Merge branch 'pm-cpuidle'

Merge cpuidle updates for 6.16-rc1:

 - Optimize bucket assignment when next_timer_ns equals KTIME_MAX in the
   menu cpuidle governor (Zhongqiu Han).

 - Convert the cpuidle PSCI driver to a faux device one (Sudeep Holla).

 - Add C1 demotion on/off sysfs knob to the intel_idle driver (Artem
   Bityutskiy).

 - Fix typos in two comments in the teo cpuidle governor (Atul Kumar
   Pant).

* pm-cpuidle:
  cpuidle: psci: Avoid initializing faux device if no DT idle states are present
  Documentation: ABI: testing: document the new cpuidle sysfs file
  Documentation: admin-guide: pm: Document intel_idle C1 demotion
  intel_idle: Add C1 demotion on/off sysfs knob
  cpuidle: psci: Transition to the faux device interface
  cpuidle: menu: Optimize bucket assignment when next_timer_ns equals KTIME_MAX
  cpuidle: teo: Fix typos in two comments

Documentation/ABI/testing/sysfs-devices-system-cpu

@@ -111,6 +111,7 @@ What:		/sys/devices/system/cpu/cpuidle/available_governors
 		/sys/devices/system/cpu/cpuidle/current_driver
 		/sys/devices/system/cpu/cpuidle/current_governor
 		/sys/devices/system/cpu/cpuidle/current_governer_ro
+		/sys/devices/system/cpu/cpuidle/intel_c1_demotion
 Date:		September 2007
 Contact:	Linux kernel mailing list <linux-kernel@vger.kernel.org>
 Description:	Discover cpuidle policy and mechanism
@@ -132,7 +133,11 @@ Description:	Discover cpuidle policy and mechanism
 
 		current_governor_ro: (RO) displays current idle policy.
 
-		See Documentation/admin-guide/pm/cpuidle.rst and
+		intel_c1_demotion: (RW) enables/disables the C1 demotion
+		feature on Intel CPUs.
+
+		See Documentation/admin-guide/pm/cpuidle.rst,
+		Documentation/admin-guide/pm/intel_idle.rst, and
 		Documentation/driver-api/pm/cpuidle.rst for more information.
 
 

Documentation/admin-guide/pm/intel_idle.rst

@@ -38,6 +38,27 @@ instruction at all.
 only way to pass early-configuration-time parameters to it is via the kernel
 command line.
 
+Sysfs Interface
+===============
+
+The ``intel_idle`` driver exposes the following ``sysfs`` attributes in
+``/sys/devices/system/cpu/cpuidle/``:
+
+``intel_c1_demotion``
+	Enable or disable C1 demotion for all CPUs in the system. This file is
+	only exposed on platforms that support the C1 demotion feature and where
+	it was tested. Value 0 means that C1 demotion is disabled, value 1 means
+	that it is enabled. Write 0 or 1 to disable or enable C1 demotion for
+	all CPUs.
+
+	The C1 demotion feature involves the platform firmware demoting deep
+	C-state requests from the OS (e.g., C6 requests) to C1. The idea is that
+	firmware monitors CPU wake-up rate, and if it is higher than a
+	platform-specific threshold, the firmware demotes deep C-state requests
+	to C1. For example, Linux requests C6, but firmware noticed too many
+	wake-ups per second, and it keeps the CPU in C1. When the CPU stays in
+	C1 long enough, the platform promotes it back to C6. This may improve
+	some workloads' performance, but it may also increase power consumption.
 
 .. _intel-idle-enumeration-of-states:
 

drivers/cpuidle/cpuidle-psci.c

@@ -16,7 +16,7 @@
 #include <linux/kernel.h>
 #include <linux/module.h>
 #include <linux/of.h>
-#include <linux/platform_device.h>
+#include <linux/device/faux.h>
 #include <linux/psci.h>
 #include <linux/pm_domain.h>
 #include <linux/pm_runtime.h>
@@ -407,14 +407,14 @@ deinit:
  * to register cpuidle driver then rollback to cancel all CPUs
  * registration.
  */
-static int psci_cpuidle_probe(struct platform_device *pdev)
+static int psci_cpuidle_probe(struct faux_device *fdev)
 {
 	int cpu, ret;
 	struct cpuidle_driver *drv;
 	struct cpuidle_device *dev;
 
 	for_each_present_cpu(cpu) {
-		ret = psci_idle_init_cpu(&pdev->dev, cpu);
+		ret = psci_idle_init_cpu(&fdev->dev, cpu);
 		if (ret)
 			goto out_fail;
 	}
@@ -434,26 +434,37 @@ out_fail:
 	return ret;
 }
 
-static struct platform_driver psci_cpuidle_driver = {
+static struct faux_device_ops psci_cpuidle_ops = {
 	.probe = psci_cpuidle_probe,
-	.driver = {
-		.name = "psci-cpuidle",
-	},
 };
 
