PM / QoS: Introcuce latency tolerance device PM QoS type (2d984ad1) · Commits · 戴 / test

Documentation/ABI/testing/sysfs-devices-power

+26 −1

Original line number	Diff line number	Diff line
		@@ -187,7 +187,7 @@ Description:
		Not all drivers support this attribute. If it isn't supported,
		attempts to read or write it will yield I/O errors.

		What: /sys/devices/.../power/pm_qos_latency_us
		What: /sys/devices/.../power/pm_qos_resume_latency_us
		Date: March 2012
		Contact: Rafael J. Wysocki <rjw@rjwysocki.net>
		Description:
		@@ -205,6 +205,31 @@ Description:
		This attribute has no effect on system-wide suspend/resume and
		hibernation.

		What: /sys/devices/.../power/pm_qos_latency_tolerance_us
		Date: January 2014
		Contact: Rafael J. Wysocki <rjw@rjwysocki.net>
		Description:
		The /sys/devices/.../power/pm_qos_latency_tolerance_us attribute
		contains the PM QoS active state latency tolerance limit for the
		given device in microseconds. That is the maximum memory access
		latency the device can suffer without any visible adverse
		effects on user space functionality. If that value is the
		string "any", the latency does not matter to user space at all,
		but hardware should not be allowed to set the latency tolerance
		for the device automatically.

		Reading "auto" from this file means that the maximum memory
		access latency for the device may be determined automatically
		by the hardware as needed. Writing "auto" to it allows the
		hardware to be switched to this mode if there are no other
		latency tolerance requirements from the kernel side.

		This attribute is only present if the feature controlled by it
		is supported by the hardware.

		This attribute has no effect on runtime suspend and resume of
		devices and on system-wide suspend/resume and hibernation.

		What: /sys/devices/.../power/pm_qos_no_power_off
		Date: September 2012
		Contact: Rafael J. Wysocki <rjw@rjwysocki.net>

Documentation/power/pm_qos_interface.txt

+50 −9

Original line number	Diff line number	Diff line
		@@ -88,17 +88,19 @@ node.

		2. PM QoS per-device latency and flags framework

		For each device, there are two lists of PM QoS requests. One is maintained
		along with the aggregated target of resume latency value and the other is for
		PM QoS flags. Values are updated in response to changes of the request list.
		For each device, there are three lists of PM QoS requests. Two of them are
		maintained along with the aggregated targets of resume latency and active
		state latency tolerance (in microseconds) and the third one is for PM QoS flags.
		Values are updated in response to changes of the request list.

		Target resume latency value is simply the minimum of the request values held in
		the parameter list elements. The PM QoS flags aggregate value is a gather
		(bitwise OR) of all list elements' values. Two device PM QoS flags are defined
		currently: PM_QOS_FLAG_NO_POWER_OFF and PM_QOS_FLAG_REMOTE_WAKEUP.
		The target values of resume latency and active state latency tolerance are
		simply the minimum of the request values held in the parameter list elements.
		The PM QoS flags aggregate value is a gather (bitwise OR) of all list elements'
		values. Two device PM QoS flags are defined currently: PM_QOS_FLAG_NO_POWER_OFF
		and PM_QOS_FLAG_REMOTE_WAKEUP.

		Note: the aggregated target value is implemented in such a way that reading the
		aggregated value does not require any locking mechanism.
		Note: The aggregated target values are implemented in such a way that reading
		the aggregated value does not require any locking mechanism.


		From kernel mode the use of this interface is the following:
		@@ -177,3 +179,42 @@ The callback is called when the aggregated value for any device is changed
		int dev_pm_qos_remove_global_notifier(notifier):
		Removes the notification callback function from the global notification tree
		of the framework.


		Active state latency tolerance

		This device PM QoS type is used to support systems in which hardware may switch
		to energy-saving operation modes on the fly. In those systems, if the operation
		mode chosen by the hardware attempts to save energy in an overly aggressive way,
		it may cause excess latencies to be visible to software, causing it to miss
		certain protocol requirements or target frame or sample rates etc.

		If there is a latency tolerance control mechanism for a given device available
		to software, the .set_latency_tolerance callback in that device's dev_pm_info
		structure should be populated. The routine pointed to by it is should implement
		whatever is necessary to transfer the effective requirement value to the
		hardware.

