From mboxrd@z Thu Jan 1 00:00:00 1970 From: Keir Fraser Subject: Re: [PATCH] X86: cpuid faulting feature enable Date: Fri, 01 Jul 2011 17:07:52 +0100 Message-ID: References: <4E0E088A020000780004BB07@nat28.tlf.novell.com> Mime-Version: 1.0 Content-Type: text/plain; charset="US-ASCII" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <4E0E088A020000780004BB07@nat28.tlf.novell.com> List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Sender: xen-devel-bounces@lists.xensource.com Errors-To: xen-devel-bounces@lists.xensource.com To: Jan Beulich , Jinsong Liu Cc: Kevin Tian , "xen-devel@lists.xensource.com" , Haitao Shan , Xin Li List-Id: xen-devel@lists.xenproject.org On 01/07/2011 16:48, "Jan Beulich" wrote: >> + if ( cpuid_faulting_flip ) >> + { >> + unsigned int enable; >> + >> + enable = !is_hvm_vcpu(next) && (next->domain->domain_id != 0); > > Excluding Dom0 here is perhaps questionable (as it could allow hiding > features not supported by Xen from applications). It's probably because of a need for raw cpuid in libxc/xc_cpuid_x86.c where we need to detect things like long mode. We could do that another way, or provide a new paravirtualised raw cpuid just for dom0. Not a barrier to this patch going in imo. I suppose HVM VCPUs are excluded because otherwise the new feature would be unexpectedly enabled in non-root mode too, and make HVM guests crash? :-) -- Keir