cc: Fix readback from non-root layers.

cc: Fix readback from non-root layers.

The readback was assuming that it was reading from the root render
pass, so using the full viewport size when flipping the result,
but it should use the size of the currently bound framebuffer. Also,
it should not flip if the readback is already in the correct
orientation.

Tests:
LayerTreeHostReadbackPixelTest.ReadbackRootLayer
LayerTreeHostReadbackPixelTest.ReadbackRootLayerWithChild
LayerTreeHostReadbackPixelTest.ReadbackNonRootLayer
LayerTreeHostReadbackPixelTest.ReadbackSmallNonRootLayer
LayerTreeHostReadbackPixelTest.ReadbackSmallNonRootLayerWithChild
RendererPixelTest.SimpleGreenRect_NonRootRenderPass

R=piman@chromium.org
BUG=179896
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=199055

[danakj, 2013-05-07 21:36:14]

This depends on green_small.png which cannot be included here for the new tests.
It is http://i.imgur.com/wJNKXvX.png

[danakj, 2013-05-07 22:00:26]

New patch also fixes y-flipping for non-root layers.

[danakj, 2013-05-07 22:03:28]

New pngs are:

green small - http://i.imgur.com/nw255b7.png
green small with blue corner - http://i.imgur.com/7Omq8aN.png

[piman, 2013-05-07 22:35:02]

Mostly LGTM.

https://codereview.chromium.org/15004009/diff/14001/cc/output/gl_renderer.cc
File cc/output/gl_renderer.cc (right):

https://codereview.chromium.org/15004009/diff/14001/cc/output/gl_renderer.cc#...
cc/output/gl_renderer.cc:2018: bool flipped_y = FlippedFramebuffer();
nit: this is always true

https://codereview.chromium.org/15004009/diff/14001/cc/output/gl_renderer.cc#...
cc/output/gl_renderer.cc:2288: SetDrawViewportSize(ViewportSize());
This should already be done by DirectRenderer::UseRenderPass which is the only
caller of BindFramebufferToOutputSurface AFAICT

[danakj, 2013-05-07 22:36:18]

https://codereview.chromium.org/15004009/diff/14001/cc/output/gl_renderer.cc
File cc/output/gl_renderer.cc (right):

https://codereview.chromium.org/15004009/diff/14001/cc/output/gl_renderer.cc#...
cc/output/gl_renderer.cc:2018: bool flipped_y = FlippedFramebuffer();
On 2013/05/07 22:35:02, piman wrote:
> nit: this is always true

That is true, but I prefer the context so you can see how this relates to
InitializeMatrices?

https://codereview.chromium.org/15004009/diff/14001/cc/output/gl_renderer.cc#...
cc/output/gl_renderer.cc:2288: SetDrawViewportSize(ViewportSize());
On 2013/05/07 22:35:02, piman wrote:
> This should already be done by DirectRenderer::UseRenderPass which is the only
> caller of BindFramebufferToOutputSurface AFAICT

oh right, i was wondering where it was.

[piman, 2013-05-07 22:42:35]

https://codereview.chromium.org/15004009/diff/14001/cc/output/gl_renderer.cc
File cc/output/gl_renderer.cc (right):

https://codereview.chromium.org/15004009/diff/14001/cc/output/gl_renderer.cc#...
cc/output/gl_renderer.cc:2018: bool flipped_y = FlippedFramebuffer();
On 2013/05/07 22:36:19, danakj wrote:
> On 2013/05/07 22:35:02, piman wrote:
> > nit: this is always true
> 
> That is true, but I prefer the context so you can see how this relates to
> InitializeMatrices?

I guess it's not trivial because InitializeMatrices gets called from
DirectRenderer, and it's hard to follow from here... Whatever you want.

[commit-bot: I haz the power, 2013-05-07 22:45:28]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/danakj@chromium.org/15004009/19001

[commit-bot: I haz the power, 2013-05-07 23:50:24]

Retried try job too often on linux_rel for step(s) cc_unittests
http://build.chromium.org/p/tryserver.chromium/buildstatus?builder=linux_rel&...

[commit-bot: I haz the power, 2013-05-07 23:57:14]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/danakj@chromium.org/15004009/35001

[commit-bot: I haz the power, 2013-05-08 14:39:08]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/danakj@chromium.org/15004009/35001

[commit-bot: I haz the power, 2013-05-08 15:01:44]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/danakj@chromium.org/15004009/35001

[commit-bot: I haz the power, 2013-05-08 19:24:20]

Failed to trigger a try job on android_dbg
HTTP Error 400: Bad Request

[commit-bot: I haz the power, 2013-05-08 19:24:42]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/danakj@chromium.org/15004009/58001

[danakj, 2013-05-08 23:52:19]

Committed patchset #10 manually as r199055 (presubmit successful).