mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
* [PATCH 0/2] cpufreq: Avoid checking governor return values that are always 0
@ 2016-05-12 13:12 Rafael J. Wysocki
  2016-05-12 13:13 ` [PATCH 1/2] cpufreq: governor: CPUFREQ_GOV_POLICY_EXIT never fails Rafael J. Wysocki
  2016-05-12 13:14 ` [PATCH 2/2] cpufreq: governor: CPUFREQ_GOV_STOP " Rafael J. Wysocki
  0 siblings, 2 replies; 5+ messages in thread
From: Rafael J. Wysocki @ 2016-05-12 13:12 UTC (permalink / raw)
  To: Linux PM list
  Cc: Srinivas Pandruvada, Viresh Kumar, Linux Kernel Mailing List

Hi,

As it turns out, all of the existing governors (and the upcoming schedutil one)
always return 0 from ->governor() for the event argument equal to _GOV_STOP
or _POLICY_EXIT.

The following two patches rearrange the core cpufreq code to take that into
account.

Thanks,
Rafael

^ permalink raw reply	[flat|nested] 5+ messages in thread

* [PATCH 1/2] cpufreq: governor: CPUFREQ_GOV_POLICY_EXIT never fails
  2016-05-12 13:12 [PATCH 0/2] cpufreq: Avoid checking governor return values that are always 0 Rafael J. Wysocki
@ 2016-05-12 13:13 ` Rafael J. Wysocki
  2016-05-13  3:13   ` Viresh Kumar
  2016-05-12 13:14 ` [PATCH 2/2] cpufreq: governor: CPUFREQ_GOV_STOP " Rafael J. Wysocki
  1 sibling, 1 reply; 5+ messages in thread
From: Rafael J. Wysocki @ 2016-05-12 13:13 UTC (permalink / raw)
  To: Linux PM list
  Cc: Srinivas Pandruvada, Viresh Kumar, Linux Kernel Mailing List

From: Rafael J. Wysocki <rafael.j.wysocki@intel.com>

None of the cpufreq governors currently in the tree will ever fail
an invocation of the ->governor() callback with the event argument
equal to CPUFREQ_GOV_POLICY_EXIT (unless invoked with incorrect
arguments which doesn't matter anyway) and it wouldn't really
make sense to fail it, because the caller won't be able to handle
that failure in a meaningful way.

Accordingly, rearrange the code in the core to make it clear that
this call never fails.

Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
---
 drivers/cpufreq/cpufreq.c |   35 ++++++++++++-----------------------
 1 file changed, 12 insertions(+), 23 deletions(-)

Index: linux-pm/drivers/cpufreq/cpufreq.c
===================================================================
--- linux-pm.orig/drivers/cpufreq/cpufreq.c
+++ linux-pm/drivers/cpufreq/cpufreq.c
@@ -78,9 +78,9 @@ static int cpufreq_governor(struct cpufr
 static unsigned int __cpufreq_get(struct cpufreq_policy *policy);
 static int cpufreq_start_governor(struct cpufreq_policy *policy);
 
-static inline int cpufreq_exit_governor(struct cpufreq_policy *policy)
+static inline void cpufreq_exit_governor(struct cpufreq_policy *policy)
 {
-	return cpufreq_governor(policy, CPUFREQ_GOV_POLICY_EXIT);
+	(void)cpufreq_governor(policy, CPUFREQ_GOV_POLICY_EXIT);
 }
 
 /**
@@ -1387,12 +1387,8 @@ static void cpufreq_offline(unsigned int
 	if (cpufreq_driver->stop_cpu)
 		cpufreq_driver->stop_cpu(policy);
 
-	/* If cpu is last user of policy, free policy */
-	if (has_target()) {
-		ret = cpufreq_exit_governor(policy);
-		if (ret)
-			pr_err("%s: Failed to exit governor\n", __func__);
-	}
+	if (has_target())
+		cpufreq_exit_governor(policy);
 
 	/*
 	 * Perform the ->exit() even during light-weight tear-down,
@@ -2049,16 +2045,15 @@ static int cpufreq_governor(struct cpufr
 
 	ret = policy->governor->governor(policy, event);
 
