All of lore.kernel.org
 help / color / mirror / Atom feed
* [PATCH] watchdog: stm32_iwdg: use managed variant of dev_pm_set_wake_irq()
@ 2026-11-16 19:01 Antonio Borneo
  2026-11-16 19:09 ` sashiko-bot
  2026-11-16 20:36 ` Guenter Roeck
  0 siblings, 2 replies; 3+ messages in thread
From: Antonio Borneo @ 2026-11-16 19:01 UTC (permalink / raw)
  To: Wim Van Sebroeck, Guenter Roeck, Maxime Coquelin,
	Alexandre Torgue, Clément Le Goffic, Marek Vasut
  Cc: Antonio Borneo, linux-watchdog, linux-stm32, linux-arm-kernel,
	linux-kernel, Sashiko

On probe failure after stm32_iwdg_irq_init() or on driver unbind,
the struct wake_irq allocated by dev_pm_set_wake_irq() is leaked
and dev->power.wakeirq remains attached to the device.
Any subsequent attempt to re-probe the device hits the
dev->power.wakeirq check in dev_pm_attach_wake_irq(), triggers a
dev_WARN_ONCE() splat, and fails probe with -EEXIST.

Use the managed version devm_pm_set_wake_irq().

This is reported by Sashiko inside the raw log of another commit
that touches the same driver.

Fixes: 3ab1663af6c1 ("watchdog: stm32_iwdg: Add pretimeout support")
Reported-by: Sashiko <sashiko-bot@kernel.org>
Link: https://sashiko.dev/#/message/20261116132416.1184447-3-antonio.borneo%40foss.st.com
Link: https://sashiko.dev/#/log/249869
Signed-off-by: Antonio Borneo <antonio.borneo@foss.st.com>
---
 drivers/watchdog/stm32_iwdg.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/watchdog/stm32_iwdg.c b/drivers/watchdog/stm32_iwdg.c
index a8d2db2d3cf8b..f1056f8100cf0 100644
--- a/drivers/watchdog/stm32_iwdg.c
+++ b/drivers/watchdog/stm32_iwdg.c
@@ -298,7 +298,7 @@ static int stm32_iwdg_irq_init(struct platform_device *pdev,
 		if (ret)
 			return ret;
 
-		ret = dev_pm_set_wake_irq(dev, irq);
+		ret = devm_pm_set_wake_irq(dev, irq);
 		if (ret)
 			return ret;
 	}

base-commit: 6628e7d4c6b25c8d18f1345e3a20d21829aae7e4
-- 
2.34.1


^ permalink raw reply related	[flat|nested] 3+ messages in thread

* Re: [PATCH] watchdog: stm32_iwdg: use managed variant of dev_pm_set_wake_irq()
  2026-11-16 19:01 [PATCH] watchdog: stm32_iwdg: use managed variant of dev_pm_set_wake_irq() Antonio Borneo
@ 2026-11-16 19:09 ` sashiko-bot
  2026-11-16 20:36 ` Guenter Roeck
  1 sibling, 0 replies; 3+ messages in thread
From: sashiko-bot @ 2026-11-16 19:09 UTC (permalink / raw)
  To: Antonio Borneo; +Cc: linux-watchdog

Thank you for your contribution! Sashiko AI review found no new issues
in this patch, but noticed 1 related pre-existing issue it does not
introduce:
- [Medium] watchdog: stm32_iwdg: NULL pointer dereference in
  stm32_iwdg_isr() [1]
--

commit faff7513f5bea59db8a5e975f7eb126068ba7dc8
Author: Antonio Borneo <antonio.borneo@foss.st.com>

watchdog: stm32_iwdg: use managed variant of dev_pm_set_wake_irq()

This commit switches stm32_iwdg_irq_init() to use devm_pm_set_wake_irq()
so the wake IRQ is automatically detached on probe failure or driver
unbind.

