[PATCH v2 23/31] watchdog: Switch to use hrtimer_setup()

[PATCH v2 23/31] watchdog: Switch to use hrtimer_setup()

[Nam Cao]

hrtimer_setup() takes the callback function pointer as argument and
initializes the timer completely.

Replace hrtimer_init() and the open coded initialization of
hrtimer::function with the new setup mechanism.

Patch was created by using Coccinelle.

Signed-off-by: Nam Cao <namcao@linutronix.de>
Cc: Wim Van Sebroeck <wim@linux-watchdog.org>
Cc: linux-watchdog@vger.kernel.org
---
 drivers/watchdog/softdog.c                     | 8 +++-----
 drivers/watchdog/watchdog_dev.c                | 4 ++--
 drivers/watchdog/watchdog_hrtimer_pretimeout.c | 4 ++--
 kernel/watchdog.c                              | 3 +--
 4 files changed, 8 insertions(+), 11 deletions(-)

diff --git a/drivers/watchdog/softdog.c b/drivers/watchdog/softdog.c
index 7a1096265f18..0820e35ad2e3 100644
--- a/drivers/watchdog/softdog.c
+++ b/drivers/watchdog/softdog.c
@@ -187,14 +187,12 @@ static int __init softdog_init(void)
 	watchdog_set_nowayout(&softdog_dev, nowayout);
 	watchdog_stop_on_reboot(&softdog_dev);
 
-	hrtimer_init(&softdog_ticktock, CLOCK_MONOTONIC, HRTIMER_MODE_REL);
-	softdog_ticktock.function = softdog_fire;
+	hrtimer_setup(&softdog_ticktock, softdog_fire, CLOCK_MONOTONIC, HRTIMER_MODE_REL);
 
 	if (IS_ENABLED(CONFIG_SOFT_WATCHDOG_PRETIMEOUT)) {
 		softdog_info.options |= WDIOF_PRETIMEOUT;
-		hrtimer_init(&softdog_preticktock, CLOCK_MONOTONIC,
-			     HRTIMER_MODE_REL);
-		softdog_preticktock.function = softdog_pretimeout;
+		hrtimer_setup(&softdog_preticktock, softdog_pretimeout, CLOCK_MONOTONIC,
+			      HRTIMER_MODE_REL);
 	}
 
 	if (soft_active_on_boot)
diff --git a/drivers/watchdog/watchdog_dev.c b/drivers/watchdog/watchdog_dev.c
index 19698d87dc57..8369fd94fc1a 100644
--- a/drivers/watchdog/watchdog_dev.c
+++ b/drivers/watchdog/watchdog_dev.c
@@ -1051,8 +1051,8 @@ static int watchdog_cdev_register(struct watchdog_device *wdd)
 	}
 
 	kthread_init_work(&wd_data->work, watchdog_ping_work);
-	hrtimer_init(&wd_data->timer, CLOCK_MONOTONIC, HRTIMER_MODE_REL_HARD);
-	wd_data->timer.function = watchdog_timer_expired;
+	hrtimer_setup(&wd_data->timer, watchdog_timer_expired, CLOCK_MONOTONIC,
+		      HRTIMER_MODE_REL_HARD);
 	watchdog_hrtimer_pretimeout_init(wdd);
 
 	if (wdd->id == 0) {
diff --git a/drivers/watchdog/watchdog_hrtimer_pretimeout.c b/drivers/watchdog/watchdog_hrtimer_pretimeout.c
index 940b53718a91..fbc7eecd8b20 100644
--- a/drivers/watchdog/watchdog_hrtimer_pretimeout.c
+++ b/drivers/watchdog/watchdog_hrtimer_pretimeout.c
@@ -23,8 +23,8 @@ void watchdog_hrtimer_pretimeout_init(struct watchdog_device *wdd)
 {
 	struct watchdog_core_data *wd_data = wdd->wd_data;
 
-	hrtimer_init(&wd_data->pretimeout_timer, CLOCK_MONOTONIC, HRTIMER_MODE_REL);
-	wd_data->pretimeout_timer.function = watchdog_hrtimer_pretimeout;
+	hrtimer_setup(&wd_data->pretimeout_timer, watchdog_hrtimer_pretimeout, CLOCK_MONOTONIC,
+		      HRTIMER_MODE_REL);
 }
 
 void watchdog_hrtimer_pretimeout_start(struct watchdog_device *wdd)
diff --git a/kernel/watchdog.c b/kernel/watchdog.c
index b2da7de39d06..6a98dbc931ac 100644
--- a/kernel/watchdog.c
+++ b/kernel/watchdog.c
@@ -797,8 +797,7 @@ static void watchdog_enable(unsigned int cpu)
 	 * Start the timer first to prevent the hardlockup watchdog triggering
 	 * before the timer has a chance to fire.
 	 */
-	hrtimer_init(hrtimer, CLOCK_MONOTONIC, HRTIMER_MODE_REL_HARD);
-	hrtimer->function = watchdog_timer_fn;
+	hrtimer_setup(hrtimer, watchdog_timer_fn, CLOCK_MONOTONIC, HRTIMER_MODE_REL_HARD);
 	hrtimer_start(hrtimer, ns_to_ktime(sample_period),
 		      HRTIMER_MODE_REL_PINNED_HARD);
 
