From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:53601) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1VEJV9-0001Gz-MC for qemu-devel@nongnu.org; Tue, 27 Aug 2013 09:40:13 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1VEJV3-0002U0-2T for qemu-devel@nongnu.org; Tue, 27 Aug 2013 09:40:07 -0400 Received: from mx1.redhat.com ([209.132.183.28]:23654) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1VEJV2-0002Tu-QY for qemu-devel@nongnu.org; Tue, 27 Aug 2013 09:40:00 -0400 Date: Tue, 27 Aug 2013 16:39:54 +0300 From: Gleb Natapov Message-ID: <20130827133953.GA29736@redhat.com> References: <20130813181713.GA5204@irqsave.net> <20130826124031.GB22977@redhat.com> <20130826124941.GA5418@irqsave.net> <20130826143320.GA22899@redhat.com> <20130826164526.GB5418@irqsave.net> <20130826204729.GW2631@otherpad.lan.raisama.net> <20130827131816.GA5436@irqsave.net> <20130827131756.GA29072@redhat.com> <20130827132008.GB29072@redhat.com> <20130827133533.GB5436@irqsave.net> MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Disposition: inline Content-Transfer-Encoding: quoted-printable In-Reply-To: <20130827133533.GB5436@irqsave.net> Subject: Re: [Qemu-devel] KVM guest cpu L3 cache and cpufreq List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: =?utf-8?Q?Beno=C3=AEt?= Canet Cc: qemu-devel@nongnu.org, pbonzini@redhat.com, Eduardo Habkost , Andreas =?utf-8?Q?F=C3=A4rber?= , Igor Mammedov On Tue, Aug 27, 2013 at 03:35:33PM +0200, Beno=C3=AEt Canet wrote: > > > > If I understand correctly changing the CPUID L3 cache infos in QEMU= will change > > > > the value displayed in the guest /proc/cpuinfo but will not change = the size of > > > > the l3 cache used by the hardware. So I am chasing a cosmetic bug. > > > > If it right ? > > > >=20 > > > Right. > > >=20 > > Well, actually not entirely. An application can query cache sizes and > > adjust its algorithms accordingly in which case performance can be > > suboptimal if information in a guest is not accurate. >=20 > So maybe what is missing is a CPUID passthrough mode triggered when -cpu = host > is used ? >=20 That is missing too, but it should not be default with other CPU models. -- Gleb.