Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Date:   Tue, 30 Jun 2020 23:56:09 +0200
From:   Miquel Raynal <miquel.raynal@bootlin.com>
To:     Andy Shevchenko <andy.shevchenko@gmail.com>
Cc:     Uwe =?UTF-8?B?S2xlaW5lLUvDtm5pZw==?= 
        <u.kleine-koenig@pengutronix.de>,
        Bartosz Golaszewski <bgolaszewski@baylibre.com>,
        Thierry Reding <thierry.reding@gmail.com>,
        Linus Walleij <linus.walleij@linaro.org>,
        linux-gpio <linux-gpio@vger.kernel.org>,
        linux-pwm@vger.kernel.org,
        Thomas Petazzoni <thomas.petazzoni@bootlin.com>,
        LKML <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH v6] gpio: pca953x: Add Maxim MAX7313 PWM support
Message-ID: <20200630235609.52611e41@xps13>
In-Reply-To: <CAHp75VcxQ-qK+RYmfULDx3cHQW9W8__-Gf8EFw7Hgqsz0oPkjA@mail.gmail.com>
References: <20200503105453.23658-1-miquel.raynal@bootlin.com>
        <20200629160844.6ecf79c1@xps13>
        <20200629195044.uvgma53cajiko3gf@pengutronix.de>
        <20200630085828.1aebdf99@xps13>
        <CAMpxmJUHy+5ynmsU6FhX=MJR6XbwoNr65NGqzuAd0Bm-JS1kWw@mail.gmail.com>
        <20200630124500.m765t276atb3wqgy@pengutronix.de>
        <CAHp75VcxQ-qK+RYmfULDx3cHQW9W8__-Gf8EFw7Hgqsz0oPkjA@mail.gmail.com>
Organization: Bootlin
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8BIT
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

Hello,

Andy Shevchenko <andy.shevchenko@gmail.com> wrote on Wed, 1 Jul 2020
00:27:31 +0300:

> On Tue, Jun 30, 2020 at 3:45 PM Uwe Kleine-König
> <u.kleine-koenig@pengutronix.de> wrote:
> > On Tue, Jun 30, 2020 at 11:13:31AM +0200, Bartosz Golaszewski wrote:
> > > I can't find anything in any of the previous threads. What was the
> > > reason to not go the MFD way here?
> >
> > MFD doesn't work because the different "modes" are not orthogonal. So
> > this is not a single device that provides several functions at the same
> > time that can be used in parallel.
> 
> Did I get this correctly that it's either *all* pins are GPIO or *all* are PWM?
> 
> Otherwise, it fits from my point of view.
> 

Sorry for not being clear in the first place : the choice between
pure GPIO and PWM is on a per output basis. As Uwe said, the two
"modes" are not orthogonal hence the MFD does not fit here IMHO.

Thanks,
Miquèl