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


[Xen-devel] Re: [PATCH] FP instruction in realmode emulation

To: <nitin.a.kamble@xxxxxxxxx>
Subject: [Xen-devel] Re: [PATCH] FP instruction in realmode emulation
From: Keir Fraser <Keir.Fraser@xxxxxxxxxxxx>
Date: Fri, 01 Feb 2008 08:47:32 +0000
Cc: xen-devel@xxxxxxxxxxxxxxxxxxx
Delivery-date: Fri, 01 Feb 2008 00:47:44 -0800
Envelope-to: www-data@xxxxxxxxxxxxxxxxxx
In-reply-to: <1201824733.9778.6.camel@xxxxxxxxxxxxxxxxxxxxxxxxxx>
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: AchkrxVUVCw5mtCiEdyXwwAWy6hiGQ==
Thread-topic: [PATCH] FP instruction in realmode emulation
User-agent: Microsoft-Entourage/
I am surprised we get there: when emulating FPU instructions we first call
the ->load_fpu_ctxt hook, which calls vmx_do_no_device_fault() ->
setup_fpu() -> clts(). Obviously there is something going wrong though...
can you provide a backtrace?

Whatever the problem is, bear in mind that we should never take a FPU fault
from within Xen and if we do that is a bug in itself.

 -- Keir

On 1/2/08 00:12, "Nitin A Kamble" <nitin.a.kamble@xxxxxxxxx> wrote:

> Hi Keir,
>   Our QA team also found that the latest base kernel was failing on the
> new real mode emulation code. I looked into it, and found out the
> culprit to be guest_mode() call in the do_device_not_available handler.
>   This guest_mode() check over there is no more correct as the guest may
> run in hypervisor context in the form of emulation. The attached patch
> fixes this issue by removing the check. Another fix would be modifying
> the guest_mode() macro. What do you think, Is this check is needed there
> now ?
> Signed-off-By: Nitin A Kamble <nitin.a.kamble@xxxxxxxxx>

Xen-devel mailing list

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