Re: Please test xen/stable-2.6.32.x in xen.git

[Jeremy Fitzhardinge <jeremy@goop.org>]

On 04/01/2010 12:38 AM, Stefan Kuhne wrote:
> Am 31.03.2010 22:31, schrieb Jeremy Fitzhardinge:
>    
>> On 03/31/2010 01:26 PM, Stefan Kuhne wrote:
>>      
>>> Am 31.03.2010 21:06, schrieb Jeremy Fitzhardinge:
>>>        
>>>> On 03/31/2010 11:14 AM, Stefan Kuhne wrote:
>>>>          
> Hi Jeremy,
>
>    
>> You mean this patch?
>> http://lists.xensource.com/archives/html/xen-devel/2010-02/msg00832.html
>>
>>      
> Yes.
>
>    
>> OK, but can you post all the output before the hang?  And also the
>> content of /proc/interrupts (ideally before and after starting the
>> domain with the passthrough WLAN card).
>>
>>      
> I've take a look on the real console.
> I got:
> irq 18: nobody cared (try booting with the "irqpoll" option)
> handlers:
> [<edb28270>] (usb_hcd_irq+0x0/0x80 [usbcore])
> Disabling IRQ #18
>    

Is irq 18 being shared with your wlan device?  Is it USB which stops 
working after this message, or does it take the whole machine out?  Does 
it work if you disable USB (unload the module, or boot with nousb)?

Konrad, how is this supposed to work?  Is there supposed to be a driver 
in dom0 which mops up "spurious" interrupts being handled in another domain?

> This is 2.6.32.x on xen-4.0.0-rc9 an 2.6.20 (Gentoo) domU kernel with
> more patches.
>
> Another question,  cpufreq isn't running with 2.6.32 from you (Intel
> T5200 CPU). 2.6.31 works fine.
>    

Hm, yeah, it seems to work sometimes and not others.  I'm not sure why.  
Is ACPI_PROCESSOR(_XEN) =y or =m in your .config?

Thanks,
     J