About Kernel Documentation Linux Kernel Contact Linux Resources Linux Blog

Documentation / power / runtime_pm.txt




Custom Search

Based on kernel version 3.15.4. Page generated on 2014-07-07 09:04 EST.

1	Runtime Power Management Framework for I/O Devices
2	
3	(C) 2009-2011 Rafael J. Wysocki <rjw@sisk.pl>, Novell Inc.
4	(C) 2010 Alan Stern <stern@rowland.harvard.edu>
5	
6	1. Introduction
7	
8	Support for runtime power management (runtime PM) of I/O devices is provided
9	at the power management core (PM core) level by means of:
10	
11	* The power management workqueue pm_wq in which bus types and device drivers can
12	  put their PM-related work items.  It is strongly recommended that pm_wq be
13	  used for queuing all work items related to runtime PM, because this allows
14	  them to be synchronized with system-wide power transitions (suspend to RAM,
15	  hibernation and resume from system sleep states).  pm_wq is declared in
16	  include/linux/pm_runtime.h and defined in kernel/power/main.c.
17	
18	* A number of runtime PM fields in the 'power' member of 'struct device' (which
19	  is of the type 'struct dev_pm_info', defined in include/linux/pm.h) that can
20	  be used for synchronizing runtime PM operations with one another.
21	
22	* Three device runtime PM callbacks in 'struct dev_pm_ops' (defined in
23	  include/linux/pm.h).
24	
25	* A set of helper functions defined in drivers/base/power/runtime.c that can be
26	  used for carrying out runtime PM operations in such a way that the
27	  synchronization between them is taken care of by the PM core.  Bus types and
28	  device drivers are encouraged to use these functions.
29	
30	The runtime PM callbacks present in 'struct dev_pm_ops', the device runtime PM
31	fields of 'struct dev_pm_info' and the core helper functions provided for
32	runtime PM are described below.
33	
34	2. Device Runtime PM Callbacks
35	
36	There are three device runtime PM callbacks defined in 'struct dev_pm_ops':
37	
38	struct dev_pm_ops {
39		...
40		int (*runtime_suspend)(struct device *dev);
41		int (*runtime_resume)(struct device *dev);
42		int (*runtime_idle)(struct device *dev);
43		...
44	};
45	
46	The ->runtime_suspend(), ->runtime_resume() and ->runtime_idle() callbacks
47	are executed by the PM core for the device's subsystem that may be either of
48	the following:
49	
50	  1. PM domain of the device, if the device's PM domain object, dev->pm_domain,
51	     is present.
52	
53	  2. Device type of the device, if both dev->type and dev->type->pm are present.
54	
55	  3. Device class of the device, if both dev->class and dev->class->pm are
56	     present.
57	
58	  4. Bus type of the device, if both dev->bus and dev->bus->pm are present.
59	
60	If the subsystem chosen by applying the above rules doesn't provide the relevant
61	callback, the PM core will invoke the corresponding driver callback stored in
62	dev->driver->pm directly (if present).
63	
64	The PM core always checks which callback to use in the order given above, so the
65	priority order of callbacks from high to low is: PM domain, device type, class
66	and bus type.  Moreover, the high-priority one will always take precedence over
67	a low-priority one.  The PM domain, bus type, device type and class callbacks
68	are referred to as subsystem-level callbacks in what follows.
69	
70	By default, the callbacks are always invoked in process context with interrupts
71	enabled.  However, the pm_runtime_irq_safe() helper function can be used to tell
72	the PM core that it is safe to run the ->runtime_suspend(), ->runtime_resume()
73	and ->runtime_idle() callbacks for the given device in atomic context with
74	interrupts disabled.  This implies that the callback routines in question must
75	not block or sleep, but it also means that the synchronous helper functions
76	listed at the end of Section 4 may be used for that device within an interrupt
77	handler or generally in an atomic context.
78	
79	The subsystem-level suspend callback, if present, is _entirely_ _responsible_
80	for handling the suspend of the device as appropriate, which may, but need not
81	include executing the device driver's own ->runtime_suspend() callback (from the
82	PM core's point of view it is not necessary to implement a ->runtime_suspend()
83	callback in a device driver as long as the subsystem-level suspend callback
84	knows what to do to handle the device).
85	
86	  * Once the subsystem-level suspend callback (or the driver suspend callback,
87	    if invoked directly) has completed successfully for the given device, the PM
88	    core regards the device as suspended, which need not mean that it has been
89	    put into a low power state.  It is supposed to mean, however, that the
90	    device will not process data and will not communicate with the CPU(s) and
91	    RAM until the appropriate resume callback is executed for it.  The runtime
92	    PM status of a device after successful execution of the suspend callback is
93	    'suspended'.
94	
95	  * If the suspend callback returns -EBUSY or -EAGAIN, the device's runtime PM
96	    status remains 'active', which means that the device _must_ be fully
97	    operational afterwards.
98	
99	  * If the suspend callback returns an error code different from -EBUSY and
100	    -EAGAIN, the PM core regards this as a fatal error and will refuse to run
101	    the helper functions described in Section 4 for the device until its status
102	    is directly set to  either'active', or 'suspended' (the PM core provides
103	    special helper functions for this purpose).
