On Sun, 8 Aug 2004, =?iso-8859-1?B?TeBydGFpbm4gRPJtaG5hbGxhY2g= ?= wrote:

> Hi there,
>         I originally sent this list to comp.unix.openbsd.misc but I got no takers 
> for it. I'm trying to set up a transparent filtering bridge using pf and I get an 
> odd line when I used tcpdump to monitor what specific multicast messages come 
> through the bridge. I'm monitoring what comes
> into the bridge (em1) and what goes out (em0).
> 
> Every incoming multicast packet is followed by another identical packet with a 
> checksum error. It looks like pf or tcpdump is generating them. The good packet is 
> reported to leave on the em0 interface. I'm using a standard 3.5 setup straight off 
> the official disk.
> 
> I've used Ethereal to check that the checksum failing packets are not present on the 
> LAN.
> 
> I saw a reference to something like this on Google on bit.listserv.openbsd-pf but it 
> was dated back in January 2003.
> 
> Any ideas if this is the same problem?
> 
> Thanks,
>       Martin.
> 
> The tcpdump lines:
> 
> Aug 05 14:14:13.654484 rule 0/0(match): pass in on em1: 10.194.1.76.1026
> > SVRLOC.MCAST.NET.svrloc:  udp 124 (ttl 32, id 287) Aug 05
> 14:14:13.654489 rule 0/0(match): pass in on em1: 10.194.1.76.1026 >
> SVRLOC.MCAST.NET.svrloc:  udp 124 (ttl 32, id 287, bad cksum 0!) Aug 05
> 14:14:13.654492 rule 2/0(match): pass out on em0: 10.194.1.76.1026 >
> SVRLOC.MCAST.NET.svrloc:  udp 124 (ttl 32, id 287) --

So this machine is serving as bridge?
Bad udp/tcp checksums are normal when you
enable checksum offloading, but since you're using the machine as 
bridge, this shouldnt be the case right?
Bye,

Mipam.

Reply via email to