[PATCH v1 0/7] Setting the scene to convert the timers into modules

[PATCH v1 0/7] Setting the scene to convert the timers into modules

[Daniel Lezcano]

The timer drivers are all compiled-in. The initial pre-requisite is to
have them available as soon as possible in the boot process. While
this statement made sense a long time ago, the platforms have today
multiple timers for different purposes along with architected timers
which are initialized very early. For example, a timer can be used as
a backup timer when the local timers are belonging to a power domain
which is shutted down, or used a watchdog timer when the counter are
shared, or also as a pulse width modulation counter. Another use case
is the platform user may want to switch to a timer different from the
architected timers because they have interesting characteristics in
the context of a dedicated platform (eg. automotive).

In some existing drivers, there is already the code to load and unload
a timer driver even if the Kconfig does not allow that. It means, the
need is there but partially upstream.

There were multiple attempts to configure the timer drivers into
modules but it faced the fact that we were unsure if it is correctly
supported by the time framework.

After investigating deeper in the core code it appears we have
everything set for the modularization of the timer drivers.

 - When a clocksource is registered with a better rating, the current
   clocksource is swapped with the new one. The userspace allows to
   change the current clocksource via sysfs

 - A clocksource can be unregistered

 - When a clockevent is registered with a better rating, it becomes
   the active one

 - A clockevent can not be unregistered

A timer driver can be loaded later because of all the supported
above. However unloading is unsupported because a clockevent can not
be unregistered and that will lead to a crash.

But if the timer driver has the module owner set, the core framework
will handle the refcount correctly and will prevent to unload the
module if a clockevent is registered. All the refcounting is working
in different use cases.

 - A clocksource is the current clocksource, the refcount is held

 - A current clocksource is switched to another one, the refcount is
   released

 - A broadcast timer is registered, the refcount is held

 - A local timer is registered, the refcount is held

Consequently, it is possible to unload a module which is only used as
a clocksource. As soon as a clockevent is registered, the refcount is
held and can not be released thus preventing the module to be
unloaded.

That mechanism ensure it is safe to convert the different timer
drivers into modules.

This series adds the module owner in the different driver which are
initialized with the module_platform_driver() function and export the
symbols for the sched_clock_register() function.

Cc: Jim Cromie <jim.cromie@gmail.com>
Cc: Daniel Lezcano <daniel.lezcano@linaro.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Maxime Coquelin <mcoquelin.stm32@gmail.com>
Cc: Alexandre Torgue <alexandre.torgue@foss.st.com>
Cc: Chen-Yu Tsai <wens@csie.org>
Cc: Jernej Skrabec <jernej.skrabec@gmail.com>
Cc: Samuel Holland <samuel@sholland.org>
Cc: Thierry Reding <thierry.reding@gmail.com>
Cc: Jonathan Hunter <jonathanh@nvidia.com>
Cc: "Peter Zijlstra (Intel)" <peterz@infradead.org>
Cc: Marco Elver <elver@google.com>
Cc: Nam Cao <namcao@linutronix.de>
Cc: linux-kernel@vger.kernel.org
Cc: linux-stm32@st-md-mailman.stormreply.com
Cc: linux-arm-kernel@lists.infradead.org
Cc: linux-sunxi@lists.linux.dev
Cc: linux-tegra@vger.kernel.org
Cc: John Stulz <jstultz@google.com>
Cc: Will McVicker <willmcvicker@google.com>
Cc: Peter Griffin <peter.griffin@linaro.org>
Cc: Saravan Kanna <saravanak@google.com>


Daniel Lezcano (7):
  clocksource/drivers/scx200: Add module owner
  clocksource/drivers/stm32-lp: Add module owner
  clocksource/drivers/sun5i: Add module owner
  clocksource/drivers/tegra186: Add module owner
  clocksource/drivers/stm: Add module owner
  clocksource/drivers/cs5535: Add module owner
  time: Export symbol for sched_clock register function

 drivers/clocksource/scx200_hrt.c     | 1 +
 drivers/clocksource/timer-cs5535.c   | 1 +
 drivers/clocksource/timer-nxp-stm.c  | 2 ++
 drivers/clocksource/timer-stm32-lp.c | 1 +
 drivers/clocksource/timer-sun5i.c    | 2 ++
 drivers/clocksource/timer-tegra186.c | 3 +++
 kernel/time/sched_clock.c            | 4 ++--
 7 files changed, 12 insertions(+), 2 deletions(-)

-- 
2.43.0

[PATCH v1 1/7] clocksource/drivers/scx200: Add module owner

[Daniel Lezcano]

The conversion to modules requires a correct handling of the module
refcount in order to prevent to unload it if it is in use. That is
especially true with the clockevents where there is no function to
unregister them.

The core time framework correctly handles the module refcount with the
different clocksource and clockevents if the module owner is set.

