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

[Xen-users] Re: Same IRQ in dom0 and domU by different hardware causes d

To: xen-users@xxxxxxxxxxxxxxxxxxx
Subject: [Xen-users] Re: Same IRQ in dom0 and domU by different hardware causes dom0 crash
From: Sylvain Le Gall <sylvain.le-gall@xxxxxxxxxxxxxxxxx>
Date: Wed, 16 May 2007 21:30:47 +0000 (UTC)
Delivery-date: Wed, 16 May 2007 15:38:32 -0700
Envelope-to: www-data@xxxxxxxxxxxxxxxxxx
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: <200705081121.24894.ibc@xxxxxxxxxxxx>
Sender: xen-users-bounces@xxxxxxxxxxxxxxxxxxx
User-agent: slrn/0.9.8.1pl1 (Debian)
On 08-05-2007, Iñaki Baz Castillo <ibc@xxxxxxxxxxxx> wrote:
> Hi, this is my first mail in this list. I've installed Xen in Debian Etch and 
> created a virtual domU Debian Etch as well.
>
> In domU I've installed an Asterisk and want to use a Beronet ISDN card:
>
>   ~# lspci
>   0a:01.0 ISDN controller: GmbH ISDN network Controller [HFC-4S] (rev 01)
>
> So I pass the iSDN card PCI to domU by adding in dom0 "/boot/grub/menu.lst":
> ---
> module /boot/vmlinuz-2.6.18-4-xen-686 root=/dev/sda1 ro pciback.hide=(0a:01.0)
> ---
>
> After that, I see the PCI in domU:
>
>   ~# lspci
>   00:00.0 ISDN controller: GmbH ISDN network Controller [HFC-4S] (rev 01)
>
>
> -- PROBLEM --
> My first problem is that dom0 and domU share IRQ 18:
>
> ** dom0 **
>   ~# cat /proc/interrupts
>               CPU0           CPU1
>    18:      14140          0        Phys-irq  uhci_hcd:usb3
>
> ** domU **
>   ~# cat /proc/interrupts
>               CPU0
>    18:      16507        Phys-irq  HFC-multi
>
> Obviously, the interrupts so big number by IRQ18 in dom0 is "false", they are 
> in fact interruptions of the HFC card in domU.
>
> If I reboot or destroy domU I get a dom0 message:
>   Disabling IRQ #18
>
> and dom0 looses net, keyboard and so. I must restart it.
>
>
> -- WORKAROUND --
> I can quit the module "uhci_hcd" in dom0 which is using IRQ 18 and the 
> problem 
> dissapears, but of course it's not a good/clean solution.
>
> But I have another problem with this workaround: I'd like to quit this module 
> for the start by adding to "/etc/modprobe.d/blacklist":
>   blacklist  uhci_hcd
>
> but it doesnt' work and the module is loaded anyway. I can't understand 
> because I use that method to avoid loading other modules in the same Debian 
> Etch and it works. Any idea about this problem?
>
>
> -- SOLUTION --
> I'd like a way to dissable IRQ 18 in dom0. Anyway I think that the only 
> solution is by editing the Bios and dissabling or chanching that IRQ (haven't 
> tryied it yet) but I'd like to listen to other suggestions.
>
>
> -- POSSIBLE BUG? --
> Waht I really care is this IRQ sharing between dom0 and domU by different 
> hardware. Is it really normal?
> I've found other case exactly the same as mine without no response:
>   http://lists.xensource.com/archives/html/xen-users/2006-05/msg00967.html
>
>

I have exactly the same kind of problems:
* hide audio card from dom0
* unhide audio card in domU
* audio card share IRQ with eth0, hda
* after a long time (more than 24 hours running), i loose audio card,
  eth0 and hda. 
* the mouse become unresponsive (not frozen, only very slow)
* eth0 is dead 
* hda: lost interrupt ide-cd: cmd 0x3 timedout hda: request sense
  failure...

I have upgraded my BIOS et al. I cannot find a good solution to this
problem...

For now, i am trying to run vanilla linux (without xen). If it perform a
week of uptime, i will consider the audio/eth0 problem as a Xen related
problem (the host is actually running fine since 40 hours which is 12
hour better than with Xen)

Regards,
Sylvain Le Gall


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

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