Add compositor support for half-float RGBA buffers and textures

Add the ResourceFormat RGBA_F16, and appropriate GL and Skia values.

Add the GPU capability color_buffer_float, to expose if it is possible
to create a RGBA_F16 render target.

When allocating a RenderPass texture, use RGBA_F16 if the output surface
is high-bit-depth (which will include HDR surfaces). Add the method
OutputSurface::IsHighBitDepth to indicate this property.

Note that this cannot be tested by the cc pixel tests (the correct way
to test this) because they use a version of Mesa that does not support
floating-point textures. Tests so far have been manual, as we work on
upgrading Mesa. This will be tested by having the pixel tests which
currently run in software and GL mode also run in "GL with high bit
depth output surface" mode.

BUG=687346
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_trusty_blink_rel;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


Review-Url: https://codereview.chromium.org/2674493003
Cr-Original-Commit-Position: refs/heads/master@{#448143}
Committed: https://chromium.googlesource.com/chromium/src/+/197ec6b2ea2af1d95c0acdf9a0ebf0cd479fdf40
Review-Url: https://codereview.chromium.org/2674493003
Cr-Commit-Position: refs/heads/master@{#448375}
Committed: https://chromium.googlesource.com/chromium/src/+/22094236e8ef012d05e313c6e77c07a71addb37e

[ccameron, 2017-02-01 23:07:58]

Description was changed from

==========
Add compositor support for half-float RGBA buffers and textures

BUG=687346
==========

to

==========
Add compositor support for half-float RGBA buffers and textures

BUG=687346
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
==========

[ccameron, 2017-02-01 23:07:58]

Description was changed from

==========
Add compositor support for half-float RGBA buffers and textures

BUG=687346
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
==========

to

==========
Add compositor support for half-float RGBA buffers and textures

BUG=687346
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_trusty_blink_rel
==========

[ccameron, 2017-02-01 23:08:00]

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

[commit-bot: I haz the power, 2017-02-01 23:08:36]

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

[ccameron, 2017-02-01 23:19:18]

Description was changed from

==========
Add compositor support for half-float RGBA buffers and textures

BUG=687346
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_trusty_blink_rel
==========

to

==========
Add the ResourceFormat RGBA_F16, and appropriate GL and Skia values.

Add the GPU capability color_buffer_float, to expose if it is possible
to create a RGBA_F16 render target.

Enable exposing GPU capabilities color_buffer_float and
texture_half_float_linear to non-WebGL GLES2 contexts (e.g, the
compositor context).

When allocating a RenderPass texture, use RGBA_F16 if the output surface
is high-bit-depth (which will include HDR surfaces). Add the method
OutputSurface::IsHighBitDepth to indicate this property.

BUG=687346
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_trusty_blink_rel
==========

[ccameron, 2017-02-01 23:19:18]

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

[ccameron, 2017-02-01 23:21:19]

Description was changed from

==========
Add the ResourceFormat RGBA_F16, and appropriate GL and Skia values.

Add the GPU capability color_buffer_float, to expose if it is possible
to create a RGBA_F16 render target.

Enable exposing GPU capabilities color_buffer_float and
texture_half_float_linear to non-WebGL GLES2 contexts (e.g, the
compositor context).

When allocating a RenderPass texture, use RGBA_F16 if the output surface
is high-bit-depth (which will include HDR surfaces). Add the method
OutputSurface::IsHighBitDepth to indicate this property.

BUG=687346
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_trusty_blink_rel
==========

to

==========
Add the ResourceFormat RGBA_F16, and appropriate GL and Skia values.

Add the GPU capability color_buffer_float, to expose if it is possible
to create a RGBA_F16 render target.

Enable exposing GPU capabilities color_buffer_float and
texture_half_float_linear to non-WebGL GLES2 contexts (e.g, the
compositor context).

When allocating a RenderPass texture, use RGBA_F16 if the output surface
is high-bit-depth (which will include HDR surfaces). Add the method
OutputSurface::IsHighBitDepth to indicate this property.

Note that this cannot be tested by the cc pixel tests (the correct way
to test this) because they use a version of Mesa that does not support
floating-point textures. Tests so far have been manual, as we work on
upgrading Mesa. This will be tested by having the pixel tests which
currently run in software and GL mode also run in "GL with high bit
depth output surface" mode.

