Hello Bill,

> > > The kernel booted, got to the point of scanning scsi busses, failed to 
> > > find any of the four SCSI devices, then reported:
> > > 
> > > panic Stuck interupt 00001000/00001000
> 
> > Hmm, I see a problem in the code - if we interrupt on a high ipl we can get stuck if there's a level-triggered IRQ with a lower
> >  priority - we'd endlessly mark it as pending but find it again in GC's level register on every loop. 
> > I think for now it would be best to just leave the loop instead of panic()ing.
> Please try the attached patch, it should prevent getting stuck because of masked-out interrupts.
> Btw. can you reproduce the panic?
Bah, I should sleep more, the patch is useless. Anyway, it would still be nice to know which IRQ caused the panic.

have fun
Michael