[v2] cpufreq: Avoid get_governor() for first policy

[PATCH v2] cpufreq: Avoid get_governor() for first policy

Posted by Zihuan Zhang 2 months, 1 week ago

When a cpufreq driver registers the first policy, it may attempt to
initialize the policy governor from `last_governor`. However, this is
meaningless for the first policy instance, because `last_governor` is
only updated when policies are removed (e.g. during CPU offline).

The `last_governor` mechanism is intended to restore the previously
used governor across CPU hotplug events. For the very first policy,
there is no "previous governor" to restore, so calling
get_governor(last_governor) is unnecessary and potentially confusing.

This patch skips looking up `last_governor` when registering the first
policy. Instead, it directly uses the default governor after all
governors have been registered and are available.

This avoids meaningless lookups, reduces unnecessary module reference
handling, and simplifies the initial policy path.

Signed-off-by: Zihuan Zhang <zhangzihuan@kylinos.cn>

---
Changes in v2:
 - Fix the case where the governor is NULL.
---
---
 drivers/cpufreq/cpufreq.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/cpufreq/cpufreq.c b/drivers/cpufreq/cpufreq.c
index d7426e1d8bdd..1aa559f53479 100644
--- a/drivers/cpufreq/cpufreq.c
+++ b/drivers/cpufreq/cpufreq.c
@@ -1122,7 +1122,8 @@ static int cpufreq_init_policy(struct cpufreq_policy *policy)
 
 	if (has_target()) {
 		/* Update policy governor to the one used before hotplug. */
-		gov = get_governor(policy->last_governor);
+		if (policy->last_governor[0] != '\0')
+			gov = get_governor(policy->last_governor);
 		if (gov) {
 			pr_debug("Restoring governor %s for cpu %d\n",
 				 gov->name, policy->cpu);
-- 
2.25.1

Re: [PATCH v2] cpufreq: Avoid get_governor() for first policy

Posted by zhenglifeng (A) 2 months, 1 week ago

On 2025/7/25 12:14, Zihuan Zhang wrote:

> When a cpufreq driver registers the first policy, it may attempt to
> initialize the policy governor from `last_governor`. However, this is
> meaningless for the first policy instance, because `last_governor` is
> only updated when policies are removed (e.g. during CPU offline).
> 
> The `last_governor` mechanism is intended to restore the previously
> used governor across CPU hotplug events. For the very first policy,
> there is no "previous governor" to restore, so calling
> get_governor(last_governor) is unnecessary and potentially confusing.
> 
> This patch skips looking up `last_governor` when registering the first
> policy. Instead, it directly uses the default governor after all
> governors have been registered and are available.
> 
> This avoids meaningless lookups, reduces unnecessary module reference
> handling, and simplifies the initial policy path.
> 
> Signed-off-by: Zihuan Zhang <zhangzihuan@kylinos.cn>
> 
> ---
> Changes in v2:
>  - Fix the case where the governor is NULL.
> ---
> ---
>  drivers/cpufreq/cpufreq.c | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/cpufreq/cpufreq.c b/drivers/cpufreq/cpufreq.c
> index d7426e1d8bdd..1aa559f53479 100644
> --- a/drivers/cpufreq/cpufreq.c
> +++ b/drivers/cpufreq/cpufreq.c
> @@ -1122,7 +1122,8 @@ static int cpufreq_init_policy(struct cpufreq_policy *policy)
>  
>  	if (has_target()) {
>  		/* Update policy governor to the one used before hotplug. */
> -		gov = get_governor(policy->last_governor);
> +		if (policy->last_governor[0] != '\0')
> +			gov = get_governor(policy->last_governor);
>  		if (gov) {
>  			pr_debug("Restoring governor %s for cpu %d\n",
>  				 gov->name, policy->cpu);

