[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: [Xen-devel] [VTD][patch 0/6] HVM device assignment using vt-d


  • To: "Muli Ben-Yehuda" <muli@xxxxxxxxxx>
  • From: "Kay, Allen M" <allen.m.kay@xxxxxxxxx>
  • Date: Fri, 6 Apr 2007 13:28:36 -0700
  • Cc: xen-devel@xxxxxxxxxxxxxxxxxxx
  • Delivery-date: Fri, 06 Apr 2007 13:27:27 -0700
  • List-id: Xen developer discussion <xen-devel.lists.xensource.com>
  • Thread-index: Acd4hSRnj7/thwQWQPeGB870ZvalOAAAMCMQ
  • Thread-topic: [Xen-devel] [VTD][patch 0/6] HVM device assignment using vt-d

Hi Muli,

I will need to check with people on HW availability outside of Intel and
get back to you.

I just checked out your iommu patch.  Yes, this is a framework we can
look into using.  It is just a matter of mapping hypercalls and
iommu_ops.  It is possible we may need to add some operations.  I would
be happy to work with you on this.

I see that your hypercall definitions do not contain
interrupt/ioport/mmio related stuff.  I'm wondering how do you handle
these on Calgary implementation.

Allen

>-----Original Message-----
>From: Muli Ben-Yehuda [mailto:muli@xxxxxxxxxx] 
>Sent: Friday, April 06, 2007 12:53 PM
>To: Kay, Allen M
>Cc: xen-devel@xxxxxxxxxxxxxxxxxxx
>Subject: Re: [Xen-devel] [VTD][patch 0/6] HVM device 
>assignment using vt-d
>
>On Fri, Apr 06, 2007 at 11:21:29AM -0700, Kay, Allen M wrote:
>> Keir,
>> 
>> Following 6 patches are for enabling PCI device assignment to a HVM
>> domain using vt-d.  We have done no-harm-done testing for PV and
>> non-vtd HVM guests on 32PAE, x86_64 and IPF.  See below for details
>> of the patch.
>
>Cool! I have a couple of questions:
>
>1. When is VT-d hardware expected to be available?
>2. We need a way for this to work with multiple IOMMUs - VT-d, AMD's
>   upcoming IOMMU and IBM's Calgary/CalIOC2 family of IOMMUs are the
>   ones that are either available or will be soon. From a quick look
>   at the code your patch is VT-d specific? What I've done so far in
>   the xen-iommu tree is to add a new platform op by which dom0 tells
>   Xen about the presence of a new IOMMU and then add a common layer
>   in Xen for different IOMMUs to hook into using an iommu_ops
>   structure. Will this work for you, or do you have something else in
>   mind?
>
>More to come as I digest your patches. It's great to finally see VT-d
>support out-there!
>
>Cheers,
>Muli
>

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


 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.