-	if (!ret) {
-		if (event == CPUFREQ_GOV_POLICY_INIT)
+	if (event == CPUFREQ_GOV_POLICY_INIT) {
+		if (ret)
+			module_put(policy->governor->owner);
+		else
 			policy->governor->initialized++;
-		else if (event == CPUFREQ_GOV_POLICY_EXIT)
-			policy->governor->initialized--;
-	}
-
-	if (((event == CPUFREQ_GOV_POLICY_INIT) && ret) ||
-			((event == CPUFREQ_GOV_POLICY_EXIT) && !ret))
+	} else if (event == CPUFREQ_GOV_POLICY_EXIT) {
+		policy->governor->initialized--;
 		module_put(policy->governor->owner);
+	}
 
 	return ret;
 }
@@ -2228,13 +2223,7 @@ static int cpufreq_set_policy(struct cpu
 				 __func__, old_gov->name, ret);
 			return ret;
 		}
-
-		ret = cpufreq_exit_governor(policy);
-		if (ret) {
-			pr_err("%s: Failed to Exit Governor: %s (%d)\n",
-			       __func__, old_gov->name, ret);
-			return ret;
-		}
+		cpufreq_exit_governor(policy);
 	}
 
 	/* start new governor */

^ permalink raw reply	[flat|nested] 5+ messages in thread

* [PATCH 2/2] cpufreq: governor: CPUFREQ_GOV_STOP never fails
  2016-05-12 13:12 [PATCH 0/2] cpufreq: Avoid checking governor return values that are always 0 Rafael J. Wysocki
  2016-05-12 13:13 ` [PATCH 1/2] cpufreq: governor: CPUFREQ_GOV_POLICY_EXIT never fails Rafael J. Wysocki
@ 2016-05-12 13:14 ` Rafael J. Wysocki
  2016-05-13  3:14   ` Viresh Kumar
  1 sibling, 1 reply; 5+ messages in thread
From: Rafael J. Wysocki @ 2016-05-12 13:14 UTC (permalink / raw)
  To: Linux PM list
  Cc: Srinivas Pandruvada, Viresh Kumar, Linux Kernel Mailing List

From: Rafael J. Wysocki <rafael.j.wysocki@intel.com>

None of the cpufreq governors currently in the tree will ever fail
an invocation of the ->governor() callback with the event argument
equal to CPUFREQ_GOV_STOP (unless invoked with incorrect arguments
which doesn't matter anyway) and it is rather difficult to imagine
a valid reason for such a failure.

Accordingly, rearrange the code in the core to make it clear that
this call never fails.

Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
---
 drivers/cpufreq/cpufreq.c |   40 +++++++++++-----------------------------
 1 file changed, 11 insertions(+), 29 deletions(-)

Index: linux-pm/drivers/cpufreq/cpufreq.c
===================================================================
--- linux-pm.orig/drivers/cpufreq/cpufreq.c
+++ linux-pm/drivers/cpufreq/cpufreq.c
@@ -83,6 +83,11 @@ static inline void cpufreq_exit_governor
 	(void)cpufreq_governor(policy, CPUFREQ_GOV_POLICY_EXIT);
 }
 
+static inline void cpufreq_stop_governor(struct cpufreq_policy *policy)
+{
+	(void)cpufreq_governor(policy, CPUFREQ_GOV_STOP);
+}
+
 /**
  * Two notifier lists: the "policy" list is involved in the
  * validation process for a new CPU frequency policy; the
@@ -1026,13 +1031,8 @@ static int cpufreq_add_policy_cpu(struct
 		return 0;
 
 	down_write(&policy->rwsem);
-	if (has_target()) {
-		ret = cpufreq_governor(policy, CPUFREQ_GOV_STOP);
-		if (ret) {
-			pr_err("%s: Failed to stop governor\n", __func__);
-			goto unlock;
-		}
-	}
+	if (has_target())
+		cpufreq_stop_governor(policy);
 
 	cpumask_set_cpu(cpu, policy->cpus);
 
@@ -1041,8 +1041,6 @@ static int cpufreq_add_policy_cpu(struct
 		if (ret)
 			pr_err("%s: Failed to start governor\n", __func__);
 	}
-
-unlock:
 	up_write(&policy->rwsem);
 	return ret;
 }
@@ -1354,11 +1352,8 @@ static void cpufreq_offline(unsigned int
 	}
 
 	down_write(&policy->rwsem);
-	if (has_target()) {
-		ret = cpufreq_governor(policy, CPUFREQ_GOV_STOP);
-		if (ret)
-			pr_err("%s: Failed to stop governor\n", __func__);
-	}
+	if (has_target())
+		cpufreq_stop_governor(policy);
 
 	cpumask_clear_cpu(cpu, policy->cpus);
 
