[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <OF57FAD1F9.00750CDD-ON8825731A.006E33C6-8825731A.006F814B@us.ibm.com>
Date: Mon, 16 Jul 2007 13:18:39 -0700
From: David Stevens <dlstevens@...ibm.com>
To: Steffen Maier <smaier@...rs.sourceforge.net>
Cc: linux-kernel@...r.kernel.org, netdev@...r.kernel.org,
netdev-owner@...r.kernel.org,
Steffen Maier <smaier@...rs.sourceforge.net>
Subject: Re: kernel module to list sockets with their multicast group joins and
respective processes
> sockets that join different groups receive messages from the respective
> other group (if they are only bound to the wildcard address). Obviously
> this is handled differently in Linux for IPv4, where the socket matching
> for incoming message is done solely on the 4-tuple of addresses and
ports,
> and IPv6, where the explicit socket joins are taken into account [1,2].
These are not different in v4 vs v6. Group membership is
per-interface,
delivery depends on the binding (for both v4 and v6), and this is the same
as BSD behavior since IP multicasting was invented.
There are two levels of checks, and I think you're confusing them.
inet6_mc_check() is checking that somebody (*anybody*) has joined the
group on the receiving interface, before delivering the packet to IP.
Delivery to a particular socket depends on source filter settings, if
they are in use, and bindings (only). The socket has a list of joined
addresses to track source filters and to check for correct join/leave
(a socket can't leave a group it hasn't joined), but that list does not
determine delivery of a multicast packet to the socket when source
filtering
is not in use.
+-DLS
-
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Powered by blists - more mailing lists