[PATCH 11/17] P2P: Allow additional operating channels for P2P GO and client

[Peer, Ilan]

Hi Jouni,

> -----Original Message-----
> From: Jouni Malinen [mailto:j at w1.fi]
> Sent: Saturday, August 15, 2015 00:33
> To: Peer, Ilan
> Cc: hostap at lists.shmoo.com
> Subject: Re: [PATCH 11/17] P2P: Allow additional operating channels for P2P
> GO and client
> 
> On Mon, Jul 27, 2015 at 10:24:28PM +0300, Ilan Peer wrote:
> > Enable using additional frequencies for P2P GO and P2P Client.
> 
> Have you considered the impacts of these changes to hostapd and
> wpa_supplicant non-P2P AP use cases?

Not really ...

> 
> > diff --git a/src/ap/hw_features.c b/src/ap/hw_features.c @@ -118,11
> > +118,16 @@ int hostapd_get_hw_features(struct hostapd_iface *iface)
> > +			} else if ((feature->channels[j].flag &
> > +				    HOSTAPD_CHAN_NO_IR) &&
> > +				   !(feature->channels[j].flag &
> > +				     (HOSTAPD_CHAN_INDOOR_ONLY |
> > +				      HOSTAPD_CHAN_GO_CONCURRENT))) {
> > +					feature->channels[j].flag |=
> > +						HOSTAPD_CHAN_DISABLED;
> >  			}
> 
> This would seem to leave no-IR channels enabled if they happen to have
> INDOOR_ONLY or GO_CONCURRENT flag. Why would INDOOR_ONLY allow no-
> IR to be ignored?

It should not its own. The implementation is in complete, i.e., buggy .. :)

There were no other changes in src/ap/* so this would
> change hostapd behavior if we had any NO-OUTDOOR + NO-IR channels.
> Though, the current wireless-regdb db.txt has no such combination. Nor does
> it have any sign of "GO concurrent" (IR_CONCURRENT in current nl80211.h).
> 

We did not intend to add this to wireless regdb. We are using this regulatory relaxation features with hw based regulatory.

> How is this supposed to be usable with the current wireless-regdb? Are there
> some pending changes to it (and CRDA?) to start using the new flags? If not,
> I'm not sure why hostapd/wpa_supplicant should be changed either in this
> way.
> 

I do not think that this should mandate changes in regdb, as this can be still used by hw based regulatory configuration.

> For me to take this change, I think would need to be able to run a hwsim test
> case that shows how this behaves, i.e., that wireless-regdb change would be
> needed..
> 

This can also be done with custom regdomain.

Anyway, this would need more work, so please drop these patches from your queue for now.

Thanks again,

Ilan.