On Thu, 24 Apr 1997, Harald Koenig wrote:
> there have been a number of reports at least for 3c905 (Boomerang) card owners
> who get driver lockups for very high netloads (even one very fast host-to-host
> transfer is sufficient) which result in lots of these kernel messages:
>
> eth0: transmit timed out, tx_status 88 status e000.
> eth0: Transmitter encountered 16 collisions -- network network cable problem?
> Flags; bus-master 1, full 1; dirty 560482 current 560498.
> Down list 07ffe940 vs. 07ffe920.
> 0: 07ffe920 length 8000002a status 8000002a
> 1: 07ffe930 length 80000052 status 80000052
> 2: 07ffe940 length 8000009e status 0000009e
> 3: 07ffe950 length 8000002a status 0000002a
> 4: 07ffe960 length 8000002a status 0000002a
Not only 3c905 cards are affected, but "Vortex" cards (3c59x), too. I was
able to get this error once on two machines with 3c59x; one machine used
v0.39, the other machine v0.40 of the driver.
[...]
> btw: we're still using the "boomerang.c" driver v0.37 because this
> was the last version which does not show these very bad receive
> performace problems as v0.39 or 0.40 do now. but at least v0.39
Yes, with 0.40 I get only about 450K/sec. when receiving data, but with
0.37 I get 9MB/sec ! But there still are other problems: When running
Harald's test program "ntest" (reading data on one host from /dev/zero and
writing it on another machine to /dev/null) other network connections
dramatically slow down. A ftp session running concurrently to ntest gets
only 3KB/sec. "ifconfig eth0" does not show any errors or overruns.
Also, v0.37 of the driver seems to hang machines with more than one
network device. I've got several reports of hanging machines with 0.37 of
the driver with a 3c905 card. It seems not to be related to the network
load but to the fact that there are other network devices (in some cases
the machine had a Boomerang card and only an ISDN card).
Going back to v0.28c solved this problem in every case I know of, but it
is very slow of course.
As v0.37 seems to be ok for a non-server setup (I'm using this driver in
two workstations for several weeks now without any problem) I would prefer
to have both drivers (v0.28c and v0.37) in 2.0.31. v0.28c should replace
the current 3c59x.c (0.25). It supports boomerang cards so people with a
3c905 could at least use their hardware. Additionally, the boomerang
driver v0.37 could be included as an experimental driver for those who are
brave.
If there are more things I could test, please let me know.
> thanks for any hint or help how this lockup in the driver
> could be solved (or for any pointer that it's useless to
> waste more time with these broken cards...)
"Me too" ;-)
> Harald
Hubert mantel@suse.de