Port-xen archive

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

Re: XEN and cardbus



On Mon, Jan 25, 2010 at 10:20:08AM +0100, Juraj Hercek wrote:
>> It looks like you don't have ioapic in your kernel config file. Is it
>> intentionnal ?
>>    
> I'm a bit confused, now, because I see "ioapic*        at mainbus? apid  
> ?" line in XEN3_DOM0 which I include to my customized kernel (and I'm  
> not deconfiguring it later). Also binary contains it, I believe:
>
> $ zcat netbsd-XEN3PAE_DOM0_MASTER.gz | strings | grep _CFG_.*ioapic
> _CFG_###> # boot messages with MPBIOS, acpi and ioapic can be quite large
> _CFG_###> ioapic*\011\011at mainbus? apid ?
>
> I also tried to enable PCI_INTR_FIXUP and also enabled "lapic" in Xen  
> command line. But it still tries to bind IRQ 255.
>
> Is there some other, way to enable ioapic?

Not that I know. This would mean that on this platform, the kernel
can't find the ioapic (or it doesn't have one but I believe all
CPUs capable of running Xen do have an ioapic).
Can you boot a GENERIC kernel and send the dmesg (and test if the
cardbus slot is working or not) ?

>
>> Looks like xen_intr_map() was called with *pirq = 255, that's wrong
>
> Ok. I tried it the other way around - to disable ioapic with "no ioapic"  
> in my kernel config. And the kernel booted, also rtk0 adapter connected  
> to cardbus was recognized, but when I did "ifconfig up rtk0" whole thing  
> gets frozen. This is excerpt from dmesg:
>
> cbb0: cacheline 0x8 lattimer 0x20
> cbb0: bhlc 0x822008
> pci_intr_map: no mapping for pin B (line=ff)
> cbb0: couldn't map interrupt
> cardslot0 at cbb0
> cardbus0 at cardslot0: bus 3
> pcmcia0 at cardslot0

So the interrupt is still 255, but now it doesn't try to register it
with Xen.

-- 
Manuel Bouyer <bouyer%antioche.eu.org@localhost>
     NetBSD: 26 ans d'experience feront toujours la difference
--


Home | Main Index | Thread Index | Old Index