[PATCH v2 0/2] cpufreq: Return -EOPNOTSUPP if no policy is boost supported

[PATCH v2 0/2] cpufreq: Return -EOPNOTSUPP if no policy is boost supported

[Lifeng Zheng]

Make cpufreq_boost_trigger_state() return error when boost isn't supported
by any policy.

Changelog:

v2:

 - change the error to -EOPNOTSUPP
 - remove the comment and blank line
 - split the optimization into another patch

---
Discussions of previous version:
v1: https://lore.kernel.org/all/20251126031916.3641176-1-zhenglifeng1@huawei.com/

Lifeng Zheng (2):
  cpufreq: Return -EOPNOTSUPP if no policy is boost supported
  cpufreq: cpufreq_boost_trigger_state() optimization

 drivers/cpufreq/cpufreq.c | 11 +++++------
 1 file changed, 5 insertions(+), 6 deletions(-)

-- 
2.33.0

[PATCH v2 1/2] cpufreq: Return -EOPNOTSUPP if no policy is boost supported

[Lifeng Zheng]

In cpufreq_boost_trigger_state(), if all the policies are boost
unsupported, policy_set_boost() will not be called and this function will
return 0. But it is better to return an error to indicate that the platform
doesn't support boost.

Signed-off-by: Lifeng Zheng <zhenglifeng1@huawei.com>
---
 drivers/cpufreq/cpufreq.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/drivers/cpufreq/cpufreq.c b/drivers/cpufreq/cpufreq.c
index e8d7544b77b8..a4399e5490da 100644
--- a/drivers/cpufreq/cpufreq.c
+++ b/drivers/cpufreq/cpufreq.c
@@ -2806,7 +2806,7 @@ static int cpufreq_boost_trigger_state(int state)
 {
 	struct cpufreq_policy *policy;
 	unsigned long flags;
-	int ret = 0;
+	int ret = -EOPNOTSUPP;
 
 	/*
 	 * Don't compare 'cpufreq_driver->boost_enabled' with 'state' here to
@@ -2826,6 +2826,10 @@ static int cpufreq_boost_trigger_state(int state)
 		if (ret)
 			goto err_reset_state;
 	}
+
+	if (ret)
+		goto err_reset_state;
+
 	cpus_read_unlock();
 
 	return 0;
-- 
2.33.0

[PATCH v2 2/2] cpufreq: cpufreq_boost_trigger_state() optimization

[Lifeng Zheng]

Simplify the error handling branch code in cpufreq_boost_trigger_state().

Signed-off-by: Lifeng Zheng <zhenglifeng1@huawei.com>
---
 drivers/cpufreq/cpufreq.c | 11 +++--------
 1 file changed, 3 insertions(+), 8 deletions(-)

diff --git a/drivers/cpufreq/cpufreq.c b/drivers/cpufreq/cpufreq.c
index a4399e5490da..a725747572c9 100644
--- a/drivers/cpufreq/cpufreq.c
+++ b/drivers/cpufreq/cpufreq.c
@@ -2824,18 +2824,13 @@ static int cpufreq_boost_trigger_state(int state)
 
 		ret = policy_set_boost(policy, state);
 		if (ret)
-			goto err_reset_state;
+			break;
 	}
 
-	if (ret)
-		goto err_reset_state;
-
 	cpus_read_unlock();
 
-	return 0;
-
-err_reset_state:
-	cpus_read_unlock();
+	if (!ret)
+		return 0;
 
 	write_lock_irqsave(&cpufreq_driver_lock, flags);
 	cpufreq_driver->boost_enabled = !state;
-- 
2.33.0

Re: [PATCH v2 2/2] cpufreq: cpufreq_boost_trigger_state() optimization

[Viresh Kumar]

On 28-11-25, 17:13, Lifeng Zheng wrote:
> Simplify the error handling branch code in cpufreq_boost_trigger_state().
> 
> Signed-off-by: Lifeng Zheng <zhenglifeng1@huawei.com>
> ---
>  drivers/cpufreq/cpufreq.c | 11 +++--------
>  1 file changed, 3 insertions(+), 8 deletions(-)
> 
> diff --git a/drivers/cpufreq/cpufreq.c b/drivers/cpufreq/cpufreq.c
> index a4399e5490da..a725747572c9 100644
> --- a/drivers/cpufreq/cpufreq.c
> +++ b/drivers/cpufreq/cpufreq.c
> @@ -2824,18 +2824,13 @@ static int cpufreq_boost_trigger_state(int state)
>  
>  		ret = policy_set_boost(policy, state);
>  		if (ret)
> -			goto err_reset_state;
> +			break;
>  	}
>  
> -	if (ret)
> -		goto err_reset_state;
> -
>  	cpus_read_unlock();
>  
> -	return 0;
> -
> -err_reset_state:
> -	cpus_read_unlock();
> +	if (!ret)