104	
105	In particular, if the driver requires remote wakeup capability (i.e. hardware
106	mechanism allowing the device to request a change of its power state, such as
107	PCI PME) for proper functioning and device_run_wake() returns 'false' for the
108	device, then ->runtime_suspend() should return -EBUSY.  On the other hand, if
109	device_run_wake() returns 'true' for the device and the device is put into a
110	low-power state during the execution of the suspend callback, it is expected
111	that remote wakeup will be enabled for the device.  Generally, remote wakeup
112	should be enabled for all input devices put into low-power states at run time.
113	
114	The subsystem-level resume callback, if present, is _entirely_ _responsible_ for
115	handling the resume of the device as appropriate, which may, but need not
116	include executing the device driver's own ->runtime_resume() callback (from the
117	PM core's point of view it is not necessary to implement a ->runtime_resume()
118	callback in a device driver as long as the subsystem-level resume callback knows
119	what to do to handle the device).
120	
121	  * Once the subsystem-level resume callback (or the driver resume callback, if
122	    invoked directly) has completed successfully, the PM core regards the device
123	    as fully operational, which means that the device _must_ be able to complete
124	    I/O operations as needed.  The runtime PM status of the device is then
125	    'active'.
126	
127	  * If the resume callback returns an error code, the PM core regards this as a
128	    fatal error and will refuse to run the helper functions described in Section
129	    4 for the device, until its status is directly set to either 'active', or
130	    'suspended' (by means of special helper functions provided by the PM core
131	    for this purpose).
132	
133	The idle callback (a subsystem-level one, if present, or the driver one) is
134	executed by the PM core whenever the device appears to be idle, which is
135	indicated to the PM core by two counters, the device's usage counter and the
136	counter of 'active' children of the device.
137	
138	  * If any of these counters is decreased using a helper function provided by
139	    the PM core and it turns out to be equal to zero, the other counter is
140	    checked.  If that counter also is equal to zero, the PM core executes the
141	    idle callback with the device as its argument.
142	
143	The action performed by the idle callback is totally dependent on the subsystem
144	(or driver) in question, but the expected and recommended action is to check
145	if the device can be suspended (i.e. if all of the conditions necessary for
146	suspending the device are satisfied) and to queue up a suspend request for the
147	device in that case.  If there is no idle callback, or if the callback returns
148	0, then the PM core will attempt to carry out a runtime suspend of the device,
149	also respecting devices configured for autosuspend.  In essence this means a
150	call to pm_runtime_autosuspend() (do note that drivers needs to update the
151	device last busy mark, pm_runtime_mark_last_busy(), to control the delay under
152	this circumstance).  To prevent this (for example, if the callback routine has
153	started a delayed suspend), the routine must return a non-zero value.  Negative
154	error return codes are ignored by the PM core.
155	
156	The helper functions provided by the PM core, described in Section 4, guarantee
157	that the following constraints are met with respect to runtime PM callbacks for
158	one device:
159	
160	(1) The callbacks are mutually exclusive (e.g. it is forbidden to execute
161	    ->runtime_suspend() in parallel with ->runtime_resume() or with another
162	    instance of ->runtime_suspend() for the same device) with the exception that
163	    ->runtime_suspend() or ->runtime_resume() can be executed in parallel with
164	    ->runtime_idle() (although ->runtime_idle() will not be started while any
165	    of the other callbacks is being executed for the same device).
166	
167	(2) ->runtime_idle() and ->runtime_suspend() can only be executed for 'active'
168	    devices (i.e. the PM core will only execute ->runtime_idle() or
169	    ->runtime_suspend() for the devices the runtime PM status of which is
170	    'active').
171	
172	(3) ->runtime_idle() and ->runtime_suspend() can only be executed for a device
173	    the usage counter of which is equal to zero _and_ either the counter of
174	    'active' children of which is equal to zero, or the 'power.ignore_children'
175	    flag of which is set.
176	
177	(4) ->runtime_resume() can only be executed for 'suspended' devices  (i.e. the
178	    PM core will only execute ->runtime_resume() for the devices the runtime
179	    PM status of which is 'suspended').
180	
181	Additionally, the helper functions provided by the PM core obey the following
182	rules:
183	
184	  * If ->runtime_suspend() is about to be executed or there's a pending request
185	    to execute it, ->runtime_idle() will not be executed for the same device.
186	
187	  * A request to execute or to schedule the execution of ->runtime_suspend()
188	    will cancel any pending requests to execute ->runtime_idle() for the same
189	    device.
190	
191	  * If ->runtime_resume() is about to be executed or there's a pending request
192	    to execute it, the other callbacks will not be executed for the same device.
193	
194	  * A request to execute ->runtime_resume() will cancel any pending or
195	    scheduled requests to execute the other callbacks for the same device,
196	    except for scheduled autosuspends.
