> -----Message d'origine----- > De : Felix Fietkau [mailto:nbd@xxxxxxxx] > Envoyé : jeudi 15 février 2018 15:46 > À : Jean Pierre TOSONI; Kalle Valo > Cc : linux-wireless@xxxxxxxxxxxxxxx; ath9k-devel@xxxxxxxxxxxxxxxx > Objet : Re: [PATCH v2] ath9k: mark RSSI as invalid if frame received > during channel setup > > On 2018-02-14 17:26, Jean Pierre TOSONI wrote: > > ath9k returns a wrong RSSI value for frames received in a 30ms > time > > window after a channel change. The correct value is typically 10dB > > below the returned value. > > > > This was found with a Atheros AR9300 Rev:3 chip (WLE350NX / > JWX6083 > > cards), during offchannel scans. > > > > Mark the signal value as invalid in this case. > > > > Signed-off-by: Jean Pierre TOSONI <jp.tosoni@xxxxxxxxx> > It seems that this will potentially cause a lot of scan results with > RSSI information missing, which I think may be worse than a 10 db > fluctuation. Hmmm... My client is moving linearly from AP to AP and if it finds that the previous AP is "better" than the current or next one, havoc will occur on throughput! In this case I prefer to wait for the next scan round than use a wrong value. Also, note that my patch sets the invalid flag but it is currently Ignored by mac80211 (I submitted another patch about that). Actually the problem shows up only on passive channels, because (1) on active channels the wrong beacon report will be soon overwritten by the probe response report (2) mac80211 waits only for 108ms on passive channels, if it waited for 133ms we would have the opportunity to obtain a 2nd correct beacon report (well, depending on the beacon interval of course) > Actually, it could also be that the hardware is reporting accurate > RSSI values (they're reported as SNR relative to the noise floor) > and this is just a blip caused by the fact that the hw is updating > its internal noise floor value based on running measurements. > > Does the same fluctuation also happen if you set update = false for > calls to ath9k_hw_start_nfcal? Thanks for the hint, I will have a look at that > > - Felix
