[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <ab853605-859d-44c6-8cbd-44391cd677e6@gmail.com>
Date: Wed, 5 Feb 2025 18:58:25 +0100
From: Artur Weber <aweber.kernel@...il.com>
To: Krzysztof Kozlowski <krzk@...nel.org>
Cc: Lee Jones <lee@...nel.org>, Rob Herring <robh@...nel.org>,
Krzysztof Kozlowski <krzk+dt@...nel.org>, Conor Dooley
<conor+dt@...nel.org>, Liam Girdwood <lgirdwood@...il.com>,
Mark Brown <broonie@...nel.org>,
Florian Fainelli <florian.fainelli@...adcom.com>, Ray Jui
<rjui@...adcom.com>, Scott Branden <sbranden@...adcom.com>,
Broadcom internal kernel review list
<bcm-kernel-feedback-list@...adcom.com>,
Stanislav Jakubek <stano.jakubek@...il.com>, devicetree@...r.kernel.org,
linux-kernel@...r.kernel.org, linux-arm-kernel@...ts.infradead.org,
~postmarketos/upstreaming@...ts.sr.ht
Subject: Re: [PATCH v3 2/7] dt-bindings: mfd: brcm,bcm59056: Add compatible
for BCM59054
On 2.02.2025 14:24, Krzysztof Kozlowski wrote:
> On Fri, Jan 31, 2025 at 07:13:50PM +0100, Artur Weber wrote:
>> ...
>> @@ -22,7 +24,6 @@ properties:
>> regulators:
>> type: object
>> description: Container node for regulators.
>> - $ref: ../regulator/brcm,bcm59056.yaml
>
> Refs should rather stay here, so I don't think keeping these devices in
> one binding makes it simpler.
>
> Simpler - drop ref and add properties compatible with enum for your
> regulator compatibles.
>
The only problem with that is that the regulator nodes have no
compatibles. I see two ways out of it: either we add a compatible then
use the compatible enum (might break existing users who don't have the
compatible, unless we make the driver ignore it), or if we just want to
simplify but not add compatibles, list all the refs under an oneOf
condition (like it's done in
Documentation/devicetree/bindings/mfd/qcom,spmi-pmic.yaml[1]).
Personally I'm in favor of keeping this as-is, though, for the benefit
of the DT linter preventing accidental mixing of bindings for different
device types (e.g. BCM59056 MFD node with BCM59054 regulators). There's
at least some precedent for this method[2]. But if making it simple is
what's considered better, I'll go with one of the aforementioned
options.
Best regards
Artur
[1] https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git/tree/Documentation/devicetree/bindings/mfd/qcom,spmi-pmic.yaml#n147
[2] https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git/tree/Documentation/devicetree/bindings/mfd/samsung,s2mps11.yaml#n69
Powered by blists - more mailing lists