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: [XenPPC] Question? Re: [PATCH] Fix paddr_to_maddr panic message

To: Jimi Xenidis <jimix@xxxxxxxxxxxxxx>
Subject: Re: [XenPPC] Question? Re: [PATCH] Fix paddr_to_maddr panic message
From: Hollis Blanchard <hollisb@xxxxxxxxxx>
Date: Thu, 31 Aug 2006 09:59:10 -0500
Cc: xen-ppc-devel@xxxxxxxxxxxxxxxxxxx
Delivery-date: Thu, 31 Aug 2006 07:58:48 -0700
Envelope-to: www-data@xxxxxxxxxxxxxxxxxx
In-reply-to: <FDAF70CC-D180-4F1F-8EE2-4EEEEB054B2A@xxxxxxxxxxxxxx>
List-help: <mailto:xen-ppc-devel-request@lists.xensource.com?subject=help>
List-id: Xen PPC development <xen-ppc-devel.lists.xensource.com>
List-post: <mailto:xen-ppc-devel@lists.xensource.com>
List-subscribe: <http://lists.xensource.com/cgi-bin/mailman/listinfo/xen-ppc-devel>, <mailto:xen-ppc-devel-request@lists.xensource.com?subject=subscribe>
List-unsubscribe: <http://lists.xensource.com/cgi-bin/mailman/listinfo/xen-ppc-devel>, <mailto:xen-ppc-devel-request@lists.xensource.com?subject=unsubscribe>
Organization: IBM Linux Technology Center
References: <20060831001030.GA3138@xxxxxxxxxxxxxxxxxxxxx> <FDAF70CC-D180-4F1F-8EE2-4EEEEB054B2A@xxxxxxxxxxxxxx>
Sender: xen-ppc-devel-bounces@xxxxxxxxxxxxxxxxxxx
On Thu, 2006-08-31 at 08:07 -0400, Jimi Xenidis wrote:
> Hollis, ASAICT the only way Amos could have gotten here (and passed  
> the BUG() in pfn2mfn()) is if one of the user_copy routines tried to  
> access a PFN_TYPE_REMOTE or PFN_TYPE_IO, that paddr is not an IO type  
> so the question I have for you is:
>    Is it possible for Xen (on behalf of a privileged domain) use a  
> user_copy routine to access another domains memory?

No. Because copy_*_guest() require a xencomm_desc structure to do
anything[1], they can only copy into areas of memory the domain expects.

However, if we need that functionality, we can create a parallel API to
do this. x86 does that right now with copy_*_user() -- that API only
exists in x86-specific code, and it directly accesses the "user" virtual

[1] If we go with IA64's proposed "inline" extension to the xencomm
stuff, this isn't strictly true and I think we could copy to physically
contiguous areas that way.

Hollis Blanchard
IBM Linux Technology Center

Xen-ppc-devel mailing list

<Prev in Thread] Current Thread [Next in Thread>