With the addition of following patch, related_cpus is required to be set by cpufreq platform drivers:
commit c1070fd743533efb54e98142252283583f379190 Author: Viresh Kumar viresh.kumar@linaro.org Date: Mon Jan 14 13:23:04 2013 +0000
cpufreq: Simplify cpufreq_add_dev()
Because this change is required by all platform drivers, why not do this in the core itself. Hence, this patch is an attempt towards fixing all broken drivers.
From now on, platforms don't really need to set related_cpus from their init()
routines, as the same work is done by core too.
If platform driver needs to set the related_cpus mask with some additional cpus, other than cpus present in policy->cpus, they are free to do it as we aren't overriding anything.
Signed-off-by: Viresh Kumar viresh.kumar@linaro.org --- Inderpal,
Can you please add your tested-by for this patch? And this will require you to drop your patch for exynos-cpufreq.c :)
drivers/cpufreq/cpufreq.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/drivers/cpufreq/cpufreq.c b/drivers/cpufreq/cpufreq.c index f5dc02b..db81382 100644 --- a/drivers/cpufreq/cpufreq.c +++ b/drivers/cpufreq/cpufreq.c @@ -554,8 +554,6 @@ static ssize_t show_cpus(const struct cpumask *mask, char *buf) */ static ssize_t show_related_cpus(struct cpufreq_policy *policy, char *buf) { - if (cpumask_empty(policy->related_cpus)) - return show_cpus(policy->cpus, buf); return show_cpus(policy->related_cpus, buf); }
@@ -945,6 +943,9 @@ static int cpufreq_add_dev(struct device *dev, struct subsys_interface *sif) goto err_unlock_policy; }
+ /* related cpus should atleast have policy->cpus */ + cpumask_or(policy->related_cpus, policy->related_cpus, policy->cpus); + /* * affected cpus must always be the one, which are online. We aren't * managing offline cpus here.
This reverts commit 956f33948b95aa91f6cbc6860087671c6ac1de4b.
With the addition of following patch, this change/variable is not required:
commit b9ba2725343ae57add3f324dfa5074167f48de96 Author: Viresh Kumar viresh.kumar@linaro.org Date: Mon Jan 14 13:23:03 2013 +0000
cpufreq: Simplify __cpufreq_remove_dev()
Signed-off-by: Viresh Kumar viresh.kumar@linaro.org --- drivers/cpufreq/cpufreq.c | 7 ------- 1 file changed, 7 deletions(-)
diff --git a/drivers/cpufreq/cpufreq.c b/drivers/cpufreq/cpufreq.c index db81382..8d521422 100644 --- a/drivers/cpufreq/cpufreq.c +++ b/drivers/cpufreq/cpufreq.c @@ -47,9 +47,6 @@ static DEFINE_PER_CPU(char[CPUFREQ_NAME_LEN], cpufreq_cpu_governor); #endif static DEFINE_SPINLOCK(cpufreq_driver_lock);
-/* Used when we unregister cpufreq driver */ -static struct cpumask cpufreq_online_mask; - /* * cpu_policy_rwsem is a per CPU reader-writer semaphore designed to cure * all cpufreq/hotplug/workqueue/etc related lock issues. @@ -951,7 +948,6 @@ static int cpufreq_add_dev(struct device *dev, struct subsys_interface *sif) * managing offline cpus here. */ cpumask_and(policy->cpus, policy->cpus, cpu_online_mask); - cpumask_and(policy->cpus, policy->cpus, &cpufreq_online_mask);
policy->user_policy.min = policy->min; policy->user_policy.max = policy->max; @@ -1133,7 +1129,6 @@ static int cpufreq_remove_dev(struct device *dev, struct subsys_interface *sif) if (unlikely(lock_policy_rwsem_write(cpu))) BUG();
- cpumask_clear_cpu(cpu, &cpufreq_online_mask); retval = __cpufreq_remove_dev(dev, sif); return retval; } @@ -1875,8 +1870,6 @@ int cpufreq_register_driver(struct cpufreq_driver *driver_data) cpufreq_driver = driver_data; spin_unlock_irqrestore(&cpufreq_driver_lock, flags);
- cpumask_setall(&cpufreq_online_mask); - ret = subsys_interface_register(&cpufreq_interface); if (ret) goto err_null_driver;
On 29 January 2013 10:09, Viresh Kumar viresh.kumar@linaro.org wrote:
With the addition of following patch, related_cpus is required to be set by cpufreq platform drivers:
commit c1070fd743533efb54e98142252283583f379190 Author: Viresh Kumar viresh.kumar@linaro.org Date: Mon Jan 14 13:23:04 2013 +0000
cpufreq: Simplify cpufreq_add_dev()Because this change is required by all platform drivers, why not do this in the core itself. Hence, this patch is an attempt towards fixing all broken drivers.
From now on, platforms don't really need to set related_cpus from their init() routines, as the same work is done by core too.
If platform driver needs to set the related_cpus mask with some additional cpus, other than cpus present in policy->cpus, they are free to do it as we aren't overriding anything.
Signed-off-by: Viresh Kumar viresh.kumar@linaro.org
Inderpal,
Can you please add your tested-by for this patch? And this will require you to drop your patch for exynos-cpufreq.c :)
ARM mail servers are broken. Patches attached.
On Tuesday, January 29, 2013 10:09:59 AM Viresh Kumar wrote:
With the addition of following patch, related_cpus is required to be set by cpufreq platform drivers:
commit c1070fd743533efb54e98142252283583f379190 Author: Viresh Kumar viresh.kumar@linaro.org Date: Mon Jan 14 13:23:04 2013 +0000
cpufreq: Simplify cpufreq_add_dev()
I've dropped this one in the meantime.
Can you please fold the $subject patch into "cpufreq: Simplify cpufreq_add_dev()" and post the result instead? That surely will be less confusing?
Rafael
On 29 January 2013 17:21, Rafael J. Wysocki rjw@sisk.pl wrote:
On Tuesday, January 29, 2013 10:09:59 AM Viresh Kumar wrote:
With the addition of following patch, related_cpus is required to be set by cpufreq platform drivers:
commit c1070fd743533efb54e98142252283583f379190 Author: Viresh Kumar viresh.kumar@linaro.org Date: Mon Jan 14 13:23:04 2013 +0000
cpufreq: Simplify cpufreq_add_dev()I've dropped this one in the meantime.
Can you please fold the $subject patch into "cpufreq: Simplify cpufreq_add_dev()" and post the result instead? That surely will be less confusing?
Okay. I will squash this one with cpufreq_add_dev() one + following line:
diff --git a/drivers/cpufreq/spear-cpufreq.c b/drivers/cpufreq/spear-cpufreq.c index 8ff26af..fc714a6 100644 --- a/drivers/cpufreq/spear-cpufreq.c +++ b/drivers/cpufreq/spear-cpufreq.c @@ -189,7 +189,6 @@ static int spear_cpufreq_init(struct cpufreq_policy *policy) policy->cur = spear_cpufreq_get(0);
cpumask_copy(policy->cpus, topology_core_cpumask(policy->cpu)); - cpumask_copy(policy->related_cpus, policy->cpus);
return 0; }
Also, because you are happy loosing your commit history in linux-next, you can drop the patch that i have reverted as 2/2 of this set.
I will resend the rest again.
On Tuesday, January 29, 2013 08:00:23 PM Viresh Kumar wrote:
On 29 January 2013 17:21, Rafael J. Wysocki rjw@sisk.pl wrote:
On Tuesday, January 29, 2013 10:09:59 AM Viresh Kumar wrote:
With the addition of following patch, related_cpus is required to be set by cpufreq platform drivers:
commit c1070fd743533efb54e98142252283583f379190 Author: Viresh Kumar viresh.kumar@linaro.org Date: Mon Jan 14 13:23:04 2013 +0000
cpufreq: Simplify cpufreq_add_dev()I've dropped this one in the meantime.
Can you please fold the $subject patch into "cpufreq: Simplify cpufreq_add_dev()" and post the result instead? That surely will be less confusing?
Okay. I will squash this one with cpufreq_add_dev() one + following line:
diff --git a/drivers/cpufreq/spear-cpufreq.c b/drivers/cpufreq/spear-cpufreq.c index 8ff26af..fc714a6 100644 --- a/drivers/cpufreq/spear-cpufreq.c +++ b/drivers/cpufreq/spear-cpufreq.c @@ -189,7 +189,6 @@ static int spear_cpufreq_init(struct cpufreq_policy *policy) policy->cur = spear_cpufreq_get(0);
cpumask_copy(policy->cpus, topology_core_cpumask(policy->cpu));
cpumask_copy(policy->related_cpus, policy->cpus); return 0;}
Also, because you are happy loosing your commit history in linux-next, you can drop the patch that i have reverted as 2/2 of this set.
Well, I'm not attached to the linux-next commit history, but also it's a pain to change it too oftern. :-)
I generally avoid changing it unless there are build issues and such that would cause pain to people doing bisection, for example.
Thanks, Rafael
-
Rafael J. Wysocki -
Viresh Kumar