[PATCH] Reorganize the cpufreq cpu hotplug locking to not be totally bizare

The patch below moves the cpu hotplugging higher up in the cpufreq
layering; this is needed to avoid recursive taking of the cpu hotplug
lock and to otherwise detangle the mess.

The new rules are:
1. you must do lock_cpu_hotplug() around the following functions:
__cpufreq_driver_target
__cpufreq_governor (for CPUFREQ_GOV_LIMITS operation only)
__cpufreq_set_policy
2. governer methods (.governer) must NOT take the lock_cpu_hotplug()
lock in any way; they are called with the lock taken already
3. if your governer spawns a thread that does things, like calling
__cpufreq_driver_target, your thread must honor rule #1.
4. the policy lock and other cpufreq internal locks nest within
the lock_cpu_hotplug() lock.

I'm not entirely happy about how the __cpufreq_governor rule ended up
(conditional locking rule depending on the argument) but basically all
callers pass this as a constant so it's not too horrible.

The patch also removes the cpufreq_governor() function since during the
locking audit it turned out to be entirely unused (so no need to fix it)

The patch works on my testbox, but it could use more testing
(otoh... it can't be much worse than the current code)

Signed-off-by: Arjan van de Ven <[e-mail]>
Signed-off-by: Linus Torvalds <[e-mail]>

drivers/cpufreq/cpufreq.c

···
 	if (ret != 1)							\
 		return -EINVAL;						\
 									\
+	lock_cpu_hotplug();						\
 	mutex_lock(&policy->lock);					\
 	ret = __cpufreq_set_policy(policy, &new_policy);		\
 	policy->user_policy.object = policy->object;			\
 	mutex_unlock(&policy->lock);					\
+	unlock_cpu_hotplug();						\
 									\
 	return ret ? ret : count;					\
 }
···
  *********************************************************************/
 
 
+/* Must be called with lock_cpu_hotplug held */
 int __cpufreq_driver_target(struct cpufreq_policy *policy,
 			    unsigned int target_freq,
 			    unsigned int relation)
 {
 	int retval = -EINVAL;
 
-	lock_cpu_hotplug();
 	dprintk("target for CPU %u: %u kHz, relation %u\n", policy->cpu,
 		target_freq, relation);
 	if (cpu_online(policy->cpu) && cpufreq_driver->target)
 		retval = cpufreq_driver->target(policy, target_freq, relation);
-
-	unlock_cpu_hotplug();
 
 	return retval;
 }
···
 	if (!policy)
 		return -EINVAL;
 
+	lock_cpu_hotplug();
 	mutex_lock(&policy->lock);
 
 	ret = __cpufreq_driver_target(policy, target_freq, relation);
 
 	mutex_unlock(&policy->lock);
+	unlock_cpu_hotplug();
 
 	cpufreq_cpu_put(policy);
 	return ret;
 }
 EXPORT_SYMBOL_GPL(cpufreq_driver_target);
 
+/*
+ * Locking: Must be called with the lock_cpu_hotplug() lock held
+ * when "event" is CPUFREQ_GOV_LIMITS
+ */
 
 static int __cpufreq_governor(struct cpufreq_policy *policy, unsigned int event)
 {
···
 
 	return ret;
 }
-
-
-int cpufreq_governor(unsigned int cpu, unsigned int event)
-{
-	int ret = 0;
-	struct cpufreq_policy *policy = cpufreq_cpu_get(cpu);
-
-	if (!policy)
-		return -EINVAL;
-
-	mutex_lock(&policy->lock);
-	ret = __cpufreq_governor(policy, event);
-	mutex_unlock(&policy->lock);
-
-	cpufreq_cpu_put(policy);
-	return ret;
-}
-EXPORT_SYMBOL_GPL(cpufreq_governor);
 
 
 int cpufreq_register_governor(struct cpufreq_governor *governor)
···
 EXPORT_SYMBOL(cpufreq_get_policy);
 
 
+/*
+ * Locking: Must be called with the lock_cpu_hotplug() lock held
+ */
 static int __cpufreq_set_policy(struct cpufreq_policy *data, struct cpufreq_policy *policy)
 {
 	int ret = 0;
···
 	if (!data)
 		return -EINVAL;
 
+	lock_cpu_hotplug();
+
 	/* lock this CPU */
 	mutex_lock(&data->lock);
 
···
 	data->user_policy.governor = data->governor;
 
 	mutex_unlock(&data->lock);
+
+	unlock_cpu_hotplug();
 	cpufreq_cpu_put(data);
 
 	return ret;
···
 	if (!data)
 		return -ENODEV;
 
+	lock_cpu_hotplug();
 	mutex_lock(&data->lock);
 
 	dprintk("updating policy for CPU %u\n", cpu);
···
 	ret = __cpufreq_set_policy(data, &policy);
 
 	mutex_unlock(&data->lock);
-
+	unlock_cpu_hotplug();
 	cpufreq_cpu_put(data);
 	return ret;
 }

drivers/cpufreq/cpufreq_conservative.c

···
 		break;
 
 	case CPUFREQ_GOV_LIMITS:
-		lock_cpu_hotplug();
 		mutex_lock(&dbs_mutex);
 		if (policy->max < this_dbs_info->cur_policy->cur)
 			__cpufreq_driver_target(
···
 					this_dbs_info->cur_policy,
 				       	policy->min, CPUFREQ_RELATION_L);
 		mutex_unlock(&dbs_mutex);
-		unlock_cpu_hotplug();
 		break;
 	}
 	return 0;

drivers/cpufreq/cpufreq_ondemand.c

···
 	if (!dbs_info->enable)
 		return;
 
+	lock_cpu_hotplug();
 	dbs_check_cpu(dbs_info);
+	unlock_cpu_hotplug();
 	queue_delayed_work_on(cpu, kondemand_wq, &dbs_info->work,
 			usecs_to_jiffies(dbs_tuners_ins.sampling_rate));
 }
···
 		break;
 
 	case CPUFREQ_GOV_LIMITS:
-		lock_cpu_hotplug();
 		mutex_lock(&dbs_mutex);
 		if (policy->max < this_dbs_info->cur_policy->cur)
 			__cpufreq_driver_target(this_dbs_info->cur_policy,
···
 			                        policy->min,
 			                        CPUFREQ_RELATION_L);
 		mutex_unlock(&dbs_mutex);
-		unlock_cpu_hotplug();
 		break;
 	}
 	return 0;

drivers/cpufreq/cpufreq_userspace.c

···
 #include <linux/spinlock.h>
 #include <linux/interrupt.h>
 #include <linux/cpufreq.h>
+#include <linux/cpu.h>
 #include <linux/types.h>
 #include <linux/fs.h>
 #include <linux/sysfs.h>
···
 
 	dprintk("cpufreq_set for cpu %u, freq %u kHz\n", policy->cpu, freq);
 
+	lock_cpu_hotplug();
 	mutex_lock(&userspace_mutex);
 	if (!cpu_is_managed[policy->cpu])
 		goto err;
···
 
  err:
 	mutex_unlock(&userspace_mutex);
+	unlock_cpu_hotplug();
 	return ret;
 }
 

include/linux/cpufreq.h

···
 				   unsigned int relation);
 
 
-/* pass an event to the cpufreq governor */
-int cpufreq_governor(unsigned int cpu, unsigned int event);
-
 int cpufreq_register_governor(struct cpufreq_governor *governor);
 void cpufreq_unregister_governor(struct cpufreq_governor *governor);