id:100001) has been deleted

[Zhenyao Mo, 2015-01-21 19:01:20]

Patchset #3 (id:140001) has been deleted

[Zhenyao Mo, 2015-01-21 23:26:04]

Patchset #6 (id:220001) has been deleted

[Zhenyao Mo, 2015-01-21 23:26:12]

Patchset #5 (id:200001) has been deleted

[Zhenyao Mo, 2015-01-21 23:26:21]

Patchset #4 (id:180001) has been deleted

[Zhenyao Mo, 2015-01-21 23:26:29]

Patchset #3 (id:160001) has been deleted

[Zhenyao Mo, 2015-01-21 23:33:30]

Patchset #3 (id:240001) has been deleted

[Zhenyao Mo, 2015-01-21 23:40:38]

Patchset #3 (id:260001) has been deleted

[Zhenyao Mo, 2015-01-21 23:46:11]

Patchset #3 (id:280001) has been deleted

[Zhenyao Mo, 2015-01-21 23:48:11]

Patchset #3 (id:300001) has been deleted

[Zhenyao Mo, 2015-01-22 00:01:46]

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

[Zhenyao Mo, 2015-01-22 00:04:03]

piman: please review.
Others: FYI (feel free to review)

[piman, 2015-01-22 00:49:32]

lgtm

https://codereview.chromium.org/859043005/diff/340001/gpu/command_buffer/clie...
File gpu/command_buffer/client/gles2_implementation.cc (right):

https://codereview.chromium.org/859043005/diff/340001/gpu/command_buffer/clie...
gpu/command_buffer/client/gles2_implementation.cc:980: GLuint my_sync =
ToGLuint(sync);
nit: anything better than my_?
Maybe uint_sync?

https://codereview.chromium.org/859043005/diff/340001/gpu/command_buffer/clie...
File gpu/command_buffer/client/gles2_implementation_impl_autogen.h (right):

https://codereview.chromium.org/859043005/diff/340001/gpu/command_buffer/clie...
gpu/command_buffer/client/gles2_implementation_impl_autogen.h:647: if (condition
!= 0x9117) {
nit: any way to have the enum rather than the value?

https://codereview.chromium.org/859043005/diff/340001/gpu/command_buffer/serv...
File gpu/command_buffer/service/context_group.h (right):

https://codereview.chromium.org/859043005/diff/340001/gpu/command_buffer/serv...
gpu/command_buffer/service/context_group.h:290: std::map<GLuint, GLsync>
syncs_id_map_;
nit: by the way, could all these be hash_maps instead, since we don't care about
the order (hash maps are usually cheaper).

[Zhenyao Mo, 2015-01-22 01:12:57]

https://codereview.chromium.org/859043005/diff/340001/gpu/command_buffer/clie...
File gpu/command_buffer/client/gles2_implementation.cc (right):

https://codereview.chromium.org/859043005/diff/340001/gpu/command_buffer/clie...
gpu/command_buffer/client/gles2_implementation.cc:980: GLuint my_sync =
ToGLuint(sync);
On 2015/01/22 00:49:32, piman (Very slow to review) wrote:
> nit: anything better than my_?
> Maybe uint_sync?

Done.

https://codereview.chromium.org/859043005/diff/340001/gpu/command_buffer/clie...
File gpu/command_buffer/client/gles2_implementation_impl_autogen.h (right):

https://codereview.chromium.org/859043005/diff/340001/gpu/command_buffer/clie...
gpu/command_buffer/client/gles2_implementation_impl_autogen.h:647: if (condition
!= 0x9117) {
On 2015/01/22 00:49:32, piman (Very slow to review) wrote:
> nit: any way to have the enum rather than the value?

Yes, we could switch to use the enum if we forwardly declare the enum in a few
files or in gles2_cmd_format_autogen.h.  I think later is a cleaner solution,
but requires some extra modification on the code generator.  Created
crbug.com/450854 to followup in a separate CL.

https://codereview.chromium.org/859043005/diff/340001/gpu/command_buffer/serv...
File gpu/command_buffer/service/context_group.h (right):

https://codereview.chromium.org/859043005/diff/340001/gpu/command_buffer/serv...
gpu/command_buffer/service/context_group.h:290: std::map<GLuint, GLsync>
syncs_id_map_;
On 2015/01/22 00:49:32, piman (Very slow to review) wrote:
> nit: by the way, could all these be hash_maps instead, since we don't care
about
> the order (hash maps are usually cheaper).

Done.

[Zhenyao Mo, 2015-01-22 01:30:55]

The CQ bit was checked by zmo@chromium.org

[commit-bot: I haz the power, 2015-01-22 01:32:41]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/859043005/380001

[commit-bot: I haz the power, 2015-01-22 02:51:05]

Committed patchset #6 (id:380001)

[commit-bot: I haz the power, 2015-01-22 02:52:35]

Patchset 6 (id:??) landed as
https://crrev.com/b5e2c68d2572cf5a38a0329e5daaee5fc27f32fe
Cr-Commit-Position: refs/heads/master@{#312544}