From: "Zhang, Yanmin" <yanmin_zhang@linux.intel.com>
To: Christoph Lameter <clameter@sgi.com>
Cc: Andi Kleen <andi@firstfloor.org>,
Pekka Enberg <penberg@cs.helsinki.fi>,
KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>,
Rik van Riel <riel@redhat.com>,
akpm@linux-foundation.org, linux-kernel@vger.kernel.org,
linux-fsdevel@vger.kernel.org, Mel Gorman <mel@skynet.ie>,
mpm@selenic.com, Matthew Wilcox <matthew@wil.cx>
Subject: Re: [patch 21/21] slab defrag: Obsolete SLAB
Date: Thu, 15 May 2008 11:26:42 +0800 [thread overview]
Message-ID: <1210822002.3177.121.camel@ymzhang> (raw)
In-Reply-To: <Pine.LNX.4.64.0805141100110.15633@schroedinger.engr.sgi.com>
On Wed, 2008-05-14 at 11:03 -0700, Christoph Lameter wrote:
> On Wed, 14 May 2008, Andi Kleen wrote:
>
> > iirc profiling analysis showed that the problem was the page lock
> > serialization (in particular the slab_lock() in __slab_free). That
> > was on 2.6.24.2
>
> Do you have an URL?
>
> > I think the problem is that this atomic operation thrashes cache lines
> > around. Really counting cycles on instructions is not that interesting,
> > but minimizing the cache thrashing is. And for that it looks like slub
> > is worse.
>
> It can thrash cachelines if objects from the same slab page are freed
> simultaneously on multiple processors. That occurred in the hackbench
> regression that we addressed with the dynamic configuration of slab sizes.
hackbench regression is because of slow allocation instead of slow freeing.
With dynamic configuration of slab sizes, fast allocation becomes 97% (the bad
one is 68%), but fast free is always 8~9% with/without the patch.
--
To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
next prev parent reply other threads:[~2008-05-15 3:29 UTC|newest]
Thread overview: 93+ messages / expand[flat|nested] mbox.gz Atom feed top
2008-05-10 3:08 [patch 00/21] Slab Fragmentation Reduction V12 Christoph Lameter
2008-05-10 3:08 ` [patch 01/21] slub: Add defrag_ratio field and sysfs support Christoph Lameter
2008-05-10 3:08 ` [patch 02/21] slub: Replace ctor field with ops field in /sys/slab/* Christoph Lameter
2008-05-10 3:08 ` [patch 03/21] slub: Add get() and kick() methods Christoph Lameter
2008-05-10 3:08 ` [patch 04/21] slub: Sort slab cache list and establish maximum objects for defrag slabs Christoph Lameter
2008-05-10 3:08 ` [patch 05/21] slub: Slab defrag core Christoph Lameter
2008-05-10 3:08 ` [patch 06/21] slub: Add KICKABLE to avoid repeated kick() attempts Christoph Lameter
2008-05-10 3:08 ` [patch 07/21] slub: Extend slabinfo to support -D and -F options Christoph Lameter
2008-05-10 3:08 ` [patch 08/21] slub: add defrag statistics Christoph Lameter
2008-05-10 3:08 ` [patch 09/21] slub: Trigger defragmentation from memory reclaim Christoph Lameter
2008-05-10 3:08 ` [patch 10/21] buffer heads: Support slab defrag Christoph Lameter
2008-05-12 0:24 ` David Chinner
2008-05-15 17:42 ` Christoph Lameter
2008-05-15 23:10 ` David Chinner
2008-05-16 17:01 ` Christoph Lameter
2008-05-19 5:45 ` David Chinner
2008-05-19 16:44 ` Christoph Lameter
2008-05-20 0:25 ` David Chinner
2008-05-20 6:56 ` Evgeniy Polyakov
2008-05-20 21:46 ` David Chinner
2008-05-20 22:25 ` Evgeniy Polyakov
2008-05-20 23:19 ` David Chinner
2008-05-20 23:28 ` Andrew Morton
2008-05-21 6:15 ` Evgeniy Polyakov
2008-05-21 6:24 ` Andrew Morton
2008-05-21 17:52 ` iput() in reclaim context Hugh Dickins
