[vortex] The mysteries of on-chip MII.

[Bogdan Costescu]

On Wed, 20 Mar 2002, Bill Cattey wrote:

>   8K byte-wide RAM 5:3 Rx:Tx split, autoselect/MII interface.

That's not the good one. You should have here "10baseT", not "MII".
I suggested that you use 3c90xcfg.exe or the modified vortex-diag because 
it would have been easier to choose the right media. If you want EEPROM 
values, they should be:

0x0d: 0x0000
0x13: 0x0000	or 0x0100 to set "Autoselect"
0x15: 0x0002

If you did use the modified vortex-diag and it set these values, it means 
there might still be some bugs in computing the right values...

If the transceiver is activated and correctly detected, but the network is 
still slow, please try (in this order), after using "debug=3" module 
option to get some more messages from vortex_timer():

1. "mii-diag -r" (then print again status with mii-diag -v) and wait for 
up to 1 minute to see in the logs any message from vortex_timer() like 
"Media selection timer tick happened" and whatever follows.

2. The same as 1. above, but using "mii-diag -R".

and let me know if any of these restore normal network functionality (I'm 
also interested in the logs and mii-diag output).

Please note that the driver only checks once a minute for changes in link 
status, so checks for network functionality should be done only after the 
"Media selection tick..." message appears, when the driver should have 
programmed the card for the correct duplex setting based on the negotiated 
one.

-- 
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