[tulip] Conexant LANfinity

[Donald Becker]

On Fri, 29 Dec 2000, Eberhard Moenkeberg wrote:

> I have a notebook Compaq Presario 1700 which has a builtin
> miniPCI modem/LAN interface.
> 
> The LAN chip has PCI ID 14F1:1803. lspci is telling:
> 00:09.0 Ethernet controller: CONEXANT: Unknown device 1803 (rev 08)
> 
> I grepped for these strings through my mirror of ftp.scyld.com,
> and sic! Donald covers that chip in network/test/tulip.c
> (but not in tulip-diag yet).

I haven't released the updated tulip-diag.c -- I made a bunch of changes,
not all of which are tested.

> So I put this 0.92p tulip driver into my 2.2.18 kernel source
...
> and compiled a kernel 2.2.18 with built-in tulip 0.92p support.
...
> eth0: Conexant LANfinity rev 8 at 0x1400, EEPROM not present,
>    00:4C:69:6E:75:79, IRQ 9.
> eth0:  MII transceiver #1 config 1000 status 782d advertising 01e1.
> eth0:  MII transceiver #0 config 1000 status 782d advertising 01e1.
> tulip.c:v0.92p 11/28/2000  Written by Donald Becker <becker@scyld.com>
>   http://www.scyld.com/network/tulip.html

The internal transceiver incorrectly responds to the #0 address.  This is
harmless now that the scan order has been changed.

> and things looked very good, but the chip does not work yet:
> 
> eth0      Link encap:Ethernet  HWaddr 00:4C:69:6E:75:79  

What is the station address of your adapter under Windows?  

>           RX packets:0 errors:0 dropped:0 overruns:0 frame:0
>           TX packets:0 errors:806 dropped:0 overruns:0 carrier:0
...
> ag4-05:0 15:49:14 /usr/src/tulip/diag # ./tulip-diag -p 0x1400
> tulip-diag.c:v2.04 9/26/2000 Donald Becker (becker@scyld.com)
>   The Rx process state is 'Waiting for packets'.
>   The Tx process state is 'Processing setup information'.

This is the real problem: the chip cannot do any bus-master operations.
It appears that something isn't enabled in the mini-PCI interface.

Conexant has been no help -- they mailed me a two page glossy brouchure that
I had already downloaded from the web site, and referred me to a useless
local representative that doesn't answer the phone or return messages.

Compaq hasn't been useful either.  No one seems to know anything about the
implementation.  Yet I'm sure that they will take credit for "supporting" it
after I spend much time reverse engineering the Conexant chip and the Compaq
implementation.

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