		Whenever the effective latency tolerance changes for the device, its
		.set_latency_tolerance() callback will be executed and the effective value will
		be passed to it. If that value is negative, which means that the list of
		latency tolerance requirements for the device is empty, the callback is expected
		to switch the underlying hardware latency tolerance control mechanism to an
		autonomous mode if available. If that value is PM_QOS_LATENCY_ANY, in turn, and
		the hardware supports a special "no requirement" setting, the callback is
		expected to use it. That allows software to prevent the hardware from
		automatically updating the device's latency tolerance in response to its power
		state changes (e.g. during transitions from D3cold to D0), which generally may
		be done in the autonomous latency tolerance control mode.

		If .set_latency_tolerance() is present for the device, sysfs attribute
		pm_qos_latency_tolerance_us will be present in the devivce's power directory.
		Then, user space can use that attribute to specify its latency tolerance
		requirement for the device, if any. Writing "any" to it means "no requirement,
		but do not let the hardware control latency tolerance" and writing "auto" to it
		allows the hardware to be switched to the autonomous mode if there are no other
		requirements from the kernel side in the device's list.

		Kernel code can use the functions described above along with the
		DEV_PM_QOS_LATENCY_TOLERANCE device PM QoS type to add, remove and update
		latency tolerance requirements for devices.

drivers/base/power/qos.c

+122 −22

Original line number	Diff line number	Diff line
		@@ -151,6 +151,14 @@ static int apply_constraint(struct dev_pm_qos_request *req,
		req);
		}
		break;
		case DEV_PM_QOS_LATENCY_TOLERANCE:
		ret = pm_qos_update_target(&qos->latency_tolerance,
		&req->data.pnode, action, value);
		if (ret) {
		value = pm_qos_read_value(&qos->latency_tolerance);
		req->dev->power.set_latency_tolerance(req->dev, value);
		}
		break;
		case DEV_PM_QOS_FLAGS:
		ret = pm_qos_update_flags(&qos->flags, &req->data.flr,
		action, value);
		@@ -194,6 +202,13 @@ static int dev_pm_qos_constraints_allocate(struct device *dev)
		c->type = PM_QOS_MIN;
		c->notifiers = n;

		c = &qos->latency_tolerance;
		plist_head_init(&c->list);
		c->target_value = PM_QOS_LATENCY_TOLERANCE_DEFAULT_VALUE;
		c->default_value = PM_QOS_LATENCY_TOLERANCE_DEFAULT_VALUE;
		c->no_constraint_value = PM_QOS_LATENCY_TOLERANCE_NO_CONSTRAINT;
		c->type = PM_QOS_MIN;

		INIT_LIST_HEAD(&qos->flags.list);

		spin_lock_irq(&dev->power.lock);
		@@ -247,6 +262,11 @@ void dev_pm_qos_constraints_destroy(struct device *dev)
		apply_constraint(req, PM_QOS_REMOVE_REQ, PM_QOS_DEFAULT_VALUE);
		memset(req, 0, sizeof(*req));
		}
		c = &qos->latency_tolerance;
		plist_for_each_entry_safe(req, tmp, &c->list, data.pnode) {
		apply_constraint(req, PM_QOS_REMOVE_REQ, PM_QOS_DEFAULT_VALUE);
		memset(req, 0, sizeof(*req));
		}
		f = &qos->flags;
		list_for_each_entry_safe(req, tmp, &f->list, data.flr.node) {
		apply_constraint(req, PM_QOS_REMOVE_REQ, PM_QOS_DEFAULT_VALUE);
		@@ -266,6 +286,40 @@ void dev_pm_qos_constraints_destroy(struct device *dev)
		mutex_unlock(&dev_pm_qos_sysfs_mtx);
		}

		static bool dev_pm_qos_invalid_request(struct device *dev,
		struct dev_pm_qos_request *req)
		{
		return !req \|\| (req->type == DEV_PM_QOS_LATENCY_TOLERANCE
		&& !dev->power.set_latency_tolerance);
		}

		static int __dev_pm_qos_add_request(struct device *dev,
		struct dev_pm_qos_request *req,
		enum dev_pm_qos_req_type type, s32 value)
		{
		int ret = 0;

		if (!dev \|\| dev_pm_qos_invalid_request(dev, req))
		return -EINVAL;

		if (WARN(dev_pm_qos_request_active(req),
		"%s() called for already added request\n", __func__))
		return -EINVAL;

		if (IS_ERR(dev->power.qos))
		ret = -ENODEV;
		else if (!dev->power.qos)
		ret = dev_pm_qos_constraints_allocate(dev);

		trace_dev_pm_qos_add_request(dev_name(dev), type, value);
		if (!ret) {
		req->dev = dev;
		req->type = type;
		ret = apply_constraint(req, PM_QOS_ADD_REQ, value);
		}
		return ret;
		}

