[PATCH] rtc: pm8xxx: switch to devm_device_init_wakeup

[PATCH] rtc: pm8xxx: switch to devm_device_init_wakeup

[alexandre.belloni]

From: Alexandre Belloni <alexandre.belloni@bootlin.com>

Switch to devm_device_init_wakeup to avoid a possible memory leak as wakeup
is never disabled.

Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
---
 drivers/rtc/rtc-pm8xxx.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/rtc/rtc-pm8xxx.c b/drivers/rtc/rtc-pm8xxx.c
index c6241a0c26e9..70cbac76147b 100644
--- a/drivers/rtc/rtc-pm8xxx.c
+++ b/drivers/rtc/rtc-pm8xxx.c
@@ -647,7 +647,7 @@ static int pm8xxx_rtc_probe(struct platform_device *pdev)
 		if (rc)
 			return rc;
 
-		device_init_wakeup(&pdev->dev, true);
+		devm_device_init_wakeup(&pdev->dev);
 	} else {
 		clear_bit(RTC_FEATURE_ALARM, rtc_dd->rtc->features);
 	}
-- 
2.48.1

Re: [PATCH] rtc: pm8xxx: switch to devm_device_init_wakeup

[Johan Hovold]

On Mon, Mar 17, 2025 at 12:13:11PM +0100, alexandre.belloni@bootlin.com wrote:
> From: Alexandre Belloni <alexandre.belloni@bootlin.com>
> 
> Switch to devm_device_init_wakeup to avoid a possible memory leak as wakeup
> is never disabled.

You should probably mention that this was due to a bad merge. The
no-alarm patch moved the previous call to device_init_wakeup() into the
conditional, but you had switched it to use devres so we ended up with
two calls when you applied the patch.

Unless you want to and can rebase your tree.

> Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
> ---
>  drivers/rtc/rtc-pm8xxx.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/drivers/rtc/rtc-pm8xxx.c b/drivers/rtc/rtc-pm8xxx.c
> index c6241a0c26e9..70cbac76147b 100644
> --- a/drivers/rtc/rtc-pm8xxx.c
> +++ b/drivers/rtc/rtc-pm8xxx.c
> @@ -647,7 +647,7 @@ static int pm8xxx_rtc_probe(struct platform_device *pdev)
>  		if (rc)
>  			return rc;
>  
> -		device_init_wakeup(&pdev->dev, true);
> +		devm_device_init_wakeup(&pdev->dev);

So you need to remove the call to devm_device_init_wakeup() above as
well (after platform_set_drvdata()) which the no-alarm patch moved here.

With that fixed:

Reviewed-by: Johan Hovold <johan+linaro@kernel.org>

>  	} else {
>  		clear_bit(RTC_FEATURE_ALARM, rtc_dd->rtc->features);
>  	}

Johan

Re: [PATCH] rtc: pm8xxx: switch to devm_device_init_wakeup

[Alexandre Belloni]

On 17/03/2025 12:58:07+0100, Johan Hovold wrote:
> On Mon, Mar 17, 2025 at 12:13:11PM +0100, alexandre.belloni@bootlin.com wrote:
> > From: Alexandre Belloni <alexandre.belloni@bootlin.com>
> > 
> > Switch to devm_device_init_wakeup to avoid a possible memory leak as wakeup
> > is never disabled.
> 
> You should probably mention that this was due to a bad merge. The
> no-alarm patch moved the previous call to device_init_wakeup() into the
> conditional, but you had switched it to use devres so we ended up with
> two calls when you applied the patch.
> 
> Unless you want to and can rebase your tree.
> 
> > Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
> > ---
> >  drivers/rtc/rtc-pm8xxx.c | 2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> > 
> > diff --git a/drivers/rtc/rtc-pm8xxx.c b/drivers/rtc/rtc-pm8xxx.c
> > index c6241a0c26e9..70cbac76147b 100644
> > --- a/drivers/rtc/rtc-pm8xxx.c
> > +++ b/drivers/rtc/rtc-pm8xxx.c
> > @@ -647,7 +647,7 @@ static int pm8xxx_rtc_probe(struct platform_device *pdev)
> >  		if (rc)
> >  			return rc;
> >  
> > -		device_init_wakeup(&pdev->dev, true);
> > +		devm_device_init_wakeup(&pdev->dev);
> 
> So you need to remove the call to devm_device_init_wakeup() above as
> well (after platform_set_drvdata()) which the no-alarm patch moved here.
> 
> With that fixed:
> 
> Reviewed-by: Johan Hovold <johan+linaro@kernel.org>

Right, I fixed my merge instead, thanks!

> 
> >  	} else {
> >  		clear_bit(RTC_FEATURE_ALARM, rtc_dd->rtc->features);
> >  	}
> 
> Johan
> 

-- 
Alexandre Belloni, co-owner and COO, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com