Messages in this thread |  | | Date | Thu, 27 Nov 2025 11:13:28 +0200 | | From | Andy Shevchenko <> | | Subject | Re: [PATCH v2 7/9] iio: adc: ad4062: Add IIO Events support |
| |
On Wed, Nov 26, 2025 at 04:00:36PM +0100, Jorge Marques wrote: > On Mon, Nov 24, 2025 at 12:33:12PM +0200, Andy Shevchenko wrote: > > On Mon, Nov 24, 2025 at 10:18:06AM +0100, Jorge Marques wrote:
...
> > > +static int __ad4062_write_event_info_value(struct ad4062_state *st, > > > + enum iio_event_direction dir, int val) > > > +{ > > > + u8 reg; > > > + > > > + if (val > 2047 || val < -2048) > > > + return -EINVAL; > > > > There was already magic '11', perhaps define it and use there and here? > > > > #define x11 11 // needs a good name > > > > if (val > BIT(x11) || val < -BIT(x11)) > > > Not magic number, but max and min signed 12-bit, maybe > > if (val != sign_extend32(val, 11))
If you go this way, the 11 still needs a definition.
> return -EINVAL; > to not look like magic numbers, or > if (val < (-BIT(11)) || val > BIT(11) - 1) > return -EINVAL; > For Hysteresis I will change from > > if (val >= BIT(7)) > to > if (val & ~GENMASK(6,0))
Not sure about this. If it's a HW-based limit, the
val > (BIT(x) - 1)
says that this is limited by x-bit size of the register (field).
So, I leave it to Jonathan (my personal preference here is BIT(x) - 1 approach).
> I believe iio only passes positive to the hysteresis, but is a little clearer. > > > > + if (dir == IIO_EV_DIR_RISING) > > > + reg = AD4062_REG_MAX_LIMIT; > > > + else > > > + reg = AD4062_REG_MIN_LIMIT; > > > + put_unaligned_be16(val, st->buf.bytes); > > > + > > > + return regmap_bulk_write(st->regmap, reg, &st->buf.be16, > > > + sizeof(st->buf.be16)); > > > +}
-- With Best Regards, Andy Shevchenko
|  |