IGMP snooping: set mrouters_only flag for IPv4 traffic
commit8904f07a1b5683c4363b225cb376f2bd1435246f
authorFernando Luis Vázquez Cao <fernando@oss.ntt.co.jp>
Mon, 13 Jun 2011 15:04:43 +0000 (13 15:04 +0000)
committerAndi Kleen <ak@linux.intel.com>
Mon, 1 Aug 2011 20:55:04 +0000 (1 13:55 -0700)
tree859182719557c52f07748c6a4406f03fb5f0b64b
parentea626c8fccc5b119bf1d398d99e7ecad4bfc95e2
IGMP snooping: set mrouters_only flag for IPv4 traffic

[ upstream commit 62b2bcb49cca72f6d3f39f831127a6ab315a475d ]
 properly

Upon reception of a IGMP/IGMPv2 membership report the kernel sets the
mrouters_only flag in a skb that may be a clone of the original skb, which
means that sometimes the bridge loses track of membership report packets (cb
buffers are tied to a specific skb and not shared) and it ends up forwading
join requests to the bridge interface.

This can cause unexpected membership timeouts and intermitent/permanent loss
of connectivity as described in RFC 4541 [2.1.1. IGMP Forwarding Rules]:

    A snooping switch should forward IGMP Membership Reports only to
    those ports where multicast routers are attached.
    [...]
    Sending membership reports to other hosts can result, for IGMPv1
    and IGMPv2, in unintentionally preventing a host from joining a
    specific multicast group.

Signed-off-by: Fernando Luis Vazquez Cao <fernando@oss.ntt.co.jp>
Tested-by: Hayato Kakuta <kakuta.hayato@oss.ntt.co.jp>
Signed-off-by: David S. Miller <davem@conan.davemloft.net>
Signed-off-by: Andi Kleen <ak@linux.intel.com>
net/bridge/br_multicast.c