Merge tag 'pm-5.15-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm

Pull power management updates from Rafael Wysocki:
 "These address some PCI device power management issues, add new
  hardware support to the RAPL power capping driver, add HWP guaranteed
  performance change notification support to the intel_pstate driver,
  replace deprecated CPU-hotplug functions in a few places, update CPU
  PM notifiers to use raw spinlocks, update the PM domains framework
  (new DT property support, Kconfig fix), do a couple of cleanups in
  code related to system sleep, and improve the energy model and the
  schedutil cpufreq governor.

  Specifics:

   - Address 3 PCI device power management issues (Rafael Wysocki).

   - Add Power Limit4 support for Alder Lake to the Intel RAPL power
     capping driver (Sumeet Pawnikar).

   - Add HWP guaranteed performance change notification support to the
     intel_pstate driver (Srinivas Pandruvada).

   - Replace deprecated CPU-hotplug functions in code related to power
     management (Sebastian Andrzej Siewior).

   - Update CPU PM notifiers to use raw spinlocks (Valentin Schneider).

   - Add support for 'required-opps' DT property to the generic power
     domains (genpd) framework and use this property for I2C on ARM64
     sc7180 (Rajendra Nayak).

   - Fix Kconfig issue related to genpd (Geert Uytterhoeven).

   - Increase energy calculation precision in the Energy Model (Lukasz
     Luba).

   - Fix kobject deletion in the exit code of the schedutil cpufreq
     governor (Kevin Hao).

   - Unmark some functions as kernel-doc in the PM core to avoid
     false-positive documentation build warnings (Randy Dunlap).

   - Check RTC features instead of ops in suspend_test Alexandre
     Belloni)"

* tag 'pm-5.15-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm:
  PM: domains: Fix domain attach for CONFIG_PM_OPP=n
  powercap: Add Power Limit4 support for Alder Lake SoC
  cpufreq: intel_pstate: Process HWP Guaranteed change notification
  thermal: intel: Allow processing of HWP interrupt
  notifier: Remove atomic_notifier_call_chain_robust()
  PM: cpu: Make notifier chain use a raw_spinlock_t
  PM: sleep: unmark 'state' functions as kernel-doc
  arm64: dts: sc7180: Add required-opps for i2c
  PM: domains: Add support for 'required-opps' to set default perf state
  opp: Don't print an error if required-opps is missing
  cpufreq: schedutil: Use kobject release() method to free sugov_tunables
  PM: EM: Increase energy calculation precision
  PM: sleep: check RTC features instead of ops in suspend_test
  PM: sleep: s2idle: Replace deprecated CPU-hotplug functions
  cpufreq: Replace deprecated CPU-hotplug functions
  powercap: intel_rapl: Replace deprecated CPU-hotplug functions
  PCI: PM: Enable PME if it can be signaled from D3cold
  PCI: PM: Avoid forcing PCI_D0 for wakeup reasons inconsistently
  PCI: Use pci_update_current_state() in pci_enable_device_flags()

kernel/cpu_pm.c

@@ -13,19 +13,32 @@
 #include <linux/spinlock.h>
 #include <linux/syscore_ops.h>
 
-static ATOMIC_NOTIFIER_HEAD(cpu_pm_notifier_chain);
+/*
+ * atomic_notifiers use a spinlock_t, which can block under PREEMPT_RT.
+ * Notifications for cpu_pm will be issued by the idle task itself, which can
+ * never block, IOW it requires using a raw_spinlock_t.
+ */
+static struct {
+	struct raw_notifier_head chain;
+	raw_spinlock_t lock;
+} cpu_pm_notifier = {
+	.chain = RAW_NOTIFIER_INIT(cpu_pm_notifier.chain),
+	.lock  = __RAW_SPIN_LOCK_UNLOCKED(cpu_pm_notifier.lock),
+};
 
 static int cpu_pm_notify(enum cpu_pm_event event)
 {
 	int ret;
 
