Delay resize of the front buffer until the swap for offscreen buffers.

Delay resize of the front buffer until the swap for offscreen buffers.

Previous code was resizing, and clearing the front buffer as soon as the back
buffer was resized. It can lead to race conditions where the compositor draws a
black texture if it triggers between the resize and the swap.

BUG=None
TEST=Pepper Flash


Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=98905

[piman, 2011-08-30 02:48:12]

Ken: I hope this is not going to affect your Mac/NVIDIA workaround. If anything
I'm hoping it will make it better (since the frame buffer recreated later, so is
less likely to be still in use).

[apatrick_chromium, 2011-08-30 19:40:47]

I think it's right. LGTM.

[Ken Russell (switch to Gerrit), 2011-08-30 22:47:13]

LGTM. Not too concerned about interactions with the NVIDIA workaround. One
comment.

http://codereview.chromium.org/7795015/diff/1/gpu/command_buffer/service/gles...
File gpu/command_buffer/service/gles2_cmd_decoder.cc (right):

http://codereview.chromium.org/7795015/diff/1/gpu/command_buffer/service/gles...
gpu/command_buffer/service/gles2_cmd_decoder.cc:6574: // TODO(piman): Is this
still necessary?
I'm quite sure it is -- if the offscreen color texture has increased in width or
height then the copy from the frame buffer will not cover it all, so we must
prevent garbage from showing up in the other regions.

[piman, 2011-08-30 22:50:58]

On Tue, Aug 30, 2011 at 3:47 PM, <kbr@chromium.org> wrote:

> LGTM. Not too concerned about interactions with the NVIDIA workaround. One
> comment.
>
>
>
> http://codereview.chromium.**org/7795015/diff/1/gpu/**
>
command_buffer/service/gles2_**cmd_decoder.cc<http://codereview.chromium.org/7795015/diff/1/gpu/command_buffer/service/gles2_cmd_decoder.cc>
> File gpu/command_buffer/service/**gles2_cmd_decoder.cc (right):
>
> http://codereview.chromium.**org/7795015/diff/1/gpu/**
>
command_buffer/service/gles2_**cmd_decoder.cc#newcode6574<http://codereview.chromium.org/7795015/diff/1/gpu/command_buffer/service/gles2_cmd_decoder.cc#newcode6574>
> gpu/command_buffer/service/**gles2_cmd_decoder.cc:6574: // TODO(piman): Is
> this still necessary?
> I'm quite sure it is -- if the offscreen color texture has increased in
> width or height then the copy from the frame buffer will not cover it
> all, so we must prevent garbage from showing up in the other regions.


But at that point, we just resized the "front" buffer to be the size of the
"back" buffer, right ?
Note, I'm happy to leave the code if there's any reason, just trying to
understand.


>
>
http://codereview.chromium.**org/7795015/<http://codereview.chromium.org/7795...
>

[Ken Russell (switch to Gerrit), 2011-08-30 23:00:32]

On 2011/08/30 22:50:58, piman wrote:
> On Tue, Aug 30, 2011 at 3:47 PM, <mailto:kbr@chromium.org> wrote:
> 
> > LGTM. Not too concerned about interactions with the NVIDIA workaround. One
> > comment.
> >
> >
> >
> > http://codereview.chromium.**org/7795015/diff/1/gpu/**
> >
>
command_buffer/service/gles2_**cmd_decoder.cc<http://codereview.chromium.org/7795015/diff/1/gpu/command_buffer/service/gles2_cmd_decoder.cc>
> > File gpu/command_buffer/service/**gles2_cmd_decoder.cc (right):
> >
> > http://codereview.chromium.**org/7795015/diff/1/gpu/**
> >
>
command_buffer/service/gles2_**cmd_decoder.cc#newcode6574<http://codereview.chromium.org/7795015/diff/1/gpu/command_buffer/service/gles2_cmd_decoder.cc#newcode6574>
> > gpu/command_buffer/service/**gles2_cmd_decoder.cc:6574: // TODO(piman): Is
> > this still necessary?
> > I'm quite sure it is -- if the offscreen color texture has increased in
> > width or height then the copy from the frame buffer will not cover it
> > all, so we must prevent garbage from showing up in the other regions.
> 
> 
> But at that point, we just resized the "front" buffer to be the size of the
> "back" buffer, right ?
> Note, I'm happy to leave the code if there's any reason, just trying to
> understand.

Upon further thought, yes, you're right.

[commit-bot: I haz the power, 2011-08-31 01:01:57]

Change committed as 98905