Add the module owner to make sure the core framework will prevent
stupid things happening when the driver will be converted into a
module.

Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
---
 drivers/clocksource/scx200_hrt.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/clocksource/scx200_hrt.c b/drivers/clocksource/scx200_hrt.c
index c3536fffbe9a..5a99801a1657 100644
--- a/drivers/clocksource/scx200_hrt.c
+++ b/drivers/clocksource/scx200_hrt.c
@@ -52,6 +52,7 @@ static struct clocksource cs_hrt = {
 	.mask		= CLOCKSOURCE_MASK(32),
 	.flags		= CLOCK_SOURCE_IS_CONTINUOUS,
 	/* mult, shift are set based on mhz27 flag */
+	.owner		= THIS_MODULE,
 };
 
 static int __init init_hrt_clocksource(void)
-- 
2.43.0

Re: [PATCH v1 1/7] clocksource/drivers/scx200: Add module owner

[William McVicker]

On 06/02/2025, Daniel Lezcano wrote:
> The conversion to modules requires a correct handling of the module
> refcount in order to prevent to unload it if it is in use. That is
> especially true with the clockevents where there is no function to
> unregister them.
> 
> The core time framework correctly handles the module refcount with the
> different clocksource and clockevents if the module owner is set.
> 
> Add the module owner to make sure the core framework will prevent
> stupid things happening when the driver will be converted into a
> module.
> 
> Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>

Reviewed-by: Will McVicker <willmcvicker@google.com>

Thanks,
Will

> ---
>  drivers/clocksource/scx200_hrt.c | 1 +
>  1 file changed, 1 insertion(+)
> 
> diff --git a/drivers/clocksource/scx200_hrt.c b/drivers/clocksource/scx200_hrt.c
> index c3536fffbe9a..5a99801a1657 100644
> --- a/drivers/clocksource/scx200_hrt.c
> +++ b/drivers/clocksource/scx200_hrt.c
> @@ -52,6 +52,7 @@ static struct clocksource cs_hrt = {
>  	.mask		= CLOCKSOURCE_MASK(32),
>  	.flags		= CLOCK_SOURCE_IS_CONTINUOUS,
>  	/* mult, shift are set based on mhz27 flag */
> +	.owner		= THIS_MODULE,
>  };
>  
>  static int __init init_hrt_clocksource(void)
> -- 
> 2.43.0
> 

[PATCH v1 2/7] clocksource/drivers/stm32-lp: Add module owner

[Daniel Lezcano]

The conversion to modules requires a correct handling of the module
refcount in order to prevent to unload it if it is in use. That is
especially true with the clockevents where there is no function to
unregister them.

The core time framework correctly handles the module refcount with the
different clocksource and clockevents if the module owner is set.

Add the module owner to make sure the core framework will prevent
stupid things happening when the driver will be converted into a
module.

Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
---
 drivers/clocksource/timer-stm32-lp.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/clocksource/timer-stm32-lp.c b/drivers/clocksource/timer-stm32-lp.c
index 928da2f6de69..cf1423ca00d0 100644
--- a/drivers/clocksource/timer-stm32-lp.c
+++ b/drivers/clocksource/timer-stm32-lp.c
@@ -159,6 +159,7 @@ static void stm32_clkevent_lp_init(struct stm32_lp_private *priv,
 	priv->clkevt.rating = STM32_LP_RATING;
 	priv->clkevt.suspend = stm32_clkevent_lp_suspend;
 	priv->clkevt.resume = stm32_clkevent_lp_resume;
+	priv->clkevt.owner = THIS_MODULE;
 
 	clockevents_config_and_register(&priv->clkevt, rate, 0x1,
 					STM32_LPTIM_MAX_ARR);
-- 
2.43.0

Re: [PATCH v1 2/7] clocksource/drivers/stm32-lp: Add module owner

[William McVicker]

On 06/02/2025, Daniel Lezcano wrote:
> The conversion to modules requires a correct handling of the module
> refcount in order to prevent to unload it if it is in use. That is
> especially true with the clockevents where there is no function to
> unregister them.
> 
> The core time framework correctly handles the module refcount with the
> different clocksource and clockevents if the module owner is set.
> 
> Add the module owner to make sure the core framework will prevent
> stupid things happening when the driver will be converted into a
> module.
> 
> Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>

