[Olsr-dev] 0.5.6-rc5 is out

[Hannes Gredler]

Bernd Petrovitsch wrote:
> Then I don't understand the bug.

there is a couple of things:

1. a tc entry may timeout, gets not refreshed.

2. a tc_entry is the attachment point for rt_paths
    --> i.e. if a tc_entry goes away then all the
    rt_paths are flushed too.

3. the hna_entry and hna_nets are independent of the
    tc_entry. the rt_path is only created when a HNA
    net is fresh created.

now consider the following series of events.

1. tc goes away (note that the HNA is still valid)
    the route shas been removed from the kernel since the
    tc is not reachable anymore.

2. the tc is updated. now the rt_path is not resinserted,
    since the hna_net was not changed, did get properly refreshed.

you only get back the hna net if it times out (is removed) and later on
gets refreshed and thehna_net is re-created.

HTH,


/hannes