3c509B fun and games at 100M on NT

[Jeff Barrow]

>presume that different parts of the 905 driver assume the card is going at
>different speeds. It's almost as if the card always performs a correct
>Auto-Detect, but the upper layers assume duplex is turned on when one tries
>to statically
>configure the speed. This would explain the situation we found of dropped
>packets - the presumption is that one of the receiver's ACK's collided
>outbound with
>the sender's next DATA packet, but neither end detected the collision
>because the 3Com card had it's collision detection circuitry disabled
>because it thought
>it was in duplex mode.


>From my albeit limited experience with the 3c905B's and the Cyclone drivers
under Linux, I'd have to say thtat the same oddity applies: the card detects
the speed correctly but doesn't tell the upper layers of the driver the
correct speed/duplex setting.  It even messes up when the card autodetects
that it is capable of full duplex but doesn't tell the upper layers the
correct duplex setting: no packets get received from the network in this
case (I don't know why...).
A work-around I've found for getting 100M/Full Duplex working under Linux
was to tell the driver to use an EXTERNAL MIB to detect speed/duplex... but
this only works if the other end is capable of full duplex :)  (this was
with the 0.99E driver)

--Jeff Barrow, Internet Connections, Inc.