[PATCH v9 0/5] Add support for AArch64 AMUv1-based average freq

[PATCH v9 0/5] Add support for AArch64 AMUv1-based average freq

[Beata Michalska]

Hi All,

This series adds support for obtaining an average CPU frequency based on
a hardware provided feedback. The average frequency is being exposed via
dedicated yet optional cpufreq sysfs attribute - cpuinfo_avg_freq.
The architecture specific bits are being provided for AArch64, caching on
existing implementation for FIE and AMUv1 support: the frequency scale
factor, updated on each sched tick, serving as a base for retrieving
the frequency for a given CPU, representing an average frequency
reported between the ticks.

The changes have been rather lightly (due to some limitations) tested on
an FVP model.

Note that [PATCH 2/4] arm64: amu: Delay allocating cpumask for AMU FIE support
can be merged independently.
Additionally, this series depends on [6]

Relevant discussions:
[1] https://lore.kernel.org/all/20240229162520.970986-1-vanshikonda@os.amperecomputing.com/
[2] https://lore.kernel.org/all/7eozim2xnepacnnkzxlbx34hib4otycnbn4dqymfziqou5lw5u@5xzpv3t7sxo3/
[3] https://lore.kernel.org/all/20231212072617.14756-1-lihuisong@huawei.com/
[4] https://lore.kernel.org/lkml/ZIHpd6unkOtYVEqP@e120325.cambridge.arm.com/T/#m4e74cb5a0aaa353c60fedc6cfb95ab7a6e381e3c
[5] https://lore.kernel.org/all/20240603081331.3829278-1-beata.michalska@arm.com/
[6] https://lore.kernel.org/all/20240827154818.1195849-1-ionela.voinescu@arm.com/

v9:
- Moved changes to arch_freq_get_on_cpu to a separate patch

v8:
- Drop introducing new function and reuse arch_freq_get_on_cpu, guarding its use
  in scaling_cur_freq sysfs handler with dedicated config for x86

v7:
- Dropping 'arch_topology: init capacity_freq_ref to 0' patch from the series
  as this one has been sent separately as an independent change
  [https://lore.kernel.org/all/20240827154818.1195849-1-ionela.voinescu@arm.com/]
- Including in the series change that introduces new sysfs entry [PATCH 1/4]
- Consequently modifying previously arch_freq_get_on_cpu to match reqs for new
  sysfs attribute
- Dropping an idea of considering a CPU that has been idle for a while as a
  valid source of information for obtaining an AMU-counter based frequency
- Some minor cosmetic changes

v6:
 - delay allocating cpumask for AMU FIE support instead of invalidating the mask
   upon failure to register cpufreq policy notifications
 - drop the change to cpufreq core (for cpuinfo_cur_freq) as this one will be
   sent as a separate change

v5:
 - Fix invalid access to cpumask
 - Reworked finding reference cpu when getting the freq

v4:
- dropping seqcount
- fixing identifying active cpu within given policy
- skipping full dynticks cpus when retrieving the freq
- bringing back plugging in arch_freq_get_on_cpu into cpuinfo_cur_freq

v3:
- dropping changes to cpufreq_verify_current_freq
- pulling in changes from Ionela initializing capacity_freq_ref to 0
  (thanks for that!)  and applying suggestions made by her during last review:
	- switching to arch_scale_freq_capacity and arch_scale_freq_ref when
	  reversing freq scale factor computation
	- swapping shift with multiplication
- adding time limit for considering last scale update as valid
- updating frequency scale factor upon entering idle

v2:
- Splitting the patches
- Adding comment for full dyntick mode
- Plugging arch_freq_get_on_cpu into cpufreq_verify_current_freq instead
  of in show_cpuinfo_cur_freq to allow the framework to stay more in sync
  with potential freq changes

CC: Jonathan Corbet <corbet@lwn.net>
CC: Thomas Gleixner <tglx@linutronix.de>
CC: Ingo Molnar <mingo@redhat.com>
CC: Borislav Petkov <bp@alien8.de>
CC: Dave Hansen <dave.hansen@linux.intel.com>
CC: H. Peter Anvin <hpa@zytor.com>
CC: Phil Auld <pauld@redhat.com>
CC: x86@kernel.org
CC: linux-doc@vger.kernel.org

*** BLURB HERE ***

Beata Michalska (5):
  cpufreq: Allow arch_freq_get_on_cpu to return an error
  cpufreq: Introduce an optional cpuinfo_avg_freq sysfs entry
  arm64: amu: Delay allocating cpumask for AMU FIE support
  arm64: Provide an AMU-based version of arch_freq_get_on_cpu
  arm64: Update AMU-based freq scale factor on entering idle

