Lines Matching +full:power +full:- +full:on +full:- +full:delay
2 Runtime Power Management Framework for I/O Devices
5 (C) 2009-2011 Rafael J. Wysocki <rjw@sisk.pl>, Novell Inc.
14 Support for runtime power management (runtime PM) of I/O devices is provided
15 at the power management core (PM core) level by means of:
17 * The power management workqueue pm_wq in which bus types and device drivers can
18 put their PM-related work items. It is strongly recommended that pm_wq be
20 them to be synchronized with system-wide power transitions (suspend to RAM,
22 include/linux/pm_runtime.h and defined in kernel/power/main.c.
24 * A number of runtime PM fields in the 'power' member of 'struct device' (which
31 * A set of helper functions defined in drivers/base/power/runtime.c that can be
53 The ->runtime_suspend(), ->runtime_resume() and ->runtime_idle() callbacks
57 1. PM domain of the device, if the device's PM domain object, dev->pm_domain,
60 2. Device type of the device, if both dev->type and dev->type->pm are present.
62 3. Device class of the device, if both dev->class and dev->class->pm are
65 4. Bus type of the device, if both dev->bus and dev->bus->pm are present.
69 dev->driver->pm directly (if present).
73 and bus type. Moreover, the high-priority one will always take precedence over
74 a low-priority one. The PM domain, bus type, device type and class callbacks
75 are referred to as subsystem-level callbacks in what follows.
79 the PM core that it is safe to run the ->runtime_suspend(), ->runtime_resume()
80 and ->runtime_idle() callbacks for the given device in atomic context with
86 The subsystem-level suspend callback, if present, is _entirely_ _responsible_
88 include executing the device driver's own ->runtime_suspend() callback (from the
89 PM core's point of view it is not necessary to implement a ->runtime_suspend()
90 callback in a device driver as long as the subsystem-level suspend callback
93 * Once the subsystem-level suspend callback (or the driver suspend callback,
96 put into a low power state. It is supposed to mean, however, that the
102 * If the suspend callback returns -EBUSY or -EAGAIN, the device's runtime PM
106 * If the suspend callback returns an error code different from -EBUSY and
107 -EAGAIN, the PM core regards this as a fatal error and will refuse to run
113 mechanism allowing the device to request a change of its power state, such as
115 device, then ->runtime_suspend() should return -EBUSY. On the other hand, if
117 low-power state during the execution of the suspend callback, it is expected
119 should be enabled for all input devices put into low-power states at run time.
121 The subsystem-level resume callback, if present, is **entirely responsible** for
123 include executing the device driver's own ->runtime_resume() callback (from the
124 PM core's point of view it is not necessary to implement a ->runtime_resume()
125 callback in a device driver as long as the subsystem-level resume callback knows
128 * Once the subsystem-level resume callback (or the driver resume callback, if
140 The idle callback (a subsystem-level one, if present, or the driver one) is
150 The action performed by the idle callback is totally dependent on the subsystem
158 device last busy mark, pm_runtime_mark_last_busy(), to control the delay under
160 started a delayed suspend), the routine must return a non-zero value. Negative
168 ->runtime_suspend() in parallel with ->runtime_resume() or with another
169 instance of ->runtime_suspend() for the same device) with the exception that
170 ->runtime_suspend() or ->runtime_resume() can be executed in parallel with
171 ->runtime_idle() (although ->runtime_idle() will not be started while any
174 (2) ->runtime_idle() and ->runtime_suspend() can only be executed for 'active'
175 devices (i.e. the PM core will only execute ->runtime_idle() or
176 ->runtime_suspend() for the devices the runtime PM status of which is
179 (3) ->runtime_idle() and ->runtime_suspend() can only be executed for a device
181 'active' children of which is equal to zero, or the 'power.ignore_children'
184 (4) ->runtime_resume() can only be executed for 'suspended' devices (i.e. the
185 PM core will only execute ->runtime_resume() for the devices the runtime
191 * If ->runtime_suspend() is about to be executed or there's a pending request
192 to execute it, ->runtime_idle() will not be executed for the same device.
