[PATCH]: wait_for_interface poll on driver interface initialization

[Jouni Malinen]

On Sun, Jan 13, 2008 at 07:41:21PM +1000, Kel Modderman wrote:
> The attached patch attempts to fix the wait_for_interface option by
> reorganizing wpa_supplicant_driver_init() and wpa_supplicant_init_iface2() so
> that polling is done before the driver interface is prepared for work.
> 
> Currently, the driver interface is prepared before wait_for_interface is
> considered, resulting in broken -w command line option, and ignoring any driver
> parameters.

Hmm.. I must have broken this at some point when moving initialization
calls around to fetch MAC address early enough for other code modules..

> I'd be just as happy to see the wait_for_interface stuff completely removed,
> as any distro should have some hotplug agent or similar support in their
> networking toolset. Therefore, I see no need for this option to exist at all,
> but while it does exist and does not work, I'll have a go at fixing it ;-)

I think I would prefer to just get rid of it completely.. It's
unnecessary extra complexity that is obviously easy to break (as seen
here..). I've never used it myself (apart from the minimal testing I've
done) and I would agree that there are better ways of handling this.
Anyway, similar functionality can be added by using a separate program
to do the waiting if someone really needs it and does not have hotplug
(etc.)..

Does anyone see any good use for keeping this functionality in
wpa_supplicant? Any objections to removing '-w' option?

-- 
Jouni Malinen                                            PGP id EFC895FA