The unicast packet rerouting code makes several assumptions. For instance it assumes that there is always exactly one destination in the TT. This breaks for multicast frames in a unicast packets in several ways:

For one thing if there is actually no TT entry and the destination node was selected due to the multicast tvlv flags it announced. Then an intermediate node will wrongly drop the packet.

For another thing if there is a TT entry but the TTVN of this entry is newer than the originally addressed destination node: Then the intermediate node will wrongly redirect the packet, leading to duplicated multicast packets at a multicast listener and missing packets at other multicast listeners or multicast routers.

Fixing this by not applying the unicast packet rerouting to batman-adv unicast packets with a multicast payload. We are not able to detect a roaming multicast listener at the moment and will just continue to send the multicast frame to both the new and old destination for a while in case of such a roaming multicast listener.

Fixes: cea194d90b11 ("batman-adv: improved client announcement mechanism") Signed-off-by: Linus Lüssing linus.luessing@c0d3.blue --- net/batman-adv/routing.c | 4 ++++ 1 file changed, 4 insertions(+)

diff --git a/net/batman-adv/routing.c b/net/batman-adv/routing.c index 27cdf5e4..9e5c71e4 100644 --- a/net/batman-adv/routing.c +++ b/net/batman-adv/routing.c @@ -826,6 +826,10 @@ static bool batadv_check_unicast_ttvn(struct batadv_priv *bat_priv, vid = batadv_get_vid(skb, hdr_len); ethhdr = (struct ethhdr *)(skb->data + hdr_len);

+ /* do not reroute multicast frames in a unicast header */ + if (is_multicast_ether_addr(ethhdr->h_dest)) + return true; + /* check if the destination client was served by this node and it is now * roaming. In this case, it means that the node has got a ROAM_ADV * message and that it knows the new destination in the mesh to re-route

Scenario: * Multicast frame send from a BLA backbone (multiple nodes with their bat0 bridged together, with BLA enabled)

Issue: * BLA backbone nodes receive the frame multiple times on bat0

For multicast frames received via batman-adv broadcast packets the originator of the broadcast packet is checked before decapsulating and forwarding the frame to bat0 (batadv_bla_is_backbone_gw()-> batadv_recv_bcast_packet()). If it came from a node which shares the same BLA backbone with us then it is not forwarded to bat0 to avoid a loop.

When sending a multicast frame in a non-4-address batman-adv unicast packet we are currently missing this check - and cannot do so because the batman-adv unicast packet has no originator address field.

However, we can simply fix this on the sender side by only sending the multicast frame via unicasts to interested nodes which do not share the same BLA backbone with us. This also nicely avoids some unnecessary transmissions on mesh side.

Note that no infinite loop was observed, probably because of dropping via batadv_interface_tx()->batadv_bla_tx(). However the duplicates still utterly confuse switches/bridges, ICMPv6 duplicate address detection and neighbor discovery and therefore leads to long delays before being able to establish TCP connections, for instance. And it also leads to the Linux bridge printing messages like: "br-lan: received packet on eth1 with own address as source address ..."

Fixes: 405cc1e5a81e ("batman-adv: Modified forwarding behaviour for multicast packets") Signed-off-by: Linus Lüssing linus.luessing@c0d3.blue --- net/batman-adv/send.c | 13 +++++++++++++ 1 file changed, 13 insertions(+)

diff --git a/net/batman-adv/send.c b/net/batman-adv/send.c index d267b948..67f493c0 100644 --- a/net/batman-adv/send.c +++ b/net/batman-adv/send.c @@ -29,6 +29,7 @@ #include <linux/stddef.h> #include <linux/workqueue.h>

+#include "bridge_loop_avoidance.h" #include "distributed-arp-table.h" #include "fragmentation.h" #include "gateway_client.h" @@ -343,6 +344,18 @@ int batadv_send_skb_unicast(struct batadv_priv *bat_priv, if (!orig_node) goto out;

