From mboxrd@z Thu Jan 1 00:00:00 1970 From: "Michael S. Tsirkin" Subject: Re: [Qemu-devel] [PATCHv4 3/4] cpuid: disable pv eoi for 1.1 and older compat types Date: Wed, 29 Aug 2012 13:31:13 +0300 Message-ID: <20120829103113.GA5825@redhat.com> References: <20120828191338.GE6223@otherpad.lan.raisama.net> <20120828213527.GB5817@redhat.com> <20120828220242.GW2886@otherpad.lan.raisama.net> <20120828222113.GA5970@redhat.com> <20120829095903.GB32031@amt.cnet> <20120829102324.GC5225@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: Eduardo Habkost , Anthony Liguori , avi@redhat.com, gleb@redhat.com, qemu-devel@nongnu.org, kvm@vger.kernel.org, Jan Kiszka , Blue Swirl To: Marcelo Tosatti Return-path: Received: from mx1.redhat.com ([209.132.183.28]:24069 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752751Ab2H2KaA (ORCPT ); Wed, 29 Aug 2012 06:30:00 -0400 Content-Disposition: inline In-Reply-To: <20120829102324.GC5225@redhat.com> Sender: kvm-owner@vger.kernel.org List-ID: On Wed, Aug 29, 2012 at 01:23:24PM +0300, Michael S. Tsirkin wrote: > > Michael, not migrating the PVEOI MSR actually crashes the guest? > > > Not at all. migration currently simply disables PV EOI, so > each EOI triggers exits after migration. I would like to add: without this patch guest will crash. But what will crash guest is *not* the fact we do not migrate the PV EOI MSR. It is the fact that PV EOI bit is set in CPUID. If guest reads CPUID before migration and then writes the MSR after migration, it will crash. > -- > MST