From owner-freebsd-net Thu Aug 16 13:33:46 2001 Delivered-To: freebsd-net@freebsd.org Received: from quarter.csl.sri.com (quarter.csl.sri.com [130.107.1.30]) by hub.freebsd.org (Postfix) with ESMTP id 6282E37B40D; Thu, 16 Aug 2001 13:33:42 -0700 (PDT) (envelope-from hogsett@csl.sri.com) Received: from glob.csl.sri.com (glob.csl.sri.com [130.107.15.161]) by quarter.csl.sri.com (8.11.1/8.11.1) with ESMTP id f7GKXgP29709; Thu, 16 Aug 2001 13:33:42 -0700 Received: from glob.csl.sri.com (hogsett@localhost) by glob.csl.sri.com (8.11.2/8.8.7) with ESMTP id f7GKXfA10365; Thu, 16 Aug 2001 13:33:41 -0700 Message-Id: <200108162033.f7GKXfA10365@glob.csl.sri.com> X-Authentication-Warning: glob.csl.sri.com: hogsett owned process doing -bs To: freebsd-net@freebsd.org Cc: Mike Smith , hogsett@csl.sri.com Subject: Question : vlan -> fec -> fxp Date: Thu, 16 Aug 2001 13:33:01 -0700 From: Mike Hogsett Sender: owner-freebsd-net@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.org Previously I had been posting to freebsd-stable regarding my questions, but this seems to be a better list for this. First let me explain my goal, then I will explain the problem. My goal is to use two fxp's "bonded" together using the fast ether channel netgraph module ( from http://people.freebsd.org/~wpaul/FEC/ ) then create n vlan interfaces with the fec0 interface as the parent dev for the vlan interfaces. fxp0 fxp1 \________/ | fec0 / | \ / | \ / | \ vlan0 vlan1 vlanN I have built and installed a new kernel with "pseudo-device vlan 1" and have built and installed the fec module. I have confirmed that fxp{0,1} interfaces work stand alone. I have confirmed that the fec0 interface works (using "set_mode_inet"). I have confirmed that the vlan interfaces work when the parent device is either fxp0 or fxp1. What does not work is when the fec0 interface is the parent for the vlan interfaces. Here is the setup : ngctl mkpeer fec dummy fec ngctl msg fec0: add_iface '"fxp0"' ngctl msg fec0: add_iface '"fxp1"' ngctl msg fec0: set_mode_mac #(I have also tried using "set_mode_inet") ifconfig fxp0 up ifconfig fxp1 up ifconfig fec0 up ifconfig vlan0 192.168.1.10 netmask 255.255.255.0 vlan 505 vlandev fec0 up mtu 1500 Nothing complains. The interfaces all look good in ifconfig, mtu's are correct, all are "UP". When pinging this machine from another machine on the lan here is the behavior that I get : Running "tcpdump -ni vlan0" shows the arp requests and the arp replies Running "tcpdump -ni fec0" shows the 802.1q tagged arp reqests and the arp replies Running "tcpdump -ni fxp{0,1}" shows the 802.1q tagged arp requests So the packets climb the chain but seem to enter a blackhole back down... Suggestions? - Michael Hogsett To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message