[PATCH 0/4] CPUFreq stats: Bug fixes

[PATCH 0/4] CPUFreq stats: Bug fixes

[Viresh Kumar]

Hi Rafael,

I know you asked me not to send any more patches before the earlier ones get
into kernel. I got to this as Nicolas Pitre required to send few CPUFreq patches
for ARM's big LITTLE In-Kernel-Switcher. And within linaro we have hacked these
bugs in a bad way..

Because of his dependency I am forced to send these.. These aren't introduced
recently and so they can be included in 3.13.

There are several problems/bugs in cpufreq-stats specially with cpufreq drivers
as modules and suspend/resume path. These are mentioned well in changelogs.

These are tested over my thinkpad (acpi-cpufreq) in following way:
[1] offline+online all CPUs except boot cpu in a while loop
[2] then do suspend resume
[3] repeat [1] and [2] several times.

No issues found..

Also tested on my exynos board:
- Added cpufreq_unregister/register while loop in exynos-cpufreq.c so that we
  continuously register/unregister driver... Stats were working fine now..
- Compile cpufreq-stats as module and insert/remove it several times after
  removing above hack (as that doesn't let linux boot :) )..

@Srivatsa: You also have fairly good idea of cpufreq now, so please give some
time to review this :)

@Nico: Can you remove the hacky code from IKS tree and test these instead to see
if we still have any issues?

--
viresh

Viresh Kumar (4):
  cpufreq: stats: handle cpufreq_unregister_driver() and suspend/resume
    properly
  cpufreq: stats: remove hotplug notifiers
  cpufreq: stats: free table and remove sysfs entry in a single routine
  cpufreq: stats: create sysfs entries when cpufreq_stats is a module

 drivers/cpufreq/cpufreq.c       |   5 ++
 drivers/cpufreq/cpufreq_stats.c | 109 ++++++++++++++++++----------------------
 include/linux/cpufreq.h         |   2 +
 3 files changed, 55 insertions(+), 61 deletions(-)

-- 
1.7.12.rc2.18.g61b472e

[PATCH 1/4] cpufreq: stats: handle cpufreq_unregister_driver() and suspend/resume properly

[Viresh Kumar]

There are several problems cpufreq stats in the way it handles
cpufreq_unregister_driver() and suspend/resume..

- We must not loose data collected so far when suspend/resume happens and so
  stats directories must not be removed/allocated during these operations, Which
  is done currently.

- cpufreq_stat has registered notifiers with both cpufreq and hotplug. It adds
  sysfs stats directory with a cpufreq notifier: CPUFREQ_NOTIFY and removes this
  directory with a notifier from hotplug core.

  In case cpufreq_unregister_driver() is called (on rmmod cpufreq driver), stats
  directories per cpu aren't removed as CPUs are still online. The only call
  cpufreq_stats gets is cpufreq_stats_update_policy_cpu for all CPUs except the
  last of each policy. And pointer to stat information is stored in the entry
  for last cpu in per-cpu cpufreq_stats_table. But policy structure would be
  freed inside cpufreq core and so that will result in memory leak inside
  cpufreq stats (as we are never freeing memory for stats).

  Now if we again insert the module cpufreq_register_driver() will be called and
  we will again allocate stats data and put it on for first cpu of every policy.
  In case we only have a single cpu per policy we will return with a error from
  cpufreq_stats_create_table() due to below code:

	if (per_cpu(cpufreq_stats_table, cpu))
		return -EBUSY;

  And so probably cpufreq stats directory would show up anymore (as it was added
  inside last policies->kobj which doesn't exist anymore). I haven't tested it
  though. Also the values in stats files wouldn't be refreshed as we are using
  the earlier stats structure.

- CPUFREQ_NOTIFY is called from cpufreq_set_policy() which is called for
  scenarios where we don't really want cpufreq_stat_notifier_policy() to get
  called. For example whenever we are changing anything related to a policy:
  min/max/current freq, etc.. cpufreq_set_policy() is called and so cpufreq
  stats is notified. Where we don't do any useful stuff other than simply
  returning with -EBUSY from cpufreq_stats_create_table(). And so this isn't the
  right notifier that cpufreq stats..

