Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
MIME-Version: 1.0
References: <20221202113226.114465-1-Mr.Bossman075@gmail.com>
 <20230119175135.GA2085792-robh@kernel.org> <Y8mt6ZCMf4YZvDYA@google.com>
In-Reply-To: <Y8mt6ZCMf4YZvDYA@google.com>
From:   Rob Herring <robh@kernel.org>
Date:   Fri, 20 Jan 2023 08:14:45 -0600
Message-ID: <CAL_JsqJQd4W5-8ej48hAebvyA6neH=2hTtzVU5HhFpQ2yKsQFw@mail.gmail.com>
Subject: Re: [PATCH v2] drivers/mfd: simple-mfd-i2c: Add generic compatible
To:     Lee Jones <lee@kernel.org>
Cc:     Jesse Taube <mr.bossman075@gmail.com>,
        linux-kernel@vger.kernel.org, linux-input@vger.kernel.org,
        lee.jones@linaro.org
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Thu, Jan 19, 2023 at 2:54 PM Lee Jones <lee@kernel.org> wrote:
>
> On Thu, 19 Jan 2023, Rob Herring wrote:
>
> > On Fri, Dec 02, 2022 at 06:32:26AM -0500, Jesse Taube wrote:
> > > Some devices may want to use this driver without having a specific
> > > compatible string. Add a generic compatible string to allow this.
> >
> > What devices need this?
> >
> > Is that no specific compatible string at all or just in the kernel?
> > Because the former definitely goes against DT requirements. The latter
> > enables the former without a schema.
> >
> > >
> > > Signed-off-by: Jesse Taube <Mr.Bossman075@gmail.com>
> > > ---
> > >  drivers/mfd/simple-mfd-i2c.c | 1 +
> > >  1 file changed, 1 insertion(+)
> > >
> > > diff --git a/drivers/mfd/simple-mfd-i2c.c b/drivers/mfd/simple-mfd-i2c.c
> > > index f4c8fc3ee463..0bda0dd9276e 100644
> > > --- a/drivers/mfd/simple-mfd-i2c.c
> > > +++ b/drivers/mfd/simple-mfd-i2c.c
> > > @@ -73,6 +73,7 @@ static const struct simple_mfd_data silergy_sy7636a = {
> > >  };
> > >
> > >  static const struct of_device_id simple_mfd_i2c_of_match[] = {
> > > +   { .compatible = "simple-mfd-i2c-generic" },
> >
> > Simple and generic? There is no such device. Anywhere.
> >
> > This is also not documented which is how I found it (make
> > dt_compatible_check). But this should be reverted or dropped rather than
> > documented IMO.
>
> I thought it would be better than having a huge list here.

What indication is there that the list would be huge? We have 2 out of
137 MFD bindings. Usually if the MFD is simple, we'd make it a single
node. It just needs to be clear what the conditions are for using it.

> Devices should *also* be allocated a specific compatible string.
>
> $ git grep simple-mfd -- arch

Why can't simple-mfd be used here?

Rob