Subject: Re: ray0 in absolutely-current as of Jan 11?
To: None <>
From: Neil A. Carson <>
List: current-users
Date: 02/13/2000 13:45:01
"Christian E. Hopps" wrote:
> "Neil A. Carson" <> writes:
> > Dave Huang wrote:
> > >
> > > > So, I built a new kernel.
> > > >
> > > > BAM!  ray0 gets 10-50 dup packets back per packet it should get.
> > >
> > > You're using cardbus, right? Have you tried setting mem8=1 in pccbb.c?
> > > (see <>)
> > > That seems to do the trick for me...
> >
> > I wonder if this is the same symptom that I get? I find:
> >   - Sometimes I bring up the card and it works fine.
> Please explain fine.  What systems does it work fine between?

Either two NetBSD systems or my laptop (under Win98) with link0 set. The
base stations have been one of the following, both with pretty much the
same behaviour:
  o An old Pentium 133 Liberty computer (standard PC in a small box)
  o An 200MHz IBM Netstation 1000 (PowerPC 600-series) with a
We've also tried the card in a 233MHz modern Pentium laptop (well, a
couple of years old). On this platform, again without cardbus built in,
the driver seems much more stable and has only hiccuped once.

> >   - Sometimes I bring up the card, nothing happens,
> >     and I need to ifconfig it down and then back up
> >     again to make it work.

> When I was developing the driver and was abusing it in random ways, I
> saw this behavoir, basically the card would fail to aquire a network.  I
> do not believe that I have seen this since before I commited the
> driver.  It is possible with all the random isa config/cardbus
> unhappyness going on that similar problems could be seen.

Interesting. This one isn't a big deal, as it's fairly easy to ifconfig