This is an archived copy of the Xen.org mailing list, which we have preserved to ensure that existing links to archives are not broken. The live archive, which contains the latest emails, can be found at http://lists.xen.org/
Home Products Support Community News


Re: [Xen-devel] [PATCH] fix memory exchange hypercall.

To: Isaku Yamahata <yamahata@xxxxxxxxxxxxx>, Keir Fraser <keir@xxxxxxxxxxxxx>
Subject: Re: [Xen-devel] [PATCH] fix memory exchange hypercall.
From: Keir Fraser <keir@xxxxxxxxxxxxx>
Date: Thu, 15 Mar 2007 12:23:20 +0000
Cc: Tim Deegan <Tim.Deegan@xxxxxxxxxxxxx>, xen-devel@xxxxxxxxxxxxxxxxxxx, xen-ia64-devel@xxxxxxxxxxxxxxxxxxx
Delivery-date: Thu, 15 Mar 2007 05:22:38 -0700
Envelope-to: www-data@xxxxxxxxxxxxxxxxxx
In-reply-to: <20070315114201.GC1819%yamahata@xxxxxxxxxxxxx>
List-help: <mailto:xen-devel-request@lists.xensource.com?subject=help>
List-id: Xen developer discussion <xen-devel.lists.xensource.com>
List-post: <mailto:xen-devel@lists.xensource.com>
List-subscribe: <http://lists.xensource.com/cgi-bin/mailman/listinfo/xen-devel>, <mailto:xen-devel-request@lists.xensource.com?subject=subscribe>
List-unsubscribe: <http://lists.xensource.com/cgi-bin/mailman/listinfo/xen-devel>, <mailto:xen-devel-request@lists.xensource.com?subject=unsubscribe>
Sender: xen-devel-bounces@xxxxxxxxxxxxxxxxxxx
Thread-index: Acdm/LeD9cke0tLvEdu/SQAX8io7RQ==
Thread-topic: [Xen-devel] [PATCH] fix memory exchange hypercall.
User-agent: Microsoft-Entourage/
On 15/3/07 11:42, "Isaku Yamahata" <yamahata@xxxxxxxxxxxxx> wrote:

> The big differrence is that the x86 p2m table (i.e. hvm domain)
> doesn't support foreign domain page mapping and grant table page mapping.
> On the other hand, the ia64 p2m supports them.
> (NOTE:The ia64 p2m is used for both para/full virtualized domain.
> Even for dom0. So they must be supported.)

Could you take references for the foreign mappings and grant mappings only?
Presumably you have some kind of metadata associated with each p2m entry to
tell you what kind of thing is mapped at each pfn?

Alternatively we could change x86 to take a reference for anything in its
physmap. It seems a bit unnecessary though. I think when we come to add
other types of things to a physmap (grant references in particular) we'll do
reference counting dependent on the type of thing that is mapped. After all,
we also sometime want to be able to assign real mmio ranges to HVM guests,
and there are no page_info structures associated with such ranges, and so no
reference count to increment!

Whichever, I think we should sort it out one way or the other and change
ia64 and/or x86 now. Or it'll never happen!

 -- Keir

Xen-devel mailing list