[OpenWrt-Devel] Q: mac80211: default distance-settings 0

Bill bmoffitt at ayrstone.com
Tue Oct 7 13:52:06 EDT 2014


Message: 1 Date: Tue, 7 Oct 2014 17:17:40 +0200 From: Bastian Bittorf 
<bittorf at bluebottle.com> To: mailinglist 
<openwrt-devel at lists.openwrt.org> Subject: Re: [OpenWrt-Devel] Q: 
mac80211: default distance-settings 0 Message-ID: 
<20141007151740.GJ29315 at medion.lan> Content-Type: text/plain; 
charset=us-ascii * Felix Fietkau <nbd at openwrt.org> [07.10.2014 13:40]:
>> On 2014-10-07 08:15, Bastian Bittorf wrote:
>>> because 0 seems to be a valid value:
>> 0 does not imply dynamic ACK, it is simply the minimum value.
>> Enabling dynack by default would be a bad idea.
> what does 0 mean? the wiki says: 0 meters away, so a short
> ack-timeout is used, or is '0' something special, eg. driver default?
>
> i tested a p2p/longshot here, where both stations are 350m away, but
> invoking on both sides:
>
> iw phy phy0 set distance 350
>
> shows, that the link gets really worse, also with 500 or 2000.
> can't it be changed during runtime?
>
> bye, bastian
>
>
> ------------------------------
>
Bastian-

I was doing some tests last week using an access point running a CC 
trunk build.

I remembered that the "right" value was supposed to be the actual 
distance*2 (essentially, counting "out and back" distance).

My test link was at about 8 km. I tried a number of values (on both AP 
and client) - distance=15000, 10000, and then just backed it off by 1000 
incrementally - and the throughput (measured with iperf) got better as I 
went down in distance. The optimal throughput was at distance=5000, 
where it was about 5 Mbps. However, when I set it at distance=4000, 
suddenly throughput went to less than 200 Kbps.

"Real world" observations, FWIW...

-Bill
_______________________________________________
openwrt-devel mailing list
openwrt-devel at lists.openwrt.org
https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel



More information about the openwrt-devel mailing list