Tulip tx and rx stopage

[Jim Bauer]

This was originally posted to the linux-kernel mailing list.
In addition to everything below, I was able to get this
to happen with tulip.c-0.90 on linux-2.1.129 as well.
It did seem to be a bit harder to reproduce with 0.90
then with 0.83.

p.s. I am not presently subscribed to this mailing list.

-----FW: <XFMail.981121165629.jfbauer@home.com>-----

Date: Mon Nov 23 22:04:17 1998
From: Jim Bauer <jfbauer@home.com>
Subject: tulip driver problem

I mentioned this before in l-k in conjunction with a UP-flu report,
but I suspect it got lost in the noise.  I may have discovered
a bug in the tulip driver in at least 2.1.127 - 2.1.129.
Under somewhat heavy load, the driver will either stop trasnmitting
or stop recieving.  This is apparently similar to a previously
reported problems with the eepro drivers.

I can easily recreate the problem by running...

        find /usr -type f -exec wc {} \;

while ping flooding the system.  I have not tried this on kernel
version earlier then 2.1.127..

Boot messages, tulip driver is *not* a module.

tulip.c:v0.83 10/19/97 becker@cesdis.gsfc.nasa.gov
eth0: Digital DS21041 Tulip at 0x6100, 21041 mode, 00 c0 f0 21 72 0b, IRQ 9.
eth0:21041 Media information at 30, default media 0800 (Autosense).
eth0:  21041 media 00 (10baseT), csr13 0004 csr14 0000 csr15 0000.
eth0:  21041 media 00 (10baseT), csr13 0000 csr14 0000 csr15 0000.

Then I get the following errors (if transmit stops working) while
doing the ping flood and the above mentioned find.

eth0: 21041 transmit timed out, status fc660000, CSR12 000001c8, CSR13 ffffef05, CSR14 ffffff3f, resetting...
eth0: 21041 transmit timed out, status fc660010, CSR12 000052c8, CSR13 ffffef09, CSR14 fffff7fd, resetting...
eth0: 21041 transmit timed out, status fc660010, CSR12 000000c8, CSR13 ffffef05, CSR14 ffffff3f, resetting...
eth0: 21041 transmit timed out, status fc660010, CSR12 000050c8, CSR13 ffffef09, CSR14 fffff7fd, resetting...
eth0: 21041 transmit timed out, status fc660010, CSR12 000000c8, CSR13 ffffef05, CSR14 ffffff3f, resetting...
eth0: 21041 media switched to 10baseT.
eth0: 21041 transmit timed out, status fc660010, CSR12 000000c8, CSR13 ffffef05, CSR14 ffffff3f, resetting...
eth0: 21041 transmit timed out, status fc660010, CSR12 000050c8, CSR13 ffffef09, CSR14 fffff7fd, resetting...

ifconfig eth0 down; ifconfig eth0 up
fixes the problem.

If receive stops working, but transmit is still ok, I get no error messages.

System is a UP 200MHz K6, all SCSI, Kingston KNE40T tulip based
ethernet card.

--------------End of forwarded message-------------------------

Jim Bauer, jfbauer@home.com