 Documentation/admin-guide/pm/cpufreq.rst |  16 ++-
 arch/arm64/kernel/topology.c             | 144 +++++++++++++++++++----
 arch/x86/kernel/cpu/aperfmperf.c         |   2 +-
 arch/x86/kernel/cpu/proc.c               |   7 +-
 drivers/cpufreq/Kconfig.x86              |  12 ++
 drivers/cpufreq/cpufreq.c                |  38 +++++-
 include/linux/cpufreq.h                  |   2 +-
 7 files changed, 189 insertions(+), 32 deletions(-)

-- 
2.25.1

[PATCH v9 2/5] cpufreq: Introduce an optional cpuinfo_avg_freq sysfs entry

[Beata Michalska]

Currently the CPUFreq core exposes two sysfs attributes that can be used
to query current frequency of a given CPU(s): namely cpuinfo_cur_freq
and scaling_cur_freq. Both provide slightly different view on the
subject and they do come with their own drawbacks.

cpuinfo_cur_freq provides higher precision though at a cost of being
rather expensive. Moreover, the information retrieved via this attribute
is somewhat short lived as frequency can change at any point of time
making it difficult to reason from.

scaling_cur_freq, on the other hand, tends to be less accurate but then
the actual level of precision (and source of information) varies between
architectures making it a bit ambiguous.

The new attribute, cpuinfo_avg_freq, is intended to provide more stable,
distinct interface, exposing an average frequency of a given CPU(s), as
reported by the hardware, over a time frame spanning no more than a few
milliseconds. As it requires appropriate hardware support, this
interface is optional.

Note that under the hood, the new attribute relies on the information
provided by arch_freq_get_on_cpu, which, up to this point, has been
feeding data for scaling_cur_freq attribute, being the source of
ambiguity when it comes to interpretation. This has been amended by
restoring the intended behavior for scaling_cur_freq, with a new
dedicated config option to maintain status quo for those, who may need
it.

CC: Jonathan Corbet <corbet@lwn.net>
CC: Thomas Gleixner <tglx@linutronix.de>
CC: Ingo Molnar <mingo@redhat.com>
CC: Borislav Petkov <bp@alien8.de>
CC: Dave Hansen <dave.hansen@linux.intel.com>
CC: H. Peter Anvin <hpa@zytor.com>
CC: Phil Auld <pauld@redhat.com>
CC: x86@kernel.org
CC: linux-doc@vger.kernel.org

Signed-off-by: Beata Michalska <beata.michalska@arm.com>
---
 Documentation/admin-guide/pm/cpufreq.rst | 16 ++++++++++++-
 drivers/cpufreq/Kconfig.x86              | 12 ++++++++++
 drivers/cpufreq/cpufreq.c                | 30 +++++++++++++++++++++++-
 3 files changed, 56 insertions(+), 2 deletions(-)

diff --git a/Documentation/admin-guide/pm/cpufreq.rst b/Documentation/admin-guide/pm/cpufreq.rst
index a21369eba034..e9969174026c 100644
--- a/Documentation/admin-guide/pm/cpufreq.rst
+++ b/Documentation/admin-guide/pm/cpufreq.rst
@@ -248,6 +248,19 @@ are the following:
 	If that frequency cannot be determined, this attribute should not
 	be present.
 
+``cpuinfo_avg_freq``
+        An average frequency (in KHz) of all CPUs belonging to a given policy,
+        derived from a hardware provided feedback and reported on a time frame
+        spanning at most few milliseconds.
+
+        This is expected to be based on the frequency the hardware actually runs
+        at and, as such, might require specialised hardware support (such as AMU
+        extension on ARM). If one cannot be determined, this attribute should
+        not be present.
+
+        Note, that failed attempt to retrieve current frequency for a given
+        CPU(s) will result in an appropriate error.
+
 ``cpuinfo_max_freq``
 	Maximum possible operating frequency the CPUs belonging to this policy
 	can run at (in kHz).
@@ -293,7 +306,8 @@ are the following:
 	Some architectures (e.g. ``x86``) may attempt to provide information
 	more precisely reflecting the current CPU frequency through this
 	attribute, but that still may not be the exact current CPU frequency as
-	seen by the hardware at the moment.
+	seen by the hardware at the moment. This behavior though, is only
+	available via c:macro:``CPUFREQ_ARCH_CUR_FREQ`` option.
 
