Re: [patch 20/35] leds: route kbd LEDs through the generic LEDs layer

[Samuel Thibault <samuel.thibault@ens-lyon.org>]

Pavel Machek, le Mon 06 Feb 2012 15:19:23 +0100, a écrit :
> On Mon 2011-11-14 05:06:13, Samuel Thibault wrote:
> > Dmitry Torokhov, le Wed 12 Jan 2011 10:27:02 -0800, a écrit :
> > > > --- a/drivers/char/Kconfig
> > > > +++ b/drivers/char/Kconfig
> > > > @@ -8,6 +8,7 @@ config VT
> > > >  	bool "Virtual terminal" if EMBEDDED
> > > >  	depends on !S390
> > > >  	select INPUT
> > > > +	select LEDS_INPUT
> > > 
> > > This will not work as LEDS_INPUT is itself depends and selects other
> > > symbols and "select" statement does not propagate dependencies.
> > > 
> > > Anyway, I was looking at the patch this way and that and I convinced
> > > myself that while input should be integrated with LED subsystem doing it
> > > via an input handler is not the right way. Input handler is, by it's
> > > nature, an optional operation or an interface that exists to transform
> > > the input core data stream into some other form. Here OTOH we have a
> > > piece of hardware that we want to manage. I think the proper conversion
> > > would involve attaching led_classdev structures directly to input_dev
> > > objects and controlling them at the core level.
> > 
> > Here is (at last!) an updated & tested patch. It actually makes things
> > simpler, which is a good sign :)
> > 
> > I however don't know how to fix the issue about VT selecting
> > INPUT_LEDS.
> 
> Any news about this one? Having keyboard leds in the common framework
> seems very welcome to me.	

Do you perhaps have an idea about the issue of VT selecting INPUT_LEDS?

Samuel