On 20/06/07, Brett Lymn <blymn@baesystems.com.au> wrote:
> On Wed, Jun 20, 2007 at 11:30:55AM +0100, Chavdar Ivanov wrote:
> >
> > Disabling ehci fixes the problem; ACPI seems to work.
> >
>
> That actually makes sense - it seems to be IRQ 7 that is causing an
> interrupt storm.  I am thinking perhaps it has something to do with
> the message about forcing the bios to give up possession of the
> device.  I did try checking for and ack'ing interrupts on the ehci
> device when forcing it out of the BIOS's clutches but that didn't seem
> to help any.  The other thing I was going to try was tweaking the bios
> to reserve IRQ 7 just to see of the problem persisted.

I'll try the latter.

In the mean time, the diagnostic kernel panics as follows:

...
root file system type: ffs
init: copying out path '/sbin/init' 11
kernel: double fault trap code=0
Stopped in pid 1.1 (init) at netbsd:Xintr_legacy7+0x48: movq %r9, 0x28(%rsp)

so you guess about interrupt 7 is probably right.

If I disable acpi (but not ehci), it boots, but the nfe driver emits
'watchdog timeout' and there is no network.

Hope that helps.

>
> --
> Brett Lymn
>

Chavdar Ivanov