Reviewed-by: Will McVicker <willmcvicker@google.com>

Thanks,
Will

> ---
>  drivers/clocksource/timer-stm32-lp.c | 1 +
>  1 file changed, 1 insertion(+)
> 
> diff --git a/drivers/clocksource/timer-stm32-lp.c b/drivers/clocksource/timer-stm32-lp.c
> index 928da2f6de69..cf1423ca00d0 100644
> --- a/drivers/clocksource/timer-stm32-lp.c
> +++ b/drivers/clocksource/timer-stm32-lp.c
> @@ -159,6 +159,7 @@ static void stm32_clkevent_lp_init(struct stm32_lp_private *priv,
>  	priv->clkevt.rating = STM32_LP_RATING;
>  	priv->clkevt.suspend = stm32_clkevent_lp_suspend;
>  	priv->clkevt.resume = stm32_clkevent_lp_resume;
> +	priv->clkevt.owner = THIS_MODULE;
>  
>  	clockevents_config_and_register(&priv->clkevt, rate, 0x1,
>  					STM32_LPTIM_MAX_ARR);
> -- 
> 2.43.0
> 

[PATCH v1 3/7] clocksource/drivers/sun5i: Add module owner

[Daniel Lezcano]

The conversion to modules requires a correct handling of the module
refcount in order to prevent to unload it if it is in use. That is
especially true with the clockevents where there is no function to
unregister them.

The core time framework correctly handles the module refcount with the
different clocksource and clockevents if the module owner is set.

Add the module owner to make sure the core framework will prevent
stupid things happening when the driver will be converted into a
module.

Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
---
 drivers/clocksource/timer-sun5i.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/drivers/clocksource/timer-sun5i.c b/drivers/clocksource/timer-sun5i.c
index 6b48a9006444..f827d3f98f60 100644
--- a/drivers/clocksource/timer-sun5i.c
+++ b/drivers/clocksource/timer-sun5i.c
@@ -185,6 +185,7 @@ static int sun5i_setup_clocksource(struct platform_device *pdev,
 	cs->clksrc.read = sun5i_clksrc_read;
 	cs->clksrc.mask = CLOCKSOURCE_MASK(32);
 	cs->clksrc.flags = CLOCK_SOURCE_IS_CONTINUOUS;
+	cs->clksrc.owner = THIS_MODULE;
 
 	ret = clocksource_register_hz(&cs->clksrc, rate);
 	if (ret) {
@@ -214,6 +215,7 @@ static int sun5i_setup_clockevent(struct platform_device *pdev,
 	ce->clkevt.rating = 340;
 	ce->clkevt.irq = irq;
 	ce->clkevt.cpumask = cpu_possible_mask;
+	ce->clkevt.owner = THIS_MODULE;
 
 	/* Enable timer0 interrupt */
 	val = readl(base + TIMER_IRQ_EN_REG);
-- 
2.43.0

Re: [PATCH v1 3/7] clocksource/drivers/sun5i: Add module owner

[Chen-Yu Tsai]

On Mon, Jun 2, 2025 at 11:19 PM Daniel Lezcano
<daniel.lezcano@linaro.org> wrote:
>
> The conversion to modules requires a correct handling of the module
> refcount in order to prevent to unload it if it is in use. That is
> especially true with the clockevents where there is no function to
> unregister them.
>
> The core time framework correctly handles the module refcount with the
> different clocksource and clockevents if the module owner is set.
>
> Add the module owner to make sure the core framework will prevent
> stupid things happening when the driver will be converted into a
> module.
>
> Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>

Acked-by: Chen-Yu Tsai <wens@csie.org>

Re: [PATCH v1 3/7] clocksource/drivers/sun5i: Add module owner

[William McVicker]

On 06/02/2025, Daniel Lezcano wrote:
> The conversion to modules requires a correct handling of the module
> refcount in order to prevent to unload it if it is in use. That is
> especially true with the clockevents where there is no function to
> unregister them.
> 
> The core time framework correctly handles the module refcount with the
> different clocksource and clockevents if the module owner is set.
> 
> Add the module owner to make sure the core framework will prevent
> stupid things happening when the driver will be converted into a
> module.
> 
> Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>

Reviewed-by: Will McVicker <willmcvicker@google.com>

Thanks,
Will

> ---
>  drivers/clocksource/timer-sun5i.c | 2 ++
>  1 file changed, 2 insertions(+)
> 
> diff --git a/drivers/clocksource/timer-sun5i.c b/drivers/clocksource/timer-sun5i.c
> index 6b48a9006444..f827d3f98f60 100644
> --- a/drivers/clocksource/timer-sun5i.c
> +++ b/drivers/clocksource/timer-sun5i.c
> @@ -185,6 +185,7 @@ static int sun5i_setup_clocksource(struct platform_device *pdev,
>  	cs->clksrc.read = sun5i_clksrc_read;
>  	cs->clksrc.mask = CLOCKSOURCE_MASK(32);
>  	cs->clksrc.flags = CLOCK_SOURCE_IS_CONTINUOUS;
> +	cs->clksrc.owner = THIS_MODULE;
>  
>  	ret = clocksource_register_hz(&cs->clksrc, rate);
>  	if (ret) {
> @@ -214,6 +215,7 @@ static int sun5i_setup_clockevent(struct platform_device *pdev,
>  	ce->clkevt.rating = 340;
>  	ce->clkevt.irq = irq;
>  	ce->clkevt.cpumask = cpu_possible_mask;
> +	ce->clkevt.owner = THIS_MODULE;
>  
>  	/* Enable timer0 interrupt */
>  	val = readl(base + TIMER_IRQ_EN_REG);
> -- 
> 2.43.0
> 

[PATCH v1 4/7] clocksource/drivers/tegra186: Add module owner

[Daniel Lezcano]

The conversion to modules requires a correct handling of the module
refcount in order to prevent to unload it if it is in use. That is
especially true with the clockevents where there is no function to
unregister them.

