In message <E1F6dcf-0006RE-00@smeg.dsg.stanford.edu>, jonathan@dsg.stanford.edu
 writes:
>
>In message <Pine.NEB.4.64.0602080939100.12868@angelic.cynic.net>,
>Curt Sampson writes:
>
>>On Tue, 7 Feb 2006, Simon Burge wrote:
>>
>>> wm0 is on the motherboard, wm1 is a PCIX card.  wm0 will only work if I
>>> hard wire it to 10mbit/sec (with "media 10baseT" in the ifconfig line).
>>
>>I have a possibly related issue with the wm0 on my Thinkpad T42. With
>>the one 10base-T hub I've tried, it autodetects that it's 10base-T, but
>>appears not to receive (and possibly not send, either) packets unless I
>>manually set the media type to 10base-T.
>
>I wonder how many of the developers hacking on gigabi-ethernet drivers
>still have 10baseT networks to test?  I know I don't.  (OK, maybe I
>have a transciever hidden in a pile of Decstations somewhere, but I
>don't have any acutal 10baseT ports to connect it to.)
>
>This may be an overly-naive approach, but since SMB says NetBSD-2.0
>handled this better for him: has anyone tried the 2.0 makphy code with
>the 3.0 wm driver?  Or vice-versa?
>

ENOTIME.

I do have some 10BaseT hubs lying around; I also have some oddball 
10BaseT gear that I need to talk to periodically.  More important,
perhaps, my daughter's dorm room has 10BaseT only; given the primary 
use for high-bandwidth connections in residence halls, I doubt they're 
going to upgrade any time soon.  They have enough trouble with traffic 
shaping as is....

		--Steven M. Bellovin, http://www.cs.columbia.edu/~smb