Re: Braindump on wpa_supplicant

[Olivier Blin <oblin mandriva com>]

Dan Williams <dcbw redhat com> writes:

> We could rely on wpa_supplicant to do scanning for us (or, perhaps, only 
> when we ask it to scan) and then read the scan data from wpa_supplicant on a 
> per-interface basis. 

It looks like a good plan!
That's how I proceed in Mandriva tools to get scanning results.
Then a small generic daemon forward the results from the socket to
user applications using D-Bus.

There's a small problem though, wpa_supplicant only allows one control
program to be attached on the control socket. So if another program
starts trying to attach on the control socket, the daemon will loose
notification of events.
It would be nice to enhance wpa_supplicant so that multiple program
can attach at the same time. Maybe I'll ask Jouni on the hostap ML.

> That's a lot of work though.

Something like 3 lines of Perl code ? :-p

> In any case, wpa_supplicant needs some major changes architecturally to deal 
> with network interfaces not being present when it starts up.  The other solution 
> would be to start 1 instance of wpa_supplicant per-interface (even wired 
> interfaces may need 802.1x authentication), and ignore the resulting memory 
> bloat.

One wpa_supplicant daemon allocates less than 200 KB of memory, that's
an acceptable compromise.

-- 
Olivier Blin
Mandriva
Mandrakesoft becomes Mandriva