From mboxrd@z Thu Jan 1 00:00:00 1970 From: Andrew Cooper Subject: Re: Xen PV PTE ABI (or lack thereof) Date: Thu, 21 Jan 2016 14:37:19 +0000 Message-ID: <56A0ED1F.90603@citrix.com> References: <569FE999.2080404@citrix.com> <56A0CA0902000078000C9899@prv-mh.provo.novell.com> <56A0BDF2.8030308@citrix.com> <56A0E42702000078000C9970@prv-mh.provo.novell.com> <56A0DA69.7090002@citrix.com> <56A0EB42.1050003@citrix.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <56A0EB42.1050003@citrix.com> List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Sender: xen-devel-bounces@lists.xen.org Errors-To: xen-devel-bounces@lists.xen.org To: David Vrabel , Jan Beulich Cc: George Dunlap , Huaitong Han , Tim Deegan , Xen-devel List List-Id: xen-devel@lists.xenproject.org On 21/01/16 14:29, David Vrabel wrote: > On 21/01/16 13:17, Andrew Cooper wrote: >> As we see with the Protection Key feature, newer hardware feature start >> using bits which were previously software available, and we absolutely >> don't want to be in a position where our ABI prevents us from ever >> supporting a new feature. > I don't see a problem in not supporting new features in /PV/ guests. I know you want to kill PV guests. I am not going to deliberately cripple PV guests in an attempt to achieve that goal. > > I think we should document the PV ABI as-is. i.e., that these two PTE > bits are not available for PV guest use. This is already known to break some guests. It shouldn't stay as it currently is. ~Andrew