Open-Mesh » batmand » vis

first of all: The routing works perfectly.

node-A 105.61.250.1 has batman running on eth1 (primary) and tap0 (secondary).
node-B 105.61.252.129 has batman running on tap0. node-B is the visualisation-server.
they reach each other directly via tap0.
batman sets the correct routes on both nodes, however the visualisation-packets don't
get sent on the right interface (despite the routing table).

following output on node-A (105.61.250.1)

batmand is started like this:

root@se:~# ps aux |grep batman
2930 root        408 S   batmand -g 10 -s 105.61.252.129 eth1:bat tap0:bat

the topoviser server is reachable via tap0, batman found this route.

root@se:~# route -n |grep 105.61.252.129
105.61.252.129  0.0.0.0         255.255.255.255 UH    0      0        0 tap0

ping works...

root@se:~# ping 105.61.252.129
PING 105.61.252.129 (105.61.252.129): 56 data bytes
64 bytes from 105.61.252.129: icmp_seq=0 ttl=64 time=106.4 ms

however batmand wants to send the vis-packets via eth1.
of course the topoviser server is not reachable via eth1 but via tap0

root@se:~# tcpdump -n -i eth1 net 105 and arp
18:04:44.293731 arp who-has 105.61.252.129 tell 105.61.250.1

no visualisation-packets on tap0.
i think that this is a bug in the visualisation-socket-code.
maybe it is a false assumption that vis-packets are only sent on the primary interface?!

batman version...

root@se:~# batmand -v
B.A.T.M.A.N.-III vIII 0.2.0 rc2 rv372 (compability version 3)