[Olsr-dev] [PATCH][BUILD] check_neighbor_link is wrong

[Henning Rogge]

On 12/11/2012 01:15 PM, Ferry Huberts wrote:

>>> I am a little bit scared to look up who use this function if it can
>>> return an asymmetric link. Asymmetric links are useless for outbound
>>> routing.
>>>
>> Hi Henning,
>>
>> the backup link is given back only if no symmetric link has been
>> found, which is the same that is done now without the patch.
>> Do you think we should return null in case there is an asymmetric link
>> but no a symmetric link?

Yes, I think we should test if this wrecks anything. Asymmetric links 
shouldn't be used at all.

>> If things stay as they are, it could be made clear that it is caller
>> responsibility to check what kind of link is the best link given by
>> the get_best_link function.
>
> In my experience, asymmetric links are evil and cause all kinds of
> unexpected, hard to diagnose, problems.
>
> I would _very much_ like to avoid them.

Its not only that asymmetric links make trouble, they should be pretty 
useless from the protocols view because in OLSR asymmetric links should 
be nearly always incoming links. You know you hear someone, but you 
don't know if he can hear you.

Which means you cannot send him traffic anyways.

Henning Rogge

-- 
Diplom-Informatiker Henning Rogge , Fraunhofer-Institut für
Kommunikation, Informationsverarbeitung und Ergonomie FKIE
Kommunikationssysteme (KOM)
Fraunhofer Straße 20, 53343 Wachtberg, Germany
Telefon +49 228 9435-961,   Fax +49 228 9435 685
mailto:(spam-protected) http://www.fkie.fraunhofer.de

-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 6169 bytes
Desc: S/MIME Cryptographic Signature
URL: <http://lists.olsr.org/pipermail/olsr-dev/attachments/20121211/6049b65e/attachment.bin>