BUG=687346
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_trusty_blink_rel
==========

[ccameron, 2017-02-01 23:21:33]

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

[ccameron, 2017-02-01 23:21:44]

ptal

[commit-bot: I haz the power, 2017-02-01 23:48:57]

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

[commit-bot: I haz the power, 2017-02-01 23:48:58]

Dry run: Try jobs failed on following builders:
  cast_shell_linux on master.tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/cast_shell_linu...)

[enne (OOO), 2017-02-01 23:55:22]

https://codereview.chromium.org/2674493003/diff/1/cc/output/output_surface.cc
File cc/output/output_surface.cc (right):

https://codereview.chromium.org/2674493003/diff/1/cc/output/output_surface.cc...
cc/output/output_surface.cc:40: return true;
I'm a little surprised you return true by default here.  <_<

https://codereview.chromium.org/2674493003/diff/1/cc/output/output_surface.h
File cc/output/output_surface.h (right):

https://codereview.chromium.org/2674493003/diff/1/cc/output/output_surface.h#...
cc/output/output_surface.h:96: virtual bool IsHighBitDepth() const;
Maybe should also be pure virtual like the rest of the interface?

[jbauman, 2017-02-01 23:59:21]

gpu/ stuff lgtm

https://codereview.chromium.org/2674493003/diff/1/cc/output/direct_renderer.cc
File cc/output/direct_renderer.cc (right):

https://codereview.chromium.org/2674493003/diff/1/cc/output/direct_renderer.c...
cc/output/direct_renderer.cc:612: texture->Allocate(size,
Is it expected that BackbufferFormat could change during the life of the
OutputSurface? If so this should check if the formats match and recreate it if
not.

[danakj, 2017-02-02 16:23:09]

https://codereview.chromium.org/2674493003/diff/1/cc/output/output_surface.h
File cc/output/output_surface.h (right):

https://codereview.chromium.org/2674493003/diff/1/cc/output/output_surface.h#...
cc/output/output_surface.h:96: virtual bool IsHighBitDepth() const;
On 2017/02/01 23:55:22, enne wrote:
> Maybe should also be pure virtual like the rest of the interface?

Yes please.

[ccameron, 2017-02-02 21:08:41]

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

[ccameron, 2017-02-02 21:08:41]

Updated. I removed the OuputSurface part because it's premature -- adding that
was so that the HDR firefight work could hook in, but they can make that change
locally for now.

Adding tsepez@ for IPC review

https://codereview.chromium.org/2674493003/diff/1/cc/output/output_surface.cc
File cc/output/output_surface.cc (right):

https://codereview.chromium.org/2674493003/diff/1/cc/output/output_surface.cc...
cc/output/output_surface.cc:40: return true;
On 2017/02/01 23:55:22, enne wrote:
> I'm a little surprised you return true by default here.  <_<

I'm not 100% sure this is the right interface, so I've removed it and replaced
it with a TODO.

[ccameron, 2017-02-02 21:16:03]

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

[commit-bot: I haz the power, 2017-02-02 21:16:33]

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

[commit-bot: I haz the power, 2017-02-02 21:48:24]

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

[commit-bot: I haz the power, 2017-02-02 21:48:25]

Dry run: Try jobs failed on following builders:
  cast_shell_linux on master.tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/cast_shell_linu...)

[enne (OOO), 2017-02-02 22:18:57]

lgtm

[ccameron, 2017-02-03 03:09:06]

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

[commit-bot: I haz the power, 2017-02-03 03:09:56]

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

[commit-bot: I haz the power, 2017-02-03 04:05:59]

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

[commit-bot: I haz the power, 2017-02-03 04:06:05]

Dry run: This issue passed the CQ dry run.

[ccameron, 2017-02-03 07:58:20]

ping tsepez@ for security review

[Tom Sepez, 2017-02-03 20:04:41]

lgtm

[ccameron, 2017-02-03 20:16:09]

The CQ bit was checked by ccameron@chromium.org

[ccameron, 2017-02-03 20:16:09]

The patchset sent to the CQ was uploaded after l-g-t-m from
jbauman@chromium.org, enne@chromium.org
Link to the patchset: https://codereview.chromium.org/2674493003/#ps40001
(title: "Update test expectations")

[commit-bot: I haz the power, 2017-02-03 20:16:44]

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

[commit-bot: I haz the power, 2017-02-03 20:36:33]

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

[commit-bot: I haz the power, 2017-02-03 20:36:35]

