Re: [PATCH v4 2/2] arm64: dts: rockchip: Add Radxa ROCK 4D device tree

From: "Heiko Stübner" <heiko@sntech.de>
To: Jonas Karlman <jonas@kwiboo.se>,
	Detlev Casanova <detlev.casanova@collabora.com>
Cc: Rob Herring <robh@kernel.org>,
	Krzysztof Kozlowski <krzk+dt@kernel.org>,
	Conor Dooley <conor+dt@kernel.org>,
	Chris Morgan <macromorgan@hotmail.com>,
	Kever Yang <kever.yang@rock-chips.com>,
	Dragan Simic <dsimic@manjaro.org>, Tim Lunn <tim@feathertop.org>,
	FUKAUMI Naoki <naoki@radxa.com>,
	Michael Riesch <michael.riesch@wolfvision.net>,
	Stephen Chen <stephen@radxa.com>,
	Elon Zhang <zhangzj@rock-chips.com>,
	Alexey Charkov <alchark@gmail.com>,
	devicetree@vger.kernel.org, linux-arm-kernel@lists.infradead.org,
	linux-rockchip@lists.infradead.org, kernel@collabora.com,
	linux-kernel@vger.kernel.org
Subject: Re: [PATCH v4 2/2] arm64: dts: rockchip: Add Radxa ROCK 4D device tree
Date: Mon, 17 Feb 2025 22:30:40 +0100	[thread overview]
Message-ID: <3568510.6YUMPnJmAY@diego> (raw)
In-Reply-To: <1914418.tdWV9SEqCh@earth>

Am Montag, 17. Februar 2025, 22:07:06 MEZ schrieb Detlev Casanova:
> On Monday, 17 February 2025 12:08:47 EST Jonas Karlman wrote:
> > On 2025-02-17 17:34, Detlev Casanova wrote:
> > > +	pmic@23 {
> > > +		compatible = "rockchip,rk806";
> > > +		reg = <0x23>;
> > > +
> > > +		gpio-controller;
> > > +
> > > +		interrupt-parent = <&gpio0>;
> > > +		interrupts = <6 IRQ_TYPE_LEVEL_LOW>;
> > > +
> > > +		pinctrl-names = "default";
> > > +		pinctrl-0 = <&pmic_pins
> > > +			     &rk806_dvs1_null
> > > +			     &rk806_dvs2_null
> > > +			     &rk806_dvs3_null>;
> > > +
> > > +		system-power-controller;
> > > +
> > > +		vcc1-supply = <&vcc_5v0_sys>;
> > > +		vcc2-supply = <&vcc_5v0_sys>;
> > > +		vcc3-supply = <&vcc_5v0_sys>;
> > > +		vcc4-supply = <&vcc_5v0_sys>;
> > > +		vcc5-supply = <&vcc_5v0_sys>;
> > > +		vcc6-supply = <&vcc_5v0_sys>;
> > > +		vcc7-supply = <&vcc_5v0_sys>;
> > > +		vcc8-supply = <&vcc_5v0_sys>;
> > > +		vcc9-supply = <&vcc_5v0_sys>;
> > > +		vcc10-supply = <&vcc_5v0_sys>;
> > > +		vcc11-supply = <&vcc_2v0_pldo_s3>;
> > > +		vcc12-supply = <&vcc_5v0_sys>;
> > > +		vcc13-supply = <&vcc_1v1_nldo_s3>;
> > > +		vcc14-supply = <&vcc_1v1_nldo_s3>;
> > > +		vcca-supply = <&vcc_5v0_sys>;
> > > +
> > > +		#gpio-cells = <2>;
> > 
> > This should probably be sorted next to gpio-controller.
> 
> It's not unusual to put # props at the end. but I can move it up if it is 
> preferred.

That is actually a multi-colored bikeshed ;-)

When sorting alphabetically, do you
- just ignore the "#", this would move #gpio-cells to gpio*  but also
  split up #address-cells and #size-cells
- count "#" as special character and move them to the bottom, but this
  would split #gpio-cells from gpio-controller

--- TL;DR
So far I've not managed to come with a 1-size-fits-all opinion, but for
#gpio* and #clock* properties, readability gets better when they are
together with other gpio* / clock* properties .
----

Also I think you could do away with all the empty lines between properties
above (pinctrl <-> system-power-controller, etc), but of course keep the
empty between the subnodes below :-)


Heiko