197	
198	3. Runtime PM Device Fields
199	
200	The following device runtime PM fields are present in 'struct dev_pm_info', as
201	defined in include/linux/pm.h:
202	
203	  struct timer_list suspend_timer;
204	    - timer used for scheduling (delayed) suspend and autosuspend requests
205	
206	  unsigned long timer_expires;
207	    - timer expiration time, in jiffies (if this is different from zero, the
208	      timer is running and will expire at that time, otherwise the timer is not
209	      running)
210	
211	  struct work_struct work;
212	    - work structure used for queuing up requests (i.e. work items in pm_wq)
213	
214	  wait_queue_head_t wait_queue;
215	    - wait queue used if any of the helper functions needs to wait for another
216	      one to complete
217	
218	  spinlock_t lock;
219	    - lock used for synchronisation
220	
221	  atomic_t usage_count;
222	    - the usage counter of the device
223	
224	  atomic_t child_count;
225	    - the count of 'active' children of the device
226	
227	  unsigned int ignore_children;
228	    - if set, the value of child_count is ignored (but still updated)
229	
230	  unsigned int disable_depth;
231	    - used for disabling the helper funcions (they work normally if this is
232	      equal to zero); the initial value of it is 1 (i.e. runtime PM is
233	      initially disabled for all devices)
234	
235	  int runtime_error;
236	    - if set, there was a fatal error (one of the callbacks returned error code
237	      as described in Section 2), so the helper funtions will not work until
238	      this flag is cleared; this is the error code returned by the failing
239	      callback
240	
241	  unsigned int idle_notification;
242	    - if set, ->runtime_idle() is being executed
243	
244	  unsigned int request_pending;
245	    - if set, there's a pending request (i.e. a work item queued up into pm_wq)
246	
247	  enum rpm_request request;
248	    - type of request that's pending (valid if request_pending is set)
249	
250	  unsigned int deferred_resume;
251	    - set if ->runtime_resume() is about to be run while ->runtime_suspend() is
252	      being executed for that device and it is not practical to wait for the
253	      suspend to complete; means "start a resume as soon as you've suspended"
254	
255	  unsigned int run_wake;
256	    - set if the device is capable of generating runtime wake-up events
257	
258	  enum rpm_status runtime_status;
259	    - the runtime PM status of the device; this field's initial value is
260	      RPM_SUSPENDED, which means that each device is initially regarded by the
261	      PM core as 'suspended', regardless of its real hardware status
262	
263	  unsigned int runtime_auto;
264	    - if set, indicates that the user space has allowed the device driver to
265	      power manage the device at run time via the /sys/devices/.../power/control
266	      interface; it may only be modified with the help of the pm_runtime_allow()
267	      and pm_runtime_forbid() helper functions
268	
269	  unsigned int no_callbacks;
270	    - indicates that the device does not use the runtime PM callbacks (see
271	      Section 8); it may be modified only by the pm_runtime_no_callbacks()
272	      helper function
273	
274	  unsigned int irq_safe;
275	    - indicates that the ->runtime_suspend() and ->runtime_resume() callbacks
276	      will be invoked with the spinlock held and interrupts disabled
277	
278	  unsigned int use_autosuspend;
279	    - indicates that the device's driver supports delayed autosuspend (see
280	      Section 9); it may be modified only by the
281	      pm_runtime{_dont}_use_autosuspend() helper functions
282	
283	  unsigned int timer_autosuspends;
284	    - indicates that the PM core should attempt to carry out an autosuspend
285	      when the timer expires rather than a normal suspend
286	
287	  int autosuspend_delay;
288	    - the delay time (in milliseconds) to be used for autosuspend
289	
290	  unsigned long last_busy;
291	    - the time (in jiffies) when the pm_runtime_mark_last_busy() helper
292	      function was last called for this device; used in calculating inactivity
293	      periods for autosuspend
294	
295	All of the above fields are members of the 'power' member of 'struct device'.
296	
297	4. Runtime PM Device Helper Functions
298	
299	The following runtime PM helper functions are defined in
300	drivers/base/power/runtime.c and include/linux/pm_runtime.h:
301	
302	  void pm_runtime_init(struct device *dev);
303	    - initialize the device runtime PM fields in 'struct dev_pm_info'
304	
305	  void pm_runtime_remove(struct device *dev);
306	    - make sure that the runtime PM of the device will be disabled after
307	      removing the device from device hierarchy
308	
309	  int pm_runtime_idle(struct device *dev);
310	    - execute the subsystem-level idle callback for the device; returns an
311	      error code on failure, where -EINPROGRESS means that ->runtime_idle() is
312	      already being executed; if there is no callback or the callback returns 0
313	      then run pm_runtime_autosuspend(dev) and return its result
314	
315	  int pm_runtime_suspend(struct device *dev);
316	    - execute the subsystem-level suspend callback for the device; returns 0 on
317	      success, 1 if the device's runtime PM status was already 'suspended', or
318	      error code on failure, where -EAGAIN or -EBUSY means it is safe to attempt
319	      to suspend the device again in future and -EACCES means that
320	      'power.disable_depth' is different from 0
321	
322	  int pm_runtime_autosuspend(struct device *dev);
323	    - same as pm_runtime_suspend() except that the autosuspend delay is taken
324	      into account; if pm_runtime_autosuspend_expiration() says the delay has
325	      not yet expired then an autosuspend is scheduled for the appropriate time
326	      and 0 is returned
327	
328	  int pm_runtime_resume(struct device *dev);
329	    - execute the subsystem-level resume callback for the device; returns 0 on