Failed to apply patch for cc/resources/resource_provider.cc:
While running git apply --index -p1;
  error: patch failed: cc/resources/resource_provider.cc:579
  error: cc/resources/resource_provider.cc: patch does not apply

Patch:       cc/resources/resource_provider.cc
Index: cc/resources/resource_provider.cc
diff --git a/cc/resources/resource_provider.cc
b/cc/resources/resource_provider.cc
index
d0733c9f095d3214c544baf558ea19bab71168b5..f73e46654747577a6095862759fec0e9c73ac2ac
100644
--- a/cc/resources/resource_provider.cc
+++ b/cc/resources/resource_provider.cc
@@ -86,6 +86,9 @@ GLenum TextureToStorageFormat(ResourceFormat format) {
     case BGRA_8888:
       storage_format = GL_BGRA8_EXT;
       break;
+    case RGBA_F16:
+      storage_format = GL_RGBA16F_EXT;
+      break;
     case RGBA_4444:
     case ALPHA_8:
     case LUMINANCE_8:
@@ -103,6 +106,7 @@ GLenum TextureToStorageFormat(ResourceFormat format) {
 bool IsFormatSupportedForStorage(ResourceFormat format, bool use_bgra) {
   switch (format) {
     case RGBA_8888:
+    case RGBA_F16:
       return true;
     case BGRA_8888:
       return use_bgra;
@@ -126,6 +130,8 @@ GrPixelConfig ToGrPixelConfig(ResourceFormat format) {
       return kBGRA_8888_GrPixelConfig;
     case RGBA_4444:
       return kRGBA_4444_GrPixelConfig;
+    case RGBA_F16:
+      return kRGBA_half_GrPixelConfig;
     default:
       break;
   }
@@ -540,6 +546,11 @@ bool
ResourceProvider::IsResourceFormatSupported(ResourceFormat format) const {
       return caps.texture_rg;
     case LUMINANCE_F16:
       return caps.texture_half_float_linear;
+    case RGBA_F16:
+      // TODO(ccameron): This will always return false on pixel tests, which
+      // makes it un-test-able until we upgrade Mesa.
+      // https://crbug.com/687720
+      return caps.texture_half_float_linear && caps.color_buffer_float;
   }
 
   NOTREACHED();
@@ -579,8 +590,8 @@ ResourceId ResourceProvider::CreateResource(
   DCHECK(!size.IsEmpty());
   switch (default_resource_type_) {
     case RESOURCE_TYPE_GPU_MEMORY_BUFFER:
-      // GPU memory buffers don't support LUMINANCE_F16.
-      if (format != LUMINANCE_F16) {
+      // GPU memory buffers don't support LUMINANCE_F16 or RGBA_F16 yet.
+      if (format != LUMINANCE_F16 && format != RGBA_F16) {
         return CreateGLTexture(
             size, hint, RESOURCE_TYPE_GPU_MEMORY_BUFFER, format,
             gfx::BufferUsage::GPU_READ_CPU_READ_WRITE, color_space);

[ccameron, 2017-02-04 00:07:39]

The CQ bit was checked by ccameron@chromium.org

[ccameron, 2017-02-04 00:07:40]

The patchset sent to the CQ was uploaded after l-g-t-m from tsepez@chromium.org,
jbauman@chromium.org, enne@chromium.org
Link to the patchset: https://codereview.chromium.org/2674493003/#ps60001
(title: "Rebase agin")

[commit-bot: I haz the power, 2017-02-04 00:08:01]

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

[commit-bot: I haz the power, 2017-02-04 02:07:38]

CQ is committing da patch.

Bot data: {"patchset_id": 60001, "attempt_start_ts": 1486166859792310,
"parent_rev": "3ceab9f0a58bef737f4685dc030757cbb330d078", "commit_rev":
"197ec6b2ea2af1d95c0acdf9a0ebf0cd479fdf40"}

[commit-bot: I haz the power, 2017-02-04 02:08:57]

Description was changed from

==========
Add the ResourceFormat RGBA_F16, and appropriate GL and Skia values.

Add the GPU capability color_buffer_float, to expose if it is possible
to create a RGBA_F16 render target.

Enable exposing GPU capabilities color_buffer_float and
texture_half_float_linear to non-WebGL GLES2 contexts (e.g, the
compositor context).

When allocating a RenderPass texture, use RGBA_F16 if the output surface
is high-bit-depth (which will include HDR surfaces). Add the method
OutputSurface::IsHighBitDepth to indicate this property.

