2.0.32 problems with eepro100 card

bwoodard@cisco.com
Mon, 05 Jan 1998 16:12:03 -0800


I seem to be having some problems with the 2.0.32 version of eepro100 driver.

Two things seem to be happening:

First of all the syslog seems to be filling with messages such as these:

Jan 5 12:51:39 borlach kernel: eth0: Transmit timed out: status 0050 command 0000.
Jan 5 12:51:39 borlach kernel: eth0: Tx timeout fill index 466446 scavenge index 466432.
Jan 5 12:51:39 borlach kernel: Tx queue 00000000 00000000 000ca000 000ca000 000ca000 0000a000 000ca000 000ca000 000ca000 000ca000 000ca000 000ca000 000ca000 400ca000 0000a000 0000a000.
Jan 5 12:51:39 borlach kernel: Rx ring 00000003 00000003 00000003 00000003
00000003 00000003 00000003 00000003 c0000003 00000003 00000003 00000003 00000003 00000003 00000003 00000003.
Jan 5 12:51:39 borlach kernel: eth0: Trying to restart the transmitter...
Jan 5 12:51:45 borlach kernel: eth0: Transmit timed out: status 0050 command 0000.
Jan 5 12:51:45 borlach kernel: eth0: Tx timeout fill index 466448 scavenge index 466433.
Jan 5 12:51:45 borlach kernel: Tx queue 0000a000 00000000 000ca000 000ca000 000ca000 0000a000 000ca000 000ca000 000ca000 000ca000 000ca000 000ca000 000ca000 000ca000 000ca000 0000a000.
Jan 5 12:51:45 borlach kernel: Rx ring 00000003 00000003 00000003 00000003
00000003 00000003 00000003 00000003 00000003 c0000003 00000003 00000003 00000003 00000003 00000003 00000003.
Jan 5 12:51:45 borlach kernel: eth0: Trying to restart the transmitter...

and secondly:

The SE in that office (borlach, being in a cisco field office is under
the close scrutiny of one of those highly trained cisco service
engineers) and he reports that it was broadcasting from hardware
address 0000000000ff every 100 to 200 micro seconds. This was bad; it
was crashing the inferior TCP/IP stack of the windows boxes on the
same subnet. A simple reboot seemed to fix the problem momentarily.

One thing that was kind of scary was when I looked at /proc/interupts

0: 1176122 timer
1: 6 keyboard
2: 0 cascade
4: 8 + serial
8: 0 + rtc
9: 99514729 Intel EtherExpress Pro 10/100 Ethernet
13: 1 math error
14: 92497 + ide0

Having about 99 times as many ethernet interrupts as timer interrupts
seems pretty bad?

Does anybody know what the reason for either one of the above and how
to fix it?