From mboxrd@z Thu Jan  1 00:00:00 1970
From: dmitry.torokhov@gmail.com (Dmitry Torokhov)
Date: Tue, 5 Oct 2010 09:16:08 -0700
Subject: [PATCH v2] Input: Make ADS7846 independent on regulator
In-Reply-To: <4CAACA63.6090202@compulab.co.il>
References: <1280560182-7071-1-git-send-email-marek.vasut@gmail.com>
	<4C862EB3.4080006@compulab.co.il>
	<20100907125335.GD25830@sirena.org.uk>
	<201009091027.18176.marek.vasut@gmail.com>
	<20100909094120.GB1988@rakim.wolfsonmicro.main>
	<4CAACA63.6090202@compulab.co.il>
Message-ID: <20101005161608.GD19730@core.coreip.homeip.net>
To: linux-arm-kernel@lists.infradead.org
List-Id: linux-arm-kernel.lists.infradead.org

On Tue, Oct 05, 2010 at 08:49:07AM +0200, Igor Grinberg wrote:
>  On 09/09/10 12:41, Mark Brown wrote:
> > On Thu, Sep 09, 2010 at 10:27:17AM +0200, Marek Vasut wrote:
> >> Dne ?t 7. z??? 2010 14:53:35 Mark Brown napsal(a):
> >>> From a regulator API usage point of view a separate implementation of
> >>> the same thing was nacked - there are regulator API facilties for hiding
> >>> missing regulators from drivers when needed to get systems going, unless
> >>> the device genuinely can cope without supplies it should be relying on
> >>> those.
> 
> I actually, don't see why ads7846 is strictly relying on the regulator
> and I don't understand, why ads7846 driver has to bail out if the regulator
> is not found? Why shouldn't the driver try to continue?
> I think it should bail out only in case communicating with the device failed.
> 
> >> Maybe these platforms should have been fixed prior to applying the patch adding 
> >> regulator goo into ads7846 driver then. What's the way to go now then ?
> > Fix the platforms and use the dummy regulators to keep them going until
> > that happens.  It's trivial to do the hookup in the platforms.
> 
> You want each platform, that does not have a special regulated power supply
> for the ads7846, to define a dummy regulator just to cope with that artificial
> dependency of the device driver?
> I think it is a waste and big code duplication in each platform
> that does not have that special regulator.
>

I tend to agree, however I think that original patch that simply ignored
failures from regulator_get() is not the best option either. Can we have
a flag in platform data indicating that the board does not employ a
regulator? Then we could retain the hard failure in cases when we expect
regulator to be present while allowing to continue on boards that do not
have it.

-- 
Dmitry