330	      success, 1 if the device's runtime PM status was already 'active' or
331	      error code on failure, where -EAGAIN means it may be safe to attempt to
332	      resume the device again in future, but 'power.runtime_error' should be
333	      checked additionally, and -EACCES means that 'power.disable_depth' is
334	      different from 0
335	
336	  int pm_request_idle(struct device *dev);
337	    - submit a request to execute the subsystem-level idle callback for the
338	      device (the request is represented by a work item in pm_wq); returns 0 on
339	      success or error code if the request has not been queued up
340	
341	  int pm_request_autosuspend(struct device *dev);
342	    - schedule the execution of the subsystem-level suspend callback for the
343	      device when the autosuspend delay has expired; if the delay has already
344	      expired then the work item is queued up immediately
345	
346	  int pm_schedule_suspend(struct device *dev, unsigned int delay);
347	    - schedule the execution of the subsystem-level suspend callback for the
348	      device in future, where 'delay' is the time to wait before queuing up a
349	      suspend work item in pm_wq, in milliseconds (if 'delay' is zero, the work
350	      item is queued up immediately); returns 0 on success, 1 if the device's PM
351	      runtime status was already 'suspended', or error code if the request
352	      hasn't been scheduled (or queued up if 'delay' is 0); if the execution of
353	      ->runtime_suspend() is already scheduled and not yet expired, the new
354	      value of 'delay' will be used as the time to wait
355	
356	  int pm_request_resume(struct device *dev);
357	    - submit a request to execute the subsystem-level resume callback for the
358	      device (the request is represented by a work item in pm_wq); returns 0 on
359	      success, 1 if the device's runtime PM status was already 'active', or
360	      error code if the request hasn't been queued up
361	
362	  void pm_runtime_get_noresume(struct device *dev);
363	    - increment the device's usage counter
364	
365	  int pm_runtime_get(struct device *dev);
366	    - increment the device's usage counter, run pm_request_resume(dev) and
367	      return its result
368	
369	  int pm_runtime_get_sync(struct device *dev);
370	    - increment the device's usage counter, run pm_runtime_resume(dev) and
371	      return its result
372	
373	  void pm_runtime_put_noidle(struct device *dev);
374	    - decrement the device's usage counter
375	
376	  int pm_runtime_put(struct device *dev);
377	    - decrement the device's usage counter; if the result is 0 then run
378	      pm_request_idle(dev) and return its result
379	
380	  int pm_runtime_put_autosuspend(struct device *dev);
381	    - decrement the device's usage counter; if the result is 0 then run
382	      pm_request_autosuspend(dev) and return its result
383	
384	  int pm_runtime_put_sync(struct device *dev);
385	    - decrement the device's usage counter; if the result is 0 then run
386	      pm_runtime_idle(dev) and return its result
387	
388	  int pm_runtime_put_sync_suspend(struct device *dev);
389	    - decrement the device's usage counter; if the result is 0 then run
390	      pm_runtime_suspend(dev) and return its result
391	
392	  int pm_runtime_put_sync_autosuspend(struct device *dev);
393	    - decrement the device's usage counter; if the result is 0 then run
394	      pm_runtime_autosuspend(dev) and return its result
395	
396	  void pm_runtime_enable(struct device *dev);
397	    - decrement the device's 'power.disable_depth' field; if that field is equal
398	      to zero, the runtime PM helper functions can execute subsystem-level
399	      callbacks described in Section 2 for the device
400	
401	  int pm_runtime_disable(struct device *dev);
402	    - increment the device's 'power.disable_depth' field (if the value of that
403	      field was previously zero, this prevents subsystem-level runtime PM
404	      callbacks from being run for the device), make sure that all of the
405	      pending runtime PM operations on the device are either completed or
406	      canceled; returns 1 if there was a resume request pending and it was
407	      necessary to execute the subsystem-level resume callback for the device
408	      to satisfy that request, otherwise 0 is returned
409	
410	  int pm_runtime_barrier(struct device *dev);
411	    - check if there's a resume request pending for the device and resume it
412	      (synchronously) in that case, cancel any other pending runtime PM requests
413	      regarding it and wait for all runtime PM operations on it in progress to
414	      complete; returns 1 if there was a resume request pending and it was
415	      necessary to execute the subsystem-level resume callback for the device to
416	      satisfy that request, otherwise 0 is returned
417	
418	  void pm_suspend_ignore_children(struct device *dev, bool enable);
419	    - set/unset the power.ignore_children flag of the device
420	
421	  int pm_runtime_set_active(struct device *dev);
422	    - clear the device's 'power.runtime_error' flag, set the device's runtime
423	      PM status to 'active' and update its parent's counter of 'active'
424	      children as appropriate (it is only valid to use this function if
425	      'power.runtime_error' is set or 'power.disable_depth' is greater than
426	      zero); it will fail and return error code if the device has a parent
427	      which is not active and the 'power.ignore_children' flag of which is unset
428	
429	  void pm_runtime_set_suspended(struct device *dev);
430	    - clear the device's 'power.runtime_error' flag, set the device's runtime
431	      PM status to 'suspended' and update its parent's counter of 'active'
432	      children as appropriate (it is only valid to use this function if
433	      'power.runtime_error' is set or 'power.disable_depth' is greater than
434	      zero)
435	
