On Tue, 28 May 2024 at 11:43, Oliver Neukum <oneukum@xxxxxxxx> wrote:
>
> On 28.05.24 01:15, Dmitry Baryshkov wrote:
> > On Mon, May 27, 2024 at 02:26:36PM +0200, Oliver Neukum wrote:
> >> On 27.05.24 12:03, Dmitry Baryshkov wrote:
>
> Hi,
>
> >>> +struct yoga_c630_psy {
> >>> + struct yoga_c630_ec *ec;
> >>> + struct device *dev;
> >>> + struct device_node *of_node;
> >>> + struct notifier_block nb;
> >>> + struct mutex lock;
> >>> +
> >>> + struct power_supply *adp_psy;
> >>> + struct power_supply *bat_psy;
> >>> +
> >>> + unsigned long last_status_update;
> >>> +
> >>> + bool adapter_online;
> >>> +
> >>> + bool unit_mA;
> >>> +
> >>> + unsigned int scale;
> >>
> >> why do you store unit_mA and scale? This looks redundant and like a source
> >> of confusion to me.
> >
> > Here we just followed the AML code in ACPI tables. The unit_mA is a
> > returned from the_BIX method, the 'scale' is used internally in the DSDT.
> > If you think that it's better, I can change all '* scale * 1000' to
> > 'if unit_mA then foo = bar * 10000 else foo = bar * 1000'.
>
> I think that would indeed be better. Implementation details of the DSDT
> should not dictate data structures in a kernel driver.
Ack.
>
> Regards
> Oliver
>
--
With best wishes
Dmitry
