net: bridge: mcast: add support for raw L2 multicast groups

Extend the bridge multicast control and data path to configure routes
for L2 (non-IP) multicast groups.

The uapi struct br_mdb_entry union u is extended with another variant,
mac_addr, which does not change the structure size, and which is valid
when the proto field is zero.

To be compatible with the forwarding code that is already in place,
which acts as an IGMP/MLD snooping bridge with querier capabilities, we
need to declare that for L2 MDB entries (for which there exists no such
thing as IGMP/MLD snooping/querying), that there is always a querier.
Otherwise, these entries would be flooded to all bridge ports and not
just to those that are members of the L2 multicast group.

Needless to say, only permanent L2 multicast groups can be installed on
a bridge port.

Signed-off-by: Nikolay Aleksandrov <[e-mail]>
Signed-off-by: Vladimir Oltean <[e-mail]>
Link: https://lore.kernel.org/r/[e-mail]
Signed-off-by: Jakub Kicinski <[e-mail]>

include/linux/if_bridge.h

···
 #if IS_ENABLED(CONFIG_IPV6)
 		struct in6_addr ip6;
 #endif
+		unsigned char	mac_addr[ETH_ALEN];
 	} dst;
 	__be16		proto;
 	__u16           vid;

include/uapi/linux/if_bridge.h

···
 		union {
 			__be32	ip4;
 			struct in6_addr ip6;
+			unsigned char mac_addr[ETH_ALEN];
 		} u;
 		__be16		proto;
 	} addr;

net/bridge/br_device.c

···
 
 		mdst = br_mdb_get(br, skb, vid);
 		if ((mdst || BR_INPUT_SKB_CB_MROUTERS_ONLY(skb)) &&
-		    br_multicast_querier_exists(br, eth_hdr(skb)))
+		    br_multicast_querier_exists(br, eth_hdr(skb), mdst))
 			br_multicast_flood(mdst, skb, false, true);
 		else
 			br_flood(br, skb, BR_PKT_MULTICAST, false, true);

net/bridge/br_input.c

···
 	case BR_PKT_MULTICAST:
 		mdst = br_mdb_get(br, skb, vid);
 		if ((mdst || BR_INPUT_SKB_CB_MROUTERS_ONLY(skb)) &&
-		    br_multicast_querier_exists(br, eth_hdr(skb))) {
+		    br_multicast_querier_exists(br, eth_hdr(skb), mdst)) {
 			if ((mdst && mdst->host_joined) ||
 			    br_multicast_is_router(br)) {
 				local_rcv = true;

net/bridge/br_mdb.c

···
 			ip->src.ip6 = nla_get_in6_addr(mdb_attrs[MDBE_ATTR_SOURCE]);
 		break;
 #endif
+	default:
+		ether_addr_copy(ip->dst.mac_addr, entry->addr.u.mac_addr);
 	}
 
 }
···
 	if (mp->addr.proto == htons(ETH_P_IP))
 		e.addr.u.ip4 = mp->addr.dst.ip4;
 #if IS_ENABLED(CONFIG_IPV6)
-	if (mp->addr.proto == htons(ETH_P_IPV6))
+	else if (mp->addr.proto == htons(ETH_P_IPV6))
 		e.addr.u.ip6 = mp->addr.dst.ip6;
 #endif
+	else
+		ether_addr_copy(e.addr.u.mac_addr, mp->addr.dst.mac_addr);
 	e.addr.proto = mp->addr.proto;
 	nest_ent = nla_nest_start_noflag(skb,
 					 MDBA_MDB_ENTRY_INFO);
···
 		}
 		break;
 #endif
+	default:
+		ether_addr_copy(e.addr.u.mac_addr, mp->addr.dst.mac_addr);
 	}
 	if (p) {
 		if (nla_put_u8(skb, MDBA_MDB_EATTR_RTPROT, p->rt_protocol))
···
 		if (mp->addr.proto == htons(ETH_P_IP))
 			ip_eth_mc_map(mp->addr.dst.ip4, mdb.addr);
 #if IS_ENABLED(CONFIG_IPV6)
-		else
+		else if (mp->addr.proto == htons(ETH_P_IPV6))
 			ipv6_eth_mc_map(&mp->addr.dst.ip6, mdb.addr);
 #endif
+		else
+			ether_addr_copy(mdb.addr, mp->addr.dst.mac_addr);
+
 		mdb.obj.orig_dev = pg->key.port->dev;
 		switch (type) {
 		case RTM_NEWMDB:
···
 			return false;
 		}
 #endif
+	} else if (entry->addr.proto == 0) {
+		/* L2 mdb */
+		if (!is_multicast_ether_addr(entry->addr.u.mac_addr)) {
+			NL_SET_ERR_MSG_MOD(extack, "L2 entry group is not multicast");
+			return false;
+		}
 	} else {
 		NL_SET_ERR_MSG_MOD(extack, "Unknown entry protocol");
 		return false;
···
 			NL_SET_ERR_MSG_MOD(extack, "Groups with sources cannot be manually host joined");
 			return -EINVAL;
 		}
+	}
+
+	if (br_group_is_l2(&group) && entry->state != MDB_PERMANENT) {
+		NL_SET_ERR_MSG_MOD(extack, "Only permanent L2 entries allowed");
+		return -EINVAL;
 	}
 
 	mp = br_mdb_ip_get(br, &group);

net/bridge/br_multicast.c

···
 		break;
 #endif
 	default:
-		return NULL;
+		ip.proto = 0;
+		ether_addr_copy(ip.dst.mac_addr, eth_hdr(skb)->h_dest);
 	}
 
 	return br_mdb_ip_get_rcu(br, &ip);
···
 		if (notify)
 			br_mdb_notify(mp->br->dev, mp, NULL, RTM_NEWMDB);
 	}
+
+	if (br_group_is_l2(&mp->addr))
+		return;
+
 	mod_timer(&mp->timer, jiffies + mp->br->multicast_membership_interval);
 }
 
···
 			break;
 	}
 
-	p = br_multicast_new_port_group(port, group, *pp, 0, src, filter_mode,
-					RTPROT_KERNEL);
+	p = br_multicast_new_port_group(port, group, *pp, 0, src,
+					filter_mode, RTPROT_KERNEL);
 	if (unlikely(!p)) {
 		p = ERR_PTR(-ENOMEM);
 		goto out;
···
 	memset(&eth, 0, sizeof(eth));
 	eth.h_proto = htons(proto);
 
-	ret = br_multicast_querier_exists(br, &eth);
+	ret = br_multicast_querier_exists(br, &eth, NULL);
 
 unlock:
 	rcu_read_unlock();

net/bridge/br_private.h

···
 void br_multicast_sg_add_exclude_ports(struct net_bridge_mdb_entry *star_mp,
 				       struct net_bridge_port_group *sg);
 
+static inline bool br_group_is_l2(const struct br_ip *group)
+{
+	return group->proto == 0;
+}
+
 #define mlock_dereference(X, br) \
 	rcu_dereference_protected(X, lockdep_is_held(&br->multicast_lock))
 
···
 }
 
 static inline bool br_multicast_querier_exists(struct net_bridge *br,
-					       struct ethhdr *eth)
+					       struct ethhdr *eth,
+					       const struct net_bridge_mdb_entry *mdb)
 {
 	switch (eth->h_proto) {
 	case (htons(ETH_P_IP)):
···
 			&br->ip6_other_query, true);
 #endif
 	default:
-		return false;
+		return !!mdb && br_group_is_l2(&mdb->addr);
 	}
 }