We are trying to run the "l3fwdr" project on the IXP1200 hardware and are facing a weird problem. The setup is: OS: Suse linux 8.0 (2.4.18) The PCI driver works correctly (ping between the host & target works) Port 0 of the IXP board is connected to another linux machine via a cross-over cable. After the NFS mount, when we try the "ixstart", tcpdump on the above mentioned linux machine's interface starts showing an "insane" number of received packets. Here is a snapshot. 04:22:38.693713 0:0:0:0:0:0 1:80:c2:0:0:1 8808 60: 0001 0f01 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 04:22:38.694229 0:0:0:0:0:0 1:80:c2:0:0:1 8808 60: 0001 0f01 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 04:22:38.694580 0:0:0:0:0:0 1:80:c2:0:0:1 8808 60: 0001 0f01 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 04:22:38.694910 0:0:0:0:0:0 1:80:c2:0:0:1 8808 60: 0001 0f01 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 04:22:38.695319 0:0:0:0:0:0 1:80:c2:0:0:1 8808 60: 0001 0f01 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 The moment we type "ixstop", this packet flow stops. We are using the "ixsys.config-l3fwdr" with 4 interfaces configured. The ping between this second linux machine and IXP board (via the cross cable) doesnt work inspite of all the route/arp entries being correct although at least some of packets seem to be getting through (the RX/TX counts of the interfaces do change) We had a similar setup working last year (with an older version of the driver and ofcourse an older linux kernel). The version we are trying to use now is the one distributed at the IXA event this year in May. Any ideas as to the cause of this problem ???? Thanks in advance, Rgds, -Nimit