+static bool __init dt_idle_state_present(void)
+{
+	struct device_node *cpu_node __free(device_node);
+	struct device_node *state_node __free(device_node);
+
+	cpu_node = of_cpu_device_node_get(cpumask_first(cpu_possible_mask));
+	if (!cpu_node)
+		return false;
+
+	state_node = of_get_cpu_state_node(cpu_node, 0);
+	if (!state_node)
+		return false;
+
+	return !!of_match_node(psci_idle_state_match, state_node);
+}
+
 static int __init psci_idle_init(void)
 {
-	struct platform_device *pdev;
-	int ret;
+	struct faux_device *fdev;
 
-	ret = platform_driver_register(&psci_cpuidle_driver);
-	if (ret)
-		return ret;
+	if (!dt_idle_state_present())
+		return 0;
 
-	pdev = platform_device_register_simple("psci-cpuidle", -1, NULL, 0);
-	if (IS_ERR(pdev)) {
-		platform_driver_unregister(&psci_cpuidle_driver);
-		return PTR_ERR(pdev);
+	fdev = faux_device_create("psci-cpuidle", NULL, &psci_cpuidle_ops);
+	if (!fdev) {
+		pr_err("Failed to create psci-cpuidle device\n");
+		return -ENODEV;
 	}
 
 	return 0;

drivers/cpuidle/governors/menu.c

@@ -255,7 +255,7 @@ static int menu_select(struct cpuidle_driver *drv, struct cpuidle_device *dev,
 		 */
 		data->next_timer_ns = KTIME_MAX;
 		delta_tick = TICK_NSEC / 2;
-		data->bucket = which_bucket(KTIME_MAX);
+		data->bucket = BUCKETS - 1;
 	}
 
 	if (unlikely(drv->state_count <= 1 || latency_req == 0) ||

drivers/cpuidle/governors/teo.c

@@ -19,7 +19,7 @@
  *
  * Of course, non-timer wakeup sources are more important in some use cases,
  * but even then it is generally unnecessary to consider idle duration values
- * greater than the time time till the next timer event, referred as the sleep
+ * greater than the time till the next timer event, referred as the sleep
  * length in what follows, because the closest timer will ultimately wake up the
  * CPU anyway unless it is woken up earlier.
  *
@@ -311,7 +311,7 @@ static int teo_select(struct cpuidle_driver *drv, struct cpuidle_device *dev,
 		struct cpuidle_state *s = &drv->states[i];
 
 		/*
-		 * Update the sums of idle state mertics for all of the states
+		 * Update the sums of idle state metrics for all of the states
 		 * shallower than the current one.
 		 */
 		intercept_sum += prev_bin->intercepts;

drivers/idle/intel_idle.c

@@ -48,9 +48,11 @@
 #include <trace/events/power.h>
 #include <linux/sched.h>
 #include <linux/sched/smt.h>
+#include <linux/mutex.h>
 #include <linux/notifier.h>
 #include <linux/cpu.h>
 #include <linux/moduleparam.h>
+#include <linux/sysfs.h>
 #include <asm/cpuid.h>
 #include <asm/cpu_device_id.h>
 #include <asm/intel-family.h>
@@ -92,9 +94,15 @@ struct idle_cpu {
 	 */
 	unsigned long auto_demotion_disable_flags;
 	bool disable_promotion_to_c1e;
+	bool c1_demotion_supported;
 	bool use_acpi;
 };
 
+static bool c1_demotion_supported;
+static DEFINE_MUTEX(c1_demotion_mutex);
+
+static struct device *sysfs_root __initdata;
+
 static const struct idle_cpu *icpu __initdata;
 static struct cpuidle_state *cpuidle_state_table __initdata;
 
@@ -1549,18 +1557,21 @@ static const struct idle_cpu idle_cpu_gmt __initconst = {
 static const struct idle_cpu idle_cpu_spr __initconst = {
 	.state_table = spr_cstates,
 	.disable_promotion_to_c1e = true,
+	.c1_demotion_supported = true,
 	.use_acpi = true,
 };
 
 static const struct idle_cpu idle_cpu_gnr __initconst = {
 	.state_table = gnr_cstates,
 	.disable_promotion_to_c1e = true,
+	.c1_demotion_supported = true,
 	.use_acpi = true,
 };
 
 static const struct idle_cpu idle_cpu_gnrd __initconst = {
 	.state_table = gnrd_cstates,
 	.disable_promotion_to_c1e = true,
+	.c1_demotion_supported = true,
 	.use_acpi = true,
 };
 