		/**
		* dev_pm_qos_add_request - inserts new qos request into the list
		* @dev: target device for the constraint
		@@ -291,31 +345,11 @@ void dev_pm_qos_constraints_destroy(struct device *dev)
		int dev_pm_qos_add_request(struct device dev, struct dev_pm_qos_request req,
		enum dev_pm_qos_req_type type, s32 value)
		{
		int ret = 0;

		if (!dev \|\| !req) /guard against callers passing in null /
		return -EINVAL;

		if (WARN(dev_pm_qos_request_active(req),
		"%s() called for already added request\n", __func__))
		return -EINVAL;
		int ret;

		mutex_lock(&dev_pm_qos_mtx);

		if (IS_ERR(dev->power.qos))
		ret = -ENODEV;
		else if (!dev->power.qos)
		ret = dev_pm_qos_constraints_allocate(dev);

		trace_dev_pm_qos_add_request(dev_name(dev), type, value);
		if (!ret) {
		req->dev = dev;
		req->type = type;
		ret = apply_constraint(req, PM_QOS_ADD_REQ, value);
		}

		ret = __dev_pm_qos_add_request(dev, req, type, value);
		mutex_unlock(&dev_pm_qos_mtx);

		return ret;
		}
		EXPORT_SYMBOL_GPL(dev_pm_qos_add_request);
		@@ -343,6 +377,7 @@ static int __dev_pm_qos_update_request(struct dev_pm_qos_request *req,

		switch(req->type) {
		case DEV_PM_QOS_RESUME_LATENCY:
		case DEV_PM_QOS_LATENCY_TOLERANCE:
		curr_value = req->data.pnode.prio;
		break;
		case DEV_PM_QOS_FLAGS:
		@@ -563,6 +598,10 @@ static void __dev_pm_qos_drop_user_request(struct device *dev,
		req = dev->power.qos->resume_latency_req;
		dev->power.qos->resume_latency_req = NULL;
		break;
		case DEV_PM_QOS_LATENCY_TOLERANCE:
		req = dev->power.qos->latency_tolerance_req;
		dev->power.qos->latency_tolerance_req = NULL;
		break;
		case DEV_PM_QOS_FLAGS:
		req = dev->power.qos->flags_req;
		dev->power.qos->flags_req = NULL;
		@@ -768,6 +807,67 @@ int dev_pm_qos_update_flags(struct device *dev, s32 mask, bool set)
		pm_runtime_put(dev);
		return ret;
		}

		/**
		* dev_pm_qos_get_user_latency_tolerance - Get user space latency tolerance.
		* @dev: Device to obtain the user space latency tolerance for.
		*/
		s32 dev_pm_qos_get_user_latency_tolerance(struct device *dev)
		{
		s32 ret;

		mutex_lock(&dev_pm_qos_mtx);
		ret = IS_ERR_OR_NULL(dev->power.qos)
		\|\| !dev->power.qos->latency_tolerance_req ?
		PM_QOS_LATENCY_TOLERANCE_NO_CONSTRAINT :
		dev->power.qos->latency_tolerance_req->data.pnode.prio;
		mutex_unlock(&dev_pm_qos_mtx);
		return ret;
		}

		/**
		* dev_pm_qos_update_user_latency_tolerance - Update user space latency tolerance.
		* @dev: Device to update the user space latency tolerance for.
		* @val: New user space latency tolerance for @dev (negative values disable).
		*/
		int dev_pm_qos_update_user_latency_tolerance(struct device *dev, s32 val)
		{
		int ret;

		mutex_lock(&dev_pm_qos_mtx);

		if (IS_ERR_OR_NULL(dev->power.qos)
		\|\| !dev->power.qos->latency_tolerance_req) {
		struct dev_pm_qos_request *req;

		if (val < 0) {
		ret = -EINVAL;
		goto out;
		}
		req = kzalloc(sizeof(*req), GFP_KERNEL);
		if (!req) {
		ret = -ENOMEM;
		goto out;
		}
		ret = __dev_pm_qos_add_request(dev, req, DEV_PM_QOS_LATENCY_TOLERANCE, val);
		if (ret < 0) {
		kfree(req);
		goto out;
		}
		dev->power.qos->latency_tolerance_req = req;
		} else {
		if (val < 0) {
		__dev_pm_qos_drop_user_request(dev, DEV_PM_QOS_LATENCY_TOLERANCE);
		ret = 0;
		} else {
		ret = __dev_pm_qos_update_request(dev->power.qos->latency_tolerance_req, val);
		}
		}

		out:
		mutex_unlock(&dev_pm_qos_mtx);
		return ret;
		}
		#else /* !CONFIG_PM_RUNTIME */
		static void __dev_pm_qos_hide_latency_limit(struct device *dev) {}
		static void __dev_pm_qos_hide_flags(struct device *dev) {}

