Re: [RFC PATCH 4/5] regulator: add 88pm88x regulators driver

[Karel Balej]

On Sun Jan 7, 2024 at 11:35 AM CET, Krzysztof Kozlowski wrote:
> On 28/12/2023 10:39, Karel Balej wrote:
> > diff --git a/drivers/mfd/88pm88x.c b/drivers/mfd/88pm88x.c
> > index 69a8e39d43b3..999d0539b720 100644
> > --- a/drivers/mfd/88pm88x.c
> > +++ b/drivers/mfd/88pm88x.c
> > @@ -68,6 +68,21 @@ static struct mfd_cell pm886_devs[] = {
> >  		.num_resources = ARRAY_SIZE(pm88x_onkey_resources),
> >  		.resources = pm88x_onkey_resources,
> >  	},
> > +	{
> > +		.name = "88pm88x-regulator",
> > +		.id = PM88X_REGULATOR_ID_LDO2,
> > +		.of_compatible = "marvell,88pm88x-regulator",
> > +	},
> > +	{
> > +		.name = "88pm88x-regulator",
> > +		.id = PM88X_REGULATOR_ID_LDO15,
> > +		.of_compatible = "marvell,88pm88x-regulator",
> > +	},
> > +	{
> > +		.name = "88pm88x-regulator",
> > +		.id = PM886_REGULATOR_ID_BUCK2,
> > +		.of_compatible = "marvell,88pm88x-regulator",
>
> Same compatible per each regulator looks suspicious, if not even wrong.
> What are these?

The original attempt for upstreaming this MFD had a different compatible
for each regulator which was not correct according to the reviewers at
the time. I have thus used the same compatible for all regulators and
make the distinction in the regulator driver (using the .id property).
But I think that the problem here is again that I have confused the
purpose of .name and .of_compatible properties of struct mfd_cell - if a
driver is probed due to the .name property then I indeed should not need
compatible for the regulator driver at all.

>
> Best regards,
> Krzysztof

Best regards,
K. B.