[webkit-reviews] review granted: [Bug 218177] REGRESSION (r268386): Flashes of inverted color when zooming the map on windy.com : [Attachment 412349] Patch

[bugzilla-daemon at webkit.org]

Dean Jackson <dino at apple.com> has granted Kimmo Kinnunen
<kkinnunen at apple.com>'s request for review:
Bug 218177: REGRESSION (r268386): Flashes of inverted color when zooming the
map on windy.com
https://bugs.webkit.org/show_bug.cgi?id=218177

Attachment 412349: Patch

https://bugs.webkit.org/attachment.cgi?id=412349&action=review




--- Comment #15 from Dean Jackson <dino at apple.com> ---
Comment on attachment 412349
  --> https://bugs.webkit.org/attachment.cgi?id=412349
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=412349&action=review

>> Source/WebCore/ChangeLog:12
>> +	    draw on top of the IOSurface even if CA was using it.
> 
> It's not clear to me that prior to r268386 / Bug 217212, that this was the
case. WebGLLayer previously contained three IOSurfaces and used a
triple-buffering scheme among them.

Yeah. The reason for the third buffer was due to the call to display happening
when the current buffer was still locked, and thus we were unable to swap it
into the slot for rendering the next frame. That issue may have gone away when
we moved to doing "prepareForDisplay" or when Kimmo cleaned up the buffer
ownership.

The effect was pretty clear - lots of blank or corrupt frames, so I'm willing
to trust the outcome of this patch if it is stable. Please check on heavy
workloads though. Fullscreen YouTube 360 video is a good one.