Note that this cannot be tested by the cc pixel tests (the correct way
to test this) because they use a version of Mesa that does not support
floating-point textures. Tests so far have been manual, as we work on
upgrading Mesa. This will be tested by having the pixel tests which
currently run in software and GL mode also run in "GL with high bit
depth output surface" mode.

BUG=687346
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_trusty_blink_rel
==========

to

==========
Add the ResourceFormat RGBA_F16, and appropriate GL and Skia values.

Add the GPU capability color_buffer_float, to expose if it is possible
to create a RGBA_F16 render target.

Enable exposing GPU capabilities color_buffer_float and
texture_half_float_linear to non-WebGL GLES2 contexts (e.g, the
compositor context).

When allocating a RenderPass texture, use RGBA_F16 if the output surface
is high-bit-depth (which will include HDR surfaces). Add the method
OutputSurface::IsHighBitDepth to indicate this property.

Note that this cannot be tested by the cc pixel tests (the correct way
to test this) because they use a version of Mesa that does not support
floating-point textures. Tests so far have been manual, as we work on
upgrading Mesa. This will be tested by having the pixel tests which
currently run in software and GL mode also run in "GL with high bit
depth output surface" mode.

BUG=687346
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_trusty_blink_rel

Review-Url: https://codereview.chromium.org/2674493003
Cr-Commit-Position: refs/heads/master@{#448143}
Committed:
https://chromium.googlesource.com/chromium/src/+/197ec6b2ea2af1d95c0acdf9a0eb...
==========

[commit-bot: I haz the power, 2017-02-04 02:08:59]

Committed patchset #4 (id:60001) as
https://chromium.googlesource.com/chromium/src/+/197ec6b2ea2af1d95c0acdf9a0eb...

[Ken Russell (switch to Gerrit), 2017-02-05 00:33:11]

A revert of this CL (patchset #4 id:60001) has been created in
https://codereview.chromium.org/2676153002/ by kbr@chromium.org.

The reason for reverting is: Broke all Windows bots on chromium.gpu.fyi
waterfall; see https://bugs.chromium.org/p/chromium/issues/detail?id=687346#c4 .
.

[ccameron, 2017-02-06 18:59:22]

Description was changed from

==========
Add the ResourceFormat RGBA_F16, and appropriate GL and Skia values.

Add the GPU capability color_buffer_float, to expose if it is possible
to create a RGBA_F16 render target.

Enable exposing GPU capabilities color_buffer_float and
texture_half_float_linear to non-WebGL GLES2 contexts (e.g, the
compositor context).

When allocating a RenderPass texture, use RGBA_F16 if the output surface
is high-bit-depth (which will include HDR surfaces). Add the method
OutputSurface::IsHighBitDepth to indicate this property.

Note that this cannot be tested by the cc pixel tests (the correct way
to test this) because they use a version of Mesa that does not support
floating-point textures. Tests so far have been manual, as we work on
upgrading Mesa. This will be tested by having the pixel tests which
currently run in software and GL mode also run in "GL with high bit
depth output surface" mode.

BUG=687346
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_trusty_blink_rel

Review-Url: https://codereview.chromium.org/2674493003
Cr-Commit-Position: refs/heads/master@{#448143}
Committed:
https://chromium.googlesource.com/chromium/src/+/197ec6b2ea2af1d95c0acdf9a0eb...
==========

to

==========
Add the ResourceFormat RGBA_F16, and appropriate GL and Skia values.

Add the GPU capability color_buffer_float, to expose if it is possible
to create a RGBA_F16 render target.

Enable exposing GPU capabilities color_buffer_float and
texture_half_float_linear to non-WebGL GLES2 contexts (e.g, the
compositor context).

When allocating a RenderPass texture, use RGBA_F16 if the output surface
is high-bit-depth (which will include HDR surfaces). Add the method
OutputSurface::IsHighBitDepth to indicate this property.

Note that this cannot be tested by the cc pixel tests (the correct way
to test this) because they use a version of Mesa that does not support
floating-point textures. Tests so far have been manual, as we work on
upgrading Mesa. This will be tested by having the pixel tests which
currently run in software and GL mode also run in "GL with high bit
depth output surface" mode.

