[webkit-reviews] review granted: [Bug 131096] Pool IOSurfaces to help with allocation cost : [Attachment 228431] patch

bugzilla-daemon at webkit.org bugzilla-daemon at webkit.org
Wed Apr 2 15:47:57 PDT 2014


Simon Fraser (smfr) <simon.fraser at apple.com> has granted Tim Horton
<thorton at apple.com>'s request for review:
Bug 131096: Pool IOSurfaces to help with allocation cost
https://bugs.webkit.org/show_bug.cgi?id=131096

Attachment 228431: patch
https://bugs.webkit.org/attachment.cgi?id=228431&action=review

------- Additional Comments from Simon Fraser (smfr) <simon.fraser at apple.com>
View in context: https://bugs.webkit.org/attachment.cgi?id=228431&action=review


> Source/WebCore/platform/graphics/cg/IOSurfacePool.cpp:108
> +	   showPoolStatistics();

Do you really want all these showPoolStatistics calls?

> Source/WebCore/platform/graphics/cg/IOSurfacePool.cpp:112
> +    for (auto surfaceIter = mapIter->value.begin(); surfaceIter !=
mapIter->value.end(); ++surfaceIter) {

auto& ? I never know.

> Source/WebCore/platform/graphics/cg/IOSurfacePool.cpp:114
> +	   RefPtr<IOSurface> surface = surfaceIter->get();
> +	   if (!surfaceMatchesParameters(*surface, size, colorSpace))

Do we have to pay the ref() cost by making the refptr when we don't actually
use the surface?

> Source/WebCore/platform/graphics/cg/IOSurfacePool.cpp:137
> +	   RefPtr<IOSurface> surface = surfaceIter->get();
> +	   if (!surfaceMatchesParameters(*surface, size, colorSpace))

Ditto.

> Source/WebCore/platform/graphics/cg/IOSurfacePool.h:91
> +    void scheduleCollectionTimer();
> +    void collectionTimerFired(Timer<IOSurfacePool>&);
> +    void collectInUseSurfaces();

Is this "collect" in the garbage collection sense? Not clear from the context.


More information about the webkit-reviews mailing list