WARNING - OLD ARCHIVES

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/
   
 
 
Xen 
 
Home Products Support Community News
 
   
 

xen-devel

Re: [Xen-devel] [PATCH] Allow programatic iomem permissions

To: Keir Fraser <keir@xxxxxxxxxxxxx>
Subject: Re: [Xen-devel] [PATCH] Allow programatic iomem permissions
From: Kieran Mansley <kmansley@xxxxxxxxxxxxxx>
Date: Fri, 13 Jul 2007 14:03:53 +0100
Cc: xen-devel@xxxxxxxxxxxxxxxxxxx
Delivery-date: Fri, 13 Jul 2007 06:01:42 -0700
Envelope-to: www-data@xxxxxxxxxxxxxxxxxx
In-reply-to: <C2BD2ED6.1267D%keir@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>
References: <C2BD2ED6.1267D%keir@xxxxxxxxxxxxx>
Sender: xen-devel-bounces@xxxxxxxxxxxxxxxxxxx
On Fri, 2007-07-13 at 13:29 +0100, Keir Fraser wrote:
> 
> 
> On 13/7/07 12:05, "Kieran Mansley" <kmansley@xxxxxxxxxxxxxx> wrote:
> 
> > Any thoughts on the above?
> > 
> > I'm happy to do some work to rectify this situation but don't want to
> > expend effort on something that won't get accepted.  Any background on
> > the reasons for the restriction (i.e. what the TLB flushing issues are)
> > would also be of help.
> 
> The issue is that the granter is informed that the grant is released before
> stale grantee TLB entries are flushed. If the grantee is multi-vcpu then he
> could theoretically still access a granted page via a stale TLB entry after
> the granter has recycled the page. The window is extremely tiny though! The
> correct fix is to reorder the unmap operation to be unmap-list-of-grants
> then TLB-flush then update-grant-entries-to-indicate-release. Then the whole
> problem disappears.

OK, that makes sense, and doesn't at first impression look too hard to
rectify.

Am I right in thinking that it's the shared grant table entry that is
the critical one in this sense (as opposed to the "active" entry).

Kieran


_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel