[PATCH v1 0/2] rtc: cmos: Do not require IRQ if ACPI alarm is used

[PATCH v1 0/2] rtc: cmos: Do not require IRQ if ACPI alarm is used

[Rafael J. Wysocki]

Hi All,

This series of patches allows the CMOS RTC alarm to be used on x86
systems that don't include a functional HPET and may not configure
an IRQ for the CMOS RTC, but have a functional ACPI RTC fixed event.

The first patch allows the ACPI RTC fixed event to be used on systems
without functional HPET because there is no fundamental dependency
between HPET and the ACPI RTC fixed event being hooked up to the CMOS
RTC.

The second patch changes the driver to stop requiring an IRQ to be
configured for the alarm functionality if the ACPI RTC fixed event
is use for signaling events because it require a separate IRQ to
be requested (the ACPI SCI is used for event signaling in that case).

Thanks!

[PATCH v1 1/2] rtc: cmos: Enable ACPI alarm if advertised in ACPI FADT

[Rafael J. Wysocki]

From: Rafael J. Wysocki <rafael.j.wysocki@intel.com>

If the ACPI_FADT_FIXED_RTC flag is unset, the platform is declaring that
it supports the ACPI RTC fixed event which should be used instead of a
dedicated CMOS RTC IRQ.  However, the driver only enables it when
is_hpet_enabled() returns true, which is questionable because there is
no clear connection between enabled HPET and signaling wakeup via the
ACPI RTC fixed event (for instance, the latter can be expected to work
on systems that don't include a functional HPET).

Moreover, since use_hpet_alarm() returns false if use_acpi_alarm is set,
the ACPI RTC fixed event is effectively used instead of the HPET alarm
if the latter is functional, but there is no particular reason why it
could not be used otherwise.

Accordingly, on x86 systems with ACPI, set use_acpi_alarm if
ACPI_FADT_FIXED_RTC is unset without looking at whether or not HPET is
enabled.

Also, do the ACPI FADT check in use_acpi_alarm_quirks() before the DMI
BIOS year checks which are more expensive and it's better to skip them
if ACPI_FADT_FIXED_RTC is set.

Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
---
 drivers/rtc/rtc-cmos.c |    5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

--- a/drivers/rtc/rtc-cmos.c
+++ b/drivers/rtc/rtc-cmos.c
@@ -817,6 +817,9 @@ static void rtc_wake_off(struct device *
 #ifdef CONFIG_X86
 static void use_acpi_alarm_quirks(void)
 {
+	if (acpi_gbl_FADT.flags & ACPI_FADT_FIXED_RTC)
+		return;
+
 	switch (boot_cpu_data.x86_vendor) {
 	case X86_VENDOR_INTEL:
 		if (dmi_get_bios_year() < 2015)
@@ -830,8 +833,6 @@ static void use_acpi_alarm_quirks(void)
 	default:
 		return;
 	}
-	if (!is_hpet_enabled())
-		return;
 
 	use_acpi_alarm = true;
 }

Re: [PATCH v1 1/2] rtc: cmos: Enable ACPI alarm if advertised in ACPI FADT

[Alexandre Belloni]

On 14/03/2026 13:11:20+0100, Rafael J. Wysocki wrote:
> From: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
> 
> If the ACPI_FADT_FIXED_RTC flag is unset, the platform is declaring that
> it supports the ACPI RTC fixed event which should be used instead of a
> dedicated CMOS RTC IRQ.  However, the driver only enables it when
> is_hpet_enabled() returns true, which is questionable because there is
> no clear connection between enabled HPET and signaling wakeup via the
> ACPI RTC fixed event (for instance, the latter can be expected to work
> on systems that don't include a functional HPET).
> 
> Moreover, since use_hpet_alarm() returns false if use_acpi_alarm is set,
> the ACPI RTC fixed event is effectively used instead of the HPET alarm
> if the latter is functional, but there is no particular reason why it
> could not be used otherwise.
> 
> Accordingly, on x86 systems with ACPI, set use_acpi_alarm if
> ACPI_FADT_FIXED_RTC is unset without looking at whether or not HPET is
> enabled.
> 
> Also, do the ACPI FADT check in use_acpi_alarm_quirks() before the DMI
> BIOS year checks which are more expensive and it's better to skip them
> if ACPI_FADT_FIXED_RTC is set.
> 
> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>

Acked-by: Alexandre Belloni <alexandre.belloni@bootlin.com>

> ---
>  drivers/rtc/rtc-cmos.c |    5 +++--
>  1 file changed, 3 insertions(+), 2 deletions(-)
> 
> --- a/drivers/rtc/rtc-cmos.c
> +++ b/drivers/rtc/rtc-cmos.c
> @@ -817,6 +817,9 @@ static void rtc_wake_off(struct device *
>  #ifdef CONFIG_X86
>  static void use_acpi_alarm_quirks(void)
>  {
> +	if (acpi_gbl_FADT.flags & ACPI_FADT_FIXED_RTC)
> +		return;
> +
>  	switch (boot_cpu_data.x86_vendor) {
>  	case X86_VENDOR_INTEL:
>  		if (dmi_get_bios_year() < 2015)
> @@ -830,8 +833,6 @@ static void use_acpi_alarm_quirks(void)
>  	default:
>  		return;
>  	}
> -	if (!is_hpet_enabled())
> -		return;
>  
>  	use_acpi_alarm = true;
>  }
> 
> 
> 

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

Re: [PATCH v1 1/2] rtc: cmos: Enable ACPI alarm if advertised in ACPI FADT

[Mario Limonciello (AMD) (kernel.org)]

On 3/14/2026 7:11 AM, Rafael J. Wysocki wrote:
> From: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
> 
> If the ACPI_FADT_FIXED_RTC flag is unset, the platform is declaring that
> it supports the ACPI RTC fixed event which should be used instead of a
> dedicated CMOS RTC IRQ.  However, the driver only enables it when
> is_hpet_enabled() returns true, which is questionable because there is
> no clear connection between enabled HPET and signaling wakeup via the
> ACPI RTC fixed event (for instance, the latter can be expected to work
> on systems that don't include a functional HPET).
> 
> Moreover, since use_hpet_alarm() returns false if use_acpi_alarm is set,
> the ACPI RTC fixed event is effectively used instead of the HPET alarm
> if the latter is functional, but there is no particular reason why it
> could not be used otherwise.
> 
> Accordingly, on x86 systems with ACPI, set use_acpi_alarm if
> ACPI_FADT_FIXED_RTC is unset without looking at whether or not HPET is
> enabled.
> 
> Also, do the ACPI FADT check in use_acpi_alarm_quirks() before the DMI
> BIOS year checks which are more expensive and it's better to skip them
> if ACPI_FADT_FIXED_RTC is set.
> 
> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>

This makes sense to me.

Sorry for my delay, I kept on meaning to check a few machines and then 
more things came up and it fell lower.

I did check 3 generations of AMD machines and I don't expect any 
problems from this change.

Reviewed-by: Mario Limonciello (AMD) <superm1@kernel.org>

> ---
>   drivers/rtc/rtc-cmos.c |    5 +++--
>   1 file changed, 3 insertions(+), 2 deletions(-)
> 
> --- a/drivers/rtc/rtc-cmos.c
> +++ b/drivers/rtc/rtc-cmos.c
> @@ -817,6 +817,9 @@ static void rtc_wake_off(struct device *
>   #ifdef CONFIG_X86
>   static void use_acpi_alarm_quirks(void)
>   {
> +	if (acpi_gbl_FADT.flags & ACPI_FADT_FIXED_RTC)
> +		return;
> +
>   	switch (boot_cpu_data.x86_vendor) {
>   	case X86_VENDOR_INTEL:
>   		if (dmi_get_bios_year() < 2015)
> @@ -830,8 +833,6 @@ static void use_acpi_alarm_quirks(void)
>   	default:
>   		return;
>   	}
> -	if (!is_hpet_enabled())
> -		return;
>   
>   	use_acpi_alarm = true;
>   }
> 
> 
> 

[PATCH v1 2/2] rtc: cmos: Do not require IRQ if ACPI alarm is used

[Rafael J. Wysocki]

From: Rafael J. Wysocki <rafael.j.wysocki@intel.com>

If the ACPI RTC fixed event is used, a dedicated IRQ is not required
for the CMOS RTC alarm to work, so allow the driver to use the alarm
without a valid IRQ in that case.

Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
---
 drivers/rtc/rtc-cmos.c |   15 ++++++++++-----
 1 file changed, 10 insertions(+), 5 deletions(-)

--- a/drivers/rtc/rtc-cmos.c
+++ b/drivers/rtc/rtc-cmos.c
@@ -216,6 +216,11 @@ static inline void cmos_write_bank2(unsi
 
 /*----------------------------------------------------------------*/
 
+static bool cmos_no_alarm(struct cmos_rtc *cmos)
+{
+	return !is_valid_irq(cmos->irq) && !cmos_use_acpi_alarm();
+}
+
 static int cmos_read_time(struct device *dev, struct rtc_time *t)
 {
 	int ret;
@@ -287,7 +292,7 @@ static int cmos_read_alarm(struct device
 	};
 
 	/* This not only a rtc_op, but also called directly */
-	if (!is_valid_irq(cmos->irq))
+	if (cmos_no_alarm(cmos))
 		return -ETIMEDOUT;
 
 	/* Basic alarms only support hour, minute, and seconds fields.
@@ -520,7 +525,7 @@ static int cmos_set_alarm(struct device
 	int ret;
 
 	/* This not only a rtc_op, but also called directly */
-	if (!is_valid_irq(cmos->irq))
+	if (cmos_no_alarm(cmos))
 		return -EIO;
 
 	ret = cmos_validate_alarm(dev, t);
@@ -1096,7 +1101,7 @@ cmos_do_probe(struct device *dev, struct
 			dev_dbg(dev, "IRQ %d is already in use\n", rtc_irq);
 			goto cleanup1;
 		}
-	} else {
+	} else if (!cmos_use_acpi_alarm()) {
 		clear_bit(RTC_FEATURE_ALARM, cmos_rtc.rtc->features);
 	}
 
@@ -1121,7 +1126,7 @@ cmos_do_probe(struct device *dev, struct
 		acpi_rtc_event_setup(dev);
 
 	dev_info(dev, "%s%s, %d bytes nvram%s\n",
-		 !is_valid_irq(rtc_irq) ? "no alarms" :
+		 cmos_no_alarm(&cmos_rtc) ? "no alarms" :
 		 cmos_rtc.mon_alrm ? "alarms up to one year" :
 		 cmos_rtc.day_alrm ? "alarms up to one month" :
 		 "alarms up to one day",
@@ -1147,7 +1152,7 @@ cleanup0:
 static void cmos_do_shutdown(int rtc_irq)
 {
 	spin_lock_irq(&rtc_lock);
-	if (is_valid_irq(rtc_irq))
+	if (!cmos_no_alarm(&cmos_rtc))
 		cmos_irq_disable(&cmos_rtc, RTC_IRQMASK);
 	spin_unlock_irq(&rtc_lock);
 }

Re: [PATCH v1 2/2] rtc: cmos: Do not require IRQ if ACPI alarm is used

[Alexandre Belloni]

On 14/03/2026 13:12:44+0100, Rafael J. Wysocki wrote:
> From: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
> 
> If the ACPI RTC fixed event is used, a dedicated IRQ is not required
> for the CMOS RTC alarm to work, so allow the driver to use the alarm
> without a valid IRQ in that case.
> 
> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>

Acked-by: Alexandre Belloni <alexandre.belloni@bootlin.com>

> ---
>  drivers/rtc/rtc-cmos.c |   15 ++++++++++-----
>  1 file changed, 10 insertions(+), 5 deletions(-)
> 
> --- a/drivers/rtc/rtc-cmos.c
> +++ b/drivers/rtc/rtc-cmos.c
> @@ -216,6 +216,11 @@ static inline void cmos_write_bank2(unsi
>  
>  /*----------------------------------------------------------------*/
>  
> +static bool cmos_no_alarm(struct cmos_rtc *cmos)
> +{
> +	return !is_valid_irq(cmos->irq) && !cmos_use_acpi_alarm();
> +}
> +
>  static int cmos_read_time(struct device *dev, struct rtc_time *t)
>  {
>  	int ret;
> @@ -287,7 +292,7 @@ static int cmos_read_alarm(struct device
>  	};
>  
>  	/* This not only a rtc_op, but also called directly */
> -	if (!is_valid_irq(cmos->irq))
> +	if (cmos_no_alarm(cmos))
>  		return -ETIMEDOUT;
>  
>  	/* Basic alarms only support hour, minute, and seconds fields.
> @@ -520,7 +525,7 @@ static int cmos_set_alarm(struct device
>  	int ret;
>  
>  	/* This not only a rtc_op, but also called directly */
> -	if (!is_valid_irq(cmos->irq))
> +	if (cmos_no_alarm(cmos))
>  		return -EIO;
>  
>  	ret = cmos_validate_alarm(dev, t);
> @@ -1096,7 +1101,7 @@ cmos_do_probe(struct device *dev, struct
>  			dev_dbg(dev, "IRQ %d is already in use\n", rtc_irq);
>  			goto cleanup1;
>  		}
> -	} else {
> +	} else if (!cmos_use_acpi_alarm()) {
>  		clear_bit(RTC_FEATURE_ALARM, cmos_rtc.rtc->features);
>  	}
>  
> @@ -1121,7 +1126,7 @@ cmos_do_probe(struct device *dev, struct
>  		acpi_rtc_event_setup(dev);
>  
>  	dev_info(dev, "%s%s, %d bytes nvram%s\n",
> -		 !is_valid_irq(rtc_irq) ? "no alarms" :
> +		 cmos_no_alarm(&cmos_rtc) ? "no alarms" :
>  		 cmos_rtc.mon_alrm ? "alarms up to one year" :
>  		 cmos_rtc.day_alrm ? "alarms up to one month" :
>  		 "alarms up to one day",
> @@ -1147,7 +1152,7 @@ cleanup0:
>  static void cmos_do_shutdown(int rtc_irq)
>  {
>  	spin_lock_irq(&rtc_lock);
> -	if (is_valid_irq(rtc_irq))
> +	if (!cmos_no_alarm(&cmos_rtc))
>  		cmos_irq_disable(&cmos_rtc, RTC_IRQMASK);
>  	spin_unlock_irq(&rtc_lock);
>  }
> 
> 
> 

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

Re: [PATCH v1 2/2] rtc: cmos: Do not require IRQ if ACPI alarm is used

[Mario Limonciello (AMD) (kernel.org)]

On 3/20/2026 1:00 PM, Alexandre Belloni wrote:
> On 14/03/2026 13:12:44+0100, Rafael J. Wysocki wrote:
>> From: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
>>
>> If the ACPI RTC fixed event is used, a dedicated IRQ is not required
>> for the CMOS RTC alarm to work, so allow the driver to use the alarm
>> without a valid IRQ in that case.
>>
>> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
> 
> Acked-by: Alexandre Belloni <alexandre.belloni@bootlin.com>