194 * A request to execute or to schedule the execution of ->runtime_suspend()
195 will cancel any pending requests to execute ->runtime_idle() for the same
198 * If ->runtime_resume() is about to be executed or there's a pending request
201 * A request to execute ->runtime_resume() will cancel any pending or
212 - timer used for scheduling (delayed) suspend and autosuspend requests
215 - timer expiration time, in jiffies (if this is different from zero, the
220 - work structure used for queuing up requests (i.e. work items in pm_wq)
223 - wait queue used if any of the helper functions needs to wait for another
227 - lock used for synchronization
230 - the usage counter of the device
233 - the count of 'active' children of the device
236 - if set, the value of child_count is ignored (but still updated)
239 - used for disabling the helper functions (they work normally if this is
244 - if set, there was a fatal error (one of the callbacks returned error code
250 - if set, ->runtime_idle() is being executed
253 - if set, there's a pending request (i.e. a work item queued up into pm_wq)
256 - type of request that's pending (valid if request_pending is set)
259 - set if ->runtime_resume() is about to be run while ->runtime_suspend() is
264 - the runtime PM status of the device; this field's initial value is
269 - if set, indicates that the user space has allowed the device driver to
270 power manage the device at run time via the /sys/devices/.../power/control
275 - indicates that the device does not use the runtime PM callbacks (see
280 - indicates that the ->runtime_suspend() and ->runtime_resume() callbacks
284 - indicates that the device's driver supports delayed autosuspend (see
289 - indicates that the PM core should attempt to carry out an autosuspend
293 - the delay time (in milliseconds) to be used for autosuspend
296 - the time (in jiffies) when the pm_runtime_mark_last_busy() helper
300 All of the above fields are members of the 'power' member of 'struct device'.
306 drivers/base/power/runtime.c and include/linux/pm_runtime.h:
309 - initialize the device runtime PM fields in 'struct dev_pm_info'
312 - make sure that the runtime PM of the device will be disabled after
316 - execute the subsystem-level idle callback for the device; returns an
317 error code on failure, where -EINPROGRESS means that ->runtime_idle() is
322 - execute the subsystem-level suspend callback for the device; returns 0 on
324 error code on failure, where -EAGAIN or -EBUSY means it is safe to attempt
325 to suspend the device again in future and -EACCES means that
326 'power.disable_depth' is different from 0
329 - same as pm_runtime_suspend() except that the autosuspend delay is taken
330 `into account;` if pm_runtime_autosuspend_expiration() says the delay has
335 - execute the subsystem-level resume callback for the device; returns 0 on
337 error code on failure, where -EAGAIN means it may be safe to attempt to
338 resume the device again in future, but 'power.runtime_error' should be
339 checked additionally, and -EACCES means that 'power.disable_depth' is
343 - run pm_runtime_resume(dev) and if successful, increment the device's
347 - submit a request to execute the subsystem-level idle callback for the
348 device (the request is represented by a work item in pm_wq); returns 0 on
352 - schedule the execution of the subsystem-level suspend callback for the
353 device when the autosuspend delay has expired; if the delay has already
356 `int pm_schedule_suspend(struct device *dev, unsigned int delay);`
357 - schedule the execution of the subsystem-level suspend callback for the
358 device in future, where 'delay' is the time to wait before queuing up a
359 suspend work item in pm_wq, in milliseconds (if 'delay' is zero, the work
360 item is queued up immediately); returns 0 on success, 1 if the device's PM
362 hasn't been scheduled (or queued up if 'delay' is 0); if the execution of
363 ->runtime_suspend() is already scheduled and not yet expired, the new
364 value of 'delay' will be used as the time to wait
367 - submit a request to execute the subsystem-level resume callback for the
368 device (the request is represented by a work item in pm_wq); returns 0 on
373 - increment the device's usage counter
376 - increment the device's usage counter, run pm_request_resume(dev) and
380 - increment the device's usage counter, run pm_runtime_resume(dev) and
382 note that it does not drop the device's usage counter on errors, so
388 - return -EINVAL if 'power.disable_depth' is nonzero; otherwise, if the
394 - return -EINVAL if 'power.disable_depth' is nonzero; otherwise, if the