 ``scaling_driver``
 	The scaling driver currently in use.
diff --git a/drivers/cpufreq/Kconfig.x86 b/drivers/cpufreq/Kconfig.x86
index 97c2d4f15d76..212e1b9afe21 100644
--- a/drivers/cpufreq/Kconfig.x86
+++ b/drivers/cpufreq/Kconfig.x86
@@ -340,3 +340,15 @@ config X86_SPEEDSTEP_RELAXED_CAP_CHECK
 	  option lets the probing code bypass some of those checks if the
 	  parameter "relaxed_check=1" is passed to the module.
 
+config CPUFREQ_ARCH_CUR_FREQ
+	default y
+	bool "Current frequency derived from HW provided feedback"
+	help
+	  This determines whether the scaling_cur_freq sysfs attribute returns
+	  the last requested frequency or a more precise value based on hardware
+	  provided feedback (as architected counters).
+	  Given that a more precise frequency can now be provided via the
+	  cpuinfo_avg_cur_freq attribute, by enabling this option,
+	  scaling_cur_freq maintains the provision of a counter based frequency,
+	  for compatibility reasons.
+
diff --git a/drivers/cpufreq/cpufreq.c b/drivers/cpufreq/cpufreq.c
index 6f45684483c4..b2a8efa83c98 100644
--- a/drivers/cpufreq/cpufreq.c
+++ b/drivers/cpufreq/cpufreq.c
@@ -733,12 +733,20 @@ __weak int arch_freq_get_on_cpu(int cpu)
 	return -EOPNOTSUPP;
 }
 
+static inline bool cpufreq_avg_freq_supported(struct cpufreq_policy *policy)
+{
+	return arch_freq_get_on_cpu(policy->cpu) != -EOPNOTSUPP;
+}
+
 static ssize_t show_scaling_cur_freq(struct cpufreq_policy *policy, char *buf)
 {
 	ssize_t ret;
 	int freq;
 
-	freq = arch_freq_get_on_cpu(policy->cpu);
+	freq = IS_ENABLED(CONFIG_CPUFREQ_ARCH_CUR_FREQ)
+		? arch_freq_get_on_cpu(policy->cpu)
+		: 0;
+
 	if (freq > 0)
 		ret = sysfs_emit(buf, "%u\n", freq);
 	else if (cpufreq_driver->setpolicy && cpufreq_driver->get)
@@ -783,6 +791,19 @@ static ssize_t show_cpuinfo_cur_freq(struct cpufreq_policy *policy,
 	return sysfs_emit(buf, "<unknown>\n");
 }
 
+/*
+ * show_cpuinfo_avg_freq - average CPU frequency as detected by hardware
+ */
+static ssize_t show_cpuinfo_avg_freq(struct cpufreq_policy *policy,
+				     char *buf)
+{
+	int avg_freq = arch_freq_get_on_cpu(policy->cpu);
+
+	if (avg_freq > 0)
+		return sysfs_emit(buf, "%u\n", avg_freq);
+	return avg_freq != 0 ? avg_freq : -EINVAL;
+}
+
 /*
  * show_scaling_governor - show the current policy for the specified CPU
  */
@@ -945,6 +966,7 @@ static ssize_t show_bios_limit(struct cpufreq_policy *policy, char *buf)
 }
 
 cpufreq_freq_attr_ro_perm(cpuinfo_cur_freq, 0400);
+cpufreq_freq_attr_ro(cpuinfo_avg_freq);
 cpufreq_freq_attr_ro(cpuinfo_min_freq);
 cpufreq_freq_attr_ro(cpuinfo_max_freq);
 cpufreq_freq_attr_ro(cpuinfo_transition_latency);
@@ -1072,6 +1094,12 @@ static int cpufreq_add_dev_interface(struct cpufreq_policy *policy)
 			return ret;
 	}
 
+	if (cpufreq_avg_freq_supported(policy)) {
+		ret = sysfs_create_file(&policy->kobj, &cpuinfo_avg_freq.attr);
+		if (ret)
+			return ret;
+	}
+
 	ret = sysfs_create_file(&policy->kobj, &scaling_cur_freq.attr);
 	if (ret)
 		return ret;
-- 
2.25.1

Re: [PATCH v9 2/5] cpufreq: Introduce an optional cpuinfo_avg_freq sysfs entry

[Viresh Kumar]

On 21-01-25, 08:44, Beata Michalska wrote:
> diff --git a/drivers/cpufreq/cpufreq.c b/drivers/cpufreq/cpufreq.c
> index 6f45684483c4..b2a8efa83c98 100644
> --- a/drivers/cpufreq/cpufreq.c
> +++ b/drivers/cpufreq/cpufreq.c
> @@ -733,12 +733,20 @@ __weak int arch_freq_get_on_cpu(int cpu)
>  	return -EOPNOTSUPP;
>  }
>  
>  static ssize_t show_scaling_cur_freq(struct cpufreq_policy *policy, char *buf)
>  {
>  	ssize_t ret;
>  	int freq;
>  
> -	freq = arch_freq_get_on_cpu(policy->cpu);
> +	freq = IS_ENABLED(CONFIG_CPUFREQ_ARCH_CUR_FREQ)
> +		? arch_freq_get_on_cpu(policy->cpu)
> +		: 0;
> +
>  	if (freq > 0)
>  		ret = sysfs_emit(buf, "%u\n", freq);
>  	else if (cpufreq_driver->setpolicy && cpufreq_driver->get)

Maybe this should be a separate commit ? And also I am not very happy
with the new kconfig option. I don't want others to use it as we want
to get rid of this for X86 too eventually. Making it a kconfig option
allows anyone to enable it and then depend on it without us knowing..

Rather just write it as "if (x86)", with a comment on what we plan to
do with it in few release cycles.

-- 
viresh

Re: [PATCH v9 2/5] cpufreq: Introduce an optional cpuinfo_avg_freq sysfs entry

[Beata Michalska]

On Tue, Jan 21, 2025 at 04:23:55PM +0530, Viresh Kumar wrote:
> On 21-01-25, 08:44, Beata Michalska wrote:
> > diff --git a/drivers/cpufreq/cpufreq.c b/drivers/cpufreq/cpufreq.c
> > index 6f45684483c4..b2a8efa83c98 100644
> > --- a/drivers/cpufreq/cpufreq.c
> > +++ b/drivers/cpufreq/cpufreq.c
> > @@ -733,12 +733,20 @@ __weak int arch_freq_get_on_cpu(int cpu)
> >  	return -EOPNOTSUPP;
> >  }
> >  
> >  static ssize_t show_scaling_cur_freq(struct cpufreq_policy *policy, char *buf)
> >  {
> >  	ssize_t ret;
> >  	int freq;
> >  
> > -	freq = arch_freq_get_on_cpu(policy->cpu);
> > +	freq = IS_ENABLED(CONFIG_CPUFREQ_ARCH_CUR_FREQ)
> > +		? arch_freq_get_on_cpu(policy->cpu)
> > +		: 0;
> > +
> >  	if (freq > 0)
> >  		ret = sysfs_emit(buf, "%u\n", freq);
> >  	else if (cpufreq_driver->setpolicy && cpufreq_driver->get)
> 
> Maybe this should be a separate commit ? And also I am not very happy
Initially it was supposed to be one, but then the rest of the series justifies
the changes so it made sense to send those in one go.
> with the new kconfig option. I don't want others to use it as we want
> to get rid of this for X86 too eventually. Making it a kconfig option
> allows anyone to enable it and then depend on it without us knowing..
> 
> Rather just write it as "if (x86)", with a comment on what we plan to
> do with it in few release cycles.
Right, those changes are based on discussion in [1].

---
[1] https://lore.kernel.org/all/CAJZ5v0gCRKzaFrwkoBpLHQUxoP_+jAyhMiCkLQaBUpduk9yxtA@mail.gmail.com/
---
BR
Beata
> 
> -- 
> viresh

Re: [PATCH v9 2/5] cpufreq: Introduce an optional cpuinfo_avg_freq sysfs entry

[Viresh Kumar]

