From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:40877) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1eYfuf-0000Jm-Rs for qemu-devel@nongnu.org; Mon, 08 Jan 2018 17:29:02 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1eYfua-0003aE-Sy for qemu-devel@nongnu.org; Mon, 08 Jan 2018 17:29:01 -0500 Received: from mx1.redhat.com ([209.132.183.28]:57058) by eggs.gnu.org with esmtps (TLS1.0:DHE_RSA_AES_256_CBC_SHA1:32) (Exim 4.71) (envelope-from ) id 1eYfua-0003Zj-KO for qemu-devel@nongnu.org; Mon, 08 Jan 2018 17:28:56 -0500 Date: Mon, 8 Jan 2018 20:28:53 -0200 From: Eduardo Habkost Message-ID: <20180108222853.GL6646@localhost.localdomain> References: <20180108205052.24385-1-vincent@bernat.im> <20180108211623.GJ6646@localhost.localdomain> <20180108221422.GK6646@localhost.localdomain> MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Disposition: inline In-Reply-To: Content-Transfer-Encoding: quoted-printable Subject: Re: [Qemu-devel] [PATCH] target-i386: add pcid to both Sandy Bridge and Ivy Bridge List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Vincent Bernat Cc: Paolo Bonzini , Richard Henderson , qemu-devel@nongnu.org On Mon, Jan 08, 2018 at 11:22:36PM +0100, Vincent Bernat wrote: > =E2=9D=A6 8 janvier 2018 20:14 -0200, Eduardo Habkost =C2=A0: >=20 > >> What are the consequences of running a too old kernel? Would KVM jus= t > >> hide PCID flag? > > > > On an old kernel, the SandyBridge and IvyBridge CPU models will > > be unexpectedly become not runnable. >=20 > But, isn't it the same for more recent models that already have PCID > enabled? Yes, the more recent models are already not runnable on those hosts. The key here is "unexpectedly": management software can assume that the CPU model won't become unrunnable when it was runnable in the past, and logic that decides if/where a VM can be started (or migrated to) might break. --=20 Eduardo