TX power - growl!

Pavel Roskin proski at gnu.org
Wed Mar 31 18:35:15 EST 2004


On Tue, 30 Mar 2004 mike-hostap at tiedyenetworks.com wrote:

> 	So why is the subject of TX power such a dark mysterious secret?
> I've got some seano/engenius cards (pcmcia) and some I use in access
> point configurations, and others I use in client devices. But in either
> mode, setting the tx power using the iwpriv wlan0 prism2_param 6 x
> command doesn't seem to have any effect whatsoever on the power output
> of the card. Yes my testing is suspect because I don't have a spectrum
> analyzer at my disposal and I was only able to test by associating with
> a cisco AP and reading the 'received signal strength', but still I can't
> detect any change wether I use 1, 50, 126, 129 or 254. I also tried the
> wlan-ng drivers (for client mode) and they claim no tx powers are
> available.. so I'm just lost.

See comments in hostap sources.  Support for txpower setting was removed
in version 0.0.4 because it wasn't implemented correctly.  The code is
still in the driver, but it's disabled.  Look for RAW_TXPOWER_SETTING.  In
particular, read this comment from the driver:

"Note! This TX power controlling is experimental and should not be used in
production use. It just sets raw power register and does not use any kind
of feedback information from the measured TX power (CR58). This is now
commented out to make sure that it is not used by accident. TX power
configuration will be enabled again after proper algorithm using feedback
has been implemented."

By the way, TX power is implemented using registers of the baseband
processor.  It's a very hackish approach because it circumvents the
firmware.  No other driver doesn't even attempt to regulate tx power on
Prism chipset.

Also, look for HFA384X_RID_TXPOWERMAX.  It's currently unused.  It can be
relevant.  Reenable prism2_ioctl_giwtxpow() and try changing that RID by
hostap_rid utility.  See iwconfig output for the tx power read from the
baseband processor.  Not tested but may work.

-- 
Regards,
Pavel Roskin



More information about the HostAP mailing list