[vortex] Problem with 3com 10/100 Mini PCI adapter (3c556B?)under 2.2.19

Bogdan Costescu bogdan.costescu@iwr.uni-heidelberg.de
Fri, 6 Jul 2001 11:23:19 +0200 (CEST)


On Fri, 6 Jul 2001, Andrew Morton wrote:

> Yes, it does appear that MII #0 is the one for this NIC.  The
> 2.4 driver looks for an MII at #24 first.  There isn't one
> there.

OK, that was just a blind supposition.

I've seen in the logs and in the vortex-diag output that MII was selected
(instead of Autonegotiate). If you didn't specifically use "options=6"
when loading the module, then MII is set in the EEPROM. Can you use the
DOS tool from 3Com to set it to NWAY/Autonegotiate ?
Reason: the documentation says somewhere that the transceiver selected in
EEPROM will be automatically activated at boot time. If this activation
also include some kind of power management or whatever for this chip, then
it might be that the transceiver is not fully activated. By setting it
properly also in EEPROM, this problem might be solved.
This is just another supposition...

.. and another one. Transceiver 0 is the next one checked after
transceiver 24, so if 24 is not the right one, 0 is picked up. What if 0
responds as in the CX case and the real transceiver is somewhere else ?
Can you try playing with "mii-diag -p x" with p between 0 and 31 and
report everything that is not "0xffff" as in the case of 24 ?

Sincerely,

Bogdan Costescu

IWR - Interdisziplinaeres Zentrum fuer Wissenschaftliches Rechnen
Universitaet Heidelberg, INF 368, D-69120 Heidelberg, GERMANY
Telephone: +49 6221 54 8869, Telefax: +49 6221 54 8868
E-mail: Bogdan.Costescu@IWR.Uni-Heidelberg.De