Re: [PATCH 2/2] regulator: rt5759: Add support for Richtek RT5759 DCDC converter

[Krzysztof Kozlowski <krzk@kernel.org>]

On 26/03/2022 08:55, ChiYuan Huang wrote:
> Mark Brown <broonie@kernel.org> 於 2022年3月26日 週六 上午9:07寫道：
>>
>> On Sat, Mar 26, 2022 at 08:58:47AM +0800, ChiYuan Huang wrote:
>>
>>> I tried to remove only __maybe_unused and build with x86 config  that
>>> CONFIG_OF=n.
>>> There's no warning or error message when compiling the rt5759 source code.
>>
>>> If so, I will remove only '__maybe_unused'.
>>> May I ask whether 'of_match_ptr'  need to be added or not?
>>
>> If you add of_match_ptr() (which is a little better, though it's
>> a tiny different either way) then you shouldn't remove
>> __maybe_unused - the thing here is that the __maybe_unused is
>> needed because of_match_ptr() is used.
> 
> Sorry, I'm confused.
> Originally, Krzysztof's opinion is to tell me why 'of_device_id' is
> declared as '__maybe_unused'.
> That's why I mentioned that the return value  about of_device_get_match_data'

Your answer is not related to my question. of_device_get_match_data()
has nothing to do with __maybe_unused...

> And now we're talking about to add 'of_match_ptr' in struct driver
> of_match_table.

Because this is one of the solutions.

> 
> Back to the original topic, two ways can solve this.
> 1) only remove '__maybe_unused' in of_device_id
> 2) keep '__maybe_unused' in of_device_id, and add of_match_ptr for
> of_match_table.
> But option 2 seems conflict with Krzysztof's gueestion.
> 
> May I ask which option you suggested?

Option two does not conflict my suggestion. I pointed out that having
ONLY maybe_unused is incorrect. I pointed the mistake. Nothing more... I
said that there are two ways to solve it later, just choose one. I don't
know why do we talk about such basic issue for so long. This should be
one email from my side and one confirmation from you...

Obviously maybe_unused it has to be removed if you do not add
of_match_ptr. But if you intend to add of_match_ptr, then things change...

Just for the record of choosing between options (which I also mentioned
that there are two solutions) - having no of_match_ptr allows to match
with ACPI PRP0001 (AFAIU also when !OF).


Best regards,
Krzysztof