Em Sun, 9 May 2021 12:33:53 +0100
Jonathan Cameron <jic23@xxxxxxxxxx> escreveu:
> From: Jonathan Cameron <Jonathan.Cameron@xxxxxxxxxx>
>
> Using this new call makes it easy to handle failures in resume as it
> doesn't hold a reference count if it exits with an error.
>
> Signed-off-by: Jonathan Cameron <Jonathan.Cameron@xxxxxxxxxx>
> Cc: Jean-Baptiste Maneyrol <jmaneyrol@xxxxxxxxxxxxxx>
> ---
> drivers/iio/pressure/icp10100.c | 4 +++-
> 1 file changed, 3 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/iio/pressure/icp10100.c b/drivers/iio/pressure/icp10100.c
> index 48759fc4bf18..80cd66688731 100644
> --- a/drivers/iio/pressure/icp10100.c
> +++ b/drivers/iio/pressure/icp10100.c
> @@ -250,7 +250,9 @@ static int icp10100_get_measures(struct icp10100_state *st,
> __be16 measures[3];
> int ret;
>
> - pm_runtime_get_sync(&st->client->dev);
> + ret = pm_runtime_resume_and_get(&st->client->dev);
> + if (ret < 0)
> + return ret;
>
> mutex_lock(&st->lock);
> cmd = &icp10100_cmd_measure[st->mode];
This one has a logic similar to a patch I commented earlier in this
series. At probing time, it does:
pm_runtime_get_noresume(&client->dev);
pm_runtime_set_active(&client->dev);
pm_runtime_enable(&client->dev);
pm_runtime_set_autosuspend_delay(&client->dev, 2000);
pm_runtime_use_autosuspend(&client->dev);
pm_runtime_put(&client->dev);
ret = devm_add_action_or_reset(&client->dev, icp10100_pm_disable,
&client->dev);
Here, get/put are balanced, but the icp10100_pm_disable logic is:
static void icp10100_pm_disable(void *data)
{
struct device *dev = data;
pm_runtime_put_sync_suspend(dev);
pm_runtime_disable(dev);
}
Not sure what RPM get() it is trying to balance there.
Thanks,
Mauro
