Re: Problem with Tulip and 3Com

On Mon, 1 Mar 2004, Robert L. Harris wrote:

>   I'm hoping someone might have an answer on this one as I haven't
> found any fixes yet other than some threads from back in 2000.

Many problems are fixed, but the person doesn't follow-up confirming
it.

> I'm running a number of servers with DLINK Quad port network cards which use
> the tulip driver as well as a 3c905C.  Since kernel 2.4.21 I've had a problem
> with what appears to be eth2.

What driver version are you using?

> NETDEV WATCHDOG: eth2: transmit timed out

It doesn't appear to be my driver release...


> root@bridge:~# grep eth /var/log/dmesg
> eth0: Digital DS21143 Tulip rev 65 at 0x2000, 00:80:C8:CD:38:95, IRQ 17.
> eth1: Digital DS21143 Tulip rev 65 at 0x2080, 00:80:C8:CD:38:96, IRQ 18.
> eth2: Digital DS21143 Tulip rev 65 at 0x2400, 00:80:C8:CD:38:97, IRQ 19.
> eth3: Digital DS21143 Tulip rev 65 at 0x2480, 00:80:C8:CD:38:98, IRQ 16.

What does
  cat /proc/interrupts
report after the interfaces have been running for a while?

> We found the tulip diag program which gives some errors on Index 3:
>   The Rx process state is 'Suspended -- no Rx buffers'.
>  Interrupt sources are pending!  CSR5 is f06981c7.

Yup, sounds like an interrupt problem.

My guess is that all interrupts are arriving on for first interface.  My
driver explicitly works around this problem on the x86 by overriding the
IRQ setting.

-- 
Donald Becker				[email protected]
Scyld Computing Corporation		http://www.scyld.com
914 Bay Ridge Road, Suite 220		Scyld Beowulf cluster systems
Annapolis MD 21403			410-990-9993

_______________________________________________
tulip mailing list, [email protected]
To change to digest mode or unsubscribe visit
http://www.scyld.com/mailman/listinfo/tulip