The core time framework correctly handles the module refcount with the
different clocksource and clockevents if the module owner is set.

Add the module owner to make sure the core framework will prevent
stupid things happening when the driver will be converted into a
module.

Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
---
 drivers/clocksource/timer-tegra186.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/drivers/clocksource/timer-tegra186.c b/drivers/clocksource/timer-tegra186.c
index e5394f98a02e..56a5342bcf78 100644
--- a/drivers/clocksource/timer-tegra186.c
+++ b/drivers/clocksource/timer-tegra186.c
@@ -373,6 +373,7 @@ static int tegra186_timer_tsc_init(struct tegra186_timer *tegra)
 	tegra->tsc.read = tegra186_timer_tsc_read;
 	tegra->tsc.mask = CLOCKSOURCE_MASK(56);
 	tegra->tsc.flags = CLOCK_SOURCE_IS_CONTINUOUS;
+	tegra->tsc.owner = THIS_MODULE;
 
 	return clocksource_register_hz(&tegra->tsc, 31250000);
 }
@@ -392,6 +393,7 @@ static int tegra186_timer_osc_init(struct tegra186_timer *tegra)
 	tegra->osc.read = tegra186_timer_osc_read;
 	tegra->osc.mask = CLOCKSOURCE_MASK(32);
 	tegra->osc.flags = CLOCK_SOURCE_IS_CONTINUOUS;
+	tegra->osc.owner = THIS_MODULE;
 
 	return clocksource_register_hz(&tegra->osc, 38400000);
 }
@@ -411,6 +413,7 @@ static int tegra186_timer_usec_init(struct tegra186_timer *tegra)
 	tegra->usec.read = tegra186_timer_usec_read;
 	tegra->usec.mask = CLOCKSOURCE_MASK(32);
 	tegra->usec.flags = CLOCK_SOURCE_IS_CONTINUOUS;
+	tegra->usec.owner = THIS_MODULE;
 
 	return clocksource_register_hz(&tegra->usec, USEC_PER_SEC);
 }
-- 
2.43.0

Re: [PATCH v1 4/7] clocksource/drivers/tegra186: Add module owner

[William McVicker]

On 06/02/2025, Daniel Lezcano wrote:
> The conversion to modules requires a correct handling of the module
> refcount in order to prevent to unload it if it is in use. That is
> especially true with the clockevents where there is no function to
> unregister them.
> 
> The core time framework correctly handles the module refcount with the
> different clocksource and clockevents if the module owner is set.
> 
> Add the module owner to make sure the core framework will prevent
> stupid things happening when the driver will be converted into a
> module.
> 
> Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>

Reviewed-by: Will McVicker <willmcvicker@google.com>

Thanks,
Will

> ---
>  drivers/clocksource/timer-tegra186.c | 3 +++
>  1 file changed, 3 insertions(+)
> 
> diff --git a/drivers/clocksource/timer-tegra186.c b/drivers/clocksource/timer-tegra186.c
> index e5394f98a02e..56a5342bcf78 100644
> --- a/drivers/clocksource/timer-tegra186.c
> +++ b/drivers/clocksource/timer-tegra186.c
> @@ -373,6 +373,7 @@ static int tegra186_timer_tsc_init(struct tegra186_timer *tegra)
>  	tegra->tsc.read = tegra186_timer_tsc_read;
>  	tegra->tsc.mask = CLOCKSOURCE_MASK(56);
>  	tegra->tsc.flags = CLOCK_SOURCE_IS_CONTINUOUS;
> +	tegra->tsc.owner = THIS_MODULE;
>  
>  	return clocksource_register_hz(&tegra->tsc, 31250000);
>  }
> @@ -392,6 +393,7 @@ static int tegra186_timer_osc_init(struct tegra186_timer *tegra)
>  	tegra->osc.read = tegra186_timer_osc_read;
>  	tegra->osc.mask = CLOCKSOURCE_MASK(32);
>  	tegra->osc.flags = CLOCK_SOURCE_IS_CONTINUOUS;
> +	tegra->osc.owner = THIS_MODULE;
>  
>  	return clocksource_register_hz(&tegra->osc, 38400000);
>  }
> @@ -411,6 +413,7 @@ static int tegra186_timer_usec_init(struct tegra186_timer *tegra)
>  	tegra->usec.read = tegra186_timer_usec_read;
>  	tegra->usec.mask = CLOCKSOURCE_MASK(32);
>  	tegra->usec.flags = CLOCK_SOURCE_IS_CONTINUOUS;
> +	tegra->usec.owner = THIS_MODULE;
>  
>  	return clocksource_register_hz(&tegra->usec, USEC_PER_SEC);
>  }
> -- 
> 2.43.0
> 

