Re: [PATCH] libertas: implement new scanning logic

[Dan Williams <dcbw@redhat.com>]

On Fri, 2007-12-07 at 18:33 +0000, David Woodhouse wrote:
> On Fri, 2007-12-07 at 13:24 -0500, Dan Williams wrote:
> > You can really just take that NOKEY line out completely, you don't need
> > anything else in this patch.
> 
> Well, there's no point in setting/clearing IW_ENCODE_DISABLED twice,
> which is why I removed that bit...
> 
> > > -	if (   adapter->secinfo.wep_enabled
> > > -	    || adapter->secinfo.WPAenabled
> > > -	    || adapter->secinfo.WPA2enabled) {
> > > -		dwrq->flags &= ~IW_ENCODE_DISABLED;
> > > -	} else {
> > > -		dwrq->flags |= IW_ENCODE_DISABLED;
> > > -	}
> > > -
> > >  	memset(extra, 0, 16);
> > >  
> > >  	mutex_lock(&adapter->lock);
> 
> 
> And since I don't see anywhere that we'd copy a WPA key into the
> response, I figured it made sense to keep IW_ENCODE_NOKEY in this bit
> too: 
> 
> > > @@ -1150,14 +1142,13 @@ static int lbs_get_encode(struct net_device *dev,
> > >  		   || (adapter->secinfo.WPA2enabled)) {
> > >  		/* return WPA enabled */
> > >  		dwrq->flags &= ~IW_ENCODE_DISABLED;
> > > +		dwrq->flags |= IW_ENCODE_NOKEY;
> > >  	} else {
> > >  		dwrq->flags |= IW_ENCODE_DISABLED;
> > >  	}
> 
> I'd want to set up an AP with WPA and test it if I was going to actually
> make it find and return the WPA key.

You could, but there's not too much of a point in spending time on this
unless you've got a really good reason?  I'm sure there are more
important things to do in the driver...

Dan