[tulip-bug] Follow-up from April Message - tcpdump workaround

[David Butcher]

My Hawking PN672TX CardBVus 10/100M Fast Ethernet PC Card (PCMCIA)
is exhibiting the behavior mentioned in the April post by Michael Sparks

OS Is SuSE 7.2 Professional, clean install, unpatched
Kernel is:
Linux chezhall 2.4.4-4GB #1 Fri May 18 14:11:12 GMT 2001 i686 unknown    

Modules:
Module                  Size  Used by
af_packet              11648   2  (autoclean)
mousedev                4032   0  (unused)
hid                    11760   0  (unused)
input                   3168   0  [mousedev hid]
usb-uhci               21840   0  (unused)
usbcore                47120   1  [hid usb-uhci]
ipv6                  126272  -1  (autoclean)
tulip_cb               33440   2
cb_enabler              2784   2  [tulip_cb]
ds                      6896   2  [cb_enabler]
i82365                 23440   2
pcmcia_core            43936   0  [cb_enabler ds i82365]
ipchains               33408   0  (unused)

Boot Messages:
Starting PCMCIA (using scheme: SuSE)cardmgr[215]: starting, version is 3.1.25
cardmgr[215]: watching 2 sockets
cardmgr[215]: initializing socket 0
cardmgr[215]: socket 0: Linksys EtherFast PCMPC200 v2
cardmgr[215]: executing: 'modprobe cb_enabler'
cardmgr[215]: executing: 'modprobe tulip_cb'
cardmgr[215]: executing: './network start eth0'

Symptoms:
The card initializes, and transmits.  Pings go out, DHCP requests go out
(seen in the DHCP server logs), but nothing ever comes back.

Setting up static routing and fixed IP address does not help.  Starting
tcpdump instantly sets the card to normal operation.

This is a very inexpensive and popular card - I will help debug in any way
possible.

Thanks, everyone, for supporting the driver for it.

David