Clean up GLSurfaceCGL

Clean up GLSurfaceCGL

This is in preparation for support for using remote CALayers.

The big change is to separate out ImageTransportSurfaceIOSurface into
ImageTransportSurfaceFBO and its StorageProvider. The IOSurface-
specific bits of ImageTransportSurfaceIOSurface are moved into the
StorageProvider. A separate CALayer-specific version will be added
later.

While in the neighborhood, don't make everything inherit from
GLSurfaceCGL (that class doesn't actually do anything). Move the helper
functions and classes from its file to more appropriate locations.

Make image_transport_surface_mac.cc a .mm file, because it will need
to include Objective C code in the near future. Move the
ImageTransportSurfaceFBO into its own file.

Also make the IOSurface handles in renderer_host use the 32-bit
IOSurfaceID handle type instead of uint64, since that is the type
actually used by the interface, and those other bits may find use in
disambiguating handle types.

BUG=312462
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=277970

[ccameron, 2014-06-17 19:17:20]

The actual diffs here are very small (promise, I know it doesn't look like it)
-- it's mostly just code moving around.

The meat of the direction I'm going with this is in
https://codereview.chromium.org/337303003.

https://codereview.chromium.org/334173006/diff/40001/content/common/gpu/image...
File content/common/gpu/image_transport_surface_fbo_mac.h (right):

https://codereview.chromium.org/334173006/diff/40001/content/common/gpu/image...
content/common/gpu/image_transport_surface_fbo_mac.h:44: // display.
We'll probably need a more substantial interface here when I add the CALayer
stuff (to, at a minimum, distinguish between the two), but this is the minimum
to work with IOSurfaces.

https://codereview.chromium.org/334173006/diff/40001/ui/gl/gl_surface_cgl.h
File ui/gl/gl_surface_cgl.h (left):

https://codereview.chromium.org/334173006/diff/40001/ui/gl/gl_surface_cgl.h#o...
ui/gl/gl_surface_cgl.h:18: static bool InitializeOneOff();
InitializeOneOff is moved into anonymous namespace in gl_surface_mac (the only
place it was used). GetPixelFormat was dead code.

https://codereview.chromium.org/334173006/diff/40001/ui/gl/gl_surface_cgl.h#o...
ui/gl/gl_surface_cgl.h:32: class GL_EXPORT NoOpGLSurfaceCGL : public
GLSurfaceCGL {
This was moved to anonymous namespace in gl_surface_mac where it may be
instantiated for testing.

[ccameron, 2014-06-18 00:18:38]

Adding piman for the content/browser/compositor bits (s/uint64/IOSurfaceID). I'm
going to want to have a central location for that conversion and one other
(IOSurfaceID versus CAContextID), but that can wait.

[ccameron, 2014-06-18 00:18:55]

Actually adding piman

[piman, 2014-06-18 00:39:55]

lgtm

https://codereview.chromium.org/334173006/diff/140001/content/browser/rendere...
File content/browser/renderer_host/render_widget_helper_mac.mm (right):

https://codereview.chromium.org/334173006/diff/140001/content/browser/rendere...
content/browser/renderer_host/render_widget_helper_mac.mm:24: IOSurfaceID
io_surface_handle = static_cast<uint32>(params.surface_handle);
nit: static_cast<IOSurfaceID> ?

[Ken Russell (switch to Gerrit), 2014-06-18 01:24:29]

lgtm

[ccameron, 2014-06-18 02:22:34]

Thanks -- done

[ccameron, 2014-06-18 02:22:39]

The CQ bit was checked by ccameron@chromium.org

[commit-bot: I haz the power, 2014-06-18 02:24:37]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/ccameron@chromium.org/334173006/160001

[commit-bot: I haz the power, 2014-06-18 07:30:41]

Change committed as 277970