On Sun, Mar 03, 2024 at 05:52:59PM +0100, Vasileios Amoiridis wrote:
> The scan mask for the BME280 device which contains humidity
> measurement needs to become different in order for the timestamp
> to be able to work. Scan masks are added for different combinations
> of measurements. The temperature measurement is needed for either
> pressure or humidity measurements.
...
> +enum bmp280_scan {
> + BMP280_TEMP,
> + BMP280_PRESS,
> + BME280_HUMID,
> +};
Hmm... Why do we need to actually copy the IIO ones? Can't we use IIO ones
directly (or in some way)?
...
> +static const unsigned long bmp280_avail_scan_masks[] = {
> + BIT(BMP280_TEMP),
> + BIT(BMP280_PRESS) | BIT(BMP280_TEMP),
> + 0,
No comma for the terminator line.
> +};
> +static const unsigned long bme280_avail_scan_masks[] = {
> + BIT(BMP280_TEMP),
> + BIT(BMP280_PRESS) | BIT(BMP280_TEMP),
> + BIT(BME280_HUMID) | BIT(BMP280_TEMP),
> + BIT(BME280_HUMID) | BIT(BMP280_PRESS) | BIT(BMP280_TEMP),
> + 0,
Ditto.
> +};
...
> const struct iio_chan_spec *channels;
> int num_channels;
> + const unsigned long *avail_scan_masks;
> unsigned int start_up_time;
Please, run `pahole` every time you are changing data structure layout.
Here you efficiently wasted 8 bytes of memory AFAICS.
--
With Best Regards,
Andy Shevchenko