436	  bool pm_runtime_active(struct device *dev);
437	    - return true if the device's runtime PM status is 'active' or its
438	      'power.disable_depth' field is not equal to zero, or false otherwise
439	
440	  bool pm_runtime_suspended(struct device *dev);
441	    - return true if the device's runtime PM status is 'suspended' and its
442	      'power.disable_depth' field is equal to zero, or false otherwise
443	
444	  bool pm_runtime_status_suspended(struct device *dev);
445	    - return true if the device's runtime PM status is 'suspended'
446	
447	  void pm_runtime_allow(struct device *dev);
448	    - set the power.runtime_auto flag for the device and decrease its usage
449	      counter (used by the /sys/devices/.../power/control interface to
450	      effectively allow the device to be power managed at run time)
451	
452	  void pm_runtime_forbid(struct device *dev);
453	    - unset the power.runtime_auto flag for the device and increase its usage
454	      counter (used by the /sys/devices/.../power/control interface to
455	      effectively prevent the device from being power managed at run time)
456	
457	  void pm_runtime_no_callbacks(struct device *dev);
458	    - set the power.no_callbacks flag for the device and remove the runtime
459	      PM attributes from /sys/devices/.../power (or prevent them from being
460	      added when the device is registered)
461	
462	  void pm_runtime_irq_safe(struct device *dev);
463	    - set the power.irq_safe flag for the device, causing the runtime-PM
464	      callbacks to be invoked with interrupts off
465	
466	  void pm_runtime_mark_last_busy(struct device *dev);
467	    - set the power.last_busy field to the current time
468	
469	  void pm_runtime_use_autosuspend(struct device *dev);
470	    - set the power.use_autosuspend flag, enabling autosuspend delays
471	
472	  void pm_runtime_dont_use_autosuspend(struct device *dev);
473	    - clear the power.use_autosuspend flag, disabling autosuspend delays
474	
475	  void pm_runtime_set_autosuspend_delay(struct device *dev, int delay);
476	    - set the power.autosuspend_delay value to 'delay' (expressed in
477	      milliseconds); if 'delay' is negative then runtime suspends are
478	      prevented
479	
480	  unsigned long pm_runtime_autosuspend_expiration(struct device *dev);
481	    - calculate the time when the current autosuspend delay period will expire,
482	      based on power.last_busy and power.autosuspend_delay; if the delay time
483	      is 1000 ms or larger then the expiration time is rounded up to the
484	      nearest second; returns 0 if the delay period has already expired or
485	      power.use_autosuspend isn't set, otherwise returns the expiration time
486	      in jiffies
487	
488	It is safe to execute the following helper functions from interrupt context:
489	
490	pm_request_idle()
491	pm_request_autosuspend()
492	pm_schedule_suspend()
493	pm_request_resume()
494	pm_runtime_get_noresume()
495	pm_runtime_get()
496	pm_runtime_put_noidle()
497	pm_runtime_put()
498	pm_runtime_put_autosuspend()
499	pm_runtime_enable()
500	pm_suspend_ignore_children()
501	pm_runtime_set_active()
502	pm_runtime_set_suspended()
503	pm_runtime_suspended()
504	pm_runtime_mark_last_busy()
505	pm_runtime_autosuspend_expiration()
506	
507	If pm_runtime_irq_safe() has been called for a device then the following helper
508	functions may also be used in interrupt context:
509	
510	pm_runtime_idle()
511	pm_runtime_suspend()
512	pm_runtime_autosuspend()
513	pm_runtime_resume()
514	pm_runtime_get_sync()
515	pm_runtime_put_sync()
516	pm_runtime_put_sync_suspend()
517	pm_runtime_put_sync_autosuspend()
518	
519	5. Runtime PM Initialization, Device Probing and Removal
520	
521	Initially, the runtime PM is disabled for all devices, which means that the
522	majority of the runtime PM helper funtions described in Section 4 will return
523	-EAGAIN until pm_runtime_enable() is called for the device.
524	
525	In addition to that, the initial runtime PM status of all devices is
526	'suspended', but it need not reflect the actual physical state of the device.
527	Thus, if the device is initially active (i.e. it is able to process I/O), its
528	runtime PM status must be changed to 'active', with the help of
529	pm_runtime_set_active(), before pm_runtime_enable() is called for the device.
530	
531	However, if the device has a parent and the parent's runtime PM is enabled,
532	calling pm_runtime_set_active() for the device will affect the parent, unless
533	the parent's 'power.ignore_children' flag is set.  Namely, in that case the
534	parent won't be able to suspend at run time, using the PM core's helper
535	functions, as long as the child's status is 'active', even if the child's
536	runtime PM is still disabled (i.e. pm_runtime_enable() hasn't been called for
537	the child yet or pm_runtime_disable() has been called for it).  For this reason,
538	once pm_runtime_set_active() has been called for the device, pm_runtime_enable()
539	should be called for it too as soon as reasonably possible or its runtime PM
540	status should be changed back to 'suspended' with the help of
541	pm_runtime_set_suspended().
542	
543	If the default initial runtime PM status of the device (i.e. 'suspended')
544	reflects the actual state of the device, its bus type's or its driver's
545	->probe() callback will likely need to wake it up using one of the PM core's
546	helper functions described in Section 4.  In that case, pm_runtime_resume()
547	should be used.  Of course, for this purpose the device's runtime PM has to be
548	enabled earlier by calling pm_runtime_enable().
549	
550	It may be desirable to suspend the device once ->probe() has finished.
551	Therefore the driver core uses the asyncronous pm_request_idle() to submit a
552	request to execute the subsystem-level idle callback for the device at that
553	time.  A driver that makes use of the runtime autosuspend feature, may want to
554	update the last busy mark before returning from ->probe().