396 or the device's usage_count is non-zero, increment the counter and
400 - decrement the device's usage counter
403 - decrement the device's usage counter; if the result is 0 then run
407 - decrement the device's usage counter; if the result is 0 then run
411 - decrement the device's usage counter; if the result is 0 then run
415 - decrement the device's usage counter; if the result is 0 then run
419 - decrement the device's usage counter; if the result is 0 then run
423 - decrement the device's 'power.disable_depth' field; if that field is equal
424 to zero, the runtime PM helper functions can execute subsystem-level
428 - increment the device's 'power.disable_depth' field (if the value of that
429 field was previously zero, this prevents subsystem-level runtime PM
431 pending runtime PM operations on the device are either completed or
433 necessary to execute the subsystem-level resume callback for the device
437 - check if there's a resume request pending for the device and resume it
439 regarding it and wait for all runtime PM operations on it in progress to
441 necessary to execute the subsystem-level resume callback for the device to
445 - set/unset the power.ignore_children flag of the device
448 - clear the device's 'power.runtime_error' flag, set the device's runtime
451 'power.runtime_error' is set or 'power.disable_depth' is greater than
453 which is not active and the 'power.ignore_children' flag of which is unset
456 - clear the device's 'power.runtime_error' flag, set the device's runtime
459 'power.runtime_error' is set or 'power.disable_depth' is greater than
463 - return true if the device's runtime PM status is 'active' or its
464 'power.disable_depth' field is not equal to zero, or false otherwise
467 - return true if the device's runtime PM status is 'suspended' and its
468 'power.disable_depth' field is equal to zero, or false otherwise
471 - return true if the device's runtime PM status is 'suspended'
474 - set the power.runtime_auto flag for the device and decrease its usage
475 counter (used by the /sys/devices/.../power/control interface to
476 effectively allow the device to be power managed at run time)
479 - unset the power.runtime_auto flag for the device and increase its usage
480 counter (used by the /sys/devices/.../power/control interface to
481 effectively prevent the device from being power managed at run time)
484 - set the power.no_callbacks flag for the device and remove the runtime
485 PM attributes from /sys/devices/.../power (or prevent them from being
489 - set the power.irq_safe flag for the device, causing the runtime-PM
493 - return true if power.irq_safe flag was set for the device, causing
494 the runtime-PM callbacks to be invoked with interrupts off
497 - set the power.last_busy field to the current time
500 - set the power.use_autosuspend flag, enabling autosuspend delays; call
502 power.autosuspend_delay is negative
505 - clear the power.use_autosuspend flag, disabling autosuspend delays;
507 power.autosuspend_delay is negative; call pm_runtime_idle
509 `void pm_runtime_set_autosuspend_delay(struct device *dev, int delay);`
510 - set the power.autosuspend_delay value to 'delay' (expressed in
511 milliseconds); if 'delay' is negative then runtime suspends are
512 prevented; if power.use_autosuspend is set, pm_runtime_get_sync may be
514 pm_runtime_idle called depending on if power.autosuspend_delay is
515 changed to or from a negative value; if power.use_autosuspend is clear,
519 - calculate the time when the current autosuspend delay period will expire,
520 based on power.last_busy and power.autosuspend_delay; if the delay time
522 nearest second; returns 0 if the delay period has already expired or
523 power.use_autosuspend isn't set, otherwise returns the expiration time
528 - pm_request_idle()
529 - pm_request_autosuspend()
530 - pm_schedule_suspend()
531 - pm_request_resume()
532 - pm_runtime_get_noresume()
533 - pm_runtime_get()
534 - pm_runtime_put_noidle()
535 - pm_runtime_put()
536 - pm_runtime_put_autosuspend()
537 - pm_runtime_enable()
538 - pm_suspend_ignore_children()
539 - pm_runtime_set_active()
540 - pm_runtime_set_suspended()
541 - pm_runtime_suspended()
542 - pm_runtime_mark_last_busy()
543 - pm_runtime_autosuspend_expiration()
548 - pm_runtime_idle()
549 - pm_runtime_suspend()
550 - pm_runtime_autosuspend()
551 - pm_runtime_resume()
552 - pm_runtime_get_sync()
553 - pm_runtime_put_sync()
554 - pm_runtime_put_sync_suspend()
555 - pm_runtime_put_sync_autosuspend()
562 -EAGAIN until pm_runtime_enable() is called for the device.
