From: Eric J Rozner <ejrozner@st...> - 2005-03-12 21:58:06
Actually, I have already created a wireless API by hacking iwconfig.c to
our purposes. However, I'm wondering if HAL will give us finer control
over the wireless interfaces. For instance, one experiment we wanted to
perform was to see how long it takes to switch from one channel to
another. In the nonBSD branch of madwifi, Ad-Hoc is broken. Thus, we
were forced to use Master or Managed mode. Whenever changing the
channel, overhead occurs in the form of reassociating a client (Master
mode) or scanning the channel for AP's (Managed mode). I downloaded the
BSD version of madwifi and was able to get ad-hoc working. I then
performed my experiments- it doesn't look like there is any over-head
this time, but I'm not sure. That's why I think it would be beneficial
to use HAL.
Thanks for your quick response.
> Do you really want to reinvent the wheel? You could instead use the
> linuxwireless extension api, which madwifi uses. Changing channels
> and so on
> is done
> through ioctl's and wireless extension/ wireless tools. Since the
> hal is
> relatively useless without a driver (madwifi) I dont see any efford
> yet.Maybe you should explain what is your goal?
> Sebastian Weitzel
From: Michael Renzmann <madwifi@no...> - 2005-03-17 16:38:36
Eric J Rozner wrote:
> over the wireless interfaces. For instance, one experiment we wanted to
> perform was to see how long it takes to switch from one channel to
If you search the archives of madwifi-devel you should be able to find a
thread that talks about this particular task. iirc Sam described how
such meassurements could be achieved, but I'm not sure.