From mboxrd@z Thu Jan 1 00:00:00 1970 From: Keir Fraser Subject: Re: 32-on-64: pvfb issue Date: Thu, 18 Jan 2007 18:31:39 +0000 Message-ID: References: <45AF93A7.50107@suse.de> Mime-Version: 1.0 Content-Type: text/plain; charset="US-ASCII" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <45AF93A7.50107@suse.de> List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Sender: xen-devel-bounces@lists.xensource.com Errors-To: xen-devel-bounces@lists.xensource.com To: Gerd Hoffmann , Markus Armbruster Cc: Xen devel list List-Id: xen-devel@lists.xenproject.org On 18/1/07 15:35, "Gerd Hoffmann" wrote: >> Excuse my ignorance, but why do you have to guess the guest's size? >> Doesn't dom0 know? > > No, right now there is no hypercall to get that information. Which > brings the discussion back onto the table: Should we add one? pvfb > isn't the only driver with that kind of problems. > > And it better shouldn't be a dom0-only hypercall, otherwise driver > domains will not be able to use it ... Do the same as you did with blkback. Add a protocol version or bitwidth field to xenstore. Update frontend driver to write that field. Make tools write a default value for pv guests for backward compat. This is even better than with blkfront/blkback because there are definitely no PV-on-HVM pvfb driver frontends out in the wild (and it's HVM that makes things harder). Speaking of which, we should add the protocol/bitwidth field to blkfront asap... -- Keir