Re: Incorrect signal levels reported using wpa_supplicant's driver_nl80211

[Brian Morrison <bdm fenrir org uk>]

On Mon, 02 Nov 2009 17:02:44 -0800
Dan Williams <dcbw redhat com> wrote:

> On Mon, 2009-11-02 at 23:53 +0000, Brian Morrison wrote:
> > On Tue, 03 Nov 2009 01:07:22 +0200
> > Maxim Levitsky <maximlevitsky gmail com> wrote:
> > 
> > > > This is *exactly* why 'qual' is there: so that the driver itself can
> > > > figure out what the hell it's signal level is, and so that NM doesn't
> > > > have to go around assuming stuff.  
> > > It appears as kernel developers want to drop quality calculations
> > > altogether, and report everything in dBms (right?) 
> > 
> > But quality does not necessarily track signal level. Consider the case
> > where a second AP is on the same frequency but weaker than the AP that
> > is associated with. The quality of the connection is the difference in
> > signal levels (roughly the SNR) which is not the same as the signal
> > level of the stronger AP.
> 
> You're only associated to one AP at a time on the same wlan device.
> Thus whatever signal quality any other AP has doesn't matter.  The only
> signal level or quality that matters is the one for the link through
> which your packets go.  When the driver roams APs, the quality should
> update to reflect the signal strength and quality of the new AP, no
> matter what channel it's on.

I'm only talking about a single AP Dan, but if another AP is
transmitting on the same frequency and there is a simultaneous
transmission with the associated AP then the quality is poorer than if
there is clear air. CCA doesn't always work you know.

The quality metric should be returned by the modem, RSSI is usually not
calculated by the modem but is instead generated from signal levels in
various parts of the receiver.

I'm only suggesting that things may not be what they seem.

-- 

Brian Morrison

                "I am not young enough to know everything"
                                                          Oscar Wilde