On Tue, 5 Dec 1995, Scott Bartram wrote:

> > Scott Bartram writes:
> > > It may have been a coincidence but my SMC WD8013 network card died
> > > during or just after booting the 1.1 kcadp kernel. Even the DOS/SMC
> > > diagnostics floppy won't find it now.
> > 
> > You will probably find that power cycling fixes it but that resets do
> > not.
> 
> No luck with this.

The same thing happened to me.  I did an installation (from the
distribution disks) of 1.1, using the Adaptec version.  My WD8013
showed up as ed2; it was set to 0x300/10/0xCC000.  A quick look at the
INSTALL file indicated that it would have been ed0 if it were set up
with 0x280/2/0xD0000, but the other default configuration it could be
jumpered to was 0x280/3/0xD0000.  I figured it would be fun to check
if the driver could handle an unexpected IRQ (didn't have the actual
config file handy just then), so I removed the serial port card to
free up IRQ 3, moved the jumper on the WD8013, and powered on.  It
came up with a complaint that ed0 was showing up on the wrong IRQ.
OK, power off, jumper back to the original setting.

And that's the last I've heard from it.  After that, the board won't
speak to NetBSD or anything else.  The little surface mounted LEDs on
it clearly indicate traffic on the network, though, so it's got power
and receives packets off the net properly.  I really can't believe
that changing that jumper back and forth while the board was inserted
into a grounded, powered-off system could fry anything -- that's just
too weird.  Probably, something just chose that time to burn out.

Still, I'll be trying it again after it's had some time out of the bus
to get its circuitry unjammed -- I _have_ experienced hardware lockups
that took up to a day to come unstuck...

-tih
-- 
Tom Ivar Helbekkmo
tih@Hamartun.Priv.NO