drivers/base/power/sysfs.c

+58 −7

Original line number	Diff line number	Diff line
		@@ -246,6 +246,40 @@ static ssize_t pm_qos_resume_latency_store(struct device *dev,
		static DEVICE_ATTR(pm_qos_resume_latency_us, 0644,
		pm_qos_resume_latency_show, pm_qos_resume_latency_store);

		static ssize_t pm_qos_latency_tolerance_show(struct device *dev,
		struct device_attribute *attr,
		char *buf)
		{
		s32 value = dev_pm_qos_get_user_latency_tolerance(dev);

		if (value < 0)
		return sprintf(buf, "auto\n");
		else if (value == PM_QOS_LATENCY_ANY)
		return sprintf(buf, "any\n");

		return sprintf(buf, "%d\n", value);
		}

		static ssize_t pm_qos_latency_tolerance_store(struct device *dev,
		struct device_attribute *attr,
		const char *buf, size_t n)
		{
		s32 value;
		int ret;

		if (kstrtos32(buf, 0, &value)) {
		if (!strcmp(buf, "auto") \|\| !strcmp(buf, "auto\n"))
		value = PM_QOS_LATENCY_TOLERANCE_NO_CONSTRAINT;
		else if (!strcmp(buf, "any") \|\| !strcmp(buf, "any\n"))
		value = PM_QOS_LATENCY_ANY;
		}
		ret = dev_pm_qos_update_user_latency_tolerance(dev, value);
		return ret < 0 ? ret : n;
		}

		static DEVICE_ATTR(pm_qos_latency_tolerance_us, 0644,
		pm_qos_latency_tolerance_show, pm_qos_latency_tolerance_store);

		static ssize_t pm_qos_no_power_off_show(struct device *dev,
		struct device_attribute *attr,
		char *buf)
		@@ -631,6 +665,17 @@ static struct attribute_group pm_qos_resume_latency_attr_group = {
		.attrs = pm_qos_resume_latency_attrs,
		};

		static struct attribute *pm_qos_latency_tolerance_attrs[] = {
		#ifdef CONFIG_PM_RUNTIME
		&dev_attr_pm_qos_latency_tolerance_us.attr,
		#endif /* CONFIG_PM_RUNTIME */
		NULL,
		};
		static struct attribute_group pm_qos_latency_tolerance_attr_group = {
		.name = power_group_name,
		.attrs = pm_qos_latency_tolerance_attrs,
		};

		static struct attribute *pm_qos_flags_attrs[] = {
		#ifdef CONFIG_PM_RUNTIME
		&dev_attr_pm_qos_no_power_off.attr,
		@@ -656,18 +701,23 @@ int dpm_sysfs_add(struct device *dev)
		if (rc)
		goto err_out;
		}

		if (device_can_wakeup(dev)) {
		rc = sysfs_merge_group(&dev->kobj, &pm_wakeup_attr_group);
		if (rc) {
		if (pm_runtime_callbacks_present(dev))
		sysfs_unmerge_group(&dev->kobj,
		&pm_runtime_attr_group);
		goto err_out;
		if (rc)
		goto err_runtime;
		}
		if (dev->power.set_latency_tolerance) {
		rc = sysfs_merge_group(&dev->kobj,
		&pm_qos_latency_tolerance_attr_group);
		if (rc)
		goto err_wakeup;
		}
		return 0;

		err_wakeup:
		sysfs_unmerge_group(&dev->kobj, &pm_wakeup_attr_group);
		err_runtime:
		sysfs_unmerge_group(&dev->kobj, &pm_runtime_attr_group);
		err_out:
		sysfs_remove_group(&dev->kobj, &pm_attr_group);
		return rc;
		@@ -710,6 +760,7 @@ void rpm_sysfs_remove(struct device *dev)

		void dpm_sysfs_remove(struct device *dev)
		{
		sysfs_unmerge_group(&dev->kobj, &pm_qos_latency_tolerance_attr_group);
		dev_pm_qos_constraints_destroy(dev);
		rpm_sysfs_remove(dev);
		sysfs_unmerge_group(&dev->kobj, &pm_wakeup_attr_group);

include/linux/pm.h

+1 −0

Original line number	Diff line number	Diff line
		@@ -582,6 +582,7 @@ struct dev_pm_info {
		unsigned long accounting_timestamp;
		#endif
		struct pm_subsys_data subsys_data; / Owned by the subsystem. */
		void (set_latency_tolerance)(struct device , s32);
		struct dev_pm_qos *qos;
		};

Admin message