[eepro100] eth0: card reports no resources

John Jones johnj@ActiveState.com
Tue Sep 17 22:05:01 2002

Donald, thanks for the response.

> you didn't specify which driver version
> you are using. 2000/5/31
I know these are the old ones.

> What is the incoming packet rate?

Unsure, since the console becomes extremely lagged during this time.  We
unplug the eth0 cable to the outside world, and it immediately 'releases'
and becomes responsive again.  During that time, we are unable to run
commands at any 'respectable' information rate.

> Have you increased the VM parameters as previously suggested on this list?

I am doing that as soon as I understand what I am changing, and how to do it
properly (in /proc/sys/vm/bdflush), yes.

We are trying to narrow down where this is happening. I have recently
enabled /proc/sys/net/ipv4/tpc_syncookies as that was at '0'.

One thing I -am- seeing is a ton of frame errors on the nic:

          RX packets:2616981 errors:0 dropped:0 overruns:0 frame:46855
          TX packets:3856006 errors:0 dropped:0 overruns:0 carrier:0

I know that these are from link-state errors with the ISP, and since this is
a 100MB/Full Duplex connection, (and the readings related the nway detection
failures of the drivers) I was initially wondering if this was what I was