From: Keir Fraser Subject: Re: [RFC Patch] Support for making an E820 PCI hole in toolstack (xl + xm) Date: Tue, 16 Nov 2010 07:40:35 +0000 Message-ID: References: <20101115231133.GA12364@dumpdata.com> Mime-Version: 1.0 Content-Type: text/plain; charset="US-ASCII" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <20101115231133.GA12364@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: Jeremy Fitzhardinge , xen-devel@lists.xensource.com, bruce.edge@gmail.com, gianni.tedesco@citrix.com, stefano.stabellini@eu.citrix.com List-Id: xen-devel@lists.xenproject.org On 15/11/2010 23:11, "Konrad Rzeszutek Wilk" wrote: >> Well, I don't know. It sounds pretty sensible to me. :-) >> >> Certain virtualisation feature sdisappearing after a save/restore/migrate -- >> or worsse, becoming unreliable -- would be a bit sad. > > So having the option of the PCI hole being passed through, and giving > the tools the value (pci_hole) would mean we could migrate an SR-IOV type > device from one machine to another. Constructing the PCI hole using the > XENMEM_machine_memory_map could generate different E820 for the two guests, > which > would be indeed a bit sad. SR-IOV is a nice example, but even assuming the user does not have pass-thru devices attached during migrate, or save/restore, it would still be nice if they could have the pci-hotplug-passthru facility available to them fully working both before and after such an event. -- Keir