From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:39160) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1UlZxL-0000hZ-1R for qemu-devel@nongnu.org; Sun, 09 Jun 2013 03:22:29 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1UlZxI-0005Rf-Fg for qemu-devel@nongnu.org; Sun, 09 Jun 2013 03:22:26 -0400 References: <51A7036A.3050407@ozlabs.ru> <51A7049F.6040207@redhat.com> <51A70B3D.90609@ozlabs.ru> <51A71705.6060009@kamp.de> <51A74D79.7040204@redhat.com> <2765FDFA-8050-4AA3-8621-7E9EA2C89F9C@kamp.de> <51A764FC.7080705@redhat.com> <51ADF122.70307@kamp.de> <51ADF637.7060804@redhat.com> <51ADFBCE.3080200@kamp.de> <51ADFC7A.7030009@redhat.com> <51ADFE4C.6050504@kamp.de> <51AE04ED.9030108@redhat.com> <51AEB267.6020504@ozlabs.ru> <49FB93E9-FBFD-4C68-9A65-76AE025B2E20@kamp.de> Mime-Version: 1.0 (1.0) In-Reply-To: Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: quoted-printable Message-Id: <0E4549E3-E42B-40AA-B87B-4BCFD193E982@kamp.de> From: Peter Lieven Date: Sun, 9 Jun 2013 09:22:16 +0200 Subject: Re: [Qemu-devel] broken incoming migration List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: liu ping fan Cc: Alexey Kardashevskiy , Paolo Bonzini , "qemu-ppc@nongnu.org" , "qemu-devel@nongnu.org" , David Gibson Am 09.06.2013 um 06:12 schrieb liu ping fan : > Hi Peter, >=20 > Is it that sending zero page mostly service the first iteration, ie > bluk-stage? And for the subsequent iteration, dirty pages are normally > not zero. >=20 Yes most Zero Pages are sent during bulk stage except for busy windows guest= s or linux with page sanitization. In these cases freed Memory is zeroed. Peter > Thanks >=20 > On Wed, Jun 5, 2013 at 2:09 PM, Peter Lieven wrote: >>=20 >> Am 05.06.2013 um 05:37 schrieb Alexey Kardashevskiy : >>=20 >>> On 06/05/2013 05:15 AM, Peter Lieven wrote: >>>>=20 >>>> Am 04.06.2013 um 17:17 schrieb Paolo Bonzini : >>>>=20 >>>>> Il 04/06/2013 16:48, Peter Lieven ha scritto: >>>>>> Still using 3.2, but strange enough the above example is also not >>>>>> increasing RSS size for me. >>>>>>=20 >>>>>> Can you try the following: >>>>>> qemu git master with 1G of memory (hanging in bios with no boot devic= e) >>>>>> and migrate it. Before migration RSS Size os somewhat >>>>>> around 16MB. After migration its RSS size is in the order of 1G. >>>>>=20 >>>>> That may be a kernel bug. The kernel did not do the copy-on-write tri= ck >>>>> on huge zero pages. It was fixed last year, maybe 3.2 is not enough. >>>>> Try adding a MADV_HUGEPAGE madvise to the testcase and see if it repro= duces. >>>>=20 >>>> that's it. thanks for the pointer. the huge zero page was introduced in= 3.8. >>>>=20 >>>> paolo, alexey: can you please verify the following works for you: >>>> https://github.com/plieven/qemu/tree/fix-migration >>>=20 >>> These two? >>> 848b796 Tue Jun 4 14:43:04 2013 +0200 Peter Lieven migration: do not >>> overwrite zero pages >>> 2206ac8 Tue Jun 4 14:25:33 2013 +0200 Peter Lieven Revert "migration: do= >>> not sent zero pages in bulk stage" >>=20 >> Yes, sorry forgot to mention this. >>=20 >>>=20 >>> That works for me (qemu 1.5, kernel 3.9-rc2). >>> Tested-by: Alexey Kardashevskiy >>=20 >> Thank you, >> Peter