[Olsr-dev] Idea for small OLSR improvement

[Mitar]

HI!

On Sat, Apr 24, 2010 at 5:14 PM, Sven-Ola Tücke <(spam-protected)> wrote:
> madwifi. Absolutely. Check you kernel log for "stuck beacon" messages. I bet
> it's the case. Have you tried the openwrt/kamikaze patchset together with the
> "adhocmode with HAL in master?" That's the nosbeacon flag when you do
> "wlanconfig create ath0..."

Stuck beacons are there but I have not noticed that it would
correlate. I will do a little bit more thorough tests/analysis. Also
pings (every second) worked without problems in both directions while
OLSR was dropping the route.

But yes, I also have it as a main suspect. Will investigate more.

But I also read here:

http://wiki.villagetelco.org/index.php?title=Information_about_cell-id_splitting,_stuck_beacons,_and_failed_IBSS_merges!

that this workaround has impact on Broadcom hardware:

It has been observed that this little TSF deviants confused devices
with Broadcom chipsets, such as the Linksys WRT54GL: They were still
working but they stopped sending beacons.

Is this true? Is there no other less hacky solution? Is there any work
being done to fix this in HAL?

This is why I am not really enthusiast about using this approach.

Thanks


Mitar