[Olsr-dev] 0.5.6-rc5 is out

[Hannes Gredler]

Bernd Petrovitsch wrote:
> On Son, 2008-06-15 at 21:02 +0200, Hannes Gredler wrote:
> [...]
>> we only insert the candidate route into the RIB on initial,
>> reception of a HNA network. now if the TC entry vanishes for some
>> reason (e.g. timeout) the candidate route is gone, while the
>> HNA entry is still alive which is not quite right.
> 
> Hmm, should the kernel also kill the route to the HNA automatically if
> the route to the announcing node is deleted?

thats what we do ... see olsr_update_rib_routes().

first check if each stored route from the previous SPF run is reachable,
if not (all rt_paths gone) denn we enqueue a route deletion.
(comment: oops, all routes are gone)

/hannes