iio: adc: Fix pm runtime clean-up in sun4i_gpadc_probe

[PATCH] iio: adc: Fix pm runtime clean-up in sun4i_gpadc_probe

Posted by Shuhao Fu 4 months ago

In `sun4i_gpadc_probe`, in case of thermal register failure, the runtime
PM usage counter would not be decreased, resulting in a possible
inconsistency of runtime PM state.

Fixes: b0a242894f11 ("iio: adc: sun4i-gpadc-iio: register in the thermal after registering in pm")
Signed-off-by: Shuhao Fu <sfual@cse.ust.hk>
---
 drivers/iio/adc/sun4i-gpadc-iio.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/iio/adc/sun4i-gpadc-iio.c b/drivers/iio/adc/sun4i-gpadc-iio.c
index 6b8d6bee1873..ae4a59a965b9 100644
--- a/drivers/iio/adc/sun4i-gpadc-iio.c
+++ b/drivers/iio/adc/sun4i-gpadc-iio.c
@@ -644,7 +644,8 @@ static int sun4i_gpadc_probe(struct platform_device *pdev)
 			dev_err(&pdev->dev,
 				"could not register thermal sensor: %ld\n",
 				PTR_ERR(info->tzd));
-			return PTR_ERR(info->tzd);
+			ret = PTR_ERR(info->tzd);
+			goto err_map;
 		}
 	}
 
-- 
2.39.5

Re: [PATCH] iio: adc: Fix pm runtime clean-up in sun4i_gpadc_probe

Posted by Andy Shevchenko 4 months ago

On Wed, Oct 8, 2025 at 7:52 PM Shuhao Fu <sfual@cse.ust.hk> wrote:
>
> In `sun4i_gpadc_probe`, in case of thermal register failure, the runtime
> PM usage counter would not be decreased, resulting in a possible
> inconsistency of runtime PM state.
>
> Fixes: b0a242894f11 ("iio: adc: sun4i-gpadc-iio: register in the thermal after registering in pm")

This might fix this problem, but it doesn't fix the whole mess in the
probe with devm/non-devm ordering.

-- 
With Best Regards,
Andy Shevchenko

Re: [PATCH] iio: adc: Fix pm runtime clean-up in sun4i_gpadc_probe

Posted by Jonathan Cameron 3 months, 3 weeks ago

On Wed, 8 Oct 2025 19:57:47 +0300
Andy Shevchenko <andy.shevchenko@gmail.com> wrote:

> On Wed, Oct 8, 2025 at 7:52 PM Shuhao Fu <sfual@cse.ust.hk> wrote:
> >
> > In `sun4i_gpadc_probe`, in case of thermal register failure, the runtime
> > PM usage counter would not be decreased, resulting in a possible
> > inconsistency of runtime PM state.
> >
> > Fixes: b0a242894f11 ("iio: adc: sun4i-gpadc-iio: register in the thermal after registering in pm")  
> 
> This might fix this problem, but it doesn't fix the whole mess in the
> probe with devm/non-devm ordering.
> 

Mostly this looks simple to fix.  Starting with devm_iio_map_register() instead
of the non devm version.
Then devm_pm_runtime_enable().  However, I have no idea why we need a pm_runtime_put()
in the exit path.  Maybe it's messing with the parent power?  There isn't a matching
get.

Anyone have this hardware to hand for testing if we try to fix this up fully?

Jonathan