On Mon, Apr 29, 2013 at 10:26 AM, Sam Leffler <[email protected]> wrote:
> On Mon, Apr 29, 2013 at 9:54 AM, Thomas Pedersen <[email protected]> wrote:
>>
>> Hi Zaki,
>>
>> On Mon, Apr 29, 2013 at 11:11:26AM +0800, Zaki wrote:
>> > Hi,
>> >
>> >
>> > I would like to highlight some observation that i notice in my mesh
>> > network.  Setting up 10 nodes is easy, so they are now working 24/7.
>> > The
>> > exercise i did is for them to ping each other.  While mostly they are
>> > stable, there are moments where Ping Timeout happen for some seconds
>> > before
>> > the ping resumes.  It happened in all the 10 nodes randomly.  Then, i
>> > use
>> > Fluke Aircheck WiFi Tester to see what is happening during that time.
>> > The
>> > results are quite surprising to me :
>> > 1) The beacon detection (beeps on the device) is very random interval
>> > and
>> > most of its interval is almost around 1 sec each. Sometimes a long pause
>> > between the beeps.
>>
>> Assuming you're using some ath9k cards for testing, this patch:
>>
>> https://github.com/cozybit/open80211s/commit/380bf3ad4c2ccd92658943e978f8d942611a3d6d
>> might help with a more stable beacon interval. If you do test and notice
>> an improvement, please let us know.
>
>
> Why would this change beacon regularity?  In the adhoc case I thought the
> beacon xmit was handled entirely in the radio while for the ap operation it
> requires the host to post the beacon frame at SWBA?   I thought using AP
> mode was to fix multi-vif support?
>
> If beacons are being delayed due to the channel being busy then that's a
> separate issue--unless the AIFS parameters for AP are different than ADHOC
> operation.

You're right, Chun-Yeow presents this patch:
https://gist.github.com/chunyeow/426d96d529584f66178f which seems to
lower the AIFS/CW parameters for mesh beaconing.

--
Thomas
_______________________________________________
Devel mailing list
[email protected]
http://lists.open80211s.org/cgi-bin/mailman/listinfo/devel

Reply via email to