Maybe we can make this `if (likely(!ret))`

> +		return 0;
>  
>  	write_lock_irqsave(&cpufreq_driver_lock, flags);
>  	cpufreq_driver->boost_enabled = !state;

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

-- 
viresh

Re: [PATCH v2 1/2] cpufreq: Return -EOPNOTSUPP if no policy is boost supported

[Viresh Kumar]

On 28-11-25, 17:13, Lifeng Zheng wrote:
> In cpufreq_boost_trigger_state(), if all the policies are boost
> unsupported, policy_set_boost() will not be called and this function will
> return 0. But it is better to return an error to indicate that the platform
> doesn't support boost.
> 
> Signed-off-by: Lifeng Zheng <zhenglifeng1@huawei.com>
> ---
>  drivers/cpufreq/cpufreq.c | 6 +++++-
>  1 file changed, 5 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/cpufreq/cpufreq.c b/drivers/cpufreq/cpufreq.c
> index e8d7544b77b8..a4399e5490da 100644
> --- a/drivers/cpufreq/cpufreq.c
> +++ b/drivers/cpufreq/cpufreq.c
> @@ -2806,7 +2806,7 @@ static int cpufreq_boost_trigger_state(int state)
>  {
>  	struct cpufreq_policy *policy;
>  	unsigned long flags;
> -	int ret = 0;
> +	int ret = -EOPNOTSUPP;
>  
>  	/*
>  	 * Don't compare 'cpufreq_driver->boost_enabled' with 'state' here to
> @@ -2826,6 +2826,10 @@ static int cpufreq_boost_trigger_state(int state)
>  		if (ret)
>  			goto err_reset_state;
>  	}
> +
> +	if (ret)
> +		goto err_reset_state;
> +
>  	cpus_read_unlock();
>  
>  	return 0;

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

-- 
viresh

Re: [PATCH v2 2/2] cpufreq: cpufreq_boost_trigger_state() optimization

[zhenglifeng (A)]

On 2025/12/1 11:42, Viresh Kumar wrote:
> On 28-11-25, 17:13, Lifeng Zheng wrote:
>> Simplify the error handling branch code in cpufreq_boost_trigger_state().
>>
>> Signed-off-by: Lifeng Zheng <zhenglifeng1@huawei.com>
>> ---
>>  drivers/cpufreq/cpufreq.c | 11 +++--------
>>  1 file changed, 3 insertions(+), 8 deletions(-)
>>
>> diff --git a/drivers/cpufreq/cpufreq.c b/drivers/cpufreq/cpufreq.c
>> index a4399e5490da..a725747572c9 100644
>> --- a/drivers/cpufreq/cpufreq.c
>> +++ b/drivers/cpufreq/cpufreq.c
>> @@ -2824,18 +2824,13 @@ static int cpufreq_boost_trigger_state(int state)
>>  
>>  		ret = policy_set_boost(policy, state);
>>  		if (ret)
>> -			goto err_reset_state;
>> +			break;
>>  	}
>>  
>> -	if (ret)
>> -		goto err_reset_state;
>> -
>>  	cpus_read_unlock();
>>  
>> -	return 0;
>> -
>> -err_reset_state:
>> -	cpus_read_unlock();
>> +	if (!ret)
> 
> Maybe we can make this `if (likely(!ret))`

For the platforms which are not boost supported, this will never be
matched. Is `likely` OK in this situation?

> 
>> +		return 0;
>>  
>>  	write_lock_irqsave(&cpufreq_driver_lock, flags);
>>  	cpufreq_driver->boost_enabled = !state;
> 
> Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
> 

Re: [PATCH v2 2/2] cpufreq: cpufreq_boost_trigger_state() optimization

[Viresh Kumar]

On 02-12-25, 09:32, zhenglifeng (A) wrote:
> On 2025/12/1 11:42, Viresh Kumar wrote:
> > On 28-11-25, 17:13, Lifeng Zheng wrote:
> >> Simplify the error handling branch code in cpufreq_boost_trigger_state().
> >>
> >> Signed-off-by: Lifeng Zheng <zhenglifeng1@huawei.com>
> >> ---
> >>  drivers/cpufreq/cpufreq.c | 11 +++--------
> >>  1 file changed, 3 insertions(+), 8 deletions(-)
> >>
> >> diff --git a/drivers/cpufreq/cpufreq.c b/drivers/cpufreq/cpufreq.c
> >> index a4399e5490da..a725747572c9 100644
> >> --- a/drivers/cpufreq/cpufreq.c
> >> +++ b/drivers/cpufreq/cpufreq.c
> >> @@ -2824,18 +2824,13 @@ static int cpufreq_boost_trigger_state(int state)
> >>  
> >>  		ret = policy_set_boost(policy, state);
> >>  		if (ret)
> >> -			goto err_reset_state;
> >> +			break;
> >>  	}
> >>  
> >> -	if (ret)
> >> -		goto err_reset_state;
> >> -
> >>  	cpus_read_unlock();
> >>  
> >> -	return 0;
> >> -
> >> -err_reset_state:
> >> -	cpus_read_unlock();
> >> +	if (!ret)
> > 
> > Maybe we can make this `if (likely(!ret))`
> 
> For the platforms which are not boost supported, this will never be
> matched. Is `likely` OK in this situation?