555	
556	Moreover, the driver core prevents runtime PM callbacks from racing with the bus
557	notifier callback in __device_release_driver(), which is necessary, because the
558	notifier is used by some subsystems to carry out operations affecting the
559	runtime PM functionality.  It does so by calling pm_runtime_get_sync() before
560	driver_sysfs_remove() and the BUS_NOTIFY_UNBIND_DRIVER notifications.  This
561	resumes the device if it's in the suspended state and prevents it from
562	being suspended again while those routines are being executed.
563	
564	To allow bus types and drivers to put devices into the suspended state by
565	calling pm_runtime_suspend() from their ->remove() routines, the driver core
566	executes pm_runtime_put_sync() after running the BUS_NOTIFY_UNBIND_DRIVER
567	notifications in __device_release_driver().  This requires bus types and
568	drivers to make their ->remove() callbacks avoid races with runtime PM directly,
569	but also it allows of more flexibility in the handling of devices during the
570	removal of their drivers.
571	
572	The user space can effectively disallow the driver of the device to power manage
573	it at run time by changing the value of its /sys/devices/.../power/control
574	attribute to "on", which causes pm_runtime_forbid() to be called.  In principle,
575	this mechanism may also be used by the driver to effectively turn off the
576	runtime power management of the device until the user space turns it on.
577	Namely, during the initialization the driver can make sure that the runtime PM
578	status of the device is 'active' and call pm_runtime_forbid().  It should be
579	noted, however, that if the user space has already intentionally changed the
580	value of /sys/devices/.../power/control to "auto" to allow the driver to power
581	manage the device at run time, the driver may confuse it by using
582	pm_runtime_forbid() this way.
583	
584	6. Runtime PM and System Sleep
585	
586	Runtime PM and system sleep (i.e., system suspend and hibernation, also known
587	as suspend-to-RAM and suspend-to-disk) interact with each other in a couple of
588	ways.  If a device is active when a system sleep starts, everything is
589	straightforward.  But what should happen if the device is already suspended?
590	
591	The device may have different wake-up settings for runtime PM and system sleep.
592	For example, remote wake-up may be enabled for runtime suspend but disallowed
593	for system sleep (device_may_wakeup(dev) returns 'false').  When this happens,
594	the subsystem-level system suspend callback is responsible for changing the
595	device's wake-up setting (it may leave that to the device driver's system
596	suspend routine).  It may be necessary to resume the device and suspend it again
597	in order to do so.  The same is true if the driver uses different power levels
598	or other settings for runtime suspend and system sleep.
599	
600	During system resume, the simplest approach is to bring all devices back to full
601	power, even if they had been suspended before the system suspend began.  There
602	are several reasons for this, including:
603	
604	  * The device might need to switch power levels, wake-up settings, etc.
605	
606	  * Remote wake-up events might have been lost by the firmware.
607	
608	  * The device's children may need the device to be at full power in order
609	    to resume themselves.
610	
611	  * The driver's idea of the device state may not agree with the device's
612	    physical state.  This can happen during resume from hibernation.
613	
614	  * The device might need to be reset.
615	
616	  * Even though the device was suspended, if its usage counter was > 0 then most
617	    likely it would need a runtime resume in the near future anyway.
618	
619	If the device had been suspended before the system suspend began and it's
620	brought back to full power during resume, then its runtime PM status will have
621	to be updated to reflect the actual post-system sleep status.  The way to do
622	this is:
623	
624		pm_runtime_disable(dev);
625		pm_runtime_set_active(dev);
626		pm_runtime_enable(dev);
627	
628	The PM core always increments the runtime usage counter before calling the
629	->suspend() callback and decrements it after calling the ->resume() callback.
630	Hence disabling runtime PM temporarily like this will not cause any runtime
631	suspend attempts to be permanently lost.  If the usage count goes to zero
632	following the return of the ->resume() callback, the ->runtime_idle() callback
633	will be invoked as usual.
634	
635	On some systems, however, system sleep is not entered through a global firmware
636	or hardware operation.  Instead, all hardware components are put into low-power
637	states directly by the kernel in a coordinated way.  Then, the system sleep
638	state effectively follows from the states the hardware components end up in
639	and the system is woken up from that state by a hardware interrupt or a similar
640	mechanism entirely under the kernel's control.  As a result, the kernel never
641	gives control away and the states of all devices during resume are precisely
642	known to it.  If that is the case and none of the situations listed above takes
643	place (in particular, if the system is not waking up from hibernation), it may
644	be more efficient to leave the devices that had been suspended before the system
645	suspend began in the suspended state.
646	
647	The PM core does its best to reduce the probability of race conditions between
648	the runtime PM and system suspend/resume (and hibernation) callbacks by carrying
649	out the following operations:
650	
651	  * During system suspend it calls pm_runtime_get_noresume() and
652	    pm_runtime_barrier() for every device right before executing the
653	    subsystem-level .suspend() callback for it.  In addition to that it calls
654	    __pm_runtime_disable() with 'false' as the second argument for every device
655	    right before executing the subsystem-level .suspend_late() callback for it.
656	
657	  * During system resume it calls pm_runtime_enable() and pm_runtime_put()
658	    for every device right after executing the subsystem-level .resume_early()
659	    callback and right after executing the subsystem-level .resume() callback
660	    for it, respectively.
661	
662	7. Generic subsystem callbacks
663	
664	Subsystems may wish to conserve code space by using the set of generic power
665	management callbacks provided by the PM core, defined in
666	driver/base/power/generic_ops.c:
667	
668	  int pm_generic_runtime_suspend(struct device *dev);
669	    - invoke the ->runtime_suspend() callback provided by the driver of this
670	      device and return its result, or return 0 if not defined
671	
672	  int pm_generic_runtime_resume(struct device *dev);
673	    - invoke the ->runtime_resume() callback provided by the driver of this
674	      device and return its result, or return 0 if not defined
675	
676	  int pm_generic_suspend(struct device *dev);
677	    - if the device has not been suspended at run time, invoke the ->suspend()
678	      callback provided by its driver and return its result, or return 0 if not
679	      defined
680	
681	  int pm_generic_suspend_noirq(struct device *dev);
682	    - if pm_runtime_suspended(dev) returns "false", invoke the ->suspend_noirq()
683	      callback provided by the device's driver and return its result, or return
684	      0 if not defined
685	
686	  int pm_generic_resume(struct device *dev);
687	    - invoke the ->resume() callback provided by the driver of this device and,
688	      if successful, change the device's runtime PM status to 'active'
689	
690	  int pm_generic_resume_noirq(struct device *dev);
691	    - invoke the ->resume_noirq() callback provided by the driver of this device
692	
693	  int pm_generic_freeze(struct device *dev);
694	    - if the device has not been suspended at run time, invoke the ->freeze()
695	      callback provided by its driver and return its result, or return 0 if not
696	      defined
697	
698	  int pm_generic_freeze_noirq(struct device *dev);
699	    - if pm_runtime_suspended(dev) returns "false", invoke the ->freeze_noirq()
700	      callback provided by the device's driver and return its result, or return
701	      0 if not defined
702	
703	  int pm_generic_thaw(struct device *dev);
704	    - if the device has not been suspended at run time, invoke the ->thaw()
705	      callback provided by its driver and return its result, or return 0 if not
706	      defined
707	
708	  int pm_generic_thaw_noirq(struct device *dev);
709	    - if pm_runtime_suspended(dev) returns "false", invoke the ->thaw_noirq()
710	      callback provided by the device's driver and return its result, or return
711	      0 if not defined
712	
713	  int pm_generic_poweroff(struct device *dev);
714	    - if the device has not been suspended at run time, invoke the ->poweroff()
715	      callback provided by its driver and return its result, or return 0 if not
716	      defined
717	
718	  int pm_generic_poweroff_noirq(struct device *dev);
719	    - if pm_runtime_suspended(dev) returns "false", run the ->poweroff_noirq()
720	      callback provided by the device's driver and return its result, or return
721	      0 if not defined
722	
723	  int pm_generic_restore(struct device *dev);
724	    - invoke the ->restore() callback provided by the driver of this device and,
725	      if successful, change the device's runtime PM status to 'active'
726	
727	  int pm_generic_restore_noirq(struct device *dev);
728	    - invoke the ->restore_noirq() callback provided by the device's driver
729	
730	These functions are the defaults used by the PM core, if a subsystem doesn't
731	provide its own callbacks for ->runtime_idle(), ->runtime_suspend(),
732	->runtime_resume(), ->suspend(), ->suspend_noirq(), ->resume(),
733	->resume_noirq(), ->freeze(), ->freeze_noirq(), ->thaw(), ->thaw_noirq(),
734	->poweroff(), ->poweroff_noirq(), ->restore(), ->restore_noirq() in the
735	subsystem-level dev_pm_ops structure.
736	
737	Device drivers that wish to use the same function as a system suspend, freeze,
738	poweroff and runtime suspend callback, and similarly for system resume, thaw,
739	restore, and runtime resume, can achieve this with the help of the
740	UNIVERSAL_DEV_PM_OPS macro defined in include/linux/pm.h (possibly setting its
741	last argument to NULL).
742	
743	8. "No-Callback" Devices
744	
745	Some "devices" are only logical sub-devices of their parent and cannot be
746	power-managed on their own.  (The prototype example is a USB interface.  Entire
747	USB devices can go into low-power mode or send wake-up requests, but neither is
748	possible for individual interfaces.)  The drivers for these devices have no
749	need of runtime PM callbacks; if the callbacks did exist, ->runtime_suspend()
750	and ->runtime_resume() would always return 0 without doing anything else and
751	->runtime_idle() would always call pm_runtime_suspend().
752	
753	Subsystems can tell the PM core about these devices by calling
754	pm_runtime_no_callbacks().  This should be done after the device structure is
755	initialized and before it is registered (although after device registration is
756	also okay).  The routine will set the device's power.no_callbacks flag and
757	prevent the non-debugging runtime PM sysfs attributes from being created.
758	
759	When power.no_callbacks is set, the PM core will not invoke the
760	->runtime_idle(), ->runtime_suspend(), or ->runtime_resume() callbacks.
761	Instead it will assume that suspends and resumes always succeed and that idle
762	devices should be suspended.
763	
764	As a consequence, the PM core will never directly inform the device's subsystem
765	or driver about runtime power changes.  Instead, the driver for the device's
766	parent must take responsibility for telling the device's driver when the
767	parent's power state changes.
768	
769	9. Autosuspend, or automatically-delayed suspends
770	
771	Changing a device's power state isn't free; it requires both time and energy.
772	A device should be put in a low-power state only when there's some reason to
773	think it will remain in that state for a substantial time.  A common heuristic
774	says that a device which hasn't been used for a while is liable to remain
775	unused; following this advice, drivers should not allow devices to be suspended
776	at runtime until they have been inactive for some minimum period.  Even when
777	the heuristic ends up being non-optimal, it will still prevent devices from
778	"bouncing" too rapidly between low-power and full-power states.
