mcr@sandelman.ottawa.on.ca (Michael C. Richardson) writes:
>   With my recent experiences with the SMC dual ethernet cards (some
> revisions worked, others failed badly) with NetBSD 1.3 alpha, I'm
> hesistant about buying something.

[ This is my second msg on this subject.  The first msg never echoed
back here.  Curious.  Did I run afoul of a spam filter or something? -wsr ]

My old msg:
    The problem may not be limited to the the dual cards.  I have two
    smc-9332DST cards (aui and rj-45) they work fine.  I also have one
    smc-9332BDT (rj-45 only).  It fails at probe time.

	de0 at pci0 dev 14 function 0: unable to map device registers

Just in case that msg did make it to most folks, I'd like to clear up
the mystery I'd noted with different SMC cards (smc-9332DST / 
smc-9332BDT).  I ended up violating one of my cardinal rules, never
tear down a working system to provide spares for testing. ;-)

The "BDT" card worked just fine in a working netbsd box.  I did an
exhaustive (exhausting too) search of various Pheonix bios setting on
my new intel AL440LX, and noted that if I chose "PnP OS = NO" then
things worked.  Now if I'd only understand what all those setting
really did, I'd be much happier.  I'm surmising that in this case the
bios sets up a few extra things for the OS when one turns this on.
I'm not sure what the down-side is.  Anyone have any good references
for this sort of crap?  I hate groping in the dark like this.

I'm still stuck with something assigning the same interrupt to "sb0"
and "wss0" as "lpt0" and "de0" respectively.  Ticking off the
interrupts that the kprintfs emit, there should be plenty of
interrupts for what I want to do.  (com1, com0, fdc0, lpt0, de0,
wdc0).

-wolfgang