> > Okay, found a DEC document on the web that says the one for the IDE 
> > controller
> > can be ignored for the 164lx.
> >
> > But the USB controller one remains.  :-(
> 
> I think you can actually ignore the USB controller error, too.  I seem 
> to recall those devices don't use PCI interrupt, but come in via the 
> ISA PIC.

Are you sure about that?  I've been poking around and it sure looks
like the USB controller is using PCI interrupts.

I couldn't get SRM to print out any numbers in the 9a range, but poking around
in arch/alpha/pci/pci_eb164.c indicates that 0x9a maps to PCI interrupt 10.
USB activity hits eb164_iointr().  I pulled the SCSI board that also
interrupts at 10 and SRM no longer complains.  (Unfortunately this
didn't fix my USB problems.)

EB164_MAX_IRQ is 24, so there shouldn't be a shortage?