Reviewed-by: Mario Limonciello (AMD) <superm1@kernel.org>

> 
>> ---
>>   drivers/rtc/rtc-cmos.c |   15 ++++++++++-----
>>   1 file changed, 10 insertions(+), 5 deletions(-)
>>
>> --- a/drivers/rtc/rtc-cmos.c
>> +++ b/drivers/rtc/rtc-cmos.c
>> @@ -216,6 +216,11 @@ static inline void cmos_write_bank2(unsi
>>   
>>   /*----------------------------------------------------------------*/
>>   
>> +static bool cmos_no_alarm(struct cmos_rtc *cmos)
>> +{
>> +	return !is_valid_irq(cmos->irq) && !cmos_use_acpi_alarm();
>> +}
>> +
>>   static int cmos_read_time(struct device *dev, struct rtc_time *t)
>>   {
>>   	int ret;
>> @@ -287,7 +292,7 @@ static int cmos_read_alarm(struct device
>>   	};
>>   
>>   	/* This not only a rtc_op, but also called directly */
>> -	if (!is_valid_irq(cmos->irq))
>> +	if (cmos_no_alarm(cmos))
>>   		return -ETIMEDOUT;
>>   
>>   	/* Basic alarms only support hour, minute, and seconds fields.
>> @@ -520,7 +525,7 @@ static int cmos_set_alarm(struct device
>>   	int ret;
>>   
>>   	/* This not only a rtc_op, but also called directly */
>> -	if (!is_valid_irq(cmos->irq))
>> +	if (cmos_no_alarm(cmos))
>>   		return -EIO;
>>   
>>   	ret = cmos_validate_alarm(dev, t);
>> @@ -1096,7 +1101,7 @@ cmos_do_probe(struct device *dev, struct
>>   			dev_dbg(dev, "IRQ %d is already in use\n", rtc_irq);
>>   			goto cleanup1;
>>   		}
>> -	} else {
>> +	} else if (!cmos_use_acpi_alarm()) {
>>   		clear_bit(RTC_FEATURE_ALARM, cmos_rtc.rtc->features);
>>   	}
>>   
>> @@ -1121,7 +1126,7 @@ cmos_do_probe(struct device *dev, struct
>>   		acpi_rtc_event_setup(dev);
>>   
>>   	dev_info(dev, "%s%s, %d bytes nvram%s\n",
>> -		 !is_valid_irq(rtc_irq) ? "no alarms" :
>> +		 cmos_no_alarm(&cmos_rtc) ? "no alarms" :
>>   		 cmos_rtc.mon_alrm ? "alarms up to one year" :
>>   		 cmos_rtc.day_alrm ? "alarms up to one month" :
>>   		 "alarms up to one day",
>> @@ -1147,7 +1152,7 @@ cleanup0:
>>   static void cmos_do_shutdown(int rtc_irq)
>>   {
>>   	spin_lock_irq(&rtc_lock);
>> -	if (is_valid_irq(rtc_irq))
>> +	if (!cmos_no_alarm(&cmos_rtc))
>>   		cmos_irq_disable(&cmos_rtc, RTC_IRQMASK);
>>   	spin_unlock_irq(&rtc_lock);
>>   }
>>
>>
>>
> 

Re: [PATCH v1 0/2] rtc: cmos: Do not require IRQ if ACPI alarm is used

[Rafael J. Wysocki]

On Sat, Mar 14, 2026 at 1:13 PM Rafael J. Wysocki <rafael@kernel.org> wrote:
>
> Hi All,
>
> This series of patches allows the CMOS RTC alarm to be used on x86
> systems that don't include a functional HPET and may not configure
> an IRQ for the CMOS RTC, but have a functional ACPI RTC fixed event.
>
> The first patch allows the ACPI RTC fixed event to be used on systems
> without functional HPET because there is no fundamental dependency
> between HPET and the ACPI RTC fixed event being hooked up to the CMOS
> RTC.
>
> The second patch changes the driver to stop requiring an IRQ to be
> configured for the alarm functionality if the ACPI RTC fixed event
> is use for signaling events because it require a separate IRQ to
> be requested (the ACPI SCI is used for event signaling in that case).

If anyone has any objections or concerns regarding these patches,
please let me know.

In the absence of any, I'll queue them up for 7.1 on top of the
previous CMOS RTC driver changes.

Thanks!