[ath9k-devel] Question about Atheros AR9220

[gtolon at inti.gob.ar <gtolon@inti.gob.ar>]

Hi,

I'm also interested in this topic because i've been doing tests with 
interference with TP-Link (AR9285 chip), and i still don't understand 
completely what happens.

Using g mode, i ran iperf between two routers in channel 11. Then i used 
another router in channel 10, also running iperf with a PC. I used 
tcpdump to get the packets between routers in channel 11, and i can see 
that when the transmission in channel 10 is on, the data rate drops. I 
see a considerable amount of retransmitted packets, but that doesn't 
explain all the loss of performance with respect to the transmission 
without interference. I don't know if some packets are trasmitted but i 
can't see them because they are corrupted at the phy layer, or if 
they're not being transmitted because of CCA.

If ath9k implements a noise CCA, do you know how to be sure if it's 
being used. I've looked at iw survey dump, in channel busy time, but 
looking into the code it seems that it just reads a register value from 
the chip to get the number of busy cycles, so i don't know how it's 
determined.

Also, i 've done some tests with noise from a signal generator, in the 
same channel that the transmission between routers, and i observed that 
the transmission completely cuts with a calculated noise level of just 
-85 dbm (occupying the whole 20 MHz bandwith, FM modulation). I would 
expect a greater noise signal to prevent the routers to transmitt if it 
was due to CCA, or a condition where the transmission dropped little by 
little if it was due to transmitted packets damaged by the noise.

I've been doing these tests because sometimes i experience throughput 
drops and i would like to understand better the reasons behind, if it's 
because of wiphy signals in the some channel, in other channels, or just 
non wiphy noise signals. Thank you!

Gabriel



El 05/03/2012 04:55 p.m., Adrian Chadd escribi?:
> Hi,
>
> You're likely much more versed in the nuances of the specification than I am. :)
>
> My understanding of the AR_PHY_CCA_THRESH62 field is that it indeed is
> the value that defines the -62 dBm CCA threshold, likely used for what
> is termed "mode 1".
>
> The other detection modes are likely engaged when the RX hardware has
> successfully detected a preamble or started decoding a frame - it then
> treats the medium as unavailable (as it's receiving.) This can be at a
> _much lower_ power level.
>
>
> Adrian
> _______________________________________________
> ath9k-devel mailing list
> ath9k-devel at lists.ath9k.org
> https://lists.ath9k.org/mailman/listinfo/ath9k-devel
>