From: Chris Mason <chris.mason@oracle.com>
To: Mel Gorman <mel@csn.ul.ie>
Cc: Frans Pop <elendil@planet.nl>,
David Rientjes <rientjes@google.com>,
KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>,
"Rafael J. Wysocki" <rjw@sisk.pl>,
Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
Kernel Testers List <kernel-testers@vger.kernel.org>,
Pekka Enberg <penberg@cs.helsinki.fi>,
Reinette Chatre <reinette.chatre@intel.com>,
Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>,
Karol Lewandowski <karol.k.lewandowski@gmail.com>,
Mohamed Abbas <mohamed.abbas@intel.com>,
Jens Axboe <jens.axboe@oracle.com>,
"John W. Linville" <linville@tuxdriver.com>,
linux-mm@kvack.org
Subject: Re: [Bug #14141] order 2 page allocation failures in iwlagn
Date: Tue, 27 Oct 2009 12:03:32 -0400 [thread overview]
Message-ID: <20091027160332.GA7776@think> (raw)
In-Reply-To: <20091027155223.GL8900@csn.ul.ie>
On Tue, Oct 27, 2009 at 03:52:24PM +0000, Mel Gorman wrote:
>
> > So, after the move to async/sync, a lot more pages are getting queued
> > for writeback - more than three times the number of pages are queued for
> > writeback with the vanilla kernel. This amount of congestion might be why
> > direct reclaimers and kswapd's timings have changed so much.
> >
>
> Or more accurately, the vanilla kernel has queued up a lot more pages for
> IO than when the patch is reverted. I'm not seeing yet why this is.
[ sympathies over confusion about congestion...lots of variables here ]
If wb_kupdate has been able to queue more writes it is because the
congestion logic isn't stopping it. We have congestion_wait(), but
before calling that in the writeback paths it says: are you congested?
and then backs off if the answer is yes.
Ideally, direct reclaim will never do writeback. We want it to be able
to find clean pages that kupdate and friends have already processed.
Waiting for congestion is a funny thing, it only tells us the device has
managed to finish some IO or that a timeout has passed. Neither event has
any relation to figuring out if the IO for reclaimable pages has
finished.
One option is to have the VM remember the hashed waitqueue for one of
the pages it direct reclaims and then wait on it.
-chris
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
next prev parent reply other threads:[~2009-10-27 16:04 UTC|newest]
Thread overview: 88+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <3onW63eFtRF.A.xXH.oMTxKB@chimera>
[not found] ` <COE24pZSBH.A.k2B.ZNTxKB@chimera>
[not found] ` <200910021111.55749.elendil@planet.nl>
2009-10-05 5:13 ` [Bug #14141] order 2 page allocation failures in iwlagn Frans Pop
2009-10-05 6:50 ` Frans Pop
2009-10-05 8:54 ` Frans Pop
2009-10-05 8:57 ` Mel Gorman
2009-10-05 21:34 ` Frans Pop
2009-10-06 0:04 ` David Rientjes
2009-10-06 1:25 ` KOSAKI Motohiro
2009-10-06 8:53 ` Mel Gorman
2009-10-06 9:14 ` David Rientjes
2009-10-06 9:22 ` Mel Gorman
2009-10-06 10:23 ` Frans Pop
2009-10-11 23:10 ` Frans Pop
2009-10-11 23:36 ` Frans Pop
2009-10-12 13:43 ` Mel Gorman
2009-10-12 17:32 ` Frans Pop
2009-10-12 18:43 ` Mel Gorman
2009-10-13 20:38 ` Frans Pop
2009-10-14 10:30 ` Mel Gorman
