2.4.22 multicast address problem, pcap workaround

[Dominic Robinson]

Please cc to d.robinson@xxxxxxxxxxxxxxxxx

My source machine sends multicast packets on all the following
addresses: 230.1.1.1, 230.1.1.2 through to 230.1.1.64

My target machine (on local network, or with cross-over cable)
joins all these multicast groups, and receives packets on all
of these addresses EXCEPT 230.1.1.18 and 230.1.1.32.  

However, if I run tcpdump at the same time as my program, I
find my program suddenly starts receiving on 230.1.1.[18,32]
If I kill tcpdump, my program stops receiving.  It's the
"pcap_open_live()" function in tcpdump which causes this, 
and as a workaround I just call it direct in my receiver.

Could this be a bug in the multicast kernel code?

NOTES: 
(1) 230.1.1.1 to 230.1.64.1 and 230.1.1.1 to 230.64.1.1
    has the same problem.
(2) 230.1.1.1 to 230.1.1.63 works
(3) 230.1.1.1 to 230.1.1.128 fails with slightly different
    behaviour
(4) It makes no difference if I run a single program 
    listening on all addresses, or 64 individual programs
    each listening on a single address.

Thanks and regards,
Dominic
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/