Thor Lancelot Simon <tls@rek.tjls.com>  wrote:
 > The old advice to "solve" duplex mismatch problems by hardwiring one or
 > both ends may have been good advice at the time, but it's _really_ bad
 > advice now, IME.  Autonegotiation works -- use it.

ok. interesting.

So I set the vr0 and fxp0 machines to 'auto', then ran ttcp between the
two hosts:

	# ttcp -r -s
	ttcp-r: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp
	ttcp-r: socket
	ttcp-r: accept from .137
	ttcp-r: 16777216 bytes in 1.75 real seconds = 9365.32 KB/sec +++
	ttcp-r: 11578 I/O calls, msec/call = 0.15, calls/sec = 6618.14
	ttcp-r: 0.0user 0.0sys 0:01real 2% 0i+0d 0maxrss 0+2pf 11573+0csw

	# ttcp -s -t mail
	ttcp-t: buflen=8192, nbuf=2048, align=16384/0, port=5001  tcp  -> mail
	ttcp-t: socket
	ttcp-t: connect
	ttcp-t: 16777216 bytes in 1.55 real seconds = 10570.01 KB/sec +++
	ttcp-t: 2048 I/O calls, msec/call = 0.78, calls/sec = 1321.25
	ttcp-t: -1.9user 0.0sys 0:01real 1% 0i+0d 0maxrss 0+2579pf 7530+0csw

So, you're right.  What makes me curious is why neither end (the switch,
or the NetBSD hosts) were showing any interface errors, collisions, or
framing errors, or what have you.

Anyway, thanks.