gpu: Enable relevant extentions on GLES3.

gpu: Enable relevant extentions on GLES3.

What’s New in OpenGL ES 3.0 among GL extentions
- Vertex array
- Element index uint
- npot texture
- float texture (1)(2)
- GLSL standard derivatives
- Blend min&max

(1) GLES3 supports float and half float texture but only half float texture is
filterable and doesn't allow them for color attachment.
(2) This CL doesn't enable half_float due to conflicting GL_OES_texture_float

BUG=405484
Committed: https://crrev.com/5d13a2b12e8cef22465edfd7113e002d257da117
Cr-Commit-Position: refs/heads/master@{#292849}

[dshwang, 2014-08-20 11:03:37]

Could you review?

I referred to GLES3 spec[1] and OpenGL ES 3.0 Programming Guide[2]
[1] http://www.khronos.org/registry/gles/specs/3.0/es_spec_3.0.0.pdf
[2] http://opengles-book.com/

[bajones, 2014-08-20 16:50:44]

On 2014/08/20 at 11:03:37, dongseong.hwang wrote:
> Could you review?
> 
> I referred to GLES3 spec[1] and OpenGL ES 3.0 Programming Guide[2]
> [1] http://www.khronos.org/registry/gles/specs/3.0/es_spec_3.0.0.pdf
> [2] http://opengles-book.com/

Overall this look good but the bot errors in gpu_unittests seem highly relevant
to this change. Please look into what's causing the failures.

[dshwang, 2014-08-27 15:38:12]

Patchset #3 (id:60001) has been deleted

[dshwang, 2014-08-28 05:43:55]

Now all test bots succeeds. Could you review?

I postpone glDrawBuffers and gl_FragDepth support because it might need to
change gles2_cmd_decoder or ANGLE.

[bajones, 2014-08-28 16:32:48]

On 2014/08/28 at 05:43:55, dongseong.hwang wrote:
> Now all test bots succeeds. Could you review?
> 
> I postpone glDrawBuffers and gl_FragDepth support because it might need to
change gles2_cmd_decoder or ANGLE.

Thanks for sorting out the tests! gpu/command_buffer/ LGTM.

[dshwang, 2014-08-28 17:44:58]

On 2014/08/28 16:32:48, bajones wrote:
> On 2014/08/28 at 05:43:55, dongseong.hwang wrote:
> > Now all test bots succeeds. Could you review?
> > 
> > I postpone glDrawBuffers and gl_FragDepth support because it might need to
> change gles2_cmd_decoder or ANGLE.
> 
> Thanks for sorting out the tests! gpu/command_buffer/ LGTM.

Thanks for review!

piman@, could you review in ui/gl/generate_bindings.py ?

[piman, 2014-08-28 18:17:40]

LGTM+nits

https://codereview.chromium.org/489023002/diff/80001/gpu/command_buffer/servi...
File gpu/command_buffer/service/feature_info.cc (right):

https://codereview.chromium.org/489023002/diff/80001/gpu/command_buffer/servi...
gpu/command_buffer/service/feature_info.cc:442: feature_flags_.npot_ok |= true;
nit: feature_flags_.npot_ok = true

https://codereview.chromium.org/489023002/diff/80001/gpu/command_buffer/servi...
gpu/command_buffer/service/feature_info.cc:756: for (GLenum i =
GL_COLOR_ATTACHMENT0_EXT;
nit: GL_COLOR_ATTACHMENT0_EXT is already in validators_.attachment (from autogen
code), hence why this started at 1.

[dshwang, 2014-08-29 07:49:09]

Thank you for review. I address nits and add additional COMPILE_ASSERT to check
if EXT variables are equal to GLES3 variables. As doing, I found shocking
reality; GL_HALF_FLOAT_OES != GL_HALF_FLOAT. See below comment.

https://codereview.chromium.org/489023002/diff/80001/gpu/command_buffer/servi...
File gpu/command_buffer/service/feature_info.cc (right):

https://codereview.chromium.org/489023002/diff/80001/gpu/command_buffer/servi...
gpu/command_buffer/service/feature_info.cc:442: feature_flags_.npot_ok |= true;
On 2014/08/28 18:17:39, piman (OOO) wrote:
> nit: feature_flags_.npot_ok = true

Done.

https://codereview.chromium.org/489023002/diff/80001/gpu/command_buffer/servi...
gpu/command_buffer/service/feature_info.cc:494:
texture_format_validators_[GL_ALPHA].AddValue(GL_HALF_FLOAT_OES);
I exclude half_float support because GL_HALF_FLOAT_OES isn't equal to
GL_HALF_FLOAT!! Currently, chromium uses GL_HALF_FLOAT_OES here and there. IMO
we replace GL_HALF_FLOAT in all places and gles2_cmd_decoder handles it in terms
of GL_HALF_FLOAT_OES

[dshwang, 2014-09-01 07:10:48]

The CQ bit was checked by dongseong.hwang@intel.com

[commit-bot: I haz the power, 2014-09-01 07:11:36]

CQ is trying da patch. Follow status at

https://chromium-status.appspot.com/cq/dongseong.hwang@intel.com/489023002/10...

[commit-bot: I haz the power, 2014-09-01 08:32:45]

Committed patchset #4 (id:100001) as 6bdb23d53a42eb7265e2318d51924cad063c6ea1

[piman, 2014-09-02 18:33:14]

On 2014/08/29 07:49:09, dshwang(Vacation) wrote:
> Thank you for review. I address nits and add additional COMPILE_ASSERT to
check
> if EXT variables are equal to GLES3 variables. As doing, I found shocking
> reality; GL_HALF_FLOAT_OES != GL_HALF_FLOAT. See below comment.
> 
>
https://codereview.chromium.org/489023002/diff/80001/gpu/command_buffer/servi...
> File gpu/command_buffer/service/feature_info.cc (right):
> 
>
https://codereview.chromium.org/489023002/diff/80001/gpu/command_buffer/servi...
> gpu/command_buffer/service/feature_info.cc:442: feature_flags_.npot_ok |=
true;
> On 2014/08/28 18:17:39, piman (OOO) wrote:
> > nit: feature_flags_.npot_ok = true
> 
> Done.
> 
>
https://codereview.chromium.org/489023002/diff/80001/gpu/command_buffer/servi...
> gpu/command_buffer/service/feature_info.cc:494:
> texture_format_validators_[GL_ALPHA].AddValue(GL_HALF_FLOAT_OES);
> I exclude half_float support because GL_HALF_FLOAT_OES isn't equal to
> GL_HALF_FLOAT!! Currently, chromium uses GL_HALF_FLOAT_OES here and there. IMO
> we replace GL_HALF_FLOAT in all places and gles2_cmd_decoder handles it in
terms
> of GL_HALF_FLOAT_OES

We should be exposing GL_HALF_FLOAT_OES to the client, since we expose
ES2+GL_OES_texture_half_float (if available). So we should be using
GL_HALF_FLOAT_OES everywhere. There is code in ui/gl/gl_gl_api_implementation.cc
to translate GL_HALF_FLOAT_OES into GL_HALF_FLOAT_ARB (== GL_HALF_FLOAT) for
desktop GL. It should be easy to extend this for es3.

[commit-bot: I haz the power, 2014-09-10 03:16:04]

Patchset 4 (id:??) landed as
https://crrev.com/5d13a2b12e8cef22465edfd7113e002d257da117
Cr-Commit-Position: refs/heads/master@{#292849}