[PATCH v1 5/7] clocksource/drivers/stm: Add module owner

[Daniel Lezcano]

The conversion to modules requires a correct handling of the module
refcount in order to prevent to unload it if it is in use. That is
especially true with the clockevents where there is no function to
unregister them.

The core time framework correctly handles the module refcount with the
different clocksource and clockevents if the module owner is set.

Add the module owner to make sure the core framework will prevent
stupid things happening when the driver will be converted into a
module.

Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
---
 drivers/clocksource/timer-nxp-stm.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/drivers/clocksource/timer-nxp-stm.c b/drivers/clocksource/timer-nxp-stm.c
index d7ccf9001729..bbc40623728f 100644
--- a/drivers/clocksource/timer-nxp-stm.c
+++ b/drivers/clocksource/timer-nxp-stm.c
@@ -201,6 +201,7 @@ static int __init nxp_stm_clocksource_init(struct device *dev, struct stm_timer
 	stm_timer->cs.resume = nxp_stm_clocksource_resume;
 	stm_timer->cs.mask = CLOCKSOURCE_MASK(32);
 	stm_timer->cs.flags = CLOCK_SOURCE_IS_CONTINUOUS;
+	stm_timer->cs.owner = THIS_MODULE;
 
 	ret = clocksource_register_hz(&stm_timer->cs, stm_timer->rate);
 	if (ret)
@@ -314,6 +315,7 @@ static int __init nxp_stm_clockevent_per_cpu_init(struct device *dev, struct stm
 	stm_timer->ced.cpumask = cpumask_of(cpu);
 	stm_timer->ced.rating = 460;
 	stm_timer->ced.irq = irq;
+	stm_timer->ced.owner = THIS_MODULE;
 
 	per_cpu(stm_timers, cpu) = stm_timer;
 
-- 
2.43.0

Re: [PATCH v1 5/7] clocksource/drivers/stm: Add module owner

[William McVicker]

On 06/02/2025, Daniel Lezcano wrote:
> The conversion to modules requires a correct handling of the module
> refcount in order to prevent to unload it if it is in use. That is
> especially true with the clockevents where there is no function to
> unregister them.
> 
> The core time framework correctly handles the module refcount with the
> different clocksource and clockevents if the module owner is set.
> 
> Add the module owner to make sure the core framework will prevent
> stupid things happening when the driver will be converted into a
> module.
> 
> Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>

Reviewed-by: Will McVicker <willmcvicker@google.com>

Thanks,
Will

> ---
>  drivers/clocksource/timer-nxp-stm.c | 2 ++
>  1 file changed, 2 insertions(+)
> 
> diff --git a/drivers/clocksource/timer-nxp-stm.c b/drivers/clocksource/timer-nxp-stm.c
> index d7ccf9001729..bbc40623728f 100644
> --- a/drivers/clocksource/timer-nxp-stm.c
> +++ b/drivers/clocksource/timer-nxp-stm.c
> @@ -201,6 +201,7 @@ static int __init nxp_stm_clocksource_init(struct device *dev, struct stm_timer
>  	stm_timer->cs.resume = nxp_stm_clocksource_resume;
>  	stm_timer->cs.mask = CLOCKSOURCE_MASK(32);
>  	stm_timer->cs.flags = CLOCK_SOURCE_IS_CONTINUOUS;
> +	stm_timer->cs.owner = THIS_MODULE;
>  
>  	ret = clocksource_register_hz(&stm_timer->cs, stm_timer->rate);
>  	if (ret)
> @@ -314,6 +315,7 @@ static int __init nxp_stm_clockevent_per_cpu_init(struct device *dev, struct stm
>  	stm_timer->ced.cpumask = cpumask_of(cpu);
>  	stm_timer->ced.rating = 460;
>  	stm_timer->ced.irq = irq;
> +	stm_timer->ced.owner = THIS_MODULE;
>  
>  	per_cpu(stm_timers, cpu) = stm_timer;
>  
> -- 
> 2.43.0
> 

[PATCH v1 6/7] clocksource/drivers/cs5535: Add module owner

[Daniel Lezcano]

The conversion to modules requires a correct handling of the module
refcount in order to prevent to unload it if it is in use. That is
especially true with the clockevents where there is no function to
unregister them.