Due to all above reasons this patch does following changes:
- Add new notifiers CPUFREQ_CREATE_POLICY and CPUFREQ_REMOVE_POLICY, which are
  only called when policy is created/destroyed. They aren't called for
  suspend/resume paths..
- Use these notifiers in cpufreq_stat_notifier_policy() to create/destory stats
  sysfs entries. And so cpufreq_unregister_driver() or suspend/resume shouldn't
  be a problem for cpufreq_stats.
- Return early from cpufreq_stat_cpu_callback() for suspend/resume sequence, so
  that we don't free stats structure.

Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
---
 drivers/cpufreq/cpufreq.c       |  5 +++++
 drivers/cpufreq/cpufreq_stats.c | 24 +++++++++++++++++-------
 include/linux/cpufreq.h         |  2 ++
 3 files changed, 24 insertions(+), 7 deletions(-)

diff --git a/drivers/cpufreq/cpufreq.c b/drivers/cpufreq/cpufreq.c
index 89b3c52..ab56531 100644
--- a/drivers/cpufreq/cpufreq.c
+++ b/drivers/cpufreq/cpufreq.c
@@ -1086,6 +1086,8 @@ static int __cpufreq_add_dev(struct device *dev, struct subsys_interface *sif,
 		ret = cpufreq_add_dev_interface(policy, dev);
 		if (ret)
 			goto err_out_unregister;
+		blocking_notifier_call_chain(&cpufreq_policy_notifier_list,
+				CPUFREQ_CREATE_POLICY, policy);
 	}
 
 	write_lock_irqsave(&cpufreq_driver_lock, flags);
@@ -1263,6 +1265,9 @@ static int __cpufreq_remove_dev_finish(struct device *dev,
 		}
 
 		if (!frozen) {
+			blocking_notifier_call_chain(&cpufreq_policy_notifier_list,
+					CPUFREQ_REMOVE_POLICY, policy);
+
 			lock_policy_rwsem_read(cpu);
 			kobj = &policy->kobj;
 			cmp = &policy->kobj_unregister;
diff --git a/drivers/cpufreq/cpufreq_stats.c b/drivers/cpufreq/cpufreq_stats.c
index 4cf0d28..0f71562 100644
--- a/drivers/cpufreq/cpufreq_stats.c
+++ b/drivers/cpufreq/cpufreq_stats.c
@@ -277,7 +277,7 @@ static void cpufreq_stats_update_policy_cpu(struct cpufreq_policy *policy)
 static int cpufreq_stat_notifier_policy(struct notifier_block *nb,
 		unsigned long val, void *data)
 {
-	int ret;
+	int ret = 0;
 	struct cpufreq_policy *policy = data;
 	struct cpufreq_frequency_table *table;
 	unsigned int cpu = policy->cpu;
@@ -287,15 +287,21 @@ static int cpufreq_stat_notifier_policy(struct notifier_block *nb,
 		return 0;
 	}
 
-	if (val != CPUFREQ_NOTIFY)
-		return 0;
 	table = cpufreq_frequency_get_table(cpu);
 	if (!table)
 		return 0;
-	ret = cpufreq_stats_create_table(policy, table);
-	if (ret)
-		return ret;
-	return 0;
+
+	if (val == CPUFREQ_CREATE_POLICY)
+		ret = cpufreq_stats_create_table(policy, table);
+	else if (val == CPUFREQ_REMOVE_POLICY) {
+		/* This might already be freed by cpu hotplug notifier */
+		if (per_cpu(cpufreq_stats_table, cpu)) {
+			cpufreq_stats_free_sysfs(cpu);
+			cpufreq_stats_free_table(cpu);
+		}
+	}
+
+	return ret;
 }
 
 static int cpufreq_stat_notifier_trans(struct notifier_block *nb,
@@ -340,6 +346,10 @@ static int cpufreq_stat_cpu_callback(struct notifier_block *nfb,
 {
 	unsigned int cpu = (unsigned long)hcpu;
 
+	/* Don't free/allocate stats during suspend/resume */
+	if (action & CPU_TASKS_FROZEN)
+		return 0;
+
 	switch (action) {
 	case CPU_DOWN_PREPARE:
 		cpufreq_stats_free_sysfs(cpu);
diff --git a/include/linux/cpufreq.h b/include/linux/cpufreq.h
index fcabc42..9132fa3 100644
--- a/include/linux/cpufreq.h
+++ b/include/linux/cpufreq.h
@@ -259,6 +259,8 @@ static inline void cpufreq_verify_within_limits(struct cpufreq_policy *policy,
 #define CPUFREQ_NOTIFY			(2)
 #define CPUFREQ_START			(3)
 #define CPUFREQ_UPDATE_POLICY_CPU	(4)
+#define CPUFREQ_CREATE_POLICY		(5)
+#define CPUFREQ_REMOVE_POLICY		(6)
 
 #ifdef CONFIG_CPU_FREQ
 int cpufreq_register_notifier(struct notifier_block *nb, unsigned int list);
-- 
1.7.12.rc2.18.g61b472e

[PATCH 2/4] cpufreq: stats: remove hotplug notifiers

[Viresh Kumar]

Either CPUs are hot-unplugged or suspend/resume occurs, cpufreq core will send
notifications to cpufreq-stats and stats structure and sysfs entries would be
correctly handled..

And so we don't actually need hotcpu notifiers in cpufreq-stats anymore. We were
only handling cpu hot-unplug events here and that are already taken care of by
POLICY notifiers.

Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
---
 drivers/cpufreq/cpufreq_stats.c | 38 ++------------------------------------
 1 file changed, 2 insertions(+), 36 deletions(-)

diff --git a/drivers/cpufreq/cpufreq_stats.c b/drivers/cpufreq/cpufreq_stats.c
index 0f71562..8fa5844 100644
--- a/drivers/cpufreq/cpufreq_stats.c
+++ b/drivers/cpufreq/cpufreq_stats.c
@@ -294,11 +294,8 @@ static int cpufreq_stat_notifier_policy(struct notifier_block *nb,
 	if (val == CPUFREQ_CREATE_POLICY)
 		ret = cpufreq_stats_create_table(policy, table);
 	else if (val == CPUFREQ_REMOVE_POLICY) {
-		/* This might already be freed by cpu hotplug notifier */
-		if (per_cpu(cpufreq_stats_table, cpu)) {
-			cpufreq_stats_free_sysfs(cpu);
-			cpufreq_stats_free_table(cpu);
-		}
+		cpufreq_stats_free_sysfs(cpu);
+		cpufreq_stats_free_table(cpu);
 	}
 
 	return ret;
@@ -340,33 +337,6 @@ static int cpufreq_stat_notifier_trans(struct notifier_block *nb,
 	return 0;
 }
 
-static int cpufreq_stat_cpu_callback(struct notifier_block *nfb,
-					       unsigned long action,
-					       void *hcpu)
-{
-	unsigned int cpu = (unsigned long)hcpu;
-
-	/* Don't free/allocate stats during suspend/resume */
-	if (action & CPU_TASKS_FROZEN)
-		return 0;
-
-	switch (action) {
-	case CPU_DOWN_PREPARE:
-		cpufreq_stats_free_sysfs(cpu);
-		break;
-	case CPU_DEAD:
-		cpufreq_stats_free_table(cpu);
-		break;
-	}
-	return NOTIFY_OK;
-}
-
-/* priority=1 so this will get called before cpufreq_remove_dev */
-static struct notifier_block cpufreq_stat_cpu_notifier __refdata = {
-	.notifier_call = cpufreq_stat_cpu_callback,
-	.priority = 1,
-};
-
 static struct notifier_block notifier_policy_block = {
 	.notifier_call = cpufreq_stat_notifier_policy
 };
@@ -386,14 +356,11 @@ static int __init cpufreq_stats_init(void)
 	if (ret)
 		return ret;
 
