On Sat, 2020-11-21 at 17:22 +0000, Jonathan Cameron wrote:
> On Thu, 19 Nov 2020 18:03:29 +0800
> Ye Xiang <xiang.ye@xxxxxxxxx> wrote:
>
> > To avoid pm_runtime_disable called repeatedly by hid sensor
> > drivers,
> > decrease runtime pm enable count after call it.
> >
> > Signed-off-by: Ye Xiang <xiang.ye@xxxxxxxxx>
> This sounds like a fix. If so please make that clear and add a fixes
> tag.
>
> If it couldn't have been triggered before, then please explain why in
> this
> patch description.
Normally it is not a problem as this is called during driver remove and
next time insmod this data structure is recreated and count will init
again. This is the artifect of the shared iio devices and one of them
is still open, so the instance will not be freed.
But patch is fine as to do opposite of what is done during alloc.
Thanks,
Srinivas
>
> Thanks,
>
> Jonathan
>
> > ---
> > drivers/iio/common/hid-sensors/hid-sensor-trigger.c | 4 +++-
> > 1 file changed, 3 insertions(+), 1 deletion(-)
> >
> > diff --git a/drivers/iio/common/hid-sensors/hid-sensor-trigger.c
> > b/drivers/iio/common/hid-sensors/hid-sensor-trigger.c
> > index ff375790b7e8..30340abcbc8d 100644
> > --- a/drivers/iio/common/hid-sensors/hid-sensor-trigger.c
> > +++ b/drivers/iio/common/hid-sensors/hid-sensor-trigger.c
> > @@ -227,8 +227,10 @@ static int
> > hid_sensor_data_rdy_trigger_set_state(struct iio_trigger *trig,
> > void hid_sensor_remove_trigger(struct iio_dev *indio_dev,
> > struct hid_sensor_common *attrb)
> > {
> > - if (atomic_read(&attrb->runtime_pm_enable))
> > + if (atomic_read(&attrb->runtime_pm_enable)) {
> > pm_runtime_disable(&attrb->pdev->dev);
> > + atomic_dec(&attrb->runtime_pm_enable);
> > + }
> >
> > pm_runtime_set_suspended(&attrb->pdev->dev);
> > pm_runtime_put_noidle(&attrb->pdev->dev);
