Re: NetworkManager should report killswitch state per device

[Dan Williams <dcbw redhat com>]

On Fri, 2008-04-25 at 07:45 +0200, Helmut Schaa wrote:
> Am Do 24 Apr 2008 19:24:33 CEST schrieb Dan Williams <dcbw redhat com>:
> 
> > On Thu, 2008-04-24 at 12:17 +0200, Helmut Schaa wrote:
> >> Hi,
> >>
> >> NetworkManager reports the current killswitch state as a global flag
> >> (WirelessHardwareEnabled). Shouldn't this state be tied to the wireless
> >> device instead?
> >
> > No, per type.  WLAN has a killswitch, WWAN has a killswitch, BT has a
> > killswitch.
> 
> Ok. So it is intended.
> 
> In which device state should the wireless devices be if wireless is  
> disabled using the killswitch? NM_DEVICE_STATE_UNAVAILABLE?

Yeah, and the manager's 'WirelessEnabled' property will be FALSE.
Basically, UNAVAILABLE means the device cannot be activated right now
due to hardware constraints, like rfkill or no cable plugged in.

The device will go from UNMANAGED to UNAVAILABLE when the system
settings service has had a chance to figure out of the device is managed
or not, and then from UNAVAILABLE to DISCONNECTED after the rfkill state
has been checked.

Dan