At 19:39 -0800 3/9/00, Jason R Thorpe wrote:

>  > While I've got your attention, there is one other thing that it seems
>  > both the tlp* and de* drivers don't quite get right with the DE500:
>  > Nway negotiation with a 10/100 FDX switch (in this case a NetGear
>  > FS524). SRM also produces some pretty funky results.
>
>This is because the switch is broken (doesn't do NWAY properly).  There's
>not much that can be done about it.  I've already discussed this on
>tech-net.

I searched the last year's worth of tech-net and while there were 
interesting discussions of how to get Nway right in our MII code (and 
complaints about 802.3u having been written by clueless hardware 
wonks with no input from software people), nowhere did you or anyone 
else assert that NetGear's switches were broken.

Besides, I get the bad result from the DE500-BA with three different 
10/100 switches: NetGear's FS108 and FS524, Dlink's DSS-8, and a 
Cornet FS8A. They're *all* broken?

According to DEC's documentation, the DE500-XA (21140) can only 
detect; but the DE500-AA (21140A) and DE500-BA (21143) can do full 
blown Nway. The Znyx ZX345 and AsanteFast cards I have with 21140A's 
in 'em seem to get it right with the switches with no trouble at all. 
The 4-port thing with the unknown PHYs and "Intel" (pass 4.0) 21143's 
get it right, too. My late-model Macintoshes with 10/100 built-in (a 
rev C iMac, and a blue&white G3/400) also do the Nway dance correctly 
with no fuss, no matter which switch I plug them into.

This feels like a driver problem to me.

I'm not trying to fault anyone - I'm trying to get complete 
information to understand this situation, and get the bugs corrected 
if they can be. It seems to me that I've also got more than enough 
gear to do an NxM test...

	Erik <fair@clock.org>