FRR OSPF hello with wrong mask

adorne commented 5 years ago

BSDRP 1.91 It shows GRE interfaces as unnumbered and consequently send "hello" messages by OSPF with 0.0.0.0 mask from an /30 interface. RouterOS on the other side doesn't accept wrong mask "hello" and routing is broken. There is how it looks like.

system:

[root@n101v01]~# ifconfig tok1
tok1: flags=8051<UP,POINTOPOINT,RUNNING,MULTICAST> metric 0 mtu 1330
        options=80000<LINKSTATE>
        tunnel inet ip1 --> ip2
        inet 10.0.0.1 --> 10.0.0.2 netmask 0xfffffffc 
        nd6 options=29<PERFORMNUD,IFDISABLED,AUTO_LINKLOCAL>
        groups: gre

FRR cli:

n101v01# show ip ospf interface tok1
tok1 is up
  ifindex 7, MTU 1330 bytes, BW 0 Mbit <UP,POINTOPOINT,RUNNING,MULTICAST>
  This interface is UNNUMBERED, Area 0.0.0.0
  MTU mismatch detection: enabled
  Router ID 10.10.1.1, Network Type POINTOPOINT, Cost: 10
  Transmit Delay is 1 sec, State Point-To-Point, Priority 1
  No backup designated router on this network
  Multicast group memberships: OSPFAllRouters
  Timer intervals configured, Hello 10s, Dead 40s, Wait 40s, Retransmit 5
    Hello due in 2.719s
  Neighbor Count is 1, Adjacent neighbor count is 1

On the other BSDRP 1.90 system it's all right with masks.

deejay2 commented 5 years ago

Not sure but it may be a FRRouting bug. If you haven't already done it, you could open a similar issue on the FRRouting project so they can sort it out.

ocochard commented 2 years ago

I think it is related to this FRR bug: https://github.com/FRRouting/frr/pull/8132

ocochard / BSDRP

FRR OSPF hello with wrong mask #27