2021-02-02 09:42:15

by Xuewen Yan

Subject: [PATCH] cpufreq: Judging new_policy before update related_cpus

From: Xuewen Yan <[email protected]>

When the policy->related_cpus are all offline, and then
bring up one cpu, this time, if the ->online is NULL,
the code would update the ->related_cpus with ->cpus,
and now ->cpus is only one online cpu, as a result, the
->related_cpus is different from the origion ->related_cpus.

Signed-off-by: Xuewen Yan <[email protected]>
---
drivers/cpufreq/cpufreq.c | 6 ++++--
1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/drivers/cpufreq/cpufreq.c b/drivers/cpufreq/cpufreq.c
index d0a3525ce27f..3d512ac463a5 100644
--- a/drivers/cpufreq/cpufreq.c
+++ b/drivers/cpufreq/cpufreq.c
@@ -1374,8 +1374,10 @@ static int cpufreq_online(unsigned int cpu)
if (ret)
goto out_exit_policy;

- /* related_cpus should at least include policy->cpus. */
- cpumask_copy(policy->related_cpus, policy->cpus);
+ if (new_policy) {
+ /* related_cpus should at least include policy->cpus. */
+ cpumask_copy(policy->related_cpus, policy->cpus);
+ }
}

down_write(&policy->rwsem);
--
2.29.0

2021-02-02 09:45:57

by Viresh Kumar

Subject: Re: [PATCH] cpufreq: Judging new_policy before update related_cpus

On 02-02-21, 17:36, Xuewen Yan wrote:
> From: Xuewen Yan <[email protected]>
>
> When the policy->related_cpus are all offline, and then
> bring up one cpu, this time, if the ->online is NULL,
> the code would update the ->related_cpus with ->cpus,
> and now ->cpus is only one online cpu, as a result, the
> ->related_cpus is different from the origion ->related_cpus.
>
> Signed-off-by: Xuewen Yan <[email protected]>
> ---
> drivers/cpufreq/cpufreq.c | 6 ++++--
> 1 file changed, 4 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/cpufreq/cpufreq.c b/drivers/cpufreq/cpufreq.c
> index d0a3525ce27f..3d512ac463a5 100644
> --- a/drivers/cpufreq/cpufreq.c
> +++ b/drivers/cpufreq/cpufreq.c
> @@ -1374,8 +1374,10 @@ static int cpufreq_online(unsigned int cpu)
> if (ret)
> goto out_exit_policy;
>
> - /* related_cpus should at least include policy->cpus. */
> - cpumask_copy(policy->related_cpus, policy->cpus);
> + if (new_policy) {
> + /* related_cpus should at least include policy->cpus. */
> + cpumask_copy(policy->related_cpus, policy->cpus);
> + }
> }

The driver's ->init() callback is responsible for filling policy->cpus
here.

--
viresh