-	register_hotcpu_notifier(&cpufreq_stat_cpu_notifier);
-
 	ret = cpufreq_register_notifier(&notifier_trans_block,
 				CPUFREQ_TRANSITION_NOTIFIER);
 	if (ret) {
 		cpufreq_unregister_notifier(&notifier_policy_block,
 				CPUFREQ_POLICY_NOTIFIER);
-		unregister_hotcpu_notifier(&cpufreq_stat_cpu_notifier);
 		for_each_online_cpu(cpu)
 			cpufreq_stats_free_table(cpu);
 		return ret;
@@ -409,7 +376,6 @@ static void __exit cpufreq_stats_exit(void)
 			CPUFREQ_POLICY_NOTIFIER);
 	cpufreq_unregister_notifier(&notifier_trans_block,
 			CPUFREQ_TRANSITION_NOTIFIER);
-	unregister_hotcpu_notifier(&cpufreq_stat_cpu_notifier);
 	for_each_online_cpu(cpu) {
 		cpufreq_stats_free_table(cpu);
 		cpufreq_stats_free_sysfs(cpu);
-- 
1.7.12.rc2.18.g61b472e

[PATCH 3/4] cpufreq: stats: free table and remove sysfs entry in a single routine

[Viresh Kumar]

We don't have code paths now, where we need to do these two things separately
and so better do them in a single routine. Just as they are allocated in a
single routine.

Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
---
 drivers/cpufreq/cpufreq_stats.c | 50 ++++++++++++++++-------------------------
 1 file changed, 19 insertions(+), 31 deletions(-)

diff --git a/drivers/cpufreq/cpufreq_stats.c b/drivers/cpufreq/cpufreq_stats.c
index 8fa5844..9dd5883 100644
--- a/drivers/cpufreq/cpufreq_stats.c
+++ b/drivers/cpufreq/cpufreq_stats.c
@@ -151,40 +151,32 @@ static int freq_table_get_index(struct cpufreq_stats *stat, unsigned int freq)
 	return -1;
 }
 
-/* should be called late in the CPU removal sequence so that the stats
- * memory is still available in case someone tries to use it.
- */
-static void cpufreq_stats_free_table(unsigned int cpu)
+static void __cpufreq_stats_free_table(struct cpufreq_policy *policy)
 {
-	struct cpufreq_stats *stat = per_cpu(cpufreq_stats_table, cpu);
+	struct cpufreq_stats *stat = per_cpu(cpufreq_stats_table, policy->cpu);
 
-	if (stat) {
-		pr_debug("%s: Free stat table\n", __func__);
-		kfree(stat->time_in_state);
-		kfree(stat);
-		per_cpu(cpufreq_stats_table, cpu) = NULL;
-	}
+	if (!stat)
+		return;
+
+	pr_debug("%s: Free stat table\n", __func__);
+
+	sysfs_remove_group(&policy->kobj, &stats_attr_group);
+	kfree(stat->time_in_state);
+	kfree(stat);
+	per_cpu(cpufreq_stats_table, policy->cpu) = NULL;
 }
 
-/* must be called early in the CPU removal sequence (before
- * cpufreq_remove_dev) so that policy is still valid.
- */
-static void cpufreq_stats_free_sysfs(unsigned int cpu)
+static void cpufreq_stats_free_table(unsigned int cpu)
 {
-	struct cpufreq_policy *policy = cpufreq_cpu_get(cpu);
+	struct cpufreq_policy *policy;
 
+	policy = cpufreq_cpu_get(cpu);
 	if (!policy)
 		return;
 
-	if (!cpufreq_frequency_get_table(cpu))
-		goto put_ref;
-
-	if (!policy_is_shared(policy)) {
-		pr_debug("%s: Free sysfs stat\n", __func__);
-		sysfs_remove_group(&policy->kobj, &stats_attr_group);
-	}
+	if (cpufreq_frequency_get_table(policy->cpu))
+		__cpufreq_stats_free_table(policy);
 