-- 
2.39.5

Re: [PATCH v2 23/31] watchdog: Switch to use hrtimer_setup()

[Guenter Roeck]

On Wed, Feb 05, 2025 at 11:39:07AM +0100, Nam Cao wrote:
> hrtimer_setup() takes the callback function pointer as argument and
> initializes the timer completely.
> 
> Replace hrtimer_init() and the open coded initialization of
> hrtimer::function with the new setup mechanism.
> 
> Patch was created by using Coccinelle.
> 
> Signed-off-by: Nam Cao <namcao@linutronix.de>
> Cc: Wim Van Sebroeck <wim@linux-watchdog.org>
> Cc: linux-watchdog@vger.kernel.org
> ---

Change log missing. I am also personally not a friend of changing
multiple drivers in a single patch for changes like this. That makes
it all but impossible to revert one of the driver changes if it was
wrong without reverting everything.

Guenter

>  drivers/watchdog/softdog.c                     | 8 +++-----
>  drivers/watchdog/watchdog_dev.c                | 4 ++--
>  drivers/watchdog/watchdog_hrtimer_pretimeout.c | 4 ++--
>  kernel/watchdog.c                              | 3 +--
>  4 files changed, 8 insertions(+), 11 deletions(-)
> 
> diff --git a/drivers/watchdog/softdog.c b/drivers/watchdog/softdog.c
> index 7a1096265f18..0820e35ad2e3 100644
> --- a/drivers/watchdog/softdog.c
> +++ b/drivers/watchdog/softdog.c
> @@ -187,14 +187,12 @@ static int __init softdog_init(void)
>  	watchdog_set_nowayout(&softdog_dev, nowayout);
>  	watchdog_stop_on_reboot(&softdog_dev);
>  
> -	hrtimer_init(&softdog_ticktock, CLOCK_MONOTONIC, HRTIMER_MODE_REL);
> -	softdog_ticktock.function = softdog_fire;
> +	hrtimer_setup(&softdog_ticktock, softdog_fire, CLOCK_MONOTONIC, HRTIMER_MODE_REL);
>  
>  	if (IS_ENABLED(CONFIG_SOFT_WATCHDOG_PRETIMEOUT)) {
>  		softdog_info.options |= WDIOF_PRETIMEOUT;
> -		hrtimer_init(&softdog_preticktock, CLOCK_MONOTONIC,
> -			     HRTIMER_MODE_REL);
> -		softdog_preticktock.function = softdog_pretimeout;
> +		hrtimer_setup(&softdog_preticktock, softdog_pretimeout, CLOCK_MONOTONIC,
> +			      HRTIMER_MODE_REL);
>  	}
>  
>  	if (soft_active_on_boot)
> diff --git a/drivers/watchdog/watchdog_dev.c b/drivers/watchdog/watchdog_dev.c
> index 19698d87dc57..8369fd94fc1a 100644
> --- a/drivers/watchdog/watchdog_dev.c
> +++ b/drivers/watchdog/watchdog_dev.c
> @@ -1051,8 +1051,8 @@ static int watchdog_cdev_register(struct watchdog_device *wdd)
>  	}
>  
>  	kthread_init_work(&wd_data->work, watchdog_ping_work);
> -	hrtimer_init(&wd_data->timer, CLOCK_MONOTONIC, HRTIMER_MODE_REL_HARD);
> -	wd_data->timer.function = watchdog_timer_expired;
> +	hrtimer_setup(&wd_data->timer, watchdog_timer_expired, CLOCK_MONOTONIC,
> +		      HRTIMER_MODE_REL_HARD);
>  	watchdog_hrtimer_pretimeout_init(wdd);
>  
>  	if (wdd->id == 0) {
> diff --git a/drivers/watchdog/watchdog_hrtimer_pretimeout.c b/drivers/watchdog/watchdog_hrtimer_pretimeout.c
> index 940b53718a91..fbc7eecd8b20 100644
> --- a/drivers/watchdog/watchdog_hrtimer_pretimeout.c
> +++ b/drivers/watchdog/watchdog_hrtimer_pretimeout.c
> @@ -23,8 +23,8 @@ void watchdog_hrtimer_pretimeout_init(struct watchdog_device *wdd)
>  {
>  	struct watchdog_core_data *wd_data = wdd->wd_data;
>  
> -	hrtimer_init(&wd_data->pretimeout_timer, CLOCK_MONOTONIC, HRTIMER_MODE_REL);
> -	wd_data->pretimeout_timer.function = watchdog_hrtimer_pretimeout;
> +	hrtimer_setup(&wd_data->pretimeout_timer, watchdog_hrtimer_pretimeout, CLOCK_MONOTONIC,
> +		      HRTIMER_MODE_REL);
>  }
>  
>  void watchdog_hrtimer_pretimeout_start(struct watchdog_device *wdd)
> diff --git a/kernel/watchdog.c b/kernel/watchdog.c
> index b2da7de39d06..6a98dbc931ac 100644
> --- a/kernel/watchdog.c
> +++ b/kernel/watchdog.c
> @@ -797,8 +797,7 @@ static void watchdog_enable(unsigned int cpu)
>  	 * Start the timer first to prevent the hardlockup watchdog triggering
>  	 * before the timer has a chance to fire.
>  	 */
> -	hrtimer_init(hrtimer, CLOCK_MONOTONIC, HRTIMER_MODE_REL_HARD);
> -	hrtimer->function = watchdog_timer_fn;
> +	hrtimer_setup(hrtimer, watchdog_timer_fn, CLOCK_MONOTONIC, HRTIMER_MODE_REL_HARD);
>  	hrtimer_start(hrtimer, ns_to_ktime(sample_period),
>  		      HRTIMER_MODE_REL_PINNED_HARD);
>  
> -- 
> 2.39.5
> 
> 

