[PATCH 0/3] hostapd: introduce Automatic Channel Selection (ACS)

[Michal Kazior]

On 25 August 2013 08:39, Andreas Hartmann <andihartmann at 01019freenet.de> wrote:
> Andreas Hartmann wrote:
>> Michal Kazior wrote:
>>> Hi,
>>>
>>> This is an attempt at ressurecting ACS effort for hostapd.
>>>
>>> This is based on the work done by Luis R. Rodriguez
>>> <lrodriguez at qca.qualcomm.com>.
>>
>> I tested it, too with ath9k (AR9300). It's working pretty fine for me
>> (at a first glance).
>
> Well, the second glance shows a problem on boot up of the machine: I'm
> always getting an error on channel 13 with the noise floor:
>
> 1377411942.774908: ACS: Survey is missing noise floor
> 1377411942.774913: ACS: Channel 13 has insufficient survey data
>
> Restarting hostapd w/o unloading wireless modules stack "resolves" the
> problem.

Hi,

Thanks for testing.


> The complete log output on error:
>
> 1377411935.488850: ACS: Automatic channel selection started, this may take a bit
> 1377411935.488862: ACS: scanning 1 / 5
> (...)
> 1377411936.934816: nl80211: Freq survey dump event (freq=2472 MHz noise=0 channel_time=110 busy_time=0 tx_time=0 rx_time=0 filled=001e)
> (...)
> 1377411942.774890: wlan1: Event SURVEY (56) received
> 1377411942.774900: ACS: Trying survey-based ACS
> 1377411942.774908: ACS: Survey is missing noise floor
> 1377411942.774913: ACS: Channel 13 has insufficient survey data
> 1377411942.774919: ACS: Surveys have insufficient data
> 1377411942.774924: ACS: All study options have failed
> 1377411942.774936: ACS: Failed to start

I'm guessing regulatory switch was taking place while the first scan
was requested. But it's strange - the driver reported channel time
without noise floor. Other scans are okay. Some kind of a race or a
bug in regulatory/ath9k? I'm not sure if that would be a concern for
hostapd. Any other ideas?


Pozdrawiam / Best regards,
Micha? Kazior.