> On Fri, 13 Sep 2019 11:07:07 +0200
> Sean Nyekjaer <sean@geanix.com> wrote:
> 
> > The interrupt source can come from multiple sources,
> > fifo and wake interrupts.
> > Enter interrupt thread to check which interrupt that has fired.
> > 
> > Signed-off-by: Sean Nyekjaer <sean@geanix.com>
> > ---
> >  drivers/iio/imu/st_lsm6dsx/st_lsm6dsx_core.c | 4 +---
> >  1 file changed, 1 insertion(+), 3 deletions(-)
> > 
> > diff --git a/drivers/iio/imu/st_lsm6dsx/st_lsm6dsx_core.c b/drivers/iio/imu/st_lsm6dsx/st_lsm6dsx_core.c
> > index 810807c52d5f..80a94335175f 100644
> > --- a/drivers/iio/imu/st_lsm6dsx/st_lsm6dsx_core.c
> > +++ b/drivers/iio/imu/st_lsm6dsx/st_lsm6dsx_core.c
> > @@ -1717,9 +1717,7 @@ static struct iio_dev *st_lsm6dsx_alloc_iiodev(struct st_lsm6dsx_hw *hw,
> >  
> >  static irqreturn_t st_lsm6dsx_handler_irq(int irq, void *private)
> >  {
> > -	struct st_lsm6dsx_hw *hw = private;
> > -
> > -	return hw->sip > 0 ? IRQ_WAKE_THREAD : IRQ_NONE;
> > +	return IRQ_WAKE_THREAD;
> 
> I missed this before.  Isn't this the same as just not providing a top half at all?
> 
> I.e. Pass null to request_threaded_irq where this function was.
> 
> Thanks,
> 
> Jonathan

Right, for the moment we do not need it. It will be probably useful adding
buffering support for sensors that do not support hw timestamping in FIFO
(e.g. LSM9DS1). I am fine both ways, so up to you.

Regards,
Lorenzo

> 
> 
> >  }
> >  
> >  static irqreturn_t st_lsm6dsx_handler_thread(int irq, void *private)
>