[tulip-bug] tulip not correctly identifying LNE100TX

Donald Becker becker@scyld.com
Tue Jul 9 11:40:01 2002

On Tue, 9 Jul 2002, Daniel Styers wrote:

> I have a 3Com 3c900 and a Linksys LNE100TX card, using kernel 2.2.21.
> In /etc/lilo.conf, I have the following...
> append="ether=0,0,0,eth1"

This shouldn't really be required.

> During the kernel load, I get the following...
> eth0: 3Com 3c900 Cyclone 10Mbps TPO at 0xfc80,  00:50:04:72:36:ce, IRQ 11
>   8K buffer 5:3 Rx:Tx split, autoselect/Autonegotiate interface.
>   MII transceiver found at address 24, status 182d.
>   Using bus-master transmits and whole-frame receives.
> 3c59x.c:v0.99X 6/21/2002 Donald Becker, becker@scyld.com
>   http://www.scyld.com/network/vortex.html

Look Fine.

> eth1: ADMtek Centaur-P rev 17 at 0xf800, 00:04:5A:6F:50:23, IRQ 9.
> eth1:  MII transceiver #1 config 1100 status 7869 advertising 05e1.
> eth1:  MII transceiver #2 config 1100 status 786d advertising 05e1.

Hmmm, that's curious.  I'll check if I have a rev. 17 chip to see if the
transceiver is misreported.

> I am also getting the following in /var/log/syslog, repeatedly, which also
> show up when the startup scripts are attempting to configure the IP for this
> card.
> kernel: eth1: PCI Fatal Bus Error, fc36e010.
> kernel: eth1: PCI Fatal Bus Error, fc67e014.
> kernel: eth1: PCI Fatal Bus Error, fc67e050.
> kernel: eth1: PCI Fatal Bus Error, fc27e014.

Yup, looks like a fatal bus error.
It appears to be a parity error.

Try moving the card to a different slot, or gently cleaning the contacts.

> Is there any way to tell the driver, that it is indeed an LNE100TX and not an
> ADMtek Centaur-P?

Uhhmmm, it's a ADMtek Centaur-P chip.  There is likely a PCI subsystem
ID programmed into the EEPROM that identifies the board as having a
Linksys label, but that has no impact on operation.

Send me the following
   The exact name and number on the NIC box or printing (including e.g. "v5.1")
   The entry from 'cat /proc/pci' or 'lspci -n -v'.
   The output of 'tulip-diag -eee'.

>  I'm hoping this will fix the bus errors.

Nope.  The bus errors are reporting a hardware problem.  You might be
able to catch and show the error state with 'tulip-diag -a'.

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