linux-kernel - Re: [PATCH 3/4] backlight: add led-backlight driver

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Fri, 5 Jul 2019 12:08:51 +0200
From:   Pavel Machek <pavel@....cz>
To:     Jean-Jacques Hiblot <jjhiblot@...com>
Cc:     Daniel Thompson <daniel.thompson@...aro.org>,
        jacek.anaszewski@...il.com, robh+dt@...nel.org,
        mark.rutland@....com, lee.jones@...aro.org, jingoohan1@...il.com,
        dmurphy@...com, linux-leds@...r.kernel.org,
        linux-kernel@...r.kernel.org, dri-devel@...ts.freedesktop.org,
        tomi.valkeinen@...com
Subject: Re: [PATCH 3/4] backlight: add led-backlight driver

Hi!

> > > > Also still relevant is whether the LED device is being correctly
> > > > modelled if the act of turning on the LED doesn't, in fact, turn the LED
> > > > on. Is it *really* a correct implementation of an LED device that
> > > > setting it to LED_FULL using sysfs doesn't cause it to light up?
> > > What I understood from the discussion between Rob and Tomi is that the
> > > child-node of the LED controller should be considered a backlight device,
> > > not a simple LED. I'm not sure if the sysfs interface is still relevant in
> > > that case. Maybe it should just be disabled by the backlight driver
> > > (possible with led_sysfs_disable())
> > led_sysfs_disable() sounds like a sensible change but that's not quite
> > what I mean.
> > 
> > It is more a thought experiment to see if the power control *should* be
> > implemented by the backlight. Consider what happens if we *don't*
> > enable CONFIG_BACKLIGHT_LED in the kernel: we would still have an LED
> > device and it would not work correctly.
> > 
> > In other words I naively expect turning on an LED using the LED API
> > (any of them, sysfs or kernel) to result in the LED turning on.
> > Implementing a workaround in the client for what appears to be
> > something missing in the LED driver strikes me as odd. Why shouldn't
> > the regulator be managed in the LED driver?
> 
> I see your point. Indeed having the regulator handled in the LED-core makes
> sense in a lot of situations
> 
> I'll think about it.

For the record, I also believe regulator and enable gpio should be
handled in the core.

									Pavel
PS please trim down the quoted text.									
-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html