BUG=687346
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_trusty_blink_rel

Review-Url: https://codereview.chromium.org/2674493003
Cr-Commit-Position: refs/heads/master@{#448143}
Committed:
https://chromium.googlesource.com/chromium/src/+/197ec6b2ea2af1d95c0acdf9a0eb...
==========

[ccameron, 2017-02-06 19:09:59]

Description was changed from

==========
Add the ResourceFormat RGBA_F16, and appropriate GL and Skia values.

Add the GPU capability color_buffer_float, to expose if it is possible
to create a RGBA_F16 render target.

Enable exposing GPU capabilities color_buffer_float and
texture_half_float_linear to non-WebGL GLES2 contexts (e.g, the
compositor context).

When allocating a RenderPass texture, use RGBA_F16 if the output surface
is high-bit-depth (which will include HDR surfaces). Add the method
OutputSurface::IsHighBitDepth to indicate this property.

Note that this cannot be tested by the cc pixel tests (the correct way
to test this) because they use a version of Mesa that does not support
floating-point textures. Tests so far have been manual, as we work on
upgrading Mesa. This will be tested by having the pixel tests which
currently run in software and GL mode also run in "GL with high bit
depth output surface" mode.

BUG=687346
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_trusty_blink_rel

Review-Url: https://codereview.chromium.org/2674493003
Cr-Commit-Position: refs/heads/master@{#448143}
Committed:
https://chromium.googlesource.com/chromium/src/+/197ec6b2ea2af1d95c0acdf9a0eb...
==========

to

==========
Add the ResourceFormat RGBA_F16, and appropriate GL and Skia values.

Add the GPU capability color_buffer_float, to expose if it is possible
to create a RGBA_F16 render target.

Enable exposing GPU capabilities color_buffer_float and
texture_half_float_linear to non-WebGL GLES2 contexts (e.g, the
compositor context).

When allocating a RenderPass texture, use RGBA_F16 if the output surface
is high-bit-depth (which will include HDR surfaces). Add the method
OutputSurface::IsHighBitDepth to indicate this property.

Note that this cannot be tested by the cc pixel tests (the correct way
to test this) because they use a version of Mesa that does not support
floating-point textures. Tests so far have been manual, as we work on
upgrading Mesa. This will be tested by having the pixel tests which
currently run in software and GL mode also run in "GL with high bit
depth output surface" mode.

BUG=687346
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_trusty_blink_rel;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


Review-Url: https://codereview.chromium.org/2674493003
Cr-Commit-Position: refs/heads/master@{#448143}
Committed:
https://chromium.googlesource.com/chromium/src/+/197ec6b2ea2af1d95c0acdf9a0eb...
==========

[ccameron, 2017-02-06 19:10:22]

Description was changed from

==========
Add the ResourceFormat RGBA_F16, and appropriate GL and Skia values.

Add the GPU capability color_buffer_float, to expose if it is possible
to create a RGBA_F16 render target.

Enable exposing GPU capabilities color_buffer_float and
texture_half_float_linear to non-WebGL GLES2 contexts (e.g, the
compositor context).

When allocating a RenderPass texture, use RGBA_F16 if the output surface
is high-bit-depth (which will include HDR surfaces). Add the method
OutputSurface::IsHighBitDepth to indicate this property.

Note that this cannot be tested by the cc pixel tests (the correct way
to test this) because they use a version of Mesa that does not support
floating-point textures. Tests so far have been manual, as we work on
upgrading Mesa. This will be tested by having the pixel tests which
currently run in software and GL mode also run in "GL with high bit
depth output surface" mode.

BUG=687346
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_trusty_blink_rel;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


Review-Url: https://codereview.chromium.org/2674493003
Cr-Commit-Position: refs/heads/master@{#448143}
Committed:
https://chromium.googlesource.com/chromium/src/+/197ec6b2ea2af1d95c0acdf9a0eb...
==========

to

==========
Add the ResourceFormat RGBA_F16, and appropriate GL and Skia values.

Add the GPU capability color_buffer_float, to expose if it is possible
to create a RGBA_F16 render target.

When allocating a RenderPass texture, use RGBA_F16 if the output surface
is high-bit-depth (which will include HDR surfaces). Add the method
OutputSurface::IsHighBitDepth to indicate this property.

Note that this cannot be tested by the cc pixel tests (the correct way
to test this) because they use a version of Mesa that does not support
floating-point textures. Tests so far have been manual, as we work on
upgrading Mesa. This will be tested by having the pixel tests which
currently run in software and GL mode also run in "GL with high bit
depth output surface" mode.

