[Olsr-dev] Hop count metric: link quality and kernel route update probelms: bug?

[Hannes Gredler]

dario,

that might be a regression introduced by myself.
let me triage this a bit.

/hannes

Dario Borriello wrote:
> Hi all,
> 
> I'm a new developer of olsrd 0.5.4, 
> 
> and as first step I'm testing the Original 0.5.4 daemon in a simple scenario composed by three nodes:
> 
> A----------B----------C 
> 
> The node B has 2 wireless interface, A and C only one. 
> 
> If I use only hop count metric (LinkQuality =0) there are 2 great problems: 
> 
> 1) The inverse link quality that every node (A, B, C) estimates between his neighbors is 0 
> 
> Example Topology Set of node A: 
> 
> Link Quality A->B  = 1 (correct, set to default value)
> 
> Link Quality A->B  = 0 (why?????)
> 
> Link Quality B->C  = 0 (why?????)
> 
> Link Quality A->B  = 0 (why?????) 
> 
> 2) Moreover, even if the topology set is 'almost correct', 
> 
>  the kernel routing table of node A and C doesn't contain the entry for the 2 hop-neighbor!
> 
>  and so they can't ping each other! 
> 
> I made the same test with Link Quality Metric, and all works correctly! 
> 
> Can someone suggest a solution?  
> 
> Thank you 
> 
> Dario
> 
> 
> ------------------------------------------------------------------------
> Per questo Natale fai i tuoi auguri con Messenger! Windows Live 
> Messenger <http://www.augurimessenger.it>
>