Bug 60864 - Suboptimal lifetime of BackingStore objects in DrawingAreaProxyImpl
Summary: Suboptimal lifetime of BackingStore objects in DrawingAreaProxyImpl
Status: RESOLVED FIXED
Alias: None
Product: WebKit
Classification: Unclassified
Component: WebKit2 (show other bugs)
Version: 528+ (Nightly build)
Hardware: All Unspecified
: P2 Normal
Assignee: Michael Saboff
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2011-05-15 17:55 PDT by Michael Saboff
Modified: 2011-05-16 10:48 PDT (History)
1 user (show)

See Also:

Attachments
Patch to reuse backing stores and reduce timeout to 2 seconds (2.17 KB, patch)
2011-05-15 22:02 PDT, Michael Saboff 		ggaren: review+
Details | Formatted Diff | Diff
View All Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description Michael Saboff 2011-05-15 17:55:57 PDT 
BackingStore objects can be reused in DrawingAreaProxyImpl::didUpdateBackingStoreState as per a FIXME: comment.  This is likely to be more efficient.
Comment 1 Michael Saboff 2011-05-15 22:02:54 PDT 
Created attachment 93604 [details]
Patch to reuse backing stores and reduce timeout to 2 seconds
Comment 2 Geoffrey Garen 2011-05-15 22:08:28 PDT 
Comment on attachment 93604 [details]
Patch to reuse backing stores and reduce timeout to 2 seconds

View in context: https://bugs.webkit.org/attachment.cgi?id=93604&action=review

r=me

> Source/WebKit2/UIProcess/DrawingAreaProxyImpl.cpp:212
> +    // If we have a backing store the right size, reuse it.

"backing store that is the right size" would make for a more grammatically correct comment.
Comment 3 Michael Saboff 2011-05-15 22:31:36 PDT 
Committed r86524: <http://trac.webkit.org/changeset/86524>
Comment 4 Adam Roben (:aroben) 2011-05-16 06:05:58 PDT 
Was any (non-memory) performance testing done of the 5 second -> 2 second change? See bug 51262 comment 5 and bug 51262 comment 6 (and other comments, too) for how the 5-second interval was arrived at.
Comment 5 Michael Saboff 2011-05-16 10:48:07 PDT 
(In reply to comment #4)

I did some bench testing looking for white flash problems.  This included running with little free memory and a large number of tabs / windows (40-50).  I didn't notice any white flashing or unreasonable delays.  With instrumentation, I noticed that the number of backing stores would grow as needed.