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-users

RE: [Xen-users] Cannot assign device in Xen-unstable [VT-d]

To: "'Bryan York'" <bryan.york@xxxxxxxxx>, <xen-users@xxxxxxxxxxxxxxxxxxx>
Subject: RE: [Xen-users] Cannot assign device in Xen-unstable [VT-d]
From: "Dustin Henning" <Dustin.Henning@xxxxxxxxxxx>
Date: Tue, 12 Aug 2008 12:54:15 -0400
Cc:
Delivery-date: Tue, 12 Aug 2008 09:54:53 -0700
Envelope-to: www-data@xxxxxxxxxxxxxxxxxxx
In-reply-to: <5bcc2bd80808081735x199d777es5c151614b057fe33@xxxxxxxxxxxxxx>
List-help: <mailto:xen-users-request@lists.xensource.com?subject=help>
List-id: Xen user discussion <xen-users.lists.xensource.com>
List-post: <mailto:xen-users@lists.xensource.com>
List-subscribe: <http://lists.xensource.com/mailman/listinfo/xen-users>, <mailto:xen-users-request@lists.xensource.com?subject=subscribe>
List-unsubscribe: <http://lists.xensource.com/mailman/listinfo/xen-users>, <mailto:xen-users-request@lists.xensource.com?subject=unsubscribe>
Organization: PRD, Inc.
References: <5bcc2bd80808081735x199d777es5c151614b057fe33@xxxxxxxxxxxxxx>
Reply-to: Dustin.Henning@xxxxxxxxxxx
Sender: xen-users-bounces@xxxxxxxxxxxxxxxxxxx
Thread-index: Acj5t/jipwxA7US3S/W2Q7+CRiHI7QC4eWiA
        I don?t currently do anything with PCI passthrough, though I would
like to in the future.  As such, I am not going to bother going through all
of the attachments, and my suggestions may be bad ideas.  However, I am
curious as to whether or not you tried only assigning 01:00.0, I mean, maybe
since they're on the same bus it attaches both and you're already good
(hence the already assigned to another domain error).  Otherwise, if it's
hidden from dom0, then even if they are on the same bus and domU doesn't see
it, nothing should be talking to it, so unless it's specifically necessary
on domU, that might not matter (based on your post, I am assuming the fact
that they are on the same bus really matters to begin with [for any reason
other than potential instability or lack of isolation if two domains share
the bus]).  Again, I really don't know, but I can't imagine there would be
any permanent harm in trying the thoughts above.  I am also assuming there
is no way to check and see what "other domain" it is "already assigned" to
based on the fact that no one has suggested trying such a thing (though the
list wasn't very busy this weekend).
        Dustin

From: xen-users-bounces@xxxxxxxxxxxxxxxxxxx
[mailto:xen-users-bounces@xxxxxxxxxxxxxxxxxxx] On Behalf Of Bryan York
Sent: Friday, August 08, 2008 20:36
To: xen-users@xxxxxxxxxxxxxxxxxxx
Subject: [Xen-users] Cannot assign device in Xen-unstable [VT-d]

Hello,

I FINALLY got Xen working with VT-d. I compiled Xen-unstable (3.3) from
Mercurial on CentOS 5.2, but compiled my own kernel-xen from the latest
CentOS SRPM. (Made sure pciback was built-in and not a module, as the stock
kernel-xen has it as a module.) I'm probably going to write a how-to soon,
so stay tuned!

I have a JMicron SATA/PATA PCI Express controller with a DQ35JO motherboard
and a Q6600 processor.

My grub entry:

title CentOS (2.6.18-92.el5.xenpcibackxen)
        root (hd0,0)
        kernel /xen-3.3.0-rc4-pre.gz vtd=1 iommu=1
        module /vmlinuz-2.6.18-92.el5.xenpcibackxen ro
root=/dev/VolGroup00/LogVol00 pci=nommconf rhgb
pciback.hide=(01:00.0)(01:00.1)
        module /initrd-2.6.18-92.el5.xenpcibackxen.img

xm create hd.hvm 
Using config file "./hd.hvm".
Error: fail to assign device(1:0.1): maybe it has already been assigned to
other domain, or maybe it doesn't exist.

My dmesg, xm dmesg, hvm, and lspci -vv are attached. What's weird is that
Xen appears to be able to tell it grabbed 01:00.0. However I cannot grab
just 01:00.0 because it is on the same PCI Express bus. I thought the
problem initially was the fact that the JMB363 ide chipset was being picked
up by the generic ATA driver, so I removed that and recompiled, but that
didn't change anything. pciback.hide seems to grab both controllers per my
dmesg. (They are on the same board.)

Any ideas?

Thanks,

-Bryan



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

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