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] multiple dom0s (was: Re: [Xen-devel] ragarding xen in sl

To: xen-users@xxxxxxxxxxxxxxxxxxx
Subject: Re: [Xen-users] multiple dom0s (was: Re: [Xen-devel] ragarding xen in sles10)
From: Mark Williamson <mark.williamson@xxxxxxxxxxxx>
Date: Wed, 26 Dec 2007 14:58:32 +0000
Cc: jim burns <jim_burn@xxxxxxxxxxxxx>
Delivery-date: Wed, 26 Dec 2007 06:59:33 -0800
Envelope-to: www-data@xxxxxxxxxxxxxxxxxx
In-reply-to: <200712251100.15788.jim_burn@xxxxxxxxxxxxx>
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/cgi-bin/mailman/listinfo/xen-users>, <mailto:xen-users-request@lists.xensource.com?subject=subscribe>
List-unsubscribe: <http://lists.xensource.com/cgi-bin/mailman/listinfo/xen-users>, <mailto:xen-users-request@lists.xensource.com?subject=unsubscribe>
References: <B6F4AA98FA77754890FC57B49D9C7037014C6CFD@xxxxxxxxxxxxxxx> <200712100159.03572.mark.williamson@xxxxxxxxxxxx> <200712251100.15788.jim_burn@xxxxxxxxxxxxx>
Sender: xen-users-bounces@xxxxxxxxxxxxxxxxxxx
User-agent: KMail/1.9.6 (enterprise 0.20070907.709405)
> On Sun December 9 2007 8:59:02 pm Mark Williamson wrote:
> > c) compile pciback statically into the kernel and then put a pciback.hide
> > directive onto your kernel command line and reboot
>
> CONFIG_XEN_BACKEND=y
> CONFIG_XEN_BLKDEV_BACKEND=m
> CONFIG_XEN_NETDEV_BACKEND=m
> CONFIG_XEN_PCIDEV_BACKEND=m
> CONFIG_XEN_PCIDEV_BACKEND_VPCI=y
>
> I have the above in my fedora xen kernel. Is that sufficient. (BTW, SuSE
> has the first line as 'm' also.)

This means that the PCI backend is built as a module (the additional 
XEN_PCIDEV_BACKEND_VPCI=y is just configuring an option, not building 
something statically into the kernel).

You can still use PCI passthrough like that, but just putting pciback.hide on 
your kernel command line won't work - you'd need to pass them to the module 
at load time.  Whatsmore, you need to make sure that pciback is able to claim 
those devices, which is only possible if they're not bound to another driver.  
So either you need to make sure pciback is loaded before other drivers for 
the hardware, or you need to unbind the hardware from whatever driver has 
grabbed it and then rebind it to pciback.

There are some instructions at: 
http://wiki.xensource.com/xenwiki/Assign_hardware_to_DomU_with_PCIBack_as_module
which may help you.  I've not actually tested them myself but last time I 
looked through them they looked fairly sane.

You ought to make sure you're familiar with the drawbacks and limitations of 
PCI passthrough before deploying it in production, as there are a few fairly 
significant caveats (e.g. that the domain with PCI hardware is effectively as 
privileged as dom0, since it can probably abuse that hardware to take over 
the machine if it acts maliciously).  VT-d will probably help mitigate some 
of these limitations eventually - support for VT-d with HVM domains is 
already in xen-unstable, but not for PV yet, I think.

Cheers,
Mark

-- 
Dave: Just a question. What use is a unicyle with no seat?  And no pedals!
Mark: To answer a question with a question: What use is a skateboard?
Dave: Skateboards have wheels.
Mark: My wheel has a wheel!

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

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