iio: accel: kxsd9: Fix runtime PM leak in kxsd9_write_raw()

[PATCH] iio: accel: kxsd9: Fix runtime PM leak in kxsd9_write_raw()

Posted by Wentao Liang 1 week, 1 day ago

pm_runtime_get_sync() increments the usage counter even on failure, so
every successful call must be paired with a put. Rejecting an invalid
scale value returned early and skipped the paired
pm_runtime_put_autosuspend(), permanently preventing runtime suspend.

Fixes: 9a9a369d6178 ("iio: accel: kxsd9: Deploy system and runtime PM")
Cc: stable@vger.kernel.org
Signed-off-by: Wentao Liang <vulab@iscas.ac.cn>
---
 drivers/iio/accel/kxsd9.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/drivers/iio/accel/kxsd9.c b/drivers/iio/accel/kxsd9.c
index 4717d80fc24a..1186d711f4d3 100644
--- a/drivers/iio/accel/kxsd9.c
+++ b/drivers/iio/accel/kxsd9.c
@@ -146,8 +146,10 @@ static int kxsd9_write_raw(struct iio_dev *indio_dev,
 
 	if (mask == IIO_CHAN_INFO_SCALE) {
 		/* Check no integer component */
-		if (val)
+		if (val) {
+			pm_runtime_put_autosuspend(st->dev);
 			return -EINVAL;
+		}
 		ret = kxsd9_write_scale(indio_dev, val2);
 	}
 
-- 
2.34.1

Re: [PATCH] iio: accel: kxsd9: Fix runtime PM leak in kxsd9_write_raw()

Posted by Linus Walleij 1 week, 1 day ago

On Wed, Sep 16, 2026 at 6:12 PM Wentao Liang <vulab@iscas.ac.cn> wrote:

> pm_runtime_get_sync() increments the usage counter even on failure, so
> every successful call must be paired with a put. Rejecting an invalid
> scale value returned early and skipped the paired
> pm_runtime_put_autosuspend(), permanently preventing runtime suspend.
>
> Fixes: 9a9a369d6178 ("iio: accel: kxsd9: Deploy system and runtime PM")
> Cc: stable@vger.kernel.org

It's NOT that important.

> Signed-off-by: Wentao Liang <vulab@iscas.ac.cn>

Reviewed-by: Linus Walleij <linusw@kernel.org>

Yours,
Linus Walleij

Re: [PATCH] iio: accel: kxsd9: Fix runtime PM leak in kxsd9_write_raw()

Posted by Jonathan Cameron 1 week, 1 day ago

On Wed, 16 Sep 2026 19:33:47 +0200
Linus Walleij <linusw@kernel.org> wrote:

> On Wed, Sep 16, 2026 at 6:12 PM Wentao Liang <vulab@iscas.ac.cn> wrote:
> 
> > pm_runtime_get_sync() increments the usage counter even on failure, so
> > every successful call must be paired with a put. Rejecting an invalid
> > scale value returned early and skipped the paired
> > pm_runtime_put_autosuspend(), permanently preventing runtime suspend.
> >
> > Fixes: 9a9a369d6178 ("iio: accel: kxsd9: Deploy system and runtime PM")
> > Cc: stable@vger.kernel.org  
> 
> It's NOT that important.
Also not the current code...
static int kxsd9_write_raw(struct iio_dev *indio_dev,
			   struct iio_chan_spec const *chan,
			   int val,
			   int val2,
			   long mask)
{
	int ret = -EINVAL;
	struct kxsd9_state *st = iio_priv(indio_dev);

	pm_runtime_get_sync(st->dev);

	if (mask == IIO_CHAN_INFO_SCALE) {
		/* Check no integer component */
		if (val)
			ret = -EINVAL;
		else
			ret = kxsd9_write_scale(indio_dev, val2);
	}

	pm_runtime_put_autosuspend(st->dev);

	return ret;
}

Which could be improved via the ACQUIRE magic but none the less doesn't
have this bug. Was fixed back in June.

Jonathan



> 
> > Signed-off-by: Wentao Liang <vulab@iscas.ac.cn>  
> 
> Reviewed-by: Linus Walleij <linusw@kernel.org>
> 
> Yours,
> Linus Walleij