Link: https://sashiko.dev/#/message/20261116132416.1184447-3-antonio.borneo%40foss.st.com
Link: https://sashiko.dev/#/log/249869

> diff --git a/drivers/watchdog/stm32_iwdg.c b/drivers/watchdog/stm32_iwdg.c
> index a8d2db2d3cf8b..f1056f8100cf0 100644
> --- a/drivers/watchdog/stm32_iwdg.c
> +++ b/drivers/watchdog/stm32_iwdg.c
> @@ -298,7 +298,7 @@ static int stm32_iwdg_irq_init(struct platform_device *pdev,
>  		if (ret)
>  			return ret;
>  
> -		ret = dev_pm_set_wake_irq(dev, irq);
> +		ret = devm_pm_set_wake_irq(dev, irq);
>  		if (ret)
>  			return ret;
>  	}

[Severity: Medium]
This problem wasn't introduced by this patch, but can stm32_iwdg_isr() hit
a NULL pointer dereference [1] if the interrupt fires before
watchdog_set_drvdata() runs?

stm32_iwdg_probe() calls stm32_iwdg_irq_init(), which requests the IRQ,
before populating wdd->driver_data:

drivers/watchdog/stm32_iwdg.c:stm32_iwdg_probe() {
    ...
	ret = stm32_iwdg_irq_init(pdev, wdt);
	if (ret)
		return ret;

	watchdog_set_drvdata(wdd, wdt);
    ...
}

If the watchdog is already running with the early-wakeup interrupt enabled
(for example across driver rebind) and the IRQ fires in that window,
stm32_iwdg_isr() gets NULL from watchdog_get_drvdata(wdd) and dereferences
wdt->regs.

[1] https://sashiko.dev/#/bug/linux-ebf60aef-da57-4fba-a11d-75c387866884

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20261116190130.1191501-1-antonio.borneo@foss.st.com?part=1

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: [PATCH] watchdog: stm32_iwdg: use managed variant of dev_pm_set_wake_irq()
  2026-11-16 19:01 [PATCH] watchdog: stm32_iwdg: use managed variant of dev_pm_set_wake_irq() Antonio Borneo
  2026-11-16 19:09 ` sashiko-bot
@ 2026-11-16 20:36 ` Guenter Roeck
  1 sibling, 0 replies; 3+ messages in thread
From: Guenter Roeck @ 2026-11-16 20:36 UTC (permalink / raw)
  To: Antonio Borneo
  Cc: Wim Van Sebroeck, Maxime Coquelin, Alexandre Torgue,
	Clément Le Goffic, Marek Vasut, linux-watchdog, linux-stm32,
	linux-arm-kernel, linux-kernel, Sashiko

On Mon, Nov 16, 2026 at 08:01:30PM +0100, Antonio Borneo wrote:
> On probe failure after stm32_iwdg_irq_init() or on driver unbind,
> the struct wake_irq allocated by dev_pm_set_wake_irq() is leaked
> and dev->power.wakeirq remains attached to the device.
> Any subsequent attempt to re-probe the device hits the
> dev->power.wakeirq check in dev_pm_attach_wake_irq(), triggers a
> dev_WARN_ONCE() splat, and fails probe with -EEXIST.
> 
> Use the managed version devm_pm_set_wake_irq().
> 
> This is reported by Sashiko inside the raw log of another commit
> that touches the same driver.
> 
> Fixes: 3ab1663af6c1 ("watchdog: stm32_iwdg: Add pretimeout support")
> Reported-by: Sashiko <sashiko-bot@kernel.org>
> Link: https://sashiko.dev/#/message/20261116132416.1184447-3-antonio.borneo%40foss.st.com
> Link: https://sashiko.dev/#/log/249869
> Signed-off-by: Antonio Borneo <antonio.borneo@foss.st.com>

Applied.

Thanks,
Guenter

> ---
>  drivers/watchdog/stm32_iwdg.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> 
> base-commit: 6628e7d4c6b25c8d18f1345e3a20d21829aae7e4
> 
> diff --git a/drivers/watchdog/stm32_iwdg.c b/drivers/watchdog/stm32_iwdg.c
> index a8d2db2d3cf8b..f1056f8100cf0 100644
> --- a/drivers/watchdog/stm32_iwdg.c
> +++ b/drivers/watchdog/stm32_iwdg.c
> @@ -298,7 +298,7 @@ static int stm32_iwdg_irq_init(struct platform_device *pdev,
>  		if (ret)
>  			return ret;
>  
> -		ret = dev_pm_set_wake_irq(dev, irq);
> +		ret = devm_pm_set_wake_irq(dev, irq);
>  		if (ret)
>  			return ret;
>  	}


^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2026-11-16 20:36 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2026-11-16 19:01 [PATCH] watchdog: stm32_iwdg: use managed variant of dev_pm_set_wake_irq() Antonio Borneo
2026-11-16 19:09 ` sashiko-bot
2026-11-16 20:36 ` Guenter Roeck

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.