WIFI Connector Application

Al Johnson openmoko at mazikeen.demon.co.uk
Thu Sep 25 21:02:18 CEST 2008


On Thursday 25 September 2008, Matthias Apitz wrote:
> El día Thursday, September 25, 2008 a las 03:50:21PM +0100, Al Johnson 
escribió:
> > On Thursday 25 September 2008, Matthias Apitz wrote:
> > > El día Thursday, September 25, 2008 a las 03:16:43PM +0100, Arigead
> >
> > escribió:
> > > > Hello all,
> > > >     I'm looking for an application which if it exists great and if it
> > > > don't exist I'll maybe try and implement it in C, if what I'm
> > > > thinking is even possible. I've not done too much investigating into
> > > > Wifi as I believe that it has its issues on the FR.
> > >
> > > 	...
> > >
> > > Isn't wpa_supplicant doing exactly this (it does for me on my normal
> > > laptop) having /etc/wpa_supplicant.conf as the database for all your
> > > known Wifi networks?
> >
> > It would do if all the debian network scripts are in place. Last time I
> > tried (some time ago now) wpa_supplican would connect, but the scripts
> > that are supposed to run on connection to sort out IP, DNS etc. were
> > missing.
>
> I think the work that wpa_supplicant daemon does is only to control the
> association to the AP and bringing up the interface; then in FreeBSD
> a devd (device daemon) lets you describe what to do else, for example
> launch a script when IF_UP; in this script you may, based on the MAC of
> the AP, do whatever is needed: ifconfig static IP or DHCP or ....
>
> it is similiar to what I have described here for the cdce0 (USB) interface:
>
> http://wiki.openmoko.org/wiki/Usb_networking#FreeBSD
>
> a logic like that must exist in Debian too, I'm wrong?
> and this is not much scripting work, or?

Exactly. The logic exists in debian. wpa_supplicant associates with the access 
point then some scripts apply the configuration you set 
in /etc/network/interfaces that matches the AP it associated with. 
Unfortunately in 2007.2 the bit that applies the settings is missing. For 
someone who knows their way around the inner workings of the debian system it 
would probably be easy to fix, but I don't. Since ASU was moving to use 
connman instead of the debian method of network management I didn't see the 
benefit in digging deeper at the time. 





More information about the community mailing list