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] VT-x and frontend/backend drivers

To: "Mark Ryden" <markryde@xxxxxxxxx>, "Xen Mailing List" <xen-devel@xxxxxxxxxxxxxxxxxxx>
Subject: RE: [Xen-devel] VT-x and frontend/backend drivers
From: "Yu, Ke" <ke.yu@xxxxxxxxx>
Date: Mon, 9 Jan 2006 17:20:04 +0800
Delivery-date: Mon, 09 Jan 2006 09:26:24 +0000
Envelope-to: www-data@xxxxxxxxxxxxxxxxxxx
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: AcYUSbo/fGaLf8vkSdmMQ5QIQ6QvwgAngVjw
Thread-topic: [Xen-devel] VT-x and frontend/backend drivers
Mark Ryden wrote:
> Hello,
>   As I saw in this mailing list, there should be some
>       emulation solution for drivers when using VT-x processors.
> See for example :
> My question is this;
>   As I understand, the VMX support in the new  VT-x processors
> enables us to prevent Guest OSs from performing
> unwanted instructions (HLT , for example). This is
> done by using the root and non-root VMX operation.
>  Thus , we don't use paravirtualization and we
> can run unmodified guest.
>   Is there a possibility or any intention of using the
> backend-frontend solution of Xen (maybe with some fixes) when using
> VT-x processors ? wouldn't it be much better in terms of
> performance than using hardware emulation (based on QEMU maybe)?

Agree :) . This is a good way to improve VMX  domain performance.

Idealy, if we can implement the xen infrastructure (e.g. hypercall,
event channel, memory map etc.) in kernel module, porting those
frontend/backend drivers should be easy and may be only small

One challenge is the hypercall parameter passing.  In DomU case,
hypervisor can use copy_from_user/copy_to_user to pass hypercall
parameter, since they are in the same address space. In VMX domain case,
the domain use different address space than hypervisor, so the
copy_**_user doesn't work. 

Best Regards
> Regards,
>       --MR
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@xxxxxxxxxxxxxxxxxxx
> http://lists.xensource.com/xen-devel

Xen-devel mailing list

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