On 21-01-25, 16:17, Beata Michalska wrote:
> On Tue, Jan 21, 2025 at 04:23:55PM +0530, Viresh Kumar wrote:
> > On 21-01-25, 08:44, Beata Michalska wrote:
> > > diff --git a/drivers/cpufreq/cpufreq.c b/drivers/cpufreq/cpufreq.c
> > > index 6f45684483c4..b2a8efa83c98 100644
> > > --- a/drivers/cpufreq/cpufreq.c
> > > +++ b/drivers/cpufreq/cpufreq.c
> > > @@ -733,12 +733,20 @@ __weak int arch_freq_get_on_cpu(int cpu)
> > >  	return -EOPNOTSUPP;
> > >  }
> > >  
> > >  static ssize_t show_scaling_cur_freq(struct cpufreq_policy *policy, char *buf)
> > >  {
> > >  	ssize_t ret;
> > >  	int freq;
> > >  
> > > -	freq = arch_freq_get_on_cpu(policy->cpu);
> > > +	freq = IS_ENABLED(CONFIG_CPUFREQ_ARCH_CUR_FREQ)
> > > +		? arch_freq_get_on_cpu(policy->cpu)
> > > +		: 0;
> > > +
> > >  	if (freq > 0)
> > >  		ret = sysfs_emit(buf, "%u\n", freq);
> > >  	else if (cpufreq_driver->setpolicy && cpufreq_driver->get)
> > 
> > Maybe this should be a separate commit ? And also I am not very happy
> Initially it was supposed to be one, but then the rest of the series justifies
> the changes so it made sense to send those in one go.
> > with the new kconfig option. I don't want others to use it as we want
> > to get rid of this for X86 too eventually. Making it a kconfig option
> > allows anyone to enable it and then depend on it without us knowing..
> > 
> > Rather just write it as "if (x86)", with a comment on what we plan to
> > do with it in few release cycles.
> Right, those changes are based on discussion in [1].

Ahh I see.. What about making it depend on X86 for now, as we really
don't want new users to use it ?

-- 
viresh

Re: [PATCH v9 2/5] cpufreq: Introduce an optional cpuinfo_avg_freq sysfs entry

[Beata Michalska]

On Wed, Jan 22, 2025 at 11:39:02AM +0530, Viresh Kumar wrote:
> On 21-01-25, 16:17, Beata Michalska wrote:
> > On Tue, Jan 21, 2025 at 04:23:55PM +0530, Viresh Kumar wrote:
> > > On 21-01-25, 08:44, Beata Michalska wrote:
> > > > diff --git a/drivers/cpufreq/cpufreq.c b/drivers/cpufreq/cpufreq.c
> > > > index 6f45684483c4..b2a8efa83c98 100644
> > > > --- a/drivers/cpufreq/cpufreq.c
> > > > +++ b/drivers/cpufreq/cpufreq.c
> > > > @@ -733,12 +733,20 @@ __weak int arch_freq_get_on_cpu(int cpu)
> > > >  	return -EOPNOTSUPP;
> > > >  }
> > > >  
> > > >  static ssize_t show_scaling_cur_freq(struct cpufreq_policy *policy, char *buf)
> > > >  {
> > > >  	ssize_t ret;
> > > >  	int freq;
> > > >  
> > > > -	freq = arch_freq_get_on_cpu(policy->cpu);
> > > > +	freq = IS_ENABLED(CONFIG_CPUFREQ_ARCH_CUR_FREQ)
> > > > +		? arch_freq_get_on_cpu(policy->cpu)
> > > > +		: 0;
> > > > +
> > > >  	if (freq > 0)
> > > >  		ret = sysfs_emit(buf, "%u\n", freq);
> > > >  	else if (cpufreq_driver->setpolicy && cpufreq_driver->get)
> > > 
> > > Maybe this should be a separate commit ? And also I am not very happy
> > Initially it was supposed to be one, but then the rest of the series justifies
> > the changes so it made sense to send those in one go.
> > > with the new kconfig option. I don't want others to use it as we want
> > > to get rid of this for X86 too eventually. Making it a kconfig option
> > > allows anyone to enable it and then depend on it without us knowing..
> > > 
> > > Rather just write it as "if (x86)", with a comment on what we plan to
> > > do with it in few release cycles.
> > Right, those changes are based on discussion in [1].
> 
> Ahh I see.. What about making it depend on X86 for now, as we really
> don't want new users to use it ?
Do you mean the new config option? If so, it is in Kconfig.x86 already.
Unless you have smth else in mind ?

---
BR
Beata
> 
> -- 
> viresh

Re: [PATCH v9 2/5] cpufreq: Introduce an optional cpuinfo_avg_freq sysfs entry

[Viresh Kumar]

On 23-01-25, 22:47, Beata Michalska wrote:
> Do you mean the new config option? If so, it is in Kconfig.x86 already.
> Unless you have smth else in mind ?