Re: [PATCH v2 23/31] watchdog: Switch to use hrtimer_setup()

[Thomas Gleixner]

On Tue, Feb 11 2025 at 08:16, Guenter Roeck wrote:
> On Wed, Feb 05, 2025 at 11:39:07AM +0100, Nam Cao wrote:
>> hrtimer_setup() takes the callback function pointer as argument and
>> initializes the timer completely.
>> 
>> Replace hrtimer_init() and the open coded initialization of
>> hrtimer::function with the new setup mechanism.
>> 
>> Patch was created by using Coccinelle.
>> 
>> Signed-off-by: Nam Cao <namcao@linutronix.de>
>> Cc: Wim Van Sebroeck <wim@linux-watchdog.org>
>> Cc: linux-watchdog@vger.kernel.org
>> ---
>
> Change log missing.

I read the above as a change log, but _I_ might be missing something.

> I am also personally not a friend of changing
> multiple drivers in a single patch for changes like this. That makes
> it all but impossible to revert one of the driver changes if it was
> wrong without reverting everything.

Come on. This is a purely mechanical change which is fully equivalent,
so splitting it up further is just increasing patch count for no value.

Thanks,

        tglx

Re: [PATCH v2 23/31] watchdog: Switch to use hrtimer_setup()

[Guenter Roeck]

On 2/13/25 03:16, Thomas Gleixner wrote:
> On Tue, Feb 11 2025 at 08:16, Guenter Roeck wrote:
>> On Wed, Feb 05, 2025 at 11:39:07AM +0100, Nam Cao wrote:
>>> hrtimer_setup() takes the callback function pointer as argument and
>>> initializes the timer completely.
>>>
>>> Replace hrtimer_init() and the open coded initialization of
>>> hrtimer::function with the new setup mechanism.
>>>
>>> Patch was created by using Coccinelle.
>>>
>>> Signed-off-by: Nam Cao <namcao@linutronix.de>
>>> Cc: Wim Van Sebroeck <wim@linux-watchdog.org>
>>> Cc: linux-watchdog@vger.kernel.org
>>> ---
>>
>> Change log missing.
> 
> I read the above as a change log, but _I_ might be missing something.
> 

I have no idea what changed since v1 or why. If you can see that
in the patch description, good for you.

Guenter

Re: [PATCH v2 23/31] watchdog: Switch to use hrtimer_setup()

[Nam Cao]

Hi Guenter,

On Thu, Feb 13, 2025 at 04:24:25AM -0800, Guenter Roeck wrote:
> I have no idea what changed since v1 or why. If you can see that
> in the patch description, good for you.

The change log is in the cover letter [1]. I just noticed that I forgot to Cc
the cover letter to the watchdog mailing list, sorry about that!

This v2 is sent due to conflict with 6.14-rc1. Most patches (including this
one) are the same.

Best regards,
Nam

[1] https://lore.kernel.org/lkml/cover.1738746821.git.namcao@linutronix.de/