572 the parent's 'power.ignore_children' flag is set. Namely, in that case the
584 ->probe() callback will likely need to wake it up using one of the PM core's
595 It may be desirable to suspend the device once ->probe() has finished.
597 request to execute the subsystem-level idle callback for the device at that
599 update the last busy mark before returning from ->probe().
610 calling pm_runtime_suspend() from their ->remove() routines, the driver core
613 drivers to make their ->remove() callbacks avoid races with runtime PM directly,
617 Drivers in ->remove() callback should undo the runtime PM changes done
618 in ->probe(). Usually this means calling pm_runtime_disable(),
621 The user space can effectively disallow the driver of the device to power manage
622 it at run time by changing the value of its /sys/devices/.../power/control
623 attribute to "on", which causes pm_runtime_forbid() to be called. In principle,
625 runtime power management of the device until the user space turns it on.
629 value of /sys/devices/.../power/control to "auto" to allow the driver to power
637 as suspend-to-RAM and suspend-to-disk) interact with each other in a couple of
641 The device may have different wake-up settings for runtime PM and system sleep.
642 For example, remote wake-up may be enabled for runtime suspend but disallowed
644 the subsystem-level system suspend callback is responsible for changing the
645 device's wake-up setting (it may leave that to the device driver's system
647 in order to do so. The same is true if the driver uses different power levels
651 power, even if they had been suspended before the system suspend began. There
654 * The device might need to switch power levels, wake-up settings, etc.
656 * Remote wake-up events might have been lost by the firmware.
658 * The device's children may need the device to be at full power in order
670 brought back to full power during resume, then its runtime PM status will have
671 to be updated to reflect the actual post-system sleep status. The way to do
674 - pm_runtime_disable(dev);
675 - pm_runtime_set_active(dev);
676 - pm_runtime_enable(dev);
679 ->suspend() callback and decrements it after calling the ->resume() callback.
682 following the return of the ->resume() callback, the ->runtime_idle() callback
685 On some systems, however, system sleep is not entered through a global firmware
686 or hardware operation. Instead, all hardware components are put into low-power
700 that the device appears to be runtime-suspended and its state is fine, so it
706 related to hibernation (see Documentation/driver-api/pm/devices.rst for more
714 right before executing the subsystem-level .prepare() callback for it and
716 subsystem-level .suspend() callback for it. In addition to that the PM core
718 device right before executing the subsystem-level .suspend_late() callback
722 every device right after executing the subsystem-level .resume_early()
723 callback and right after executing the subsystem-level .complete() callback
728 Subsystems may wish to conserve code space by using the set of generic power
730 driver/base/power/generic_ops.c:
733 - invoke the ->runtime_suspend() callback provided by the driver of this
737 - invoke the ->runtime_resume() callback provided by the driver of this
741 - if the device has not been suspended at run time, invoke the ->suspend()
746 - if pm_runtime_suspended(dev) returns "false", invoke the ->suspend_noirq()
751 - invoke the ->resume() callback provided by the driver of this device and,
755 - invoke the ->resume_noirq() callback provided by the driver of this device
758 - if the device has not been suspended at run time, invoke the ->freeze()
763 - if pm_runtime_suspended(dev) returns "false", invoke the ->freeze_noirq()
768 - if the device has not been suspended at run time, invoke the ->thaw()
773 - if pm_runtime_suspended(dev) returns "false", invoke the ->thaw_noirq()
778 - if the device has not been suspended at run time, invoke the ->poweroff()
783 - if pm_runtime_suspended(dev) returns "false", run the ->poweroff_noirq()
788 - invoke the ->restore() callback provided by the driver of this device and,
792 - invoke the ->restore_noirq() callback provided by the device's driver
795 provide its own callbacks for ->runtime_idle(), ->runtime_suspend(),
796 ->runtime_resume(), ->suspend(), ->suspend_noirq(), ->resume(),
797 ->resume_noirq(), ->freeze(), ->freeze_noirq(), ->thaw(), ->thaw_noirq(),
798 ->poweroff(), ->poweroff_noirq(), ->restore(), ->restore_noirq() in the
799 subsystem-level dev_pm_ops structure.
