Re: linux-next: Tree for June 3 (rfkill)

[Johannes Berg]

On Wed, 2009-06-03 at 17:47 +0200, GÃbor Stefanik wrote:

> >> CFG80211=y
> >> MAC80211=y
> >> RFKILL=m
> >>
> >> net/built-in.o: In function `cfg80211_netdev_notifier_call':
> >> core.c:(.text+0xa678b): undefined reference to `rfkill_blocked'
> >> net/built-in.o: In function `cfg80211_dev_free':
> >
> > Hrm. I thought
> >
> > config CFG80211
> >        tristate "Improved wireless configuration API"
> >        depends on RFKILL || !RFKILL
> >
> > would avoid that. Why doesn't it?
> >
> > johannes
> >
> 
> Maybe the "y" state of CFG80211 specifically needs to depend on
> RFKILL=y || !RFKILL.

Maybe doesn't help me at all.

> BTW should CFG80211=y really be blocked when RFKILL=m? 

Yes.

> Shouldn't we
> just disable CFG80211 RFKILL support in this case (perhaps via a
> separate CONFIG_CFG80211_RFKILL automatically configured depending on
> CONFIG_RFKILL)?

That would be immensely stupid. You might as well just make cfg80211 a
module then.

johannes

Attachment: signature.asc
Description: This is a digitally signed message part