From mboxrd@z Thu Jan 1 00:00:00 1970 From: Gilles PIETRI Subject: Re: KSM not working on 2.6.28.7 and qemu-kvm 0.12.2 ? Date: Tue, 23 Feb 2010 19:21:55 +0100 Message-ID: <4B841CC3.7070707@gilouweb.com> References: <4B81A70A.1010407@gilouweb.com> <20100222054031.GD30897@develbox.linuxbox.cz> <4B82F0E4.5030609@gilouweb.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: kvm@vger.kernel.org To: Nikola Ciprich Return-path: Received: from dingo.eclaireurs.org ([91.121.140.153]:36424 "EHLO dingo.eclaireurs.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751867Ab0BWSWb (ORCPT ); Tue, 23 Feb 2010 13:22:31 -0500 In-Reply-To: <4B82F0E4.5030609@gilouweb.com> Sender: kvm-owner@vger.kernel.org List-ID: Le 22/02/2010 22:02, Gilles PIETRI a =E9crit : > Le 22/02/2010 06:40, Nikola Ciprich a =E9crit : >> Hi Gilles, maybe You have the same problem I experienced? >> see http://www.mail-archive.com/kvm@vger.kernel.org/msg28572.html >> nik >> >=20 > Hi, >=20 > The linux headers do include the MADV_MERGEABLE settings, and the lib= c=20 > is the debian one which happened to work fine earlier in this regard.= =20 > Anyway I still added the flags just to make sure, and nothing more ha= ppens. >=20 > Cheers, >=20 > Gilou OK, I had a doubt, so I did it again, and made sure there wasn't a=20 chance for the MADV_* to be missing... And it was the issue actually..=20 Not sure what happened there, given the linux headers are here.. It=20 works fine finally, thanks Nikola for your pointer. Regards, Gilou >=20 >> On Sun, Feb 21, 2010 at 10:35:06PM +0100, Gilles PIETRI wrote: >>> Hi, >>> >>> I'm running a 2.6.32.7 kernel with KSM enabled (CONFIG_KSM=3Dy), wi= th >>> qemu-kvm 0.12.2 that has support for KSM (exec.c checks for >>> MADV_MERGEABLE).. >>> >>> So I have KSM running, /sys/kernel/mm/ksm/run is at 1, and after a >>> few hours running a dozen of VMs, many of them running the same >>> OS/base image, I see no page being shared, KSM seems inactive. >>> >>> Values are pretty much to their default value, and it seems KSM is >>> simply not working.. I see ksmd in the process list though, but her= e >>> are the sysfs values reported: >>> =3D=3D> /sys/kernel/mm/ksm/full_scans <=3D=3D >>> 0 >>> >>> =3D=3D> /sys/kernel/mm/ksm/max_kernel_pages <=3D=3D >>> 1027257 >>> >>> =3D=3D> /sys/kernel/mm/ksm/pages_shared <=3D=3D >>> 0 >>> >>> =3D=3D> /sys/kernel/mm/ksm/pages_sharing <=3D=3D >>> 0 >>> >>> =3D=3D> /sys/kernel/mm/ksm/pages_to_scan <=3D=3D >>> 100 >>> >>> =3D=3D> /sys/kernel/mm/ksm/pages_unshared <=3D=3D >>> 0 >>> >>> =3D=3D> /sys/kernel/mm/ksm/pages_volatile <=3D=3D >>> 0 >>> >>> =3D=3D> /sys/kernel/mm/ksm/run <=3D=3D >>> 1 >>> >>> =3D=3D> /sys/kernel/mm/ksm/sleep_millisecs <=3D=3D >>> 20 >>> >>> Any clues? >>> >>> Cheers, >>> Gilou >>> --=20 >>> To unsubscribe from this list: send the line "unsubscribe kvm" in >>> the body of a message to majordomo@vger.kernel.org >>> More majordomo info at http://vger.kernel.org/majordomo-info.html >>> >> >=20 >=20