From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:46388) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1ZVL6G-0002w5-LM for qemu-devel@nongnu.org; Fri, 28 Aug 2015 10:57:53 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1ZVL6D-0006cO-As for qemu-devel@nongnu.org; Fri, 28 Aug 2015 10:57:52 -0400 Received: from mailpro.odiso.net ([89.248.209.98]:38985) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1ZVL6D-0006c6-4X for qemu-devel@nongnu.org; Fri, 28 Aug 2015 10:57:49 -0400 Date: Fri, 28 Aug 2015 16:57:46 +0200 (CEST) From: Alexandre DERUMIER Message-ID: <1978979044.5160386.1440773866768.JavaMail.zimbra@oxygem.tv> In-Reply-To: <557A00AA.2060800@msgid.tls.msk.ru> References: <20150603121201.GB28406@marvin.ws.skroutz.gr> <20150611204616.GB19602@marvin.ws.skroutz.gr> <557A00AA.2060800@msgid.tls.msk.ru> MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: quoted-printable Subject: Re: [Qemu-devel] Steal time MSR not set properly during live migration? List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Michael Tokarev Cc: debian-admin@lists.debian.org, Marcelo Tosatti , qemu-devel , Gleb Natapov Hi, I have add this bug today on 3 debian jessie guests (kernel 3.16), after mi= gration from qemu 2.3 to qemu 2.4. Is it a qemu bug or guest kernel 3.16 ? Regards, Alexandre Derumier ----- Mail original ----- De: "Michael Tokarev" =C3=80: "qemu-devel" , debian-admin@lists.debian.org Cc: "Marcelo Tosatti" , "Gleb Natapov" Envoy=C3=A9: Jeudi 11 Juin 2015 23:42:02 Objet: Re: [Qemu-devel] Steal time MSR not set properly during live=09migra= tion? 11.06.2015 23:46, Apollon Oikonomopoulos wrote:=20 > On 15:12 Wed 03 Jun , Apollon Oikonomopoulos wrote:=20 >> Any ideas?=20 >=20 > As far as I understand, there is an issue when reading the MSR on the=20 > incoming side: there is a KVM_SET_MSRS vcpu ioctl issued by the main=20 > thread during initialization, that causes the initial vCPU steal time=20 > value to be set using the main thread's (and not the vCPU thread's)=20 > run_delay. Then, upon resuming execution, kvm_arch_load_vcpu uses the=20 > vCPU thread's run_delay to determine steal time, causing an overflow.=20 > The issue was introduced by commit=20 > 917367aa968fd4fef29d340e0c7ec8c608dffaab.=20 >=20 > For the full analysis, see https://bugs.debian.org/785557#64 and the=20 > followup e-mail.=20 Adding Cc's...=20 Thanks,=20 /mjt=20