> Actually, the report from the person with a 3c905 was different > altogether. Every report of this bug I've seen has happened with the > eepro100 driver loaded. Might be worth moving to a more recent eepro100 driver. One to merge in a future -ac release to see