Ok,

thanks for the reports.  I dug into the "pcmcia patches in
current (6/30/97)" thread and actually read Chris' initial
message on that thread, and it was right on target for fixing my
problem.  You may remember me reporting the machine had a 3c595,
and since it uses the elink3 back-end driver...

So it's not weirdness of the Cyrix chips, but the compiler bug
which hosed me.  The machine in question is now up and running
just fine.

I saw someone else asked for a GENERIC i386 kernel with the
elink3 patches in them; I've started compiling one from sources a
little less than a week old and will make it available shortly.

Regards,

- H=E5vard