@@ -1622,7 +1617,6 @@ EXPORT_SYMBOL(cpufreq_generic_suspend);
 void cpufreq_suspend(void)
 {
 	struct cpufreq_policy *policy;
-	int ret;
 
 	if (!cpufreq_driver)
 		return;
@@ -1635,14 +1629,8 @@ void cpufreq_suspend(void)
 	for_each_active_policy(policy) {
 		if (has_target()) {
 			down_write(&policy->rwsem);
-			ret = cpufreq_governor(policy, CPUFREQ_GOV_STOP);
+			cpufreq_stop_governor(policy);
 			up_write(&policy->rwsem);
-
-			if (ret) {
-				pr_err("%s: Failed to stop governor for policy: %p\n",
-					__func__, policy);
-				continue;
-			}
 		}
 
 		if (cpufreq_driver->suspend && cpufreq_driver->suspend(policy))
@@ -2216,13 +2204,7 @@ static int cpufreq_set_policy(struct cpu
 	old_gov = policy->governor;
 	/* end old governor */
 	if (old_gov) {
-		ret = cpufreq_governor(policy, CPUFREQ_GOV_STOP);
-		if (ret) {
-			/* This can happen due to race with other operations */
-			pr_debug("%s: Failed to Stop Governor: %s (%d)\n",
-				 __func__, old_gov->name, ret);
-			return ret;
-		}
+		cpufreq_stop_governor(policy);
 		cpufreq_exit_governor(policy);
 	}
 

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [PATCH 1/2] cpufreq: governor: CPUFREQ_GOV_POLICY_EXIT never fails
  2016-05-12 13:13 ` [PATCH 1/2] cpufreq: governor: CPUFREQ_GOV_POLICY_EXIT never fails Rafael J. Wysocki
@ 2016-05-13  3:13   ` Viresh Kumar
  0 siblings, 0 replies; 5+ messages in thread
From: Viresh Kumar @ 2016-05-13  3:13 UTC (permalink / raw)
  To: Rafael J. Wysocki
  Cc: Linux PM list, Srinivas Pandruvada, Linux Kernel Mailing List

On 12-05-16, 15:13, Rafael J. Wysocki wrote:
> From: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
> 
> None of the cpufreq governors currently in the tree will ever fail
> an invocation of the ->governor() callback with the event argument
> equal to CPUFREQ_GOV_POLICY_EXIT (unless invoked with incorrect
> arguments which doesn't matter anyway) and it wouldn't really
> make sense to fail it, because the caller won't be able to handle
> that failure in a meaningful way.
> 
> Accordingly, rearrange the code in the core to make it clear that
> this call never fails.
> 
> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
> ---
>  drivers/cpufreq/cpufreq.c |   35 ++++++++++++-----------------------
>  1 file changed, 12 insertions(+), 23 deletions(-)

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

-- 
viresh

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [PATCH 2/2] cpufreq: governor: CPUFREQ_GOV_STOP never fails
  2016-05-12 13:14 ` [PATCH 2/2] cpufreq: governor: CPUFREQ_GOV_STOP " Rafael J. Wysocki
@ 2016-05-13  3:14   ` Viresh Kumar
  0 siblings, 0 replies; 5+ messages in thread
From: Viresh Kumar @ 2016-05-13  3:14 UTC (permalink / raw)
  To: Rafael J. Wysocki
  Cc: Linux PM list, Srinivas Pandruvada, Linux Kernel Mailing List

On 12-05-16, 15:14, Rafael J. Wysocki wrote:
> From: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
> 
> None of the cpufreq governors currently in the tree will ever fail
> an invocation of the ->governor() callback with the event argument
> equal to CPUFREQ_GOV_STOP (unless invoked with incorrect arguments
> which doesn't matter anyway) and it is rather difficult to imagine
> a valid reason for such a failure.
> 
> Accordingly, rearrange the code in the core to make it clear that
> this call never fails.
> 
> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
> ---
>  drivers/cpufreq/cpufreq.c |   40 +++++++++++-----------------------------
>  1 file changed, 11 insertions(+), 29 deletions(-)

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

-- 
viresh

^ permalink raw reply	[flat|nested] 5+ messages in thread

end of thread, other threads:[~2016-05-13  3:14 UTC | newest]

Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2016-05-12 13:12 [PATCH 0/2] cpufreq: Avoid checking governor return values that are always 0 Rafael J. Wysocki
2016-05-12 13:13 ` [PATCH 1/2] cpufreq: governor: CPUFREQ_GOV_POLICY_EXIT never fails Rafael J. Wysocki
2016-05-13  3:13   ` Viresh Kumar
2016-05-12 13:14 ` [PATCH 2/2] cpufreq: governor: CPUFREQ_GOV_STOP " Rafael J. Wysocki
2016-05-13  3:14   ` Viresh Kumar

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

all inboxes | Powered by JetHome®