[ltp] network/pcmcia problems

[Friedemann Baitinger]

On Tue, 31 Aug 1999, Peter Soderling wrote:

> Yes . . . my messages file is the same up to this point. But as soon as it
> starts to load the driver, it hangs . . . so I never see anthing more in the
> log. I would've thought it was a corrupt driver or something, except that
> I've gone through the whole compliling process w/ fresh source dozens (or so
> it seems) of times now.

I have the EtherJet 10/100 now running on my TP770X with linux-2.2.12
and pcmcia-3.0.14 by following all the advice given in this thread.
However, there is one problem left: Something is wrong with the driver,
if I ping a remote host im my LAN I experience situations like:
(Please look what happens to packets 7 to 12, the replies show up 18s
later) 

PING zenon.baiti.net (195.2.161.1): 56 data bytes
64 bytes from 195.2.161.1: icmp_seq=0 ttl=64 time=0.7 ms
64 bytes from 195.2.161.1: icmp_seq=1 ttl=64 time=0.4 ms
64 bytes from 195.2.161.1: icmp_seq=2 ttl=64 time=0.4 ms
64 bytes from 195.2.161.1: icmp_seq=3 ttl=64 time=0.4 ms
64 bytes from 195.2.161.1: icmp_seq=4 ttl=64 time=0.4 ms
64 bytes from 195.2.161.1: icmp_seq=5 ttl=64 time=0.4 ms
64 bytes from 195.2.161.1: icmp_seq=6 ttl=64 time=0.4 ms
64 bytes from 195.2.161.1: icmp_seq=13 ttl=64 time=0.4 ms
64 bytes from 195.2.161.1: icmp_seq=14 ttl=64 time=0.4 ms
64 bytes from 195.2.161.1: icmp_seq=15 ttl=64 time=0.4 ms
64 bytes from 195.2.161.1: icmp_seq=16 ttl=64 time=0.4 ms
64 bytes from 195.2.161.1: icmp_seq=17 ttl=64 time=0.4 ms
64 bytes from 195.2.161.1: icmp_seq=18 ttl=64 time=0.4 ms
64 bytes from 195.2.161.1: icmp_seq=19 ttl=64 time=0.4 ms
64 bytes from 195.2.161.1: icmp_seq=20 ttl=64 time=0.4 ms
64 bytes from 195.2.161.1: icmp_seq=21 ttl=64 time=0.4 ms
64 bytes from 195.2.161.1: icmp_seq=22 ttl=64 time=0.4 ms
64 bytes from 195.2.161.1: icmp_seq=23 ttl=64 time=0.4 ms
64 bytes from 195.2.161.1: icmp_seq=24 ttl=64 time=0.4 ms
64 bytes from 195.2.161.1: icmp_seq=25 ttl=64 time=0.4 ms
64 bytes from 195.2.161.1: icmp_seq=7 ttl=64 time=18540.1 ms
64 bytes from 195.2.161.1: icmp_seq=8 ttl=64 time=17540.0 ms
64 bytes from 195.2.161.1: icmp_seq=9 ttl=64 time=16539.9 ms
64 bytes from 195.2.161.1: icmp_seq=10 ttl=64 time=15539.8 ms
64 bytes from 195.2.161.1: icmp_seq=11 ttl=64 time=14539.7 ms
64 bytes from 195.2.161.1: icmp_seq=12 ttl=64 time=13539.5 ms
64 bytes from 195.2.161.1: icmp_seq=32 ttl=64 time=0.4 ms
64 bytes from 195.2.161.1: icmp_seq=33 ttl=64 time=0.4 ms

Working with the remote host in question is hardly possible in this
situation. If I just unlug the card and replug my old NE2000 compatible
card I can continue working even with the existing running applications.

Does anybody have an idea what could cause this? BTW: for both cards
IRQ 3 is used although IRQ 3,4,7 are excluded in the
/etc/pcmcia/config.opts. In the old NE2000 case cat /proc/interupts
shows:

  3:      12393          XT-PIC  pcnet_cs

for the Etherjet it says:

  3:      12393          XT-PIC  eth0

Ideas? Suggestions?


-- 
Friedemann Baitinger       fb@baiti.net      http://baiti.net/fb

Those "other" dudes seem to have forgotten where they came from.  As a
result, they don't have a clue as to where they are going, either...