Date: Mon, 11 Aug 1997 17:35:40 -0700 (PDT) From: Tom Samplonius <tom@sdf.com> To: David Greenman <dg@root.com> Cc: hackers@freebsd.org Subject: Re: fxp driver full duplex packet loss problem Message-ID: <Pine.BSF.3.95q.970811173306.8572A-100000@misery.sdf.com> In-Reply-To: <199708120004.RAA19545@implode.root.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Mon, 11 Aug 1997, David Greenman wrote: > > I have a Intel Etherexpress Pro100/B conneted into a 10BT port on a > >lightly loaded Catlyst 1900 etherswitch. The system is running 2.2-stable > >(last week), and traffic ranges from about 50 to 150KB/s output (as > >measured with netstat -I fxp0 -w 1) > > > > When I use "link0 link2" to enable full-duplex operation, I start seeing > >about 1 to 4% packet loss (measured with ping to and from the server). > >When I disable full-duplex, the packet loss disappears. > > > > Is it safe to change the full-duplex vs. half-duplex on-the-fly? Should > >I power-cycle/reboot between changes? > > > > Anyone else running this card in this kind of configuration? > > At least some Cisco fast ethernet interfaces (those on the 7000/7500 series > routers, for instance), don't support auto-negotiation. So when you force full > duplex on your end, the Cisco is still in half duplex. The behavior you're > describing is indicative of this. I'm only running 10BT ethernet, not fast. That isn't a problem is it? There seems to be a lot of beating around the bush in the Cisco docs about full-duplex on 10BT ports. The Catalyst 1900 switch is supposed to auto-negotiate full-duplex. It > To fix this, you'll need to set full duplex in the Cisco as well. > > -DG > > David Greenman > Core-team/Principal Architect, The FreeBSD Project > > Tom
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.BSF.3.95q.970811173306.8572A-100000>