lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite for Android: free password hash cracker in your pocket
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20131204130038.GY11295@suse.de>
Date:	Wed, 4 Dec 2013 13:00:38 +0000
From:	Mel Gorman <mgorman@...e.de>
To:	Florian Fainelli <f.fainelli@...il.com>
Cc:	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
	mhocko@...e.cz, hannes@...xchg.org, riel@...hat.com,
	linux-mm@...ck.org, m.szyprowski@...sung.com, marc.ceeeee@...il.com
Subject: Re: high kswapd CPU usage when executing binaries from NFS w/ CMA
 and COMPACTION

On Tue, Dec 03, 2013 at 06:30:28PM -0800, Florian Fainelli wrote:
> Hi all,
> 
> I am experiencing high kswapd CPU usage on an ARMv7 system running
> 3.8.13 when executing relatively large binaries from NFS. When this
> happens kswapd consumes around 55-60% CPU usage and the applications
> takes a huge time to load.
> 

There were a number of changes made related to how and when kswapd
stalls, particularly when pages are dirty. Brief check confirms that

git log v3.8..v3.12 --pretty=one --author "Mel Gorman" mm/vmscan.c

NFS dirty pages are problematic for compaction as dirty pages cannot be
migrated until cleaned. I'd suggest checking if current mainline suffers
the same problem and if not, focus on patches related to dirty page
handling and kswapd throttling in mm/vmscan.c as backport candidates.

-- 
Mel Gorman
SUSE Labs
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