• mrouter (router port)
• subscriber (port belongs to established multicast group)
• none

• multicast traffic received on “none” port
  • if no mrouter present, flood to all ports
  • otherwise flood to mrouter ports
• igmp received on “none” port
  • if no mrouter present, ignore frame (no group is created)
  • otherwise
    • creates multicast group
    • forwards to mrouter ports only
    • port becomes “subscriber”
• multicast traffic received on “mrouter” port
  • send to established group members only
• igmp received on “mrouter” port
  • creates multicast group
  • forwards to mrouter ports only

• if igmp snooping is disabled, multicast traffic is flooded to all ports in vlan, regardless of frame nature (i.e. it is igmp or not); static multicast fdb entries are ignored;
• mrouters can be statically configured while igmp snooping is disabled on vlan, but they are ignored until igmp snooping is enabled; when igmp snooping is enabled, the rules listed above apply;

Analyzing the forwarding cases, it turns out that when incoming packets are forwarded to more than one port at the same time, they are forwarded to either all ports in a specific igmp group or all mrouter ports (but not a mix of the two). In other words, depending on the igmp snooping logic, either an igmp group or a list of mrouter ports is sufficient to determine which ports to forward to.

• igmp group membership is stored as special fdb entry
• is_static field in struct net_switch_fdb_entry is change to a bitwise field:
  • b0 tells if entry is dynamic or static
  • b1 tells if entry is regular fdb entry or igmp membership entry
  • for igmp membership entries, a b0 value corresponding to “static” means user-configured group membership
• all igmp membership entries have a virtual MAC address formed as 01-00 plus the group address (4 bytes)
• when flooding to all ports in a group, the following actions are taken:
  • first the virtual mac corresponding to the packet destination is build;
  • the fdb bucket corresponding to the virtual mac is determined;
  • the fdb bucket is walked to determine the ports to forward to:
    • virtual mac must be equal (between fdb entry and packet)
    • vlan must be equal (between fdb entry and packet)

The following changes need to be made to existing code:

• fdb queries from userspace need to filter out igmp membership entries;
• standard (non-multicast) packet fdb lookup needs to ignore igmp membership entries;

exec#

sh ip igmp snooping [vlan x]
sh ip igmp snooping groups [vlan x] [dynamic|user] [count] etc
sh ip igmp snooping mrouter [vlan x]

config#

[no] ip igmp snooping vlan x mrouter interface y
[no] ip igmp snooping vlan x static a.b.c.d interface y

• RFC 4541 - Considerations for IGMP and MLD Snooping Switches
• Configuring IGMP Snooping and MVR (Catalyst 3560)
• Mapping IP Multicast to MAC-Layer Multicast
• Some experiments with IGMP snooping behavior of a couple of switches
• Changing IGMP version on Debian Linux Systems