On Thu, Jul 18, 2024 at 10:53:19AM -0700, Guenter Roeck wrote:
> On 7/18/24 10:48, Guenter Roeck wrote:
> > On 7/18/24 09:39, Tzung-Bi Shih wrote:
> > > On Wed, Jul 17, 2024 at 08:39:31PM -0700, Guenter Roeck wrote:
> > > > @@ -471,10 +472,7 @@ static ssize_t update_interval_store(struct device *dev,
> > > > if (ret < 0)
> > > > return ret;
> > > > - for (regval = 0; regval < 3; regval++) {
> > > > - if (val <= update_intervals[regval])
> > > > - break;
> > > > - }
> > > > + regval = find_closest(val, update_intervals, ARRAY_SIZE(update_intervals));
> > >
> > > The behavior changed.
> > >
> > > static u16 update_intervals[] = { 143, 364, 1000, 2500 };
> > >
> > > If val = 144,
> > > * Originally, regval = 1.
> > > * After applying the patch, regval = 0.
> > >
> >
> >
> > Yes, find_closest() rounds the value instead of using the lower match.
> > That was intentional. I'll add an explicit note to the commit message.
> >
>
> I added this to the commit message:
>
> Since find_closest() uses rounding to find the best match, the resulting
> update interval will now reflect the update interval that is closest to
> the requested value, not the value that is lower or equal to the requested
> value.
With that,
Reviewed-by: Tzung-Bi Shih <tzungbi@xxxxxxxxxx>
