Replace WGC3D visibility extension with resource_usage extension. [Part 2 of 3]

Replace WGC3D visibility extension with resource_usage extension. [Part 2 of 3]

This is the second of a three part patch (WebKit, Chromium, Webkit-Touchups) to replace the WebGraphicsContext3D setVisibility extension with a setResourceUsage extension that is both more accurate and more flexible.
This change is a prereq for some other work on reducing resource usage of background tabs in certain scenarios.


BUG=None
TEST=Manual

[mmocny, 2012-01-19 15:46:43]

Part 1(webkit patch) is https://bugs.webkit.org/show_bug.cgi?id=76634.

Ken, Al has been more in the loop with my work in this area and I think overlaps
with you on OWNERS here.  Since I asked you to review the webkit patch I've
added you here, too.

[jonathan.backer, 2012-01-19 16:11:19]

http://codereview.chromium.org/9192014/diff/3001/ui/gfx/gl/gl_surface.h
File ui/gfx/gl/gl_surface.h (right):

http://codereview.chromium.org/9192014/diff/3001/ui/gfx/gl/gl_surface.h#newco...
ui/gfx/gl/gl_surface.h:83: virtual void SetResourceUsage(ResourceUsage
resourceUsage);
A lot has changed since SetVisible was introduced. I think that it's possible to
make SetResourceUsage an ImageTransportSurface only method because it's being
called by GpuCommandBufferStub, which now knows it's an ImageTransportSurface.

[mmocny, 2012-01-19 16:14:07]

This is excellent.  I will make the change.

On 2012/01/19 16:11:19, jonathan.backer wrote:
> http://codereview.chromium.org/9192014/diff/3001/ui/gfx/gl/gl_surface.h
> File ui/gfx/gl/gl_surface.h (right):
> 
>
http://codereview.chromium.org/9192014/diff/3001/ui/gfx/gl/gl_surface.h#newco...
> ui/gfx/gl/gl_surface.h:83: virtual void SetResourceUsage(ResourceUsage
> resourceUsage);
> A lot has changed since SetVisible was introduced. I think that it's possible
to
> make SetResourceUsage an ImageTransportSurface only method because it's being
> called by GpuCommandBufferStub, which now knows it's an ImageTransportSurface.

[mmocny, 2012-01-19 17:40:08]

An ImageTransportSurface is not itself a GLSurface, and changing
GpuCommandBufferStub to using an ImageTransportSurface was uglier than the
current state.

It is true that at the moment only the ImageTransportSurfaces will make sure of
the SetResourceUsage call, but leaving others with stubbed implementations is
clearly the lesser evil here.

On 2012/01/19 16:14:07, mmocny wrote:
> This is excellent.  I will make the change.
> 
> On 2012/01/19 16:11:19, jonathan.backer wrote:
> > http://codereview.chromium.org/9192014/diff/3001/ui/gfx/gl/gl_surface.h
> > File ui/gfx/gl/gl_surface.h (right):
> > 
> >
>
http://codereview.chromium.org/9192014/diff/3001/ui/gfx/gl/gl_surface.h#newco...
> > ui/gfx/gl/gl_surface.h:83: virtual void SetResourceUsage(ResourceUsage
> > resourceUsage);
> > A lot has changed since SetVisible was introduced. I think that it's
possible
> to
> > make SetResourceUsage an ImageTransportSurface only method because it's
being
> > called by GpuCommandBufferStub, which now knows it's an
ImageTransportSurface.

[Ken Russell (switch to Gerrit), 2012-01-19 18:31:19]

This seems fine to me. Since it has to wait for the webkit patch to land and
DEPS rolled forward, please send try jobs and ask for re-review once that's
done.