[Sheflug] Network/cardmgr issues

[Richard Balthazor <R.Balthazor [at] xxxxxxxxxxxxxxx>]

Hello folks,

Hopefully someone can point me in the right direction.  I've been
struggling to put Redhat 7.3 on an old Viglen Dossier V (a cheap far
eastern Pent MMX), and I'm having network issues.  The network card is a
TDK 5670 combo ethernet/modem which uses pcnet_cs.  All the relevant
modules are present and loaded.  pcmcia starts apparently successfully.  
cardmgr starts apparently successfully.  eth0 comes up apparently
successfully.  And the only hint of an error so far is in
/var/log/messages:

cardmgr[250]: config error, file `./config.opts` line 8: no function
bindings

The file in question is /etc/pcmcia/config.opts, and line 8 is

include port 0x100-0x4ff, port 0xc00-0xcff


And at this point I find that my network is not happy: 

- pings to a numeric IP return "destination host unreachable" for a while
until finally a few packets scrape through all of a bundle, but there's
something like 95%+ packet loss.

- tcpdump shows lots of "arp who has <ping destination> tell <me>" and
"<me> > <ping destination> icmp: echo request (DF)".  It also shows a
smattering of other network activity (broadcasts from other
machines)

- other network processes (telnet,  DNS lookups, etc) fail, presumably
through this 95%+ packet loss.

My own thoughts: there's obviously massive packet loss but where?  Given
that pcmcia, cardmgr and eth0 are declared up and allegedly running OK,
what's the best way to check they are in fact OK, given that I don't trust
the [OK] tag?  Any thoughts as to how to chase this down?

cheers in advance,
Richard.



___________________________________________________________________

Sheffield Linux User's Group -
http://www.sheflug.co.uk/mailfaq.html

  GNU the choice of a complete generation.