-put_ref:
 	cpufreq_cpu_put(policy);
 }
 
@@ -293,10 +285,8 @@ static int cpufreq_stat_notifier_policy(struct notifier_block *nb,
 
 	if (val == CPUFREQ_CREATE_POLICY)
 		ret = cpufreq_stats_create_table(policy, table);
-	else if (val == CPUFREQ_REMOVE_POLICY) {
-		cpufreq_stats_free_sysfs(cpu);
-		cpufreq_stats_free_table(cpu);
-	}
+	else if (val == CPUFREQ_REMOVE_POLICY)
+		__cpufreq_stats_free_table(policy);
 
 	return ret;
 }
@@ -376,10 +366,8 @@ static void __exit cpufreq_stats_exit(void)
 			CPUFREQ_POLICY_NOTIFIER);
 	cpufreq_unregister_notifier(&notifier_trans_block,
 			CPUFREQ_TRANSITION_NOTIFIER);
-	for_each_online_cpu(cpu) {
+	for_each_online_cpu(cpu)
 		cpufreq_stats_free_table(cpu);
-		cpufreq_stats_free_sysfs(cpu);
-	}
 }
 
 MODULE_AUTHOR("Zou Nan hai <nanhai.zou@intel.com>");
-- 
1.7.12.rc2.18.g61b472e

[PATCH 4/4] cpufreq: stats: create sysfs entries when cpufreq_stats is a module

[Viresh Kumar]

When cpufreq_stats is compiled in as a module, cpufreq driver would have already
been registered. And so the CPUFREQ_CREATE_POLICY notifiers wouldn't be called
for it. Hence no sysfs entries for stats. :(

This patch calls cpufreq_stats_create_table() for each online cpu from
cpufreq_stats_init() and so if policy is already created for CPUx then we will
register sysfs stats for it.

When its not compiled as module, we will return early as policy wouldn't be
found for any of the CPUs.

Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
---
 drivers/cpufreq/cpufreq_stats.c | 27 +++++++++++++++++++++++++--
 1 file changed, 25 insertions(+), 2 deletions(-)

diff --git a/drivers/cpufreq/cpufreq_stats.c b/drivers/cpufreq/cpufreq_stats.c
index 9dd5883..5793e14 100644
--- a/drivers/cpufreq/cpufreq_stats.c
+++ b/drivers/cpufreq/cpufreq_stats.c
@@ -180,7 +180,7 @@ static void cpufreq_stats_free_table(unsigned int cpu)
 	cpufreq_cpu_put(policy);
 }
 
-static int cpufreq_stats_create_table(struct cpufreq_policy *policy,
+static int __cpufreq_stats_create_table(struct cpufreq_policy *policy,
 		struct cpufreq_frequency_table *table)
 {
 	unsigned int i, j, count = 0, ret = 0;
@@ -253,6 +253,26 @@ error_get_fail:
 	return ret;
 }
 
+static void cpufreq_stats_create_table(unsigned int cpu)
+{
+	struct cpufreq_policy *policy;
+	struct cpufreq_frequency_table *table;
+
+	/*
+	 * "likely(!policy)" because normally cpufreq_stats will be registered
+	 * before cpufreq driver
+	 */
+	policy = cpufreq_cpu_get(cpu);
+	if (likely(!policy))
+		return;
+
+	table = cpufreq_frequency_get_table(policy->cpu);
+	if (likely(table))
+		__cpufreq_stats_create_table(policy, table);
+
+	cpufreq_cpu_put(policy);
+}
+
 static void cpufreq_stats_update_policy_cpu(struct cpufreq_policy *policy)
 {
 	struct cpufreq_stats *stat = per_cpu(cpufreq_stats_table,
@@ -284,7 +304,7 @@ static int cpufreq_stat_notifier_policy(struct notifier_block *nb,
 		return 0;
 
 	if (val == CPUFREQ_CREATE_POLICY)
-		ret = cpufreq_stats_create_table(policy, table);
+		ret = __cpufreq_stats_create_table(policy, table);
 	else if (val == CPUFREQ_REMOVE_POLICY)
 		__cpufreq_stats_free_table(policy);
 