Ideally they won't have a `boost` file in sysfs, and if they have it, we don't
really need to optimize the failure case.

-- 
viresh

Re: [PATCH v2 2/2] cpufreq: cpufreq_boost_trigger_state() optimization

[zhenglifeng (A)]

On 2025/12/2 12:58, Viresh Kumar wrote:
> On 02-12-25, 09:32, zhenglifeng (A) wrote:
>> On 2025/12/1 11:42, Viresh Kumar wrote:
>>> On 28-11-25, 17:13, Lifeng Zheng wrote:
>>>> Simplify the error handling branch code in cpufreq_boost_trigger_state().
>>>>
>>>> Signed-off-by: Lifeng Zheng <zhenglifeng1@huawei.com>
>>>> ---
>>>>  drivers/cpufreq/cpufreq.c | 11 +++--------
>>>>  1 file changed, 3 insertions(+), 8 deletions(-)
>>>>
>>>> diff --git a/drivers/cpufreq/cpufreq.c b/drivers/cpufreq/cpufreq.c
>>>> index a4399e5490da..a725747572c9 100644
>>>> --- a/drivers/cpufreq/cpufreq.c
>>>> +++ b/drivers/cpufreq/cpufreq.c
>>>> @@ -2824,18 +2824,13 @@ static int cpufreq_boost_trigger_state(int state)
>>>>  
>>>>  		ret = policy_set_boost(policy, state);
>>>>  		if (ret)
>>>> -			goto err_reset_state;
>>>> +			break;
>>>>  	}
>>>>  
>>>> -	if (ret)
>>>> -		goto err_reset_state;
>>>> -
>>>>  	cpus_read_unlock();
>>>>  
>>>> -	return 0;
>>>> -
>>>> -err_reset_state:
>>>> -	cpus_read_unlock();
>>>> +	if (!ret)
>>>
>>> Maybe we can make this `if (likely(!ret))`
>>
>> For the platforms which are not boost supported, this will never be
>> matched. Is `likely` OK in this situation?
> 
> Ideally they won't have a `boost` file in sysfs, and if they have it, we don't
> really need to optimize the failure case.

I see. Then I think the `if (ret)` in the loop should be
`if (unlikely(ret))` too.

Re: [PATCH v2 2/2] cpufreq: cpufreq_boost_trigger_state() optimization

[Viresh Kumar]

On 02-12-25, 14:24, zhenglifeng (A) wrote:
> On 2025/12/2 12:58, Viresh Kumar wrote:
> > On 02-12-25, 09:32, zhenglifeng (A) wrote:
> >> On 2025/12/1 11:42, Viresh Kumar wrote:
> >>> On 28-11-25, 17:13, Lifeng Zheng wrote:
> >>>> Simplify the error handling branch code in cpufreq_boost_trigger_state().
> >>>>
> >>>> Signed-off-by: Lifeng Zheng <zhenglifeng1@huawei.com>
> >>>> ---
> >>>>  drivers/cpufreq/cpufreq.c | 11 +++--------
> >>>>  1 file changed, 3 insertions(+), 8 deletions(-)
> >>>>
> >>>> diff --git a/drivers/cpufreq/cpufreq.c b/drivers/cpufreq/cpufreq.c
> >>>> index a4399e5490da..a725747572c9 100644
> >>>> --- a/drivers/cpufreq/cpufreq.c
> >>>> +++ b/drivers/cpufreq/cpufreq.c
> >>>> @@ -2824,18 +2824,13 @@ static int cpufreq_boost_trigger_state(int state)
> >>>>  
> >>>>  		ret = policy_set_boost(policy, state);
> >>>>  		if (ret)
> >>>> -			goto err_reset_state;
> >>>> +			break;
> >>>>  	}
> >>>>  
> >>>> -	if (ret)
> >>>> -		goto err_reset_state;
> >>>> -
> >>>>  	cpus_read_unlock();
> >>>>  
> >>>> -	return 0;
> >>>> -
> >>>> -err_reset_state:
> >>>> -	cpus_read_unlock();
> >>>> +	if (!ret)
> >>>
> >>> Maybe we can make this `if (likely(!ret))`
> >>
> >> For the platforms which are not boost supported, this will never be
> >> matched. Is `likely` OK in this situation?
> > 
> > Ideally they won't have a `boost` file in sysfs, and if they have it, we don't
> > really need to optimize the failure case.
> 
> I see. Then I think the `if (ret)` in the loop should be
> `if (unlikely(ret))` too.

That can be done too.

-- 
viresh