[tulip] Errors using 0.92wax tulip driver on Conexant

[Donald Becker]

On Wed, 8 Aug 2001, Rene Klootwijk wrote:

> When I use the 0.92wax release of the drivers on a Conexant LanFinity NIC I 
> get the following error message after a while:
> 
> kernel: eth0: PCI Fatal Bus Error, f46ea000.

The chip is reporting a fatal parity error.  You have a hardware problem.

The behavior of the chip when this occurs isn't documented.
What do you observe?
What does 'tulip-diag -af' report?

> Also when downloading several files concurrently I get the following error 
> messages repeatedly:
> kernel: eth0: Too much work during an interrupt, csr5=
> kernel: eth0: Restarted Rx at 11565 / 11565.

What was the csr5 status value when this occured?

You can avoid the message by setting the module parameter
     max_interrupt_work=25
to a higher value.  But this just masks the problem, which is likely
that some other device driver is masking interrupts for too long.

Donald Becker				becker@scyld.com
Scyld Computing Corporation		http://www.scyld.com
410 Severn Ave. Suite 210		Second Generation Beowulf Clusters
Annapolis MD 21403			410-990-9993