The core time framework correctly handles the module refcount with the
different clocksource and clockevents if the module owner is set.

Add the module owner to make sure the core framework will prevent
stupid things happening when the driver will be converted into a
module.

Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
---
 drivers/clocksource/timer-cs5535.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/clocksource/timer-cs5535.c b/drivers/clocksource/timer-cs5535.c
index d47acfe848ae..8af666c39890 100644
--- a/drivers/clocksource/timer-cs5535.c
+++ b/drivers/clocksource/timer-cs5535.c
@@ -101,6 +101,7 @@ static struct clock_event_device cs5535_clockevent = {
 	.tick_resume = mfgpt_shutdown,
 	.set_next_event = mfgpt_next_event,
 	.rating = 250,
+	.owner = THIS_MODULE,
 };
 
 static irqreturn_t mfgpt_tick(int irq, void *dev_id)
-- 
2.43.0

Re: [PATCH v1 6/7] clocksource/drivers/cs5535: Add module owner

[William McVicker]

On 06/02/2025, Daniel Lezcano wrote:
> The conversion to modules requires a correct handling of the module
> refcount in order to prevent to unload it if it is in use. That is
> especially true with the clockevents where there is no function to
> unregister them.
> 
> The core time framework correctly handles the module refcount with the
> different clocksource and clockevents if the module owner is set.
> 
> Add the module owner to make sure the core framework will prevent
> stupid things happening when the driver will be converted into a
> module.
> 
> Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>

Reviewed-by: Will McVicker <willmcvicker@google.com>

Thanks,
Will

> ---
>  drivers/clocksource/timer-cs5535.c | 1 +
>  1 file changed, 1 insertion(+)
> 
> diff --git a/drivers/clocksource/timer-cs5535.c b/drivers/clocksource/timer-cs5535.c
> index d47acfe848ae..8af666c39890 100644
> --- a/drivers/clocksource/timer-cs5535.c
> +++ b/drivers/clocksource/timer-cs5535.c
> @@ -101,6 +101,7 @@ static struct clock_event_device cs5535_clockevent = {
>  	.tick_resume = mfgpt_shutdown,
>  	.set_next_event = mfgpt_next_event,
>  	.rating = 250,
> +	.owner = THIS_MODULE,
>  };
>  
>  static irqreturn_t mfgpt_tick(int irq, void *dev_id)
> -- 
> 2.43.0
> 

[PATCH v1 7/7] time: Export symbol for sched_clock register function

[Daniel Lezcano]

The timer drivers could be converted into modules. The different
functions to register the clocksource or the clockevent are already
exporting their symbols for modules but the sched_clock_register()
function is missing.

Export the symbols so the drivers using this function can be converted
into modules.

Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
---
 kernel/time/sched_clock.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/kernel/time/sched_clock.c b/kernel/time/sched_clock.c
index cc15fe293719..cc1afec306b3 100644
--- a/kernel/time/sched_clock.c
+++ b/kernel/time/sched_clock.c
@@ -174,8 +174,7 @@ static enum hrtimer_restart sched_clock_poll(struct hrtimer *hrt)
 	return HRTIMER_RESTART;
 }
 
-void __init
-sched_clock_register(u64 (*read)(void), int bits, unsigned long rate)
+void sched_clock_register(u64 (*read)(void), int bits, unsigned long rate)
 {
 	u64 res, wrap, new_mask, new_epoch, cyc, ns;
 	u32 new_mult, new_shift;
@@ -247,6 +246,7 @@ sched_clock_register(u64 (*read)(void), int bits, unsigned long rate)
 
 	pr_debug("Registered %pS as sched_clock source\n", read);
 }
+EXPORT_SYMBOL_GPL(sched_clock_register);
 
 void __init generic_sched_clock_init(void)
 {
-- 
2.43.0

Re: [PATCH v1 7/7] time: Export symbol for sched_clock register function

[William McVicker]

On 06/02/2025, Daniel Lezcano wrote:
> The timer drivers could be converted into modules. The different
> functions to register the clocksource or the clockevent are already
> exporting their symbols for modules but the sched_clock_register()
> function is missing.
> 
> Export the symbols so the drivers using this function can be converted
> into modules.
> 
> Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>