779	
780	The term "autosuspend" is an historical remnant.  It doesn't mean that the
781	device is automatically suspended (the subsystem or driver still has to call
782	the appropriate PM routines); rather it means that runtime suspends will
783	automatically be delayed until the desired period of inactivity has elapsed.
784	
785	Inactivity is determined based on the power.last_busy field.  Drivers should
786	call pm_runtime_mark_last_busy() to update this field after carrying out I/O,
787	typically just before calling pm_runtime_put_autosuspend().  The desired length
788	of the inactivity period is a matter of policy.  Subsystems can set this length
789	initially by calling pm_runtime_set_autosuspend_delay(), but after device
790	registration the length should be controlled by user space, using the
791	/sys/devices/.../power/autosuspend_delay_ms attribute.
792	
793	In order to use autosuspend, subsystems or drivers must call
794	pm_runtime_use_autosuspend() (preferably before registering the device), and
795	thereafter they should use the various *_autosuspend() helper functions instead
796	of the non-autosuspend counterparts:
797	
798		Instead of: pm_runtime_suspend    use: pm_runtime_autosuspend;
799		Instead of: pm_schedule_suspend   use: pm_request_autosuspend;
800		Instead of: pm_runtime_put        use: pm_runtime_put_autosuspend;
801		Instead of: pm_runtime_put_sync   use: pm_runtime_put_sync_autosuspend.
802	
803	Drivers may also continue to use the non-autosuspend helper functions; they
804	will behave normally, not taking the autosuspend delay into account.
805	Similarly, if the power.use_autosuspend field isn't set then the autosuspend
806	helper functions will behave just like the non-autosuspend counterparts.
807	
808	Under some circumstances a driver or subsystem may want to prevent a device
809	from autosuspending immediately, even though the usage counter is zero and the
810	autosuspend delay time has expired.  If the ->runtime_suspend() callback
811	returns -EAGAIN or -EBUSY, and if the next autosuspend delay expiration time is
812	in the future (as it normally would be if the callback invoked
813	pm_runtime_mark_last_busy()), the PM core will automatically reschedule the
814	autosuspend.  The ->runtime_suspend() callback can't do this rescheduling
815	itself because no suspend requests of any kind are accepted while the device is
816	suspending (i.e., while the callback is running).
817	
818	The implementation is well suited for asynchronous use in interrupt contexts.
819	However such use inevitably involves races, because the PM core can't
820	synchronize ->runtime_suspend() callbacks with the arrival of I/O requests.
821	This synchronization must be handled by the driver, using its private lock.
822	Here is a schematic pseudo-code example:
823	
824		foo_read_or_write(struct foo_priv *foo, void *data)
825		{
826			lock(&foo->private_lock);
827			add_request_to_io_queue(foo, data);
828			if (foo->num_pending_requests++ == 0)
829				pm_runtime_get(&foo->dev);
830			if (!foo->is_suspended)
831				foo_process_next_request(foo);
832			unlock(&foo->private_lock);
833		}
834	
835		foo_io_completion(struct foo_priv *foo, void *req)
836		{
837			lock(&foo->private_lock);
838			if (--foo->num_pending_requests == 0) {
839				pm_runtime_mark_last_busy(&foo->dev);
840				pm_runtime_put_autosuspend(&foo->dev);
841			} else {
842				foo_process_next_request(foo);
843			}
844			unlock(&foo->private_lock);
845			/* Send req result back to the user ... */
846		}
847	
848		int foo_runtime_suspend(struct device *dev)
849		{
850			struct foo_priv foo = container_of(dev, ...);
851			int ret = 0;
852	
853			lock(&foo->private_lock);
854			if (foo->num_pending_requests > 0) {
855				ret = -EBUSY;
856			} else {
857				/* ... suspend the device ... */
858				foo->is_suspended = 1;
859			}
860			unlock(&foo->private_lock);
861			return ret;
862		}
863	
864		int foo_runtime_resume(struct device *dev)
865		{
866			struct foo_priv foo = container_of(dev, ...);
867	
868			lock(&foo->private_lock);
869			/* ... resume the device ... */
870			foo->is_suspended = 0;
871			pm_runtime_mark_last_busy(&foo->dev);
872			if (foo->num_pending_requests > 0)
873				foo_process_next_request(foo);
874			unlock(&foo->private_lock);
875			return 0;
876		}
877	
878	The important point is that after foo_io_completion() asks for an autosuspend,
879	the foo_runtime_suspend() callback may race with foo_read_or_write().
880	Therefore foo_runtime_suspend() has to check whether there are any pending I/O
881	requests (while holding the private lock) before allowing the suspend to
882	proceed.
883	
884	In addition, the power.autosuspend_delay field can be changed by user space at
885	any time.  If a driver cares about this, it can call
886	pm_runtime_autosuspend_expiration() from within the ->runtime_suspend()
887	callback while holding its private lock.  If the function returns a nonzero
888	value then the delay has not yet expired and the callback should return
889	-EAGAIN.
Hide Line Numbers
About Kernel Documentation Linux Kernel Contact Linux Resources Linux Blog

Information is copyright its respective author. All material is available from the Linux Kernel Source distributed under a GPL License. This page is provided as a free service by mjmwired.net.