Subject: another weird tulip clone chip
To: None <tech-net@netbsd.org>
From: Matthias Drochner <M.Drochner@fz-juelich.de>
List: tech-net
Date: 09/20/2002 19:05:27
Hi -
is there any hope for this one?

tlp0 at pci2 dev 5 function 0: ADMtek AN985 Ethernet, pass 1.1
tlp0: broken MicroWire interface detected; setting SROM size to 1Kb
tlp0: interrupting at irq 10
tlp0: reset failed to complete
tlp0: Ethernet address ff:ff:ff:ff:ff:ff

The PCI ID matches
product ACCTON EN2242	0x1216	EN2242 10/100 Ethernet

There is a subsystem ID of SMC/0xc255.
The chip itself is labeled "SMC EN5251B".

Setting "tlp_srom_debug" shows that only "0xff"s are read from
the EEPROM.
However, there _is_ an EEPROM on the board (93LC46).

The board is a new SMC one  with a 100TX and a fibre connector,
its name is "SMC 1255FTX-SC".
There is nothing which looks like an external MII PHY chip.

The Linux "tulip" driver recognizes it as:
ADMtek Comet rev 17
and gets the ethernet address right.
100TX seems to work; connecting a fibre hung the box.

A brief look at the Linux driver made me believe that the ethernet address
is read from io registers at offset 0xa4/0xa8, but I couldn't reproduce this
on NetBSD - perhaps an initialization matter, or a misunderstanding of the
Linux driver.

best regards
Matthias