@@ -346,6 +366,9 @@ static int __init cpufreq_stats_init(void)
 	if (ret)
 		return ret;
 
+	for_each_online_cpu(cpu)
+		cpufreq_stats_create_table(cpu);
+
 	ret = cpufreq_register_notifier(&notifier_trans_block,
 				CPUFREQ_TRANSITION_NOTIFIER);
 	if (ret) {
-- 
1.7.12.rc2.18.g61b472e

Re: [PATCH 0/4] CPUFreq stats: Bug fixes

[Nicolas Pitre]

On Thu, 26 Sep 2013, Viresh Kumar wrote:

> There are several problems/bugs in cpufreq-stats specially with cpufreq drivers
> as modules and suspend/resume path. These are mentioned well in changelogs.
> 
> These are tested over my thinkpad (acpi-cpufreq) in following way:
> [1] offline+online all CPUs except boot cpu in a while loop
> [2] then do suspend resume
> [3] repeat [1] and [2] several times.
> 
> No issues found..
> 
> Also tested on my exynos board:
> - Added cpufreq_unregister/register while loop in exynos-cpufreq.c so that we
>   continuously register/unregister driver... Stats were working fine now..
> - Compile cpufreq-stats as module and insert/remove it several times after
>   removing above hack (as that doesn't let linux boot :) )..
> 
> @Nico: Can you remove the hacky code from IKS tree and test these instead to see
> if we still have any issues?

My testing of those patches took a long and winding path but finally I 
was able to do it today.

So for the 4 patches:

Tested-by: Nicolas Pitre <nico@linaro.org>
Acked-by: Nicolas Pitre <nico@linaro.org>


Nicolas

Re: [PATCH 0/4] CPUFreq stats: Bug fixes

[Viresh Kumar]

On 23 October 2013 02:36, Nicolas Pitre <nicolas.pitre@linaro.org> wrote:
> My testing of those patches took a long and winding path but finally I
> was able to do it today.
>
> So for the 4 patches:
>
> Tested-by: Nicolas Pitre <nico@linaro.org>
> Acked-by: Nicolas Pitre <nico@linaro.org>

Ahh.. Thanks Nico.. that would be pretty helpful :)

Re: [PATCH 0/4] CPUFreq stats: Bug fixes

[Rafael J. Wysocki]

On Thursday, September 26, 2013 03:59:50 PM Viresh Kumar wrote:
> Hi Rafael,
> 
> I know you asked me not to send any more patches before the earlier ones get
> into kernel. I got to this as Nicolas Pitre required to send few CPUFreq patches
> for ARM's big LITTLE In-Kernel-Switcher. And within linaro we have hacked these
> bugs in a bad way..
> 
> Because of his dependency I am forced to send these.. These aren't introduced
> recently and so they can be included in 3.13.
> 
> There are several problems/bugs in cpufreq-stats specially with cpufreq drivers
> as modules and suspend/resume path. These are mentioned well in changelogs.
> 
> These are tested over my thinkpad (acpi-cpufreq) in following way:
> [1] offline+online all CPUs except boot cpu in a while loop
> [2] then do suspend resume
> [3] repeat [1] and [2] several times.
> 
> No issues found..
> 
> Also tested on my exynos board:
> - Added cpufreq_unregister/register while loop in exynos-cpufreq.c so that we
>   continuously register/unregister driver... Stats were working fine now..
> - Compile cpufreq-stats as module and insert/remove it several times after
>   removing above hack (as that doesn't let linux boot :) )..
> 
> @Srivatsa: You also have fairly good idea of cpufreq now, so please give some
> time to review this :)
> 
> @Nico: Can you remove the hacky code from IKS tree and test these instead to see
> if we still have any issues?

Can you please rebase this series on top of linux-pm.git/bleeding-edge and
resend it?  Chances are it might make it into 3.14.

Thanks,
Rafael