Bug report, as much information as I can give...

Winged Wolf winged@spastic.org
Mon Dec 27 16:43:53 1999


On Mon, 27 Dec 1999, Donald Becker wrote:

> On Mon, 27 Dec 1999, Winged Wolf wrote:
> 
> > With the L version of the driver, I get the same output from mii-diag, but
> > when I try to do the ping, I get:
> > 
> > eth0: Host error, FIFO diagnostic register 0000.
> 
> Hmmmm, that indicates a problem, usually a hardware problem.
> 
> > I get losts of these, then:
> > 
> > eth0: Too much work in interrupt, status 8003.  Temporarily disabling
> > functions (7ffc).
> 
> This is bad.  It's reporting that the machine is too slow to process the
> packets.  Do you have some sort of APM feature turned on?

I have no APM turned on, completely disabled as a matter of fact.

> > Then:
> > eth0: transmit timed out, tx_status 00 status e000.
> 
> This is because the card has most features disabled at this point.
> 
> You can work around the symptoms by setting the following module parameter:
> /* Maximum events (Rx packets, etc.) to handle at each interrupt. */
> static int max_interrupt_work = 20;
> 
> options 3c59x max_interrupt_work=200

I'm not using modprobe, rather installing it manually.  I just turned on
debug=6; I'm getting some interesting results in my log:

eth0: transmit timed out, tx_status 00 status 8000.
  Flags; bust-master 1, full 1; dirty 0 current 16.
  Transmit list 4327aa10 vs. c327aa00.
  0: @c327aa00  length 8000002a status 8000002a
  1: @c327aa10  length 8000002a status 8000002a
  2: @c327aa20  length 8000002a status 8000002a
...

and so on, with the @ numbers increasing by 0x00000010 from 0 to 15
(@c327aaf0).

> 
> Donald Becker
> Scyld Computing Corporation, and
> USRA-CESDIS,   becker@cesdis.gsfc.nasa.gov
>