to both the EEPROM *and* the PHY on the Ethernet interface now, at least, though it is still not completely working. Many thanks to Stephen Goadhouse at ADI for some hints.