How to you get the supplicant/linux wireless to perform active scans rather
than passive scans? Are active scans faster?
Thank you,
Chuck
On Mon, 2010-09-27 at 17:09 +0200, ext Johannes Berg wrote:
> On Mon, 2010-09-27 at 18:06 +0300, Luciano Coelho wrote:
>
> > Usually the drivers stay for a certain period of time in each channel,
> > and that normally doesn't change if you have active or passive scan. So
> > there will be no improvement in the overall scan speed. Of course you
> > could do some tweaks that might improve scanning performance in specific
> > cases.
>
> It's certainly different with mac80211:
>
> #define IEEE80211_CHANNEL_TIME (HZ / 33)
> #define IEEE80211_PASSIVE_CHANNEL_TIME (HZ / 8)
Right... Even in the wl1271 driver's hw scan we stay longer on each
channel for passive scans. Obviously just listening to beacons we need
to wait more, because probe_reqs should be responded quickly.
--
Cheers,
Luca.
Hi Chuck,
On Mon, 2010-09-27 at 16:03 +0200, ext Chuck Crisler wrote:
> How to you get the supplicant/linux wireless to perform active scans rather
> than passive scans? Are active scans faster?
We do active scans by default, unless specified otherwise in the
regulatory database.
The main difference between active and passive scans is that in active
scans we send broadcast probe requests on the channel and with passive
we don't. In passive scan we just listen to beacons.
Usually the drivers stay for a certain period of time in each channel,
and that normally doesn't change if you have active or passive scan. So
there will be no improvement in the overall scan speed. Of course you
could do some tweaks that might improve scanning performance in specific
cases.
--
Cheers,
Luca.
On Mon, 2010-09-27 at 18:06 +0300, Luciano Coelho wrote:
> Usually the drivers stay for a certain period of time in each channel,
> and that normally doesn't change if you have active or passive scan. So
> there will be no improvement in the overall scan speed. Of course you
> could do some tweaks that might improve scanning performance in specific
> cases.
It's certainly different with mac80211:
#define IEEE80211_CHANNEL_TIME (HZ / 33)
#define IEEE80211_PASSIVE_CHANNEL_TIME (HZ / 8)
johannes