From mboxrd@z Thu Jan 1 00:00:00 1970 From: Christoph Egger Subject: Re: xen crash during Dom0 boot Date: Thu, 14 Apr 2011 16:44:20 +0200 Message-ID: <4DA70844.9060106@amd.com> References: <201104141318.31679.Christoph.Egger@amd.com> <20110414133326.GI5548@dumpdata.com> Mime-Version: 1.0 Content-Type: text/plain; charset="ISO-8859-1"; format=flowed Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <20110414133326.GI5548@dumpdata.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: Konrad Rzeszutek Wilk Cc: Tim, "xen-devel@lists.xensource.com" , Deegan List-Id: xen-devel@lists.xenproject.org On 04/14/11 15:33, Konrad Rzeszutek Wilk wrote: >> >> >> The reason is that guest_walk_tables() in xen/arch/x86/mm/guest_walk.c fakes >> l1e entries for 2mb pages but not for 1gb pages. >> A workaround is to assign the Dom0 not more than 2 GB RAM via >> the dom0_mem xen boot parameter. > > But if you use 4KB pages then the workaround is not neccessary, right? Right. Same counts for 2MB pages. >> This can only be reproduced with Dom0 kernels which use the autotranslation >> feature (Linux Dom0 does, NetBSD Dom0 doesn't, don't know about other Dom0). > > Autotranslation feature.. you mean that pagetables are not updated via > the Linux kernel, but instead by the Xen hypervisor? With the autotranslation feature the pv guest does not need to be aware about the fact that the guest physical address space is not the host physical address. Xen maintains a shadow page table for it. W/o the autotranslation feature the pv guest is aware of it and handles the translation itself. Christoph -- ---to satisfy European Law for business letters: Advanced Micro Devices GmbH Einsteinring 24, 85689 Dornach b. Muenchen Geschaeftsfuehrer: Alberto Bozzo, Andrew Bowd Sitz: Dornach, Gemeinde Aschheim, Landkreis Muenchen Registergericht Muenchen, HRB Nr. 43632