 	/*
-	 * atomic_notifier_call_chain has a RCU read critical section, which
-	 * could be disfunctional in cpu idle. Copy RCU_NONIDLE code to let
-	 * RCU know this.
+	 * This introduces a RCU read critical section, which could be
+	 * disfunctional in cpu idle. Copy RCU_NONIDLE code to let RCU know
+	 * this.
 	 */
 	rcu_irq_enter_irqson();
-	ret = atomic_notifier_call_chain(&cpu_pm_notifier_chain, event, NULL);
+	rcu_read_lock();
+	ret = raw_notifier_call_chain(&cpu_pm_notifier.chain, event, NULL);
+	rcu_read_unlock();
 	rcu_irq_exit_irqson();
 
 	return notifier_to_errno(ret);
@@ -33,10 +46,13 @@ static int cpu_pm_notify(enum cpu_pm_event event)
 
 static int cpu_pm_notify_robust(enum cpu_pm_event event_up, enum cpu_pm_event event_down)
 {
+	unsigned long flags;
 	int ret;
 
 	rcu_irq_enter_irqson();
-	ret = atomic_notifier_call_chain_robust(&cpu_pm_notifier_chain, event_up, event_down, NULL);
+	raw_spin_lock_irqsave(&cpu_pm_notifier.lock, flags);
+	ret = raw_notifier_call_chain_robust(&cpu_pm_notifier.chain, event_up, event_down, NULL);
+	raw_spin_unlock_irqrestore(&cpu_pm_notifier.lock, flags);
 	rcu_irq_exit_irqson();
 
 	return notifier_to_errno(ret);
@@ -49,12 +65,17 @@ static int cpu_pm_notify_robust(enum cpu_pm_event event_up, enum cpu_pm_event ev
  * Add a driver to a list of drivers that are notified about
  * CPU and CPU cluster low power entry and exit.
  *
- * This function may sleep, and has the same return conditions as
- * raw_notifier_chain_register.
+ * This function has the same return conditions as raw_notifier_chain_register.
  */
 int cpu_pm_register_notifier(struct notifier_block *nb)
 {
-	return atomic_notifier_chain_register(&cpu_pm_notifier_chain, nb);
+	unsigned long flags;
+	int ret;
+
+	raw_spin_lock_irqsave(&cpu_pm_notifier.lock, flags);
+	ret = raw_notifier_chain_register(&cpu_pm_notifier.chain, nb);
+	raw_spin_unlock_irqrestore(&cpu_pm_notifier.lock, flags);
+	return ret;
 }
 EXPORT_SYMBOL_GPL(cpu_pm_register_notifier);
 
@@ -64,12 +85,17 @@ EXPORT_SYMBOL_GPL(cpu_pm_register_notifier);
  *
  * Remove a driver from the CPU PM notifier list.
  *
- * This function may sleep, and has the same return conditions as
- * raw_notifier_chain_unregister.
+ * This function has the same return conditions as raw_notifier_chain_unregister.
  */
 int cpu_pm_unregister_notifier(struct notifier_block *nb)
 {
-	return atomic_notifier_chain_unregister(&cpu_pm_notifier_chain, nb);
+	unsigned long flags;
+	int ret;
+
+	raw_spin_lock_irqsave(&cpu_pm_notifier.lock, flags);
+	ret = raw_notifier_chain_unregister(&cpu_pm_notifier.chain, nb);
+	raw_spin_unlock_irqrestore(&cpu_pm_notifier.lock, flags);
+	return ret;
 }
 EXPORT_SYMBOL_GPL(cpu_pm_unregister_notifier);
 

kernel/notifier.c

@@ -172,25 +172,6 @@ int atomic_notifier_chain_unregister(struct atomic_notifier_head *nh,
 }
 EXPORT_SYMBOL_GPL(atomic_notifier_chain_unregister);
 