Reviewed-by: Lifeng Zheng <zhenglifeng1@huawei.com>

Re: [PATCH v2] cpufreq: Avoid get_governor() for first policy

Posted by Rafael J. Wysocki 1 month, 2 weeks ago

On Fri, Jul 25, 2025 at 11:11 AM zhenglifeng (A)
<zhenglifeng1@huawei.com> wrote:
>
> On 2025/7/25 12:14, Zihuan Zhang wrote:
>
> > When a cpufreq driver registers the first policy, it may attempt to
> > initialize the policy governor from `last_governor`. However, this is
> > meaningless for the first policy instance, because `last_governor` is
> > only updated when policies are removed (e.g. during CPU offline).
> >
> > The `last_governor` mechanism is intended to restore the previously
> > used governor across CPU hotplug events. For the very first policy,
> > there is no "previous governor" to restore, so calling
> > get_governor(last_governor) is unnecessary and potentially confusing.
> >
> > This patch skips looking up `last_governor` when registering the first
> > policy. Instead, it directly uses the default governor after all
> > governors have been registered and are available.
> >
> > This avoids meaningless lookups, reduces unnecessary module reference
> > handling, and simplifies the initial policy path.
> >
> > Signed-off-by: Zihuan Zhang <zhangzihuan@kylinos.cn>
> >
> > ---
> > Changes in v2:
> >  - Fix the case where the governor is NULL.
> > ---
> > ---
> >  drivers/cpufreq/cpufreq.c | 3 ++-
> >  1 file changed, 2 insertions(+), 1 deletion(-)
> >
> > diff --git a/drivers/cpufreq/cpufreq.c b/drivers/cpufreq/cpufreq.c
> > index d7426e1d8bdd..1aa559f53479 100644
> > --- a/drivers/cpufreq/cpufreq.c
> > +++ b/drivers/cpufreq/cpufreq.c
> > @@ -1122,7 +1122,8 @@ static int cpufreq_init_policy(struct cpufreq_policy *policy)
> >
> >       if (has_target()) {
> >               /* Update policy governor to the one used before hotplug. */
> > -             gov = get_governor(policy->last_governor);
> > +             if (policy->last_governor[0] != '\0')
> > +                     gov = get_governor(policy->last_governor);
> >               if (gov) {
> >                       pr_debug("Restoring governor %s for cpu %d\n",
> >                                gov->name, policy->cpu);
>
> Reviewed-by: Lifeng Zheng <zhenglifeng1@huawei.com>

Applied as 6.18 material, thanks!

Re: [PATCH v2] cpufreq: Avoid get_governor() for first policy

Posted by Viresh Kumar 2 months, 1 week ago

On 25-07-25, 12:14, Zihuan Zhang wrote:
> When a cpufreq driver registers the first policy, it may attempt to
> initialize the policy governor from `last_governor`. However, this is
> meaningless for the first policy instance, because `last_governor` is
> only updated when policies are removed (e.g. during CPU offline).
> 
> The `last_governor` mechanism is intended to restore the previously
> used governor across CPU hotplug events. For the very first policy,
> there is no "previous governor" to restore, so calling
> get_governor(last_governor) is unnecessary and potentially confusing.
> 
> This patch skips looking up `last_governor` when registering the first
> policy. Instead, it directly uses the default governor after all
> governors have been registered and are available.
> 
> This avoids meaningless lookups, reduces unnecessary module reference
> handling, and simplifies the initial policy path.
> 
> Signed-off-by: Zihuan Zhang <zhangzihuan@kylinos.cn>
> 
> ---
> Changes in v2:
>  - Fix the case where the governor is NULL.
> ---
> ---
>  drivers/cpufreq/cpufreq.c | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/cpufreq/cpufreq.c b/drivers/cpufreq/cpufreq.c
> index d7426e1d8bdd..1aa559f53479 100644
> --- a/drivers/cpufreq/cpufreq.c
> +++ b/drivers/cpufreq/cpufreq.c
> @@ -1122,7 +1122,8 @@ static int cpufreq_init_policy(struct cpufreq_policy *policy)
>  
>  	if (has_target()) {
>  		/* Update policy governor to the one used before hotplug. */
> -		gov = get_governor(policy->last_governor);
> +		if (policy->last_governor[0] != '\0')
> +			gov = get_governor(policy->last_governor);
>  		if (gov) {
>  			pr_debug("Restoring governor %s for cpu %d\n",
>  				 gov->name, policy->cpu);

Acked-by: Viresh Kumar <viresh.kumar@linaro.org>

-- 
viresh