BUG=687346
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_trusty_blink_rel;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


Review-Url: https://codereview.chromium.org/2674493003
Cr-Commit-Position: refs/heads/master@{#448143}
Committed:
https://chromium.googlesource.com/chromium/src/+/197ec6b2ea2af1d95c0acdf9a0eb...
==========

[ccameron, 2017-02-06 19:10:41]

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

[commit-bot: I haz the power, 2017-02-06 19:11:19]

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

[ccameron, 2017-02-06 19:15:02]

I've removed the part that enables this on non-es3 contexts.

We'll need to enable this on non-es3 contexts, but the patch to do that will be
smaller.

Not sure what was causing the issues on Windows (and those particular Windows
bots), but I'd put my first bet on it being R11F_G11F_B10F as an invalid enum.

[commit-bot: I haz the power, 2017-02-06 20:17:51]

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

[commit-bot: I haz the power, 2017-02-06 20:17:52]

Dry run: This issue passed the CQ dry run.

[ccameron, 2017-02-06 20:23:06]

The CQ bit was checked by ccameron@chromium.org

[ccameron, 2017-02-06 20:23:06]

The patchset sent to the CQ was uploaded after l-g-t-m from tsepez@chromium.org,
jbauman@chromium.org, enne@chromium.org
Link to the patchset: https://codereview.chromium.org/2674493003/#ps80001
(title: "Don't enable on non-ES3 yet")

[commit-bot: I haz the power, 2017-02-06 20:23:33]

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

[commit-bot: I haz the power, 2017-02-06 20:29:26]

CQ is committing da patch.

Bot data: {"patchset_id": 80001, "attempt_start_ts": 1486412586037130,
"parent_rev": "a3fc66293215cbdfce5a9aaa62bafc2ae80534eb", "commit_rev":
"22094236e8ef012d05e313c6e77c07a71addb37e"}

[commit-bot: I haz the power, 2017-02-06 20:30:00]

Description was changed from

==========
Add the ResourceFormat RGBA_F16, and appropriate GL and Skia values.

Add the GPU capability color_buffer_float, to expose if it is possible
to create a RGBA_F16 render target.

When allocating a RenderPass texture, use RGBA_F16 if the output surface
is high-bit-depth (which will include HDR surfaces). Add the method
OutputSurface::IsHighBitDepth to indicate this property.

Note that this cannot be tested by the cc pixel tests (the correct way
to test this) because they use a version of Mesa that does not support
floating-point textures. Tests so far have been manual, as we work on
upgrading Mesa. This will be tested by having the pixel tests which
currently run in software and GL mode also run in "GL with high bit
depth output surface" mode.

BUG=687346
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_trusty_blink_rel;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


Review-Url: https://codereview.chromium.org/2674493003
Cr-Commit-Position: refs/heads/master@{#448143}
Committed:
https://chromium.googlesource.com/chromium/src/+/197ec6b2ea2af1d95c0acdf9a0eb...
==========

to

==========
Add the ResourceFormat RGBA_F16, and appropriate GL and Skia values.

Add the GPU capability color_buffer_float, to expose if it is possible
to create a RGBA_F16 render target.

When allocating a RenderPass texture, use RGBA_F16 if the output surface
is high-bit-depth (which will include HDR surfaces). Add the method
OutputSurface::IsHighBitDepth to indicate this property.

Note that this cannot be tested by the cc pixel tests (the correct way
to test this) because they use a version of Mesa that does not support
floating-point textures. Tests so far have been manual, as we work on
upgrading Mesa. This will be tested by having the pixel tests which
currently run in software and GL mode also run in "GL with high bit
depth output surface" mode.

BUG=687346
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_trusty_blink_rel;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


Review-Url: https://codereview.chromium.org/2674493003
Cr-Original-Commit-Position: refs/heads/master@{#448143}
Committed:
https://chromium.googlesource.com/chromium/src/+/197ec6b2ea2af1d95c0acdf9a0eb...
Review-Url: https://codereview.chromium.org/2674493003
Cr-Commit-Position: refs/heads/master@{#448375}
Committed:
https://chromium.googlesource.com/chromium/src/+/22094236e8ef012d05e313c6e77c...
==========

[commit-bot: I haz the power, 2017-02-06 20:30:02]

Committed patchset #5 (id:80001) as
https://chromium.googlesource.com/chromium/src/+/22094236e8ef012d05e313c6e77c...