Re: [PATCH] hwmon: (max77705) add initial support

[Guenter Roeck]

On 3/4/25 02:03, Dzmitry Sankouski wrote:
> ср, 26 февр. 2025 г. в 16:53, Guenter Roeck <linux@xxxxxxxxxxxx>:
> > On 2/25/25 11:11, Dzmitry Sankouski wrote:
> > > Add support for max77705 hwmon. Includes charger input, system bus, and
> > > vbyp measurements.
> > >
> > > Signed-off-by: Dzmitry Sankouski <dsankouski@xxxxxxxxx>
> > > ---
> > > Maxim MAX77705 is a Companion Power Management and Type-C interface IC.
> > > It includes charger and fuel gauge blocks, and is capable of measuring
> > > charger input current, system bus volatage and current, and bypass
> > > voltage.
> > >
> > > This patch add support for mentioned measurements.
> (...)
> > > +
> > > +struct channel_desc {
> > > +     u8 reg;
> > > +     u8 avg_reg;
> > > +     const char *const label;
> > > +     // register resolution. nano Volts for voltage, nano Amperes for current
> > > +     u64 resolution;
> >
> > Why u64 ?
>
> Resolution is in nano Volts, and VBYP value is typically 5v, which leads to
> u32 overflow.

That is not a valid reason. The resolution itself does not overflow.
The conversion to 64-bit calculations can be done there.

> (...)
> > This is unusual for an mfd driver. Why not pass regmap from the parent
> > driver ?
>
> This limits access to shared device register map by using readable table
> with related registers only.
I do not see that as valid reason to implement separate regmaps. That
only complicates the code.

Guenter