[vortex] MII ident

Bogdan Costescu Bogdan.Costescu@IWR.Uni-Heidelberg.De
Thu, 6 Jul 2000 11:33:16 +0200 (CEST)


On Wed, 5 Jul 2000, Donald Becker wrote:

> As I said before, v99H wasn't designed for the newer 3Com chips and setup
> programs.  The driver should now not use the transceiver setting in the
> EEPROM directly.  Instead it's expected to know that there is no SYM
> transceiver connection on the new chip.  When it sees the "100baseTx"
> transceiver type it should set the transceiver type to
> "Autonegotiation/MII", and then configure the MII management registers to
> force the specified media type.

Can you tell us exactly where this is done in your driver? I'm not able to
find it!
But the mechanism is correct. In fact, I was thinking of doing in all
cases: if a card had NWAY/MII, we should find it and use it, even if the
EEPROM or options say "forced".

> I believe what was happening was that the transceiver was being disabled
> when the v99H driver set the chip to the invalid "100baseTx" transceiver
> configuration.  The MII transceiver scan could not detect the disabled
> transceiver.

IMHO, what happens is that vp->phys[0] (which holds the address of the
MII transceiver) is not set if the card has a forced media type. I can
only find code which sets vp->phys[0] protected by
	if (XCVR_NWAY or XVCR_MII)
so if the EEPROM says that it's forced to 100baseTx (XCVR_100baseTx) the
MII transceiver is not searched for (in vortex_probe1), thus
vp->phys[0]=0.

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