+ /* Avoid sending multicast-in-unicast packets to other BLA + * gateways - they already got the frame from the LAN side + * we share with them. + * TODO: Refactor multicast code to anticipate this, to + * avoid this check here. + */ + if (is_multicast_ether_addr(eth_hdr(skb)->h_dest) && + batadv_bla_is_backbone_gw_orig(bat_priv, orig_node->orig, vid)) { + dev_kfree_skb(skb); + return NET_XMIT_SUCCESS; + } + switch (packet_type) { case BATADV_UNICAST: if (!batadv_send_skb_prepare_unicast(skb, orig_node))

Scenario: * Multicast frame send from mesh to a BLA backbone (multiple nodes with their bat0 bridged together, with BLA enabled)

Issue: * BLA backbone nodes receive the frame multiple times on bat0, once from mesh->bat0 and once from each backbone_gw from LAN

For unicast, a node will send only to the best backbone gateway according to the TQ. However for multicast we currently cannot determine if multiple destination nodes share the same backbone if they don't share the same backbone with us. So we need to keep sending the unicasts to all backbone gateways and let the backbone gateways decide which one will forward the frame. We can use the CLAIM mechanism to make this decision.

One catch: The batman-adv gateway feature for DHCP packets potentially sends multicast packets in the same batman-adv unicast header as the multicast optimizations code. And we are not allowed to drop those even if we did not claim the source address of the sender, as for such packets there is only this one multicast-in-unicast packet.

How can we distinguish the two cases?

For DHCPv4: Here the broadcast MAC address is used and the multicast optimizations will never send a broadcast frame via batman-adv unicast packets (see the !is_broadcast_ether_addr() check in after the goto-send in batadv_interface_tx().

For DHCPv6: This is even trickier... DHCPv6 potentially uses non-broadcast multicast addresses. However according to RFC8415, section 7.1 it seems that currently multicast is only used from a DHCPv6 client to a DHCPv6 server, but not the other way round.

Working through the gateway feature part in batadv_interface_tx() it can be inferred that a DHCPv6 packet to a DHCP client would have been the only option for a DHCPv6 multicast packet to be sent via unicast through the gateway feature. Ergo, the newly introduced claim check won't wrongly drop a DHCPv6 packet received via the gateway feature either.

Fixes: e32470167379 ("batman-adv: check incoming packet type for bla") Signed-off-by: Linus Lüssing linus.luessing@c0d3.blue --- net/batman-adv/bridge_loop_avoidance.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/net/batman-adv/bridge_loop_avoidance.c b/net/batman-adv/bridge_loop_avoidance.c index d8c5d317..9603a6d0 100644 --- a/net/batman-adv/bridge_loop_avoidance.c +++ b/net/batman-adv/bridge_loop_avoidance.c @@ -1848,7 +1848,8 @@ bool batadv_bla_rx(struct batadv_priv *bat_priv, struct sk_buff *skb,

if (unlikely(atomic_read(&bat_priv->bla.num_requests))) /* don't allow broadcasts while requests are in flight */ - if (is_multicast_ether_addr(ethhdr->h_dest) && is_bcast) + if (is_multicast_ether_addr(ethhdr->h_dest) && + (!is_broadcast_ether_addr(ethhdr->h_dest) || is_bcast)) goto handled;

ether_addr_copy(search_claim.addr, ethhdr->h_source); @@ -1885,7 +1886,8 @@ bool batadv_bla_rx(struct batadv_priv *bat_priv, struct sk_buff *skb, }

/* if it is a broadcast ... */ - if (is_multicast_ether_addr(ethhdr->h_dest) && is_bcast) { + if (is_multicast_ether_addr(ethhdr->h_dest) && + (!is_broadcast_ether_addr(ethhdr->h_dest) || is_bcast)) { /* ... drop it. the responsible gateway is in charge. * * We need to check is_bcast because with the gateway