[tulip] Compaq Presario laptop and tulip PCMCIA card

[Donald Becker]

On 8 Mar 2002, Jesse F. Hughes wrote:

> > Very important: what is the detection message?  I want to confirm that
> > the new device ID is really a Centaur chip.
> 
> Here's what I see in /var/log/messages (using my edited version of
> tulip.c -- I haven't tried the new version yet).
>
> Mar  8 17:34:32 diogenes cardmgr[82]: socket 0: Linksys EtherFast PCM200 v2
> Mar  8 17:34:32 diogenes kernel: cs: cb_alloc(bus 32): vendor 0x13d1, device 0xab08
> Mar  8 17:34:32 diogenes kernel: tulip.c:v0.93 11/7/2001  Written by Donald Becker <becker@scyld.com>
...
> Mar  8 17:34:32 diogenes kernel: eth0: ADMtek Centaur-C (Linksys) rev 17 at 0xc7902000, 00:E0:98:9A:E2:05, IRQ 9.

OK, the driver does seem to be working OK.

The reasonable station address hints pretty strongly that it really is a
ADMtek chip, but we haven't confirmed that it is definitely a ADMtek
Centaur-P.
(The chip type message above says "Centaur" only because we added that text
to the table.)

> Mar  8 17:34:32 diogenes kernel: eth0:  MII transceiver #1 config 3100 status 7849 advertising 05e1.
> Mar  8 17:34:32 diogenes kernel: eth0:  MII transceiver #2 config 1100 status 7849 advertising 05e1.
> Mar  8 17:34:32 diogenes kernel: eth0:  MII transceiver #3 config 1100 status 7849 advertising 05e1.
> Mar  8 17:34:32 diogenes kernel: eth0:  MII transceiver #4 config 1100 status 7849 advertising 05e1.

Hmmm, curious: it's a common MII hardware buglet to report the status on
multiple MII addresses(*), but it's unusual to report different status
values for the same transceiver.  One of the ADMtek chips does have
multiple transceivers, with the second being for HomePNA, but the default
advertised capabilities for HomePNA is much different than 05e1.

(*) I suspect the hardware designer do this deliberately so that bogus
drivers that don't scan for transceivers and just use address '1' will
work.

> Thanks again for the help.  I'm thrilled to have the card running.

Send the output of the new 'tulip-diag -ee' when you get a chance, just
so that we can confirm the EEPROM contents look reasonable for a Centaur
chip.
  http://www.scyld.com/diag/index.html
  ftp://www.scyld.com/pub/diag/tulip-diag.c

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