From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:41865) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1Z5su6-0001JU-2n for qemu-devel@nongnu.org; Fri, 19 Jun 2015 05:48:07 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1Z5su1-0008PN-Ts for qemu-devel@nongnu.org; Fri, 19 Jun 2015 05:48:06 -0400 Received: from mx1.redhat.com ([209.132.183.28]:54546) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1Z5su1-0008PF-OE for qemu-devel@nongnu.org; Fri, 19 Jun 2015 05:48:01 -0400 Date: Fri, 19 Jun 2015 11:47:57 +0200 From: Radim =?utf-8?B?S3LEjW3DocWZ?= Message-ID: <20150619094756.GB10217@potion.brq.redhat.com> References: <1434641064-8405-1-git-send-email-rkrcmar@redhat.com> <1434641064-8405-3-git-send-email-rkrcmar@redhat.com> <20150618155046.GC3874@thinpad.lan.raisama.net> MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Disposition: inline In-Reply-To: <20150618155046.GC3874@thinpad.lan.raisama.net> Content-Transfer-Encoding: quoted-printable Subject: Re: [Qemu-devel] [PATCH 2/2] target-i386: automatically raise cpuid level to 0xd List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Eduardo Habkost Cc: pbonzini@redhat.com, bsd@redhat.com, qemu-devel@nongnu.org, rth@twiddle.net 2015-06-18 12:50-0300, Eduardo Habkost: > On Thu, Jun 18, 2015 at 05:24:24PM +0200, Radim Kr=C4=8Dm=C3=A1=C5=99 w= rote: > > We already bump to level 7 if features there are requested, so do the > > same for 0xD. > >=20 > > Signed-off-by: Radim Kr=C4=8Dm=C3=A1=C5=99 >=20 > This breaks guest ABI and live-migration, as CPUID data is not part of > the migration stream (although we have considered including it in the > future). >=20 > If we are going to add more special cases like this, we must provide a > way to make QEMU honour an explicit "level" option from the config file > or command-line. Thanks, I'll drop this patch. > I have considered introducing "min-[x]level" and "max-{x]level" > properties to control automatic increasing of level/xlevel. The existin= g > X86CPUDefinition.level field could just control min_level, while > explicit "level=3D" on the command-line or config file would explicitly > force a specific value. Probably setting "max-level" on machine-type > compat code would be enough to restore the previous behavior. We'd need to set min-level at least to 7, to capture the raising we do now, but a feature in level between default and 7 would result in a different behavior, so we need to make it much uglier :/ We can add 'compat-level' bit for old machine types and raise to highest habited function otherwise, optionally with controls you described.