Re: [PATCH V4 2/6] dt-bindings: regulator: Add pm8008 regulator bindings

From: "Satya Priya Kakitapalli (Temp)" <quic_c_skakit@quicinc.com>
To: Mark Brown <broonie@kernel.org>
Cc: Bjorn Andersson <bjorn.andersson@linaro.org>,
	Rob Herring <robh+dt@kernel.org>,
	Liam Girdwood <lgirdwood@gmail.com>, <swboyd@chromium.org>,
	<collinsd@codeaurora.org>, <subbaram@codeaurora.org>,
	Das Srinagesh <gurus@codeaurora.org>,
	<linux-arm-msm@vger.kernel.org>,
	"Lee Jones" <lee.jones@linaro.org>, <devicetree@vger.kernel.org>,
	<linux-kernel@vger.kernel.org>
Subject: Re: [PATCH V4 2/6] dt-bindings: regulator: Add pm8008 regulator bindings
Date: Mon, 20 Dec 2021 18:14:00 +0530	[thread overview]
Message-ID: <30ec6b4c-f2a8-d80e-a542-1c2b3f30c049@quicinc.com> (raw)
In-Reply-To: <Ya4UcxxEq9t+isxS@sirena.org.uk>

On 12/6/2021 7:17 PM, Mark Brown wrote:
> On Mon, Dec 06, 2021 at 07:13:02PM +0530, Satya Priya Kakitapalli (Temp) wrote:
>> On 11/25/2021 8:54 PM, Mark Brown wrote:
>>> On Fri, Nov 19, 2021 at 03:12:29PM +0530, Satya Priya wrote:
>>>> +properties:
>>>> +  compatible:
>>>> +    const: qcom,pm8008-regulators
>>> Why are we adding a separate compatible for this when we already know
>>> that this is a pm8008 based on the parent?
>> For the regulator driver to be probed we do need a separate compatible
>> right? may be I didn't get your question..
>> My understanding is we should have a separate compatible for each peripheral
>> under the parent mfd node.. like gpios, temp alarm, regulators etc..
> No, the MFD can register whatever children it likes without needing any
> help from the DT.

I think this is possible by using of_platform_bus_probe() API. But, the 
mfd driver uses of_platform_populate() API, this needs all device nodes 
to have a 'compatible' property unlike the of_platform_bus_probe() API.

All other MFD upstream drivers are also using the same API and 
registering the child regulators by using separate compatible strings.

>>>> +  vdd_l1_l2-supply:
>>>> +    description: Input supply phandle of ldo1 and ldo2 regulators.
>>> These supply nodes should be chip level, they're going into the chip and
>>> in general the expectation is that you should be able to describe the
>>> supplies going into a device without worrying about how or if any
>>> particular OS splits things up.
>> So, if i understand correctly, we don't have to mention these in the
>> documentation as these are handled at framework level?
> No.  I'm saying you should document these at the chip level, they do
> need to be documented though.

By chip level do you mean "pm8008.yaml" documentation? If so, yes, I can 
move these to pm8008.yaml and change DT accordingly.