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;
Message-ID: <feafad6d-736d-27a5-380d-e8a531b4c46b@rasmusvillemoes.dk>
Date:   Mon, 14 Nov 2022 09:10:54 +0100
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:102.0) Gecko/20100101
 Thunderbird/102.2.2
Subject: Re: [PATCH 2/5] dt-bindings: iio: ad74413r: make refin-supply
 optional
Content-Language: en-US, da
To:     Jonathan Cameron <jic23@kernel.org>
Cc:     Cosmin Tanislav <cosmin.tanislav@analog.com>,
        Lars-Peter Clausen <lars@metafoo.de>,
        Michael Hennerich <Michael.Hennerich@analog.com>,
        devicetree@vger.kernel.org, Rob Herring <robh+dt@kernel.org>,
        linux-iio@vger.kernel.org, linux-kernel@vger.kernel.org
References: <20221111143921.742194-1-linux@rasmusvillemoes.dk>
 <20221111143921.742194-3-linux@rasmusvillemoes.dk>
 <20221112165448.17489ef1@jic23-huawei>
From:   Rasmus Villemoes <linux@rasmusvillemoes.dk>
In-Reply-To: <20221112165448.17489ef1@jic23-huawei>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit
Precedence: bulk

On 12/11/2022 17.54, Jonathan Cameron wrote:
> On Fri, 11 Nov 2022 15:39:18 +0100
> Rasmus Villemoes <linux@rasmusvillemoes.dk> wrote:
> 
>> The ad74412r/ad74413r has an internal 2.5V reference output, which (by
>> tying the REFOUT pin to the REFIN pin) can be used in lieu of an
>> external 2.5V input reference. So stop marking refin-supply as
>> required.
>>
>> Signed-off-by: Rasmus Villemoes <linux@rasmusvillemoes.dk>
> 
> 
> Interesting corner case.  Given we have no way of knowing if the
> wiring has REFOUT connected to REFIN I see two ways we should fix this.
> 
> 1) Just have any DT doing this provide a fixed regulator.
> 2) Have the REFOUT supported as an actual regulator - in theory it might be
>    wired to other devices.  This might get a little interesting ordering
>    wise as we'll want to register the regulator before we try to consume
>    it in the same driver.  I'm also not 100% sure there are no other issues
>    in a driver providing and consuming the same regulator.

Hm, I don't like the idea of exposing REFOUT as a real regulator. As you
write, there's gonna be interesting chicken-and-egg problems, and I also
don't think it can actually deliver any meaningful current, i.e. it
can't really (and shouldn't) be used for supplying other peripherals.

A third option is to have a boolean property to explicitly indicate that
"yes, we're using refout as refin", and then make the requirement in the
schema be "refin-supply XOR refout-as-refin".

But I think the simplest is (1), I will just add a fixed-regulator with
a suitable comment in my .dts, and patches 2,3 can be ignored.

Thanks,
Rasmus