Its good then :)

-- 
viresh

Re: [PATCH v9 2/5] cpufreq: Introduce an optional cpuinfo_avg_freq sysfs entry

[Prasanna Kumar T S M]

On 21-01-2025 14:14, Beata Michalska wrote:
> Currently the CPUFreq core exposes two sysfs attributes that can be used
> to query current frequency of a given CPU(s): namely cpuinfo_cur_freq
> and scaling_cur_freq. Both provide slightly different view on the
> subject and they do come with their own drawbacks.
>
> cpuinfo_cur_freq provides higher precision though at a cost of being
> rather expensive. Moreover, the information retrieved via this attribute
> is somewhat short lived as frequency can change at any point of time
> making it difficult to reason from.
>
> scaling_cur_freq, on the other hand, tends to be less accurate but then
> the actual level of precision (and source of information) varies between
> architectures making it a bit ambiguous.
>
> The new attribute, cpuinfo_avg_freq, is intended to provide more stable,
> distinct interface, exposing an average frequency of a given CPU(s), as
> reported by the hardware, over a time frame spanning no more than a few
> milliseconds. As it requires appropriate hardware support, this
> interface is optional.
>
> Note that under the hood, the new attribute relies on the information
> provided by arch_freq_get_on_cpu, which, up to this point, has been
> feeding data for scaling_cur_freq attribute, being the source of
> ambiguity when it comes to interpretation. This has been amended by
> restoring the intended behavior for scaling_cur_freq, with a new
> dedicated config option to maintain status quo for those, who may need
> it.
>
> CC: Jonathan Corbet <corbet@lwn.net>
> CC: Thomas Gleixner <tglx@linutronix.de>
> CC: Ingo Molnar <mingo@redhat.com>
> CC: Borislav Petkov <bp@alien8.de>
> CC: Dave Hansen <dave.hansen@linux.intel.com>
> CC: H. Peter Anvin <hpa@zytor.com>
> CC: Phil Auld <pauld@redhat.com>
> CC: x86@kernel.org
> CC: linux-doc@vger.kernel.org
>
> Signed-off-by: Beata Michalska <beata.michalska@arm.com>
> ---
>   Documentation/admin-guide/pm/cpufreq.rst | 16 ++++++++++++-
>   drivers/cpufreq/Kconfig.x86              | 12 ++++++++++
>   drivers/cpufreq/cpufreq.c                | 30 +++++++++++++++++++++++-
>   3 files changed, 56 insertions(+), 2 deletions(-)
>
> diff --git a/Documentation/admin-guide/pm/cpufreq.rst b/Documentation/admin-guide/pm/cpufreq.rst
> index a21369eba034..e9969174026c 100644
> --- a/Documentation/admin-guide/pm/cpufreq.rst
> +++ b/Documentation/admin-guide/pm/cpufreq.rst
> @@ -248,6 +248,19 @@ are the following:
>   	If that frequency cannot be determined, this attribute should not
>   	be present.
>   
> +``cpuinfo_avg_freq``
> +        An average frequency (in KHz) of all CPUs belonging to a given policy,
> +        derived from a hardware provided feedback and reported on a time frame
> +        spanning at most few milliseconds.
> +
> +        This is expected to be based on the frequency the hardware actually runs
> +        at and, as such, might require specialised hardware support (such as AMU
> +        extension on ARM). If one cannot be determined, this attribute should
> +        not be present.
> +
> +        Note, that failed attempt to retrieve current frequency for a given
> +        CPU(s) will result in an appropriate error.
> +
>   ``cpuinfo_max_freq``
>   	Maximum possible operating frequency the CPUs belonging to this policy
>   	can run at (in kHz).
> @@ -293,7 +306,8 @@ are the following:
>   	Some architectures (e.g. ``x86``) may attempt to provide information
