At 11:42 AM 2/29/00 , Manuel Bouyer wrote:
>On Tue, Feb 29, 2000 at 08:33:13AM -0500, Paul J. Lavoie wrote:
> > 
> > This problem seems to exist with other Tulip-based cards. What I've had to do 
> > is hard-code both ends to be half-duplex, you should see much better results. 
> > Auto-negotiation seems to always fail, and full-duplex never seems to work.
>
>That's my conclusions as well.
>I've yet to try this board (dlink 570) in a -current machine :)

I have made fixes to -current for the 21143 and fd with mii.  so you might
want to try it

-- 
Matt Thomas               Internet:   matt@3am-software.com
3am Software Foundry      WWW URL:    http://www.3am-software.com/bio/matt/
Cupertino, CA             Disclaimer: I avow all knowledge of this message