Re: idea: WWAN enable/disable

[Dan Williams <dcbw redhat com>]

On Tue, 2009-06-30 at 17:20 -0700, Marcel Holtmann wrote:
> Hi Dan,
> 
> > > > > it is pretty simple actually. Check include/linux/rfkill.h and it is
> > > > > just poll, read and write. Only one ioctl for rfkill-input replacement
> > > > > in the future, but that is unimportant and will go away. That is it.
> > > > > 
> > > > > http://git.sipsolutions.net/gitweb.cgi?p=rfkill.git;a=summary
> > > > > 
> > > > > This contains a really simple rfkill userspace tool.
> > > > 
> > > > Looks good. Hopefully one of Dan or myself will pick it up and wrap it
> > > > in GObject goodness to share in NM and gnome-bluetooth.
> > > 
> > > FYI, filed this for gnome-bluetooth:
> > > http://bugzilla.gnome.org/show_bug.cgi?id=585765
> > 
> > I'm vaguely hoping I don't have to care about rfkill too much, because
> > in the New World with the kernel rfkill goodness in 2.6.31, since there
> > are a lot of rfkill methods [ wifi, wwan, wimax, bt ] and since Fn+F5
> > (ie, input-only kill) needs to be handled too, maybe somebody will write
> > an app that will just handle the rfkill problem, and NM can simply
> > listen for events and do the right thing.  But not actually ever have to
> > *set* rfkill on.
> 
> that is exactly the idea. However if you wanna offer a WiFi off switch
> for example then just calling OP_CHANGE_ALL with TYPE_WIFI is all you
> need to do.

Right.

> In the end we need a proper application that integrates with GNOME and
> handles RFKILL soft keys and its interaction with user etc. Similar to
> what Lenovo offers under Windows. This might be a good thing for next
> GSoC if nobody else picks it up.

Yeah, though we probably want it before then.

Dan