Re: [PATCH 1/1] iio: imu: adis16475.c: Add buffer padding after temp channel

From: Jonathan Cameron
Date: Sat Sep 30 2023 - 12:39:18 EST


On Tue, 26 Sep 2023 12:50:38 +0200
Nuno Sá <noname.nuno@xxxxxxxxx> wrote:

> On Tue, 2023-09-26 at 11:57 +0300, Ramona Gradinariu wrote:
> > The temperature channel has 16-bit storage size. We need to perform
> > the padding to have the buffer elements naturally aligned in case
> > the temperature channel is enabled and there are any 32-bit storage
> > size channels enabled which have a scan index higher than the
> > temperature channel scan index.
> >
> > Fixes: 8f6bc87d67c0 ("iio: imu: adis16475.c: Add delta angle and delta
> > velocity channels")
> > Signed-off-by: Ramona Gradinariu <ramona.gradinariu@xxxxxxxxxx>
> > ---
>
> Reviewed-by: Nuno Sa <nuno.sa@xxxxxxxxxx>
Applied to the togreg branch of iio.git.

Jonathan

>
> >  drivers/iio/imu/adis16475.c | 10 ++++++++++
> >  1 file changed, 10 insertions(+)
> >
> > diff --git a/drivers/iio/imu/adis16475.c b/drivers/iio/imu/adis16475.c
> > index 00e4e09cdafb..9af07fec0d89 100644
> > --- a/drivers/iio/imu/adis16475.c
> > +++ b/drivers/iio/imu/adis16475.c
> > @@ -1197,6 +1197,16 @@ static irqreturn_t adis16475_trigger_handler(int irq,
> > void *p)
> >                 switch (bit) {
> >                 case ADIS16475_SCAN_TEMP:
> >                         st->data[i++] = buffer[offset];
> > +                       /*
> > +                        * The temperature channel has 16-bit storage size.
> > +                        * We need to perform the padding to have the buffer
> > +                        * elements naturally aligned in case there are any
> > +                        * 32-bit storage size channels enabled which have a
> > +                        * scan index higher than the temperature channel scan
> > +                        * index.
> > +                        */
> > +                       if (*indio_dev->active_scan_mask &
> > GENMASK(ADIS16475_SCAN_DELTVEL_Z, ADIS16475_SCAN_DELTANG_X))
> > +                               st->data[i++] = 0;
>
> You could use ADIS16500_BURST_DATA_SEL_1_CHN_MASK but I kind of agree like this
> is more obvious what's going on...
>
> >                         break;
> >                 case ADIS16475_SCAN_DELTANG_X ... ADIS16475_SCAN_DELTVEL_Z:
> >                         buff_offset = ADIS16475_SCAN_DELTANG_X;
>