On Fri, Oct 20, 2023 at 11:18:44AM +0000, Balas, Eliza wrote:
> > > +static int adi_axi_tdd_parse_ms(struct adi_axi_tdd_state *st,
> > > + const char *buf,
> > > + u64 *res)
> > > +{
> > > + u64 clk_rate = READ_ONCE(st->clk.rate);
> > > + char *orig_str, *modf_str, *int_part, frac_part[7];
> > > + long ival, frac;
> > > + int ret;
> > > +
> > > + orig_str = kstrdup(buf, GFP_KERNEL);
> > > + int_part = strsep(&orig_str, ".");
> >
> > Why are we parsing floating point in the kernel? Please just keep the
> > api simple so that we don't have to try to do any type of parsing other
> > than turning a single text number into a value.
> >
>
> The adi_axi_tdd_parse_ms function does almost the same thing as the
> iio_str_to_fixpoint() function which already exists in kernel.
That does not mean that this is a valid api for your device as you are
not an iio driver (why aren't you an iio driver?)
> It parses a fixed-point number from a string.
And as such, you shouldn't duplicate existing logic.
> The __iio_str_to_fixpoint is used in a similar way, as I intend to use adi_axi_tdd_parse_ms.
> It writes to a channel the corresponding scale (micro,nano) for a value.
Why not just have the api accept values in nanoseconds and then no
parsing is needed?
> Since the device is not an iio device, using an iio function would be confusing.
Why isn't this an iio device?
thanks,
greg k-h
