mrg@eterna.com.au said:
   > pci_find_device() fails in the case of >1 matching device.  is this
   > going to be a problem for this change?
   
   It looks for the ISA bridge, there should be exactly 1.
   
   > i'd expect it to be onboard ide
   
   Yes, part of the chipset - it appears that meaningful revision
   information is only in the PCI-IDE configuration header.


ok... this hack is probably "ok" in that it isn't the first of
it's kind :-)