807 8. "No-Callback" Devices
810 Some "devices" are only logical sub-devices of their parent and cannot be
811 power-managed on their own. (The prototype example is a USB interface. Entire
812 USB devices can go into low-power mode or send wake-up requests, but neither is
814 need of runtime PM callbacks; if the callbacks did exist, ->runtime_suspend()
815 and ->runtime_resume() would always return 0 without doing anything else and
816 ->runtime_idle() would always call pm_runtime_suspend().
821 also okay). The routine will set the device's power.no_callbacks flag and
822 prevent the non-debugging runtime PM sysfs attributes from being created.
824 When power.no_callbacks is set, the PM core will not invoke the
825 ->runtime_idle(), ->runtime_suspend(), or ->runtime_resume() callbacks.
830 or driver about runtime power changes. Instead, the driver for the device's
832 parent's power state changes.
837 domain could get attached to the device or that the device is power managed
843 9. Autosuspend, or automatically-delayed suspends
846 Changing a device's power state isn't free; it requires both time and energy.
847 A device should be put in a low-power state only when there's some reason to
852 the heuristic ends up being non-optimal, it will still prevent devices from
853 "bouncing" too rapidly between low-power and full-power states.
860 Inactivity is determined based on the power.last_busy field. Drivers should
866 /sys/devices/.../power/autosuspend_delay_ms attribute.
871 instead of the non-autosuspend counterparts::
878 Drivers may also continue to use the non-autosuspend helper functions; they
879 will behave normally, which means sometimes taking the autosuspend delay into
884 autosuspend delay time has expired. If the ->runtime_suspend() callback
885 returns -EAGAIN or -EBUSY, and if the next autosuspend delay expiration time is
888 autosuspend. The ->runtime_suspend() callback can't do this rescheduling
894 synchronize ->runtime_suspend() callbacks with the arrival of I/O requests.
896 Here is a schematic pseudo-code example::
900 lock(&foo->private_lock);
902 if (foo->num_pending_requests++ == 0)
903 pm_runtime_get(&foo->dev);
904 if (!foo->is_suspended)
906 unlock(&foo->private_lock);
911 lock(&foo->private_lock);
912 if (--foo->num_pending_requests == 0) {
913 pm_runtime_mark_last_busy(&foo->dev);
914 pm_runtime_put_autosuspend(&foo->dev);
918 unlock(&foo->private_lock);
927 lock(&foo->private_lock);
928 if (foo->num_pending_requests > 0) {
929 ret = -EBUSY;
932 foo->is_suspended = 1;
934 unlock(&foo->private_lock);
942 lock(&foo->private_lock);
944 foo->is_suspended = 0;
945 pm_runtime_mark_last_busy(&foo->dev);
946 if (foo->num_pending_requests > 0)
948 unlock(&foo->private_lock);
958 In addition, the power.autosuspend_delay field can be changed by user space at
960 pm_runtime_autosuspend_expiration() from within the ->runtime_suspend()
962 value then the delay has not yet expired and the callback should return
963 -EAGAIN.