Bug 115631

Summary:

Removing m_maxDeadCapacity condition in fast path in MemoryCache::prune().

Product:

WebKit

Reporter:

Marcin Bychawski <m.bychawski>

Component:

WebCore Misc.

Assignee:

Nobody <webkit-unassigned>

Status:

RESOLVED FIXED

Severity:

Normal

CC:

ap, beidson, commit-queue, g.czajkowski, japhet, koivisto, m.leszko, mnaganov, w.bielawski

Priority:

Version:

528+ (Nightly build)

Hardware:

Unspecified

OS:

Unspecified

Attachments:

Description	Flags
patch	bfulgham: review+, commit-queue: commit-queue-
Patch	none
Patch	none

Marcin Bychawski

Reported 2013-05-06 03:55:34 PDT

If the m_maxDeadCapacity == 0, MemoryCache::prune() always tries to prune Resources. But if m_deadSize is also 0, pruning is unnecessary. I removed one condition in the "fast path". Is it possible that in some case this condition is required for properly work?

Attachments
patch (1.37 KB, patch) 2013-05-06 04:07 PDT, Marcin Bychawski	bfulgham: review+ commit-queue: commit-queue-	Details Formatted Diff Diff
Patch (1.52 KB, patch) 2013-10-31 02:29 PDT, Wojciech Bielawski	no flags	Details Formatted Diff Diff
Patch (1.57 KB, patch) 2013-10-31 06:39 PDT, Wojciech Bielawski	no flags	Details Formatted Diff Diff
Show Obsolete (1) View All Add attachment proposed patch, testcase, etc.

Marcin Bychawski

Comment 1 2013-05-06 04:07:40 PDT

Created attachment 200652 [details] patch

Alexey Proskuryakov

Comment 2 2013-05-06 10:03:41 PDT

This patch is not currently marked for review. Did you intend to submit it for review? You can still do that via Details link to the right of the patch.

Brent Fulgham

Comment 3 2013-05-08 23:23:03 PDT

Comment on attachment 200652 [details] patch This looks fine to me. Is there any case we now fail, such as 'm_maxDeadCapacity == 0' and 'm_deadSize < 0' that might now get through and cause a problem?

Marcin Bychawski

Comment 4 2013-05-10 04:11:54 PDT

(In reply to comment #3) > (From update of attachment 200652 [details]) > This looks fine to me. Is there any case we now fail, such as 'm_maxDeadCapacity == 0' and 'm_deadSize < 0' that might now get through and cause a problem? I don't see any case we now can fail. m_deadSize is unsigned, so it can't be < 0. When m_maxDeadCapacity == 0, we need to prune resources only when m_deadSize > 0, so there condition 'm_deadSize <= m_maxDeadCapacity' is enough.

Brent Fulgham

Comment 5 2013-10-30 10:35:16 PDT

Comment on attachment 200652 [details] patch I forgot about this! Looks good. r=me

WebKit Commit Bot

Comment 6 2013-10-30 10:47:21 PDT

Comment on attachment 200652 [details] patch Rejecting attachment 200652 [details] from commit-queue. Failed to run "['/Volumes/Data/EWS/WebKit/Tools/Scripts/webkit-patch', '--status-host=webkit-queues.appspot.com', '--bot-id=webkit-cq-02', 'validate-changelog', '--check-oops', '--non-interactive', 200652, '--port=mac']" exit_code: 1 cwd: /Volumes/Data/EWS/WebKit ChangeLog entry in Source/WebCore/ChangeLog contains OOPS!. Full output: http://webkit-queues.appspot.com/results/17138231

Wojciech Bielawski

Comment 7 2013-10-31 02:29:01 PDT

Created attachment 215636 [details] Patch

Wojciech Bielawski

Comment 8 2013-10-31 06:39:41 PDT

Created attachment 215649 [details] Patch

WebKit Commit Bot

Comment 9 2013-10-31 07:06:01 PDT

Comment on attachment 215649 [details] Patch Clearing flags on attachment: 215649 Committed r158357: <http://trac.webkit.org/changeset/158357>

Note You need to log in before you can comment on or make changes to this bug.