-int atomic_notifier_call_chain_robust(struct atomic_notifier_head *nh,
-		unsigned long val_up, unsigned long val_down, void *v)
-{
-	unsigned long flags;
-	int ret;
-
-	/*
-	 * Musn't use RCU; because then the notifier list can
-	 * change between the up and down traversal.
-	 */
-	spin_lock_irqsave(&nh->lock, flags);
-	ret = notifier_call_chain_robust(&nh->head, val_up, val_down, v);
-	spin_unlock_irqrestore(&nh->lock, flags);
-
-	return ret;
-}
-EXPORT_SYMBOL_GPL(atomic_notifier_call_chain_robust);
-NOKPROBE_SYMBOL(atomic_notifier_call_chain_robust);
-
 /**
  *	atomic_notifier_call_chain - Call functions in an atomic notifier chain
  *	@nh: Pointer to head of the atomic notifier chain

kernel/power/energy_model.c

@@ -170,7 +170,9 @@ static int em_create_perf_table(struct device *dev, struct em_perf_domain *pd,
 	/* Compute the cost of each performance state. */
 	fmax = (u64) table[nr_states - 1].frequency;
 	for (i = 0; i < nr_states; i++) {
-		table[i].cost = div64_u64(fmax * table[i].power,
+		unsigned long power_res = em_scale_power(table[i].power);
+
+		table[i].cost = div64_u64(fmax * power_res,
 					  table[i].frequency);
 	}
 

kernel/power/main.c

@@ -577,7 +577,7 @@ static inline void pm_print_times_init(void) {}
 
 struct kobject *power_kobj;
 
-/**
+/*
  * state - control system sleep states.
  *
  * show() returns available sleep state labels, which may be "mem", "standby",

kernel/power/suspend.c

@@ -96,7 +96,7 @@ static void s2idle_enter(void)
 	s2idle_state = S2IDLE_STATE_ENTER;
 	raw_spin_unlock_irq(&s2idle_lock);
 
-	get_online_cpus();
+	cpus_read_lock();
 	cpuidle_resume();
 
 	/* Push all the CPUs into the idle loop. */
@@ -106,7 +106,7 @@ static void s2idle_enter(void)
 		    s2idle_state == S2IDLE_STATE_WAKE);
 
 	cpuidle_pause();
-	put_online_cpus();
+	cpus_read_unlock();
 
 	raw_spin_lock_irq(&s2idle_lock);
 

kernel/power/suspend_test.c

@@ -129,7 +129,7 @@ static int __init has_wakealarm(struct device *dev, const void *data)
 {
 	struct rtc_device *candidate = to_rtc_device(dev);
 
-	if (!candidate->ops->set_alarm)
+	if (!test_bit(RTC_FEATURE_ALARM, candidate->features))
 		return 0;
 	if (!device_may_wakeup(candidate->dev.parent))
 		return 0;

kernel/sched/cpufreq_schedutil.c

@@ -537,9 +537,17 @@ static struct attribute *sugov_attrs[] = {
 };
 ATTRIBUTE_GROUPS(sugov);
 
+static void sugov_tunables_free(struct kobject *kobj)
+{
+	struct gov_attr_set *attr_set = container_of(kobj, struct gov_attr_set, kobj);
+
+	kfree(to_sugov_tunables(attr_set));
+}
+
 static struct kobj_type sugov_tunables_ktype = {
 	.default_groups = sugov_groups,
 	.sysfs_ops = &governor_sysfs_ops,
+	.release = &sugov_tunables_free,
 };
 
 /********************** cpufreq governor interface *********************/
@@ -639,12 +647,10 @@ static struct sugov_tunables *sugov_tunables_alloc(struct sugov_policy *sg_polic
 	return tunables;
 }
 
-static void sugov_tunables_free(struct sugov_tunables *tunables)
+static void sugov_clear_global_tunables(void)
 {
 	if (!have_governor_per_policy())
 		global_tunables = NULL;
-
-	kfree(tunables);
 }
 
 static int sugov_init(struct cpufreq_policy *policy)
@@ -707,7 +713,7 @@ out:
 fail:
 	kobject_put(&tunables->attr_set.kobj);
 	policy->governor_data = NULL;
-	sugov_tunables_free(tunables);
+	sugov_clear_global_tunables();
 
 stop_kthread:
 	sugov_kthread_stop(sg_policy);
@@ -734,7 +740,7 @@ static void sugov_exit(struct cpufreq_policy *policy)
 	count = gov_attr_set_put(&tunables->attr_set, &sg_policy->tunables_hook);
 	policy->governor_data = NULL;
 	if (!count)
-		sugov_tunables_free(tunables);
+		sugov_clear_global_tunables();
 
 	mutex_unlock(&global_tunables_lock);