2008-05-21 17:58 ` Evgeniy Polyakov
2008-05-21 18:12 ` Andrew Morton
2008-05-20 23:22 ` [patch 10/21] buffer heads: Support slab defrag Evgeniy Polyakov
2008-05-20 23:30 ` David Chinner
2008-05-21 6:20 ` Evgeniy Polyakov
2008-05-21 1:56 ` Christoph Lameter
2008-05-20 22:53 ` Jamie Lokier
2008-05-10 3:08 ` [patch 11/21] inodes: Support generic defragmentation Christoph Lameter
2008-05-10 3:08 ` [patch 12/21] Filesystem: Ext2 filesystem defrag Christoph Lameter
2008-05-10 3:08 ` [patch 13/21] Filesystem: Ext3 " Christoph Lameter
2008-05-10 3:08 ` [patch 14/21] Filesystem: Ext4 " Christoph Lameter
2008-05-10 3:08 ` [patch 15/21] Filesystem: XFS slab defragmentation Christoph Lameter
2008-05-10 6:55 ` Christoph Hellwig
2008-05-10 3:08 ` [patch 16/21] Filesystem: /proc filesystem support for slab defrag Christoph Lameter
2008-05-10 3:08 ` [patch 17/21] Filesystem: Slab defrag: Reiserfs support Christoph Lameter
2008-05-10 3:08 ` [patch 18/21] Filesystem: Socket inode defragmentation Christoph Lameter
2008-05-13 13:28 ` Evgeniy Polyakov
2008-05-15 17:40 ` Christoph Lameter
2008-05-15 18:23 ` Evgeniy Polyakov
2008-05-10 3:08 ` [patch 19/21] dentries: Add constructor Christoph Lameter
2008-05-10 3:08 ` [patch 20/21] dentries: dentry defragmentation Christoph Lameter
2008-05-10 3:08 ` [patch 21/21] slab defrag: Obsolete SLAB Christoph Lameter
2008-05-10 9:53 ` Andi Kleen
2008-05-11 2:15 ` Rik van Riel
2008-05-12 7:38 ` KOSAKI Motohiro
2008-05-12 7:54 ` Pekka Enberg
2008-05-12 10:08 ` Andi Kleen
2008-05-12 10:23 ` Pekka Enberg
2008-05-14 17:30 ` Christoph Lameter
2008-05-14 17:29 ` Christoph Lameter
2008-05-14 17:49 ` Andi Kleen
2008-05-14 18:03 ` Christoph Lameter
2008-05-14 18:18 ` Matt Mackall
2008-05-14 19:21 ` Christoph Lameter
2008-05-14 19:49 ` Matt Mackall
2008-05-14 20:33 ` Christoph Lameter
2008-05-14 21:02 ` Matt Mackall
2008-05-14 21:26 ` Christoph Lameter
2008-05-14 21:54 ` Matt Mackall
2008-05-15 17:15 ` Christoph Lameter
2008-05-15 3:26 ` Zhang, Yanmin [this message]
2008-05-15 17:05 ` Christoph Lameter
2008-05-15 17:49 ` Matthew Wilcox
2008-05-15 17:58 ` Christoph Lameter
2008-05-15 18:13 ` Matthew Wilcox
2008-05-15 18:43 ` Christoph Lameter
2008-05-15 18:51 ` Matthew Wilcox
2008-05-15 19:09 ` Christoph Lameter
2008-05-15 19:29 ` Matthew Wilcox
2008-05-15 20:14 ` Matthew Wilcox
2008-05-15 20:30 ` Pekka Enberg
2008-05-16 19:17 ` Christoph Lameter
2008-05-16 19:06 ` Christoph Lameter
2008-05-15 18:19 ` Eric Dumazet
2008-05-15 18:29 ` Vegard Nossum
2008-05-16 5:16 ` Zhang, Yanmin
2008-05-14 18:05 ` Christoph Lameter
2008-05-14 20:46 ` Christoph Lameter
2008-05-14 20:58 ` Matthew Wilcox
2008-05-14 21:00 ` Christoph Lameter
2008-05-14 21:21 ` Matthew Wilcox
2008-05-14 21:33 ` Christoph Lameter
2008-05-14 21:43 ` Matthew Wilcox
2008-05-14 21:53 ` Christoph Lameter
2008-05-14 22:00 ` Matthew Wilcox
2008-05-14 22:32 ` Christoph Lameter
2008-05-14 22:34 ` Christoph Lameter
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=1210822002.3177.121.camel@ymzhang \
--to=yanmin_zhang@linux.intel.com \
--cc=akpm@linux-foundation.org \
--cc=andi@firstfloor.org \
--cc=clameter@sgi.com \
--cc=kosaki.motohiro@jp.fujitsu.com \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=matthew@wil.cx \
--cc=mel@skynet.ie \
--cc=mpm@selenic.com \
--cc=penberg@cs.helsinki.fi \
--cc=riel@redhat.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).