Lots of rx overrun errors

[Donald Becker]

On Wed, 5 Apr 2000, Bogdan Costescu wrote:
> On Wed, 5 Apr 2000, Beger Tomasz wrote:
> > I'm using 3c509C card with driver 0.99L on Red Hat Linux 6.1 and under a

You mean "3c905C", right?  The 3c509 is the ISA card series.

> > heavy load, some 23kpps received, and after some hard disk activity
> > (probably cache flushing) the card stops receiving packets and starts
> > incrementing RX overrun errors. From that moment it seems all received
> > packets are counted up as an overrun error. Usually message "too much work
> > in the interrupt" appears on the console and in the /var/log/messages file.
...
> I've seen similar problem descriptions on tulip-bug list which were
> answered by suggesting to set the parameter "max_interrupt_work" to
> something bigger (the default in 3c59x 0.99L is 20); you can try setting
> it to 40, 100, 200 or more, but if becomes too big it might have some
> adverse efects. You can add this as a module option (in
> /etc/conf.modules), but if you're not using modules, might as well modify
> the driver source. This parameter is defined in the source as "Maximum
> events (Rx packets, etc.) to handle at each interrupt"...

Yes, this is the correct approach for the 99L version.
The 99M version should recover from this itself at the next timer tick.

Donald Becker
Scyld Computing Corporation, becker@scyld.com

-------------------------------------------------------------------
To unsubscribe send a message body containing "unsubscribe"
to linux-vortex-bug-request@beowulf.org