> > Well, you were suggesting the use of multiple SSIDs, which we don't
> > support, we only support VLANs within a BSS/single SSID. Not that I've
> > been able to test it, hostapd needs radius stuff set up for VLANs...
>
> I think it's transparent since it's handled by MLME in hostapd so
> multiple SSID is supported
> Maybe we did some minor changes in mac to support that.. Will check again.
> I think it's always done by means of VLANs
Oh, hm, could be true.
> > Right. I was just saying that the way it currently is I think you could
> > detect it that way. b43 simply assumes WEP keys are always 'static'
> > which seems to mostly work well in practice.
> > I suppose then set_key needs a new argument key_type:
> >
> > enum ieee80211_key_type {
> > KEY_TYPE_PAIRWISE,
> > KEY_TYPE_GROUP,
> > KEY_TYPE_TXONLY, /* group key in an AP */
>
> Can we drop it? Hm still not sure why you like it so much.
Well, we don't want an AP to actually decrypt things, so we need to
distinguish between these things so that the driver doesn't somehow try
to use that key for decryption.
> > KEY_TYPE_STATIC,
> > }
> >
> > where the MAC address pointer would only be non-NULL when the key type
> > is PAIRWISE, and STATIC can only be used for WEP keys.
> >
>
>
> Do you know anything about mesh security are we breaking here anything?
No, Javier, any comments? I think basically you have peer links that are
encrypted, which is just pairwise keys.
johannes
Attachment:
signature.asc
Description: This is a digitally signed message part
