cc: Be robust against invalid RenderPassDrawQuads.

cc: Be robust against invalid RenderPassDrawQuads.

When the LayerTreeHostImpl drops RenderPasses from its output, it does
not also drop the RenderPassDrawQuads that point to them. The direct
renderers ignore these quads, so this has never been a problem. However
the DelegatedRendererLayerImpl was not able to deal with this correctly
and would crash. More problematic is that a compromised renderer could
send invalid RenderPassDrawQuads, which the DelegatedRendererLayerImpl
must be able to handle gracefully.

So, for both cases, we here make the DelegatedRendererLayerImpl ignore
invalid RenderPassDrawQuads, and just drop the from its own output.

Tests:
DelegatedRendererLayerImplTest.InvalidRenderPassDrawQuad

R=piman
BUG=283630
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=221493

[piman, 2013-09-04 19:03:11]

LGTM, nice!

https://codereview.chromium.org/23891003/diff/1/cc/layers/delegated_renderer_...
File cc/layers/delegated_renderer_layer_impl.cc (right):

https://codereview.chromium.org/23891003/diff/1/cc/layers/delegated_renderer_...
cc/layers/delegated_renderer_layer_impl.cc:244: // frame and can be converted.
nit: should this be in the header instead?

https://codereview.chromium.org/23891003/diff/1/cc/layers/delegated_renderer_...
File cc/layers/delegated_renderer_layer_impl_unittest.cc (right):

https://codereview.chromium.org/23891003/diff/1/cc/layers/delegated_renderer_...
cc/layers/delegated_renderer_layer_impl_unittest.cc:1250:
host_impl_->active_tree(), 1).PassAs<LayerImpl>();
nit: this one doesn't draw content, right? Should we make it a LayerImpl
directly? I was getting confused below whether the expected quad was from the
DRL or this.

[commit-bot: I haz the power, 2013-09-05 16:46:52]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/danakj@chromium.org/23891003/12001

[danakj, 2013-09-05 16:46:55]

https://codereview.chromium.org/23891003/diff/1/cc/layers/delegated_renderer_...
File cc/layers/delegated_renderer_layer_impl.cc (right):

https://codereview.chromium.org/23891003/diff/1/cc/layers/delegated_renderer_...
cc/layers/delegated_renderer_layer_impl.cc:244: // frame and can be converted.
On 2013/09/04 19:03:12, piman wrote:
> nit: should this be in the header instead?

Ya.. thanks.

https://codereview.chromium.org/23891003/diff/1/cc/layers/delegated_renderer_...
File cc/layers/delegated_renderer_layer_impl_unittest.cc (right):

https://codereview.chromium.org/23891003/diff/1/cc/layers/delegated_renderer_...
cc/layers/delegated_renderer_layer_impl_unittest.cc:1250:
host_impl_->active_tree(), 1).PassAs<LayerImpl>();
On 2013/09/04 19:03:12, piman wrote:
> nit: this one doesn't draw content, right? Should we make it a LayerImpl
> directly? I was getting confused below whether the expected quad was from the
> DRL or this.

Oh, ya. That's a great idea. Thank you.

[commit-bot: I haz the power, 2013-09-05 19:29:53]

Change committed as 221493