Re: [PATCH v4 05/14] iio: accel: adxl345: add single tap feature

[Jonathan Cameron <jic23@kernel.org>]

On Wed, 19 Mar 2025 00:08:24 +0100
Lothar Rubusch <l.rubusch@gmail.com> wrote:

> On Sun, Mar 16, 2025 at 12:22 PM Jonathan Cameron <jic23@kernel.org> wrote:
> >
> > On Thu, 13 Mar 2025 16:50:40 +0000
> > Lothar Rubusch <l.rubusch@gmail.com> wrote:
> >  
> > > Add the single tap feature with a threshold in 62.5mg/LSB points and a
> > > scaled duration in us. Keep singletap threshold in regmap cache but
> > > the scaled value of duration in us as member variable.
> > >
> > > Both use IIO channels for individual enable of the x/y/z axis. Initializes
> > > threshold and duration with reasonable content. When an interrupt is
> > > caught it will be pushed to the according IIO channel.
> > >  
> >  
> > > Signed-off-by: Lothar Rubusch <l.rubusch@gmail.com>  
> >
> > Hi Lothar,
> >
> > A few things in here are from the discussion that was continuing
> > on v3 so I may have said more replying to that.
> >
> > Anyhow, for now I'll hold off on applying from this point on as
> > a few more things to respond to inline.
> >
> > Jonathan
> >  
> > >
> > >  #include "adxl345.h"
> > > @@ -31,6 +33,33 @@
> > >  #define ADXL345_INT1                 0
> > >  #define ADXL345_INT2                 1
> > >
> > > +#define ADXL345_REG_TAP_AXIS_MSK     GENMASK(2, 0)
> > > +
> > > +enum adxl345_axis {
> > > +     ADXL345_Z_EN = BIT(0),
> > > +     ADXL345_Y_EN = BIT(1),
> > > +     ADXL345_X_EN = BIT(2),
> > > +     /* Suppress double tap detection if value > tap threshold */
> > > +     ADXL345_TAP_SUPPRESS = BIT(3),
> > > +};  
> > As per feedback (after you sent this!) on v3, I'd drop
> > the last value out of the enum, or just use defines and a u8 for
> > the one place this is used for local variable storage.
> >
> >  
> > > @@ -198,6 +387,132 @@ static int adxl345_write_raw(struct iio_dev *indio_dev,
> > >       return -EINVAL;
> > >  }
> > >
> > > +static int adxl345_read_event_config(struct iio_dev *indio_dev,
> > > +                                  const struct iio_chan_spec *chan,
> > > +                                  enum iio_event_type type,
> > > +                                  enum iio_event_direction dir)
> > > +{
> > > +     struct adxl345_state *st = iio_priv(indio_dev);
> > > +     bool int_en;
> > > +     int ret = -EFAULT;  
> > Not used?
> >  
> > > +
> > > +     switch (type) {
> > > +     case IIO_EV_TYPE_GESTURE:
> > > +             switch (dir) {
> > > +             case IIO_EV_DIR_SINGLETAP:
> > > +                     ret = adxl345_is_tap_en(st, chan->channel2,
> > > +                                             ADXL345_SINGLE_TAP, &int_en);
> > > +                     if (ret)
> > > +                             return ret;
> > > +                     return int_en;
> > > +             default:
> > > +                     return -EINVAL;
> > > +             }
> > > +     default:
> > > +             return -EINVAL;
> > > +     }
> > > +}  
> >  
> > > +static int adxl345_write_event_value(struct iio_dev *indio_dev,
> > > +                                  const struct iio_chan_spec *chan,
> > > +                                  enum iio_event_type type,
> > > +                                  enum iio_event_direction dir,
> > > +                                  enum iio_event_info info,
> > > +                                  int val, int val2)
> > > +{
> > > +     struct adxl345_state *st = iio_priv(indio_dev);
> > > +     int ret;
> > > +
> > > +     ret = adxl345_set_measure_en(st, false);
> > > +     if (ret)
> > > +             return ret;
> > > +  
> > So in my brief reply to the v3 discussion I suggested perhaps
> > factoring out everything from here...  
> > > +     switch (type) {
> > > +     case IIO_EV_TYPE_GESTURE:
> > > +             switch (info) {
> > > +             case IIO_EV_INFO_VALUE:
> > > +                     ret = regmap_write(st->regmap, ADXL345_REG_THRESH_TAP,
> > > +                                        min(val, 0xFF));
> > > +                     break;
> > > +             case IIO_EV_INFO_TIMEOUT:
> > > +                     ret = adxl345_set_tap_duration(st, val, val2);
> > > +                     break;
> > > +             default:
> > > +                     ret = -EINVAL;
> > > +                     break;
> > > +             }
> > > +             break;
> > > +     default:
> > > +             ret = -EINVAL;
> > > +             break;
> > > +     }  
> > to here, so as to allow simple direct returns.
> >
> > I think that will make the code more readable given the need to reenable
> > measurements and that you want to leave it off on error.
> >  
> 
> Sorry for replying again on this topic. Pls, find my solution in v5.
> 
> After some thinking, I implemented it now using returns directly leaving the
> measurement on/off as is. I'm unsure if it actually makes sense, after an error
> here to turn measurement on again? I can imagine a situation where a wrong
> input might result in an error. Nothing is changed, and measurement
> could/should continue. Now, it will probably stop, in case of wrong
> input. But is
> wrong input actually an issue here?

If a userspace control input is out of range etc, then returning an error
but leaving things on makes sense.  If what we see is a comms error
it gets less clear on what we should do.  

> 
> As other alternative, I can think of is to shift measurement on/off
> into the called
> functions directly. I think, this approach was used also in the
> ADXL380 and seems
> to be common. Let me know what you think.
> 
Moving it up or down a layer can work by allowing direct returns and always
trying to reenable if that makes sense.

Sadly in error handling there is often not a right answer on what to do!

Jonathan

> > > +
> > > +     if (ret)
> > > +             return ret; /* measurement stays off */
> > > +
> > > +     return adxl345_set_measure_en(st, true);
> > > +}  
> >