On Mon, 7 Aug 2023 20:50:43 +0100
Jonathan Cameron <jic23@xxxxxxxxxx> wrote:
> From: Jonathan Cameron <Jonathan.Cameron@xxxxxxxxxx>
>
> In practice it's unlikely anyone would try to set the integration time
> to 0 which isn't in the available list and if they did then they would
> get index 12 which whilst reserved on the device fits in the field.
> However a compiler might get half way through this reasoning and that
> might be the cause of
>
> > >> drivers/iio/light/opt4001.c:215:9: error: call to '__compiletime_assert_355' declared with 'error' attribute: FIELD_PREP: value too large for the field
> > 215 | reg |= FIELD_PREP(OPT4001_CTRL_CONV_TIME_MASK, chip->int_time);
>
> Even if this isn't the cause, it looks like a bug to me.
>
> Fixes: 9a9608418292 ("iio: light: Add support for TI OPT4001 light sensor")
> Reported-by: kernel test robot <lkp@xxxxxxxxx>
> Closes: https://lore.kernel.org/oe-kbuild-all/202308061902.0gVz6bSe-lkp@xxxxxxxxx/
> Cc: Stefan Windfeldt-Prytz <stefan.windfeldt-prytz@xxxxxxxx>
> Signed-off-by: Jonathan Cameron <Jonathan.Cameron@xxxxxxxxxx>
Oops. Clearly half asleep last night. This is garbage so I'm back to the drawing board
on what is causing clang to believe int_time can be bigger than 4 bits.
Any suggestions welcome!
> ---
> drivers/iio/light/opt4001.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/iio/light/opt4001.c b/drivers/iio/light/opt4001.c
> index 502946bf9f94..83f978706758 100644
> --- a/drivers/iio/light/opt4001.c
> +++ b/drivers/iio/light/opt4001.c
> @@ -137,7 +137,7 @@ static int opt4001_als_time_to_index(const u32 als_integration_time)
> {
> int i;
>
> - for (i = 0; i < ARRAY_SIZE(opt4001_int_time_available); i++) {
> + for (i = 0; i < ARRAY_SIZE(opt4001_int_time_available) / 2; i++) {
> if (als_integration_time == opt4001_int_time_available[i][1])
> return i;
> }
