Re: [PATCH v4 07/15] rtc: rzn1: Replace remove callback with devm_add_action_or_reset()

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:

Pre-existing issues:
- [High] Asynchronous pm_runtime_put() in managed teardown action races with pm_runtime_disable(), leaving hardware powered on.
--

--- Patch [7]: [PATCH v4 07/15] rtc: rzn1: Replace remove callback with devm_add_action_or_reset() ---
commit c31bf08fd16c6f4ea71e35f49bac8367fb7f3cce
Author: Lad Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>

rtc: rzn1: Replace remove callback with devm_add_action_or_reset()

This commit simplifies driver teardown by registering a managed cleanup action.
It uses devm_add_action_or_reset() to disable hardware and drop the runtime PM
usage count, eliminating the explicit remove callback and probe error paths.

> diff --git a/drivers/rtc/rtc-rzn1.c b/drivers/rtc/rtc-rzn1.c
> index 94d592e88a044..da7d13337c0da 100644
> --- a/drivers/rtc/rtc-rzn1.c
> +++ b/drivers/rtc/rtc-rzn1.c
> @@ -371,6 +371,17 @@ static const struct rtc_class_ops rzn1_rtc_ops_scmp = {
>  	.alarm_irq_enable = rzn1_rtc_alarm_irq_enable,
>  };
>  
> +static void rzn1_rtc_disable_hardware(void *data)
> +{
> +	struct device *dev = data;
> +	struct rzn1_rtc *rtc = dev_get_drvdata(dev);
> +
> +	/* Disable all interrupts */
> +	writel(0, rtc->base + RZN1_RTC_CTL1);
> +
> +	pm_runtime_put(dev);

[Severity: High]
This isn't a bug introduced by this patch, but does this asynchronous
pm_runtime_put() risk leaving the hardware powered on?

Because rzn1_rtc_disable_hardware() is registered as a devres action, it will
execute during driver unbind just before the automatic cleanup of
pm_runtime_enable(). 

When the devres cleanup calls pm_runtime_disable(), any pending asynchronous
suspend work queued by pm_runtime_put() may be cancelled, leaving the hardware
clocks and power domains permanently active. 

Would it be safer to use pm_runtime_put_sync() here to ensure the device 
reaches a suspended state before the teardown continues?

> +}
> +
>  static int rzn1_rtc_probe(struct platform_device *pdev)
>  {

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260819103926.2136712-1-prabhakar.mahadev-lad.rj@bp.renesas.com?part=7