2009-10-14 13:10 ` Frans Pop
2009-10-14 15:40 ` Mel Gorman
2009-10-14 16:13 ` Frans Pop
2009-10-14 18:34 ` Frans Pop
2009-10-14 23:56 ` Mel Gorman
2009-10-15 20:15 ` Frans Pop
2009-10-16 9:39 ` Mel Gorman
2009-10-14 16:30 ` reinette chatre
2009-10-18 23:33 ` Frans Pop
2009-10-19 0:36 ` Pekka Enberg
2009-10-19 2:44 ` Frans Pop
2009-10-19 9:49 ` [Bug #14141] order 2 page allocation failures (generic) Tobi Oetiker
2009-10-19 9:54 ` Pekka Enberg
2009-10-19 14:01 ` Karol Lewandowski
2009-10-19 14:06 ` Mel Gorman
2009-10-19 17:09 ` Karol Lewandowski
2009-10-20 1:47 ` Karol Lewandowski
2009-10-19 13:31 ` Mel Gorman
2009-10-19 13:40 ` Tobias Oetiker
2009-10-19 14:09 ` Mel Gorman
2009-10-19 14:16 ` Tobias Oetiker
2009-10-19 14:59 ` Mel Gorman
2009-10-19 20:12 ` Tobias Oetiker
2009-10-19 20:17 ` Tobias Oetiker
2009-10-20 10:57 ` Mel Gorman
2009-10-20 11:44 ` Tobias Oetiker
2009-10-20 12:51 ` Mel Gorman
2009-10-20 12:58 ` Tobias Oetiker
2009-10-20 13:39 ` Mel Gorman
2009-10-20 13:50 ` Tobias Oetiker
2009-10-20 14:14 ` Mel Gorman
2009-10-20 14:20 ` Tobias Oetiker
2009-10-22 10:27 ` Tobias Oetiker
2009-10-19 2:52 ` [Bug #14141] order 2 page allocation failures in iwlagn Jens Axboe
2009-10-19 14:01 ` Mel Gorman
2009-10-19 16:18 ` Chris Mason
2009-10-19 17:01 ` Christoph Hellwig
2009-10-19 21:57 ` Chris Mason
2009-10-19 17:01 ` Christoph Hellwig
2009-10-20 10:48 ` Mel Gorman
2009-10-26 21:06 ` Frans Pop
2009-10-27 14:54 ` Mel Gorman
2009-10-27 15:16 ` KOSAKI Motohiro
2009-10-27 15:21 ` Mel Gorman
2009-10-27 15:52 ` Mel Gorman
2009-10-27 16:03 ` Chris Mason [this message]
2009-10-27 17:21 ` Frans Pop
2009-10-27 17:21 ` Frans Pop
2009-11-05 20:14 ` Frans Pop
2009-11-06 9:51 ` Frans Pop
2009-11-09 19:00 ` Mel Gorman
2009-10-20 10:48 ` Mel Gorman
2009-10-25 18:54 ` Frans Pop
2009-10-14 16:28 ` reinette chatre
2009-10-14 16:50 ` Mel Gorman
2009-10-14 20:41 ` reinette chatre
2009-10-14 21:33 ` Frans Pop
2009-10-14 21:55 ` reinette chatre
2009-10-15 2:02 ` Frans Pop
2009-10-15 15:29 ` reinette chatre
2009-10-15 19:41 ` Frans Pop
2009-10-16 17:21 ` reinette chatre
2009-10-17 5:42 ` reinette chatre
2009-10-27 11:10 ` Frans Pop
2009-10-27 16:15 ` reinette chatre
[not found] ` <COE24pZSBH.A.rP.2MTxKB@chimera>
2009-10-21 20:04 ` [PATCH] SLUB: Don't drop __GFP_NOFAIL completely from allocate_slab() (was: Re: [Bug #14265] ifconfig: page allocation failure. order:5,ode:0x8020 w/ e100) Karol Lewandowski
2009-10-21 21:06 ` David Rientjes
2009-10-21 21:20 ` Karol Lewandowski
2009-10-22 10:20 ` Mel Gorman
2009-10-22 21:33 ` Karol Lewandowski
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=20091027160332.GA7776@think \
--to=chris.mason@oracle.com \
--cc=bzolnier@gmail.com \
--cc=elendil@planet.nl \
--cc=jens.axboe@oracle.com \
--cc=karol.k.lewandowski@gmail.com \
--cc=kernel-testers@vger.kernel.org \
--cc=kosaki.motohiro@jp.fujitsu.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=linville@tuxdriver.com \
--cc=mel@csn.ul.ie \
--cc=mohamed.abbas@intel.com \
--cc=penberg@cs.helsinki.fi \
--cc=reinette.chatre@intel.com \
--cc=rientjes@google.com \
--cc=rjw@sisk.pl \
/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).