diff mbox

[v3,1/2] cpufreq: Don't destroy/realloc policy/sysfs on hotplug/suspend

Message ID CAKohpomt34NcMH=H5L_fhf-6R6p27_h=aCNcpT6T12u35d=Z+w@mail.gmail.com (mailing list archive)
State New, archived
Headers show

Commit Message

Viresh Kumar July 16, 2014, 8:30 a.m. UTC
On 16 July 2014 05:58, Saravana Kannan <skannan@codeaurora.org> wrote:
>> +       if (!cpus && cpufreq_driver->stop_cpu &&
>> cpufreq_driver->setpolicy) {
>> +               cpufreq_driver->stop_cpu(policy);
>> +       }
>>
>
> Viresh, I tried your suggestion (and my initial thought too) to combine this
> as an if/else with the previous if. But the indentation got nasty and made
> it hard to read. I'm sure the compiler will optimize it. So, I would prefer
> to leave it this way.

Okay, I gave that comment again :)

Try this:

                        ret = __cpufreq_governor(policy, CPUFREQ_GOV_LIMITS);
@@ -1001,9 +1003,7 @@ static int cpufreq_change_policy_cpus(struct
cpufreq_policy *policy,
                        pr_err("%s: Failed to start governor\n", __func__);
                        goto unlock;
                }
-       }
-
-       if (!cpus && cpufreq_driver->stop_cpu && cpufreq_driver->setpolicy) {
+       } else if (cpufreq_driver->stop_cpu && cpufreq_driver->setpolicy) {
                cpufreq_driver->stop_cpu(policy);
        }
--
To unsubscribe from this list: send the line "unsubscribe linux-arm-msm" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Comments

Saravana Kannan July 16, 2014, 7:19 p.m. UTC | #1
On 07/16/2014 01:30 AM, Viresh Kumar wrote:
> On 16 July 2014 05:58, Saravana Kannan <skannan@codeaurora.org> wrote:
>>> +       if (!cpus && cpufreq_driver->stop_cpu &&
>>> cpufreq_driver->setpolicy) {
>>> +               cpufreq_driver->stop_cpu(policy);
>>> +       }
>>>
>>
>> Viresh, I tried your suggestion (and my initial thought too) to combine this
>> as an if/else with the previous if. But the indentation got nasty and made
>> it hard to read. I'm sure the compiler will optimize it. So, I would prefer
>> to leave it this way.
>
> Okay, I gave that comment again :)
>
> Try this:
>
> diff --git a/drivers/cpufreq/cpufreq.c b/drivers/cpufreq/cpufreq.c
> index f72b2b7..092a0ba 100644
> --- a/drivers/cpufreq/cpufreq.c
> +++ b/drivers/cpufreq/cpufreq.c
> @@ -991,8 +991,10 @@ static int cpufreq_change_policy_cpus(struct
> cpufreq_policy *policy,
>                                          CPUFREQ_UPDATE_POLICY_CPU, policy);
>          }
>
> -       cpus = !cpumask_empty(policy->cpus);
> -       if (has_target() && cpus) {
> +       if (!cpumask_empty(policy->cpus)) {
> +               if (!has_target())
> +                       goto unlock;
> +
>                  ret = __cpufreq_governor(policy, CPUFREQ_GOV_START);
>                  if (!ret)
>                          ret = __cpufreq_governor(policy, CPUFREQ_GOV_LIMITS);
> @@ -1001,9 +1003,7 @@ static int cpufreq_change_policy_cpus(struct
> cpufreq_policy *policy,
>                          pr_err("%s: Failed to start governor\n", __func__);
>                          goto unlock;
>                  }
> -       }
> -
> -       if (!cpus && cpufreq_driver->stop_cpu && cpufreq_driver->setpolicy) {
> +       } else if (cpufreq_driver->stop_cpu && cpufreq_driver->setpolicy) {
>                  cpufreq_driver->stop_cpu(policy);
>          }
>

Not bad :) I'll take it.

-Saravana
diff mbox

Patch

diff --git a/drivers/cpufreq/cpufreq.c b/drivers/cpufreq/cpufreq.c
index f72b2b7..092a0ba 100644
--- a/drivers/cpufreq/cpufreq.c
+++ b/drivers/cpufreq/cpufreq.c
@@ -991,8 +991,10 @@  static int cpufreq_change_policy_cpus(struct
cpufreq_policy *policy,
                                        CPUFREQ_UPDATE_POLICY_CPU, policy);
        }

-       cpus = !cpumask_empty(policy->cpus);
-       if (has_target() && cpus) {
+       if (!cpumask_empty(policy->cpus)) {
+               if (!has_target())
+                       goto unlock;
+
                ret = __cpufreq_governor(policy, CPUFREQ_GOV_START);
                if (!ret)