Reviewed-by: Will McVicker <willmcvicker@google.com>

Thanks,
Will

> ---
>  kernel/time/sched_clock.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/kernel/time/sched_clock.c b/kernel/time/sched_clock.c
> index cc15fe293719..cc1afec306b3 100644
> --- a/kernel/time/sched_clock.c
> +++ b/kernel/time/sched_clock.c
> @@ -174,8 +174,7 @@ static enum hrtimer_restart sched_clock_poll(struct hrtimer *hrt)
>  	return HRTIMER_RESTART;
>  }
>  
> -void __init
> -sched_clock_register(u64 (*read)(void), int bits, unsigned long rate)
> +void sched_clock_register(u64 (*read)(void), int bits, unsigned long rate)
>  {
>  	u64 res, wrap, new_mask, new_epoch, cyc, ns;
>  	u32 new_mult, new_shift;
> @@ -247,6 +246,7 @@ sched_clock_register(u64 (*read)(void), int bits, unsigned long rate)
>  
>  	pr_debug("Registered %pS as sched_clock source\n", read);
>  }
> +EXPORT_SYMBOL_GPL(sched_clock_register);
>  
>  void __init generic_sched_clock_init(void)
>  {
> -- 
> 2.43.0
> 

Re: [PATCH v1 7/7] time: Export symbol for sched_clock register function

[John Stultz]

On Mon, Jun 2, 2025 at 8:19 AM Daniel Lezcano <daniel.lezcano@linaro.org> wrote:
>
> The timer drivers could be converted into modules. The different
> functions to register the clocksource or the clockevent are already
> exporting their symbols for modules but the sched_clock_register()
> function is missing.
>
> Export the symbols so the drivers using this function can be converted
> into modules.

Thanks for chasing down all the details around this issue Daniel!

Acked-by: John Stultz <jstultz@google.com>

Re: [PATCH v1 7/7] time: Export symbol for sched_clock register function

[Thomas Gleixner]

On Mon, Jun 02 2025 at 17:18, Daniel Lezcano wrote:

$Subject:  s@time:@time/sched_clock:@

please

Other than that:

Reviewed-by: Thomas Gleixner <tglx@linutronix.de>

Re: [PATCH v1 0/7] Setting the scene to convert the timers into modules

[William McVicker]

On 06/02/2025, Daniel Lezcano wrote:
> The timer drivers are all compiled-in. The initial pre-requisite is to
> have them available as soon as possible in the boot process. While
> this statement made sense a long time ago, the platforms have today
> multiple timers for different purposes along with architected timers
> which are initialized very early. For example, a timer can be used as
> a backup timer when the local timers are belonging to a power domain
> which is shutted down, or used a watchdog timer when the counter are
> shared, or also as a pulse width modulation counter. Another use case
> is the platform user may want to switch to a timer different from the
> architected timers because they have interesting characteristics in
> the context of a dedicated platform (eg. automotive).
> 
> In some existing drivers, there is already the code to load and unload
> a timer driver even if the Kconfig does not allow that. It means, the
> need is there but partially upstream.
> 
> There were multiple attempts to configure the timer drivers into
> modules but it faced the fact that we were unsure if it is correctly
> supported by the time framework.
> 
> After investigating deeper in the core code it appears we have
> everything set for the modularization of the timer drivers.
> 
>  - When a clocksource is registered with a better rating, the current
>    clocksource is swapped with the new one. The userspace allows to
>    change the current clocksource via sysfs
> 
>  - A clocksource can be unregistered
> 
>  - When a clockevent is registered with a better rating, it becomes
>    the active one
> 
>  - A clockevent can not be unregistered
> 
> A timer driver can be loaded later because of all the supported
> above. However unloading is unsupported because a clockevent can not
> be unregistered and that will lead to a crash.
> 
> But if the timer driver has the module owner set, the core framework
> will handle the refcount correctly and will prevent to unload the
> module if a clockevent is registered. All the refcounting is working
> in different use cases.
> 
>  - A clocksource is the current clocksource, the refcount is held
> 
>  - A current clocksource is switched to another one, the refcount is
>    released
> 
>  - A broadcast timer is registered, the refcount is held
> 
>  - A local timer is registered, the refcount is held
> 
> Consequently, it is possible to unload a module which is only used as
> a clocksource. As soon as a clockevent is registered, the refcount is
> held and can not be released thus preventing the module to be
> unloaded.
> 
> That mechanism ensure it is safe to convert the different timer
> drivers into modules.
> 
> This series adds the module owner in the different driver which are
> initialized with the module_platform_driver() function and export the
> symbols for the sched_clock_register() function.
> 

Thanks Daniel for taking the time to dig into this deeper to help identify how
we can safely convert the timer drivers to modules! The series LGTM. I'll go
ahead and address the review comments on my MCT series and rebase it on top of
your patch series.

Thanks,
Will

<cut>

Re: [PATCH v1 0/7] Setting the scene to convert the timers into modules

[Daniel Lezcano]

Hi Thomas,

was this series dropped ?

