When a socket joins a multicast group, the structure that is passed to the Linux kernel includes a network interface. Is this interface only used by the kernel to decide which interface to use to issue the IGMP join, or is it also used by the kernel when a packet arrives to decide whether or not a socket gets a copy of the packet? For example: socket s0 joins a multicast group on interface eth0. socket s1 joins the same multicast group on interface eth1. Someone publishes a packet to the multicast group. The driver for eth0 delivers the packet to the kernel. Does the kernel deliver the packet to both socket s0 and s1, or does it consider the interface through which each socket joined the multicast group, and therefore only deliver the packet to socket s0? Similarly, the driver for eth1 delivers the packet to the kernel. Does the kernel deliver the packet to both socket s0 and s1, or does it consider the interface through which each socket joined the multicast group, and therefore only deliver the packet to socket s1? My experience indicates that in each case, the packet is delivered to both sockets, so each socket receives two copies of the packet. This is not what I was expecting. Thanks, Jon -- To unsubscribe from this list: send the line "unsubscribe linux-net" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
