On Mon, 2 Oct 2000, Andrew Gillham wrote:
> Jared D. McNeill writes:
> > I got it this far right now, but I'm almost 100% positive that I'm calling
> > i386_set_ioperm wrong (is that even the equivalent of Linux's ioperm?)
> 
> This is probably completely "wrong", but:

I did it your way, and here's what happens (note that the dmesg errors
could have happened before, I didn't notice):

[ snip ]

(gdb) run -f
Starting program: /usr/home/jmcneill/3c5x9utils-1.1.orig/./3c5x9 -f
3c5*9 not found at 0x300, status 3404.
If there is a 3c5*9 card in the machine, explicitly set the I/O port
address
  using '-p <ioaddr>

Program exited with code 01.

And I noticed the following in dmesg:
ep0: adapter failure (2003)
ep0: adapter failure (2003)

So I tried the other card:

blackhole:/usr/home/jmcneill/3c5x9utils-1.1.orig# ./3c5x9 -f -p 0x210
3c5*9 not found at 0x210, status 09c5.
If there is a 3c5*9 card in the machine, explicitly set the I/O port
address
  using '-p <ioaddr>

blackhole:/usr/home/jmcneill/3c5x9utils-1.1.orig# dmesg | grep ^ep1 | grep
2003
ep1: adapter failure (2003)
ep1: adapter failure (2003)

Jared