reporting received signal strength to remote side

[Jouni Malinen]

On Tue, Nov 18, 2003 at 08:09:31PM -0500, Jason Straight wrote:

> I've operated WISP's for years, we are (right now) using some orinoco/proxim 
> OR's (Outdoor Routers) which serve as access points for our clients. The OR's 
> have a "Wireless Link Test" function which allows you to see the signal that 
> is being received by the Access point, _and also what's being received by the 
> client from the AP!

> So I'd have to guess that the recv'd signal strength is being reported to the 
> other side by the card firmware on the PHY layer, but when I mention this to 
> anyone they think I'm a dumbass because 802.11 just doesn't support that. So 
> can someone enlighten me as to how this received signal strength gets sent to 
> the other side? Because I sure would love to be able to view the signal being 
> received by the other side from linux as I can from windows or from the AP :)

"Wireless Link Test" uses proprietary data frames (i.e., IEEE 802.11
data with own LLC header (if I remember correctly)) for requesting and
sending this signal information. Processing of these requests is
implemented in firmware and the driver does not get any information
about requests. However, I have not tried to send the request myself and
what happens with the response.. You can use a wireless sniffer to see
those frames, so it should be relatively easy to send identifical frame
to test this. Of course, it would be even easier, if a specification of
the used frame format were available.

I remember some changes in the recent Prism2/2.5/3 firmware versions on
this area and it might be possible to receive some of these frames by
configuring the card in special mode (promiscuous?). I'm not sure
whether this is required for receiving the replies, though.

It has been a while since I last experimented with these frames, so I'm
not sure if I remembered everything correctly.. Anyway, if you can
easily make a sniffer log of some of the frames from Orinoco/Proxim AP
(and replies from a STA), they could be useful for making some quick
tests.

-- 
Jouni Malinen                                            PGP id EFC895FA