I'm not able to find it in v6.17-rc1 but it is in the master branch of 
the tip tree (v6.16-rc7)


On 02/06/2025 17:18, Daniel Lezcano wrote:
> The timer drivers are all compiled-in. The initial pre-requisite is to
> have them available as soon as possible in the boot process. While
> this statement made sense a long time ago, the platforms have today
> multiple timers for different purposes along with architected timers
> which are initialized very early. For example, a timer can be used as
> a backup timer when the local timers are belonging to a power domain
> which is shutted down, or used a watchdog timer when the counter are
> shared, or also as a pulse width modulation counter. Another use case
> is the platform user may want to switch to a timer different from the
> architected timers because they have interesting characteristics in
> the context of a dedicated platform (eg. automotive).
> 
> In some existing drivers, there is already the code to load and unload
> a timer driver even if the Kconfig does not allow that. It means, the
> need is there but partially upstream.
> 
> There were multiple attempts to configure the timer drivers into
> modules but it faced the fact that we were unsure if it is correctly
> supported by the time framework.
> 
> After investigating deeper in the core code it appears we have
> everything set for the modularization of the timer drivers.
> 
>   - When a clocksource is registered with a better rating, the current
>     clocksource is swapped with the new one. The userspace allows to
>     change the current clocksource via sysfs
> 
>   - A clocksource can be unregistered
> 
>   - When a clockevent is registered with a better rating, it becomes
>     the active one
> 
>   - A clockevent can not be unregistered
> 
> A timer driver can be loaded later because of all the supported
> above. However unloading is unsupported because a clockevent can not
> be unregistered and that will lead to a crash.
> 
> But if the timer driver has the module owner set, the core framework
> will handle the refcount correctly and will prevent to unload the
> module if a clockevent is registered. All the refcounting is working
> in different use cases.
> 
>   - A clocksource is the current clocksource, the refcount is held
> 
>   - A current clocksource is switched to another one, the refcount is
>     released
> 
>   - A broadcast timer is registered, the refcount is held
> 
>   - A local timer is registered, the refcount is held
> 
> Consequently, it is possible to unload a module which is only used as
> a clocksource. As soon as a clockevent is registered, the refcount is
> held and can not be released thus preventing the module to be
> unloaded.
> 
> That mechanism ensure it is safe to convert the different timer
> drivers into modules.
> 
> This series adds the module owner in the different driver which are
> initialized with the module_platform_driver() function and export the
> symbols for the sched_clock_register() function.
> 
> Cc: Jim Cromie <jim.cromie@gmail.com>
> Cc: Daniel Lezcano <daniel.lezcano@linaro.org>
> Cc: Thomas Gleixner <tglx@linutronix.de>
> Cc: Maxime Coquelin <mcoquelin.stm32@gmail.com>
> Cc: Alexandre Torgue <alexandre.torgue@foss.st.com>
> Cc: Chen-Yu Tsai <wens@csie.org>
> Cc: Jernej Skrabec <jernej.skrabec@gmail.com>
> Cc: Samuel Holland <samuel@sholland.org>
> Cc: Thierry Reding <thierry.reding@gmail.com>
> Cc: Jonathan Hunter <jonathanh@nvidia.com>
> Cc: "Peter Zijlstra (Intel)" <peterz@infradead.org>
> Cc: Marco Elver <elver@google.com>
> Cc: Nam Cao <namcao@linutronix.de>
> Cc: linux-kernel@vger.kernel.org
> Cc: linux-stm32@st-md-mailman.stormreply.com
> Cc: linux-arm-kernel@lists.infradead.org
> Cc: linux-sunxi@lists.linux.dev
> Cc: linux-tegra@vger.kernel.org
> Cc: John Stulz <jstultz@google.com>
> Cc: Will McVicker <willmcvicker@google.com>
> Cc: Peter Griffin <peter.griffin@linaro.org>
> Cc: Saravan Kanna <saravanak@google.com>
> 
> 
> Daniel Lezcano (7):
>    clocksource/drivers/scx200: Add module owner
>    clocksource/drivers/stm32-lp: Add module owner
>    clocksource/drivers/sun5i: Add module owner
>    clocksource/drivers/tegra186: Add module owner
>    clocksource/drivers/stm: Add module owner
>    clocksource/drivers/cs5535: Add module owner
>    time: Export symbol for sched_clock register function
> 
>   drivers/clocksource/scx200_hrt.c     | 1 +
>   drivers/clocksource/timer-cs5535.c   | 1 +
>   drivers/clocksource/timer-nxp-stm.c  | 2 ++
>   drivers/clocksource/timer-stm32-lp.c | 1 +
>   drivers/clocksource/timer-sun5i.c    | 2 ++
>   drivers/clocksource/timer-tegra186.c | 3 +++
>   kernel/time/sched_clock.c            | 4 ++--
>   7 files changed, 12 insertions(+), 2 deletions(-)
> 


-- 
<http://www.linaro.org/> Linaro.org │ Open source software for ARM SoCs

Follow Linaro:  <http://www.facebook.com/pages/Linaro> Facebook |
<http://twitter.com/#!/linaroorg> Twitter |
<http://www.linaro.org/linaro-blog/> Blog