>   	more precisely reflecting the current CPU frequency through this
>   	attribute, but that still may not be the exact current CPU frequency as
> -	seen by the hardware at the moment.
> +	seen by the hardware at the moment. This behavior though, is only
> +	available via c:macro:``CPUFREQ_ARCH_CUR_FREQ`` option.
>   
>   ``scaling_driver``
>   	The scaling driver currently in use.
> diff --git a/drivers/cpufreq/Kconfig.x86 b/drivers/cpufreq/Kconfig.x86
> index 97c2d4f15d76..212e1b9afe21 100644
> --- a/drivers/cpufreq/Kconfig.x86
> +++ b/drivers/cpufreq/Kconfig.x86
> @@ -340,3 +340,15 @@ config X86_SPEEDSTEP_RELAXED_CAP_CHECK
>   	  option lets the probing code bypass some of those checks if the
>   	  parameter "relaxed_check=1" is passed to the module.
>   
> +config CPUFREQ_ARCH_CUR_FREQ
> +	default y
> +	bool "Current frequency derived from HW provided feedback"
> +	help
> +	  This determines whether the scaling_cur_freq sysfs attribute returns
> +	  the last requested frequency or a more precise value based on hardware
> +	  provided feedback (as architected counters).
> +	  Given that a more precise frequency can now be provided via the
> +	  cpuinfo_avg_cur_freq attribute, by enabling this option,
> +	  scaling_cur_freq maintains the provision of a counter based frequency,
> +	  for compatibility reasons.
> +
> diff --git a/drivers/cpufreq/cpufreq.c b/drivers/cpufreq/cpufreq.c
> index 6f45684483c4..b2a8efa83c98 100644
> --- a/drivers/cpufreq/cpufreq.c
> +++ b/drivers/cpufreq/cpufreq.c
> @@ -733,12 +733,20 @@ __weak int arch_freq_get_on_cpu(int cpu)
>   	return -EOPNOTSUPP;
>   }
>   
> +static inline bool cpufreq_avg_freq_supported(struct cpufreq_policy *policy)
> +{
> +	return arch_freq_get_on_cpu(policy->cpu) != -EOPNOTSUPP;
> +}
> +
>   static ssize_t show_scaling_cur_freq(struct cpufreq_policy *policy, char *buf)
>   {
>   	ssize_t ret;
>   	int freq;
>   
> -	freq = arch_freq_get_on_cpu(policy->cpu);
> +	freq = IS_ENABLED(CONFIG_CPUFREQ_ARCH_CUR_FREQ)
> +		? arch_freq_get_on_cpu(policy->cpu)
> +		: 0;
> +
>   	if (freq > 0)
>   		ret = sysfs_emit(buf, "%u\n", freq);
>   	else if (cpufreq_driver->setpolicy && cpufreq_driver->get)
> @@ -783,6 +791,19 @@ static ssize_t show_cpuinfo_cur_freq(struct cpufreq_policy *policy,
>   	return sysfs_emit(buf, "<unknown>\n");
>   }
>   
> +/*
> + * show_cpuinfo_avg_freq - average CPU frequency as detected by hardware
> + */
> +static ssize_t show_cpuinfo_avg_freq(struct cpufreq_policy *policy,
> +				     char *buf)
> +{
> +	int avg_freq = arch_freq_get_on_cpu(policy->cpu);
> +
> +	if (avg_freq > 0)
> +		return sysfs_emit(buf, "%u\n", avg_freq);
> +	return avg_freq != 0 ? avg_freq : -EINVAL;
> +}
> +
>   /*
>    * show_scaling_governor - show the current policy for the specified CPU
>    */
> @@ -945,6 +966,7 @@ static ssize_t show_bios_limit(struct cpufreq_policy *policy, char *buf)
>   }
>   
>   cpufreq_freq_attr_ro_perm(cpuinfo_cur_freq, 0400);
> +cpufreq_freq_attr_ro(cpuinfo_avg_freq);
>   cpufreq_freq_attr_ro(cpuinfo_min_freq);
>   cpufreq_freq_attr_ro(cpuinfo_max_freq);
>   cpufreq_freq_attr_ro(cpuinfo_transition_latency);
> @@ -1072,6 +1094,12 @@ static int cpufreq_add_dev_interface(struct cpufreq_policy *policy)
>   			return ret;
>   	}
>   
> +	if (cpufreq_avg_freq_supported(policy)) {
> +		ret = sysfs_create_file(&policy->kobj, &cpuinfo_avg_freq.attr);
> +		if (ret)
> +			return ret;
> +	}
> +
>   	ret = sysfs_create_file(&policy->kobj, &scaling_cur_freq.attr);
>   	if (ret)
>   		return ret;

