[PATCH v6 12/13] leds: flash: mt6370: Add MediaTek MT6370 flashlight support
Andy Shevchenko
andy.shevchenko at gmail.com
Mon Jul 25 23:10:55 PDT 2022
On Tue, Jul 26, 2022 at 6:15 AM szuni chen <szunichen at gmail.com> wrote:
...
> > > +#define MT6370_ITORCH_MIN_UA 25000
> > > +#define MT6370_ITORCH_STEP_UA 12500
> > > +#define MT6370_ITORCH_MAX_UA 400000
> > > +#define MT6370_ITORCH_DOUBLE_MAX_UA 800000
> > > +#define MT6370_ISTRB_MIN_UA 50000
> > > +#define MT6370_ISTRB_STEP_UA 12500
> > > +#define MT6370_ISTRB_MAX_UA 1500000
> > > +#define MT6370_ISTRB_DOUBLE_MAX_UA 3000000
> >
> > Perhaps _uA would be better and consistent across your series
> > regarding current units.
>
> Yes, _uA will be more consistent, but in general, using upper case in
> the define macro is a convention, doesn't it?
There is general convention, but there are also:
1) common sense;
2) usage in practice (e.g. _US, etc for *seconds and _HZ for *frequency).
My common sense tells me that it is convenient to use mA,uA, etc.
Plus "in practice" it's related to use as in your series and elsewhere.
But of course it's minor to me, decide yourself.
...
> > > + /*
> > > + * For the flash to turn on/off, need to wait for HW ramping up/down time
> >
> > we need
> >
> > > + * 5ms/500us to prevent the unexpected problem.
> > > + */
> > > + if (!prev && curr)
> > > + usleep_range(5000, 6000);
> > > + else if (prev && !curr)
> > > + udelay(500);
> >
> > This still remains unanswered, why in the first place we allow
> > switching, and a busy loop in the other place?
>
> If I refine the description to
> "For the flash to turn on/off, need to wait for 5ms/500us analog settling time.
> If any flash led is already used, then the analog is settled done, we
> don't need to wait again."
> is it answer the question?
No. I'm talking from the Linux APIs perspective. There is a huge
difference between those branches. Please, conduct research, read
documentation to understand what my question is about.
--
With Best Regards,
Andy Shevchenko
More information about the Linux-mediatek
mailing list