[Madwifi-devel] RSSI units

[James S. <js...@cs...>]

Hi,

Like a handful of others who have previously posted, I am interested in 
extracting RSSI from an Atheros chip.

I've successfully done this using Kismet; I had to hack the Kismet server to 
propogate the information to clients, but I can now read RSSI information per 
packet, which is great. 

Now, I want to convert the information to dBm. According to the source code, 
that pdf report floating around, and previous posters, Atheros RSSI is reported 
as the dBm above the noise floor. The noise floor is usually assumed to be -
96dBm according to the source (ah_desc.h:83):

 * rx_rssi is in units of dbm above the noise floor.  This value
 * is measured during the preamble and PLCP; i.e. with the initial
 * 4us of detection.  The noise floor is typically a consistent
 * -96dBm absolute power in a 20MHz channel.

My problem is, the units I am seeing in Kismet appear to be the opposite - they 
appear to be actually just dBm, as they are negative numbers!

Heres a fragment of obfuscated log of two APs sharing a channel broadcasting 
nothing but beacons at each others (the signal:%d is what interests me)

887804 type: 0 subtype: 8 channel: 3 beacon: 25600 smac: 00:09:5B:F9:F4:1E dmac: 
FF:FF:FF:FF:FF:FF ssid: SALTERCORP quality: 0 signal: -45 noise -95
887864 type: 0 subtype: 8 channel: 3 beacon: 25600 smac: 00:02:2D:5E:AA:5F dmac: 
FF:FF:FF:FF:FF:FF ssid: AP2000 quality: 0 signal: -60 noise -95
887904 type: 0 subtype: 8 channel: 3 beacon: 25600 smac: 00:09:5B:F9:F4:1E dmac: 
FF:FF:FF:FF:FF:FF ssid: SALTERCORP quality: 0 signal: -43 noise -95
887964 type: 0 subtype: 8 channel: 3 beacon: 25600 smac: 00:02:2D:5E:AA:5F dmac: 
FF:FF:FF:FF:FF:FF ssid: AP2000 quality: 0 signal: -61 noise -95

I have checked the code path in madwifi doesn't mangle the raw signal level. It 
looks like in monitor mode the data is passed through unchanged (as opposed to 
when being passed to various tools, where the code does the conversion to dBm). 
Kismet doesn't mangle it either.

So it appears this contradicts the comment in the source, and the units are 
probably be dBm, but I need to be certain, and I guess at this point I am 
curious - Is it possible that different Atheros chipsets/firmwares report RSSI 
differently? 

If it helps, lspci reports the chip as an Atheros AR5212, and the unit is a Sony 
VAIO U8G palmtop - not sure how to check the firmware version.

Thanks!

James