@@ -1599,12 +1610,14 @@ static const struct idle_cpu idle_cpu_snr __initconst = {
 static const struct idle_cpu idle_cpu_grr __initconst = {
 	.state_table = grr_cstates,
 	.disable_promotion_to_c1e = true,
+	.c1_demotion_supported = true,
 	.use_acpi = true,
 };
 
 static const struct idle_cpu idle_cpu_srf __initconst = {
 	.state_table = srf_cstates,
 	.disable_promotion_to_c1e = true,
+	.c1_demotion_supported = true,
 	.use_acpi = true,
 };
 
@@ -2324,6 +2337,88 @@ static void __init intel_idle_cpuidle_devices_uninit(void)
 		cpuidle_unregister_device(per_cpu_ptr(intel_idle_cpuidle_devices, i));
 }
 
+static void intel_c1_demotion_toggle(void *enable)
+{
+	unsigned long long msr_val;
+
+	rdmsrl(MSR_PKG_CST_CONFIG_CONTROL, msr_val);
+	/*
+	 * Enable/disable C1 undemotion along with C1 demotion, as this is the
+	 * most sensible configuration in general.
+	 */
+	if (enable)
+		msr_val |= NHM_C1_AUTO_DEMOTE | SNB_C1_AUTO_UNDEMOTE;
+	else
+		msr_val &= ~(NHM_C1_AUTO_DEMOTE | SNB_C1_AUTO_UNDEMOTE);
+	wrmsrl(MSR_PKG_CST_CONFIG_CONTROL, msr_val);
+}
+
+static ssize_t intel_c1_demotion_store(struct device *dev,
+				       struct device_attribute *attr,
+				       const char *buf, size_t count)
+{
+	bool enable;
+	int err;
+
+	err = kstrtobool(buf, &enable);
+	if (err)
+		return err;
+
+	mutex_lock(&c1_demotion_mutex);
+	/* Enable/disable C1 demotion on all CPUs */
+	on_each_cpu(intel_c1_demotion_toggle, (void *)enable, 1);
+	mutex_unlock(&c1_demotion_mutex);
+
+	return count;
+}
+
+static ssize_t intel_c1_demotion_show(struct device *dev,
+				      struct device_attribute *attr, char *buf)
+{
+	unsigned long long msr_val;
+
+	/*
+	 * Read the MSR value for a CPU and assume it is the same for all CPUs. Any other
+	 * configuration would be a BIOS bug.
+	 */
+	rdmsrl(MSR_PKG_CST_CONFIG_CONTROL, msr_val);
+	return sysfs_emit(buf, "%d\n", !!(msr_val & NHM_C1_AUTO_DEMOTE));
+}
+static DEVICE_ATTR_RW(intel_c1_demotion);
+
+static int __init intel_idle_sysfs_init(void)
+{
+	int err;
+
+	if (!c1_demotion_supported)
+		return 0;
+
+	sysfs_root = bus_get_dev_root(&cpu_subsys);
+	if (!sysfs_root)
+		return 0;
+
+	err = sysfs_add_file_to_group(&sysfs_root->kobj,
+				      &dev_attr_intel_c1_demotion.attr,
+				      "cpuidle");
+	if (err) {
+		put_device(sysfs_root);
+		return err;
+	}
+
+	return 0;
+}
+
+static void __init intel_idle_sysfs_uninit(void)
+{
+	if (!sysfs_root)
+		return;
+
+	sysfs_remove_file_from_group(&sysfs_root->kobj,
+				     &dev_attr_intel_c1_demotion.attr,
+				     "cpuidle");
+	put_device(sysfs_root);
+}
+
 static int __init intel_idle_init(void)
 {
 	const struct x86_cpu_id *id;
@@ -2374,6 +2469,8 @@ static int __init intel_idle_init(void)
 		auto_demotion_disable_flags = icpu->auto_demotion_disable_flags;
 		if (icpu->disable_promotion_to_c1e)
 			c1e_promotion = C1E_PROMOTION_DISABLE;
+		if (icpu->c1_demotion_supported)
+			c1_demotion_supported = true;
 		if (icpu->use_acpi || force_use_acpi)
 			intel_idle_acpi_cst_extract();
 	} else if (!intel_idle_acpi_cst_extract()) {
@@ -2387,6 +2484,10 @@ static int __init intel_idle_init(void)
 	if (!intel_idle_cpuidle_devices)
 		return -ENOMEM;
 
+	retval = intel_idle_sysfs_init();
+	if (retval)
+		pr_warn("failed to initialized sysfs");
+
 	intel_idle_cpuidle_driver_init(&intel_idle_driver);
 
 	retval = cpuidle_register_driver(&intel_idle_driver);
@@ -2411,6 +2512,7 @@ hp_setup_fail:
 	intel_idle_cpuidle_devices_uninit();
 	cpuidle_unregister_driver(&intel_idle_driver);
 init_driver_fail:
+	intel_idle_sysfs_uninit();
 	free_percpu(intel_idle_cpuidle_devices);
 	return retval;