Re: [RFC PATCH] led: add Cycle LED trigger.

From: Pavel Machek <pavel@ucw.cz>
To: "Gaël PORTAY" <g.portay@overkiz.com>,
	"Rob Landley" <rob@landley.net>, "Bryan Wu" <cooloney@gmail.com>,
	"Richard Purdie" <rpurdie@rpsys.net>,
	"Milo(Woogyom) Kim" <milo.kim@ti.com>,
	linux-doc@vger.kernel.org, linux-kernel@vger.kernel.org,
	linux-leds@vger.kernel.org
Subject: Re: [RFC PATCH] led: add Cycle LED trigger.
Date: Sun, 23 Jun 2013 11:53:07 +0200	[thread overview]
Message-ID: <20130623095305.GA3669@amd.pavel.ucw.cz> (raw)
In-Reply-To: <20130622164300.GB19420@earth.universe>

On Sat 2013-06-22 18:43:01, Sebastian Reichel wrote:
> On Sat, Jun 22, 2013 at 01:26:20PM +0200, Pavel Machek wrote:
> > On Tue 2013-06-18 18:24:23, Gaël PORTAY wrote:
> > > Currently, none of available triggers supports playing with the LED brightness
> > > level.  The cycle trigger provides a way to define custom brightness cycle.
> > > For example, it is easy to customize the cycle to mock up the rhythm of human
> > > breathing which is a nice cycle to tell the user the system is doing something.
> > > 
> > > This trigger is meant to be usable for waiting an event to happen, for example
> > > when the system gets ready.  Those cycles may be used to reflect well known
> > > system status (e.g. idle mode, startup...).
> > > 
> > > This implementation provides several interfaces:
> > >  - to define the cycle itself:
> > >    * plot: definition of plot points using plot or rawplot files,
> > >            each points defines the brightness level
> > >    * interval: constant time interval between each plot point
> > >  - to control the cycle:
> > >    * repeat: the number of repetition of the whole plot cycle
> > >              0 for an infinite loop
> > >    * control: used to control the cycle trigger
> > >      + "start"/"stop": to start/stop the cycle
> > >      + "reset" to clear the cycle counter and the internal plot point index
> > >      + "pause"/"resume" to pause/resume the cycle
> > > 
> > > Signed-off-by: Gaël PORTAY <g.portay@overkiz.com>
> > 
> > I'd say this should go to userspace.... and maybe should handle RGB
> > leds. ... like the one on n900/HTC dream/....
> > 
> > Actually, there's probably some daemon in maemo that already does
> > this.
> 
> Actually the n900 has hardware support for this. There's a
> programmable LED driver on the board, which is called LP5523.
> 
> So... I don't think there's a daemon in maemo ;)

Uh. It seems that LP5523 should get gcc support, not kernel
support. It has limited memory, but 3 cores, and seems turing
complete.

Not sure what is reasonable interface for that, but I guess subset
that "makes sense" is enough.
									Pavel
-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html