[Fwd: Re: [vortex] IBM-Thinkpad with Mini-PCI Card 3c556B Kernel2.2.16- 2.4.1]

Uwe Tangen tangen@gmd.de
Thu, 08 Feb 2001 18:44:09 +0100


Dear Donald, 


thanks a lot for your help. I just found out what the problem
has been.
I got fooled by two concurrent mistakes. You had been completely
right with name-server and tcpdump. Though a name-server was
accessible
via gateway and yp, tcpdump probably only looks in /etc/hosts.
The 
second problem was that I made tcpdump -i eth1 and then pulled
the
cable on eth0 which immediately has frozen tcpdump. This problem
would
have not been apparent if I had shut eth0 down before pulling
the cable.

Still, what is not understood by me, is, why did ping {ip of the
gateway}
also fail? ping should not do a name-server lookup? It might be
possible
that Windows 2000 (the driver from 3Com) changed something in
the EEPROM
of of card. But vortex-diag and mii-diag give the same answer,
so I still
don't really understand why I got so much trouble.


I am happy that things are working now and if you ever will
visit
Bonn (former capital of Germany) let me know - I would be happy 
to invite you for a drink dinner or what else.


Thanks a lot

Uwe


Donald Becker wrote:
> 
> On Thu, 8 Feb 2001, Uwe Tangen wrote:
> 
> > thanks a lot for your information. I do have the same output in
> > /proc/pci (concerning
> > the 3Com card). Only the /var/log/messages section is slightly
> > different:
> >
> > Feb  8 17:51:59 rihm kernel: 3c59x.c 19Oct00 Donald Becker and
> > others http://www.scyld.com/network/vortex.html
> > Feb  8 17:51:59 rihm kernel: eth1: 3Com 3c556B Laptop Hurricane
> > at 0x1800,  00:00:86:4b:36:8d, IRQ 5
> >
> > ^^^^^ yours is IRQ 11
> > Feb  8 17:51:59 rihm kernel: eth1: CardBus functions mapped
> > e8101000->d80d7000 (PCMCIA committee brain-damage).
> > Feb  8 17:51:59 rihm kernel:   8K byte-wide RAM 5:3 Rx:Tx split,
> > MII interface.
> > Feb  8 17:51:59 rihm kernel:   MII transceiver found at address
> > 0, status 786d.
> >
> > ^^^^^  yours is 7869
> 
> 786d means that the card has link beat.
> 7869 means no link beat.
> 
> The transceiver address should be either #1 or #24.  You should get a
> driver update.
> 
> Donald Becker                           becker@scyld.com
> Scyld Computing Corporation             http://www.scyld.com
> 410 Severn Ave. Suite 210               Second Generation Beowulf Clusters
> Annapolis MD 21403                      410-990-9993


-- 
Dr. Uwe Tangen    GMD - German National Research
                  Center for Information Technology
                  Schlo"s Birlinghoven, D-53754 Sankt Augustin,
Germany
Tel.: +49 2241 14 1530; Fax.: +49 2241 14 1511; e-mail:
Uwe.Tangen at gmd.de