Looks good to me.

Reviewed-by: Prasanna Kumar T S M <ptsm@linux.microsoft.com>

Re: [PATCH v9 2/5] cpufreq: Introduce an optional cpuinfo_avg_freq sysfs entry

[Sumit Gupta]

On 21/01/25 14:14, Beata Michalska wrote:
> 
> 
> Currently the CPUFreq core exposes two sysfs attributes that can be used
> to query current frequency of a given CPU(s): namely cpuinfo_cur_freq
> and scaling_cur_freq. Both provide slightly different view on the
> subject and they do come with their own drawbacks.
> 
> cpuinfo_cur_freq provides higher precision though at a cost of being
> rather expensive. Moreover, the information retrieved via this attribute
> is somewhat short lived as frequency can change at any point of time
> making it difficult to reason from.
> 
> scaling_cur_freq, on the other hand, tends to be less accurate but then
> the actual level of precision (and source of information) varies between
> architectures making it a bit ambiguous.
> 
> The new attribute, cpuinfo_avg_freq, is intended to provide more stable,
> distinct interface, exposing an average frequency of a given CPU(s), as
> reported by the hardware, over a time frame spanning no more than a few
> milliseconds. As it requires appropriate hardware support, this
> interface is optional.
> 
> Note that under the hood, the new attribute relies on the information
> provided by arch_freq_get_on_cpu, which, up to this point, has been
> feeding data for scaling_cur_freq attribute, being the source of
> ambiguity when it comes to interpretation. This has been amended by
> restoring the intended behavior for scaling_cur_freq, with a new
> dedicated config option to maintain status quo for those, who may need
> it.
> 
> CC: Jonathan Corbet <corbet@lwn.net>
> CC: Thomas Gleixner <tglx@linutronix.de>
> CC: Ingo Molnar <mingo@redhat.com>
> CC: Borislav Petkov <bp@alien8.de>
> CC: Dave Hansen <dave.hansen@linux.intel.com>
> CC: H. Peter Anvin <hpa@zytor.com>
> CC: Phil Auld <pauld@redhat.com>
> CC: x86@kernel.org
> CC: linux-doc@vger.kernel.org
> 
> Signed-off-by: Beata Michalska <beata.michalska@arm.com>
> ---
>   Documentation/admin-guide/pm/cpufreq.rst | 16 ++++++++++++-
>   drivers/cpufreq/Kconfig.x86              | 12 ++++++++++
>   drivers/cpufreq/cpufreq.c                | 30 +++++++++++++++++++++++-
>   3 files changed, 56 insertions(+), 2 deletions(-)
> 
> diff --git a/Documentation/admin-guide/pm/cpufreq.rst b/Documentation/admin-guide/pm/cpufreq.rst
> index a21369eba034..e9969174026c 100644
> --- a/Documentation/admin-guide/pm/cpufreq.rst
> +++ b/Documentation/admin-guide/pm/cpufreq.rst
> @@ -248,6 +248,19 @@ are the following:
>          If that frequency cannot be determined, this attribute should not
>          be present.
> 
> +``cpuinfo_avg_freq``
> +        An average frequency (in KHz) of all CPUs belonging to a given policy,
> +        derived from a hardware provided feedback and reported on a time frame
> +        spanning at most few milliseconds.
> +
> +        This is expected to be based on the frequency the hardware actually runs
> +        at and, as such, might require specialised hardware support (such as AMU
> +        extension on ARM). If one cannot be determined, this attribute should
> +        not be present.
> +
> +        Note, that failed attempt to retrieve current frequency for a given
> +        CPU(s) will result in an appropriate error.
> +

Minor nit:
Should we also add: Idle CPU's on ARM will return EAGAIN (Resource 
temporarily unavailable) error?

>   ``cpuinfo_max_freq``
>          Maximum possible operating frequency the CPUs belonging to this policy
>          can run at (in kHz).
> @@ -293,7 +306,8 @@ are the following:
>          Some architectures (e.g. ``x86``) may attempt to provide information
>          more precisely reflecting the current CPU frequency through this
>          attribute, but that still may not be the exact current CPU frequency as
> -       seen by the hardware at the moment.
> +       seen by the hardware at the moment. This behavior though, is only
> +       available via c:macro:``CPUFREQ_ARCH_CUR_FREQ`` option.
> 
>   ``scaling_driver``
>          The scaling driver currently in use.
> diff --git a/drivers/cpufreq/Kconfig.x86 b/drivers/cpufreq/Kconfig.x86
> index 97c2d4f15d76..212e1b9afe21 100644
> --- a/drivers/cpufreq/Kconfig.x86
> +++ b/drivers/cpufreq/Kconfig.x86
> @@ -340,3 +340,15 @@ config X86_SPEEDSTEP_RELAXED_CAP_CHECK
>            option lets the probing code bypass some of those checks if the
>            parameter "relaxed_check=1" is passed to the module.
> 
> +config CPUFREQ_ARCH_CUR_FREQ
> +       default y
> +       bool "Current frequency derived from HW provided feedback"
> +       help
> +         This determines whether the scaling_cur_freq sysfs attribute returns
> +         the last requested frequency or a more precise value based on hardware
> +         provided feedback (as architected counters).
> +         Given that a more precise frequency can now be provided via the
> +         cpuinfo_avg_cur_freq attribute, by enabling this option,

s/cpuinfo_avg_cur_freq/cpuinfo_cur_freq/?

Overall looks good to me.

Reviewed-by: Sumit Gupta <sumitg@nvidia.com>