In message <199912301739.LAA10663@guild.plethora.net>Peter Seebach writes:

>In message <199912292035.OAA00598@kenku.warthog.com>, Rob Windsor writes:
>>I'm using one on a 1.4.2-ish box, and it appears to be working ok (I'm having
>>problem with NAT dropping packets or something, but I'm still debugging that).
>>
>>Which vendor?  Mine is a D-Link.
>
>Mine too.
>
>Actually, turns out not to be the vr0 - any network card I try has the
>same problem on this board.  ASUS P3B-F, I believe.
>
>What settings are likely to be relevant to attempts to "map device registers"?

If (and that's a big if) the cards ended up in this state due to 
botches in APCI, try this:

	a) shut down machine.
	b) Remove mains power for 20-30 seconds.   ATX motherboards
		capable of wake-on-lan still provide some voltage to the
		PCI bus even when the front-panel ligth says they're off.
	c) restore mains power.
	d) when the machine comes back up, disable ACPI powersave modes
	   (to try and stop this happening again).

this recipe is for the 3c90x cards which do power-save. I dont know
enuf about Intel PC design to tell if it's a BIOS bug or a bug with
the PCI cards. If its really a BIOS bug, that'd explain why it screws
up multiple cards.

Could you let the list know if this helps, for any of your cards?
thanks!

--Jonathan