From: "Jan Beulich" Subject: Re: [PATCH] x86: adjust reserved bit page fault handling Date: Fri, 01 Feb 2008 11:24:41 +0000 Message-ID: <47A30F89.76E4.0078.0@novell.com> References: <47A304B5.76E4.0078.0@novell.com> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: quoted-printable Return-path: In-Reply-To: Content-Disposition: inline List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Sender: xen-devel-bounces@lists.xensource.com Errors-To: xen-devel-bounces@lists.xensource.com To: Keir Fraser Cc: xen-devel@lists.xensource.com, Tim Deegan List-Id: xen-devel@lists.xenproject.org >>> Keir Fraser 01.02.08 11:54 >>> >On 1/2/08 10:38, "Jan Beulich" wrote: > >>> The shadow pagetable code deliberately introduces invalid pagetable >>> entries as part of its fast-path treatment of MMIO and not-present >>> entries in the guest tables, so paging_fault needs to be called for >>> PFEC_reserved_bit faults. >>=20 >> Here's a replacement patch then, it would still be questionable whether >> reserved bit faults should ever be propagated to a guest... >>=20 >> Signed-off-by: Jan Beulich > >What problem does this patch solve? When debugging the xen_change_pte_range() problem I mistakenly interpreted a page fault error code 15 as being decimal (it really was hex) and looked at the reserved bit fault handling. The patch is the outcome of this; despite it turned out to not be a fault of this kind I thought that tightening the handling (and adding the printing, since reserved bit faults are [except in the case mentioned by Tim] always bad) wouldn't be a bad idea. Jan